Derivative Securities, Financial Markets, and Risk Management
Uploaded by
dr.akumar14Derivative Securities, Financial Markets, and Risk Management
Uploaded by
dr.akumar14AN INTRODUCTION TO
Derivative Securities, Financial Markets,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and Risk Management
Second Edition
ROBERT JARROW
ARKADEV CHATTERJEA
World Scientific
AN INTRODUCTION TO
Derivative Securities, Financial Markets,
and Risk Management
Second Edition
This page intentionally left blank
AN INTRODUCTION TO
Derivative Securities, Financial
Markets, and Risk Management
Second Edition
ROBERT A. JARROW
Cornell University
ARKADEV CHATTERJEA
Indiana University Bloomington
World Scientific
NEW JERSEY • LONDON • SINGAPORE • BEIJING • SHANGHAI • HONG KONG • TAIPEI • CHENNAI • TOKYO
Published by
World Scientific Publishing Co. Inc.
27 Warren Street, Suite 401-402, Hackensack, NJ 07601, USA
Head office: 5 Toh Tuck Link, Singapore 596224
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
Library of Congress Cataloging-in-Publication Data
Names: Jarrow, Robert A., author. | Chatterjea, Arkadev, author.
Title: An introduction to derivative securities, financial markets, and risk management /
Robert Jarrow (Cornell University), Arkadev Chatterjea (Indiana University).
Description: 2nd edition. | New Jersey : World Scientific, [2019] | Includes bibliographical references and index.
Identifiers: LCCN 2018022534 | ISBN 9781944659554
Subjects: LCSH: Derivative securities. | Financial institutions. | Capital market. | Risk management.
Classification: LCC HG6024.A3 J3747 2018 | DDC 332.64/57--dc23
LC record available at https://lccn.loc.gov/2018022534
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
Copyright © 2019 by Robert Jarrow and Arkadev Chatterjea
All rights reserved.
For any available supplementary material, please visit
https://www.worldscientific.com/worldscibooks/10.1142/Y0018#t=suppl
Desk Editors: Karimah Samsudin and Shreya Gopi
Typeset by diacriTech
Printed in Singapore
Bob: To my wife Gail for her patience and understanding.
Arka: To my wife Sudeshna for her cheerful and steadfast support, and to our daughters
Rushtri, Tvisha, and Roudra (all younger than the book!), who also cheerfully and
proudly supported my writing.
About the Authors
Robert A. Jarrow is the Ronald P. and Susan E. Lynch
Professor of Investment Management at the Samuel Curtis
Johnson Graduate School of Management, Cornell SC Johnson
College of Business. He is among the most distinguished finance
scholars of his generation. Jarrow has done research in nearly
all areas of derivatives pricing. He is the co-developer of two
widely used pricing models in finance, the Heath–Jarrow–
Morton (HJM) model for pricing interest-rate derivatives and the reduced form
model for pricing securities with credit risk. He is the author of more than 200
academic publications, seven books including Option Pricing (with Andrew Rudd,
1983), Modelling Fixed Income Securities and Interest Rate Options (1996), and Derivative
Securities (with Stuart Turnbull, 2000), and several edited volumes.
Arkadev Chatterjea is a Visiting Professor of Finance, Kelley
School of Business, Indiana University Bloomington. He is also
a Research Fellow at UNC Chapel Hill and a Visiting Fellow at
CHERI, Cornell University. He did his Ph.D. at Cornell, where
he was a student of Jarrow. Earlier, he was a professor of finance
at the Indian Institute of Management Calcutta. A winner of
research and teaching awards in the USA, Chatterjea has taught
derivatives at the above universities and at other institutions including CU Boulder,
the Helsinki School, Hong Kong UST, and IIM Ahmedabad. Photo by Mr. Kallol Nath.
Brief Contents
About the Authors VI 17 Single-Period Binomial Model 367
Preface to Second Edition XXVII
18 Multiperiod Binomial Model 392
Preface to First Edition XXXI
19 The Black–Scholes–Merton Model 422
PART I Introduction 20 Using the Black–Scholes–Merton
Model 462
1 Derivatives and Risk Management 2
2 Interest Rates 22 PART IV Interest Rate
3 Stocks 53 Derivatives
4 Forwards and Futures 72 21 Yields and Forward Rates 496
5 Options 92 22 Interest Rate Swaps 538
6 Arbitrage and Trading 112 23 Single-Period Binomial
7 Financial Engineering and Swaps 129 Heath–Jarrow–Morton Model 558
24 Multiperiod Binomial HJM Model 592
PART II Forwards and Futures 25 The Heath–Jarrow–Morton Libor
Model 618
8 Forwards and Futures Markets 154
26 Risk Management Models 655
9 Futures Trading 172
10 Futures Regulations 192 Appendix A: Mathematics and Statistics
11 The Cost-of-Carry Model 209 684
12 The Extended Cost-of-Carry Model Glossary 693
229 References 704
13 Futures Hedging 256 Notation 710
Additional Sources and Websites 712
PART III Options
Books on Derivatives and Risk
14 Options Markets and Trading 288 Management 715
15 Option Trading Strategies 310 Name-Index 718
16 Option Relations 336 Subject-Index 720
This page intentionally left blank
Contents
About the Authors VI
Preface to Second Edition XXVII
Preface to First Edition XXXI
PA R T I Introduction
CHAPTER 1 Derivatives and Risk Management ..................... 2
1.1 Introduction 3
1.2 Financial Innovation 5
Expanding Derivatives Markets 5
Two Economic Motives 7
1.3 Traded Derivative Securities 7
EXTENSION 1.1: The Influence of Regulations, Taxes, and
Transaction Costs on Financial Innovation 8
Diverse Views on Derivatives 9
Applications and Uses of Derivatives 10
A Quest for Better Models 12
1.4 Defining, Measuring, and Managing Risk 12
1.5 The Regulator’s Classification of Risk 12
1.6 Portfolio Risk Management 14
1.7 Corporate Financial Risk Management 14
Risks That Businesses Face 14
Nonhedged Risks 16
Risk Management in a Blue Chip Company 16
1.8 Risk Management Perspectives in This Book 17
1.9 Summary 18
1.10 Cases 19
1.11 Questions and Problems 19
IX
X CONTENTS
CHAPTER 2 Interest Rates .............................................................. 22
2.1 Introduction 23
2.2 Rate of Return 24
2.3 Basic Interest Rates: Simple, Compound, and Continuously
Compounded 25
EXTENSION 2.1: Conventions and Rules for Rounding, Reporting
Numbers, and Measuring Time 26
2.4 Discounting (PV) and Compounding (FV): Moving Money across Time 31
EXTENSION 2.2: Moving Multiple Cash Flows across Time 34
2.5 US Treasury Securities 37
2.6 US Federal Debt Auction Markets 38
2.7 Different Ways of Investing in Treasury Securities 40
The Treasury Auction and Its Associated Markets 40
The Repo and the Reverse Repo Market 41
EXTENSION 2.3: A Repurchase Agreement 41
Interest Rate Derivatives 43
2.8 Treasury Bills, Notes, Bonds, and STRIPS 43
2.9 Libor versus a Libor Rate Index 47
2.10 Summary 48
2.11 Cases 50
2.12 Questions and Problems 50
CHAPTER 3 Stocks ............................................................................. 53
3.1 Introduction 54
3.2 Primary and Secondary Markets, Exchanges, and Over-the-Counter
Markets 54
3.3 Brokers, Dealers, and Traders in Securities Markets 56
3.4 Automation of Trading 58
3.5 The Three-Step Process of Transacting Exchange-Traded Securities 58
3.6 Buying and Selling Stocks 59
Trading at the New York Stock Exchange 60
Over-the-Counter Trading 61
Alternative Trading Systems: Dark Pools and Electronic
Communications Networks 61
3.7 Dollar Dividends and Dividend Yields 62
3.8 Short Selling Stocks 64
CONTENTS XI
3.9 Margin–-Security Deposits That Facilitate Trading 66
EXTENSION 3.1: Margin and Stock Trading 67
3.10 Summary 68
3.11 Cases 70
3.12 Questions and Problems 70
CHAPTER 4 Forwards and Futures .............................................. 72
4.1 Introduction 73
4.2 Forward Contracts 73
4.3 The Over-the-Counter Market for Trading Forwards 78
4.4 Futures Contracts 81
4.5 Exchange Trading of a Futures Contract 83
EXTENSION 4.1: Futures Exchanges in China and India 86
4.6 Hedging with Forwards and Futures 88
4.7 Summary 89
4.8 Cases 90
4.9 Questions and Problems 90
CHAPTER 5 Options .......................................................................... 92
5.1 Introduction 93
5.2 Options 93
5.3 Call Options 95
Call Payoffs and Profit Diagrams 95
The Call’s Intrinsic and Time Values 98
Price Bounds for American Calls 99
5.4 Put Options 100
Put Payoffs and Profit Diagrams 100
The Put’s Intrinsic and Time Values 102
Price Bounds for American Puts 103
5.5 Exchange-Traded Options 105
5.6 Longs and Shorts in Different Markets 107
5.7 Order Placement Strategies 107
5.8 Summary 108
5.9 Cases 109
5.10 Questions and Problems 109
XII CONTENTS
CHAPTER 6 Arbitrage and Trading ........................................... 112
6.1 Introduction 113
6.2 The Concept of Arbitrage 113
6.3 The No-Arbitrage Principle for Derivative Pricing 117
The Law of One Price 117
Nothing Comes from Nothing 117
6.4 Efficient Markets 118
6.5 In Pursuit of Arbitrage Opportunities 119
The Closed-End Fund Puzzle 119
Spread Trading 120
Index Arbitrage 121
EXTENSION 6.1: Stock Indexes 121
EXTENSION 6.2: Index Arbitrage 123
6.6 Illegal Arbitrage Opportunities 124
6.7 Summary 126
6.8 Cases 127
6.9 Questions and Problems 127
CHAPTER 7 Financial Engineering and Swaps ..................... 129
7.1 Introduction 130
7.2 The Build and Break Approach 130
7.3 Financial Engineering 131
Cash Flows versus Asset Values 131
Examples 132
7.4 An Introduction to Swaps 135
7.5 Applications and Uses of Swaps 135
7.6 Types of Swaps 136
Interest Rate Swaps 136
Forex Swaps 138
Currency Swaps 139
EXTENSION 7.1: Valuing Fixed-for-Fixed Currency Swaps 142
Commodity and Equity Swaps 146
Credit Default Swaps 147
7.7 Summary 148
7.8 Cases 149
7.9 Questions and Problems 149
CONTENTS XIII
PA R T II Forwards and Futures
CHAPTER 8 Forwards and Futures Markets ....................... 154
8.1 Introduction 155
8.2 Applications and Uses of Forwards and Futures 155
8.3 A Brief History of Forwards and Futures 156
Early Trading of Forward and Futures-Type Contracts 156
US Futures Exchanges and the Evolution of the Modern Futures
Contract, 1848–1926 157
Recent Developments, 1970 Onward 159
8.4 Futures Contract Features and Price Quotes 161
Commodity and Financial Futures Contracts 161
The Gold Futures Contract 161
Gold Futures Price Quotes 164
The Exchange and Clearinghouse 166
8.5 Commodity Price Indexes 167
EXTENSION 8.1: Doomsters and Boomsters 168
8.6 Summary 169
8.7 Cases 170
8.8 Questions and Problems 170
CHAPTER 9 Futures Trading ...................................................... 172
9.1 Introduction 173
9.2 Brokers, Dealers, and the Futures Industry 173
9.3 Trading Futures 174
9.4 Margin Accounts and Daily Settlement 176
Daily Settlement 176
9.5 Futures and Forward Price Relations 180
Equality of Futures and Spot Price at Maturity 183
Equality of Forward and Futures Prices before Maturity 183
Convergence of the Basis 185
9.6 Trading Spreads 185
9.7 Summary 188
9.8 Cases 189
9.9 Questions and Problems 189
XIV CONTENTS
CHAPTER 10 Futures Regulations ............................................. 192
10.1 Introduction 193
10.2 Which Markets Have Futures? 193
Price Discovery 194
Hedging 194
Speculation 195
EXTENSION 10.1: When Sellers Have Market Power 195
10.3 Regulation of US Futures Markets 196
Major Regulatory Acts 196
The CFTC, the NFA, and the Regulatory Role of Exchanges 199
Enforcement Actions by the CFTC and the NFA 201
10.4 Manipulation in Futures Markets 202
The Hunts Silver Case 202
Salomon Brothers and US Treasury Securities 203
Banging the Close 204
10.5 Managing Commodity Markets 205
10.6 Summary 206
10.7 Cases 206
10.8 Questions and Problems 207
CHAPTER 11 The Cost-of-Carry Model .................................. 209
11.1 Introduction 210
11.2 A Cost-of-Carry Example 210
11.3 The Assumptions 214
11.4 The Cost-of-Carry Model 216
The Model Setup 216
Using the Law of One Price 217
Using Nothing Comes from Nothing 217
Different Methods for Computing Interest 218
The Arbitrage Table Approach 220
11.5 Valuing a Forward Contract at Intermediate Dates 221
11.6 Linking Forward Prices of Different Maturities 224
11.7 Summary 225
11.8 Cases 226
11.9 Questions and Problems 226
CONTENTS XV
CHAPTER 12 The Extended Cost-of-Carry Model ............. 229
12.1 Introduction 230
12.2 A Family of Forward Pricing Models 230
12.3 Forwards on Dividend-Paying Stocks 232
The Cost-of-Carry Model with Dollar Dividends 232
The Cost-of-Carry Model with a Dividend Yield 235
A Synthetic Index 236
The Foreign Currency Forward Price 239
12.4 Extended Cost-of-Carry Models 243
12.5 Backwardation, Contango, Normal Backwardation, and Normal
Contango 246
12.6 Market Imperfections 249
12.7 Summary 252
12.8 Cases 253
12.9 Questions and Problems 253
CHAPTER 13 Futures Hedging ................................................... 256
13.1 Introduction 257
13.2 To Hedge or Not to Hedge 257
Should the Firm or the Individual Hedge? 257
The Costs and Benefits of Corporate Hedging 259
EXTENSION 13.1: Airlines and Fuel Price Risk 260
EXTENSION 13.2: A Hedged Firm Capturing a Tax Loss 262
13.3 Hedging with Futures 264
Perfect and Imperfect Hedges 264
Basis Risk 264
Guidelines for Futures Hedging 267
13.4 Risk-Minimization Hedging 268
The Mean-Variance Approach 268
Limitations of Risk-Minimization Hedging 272
13.5 Futures versus Forward Hedging 272
13.6 Spreadsheet Applications: Computing h 273
13.7 Summary 275
13.8 Appendix 276
Deriving the Minimum-Variance Hedge Ratio (h) 276
Computing the Minimum-Variance Hedge Ratio (h) 277
XVI CONTENTS
Statistical Approach 277
Econometric Approach: A Linear Regression Model 281
13.9 Cases 281
13.10 Questions and Problems 282
PA R T III Options
CHAPTER 14 Options Markets and Trading ......................... 288
14.1 Introduction 289
14.2 Exchange-Traded Options 289
14.3 A History of Options 290
Early Trading of Options 291
The Year 1973: The Watershed Year and After 293
EXTENSION 14.1: Options on Futures 295
14.4 Option Contract Features 296
Maturity Dates 297
Strike Prices 297
Margin Requirements and Position Limits 298
Dividends and Stock Splits 299
14.5 Options Trading, Exercising, and the Expiration Process 300
The Trading Process 300
Exercising the Option 300
The Expiration Process 301
14.6 Options Price Quotes 301
14.7 Regulation and Manipulation in Options Markets 305
14.8 Summary 306
14.9 Cases 307
14.10 Questions and Problems 307
CHAPTER 15 Option Trading Strategies ................................ 310
15.1 Introduction 311
15.2 Traders in Options Markets 311
15.3 Profit Diagrams 312
15.4 Options Strategies 313
CONTENTS XVII
Common Options Price Data 313
Naked Trades: The Basic Building Blocks 314
15.5 Hedged Strategies 317
15.6 Spread Strategies 320
EXTENSION 15.1: Collars, Ratio Spreads, Butterflies, and Condors 323
EXTENSION 15.2: Reinsurance Contracts and Call Spreads 327
15.7 Combination Strategies 329
15.8 Summary 331
15.9 Cases 332
15.10 Questions and Problems 332
CHAPTER 16 Option Relations ................................................... 336
16.1 Introduction 337
16.2 A Graphical Approach to Put–Call Parity 337
16.3 Put–Call Parity for European Options 338
16.4 Market Imperfections 342
EXTENSION 16.1: Put–Call Parity in Imperfect Markets 343
EXTENSION 16.2: Dividends and American Options 344
16.5 Options Price Restrictions 346
EXTENSION 16.3: The Superglue Argument 352
16.6 Early Exercise of American Options 357
No Dividends 357
Dividends and Early Exercise 360
16.7 Summary 362
16.8 Cases 364
16.9 Questions and Problems 364
CHAPTER 17 Single-Period Binomial Model ........................ 367
17.1 Introduction 368
17.2 Applications and Uses of the Binomial Model 368
17.3 A Brief History of Options Pricing Models 369
Option Pricing Pre-1973 369
Options Pricing, 1973 and After 370
17.4 An Example 372
XVIII CONTENTS
17.5 The Assumptions 374
17.6 The Single-Period Model 376
The No-Arbitrage Principle 376
Building Binomial Trees 377
Arbitrage-Free Trees 378
Stock Prices and Martingales 378
The Pricing Model 379
The Hedge Ratio 380
Risk-Neutral Valuation 381
Actual versus Pseudo-probabilities 382
17.7 Summary 383
17.8 Appendix 384
Proving the No-Arbitrage Argument 384
The Probabilities and Risk Premium 385
17.9 Cases 388
17.10 Questions and Problems 388
CHAPTER 18 Multiperiod Binomial Model ............................ 392
18.1 Introduction 393
18.2 Toward a Multiperiod Binomial Option Pricing Model 393
The Stock Price Evolution 393
Binomial Option Price Data 394
The Stock Price Tree 395
18.3 A Two-Period Binomial Model 396
Backward Induction 396
Option Pricing via Synthetic Construction (Method 1) 397
Repeat, Repeat: Risk-Neutral Pricing (Method 2) 400
One-Step Valuation: Prelude to the Multiperiod Model (Method 3) 401
18.4 The Multiperiod Binomial Option Pricing Model 403
Binomial Coefficients and Pseudo-probabilities 403
Recasting the Two-Period Example in the Multiperiod Framework 403
The n-Period Binomial Option Pricing Model 405
EXTENSION 18.1: Linking the Binomial Model to the
Black–Scholes–Merton Model 407
18.5 Extending the Binomial Model 409
Known Dividends 409
Valuing American Options 410
CONTENTS XIX
18.6 Spreadsheet Applications 413
A Two-Period Binomial Example 413
A Sixteen-Period Example 416
18.7 Summary 418
18.8 Cases 419
18.9 Questions and Problems 419
CHAPTER 19 The Black–Scholes–Merton Model ............... 422
19.1 Introduction 423
19.2 Nobel Prize–Winning Works (1973) 423
19.3 The Assumptions 424
EXTENSION 19.1: Bubbles and Option Pricing 424
19.4 The Pricing and Hedging Argument 428
19.5 The Black–Scholes–Merton Formula 429
19.6 Understanding the Black–Scholes–Merton Model 432
Stock Prices and Martingales 432
Risk-Neutral Valuation 433
Actual versus Pseudo-probabilities 433
19.7 The Greeks 434
Interpreting the Greeks 434
Some Road Bumps Ahead 439
EXTENSION 19.2: Market Manipulation and Option Pricing 439
19.8 The Inputs 440
Observable Inputs 441
Volatility: The Elusive Input 441
19.9 Extending the Black–Scholes–Merton Model 444
Adjusting for Dividends 444
Foreign Currency Options 446
Valuing American Options 447
EXTENSION 19.3: Exotic Options 448
19.10 Summary 449
19.11 Appendix 451
Modeling the Stock Price Evolution 451
Continuously Compounded (Logarithmic) Return 451
Introducing Uncertainty 452
The Central Limit Theorem 453
XX CONTENTS
First Derivation of the Black–Scholes–Merton Formula (as a Limit of the
Binomial Model) 454
Second Derivation of the Black–Scholes–Merton Formula (Using
Risk-Neutral Valuation) 457
The Probabilities and the Risk Premium 457
Proof of (24) 458
19.12 Cases 459
19.13 Questions and Problems 459
CHAPTER 20 Using the Black–Scholes–Merton Model ... 462
20.1 Introduction 463
20.2 Hedging the Greeks 463
Delta Hedging 463
Gamma Hedging 466
20.3 Hedging a portfolio of options 469
20.4 Vega Hedging 470
EXTENSION 20.1: Stochastic Volatility Option Models 473
20.5 Calibration 474
Theoretical and Econometric Models 474
Using Calibration 475
Implied Volatilities 477
EXTENSION 20.2: VIX: The Fear Index 480
20.6 The Black–Scholes–Merton Model: A Postscript 480
20.7 Summary 481
20.8 Appendix 482
The Mathematics of Delta, Gamma, and Vega Hedging 482
A Third Derivation of the Black–Scholes–Merton Formula 486
20.9 Cases 487
20.10 Questions and Problems 488
Overview 490
Data (from the Wall Street Journal, March 3, 2000) 490
The Project 491
CONTENTS XXI
PA R T IV Interest Rate Derivatives
CHAPTER 21 Yields and Forward Rates .................................. 496
21.1 Introduction 497
EXTENSION 21.1: Home Ownership and Derivatives 498
21.2 Yields 499
Revisiting Zeros 499
Yields and the Yield Curve 500
EXTENSION 21.2: The Expectations Hypothesis 504
21.3 The Traditional Approach 507
Duration 507
Modified Duration Hedging 510
Applications and Limitations 512
21.4 Forward Rates 513
The Definition 514
Understanding Forward Rates 516
Using Forward Rates 517
EXTENSION 21.3: Computing Forward Rates from Coupon Bond
Prices 519
21.5 The Basic Interest Rate Derivatives Contracts 520
A Brief History 520
Forward Rate Agreements 521
Interest Rate Futures 524
EXTENSION 21.4: Treasury Futures 529
The Equivalence between Forward and FRA Rates 530
21.6 Summary 531
21.7 Cases 534
21.8 Questions and Problems 535
CHAPTER 22 Interest Rate Swaps ............................................. 538
22.1 Introduction 539
22.2 A Brief History 539
The Introduction of Swap Contracts 540
From a Brokerage to a Dealership Market 541
ISDA and Standardization of Contracts 541
XXII CONTENTS
22.3 Institutional Features 542
Entering a Swap Contract 542
Documentation 543
Closing a Swap Position 544
22.4 Valuation 544
22.5 Variations of Interest Rate Swaps 549
22.6 Swaps and FRAs 550
Synthesizing Swaps with Eurodollars and FRAs 550
The Yield and Swap Curve 552
EXTENSION 22.1: Computing Forward Rates from Swap Rates 553
22.7 Summary 555
22.8 Cases 555
22.9 Questions and Problems 555
CHAPTER 23 Single-Period Binomial
Heath–Jarrow–Morton Model ......................... 558
23.1 Introduction 559
23.2 The Basic Interest Rate Derivatives 560
Uses of Caps, Floors, and Collars 561
23.3 A Brief History of Interest Rate Derivatives Models 564
23.4 The Assumptions 565
23.5 The Single-Period Model 570
Arbitrage-Free Evolutions 570
Zero-Coupon Bond Prices and Martingales 573
Understanding the Equal Pseudo-probability Condition 575
Actual versus Pseudo-probabilities 576
Caplet Pricing 576
The Hedge Ratio (the Holy Grail) 581
Risk-Neutral Valuation 581
Valuing a Floorlet 582
Valuing Various Interest Rate Derivatives 584
Multiple Factors 584
23.6 Summary 585
23.7 Appendix 586
The Equal Pseudo-probability Condition 586
Proof of Sufficiency 587
Proof of Caplet and Floorlet Parity 588
CONTENTS XXIII
23.8 Cases 589
23.9 Questions and Problems 589
CHAPTER 24 Multiperiod Binomial HJM Model ................. 592
24.1 Introduction 593
24.2 The Assumptions 593
24.3 The Two-Period Model 597
Arbitrage-Free Evolutions 597
Zero-Coupon Bond Prices and Martingales 599
Caplet Pricing 601
Floorlet Pricing 605
Valuing Various Interest Rate Derivatives 606
Multiple Factors 606
24.4 The Multiperiod Model 607
24.5 Forwards, Futures, and Swaptions 607
Forward Rate Agreements 607
Eurodollar Futures 608
Comparing Forward and Futures Rates 610
Swaptions 611
24.6 Summary 612
24.7 Appendix 613
Linking the Binomial HJM Model and the HJM Libor Model 613
24.8 Cases 615
24.9 Questions and Problems 616
CHAPTER 25 The Heath–Jarrow–Morton Libor Model .... 618
25.1 Introduction 619
25.2 Why Caplets? 619
25.3 A History of Caplet Pricing 520
Black’s Model 620
The Heath–Jarrow–Morton Model 621
The HJM Libor Model 621
EXTENSION 25.1: Black’s Model for Pricing Commodity Options and
Caplets 621
25.4 The Assumptions 622
XXIV CONTENTS
25.5 Pricing Caplets 627
Caplet and Floorlet Pricing (CFP) Data 627
Caplet Pricing Model 627
25.6 The Inputs 631
Observable Inputs 631
The Average Forward Rate Volatility 632
25.7 Understanding the HJM Libor Model 637
Forward Rates and Martingales 637
Risk-Neutral Valuation 637
Actual versus Pseudo-probabilities 638
25.8 Caps and Floors 638
Caps 639
Floorlets 639
Caplet–Floorlet Parity 639
Floors 641
25.9 Using the HJM Libor Model 641
The Greeks 642
Delta Hedging 643
Gamma Hedging 646
25.10 American Options, Futures, Swaptions, and Other Derivatives 647
25.11 Summary 647
25.12 Appendix 649
25.13 Cases 652
25.14 Questions and Problems 652
CHAPTER 26 Risk Management Models ................................ 655
26.1 Introduction 656
26.2 A Framework for Financial Risk Management 658
26.3 Computing the Loss Distribution 660
26.4 Value-at-Risk and Scenario Analysis 662
Value-at-Risk 662
Scenario Analysis 666
EXTENSION 26.1: Risk Measures 667
26.5 The Four Risks 668
Market Risk 668
Credit Risk 669
Structural Models 670
CONTENTS XXV
EXTENSION 26.2: Real Options 672
Reduced-Form Models 673
EXTENSION 26.3: Credit Default Swaps 676
Liquidity Risk 677
Operational Risk 678
26.6 The Future of Models and Traded Derivatives 679
Model Risk 679
Derivatives 680
26.7 Summary 680
26.8 Cases 681
26.9 Questions and Problems 681
APPE N DI X A Mathematics and Statistics ............................... 684
A.1 Exponents and Logarithms 685
Exponents 685
Logarithms 685
Continuous Compounding 687
A.2 The Binomial Theorem 687
A.3 The Normal Distribution 688
A.4 A Taylor Series Expansion 691
Single Variables 691
Multiple Variables 692
Glossary 693
References 704
Notation 710
Additional Sources and Websites 712
Books on Derivatives and Risk Management 715
Name-Index 718
Subject-Index 720
This page intentionally left blank
Preface to Second Edition
We are pleased by the global reception of the first edition. The book has seen
classroom usage in over 100 courses in 75 universities in 22 countries. It has been
used widely as: (a) an introductory textbook for undergraduates (in business schools,
in Arts & Sciences, as well as other programs), (b) an introductory text for MBAs,
MS, and other masters students, (c) a recommended text in derivatives courses, and
(d) a reference text in specialized courses.
Changes in the Second Edition
Based on feedback from colleagues, students, and our own teaching experience
we focused on updating institutional facts, reducing the historical and institutional
material, and cleaning typos.
A Solutions Manual for end-of-chapter Questions and Problems remains avail-
able for lecturers, together with PowerPoint slides and a Test Bank. These ancil-
laries can be downloaded from https://www.worldscientific.com/worldscibooks/
10.1142/y0018-sm
Although our free-software PRICED! still works in most computers, we find
that many professors and students are more comfortable developing their own
pricing models through Excel (we also make our Excel files available to adopters).
Consequently, we continue to make PRICED! available but we do not provide
any support. Lecturers and students who wish to download PRICED! can visit
https://www.worldscientific.com/worldscibooks/10.1142/y0018 and click on the
“Supplementary” tab.
Acknowledgements
We owe many thanks to colleagues and students for their help, support, and
suggestions. We especially thank two colleagues for their extraordinary help and
support: Scott Fung (California State University, East Bay) and Thijs van der
Heijden (University of Melbourne). We also thank Binay Bhushan Chakrabarti,
Prem Chandrani, Ronald Ehrenberg, Craig Holden, Sreenivas Kamma, Ravindra
Patil, and Zhenyu Wang for their comments, help, and support. We thank Yu Yan
who translated our book into Chinese.
XXVII
XXVIII PREFACE TO SECOND EDITION
We thank the following colleagues who, our records show, have adopted the book
for classroom use as a required or recommended reading:
Sofiane Aboura, Université de Paris XIII, Sorbonne Paris Cité;
̆
Vedat Akgiray, Bogaziçi Üniversitesi;
Ali Akyol, University of Melbourne;
Joao Freire de Andrade, Universidade NOVA de Lisboa;
Kostas Andriosopoulos, ESCP Europe, London;
Kevin Aretz, University of Manchester;
C Bülent Aybar, Southern New Hampshire University;
Giovanni Barone-Adesi, Università della Svizzera italiana at Lugano;
Basabi Bhattacharya, Jadavpur University;
Nalinaksha Bhattacharyya, University of Alaska Anchorage;
Jedrzej Bialkowski, University of Canterbury;
Ben Chongyang Chen, University of Memphis;
Hsin-Fu Chen, Lunghwa University of Science & Technology;
William Cheung, Waseda University;
Paul Moon Sub Choi, Ewha Womans University;
Steven P Clark, University of North Carolina at Charlotte;
Francisco José Climent Diranzo, Universitat de València;
Vladislav Damjanovic, University of Essex;
Theodoros Diasakos, University of Stirling;
Sina Erdal, University of Essex;
Shih-Chuan Fang, Chaoyang University of Technology;
Scott Fung, California State University East Bay;
Gianluca Fusai, City, University London;
Neal Galpin, University of Melbourne;
George P Gao, T Rowe Price;
Rajesh Ghai, University of Texas at Dallas;
Chinmoy Ghosh, University of Connecticut;
Bruce Grundy, University of Melbourne;
Anthony Hall, University of Technology, Sydney;
Joongho Han, Sungkyunkwan University;
Barry Harrison, Nottingham Trent University;
Mia Hinnerich, Stockholms Universitet;
Darien Huang, Cornell University;
Nagaratnam Jeyasreedharan, University of Tasmania;
Timothy C Johnson, Heriot-Watt University;
Jeffrey Jolley, California State University Fullerton;
Andreas Kaeck, University of Sussex;
Hui-Chuan Kao, Ling Tung University;
Cenk Cevat Karahan, Bogaziçi ̆ Üniversitesi;
Roger Klee, Cleveland State University;
Ioannis Kyriakou, City, University London;
Pascal Letourneau, University of Wisconsin-Whitewater;
Hung-Chi Li, National Cheng Kung University;
PREFACE TO SECOND EDITION XXIX
Hai Lin, Victoria University of Wellington;
Ravi Mateti, Concordia University;
João Amaro de Matos, Universidade NOVA de Lisboa;
Hongghi Min, Korea Advanced Institute of Science and Technology;
Nikos Nomikos, City, University London;
Kurtay Ogunc, Texas A & M University-Commerce;
Jacques Olivier, HEC Paris;
Late Philip Palm, who was at University of Washington Bothell;
Ging-Ginq Pan, National Pingtung University of Science and Technology;
Christophe Perignon, HEC Paris;
Ser-Huang Poon, University of Manchester;
Marcel Prokopczuk, Zeppelin University;
Paulo J M Rodrigues, Universiteit Maastricht;
Juan Carlos Rodriguez, Tilburg University;
̆
Mehmet Saglam, University of Cincinnati;
Rafael de Santiago, Universidad de Navarra;
Ravi Sastry, Southern Methodist University;
Norman Seeger, Vrije Universiteit Amsterdam;
Elena Sernova, HEC Lausanne;
Kaisheng Song, University of Texas at Dallas;
Erling Steigum, BI Norwegian Business School;
Liya Shen, University of Essex;
Manouchehr Tavakoli, University of St Andrews;
Siegfried Trautmann, Johannes Gutenberg Universitat Mainz;
Jiao Tong, Robert Gordon University;
Hui-Huang Tsai, National United University;
Jen-Tsung Tsu, Chien Hsin University of Science and Technology;
Calum Turvey, Cornell University;
Eleuterio Vallelado Gonzalez, Universidad de Valladolid;
Thijs van der Heijden, University of Melbourne;
Dmitry Vedenov, Texas A & M University;
Lee Wakeman, Ohio University;
Nathan Walcott, Southern Methodist University;
Costas Xiouros, BI Norwegian Business School;
Lulu Zeng, Purdue University;
J Feng Zhao, University of Texas at Dallas;
Lin Zou, Texas Woman’s University.
We thank our editor Jack Repcheck of W. W. Norton & Company and his
colleagues Dorothy Cook, Edward Crutchley, Theresia Kowara, Lindsey Osteen,
Lauren Quantrill, Eric Svendsen, Janise Turso, Dina Vakser, and others who have
faciliated the global acceptance that the book enjoys today. After Jack’s sudden and
sad death, we moved to World Scientific Publishing. Our special thanks to our new
editor Yubing Zhai and her colleagues Shreya Gopi, Karimah Samsudin and others
who have made possible a smooth transition and a fine production of the second
edition.
This page intentionally left blank
Preface to First Edition
“History has many cunning passages, contrived corridors,” wrote T. S. Eliot in
the poem “Gerontion.” The history of options and futures can be so described.
Derivatives have traded for centuries in small, over-the-counter markets in London,
New York, and several cities of continental Europe. In 1688, a sophisticated rice
futures market was established in Osaka, Japan, that thrived for 250 years. Modern
commodity futures markets began in 1848 with the founding of the Chicago Board
of Trade to help protect farmers from commodity price swings. After many placid
decades, the calm was again broken with the inauguration of the world’s first financial
futures market in 1972 at the Chicago Mercantile Exchange and in 1973 with the
opening of the Chicago Board Options Exchange. Serendipitously, 1973 also saw
the publication of the Nobel Prize–winning Black–Scholes–Merton (BSM) option
pricing model, which further spurred market expansion by enabling better pricing
and hedging.
Sometimes viewed as harmful and sometimes viewed as beneficial by the financial
press and the public, derivatives have nonetheless played a significant role in financial
markets across the centuries. Following the 2007 credit crisis, derivatives are most
recently playing the role of the harmful security again, and new financial regulations
have been proposed to rein them in. Although perhaps well intentioned, these
attacks on derivatives are hurled by those unfamiliar with their proper use. Derivative
securities, if used properly, reduce risk and facilitate real economic growth in the
economy. In today’s complex world, modern financial institutions cannot succeed
without the use of derivatives for managing the varied risks of their assets and
liabilities. An understanding of their proper usage comes through a careful study of
derivatives, which brings us to the purpose for writing this book.
This book was written to be the first book read on derivatives and not the last. Our aim has
been to design a book that is closely connected to real markets, examines the uses of derivatives
but warns against their abuses, and presents only the necessary quantitative material in an easily
digestible form—and no more!
Given our purpose, this book differs from all existing derivatives textbooks in
several important ways:
■ First, it is an introduction. We wanted to create a textbook accessible to MBAs
and undergraduates both in terms of the concepts and mathematics. Option
pricing is normally thought of as a complex, mathematical, and difficult topic.
Our experience is that this topic can be presented simply and intuitively.
XXXI
XXXII PREFACE TO FIRST EDITION
■ Second, it is about financial markets. We wanted to write an economics, not a
quantitative, book on derivatives. Since the first textbooks on option pricing,1
derivatives textbooks have taken a quantitative approach to the topic and are often
encyclopedic in presentation. Little if any effort was spent on the underlying
economics. In contrast, an economics perspective relates the market structure to
the assumptions underlying the models. An understanding of when to use and
when not to use a model based on its assumptions must be included. Our book
does this.
■ Third, it is a book about risk management. Generally speaking, there are four risks
to be managed: (1) market risk, which includes commodity (including equity)
price risk, interest rate risk, and foreign currency price risk; (2) credit risk;
(3) liquidity risk; and (4) operational risk. We walk you through these different
risks, with an emphasis on market risk, and discuss how they can be managed in
business as well as in one’s personal life.
■ Fourth, not only do governments regulate the markets, but many government
entities use derivatives to promote the public’s welfare. The book often discusses
the relevant issues from a regulator’s point of view and from a public policy
perspective.
■ Fifth, another unique feature of this book is an intuitive and accessible presentation
of the Heath–Jarrow–Morton (HJM) model, which is the most advanced as well
as a widely used derivatives pricing model. To make this model accessible, we first
present the classical option pricing theory centering on the BSM model in a user-
friendly fashion. Our presentation of the BSM model, however, is done with an
eye toward the HJM model, emphasizing those aspects of the BSM model that are
needed later. Then, when we study the HJM model, less development is needed.
This approach enables us to present the HJM model in a parallel fashion to that of
the BSM model, so if you see one, you see them both!
■ Sixth, option pricing has a fascinating history, filled with colorful people and
events. We share this history with the reader. This history is obtained from
century-old books (now, uniquely available via the Internet), recent books,
newspaper and magazine articles, websites, and personal experiences.
■ Seventh, we have included enrichment material for the advanced reader through
the use of inserts and appendixes. Many of these inserts include current research
insights not available in existing textbooks.
The organization of this book is purposely designed to facilitate its use in many
courses relating to options, futures, derivatives, risk management, investments, fixed
income securities, and financial institutions. Three major courses that can be taught
from this book are: (1) derivatives, (2) futures and commodities, and (3) fixed income
securities and interest rate derivatives. In a sense, there are many books within this
one cover.
1
Jarrow and Rudd (1983) and Cox and Rubinstein (1984).
PREFACE TO FIRST EDITION XXXIII
■ Derivatives. This is a standard course on basic derivatives that gives an
introduction to forwards, futures, options, and swaps but excludes interest rate
derivatives for the most part. For this, use chapters 1–20 and most of Chapter
22. This would be for first-year MBAs and masters of financial engineering
and for upper-level undergraduates in business schools, engineering schools, and
economics departments.
■ Futures and commodities. This is a course on futures and commodities,
excluding option pricing. For this, use Chapters 1–6 and 8–13. The target audience
is as for the prior item.
■ Fixed-income securities and interest rate derivatives. This would be a
models-based approach to teaching this material. Integrating ongoing research
developments, Jarrow has been teaching such a course at Cornell University for
over twenty years. For this, use Chapters 1, 2, 4, 6, 9, and 15–26. This course can
be taught to MBAs and more mathematically inclined upper level undergraduates.
■ Case-based courses. We have also recommended cases at the end of each chapter
so that instructors can easily develop a case study–oriented derivatives course.
■ Spreadsheet-based courses. Unlike other textbooks, we do not provide black
boxes in which you input data and get a derivative price! In line with current
teaching trends, we have woven spreadsheet applications throughout the text. Our
aim is for students to achieve self-sufficiency so that they can generate all the
models and graphs in this book via Excel. In addition, spreadsheet software—called
Priced!—especially designed for this textbook is available to facilitate learning and
to teach the course material.
■ Use in courses in other areas. Courses in accounting and law can use this as
background material.
This book is a “download” of our understandings of markets and derivatives
obtained from decades of research, teaching, and consulting. The material has
been class tested at the Cornell University, Helsinki School of Economics and
Business Administration, Hong Kong University of Science and Technology, the
Indian Institutes of Management at Ahmedabad and at Calcutta, Indiana University
(Bloomington), University of Colorado at Boulder, and the University of North
Carolina at Chapel Hill. We hope that the reader will have a better understanding of
derivative securities and risk management models after reading this book. Although
we have tried to make this textbook error free, please notify us if you discover any
errors.
The Ancillaries
This text is accompanied by a number of important ancillaries, each intended to
enhance the learning experience for the student and the teaching experience for the
instructor.
XXXIV PREFACE TO FIRST EDITION
For Students and Instructors
PRICED! Developed by a Cornell computer science Ph.D. Tibor Jánosi specifically
for this text and in collaboration with book authors Robert Jarrow and Arkadev
Chatterjea, Priced! is Excel-based software that makes the models real rather than
hypothetical.
The software computes prices and hedge ratios for the four key derivative security
models contained in the book: (1) the binomial model, (2) the Black–Scholes–
Merton (BSM) model, (3) the discrete time Heath–Jarrow–Morton (HJM) model,
and (4) the HJM Libor model.
The prices and hedge ratios are represented visually—both graphically and in
trees—for easy analysis by students and instructors. For quick recognition and clarity,
the software’s inputs and outputs are color-coded. Priced! is also completely dynamic;
when changes are made to inputs, outputs are updated instantaneously.
This software can be used to illustrate all of the key concepts associated with the
derivative models discussed in the course. Instructors can illustrate how a model’s
prices and hedge ratios change when inputs are changed. Students can even use
Priced! and current market prices from the financial press to compute actual prices.
For Students
SOLUTIONS MANUAL Written entirely by the text’s authors, Robert Jarrow and
Arkadev Chatterjea, the solutions manual provides completely worked solutions for
all the problems included with the book. ISBN: 978-0-393-92094-9
For Instructors
POWERPOINTS Created by coauthor Arkadev Chatterjea. The slides include
lecture slides and all art from the book. There is also a separate set of PowerPoints
created by Robert Jarrow for a fixed income course based on selected chapters in the
book. Downloadable from wwnorton.com/instructors.
TEST BANK Written by the text’s authors, Robert Jarrow and Arkadev Chatterjea.
Downloadable formats available on wwnorton.com/instructors.
■ PDF
■ Word
®
■ ExamView Assessment Suite
Acknowledgments
Bob thanks Peter Carr, Dilip Madan, Philip Protter, Siegfried Trautmann, Stuart
Turnbull, and Don van Deventer for many conversations about derivatives over the
years.
For the many helpful discussions, comments, and material that helped improve
this book, Bob and Arka would both like to thank Jeffery Abarbanell, Sobhesh K.
Agarwalla, Jiyoun An, Warren B. Bailey, Gregory Besharov, Sommarat Chantarat,
PREFACE TO FIRST EDITION XXXV
Surjamukhi Chatterjea, Soikot Chatterjee, Sudheer Chava, Paul Moon Sub Choi,
Steve Choi, Hazem Daouk, Werner Freystätter, Suman Ganguli, Nilanjan Ghosh,
Michael A. Goldstein, Jason Harlow, Philip Ho, Michael F. Imhoff, Keon Hee Kim,
Robert C. Klemkosky, Junghan Koo, Hao Li, Banikanta Mishra, Debi P. Mohapatra,
Gillian Mulley, B. V. Phani, George Robinson, Ambar Sengupta, Asha Ram Sihag,
Yusuke Tateno, and Han Zheng.
We thank our family and friends for supporting us during this long project. In
particular, Kaushik Basu, Nathaniel S. Behura, Amitava Bose, Alok Chakrabarti,
Jennifer Conrad, Judson Devall, Ram Sewak Dubey, David Easley, Diego Garcia,
Robert H. Jennings, Surendra Mansinghka, Robert T. Masson, Indranil Maulik,
Tapan Mitra, Peter D. McClelland, Uri M. Possen, Erik Thorbecke, and Daniel J.
Wszolek.
We were fortunate to have a number of highly conscientious formal review-
ers of the manuscript as it was taking shape. Their astute advice had a major
impact on the final version of the manuscript. We cannot thank these review-
ers enough: Farid AitSahlia (University of Florida), Gregory W. Brown (The
University of North Carolina at Chapel Hill), Michael Ferguson (University of
Cincinnati), Stephen Figlewski (New York University), Scott Fung (California State
University, East Bay), Richard Rendleman (Dartmouth College), Nejat Seyhun
(University of Michigan), Joel Vanden (The Pennsylvania State University), Kelly
Welch (The University of Kansas), and Youchang Wu (University of Wisconsin–
Madison).
Finally, but certainly not least, special thanks to the editorial staff at Norton
who have helped turn our manuscript into the finished book you are now reading:
Hannah Bachman, Cassie del Pilar, Jack Repcheck, Nicole Sawa, and Amy Wein-
traub.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
I
Introduction
CHAPTER 1
Derivatives and Risk
Management
CHAPTER 2
Interest Rates
CHAPTER 4
CHAPTER 3
Forwards and
Stocks
Futures
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
CHAPTER 6
CHAPTER 5
Arbitrage and
Options
Trading
CHAPTER 7
Financial Engineering
and Swaps
1
Derivatives and Risk
Management
1.1 Introduction 1.5 The Regulator’s Classification
of Risk
1.2 Financial Innovation
1.6 Portfolio Risk Management
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Expanding Derivatives Markets
Two Economic Motives 1.7 Corporate Financial Risk
1.3 Traded Derivative Securities Management
EXTENSION 1.1 The Influence Risks That Businesses Face
of Regulations, Taxes, and Nonhedged Risks
Transaction Costs on Financial Risk Management in a Blue Chip
Innovation Company
Diverse Views on Derivatives
1.8 Risk Management
Applications and Uses of Perspectives in This Book
Derivatives
1.9 Summary
A Quest for Better Models
1.10 Cases
1.4 Defining, Measuring, and
Managing Risk 1.11 Questions and Problems
INTRODUCTION 3
1.1 Introduction
The bursting of the housing price bubble, the credit crisis of 2007, the resulting losses
of hundreds of billions of dollars on credit derivatives, and the failure of prominent
financial institutions have forever changed the way the world views derivatives. Today
derivatives are of interest not only to Wall Street but also to Main Street. Derivatives
are cursed as one of the causes of the Great Recession of 2007–2009, a period of
decreased economic output and high unemployment.
But what are derivatives? A derivative security or a derivative is a financial
contract that derives its value from an underlying asset’s price, such as a stock or
a commodity, or even from an underlying financial index like an interest rate. A
derivative can both reduce risk, by providing insurance (which, in financial parlance,
is referred to as hedging), and magnify risk, by speculating on future events. Derivatives
provide unique and different ways of investing and managing wealth that ordinary
securities do not.
Derivatives have a long and checkered past. In the 1960s, only a handful of
individuals studied derivatives. No academic books covered the topic, and no college
or university courses were available. Derivatives markets were small, located mostly
in the US and Western Europe. Derivative users included only a limited number
of traders in futures markets and on Wall Street. The options market existed as
trading between professional traders (called the over-the-counter [OTC] market) with
little activity. In addition, cheating charges often gave the options market disrepute.
Derivatives discussion did not add sparkle to cocktail conversations, nor did it generate
the allegations and condemnations that it does today. Brash young derivatives traders
who drive exotic cars and move millions of dollars with the touch of a computer key
didn’t exist. Although Einstein had developed the theory of relativity and astronauts
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
had landed on the moon, no one knew how to price an option. That’s because in
the 1960s, nobody cared, and derivatives were unimportant.
What a difference the following decades have made! Beginning in the early 1970s,
derivatives have undergone explosive growth in the types of contracts traded and
in their importance to the financial and real economy. According to the Bank
for International Settlements (known as the BIS), the markets are now global and
measured in trillions of dollars. Indeed, as depicted in Figure 1.1, in December 2016
the total outstanding US dollar notional value for exchange-traded derivatives was
(26,172 + 41,072 =) 67,245 billion and for OTC derivatives a staggering 500,419
billion. Hundreds of academics study derivatives, and thousands of articles have been
written on the topic of pricing derivatives. Colleges and universities now offer
numerous derivatives courses using textbooks written on the subject. Derivatives
experts are in great demand. In fact, Wall Street firms hire PhDs in mathematics,
engineering, and the natural sciences to understand derivatives—these folks are
admirably called “rocket scientists” (“quants” is another name). If you understand
derivatives, then you know cool stuff; you are hot and possibly dangerous. Today
understanding derivatives is an integral part of the knowledge needed in the risk
management of financial institutions.
4 CHAPTER 1: DERIVATIVES AND RISK MANAGEMENT
FIGURE 1.1: Global Derivatives Market
Derivatives Notional Amounts
(December 2016, in billions of US dollars)
10015 27864
96 26172
1350
6140
41073
68598
386356
Futures (Exchange-Traded) Options (Exchange-Traded)
Foreign Exchange Contracts (Global OTC) Interest Rate Contracts (Global OTC)
Equity-Linked Contracts (Global OTC) Commodity Contracts (Global OTC)
Credit Derivatives (Global OTC) Other Derivatives (Global OTC)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Unallocated (Global OTC)
Source: https://www.bis.org/statistics/extderiv.htm, Table D1: Exchange-traded futures and options, by
location of exchange; and https://www.bis.org/statistics/derstats.htm, Table D5: Global OTC derivatives
market.
Markets have changed to accommodate derivatives trading in three related ways:
the introduction of new contracts and new exchanges, the consolidation and linking
of exchanges, and the introduction of computer technology. Sometimes these
changes happened with astounding quickness. For example, when twelve European
nations replaced their currencies with the euro in 2002, financial markets for euro-
denominated interest rate derivatives sprang up almost overnight, and in some
cases, they quickly overtook the dollar-denominated market for similar interest rate
derivatives.
This chapter tells the fascinating story of this expansion in derivatives trading and
the controversy surrounding its growth. An understanding of the meaning of financial
risk is essential in fully understanding this story. Hence a discussion of financial
risk comes next, from the regulator’s, the portfolio manager’s, and the corporate
financial manager’s points of view. We explain each of these unique perspectives,
FINANCIAL INNOVATION 5
using them throughout the book to increase our understanding of the uses and abuses
of derivatives. A summary completes the chapter.
1.2 Financial Innovation
Derivatives are at the core of financial innovation, for better or for worse. They are
the innovations to which columnist David Wessel’s Wall Street Journal article titled “A
Source of Our Bubble Trouble,” dated January 17, 2008, alludes:
Modern finance is, truly, as powerful and innovative as modern science. More people
own homes—many of them still making their mortgage payments—because mortgages
were turned into securities sold around the globe. More workers enjoy stable jobs
because finance shields their employers from the ups and downs of commodity prices.
More genius inventors see dreams realized because of venture capital. More consumers
get better, cheaper insurance or fatter retirement checks because of Wall Street wizardry.
Expressed at a time when most of the world was in the Great Recession, this view is
challenged by those who blame derivatives for the crisis. Indeed, this article goes on to
say that “tens of billions of dollars of losses in new-fangled investments [in derivatives
and other complex securities] at the largest US financial institutions—and the belated
realization that some of those Ph.D. wielding, computer-enhanced geniuses were
overconfident in the extreme—strongly suggests some of the brainpower drawn to
Wall Street would have been more productively employed elsewhere in the economy.”
But derivatives have been trading in various guises for over two thousand years.
They have continued to trade because, when used properly, they enable market
participants to reduce risk from their portfolios and to earn financial rewards from
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
trading on special skills and information. Indeed, derivatives help to advance or
postpone cash flows (borrowing and lending), to accumulate wealth (saving),
to protect against unfavorable outcomes (insurance or hedging), to commit funds
to earn a financial return (investment), and to accept high risks in the hope of big
returns (speculation or gambling), which often goes along with magnifying the
scale of one’s financial returns (leverage). Financial markets grow and real economic
activity prospers because derivatives make financial markets more efficient. This is a
theme to which we return repeatedly throughout the book.
Expanding Derivatives Markets
Many factors have fueled the growth of derivatives markets. These include regulatory
reforms, an increase in international commerce, population growth, political changes,
the integration of the world’s economy, and revolutionary strides in information
technology (IT). The interrelated financial markets are now more susceptible to
global shocks and financial crises. The financial world has become a mad, bad, and
dangerous place—financially speaking! More pronounced business cycles, default by
sovereign nations, high-risk leveraged bets by hedge funds, imprudent investment
in complicated securities by unsophisticated investors, and fraudulent actions by
rogue traders have the potential to shake financial markets to their core. Financial
6 CHAPTER 1: DERIVATIVES AND RISK MANAGEMENT
regulators exist to help prevent these catastrophes from happening. And if used
properly, derivatives can also help to mitigate their effects on aggregate wealth.
To help achieve economic stability, the central bank of the US, the Federal
Reserve System (often referred to as the Federal Reserve or simply the Fed),
historically used monetary policy tools to keep interest rates stable. In 1979, the
Fed also began targeting money supply growth. Despite this oversight, oil shocks
and other supply-side disturbances created double-digit inflation rates in the 1970s
and 1980s, which in turn led to double-digit US interest rates that wiggled more
than ever before. These highly volatile interest rates created a need for securities to
help corporations hedge this risk. Interest rate derivatives arose. The Chicago Board
of Trade (CBOT; now part of the CME Group) developed the first interest rate
derivative contract, the Ginnie Mae futures, in 1975 and the highly popular Treasury
bond futures in 1977.
The foreign exchange market is one of the world’s largest financial markets,
where billions of dollars change hands daily. From the mid-1940s until the early
1970s, the world economy operated under the Bretton Woods system of fixed
exchange rates—all the currencies were pegged to the US dollar, and the dollar
was pegged to gold at $35 per ounce. This stable monetary system worked well for
decades. However, problems arose when gold prices soared. Countries converted
their currencies into dollars and bought cheap gold from the US at the bargain price
of $35 per ounce, making huge profits. This was an arbitrage, a trade that makes
riskless profits with no investment. Consequently, US gold reserves suffered a terrible
decline. Because all currencies were tied to the dollar, the United States could not
adjust the dollar’s exchange rate to fix the problem. Instead, US president Richard
Nixon abandoned the Bretton Woods system in 1971.
Currencies now float vis-à-vis one another in a so-called free market, although their
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
values are frequently managed by central banks. Floating exchange rates are more
volatile than fixed exchange rates, and to hedge this newly created currency risk, the
huge foreign exchange derivatives market was created. In this regard, 1972 saw the
opening of the International Monetary Market, a division of the Chicago Mercan-
tile Exchange (CME or Merc; now part of the CME Group) to trade foreign currency
futures. The world’s first exchange-traded financial derivatives contract was born!
Given these regulatory changes and well-functioning interest rate and foreign
currency derivatives markets, in the recent past, many economists believed that a new
era of greater macroeconomic stability had dawned, dubbed the Great Moderation.
During the two decades before the new millennium, fluctuations in the growth of real
output and inflation had declined, stock market volatility was reduced, and business
cycles were tamed.
However, the tide soon turned. In 2007–2009, many nations were mired in the
Great Recession, with declining economic output and large unemployment. Stock
market volatility, as measured by the widely followed VIX Index, shot up from 10
percent to an astonishing 89.53 percent in October 2008.1 Volatility had returned
with a vengeance!
1
See the Chicago Board Options Exchange’s website, www.cboe.com/.
TRADED DERIVATIVE SECURITIES 7
Two Economic Motives
Regulatory changes have powerful impacts on markets. In fact, economics Nobel
laureate Merton Miller argued in a 1986 article that regulations and taxes cause financial
innovation. The reason is because derivative securities are often created to circumvent
government regulations that prohibit otherwise lucrative transactions. And because
most countries tax income from different sources (and uses) at different rates, financial
innovations are often designed to save tax dollars as well. The desire to lower transactions
costs also influences financial innovation. This perspective comes from another Nobel
Prize–winning University of Chicago economist, Ronald Coase. Financial institu-
tions often devise derivatives so that brokerage commissions, the difference between
a securities dealer’s buying and selling prices, are minimized (see Extension 1.1).
This leads us to an (almost) axiomatic truth that will guide us throughout the book:
trading moves to those markets where transaction costs and regulatory constraints are minimized.
1.3 Traded Derivative Securities
A security, such as a stock or bond, gives its holder ownership rights over some
assets and cash flows. It exists as a paper document or an electronic entry record and
usually trades in an organized market. You may be familiar with bonds, which are
debts of the issuer, and stocks, which give investors equity in the issuing company.
Bonds and stocks require an initial investment. Most bonds pay back a promised
amount (principal or par value) at maturity. Some bonds pay interest (coupons)
on a regular basis, typically semiannually, while others are zero-coupon bonds
that pay no interest but are sold at a discount from the principal. The investment in
stock is never repaid. Stockholders usually get paid dividends on a quarterly basis as
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
compensation for their stock ownership. Stock prices can increase and create capital
gains for investors, and this profit is realized by selling the stock. Alternatively, stock
prices can decrease and create capital losses.
Stocks and bonds are often used to create new classes of securities called derivatives,
and that’s where the variety comes in. As previously noted, a derivative security is a
financial contract whose value is derived from one or more underlying assets or
indexes—a stock, a bond, a commodity, a foreign currency, an index, an interest rate,
or even another derivative security. Forwards, futures, and options are the basic types
of derivatives. These are explained later in the book.
Some common terminology will help us understand the various derivative
contracts traded:
1. Real assets include land, buildings, machines, and commodities, whereas finan-
cial assets include stocks, bonds, and currencies—both real and financial assets
have tangible values.
2. Notional variables include interest rates, inflation rates, and security indexes,
which exist as notions rather than as tangible assets.
8 CHAPTER 1: DERIVATIVES AND RISK MANAGEMENT
EXTENSION 1.1: The Influence of Regulations, Taxes, and Transaction Costs on
Financial Innovation
In the old days, finance mainly consisted of legal issues, institutional description, and investment rules of thumb.
This changed in the middle of the twentieth century, when financial economics sprang to life as an offshoot of
economics. In a little over two decades, a new finance based on a rigorous analytics emerged. James Tobin and
Harry Markowitz’s portfolio theory (late 1940s and early 1950s); Franco Modigliani and Merton Miller’s M&M
propositions concerning the irrelevance of a firm’s capital structure and dividend policy (late 1950s and early 1960s); William
Sharpe, John Lintner, and Jan Mossin’s capital asset pricing model (mid-1960s) and Fischer Black, Myron Scholes,
and Robert Merton’s option pricing model (early 1970s) established the basic theories. All these works have been
celebrated with the Sveriges Riksbank (Bank of Sweden) Prize in Economic Sciences in Memory of Alfred
Nobel, popularly known as the Nobel Prize in Economics. Other Nobel laureates, including Kenneth Arrow,
Ronald Coase, Gerard Debreu, John Hicks, Paul Samuelson, and a well known economist, John Maynard Keynes,
also contributed to finance. Two of these Nobel laureates’ views concerned financial innovation.
Miller’s View: Regulations and Taxes Spur Financial Innovation
What leads to revolutionary change in financial institutions and instruments? University of Chicago professor
Merton Miller (1923–2000) argued that a major cause of financial innovation is regulations and taxes. Innovative
financial securities are often created to avoid government regulations that prohibit otherwise lucrative transactions.
And because most countries tax income from different sources (and uses) at different rates, derivative securities
can be designed to save tax dollars.
Miller (1986) gives several examples:
■ Regulation Q of the US placed a ceiling on the interest rate that commercial banks could pay on time deposits.
Although this wasn’t a problem during much of the postwar period, US interest rates rose above this ceiling
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
during the late 1960s and early 1970s. When this occurred, US banks started losing customers. US banks
realized, however, that Regulation Q did not apply to dollar-denominated time deposits in their overseas
branches, and they soon began offering attractive rates via Eurodollar accounts. Interestingly, Regulation Q
has long been repealed, but the Eurodollar market still continues to flourish.
■ In the late 1960s, the US government imposed a 30 percent withholding tax on interest payments to bonds
sold in the US to overseas investors. Consequently, for non-US citizens, the market for dollar-denominated
bonds moved overseas to London and other money centers on the continent. This created the Eurobond
market that still continues to grow today.
■ The British government restricted dollar financing by British firms and sterling financing by non-British firms.
Swaps, transactions in which counterparties exchange one form of cash flow for another, were developed to
circumvent these restrictions.
■ It was found in 1981 that US tax laws allowed a linear approximation for computing the implicit interest
for long-term deep discount zero-coupon bonds. This inflated the present value of the interest deductions so
much that a taxable corporation could actually profit by issuing a zero-coupon bond and giving it away! Not
surprisingly, US corporations started issuing zero-coupon bonds in large numbers. This supply dwindled after
the US Treasury fixed this blunder.
TRADED DERIVATIVE SECURITIES 9
Coase’s View: Transaction Costs are a Key Determinant of Economic Activity
In 1937, Ronald Coase, a twenty-seven-year-old faculty member at the London School of Economics, published
a simple but profound article titled “The Nature of the Firm.” Coase argued that transactions incur costs,
which come from “negotiations to be undertaken, contracts have to be drawn up, inspections have to be made,
arrangements have to be made to settle disputes, and so on,” and firms often appear when they can lower these
transaction costs. With respect to financial markets, this logic implies that market participants often trade where
they can achieve their objectives at minimum cost. Financial derivatives are often created so that these costs are
minimized as well.
The lowering of transaction costs as an economic motivation was especially important during the 1990s and
the new millennium. Changes in the economic and political landscape and the IT revolution have made it possible
to significantly lower transaction costs, even eliminating age-old professions like brokers and dealers from many
trading processes. For example, this motive was a major reason why traders migrated from Treasury securities
to Eurodollar markets. Eurodollar markets, being free from Fed regulations and the peculiarities of the Treasury
security auction cycle, have fewer market imperfections and lower liquidity costs.
Writing financial contracts on values or cash flows determined by future realiza-
tions of real asset prices, financial asset prices, or notional variables creates a derivative
security. Early derivatives were created solely from financial and real assets. For
example, in the 1960s, agricultural commodity–based futures were the most actively
traded derivative contracts in the US. However, as the economy has evolved, notional
variables and derivative securities based on these notional variables have also been
created.
Notional variables are often introduced to help summarize the state of the
economy. Just open the Wall Street Journal, and you will be amazed by the variety
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
of indexes out there. You will find not only regular stock price indexes, such as the
Dow-Jones and Standard and Poor’s, but a whole range of other indexes, including
those based on technology stocks, pharmaceuticals stocks, Mexican stocks, utility
stocks, bonds, and interest rates. In addition, notional values are often useful for the
creation of various derivatives. Perhaps the most famous example of this is a plain
vanilla interest rate swap whose underlyings are floating and fixed interest rates.
Diverse Views on Derivatives
It is easy to speculate with derivatives: buy them, take huge one-sided bets, and laugh
on the way to the bank or cry on the way to bankruptcy! Imprudent risk taking
using derivatives is not uncommon, as illustrated by huge losses at many institutions,
including Procter and Gamble (P&G) and Barings Bank. People are uncomfortable
with derivatives because they are complex instruments: difficult to understand, highly
leveraged, and oftentimes not backed by sufficient collateral. High leverage means
that small changes in the underlying security’s price can cause large swings in the
derivative’s value.
10 CHAPTER 1: DERIVATIVES AND RISK MANAGEMENT
For these reasons, derivatives attract strong views from both sides of the aisle.
The renowned investors Warren Buffett and Peter Lynch dislike derivatives. In his
“Chairman’s Letter” in Berkshire Hathaway’s 2002 annual report, Buffett charac-
terized derivatives as “time bombs, both for the parties that deal in them and the
economic system.” Lynch (1989) once stated that options and futures on stocks should
be outlawed. These concerns were vindicated by the hundreds of billions of dollars
of derivatives-related losses suffered by financial institutions during 2007 and 2008,
which contributed to the severe economic downturn.
By contrast, former Fed chairman Alan Greenspan opined in a speech delivered
before the Futures Industry Association in 1999 that derivatives “unbundle” risks by
carefully measuring and allocating them “to those investors most able and willing to
take it,” a phenomenon that has contributed to a more efficient allocation of capital.
And in Merton Miller on Derivatives, the Nobel laureate (Miller 1997, ix) assessed the
impact of the “derivatives revolution” in glowing terms:
Contrary to the widely held perception, derivatives have made the world a safer place,
not a more dangerous one. They have made it possible for firms and institutions to deal
efficiently and cost effectively with risks and hazards that have plagued them for decades,
if not for centuries. True, some firms and some financial institutions have managed to
lose substantial sums on derivatives, but some firms and institutions will always find ways
to lose money. Good judgment and good luck cannot be taken for granted.
Nobel laureate physicist David Gross provides a nice analogy for this disagreement.
He views knowledge as expanding outward like a growing sphere and ignorance as the
surface of that sphere. With respect to derivatives, we have accumulated significant
knowledge over the past thirty years, but with respect to the causes of tsunami-like
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
financial crises, there is still much for us to learn.
Applications and Uses of Derivatives
Derivatives trade in zero net supply markets, where each buyer has a match-
ing seller. When hedging with a derivative, the other side of the transaction
(counterparty) may be using it for speculative reasons. Hedging and speculation
are often two sides of the same coin, and it is a zero-sum game because one trader’s
gain is the other’s loss. Example 1.1 illustrates these concepts.
EXAMPLE 1.1: Hedging and Speculation in a Derivative Transaction
■ April is the beginning of the corn-growing season, the commodity used in our example.
■ Consider Mr. Short, a farmer in the midwestern United States. Short combines land, labor, seeds,
fertilizers, and pesticides to produce cheap corn. He hopes to sell his corn harvest in September.
■ Expertise in growing corn does not provide a crystal ball for forecasting September corn prices. If
Short likes sleeping peacefully at night, he may decide to lock in the selling price for September corn
when he plants it in April.
TRADED DERIVATIVE SECURITIES 11
■ To see how this is done, let’s assume that everyone expects corn to be worth $10.00 per bushel in
September. Ms. Long, a trader and speculator, offers to buy Short’s corn in September for $9.95 per
bushel, which is the forward price.
■ To remove his risk, Short readily agrees to this forward price. Together they have created a forward
contract—a promise to trade at a fixed forward price in the future. Although Short expects to lose 5
cents, he is happy to fix the selling price. The forward contract has removed output price uncertainty
from his business. Short sees 5 cents as the insurance premium he pays for avoiding unfavorable future
outcomes. Having hedged his corn selling price, Short can focus on what he knows best, which is
growing corn.
■ Ms. Long is also happy—not that she is wild about taking risks, but she is rational and willing to accept
some unwanted risk, expecting to earn 5 cents as compensation for this activity. Later in the book,
you will see how a speculator may manage her risk by entering into another transaction at a better
price but on the other side of the market.
This simple example illustrates a key principle of historical importance—it was
precisely to help farmers hedge grain prices that the first modern derivatives
exchange, the CBOT, was established in 1848. Even the bitterest critics of derivatives
have acknowledged this beneficial role.
Nowadays, the uses of derivatives are many. Consider the following possibilities:
■ Hedging output price risk: A gold-mining company can fix the selling price of gold
by selling gold futures.
■ Hedging input cost: A sausage maker can hedge input prices by buying pork belly
futures.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Hedging currency risk: An American manufacturer buying machines from Germany
for which the payment is due in three months can remove price risk from the
dollar–euro exchange rates by buying a currency forward contract.
■ Hedging interest rate risk: A pension fund manager who is worried that her bond
portfolio will get clobbered by rising interest rates can use Eurodollar futures (or
options on those futures) to protect her portfolio.
■ Protecting a portfolio against a market meltdown: A money manager whose portfolio
has reaped huge gains can protect these gains by buying put options.
■ Avoiding market restrictions: A trader can avoid short selling restrictions that an
exchange might impose by taking a sell position in the options or the futures
market.
■ Ruining oneself: A wretch can gamble away his inheritance with derivatives.
Before you finish this book, you will have a better understanding of the role
derivatives play in each of these scenarios, except for the last one. The proclivity
to gamble and lose arises from the depths of the human psyche, of which we have no
special understanding.
12 CHAPTER 1: DERIVATIVES AND RISK MANAGEMENT
A Quest for Better Models
Great recognition for derivatives came in 1997, when Robert Merton and Myron
Scholes won the Nobel Prize in Economics for developing the Black–Scholes–
Merton (BSM) model for option valuation. A co-originator of the model, Fischer
Black, died earlier and missed receiving the prize, which is not given posthumously.
This formula has become a staple of modern option pricing. Although extremely
useful—options traders have it programmed into their computers—the BSM model
makes a number of restrictive assumptions. These restrictive assumptions limit the
application of the model to particular derivatives, those for which interest rate risk is
not relevant (the BSM assumptions are discussed in later chapters).
Subsequent researchers have relaxed these restrictive assumptions and developed
pricing models suitable for derivatives with interest rate risk. The standard model used
on Wall Street for this application is the Heath–Jarrow–Morton (HJM) model.
The HJM model can be used to price interest rate derivatives, long-lived financial
contracts, and credit derivatives. Credit derivatives are among the newest class of
derivatives to hit Wall Street. Started in the early 1990s, they now compose a huge
market, and they are the subject of intense public debate following the 2007 credit
crisis and their role in it.
Why develop and study better pricing models? Suppose you trade a variety of
securities many times a day. If your models are better, you have an edge over your
competitors in determining fair value. Over many transactions, when trading based
on fair value—selling above fair value and buying below—your edge will pay off.
Investment banking and Wall Street firms have no choice but to use the best models.
It’s just too important for their survival. Powerful models also help traders manage
risk.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1.4 Defining, Measuring, and
Managing Risk
Risk is an elusive concept and hard to define. In finance and business, risk can be
defined and measured in many ways, none of them completely universal. Risk often
depends on the user’s perspective. We begin by looking at risk through a regulator’s
eyes. Next, we discuss risk from an individual’s as well as an institutional trader’s
viewpoint. Finally, we discuss risk from a corporate management perspective.
1.5 The Regulator’s Classification of Risk
In 1994, the international banking and securities regulators issued guidelines for
supervising the booming derivatives market. The Basel Committee on Banking
Supervision and the International Organization of Securities Commissions (IOSCO)
recommended that for a stable world financial system, the national regulators had to
THE REGULATOR’S CLASSIFICATION OF RISK 13
ensure that banks and securities firms have adequate controls over the risks they incur
when trading derivatives.2
The Basel Committee’s Risk Management Guidelines for Derivatives (July 1994,
10–17) identified the following risks in connection with an institution’s deriva-
tive activities. IOSCO’s Technical Committee also issued a similar paper at the
time:3
■ Credit risk (including settlement risk) is the risk that a counterparty will fail to
perform on an obligation.
■ Market risk is the risk to an institution’s financial condition resulting from adverse
movements in the level or volatility of market prices. This is the same as price risk.
■ Liquidity risk in derivative activities can be of two types: one related to specific
products or markets and the other related to the general funding of the institution’s
derivative activities. The former is the risk that an institution may not be able
to, or cannot easily, unwind or offset a particular position at or near the previous
market price because of inadequate market depth or disruptions in the marketplace.
Funding liquidity risk is the risk that the institution will be unable to meet its
payment obligations on settlement dates or in the event of margin calls (which, we
explain later, is equivalent to coming up with more security deposits).
■ Operational risk (also known as operations risk) is the risk that deficiencies
in information systems or internal controls will result in unexpected loss. This
risk is associated with human error, system failures, and inadequate procedures and
controls.
■ Legal risk is the risk that contracts are not legally enforceable or documented
correctly.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Managing market or price risk is the subject of this book. Initially, this topic
attracted the sole attention of academics and practitioners alike. It still remains the
most important risk for us to understand and to manage. The other risks only appear
when normal market activity ceases. Credit risk evaluation is currently a subject
of advanced research. As we explain later, exchange-traded derivatives markets are
so designed that they are nearly free from credit risk. Liquidity risk is a persistent
problem for traders who choose markets in which securities are not easily bought
and sold. Chapter 3 discusses what makes markets illiquid. Operational risk is a
reality with which one has to live. It is part of running a business, and appropriate
management checks and balances reduce it. Legal risk isn’t a problem for exchange-
2
In 1974, the Group of Ten countries’ central-bank governors established the Basel Committee on
Banking Supervision. The Basel (or Basle) Committee formulates broad supervisory standards and
recommends statements of best practice in the expectation that individual authorities will implement
(see “History of the Basel Committee and Its Membership,” www.bis.org/bcbs/history.htm). The
International Organization of Securities Commissions, which originated as an inter-American
regional association in 1974, was restructured as IOSCO in 1983. It has evolved into a truly international
cooperative body of securities regulators (see www.iosco.org/about/about).
3
For both reports, see www.bis.org/publ. These definitions are modified only slightly from those
contained in the referenced report.
14 CHAPTER 1: DERIVATIVES AND RISK MANAGEMENT
traded contracts; however, it’s a genuine problem in OTC markets. We leave this
topic for the courts and law schools.
This jargon-laden Basel Committee report also cited the need for appropriate
oversight of derivatives trading operations by boards of directors and senior man-
agement and the need for comprehensive internal control and audit procedures. It
urged national regulators to ensure that firms and banks operate on a basis of prudent
risk limits, sound measurement procedures and information systems, continuous risk
monitoring, and frequent management reporting. The Basel Committee reports have
been extremely influential in terms of their impact on derivatives regulation. We
return to the issues again in the last chapter of the book, after mastering the basics of
derivatives securities.
1.6 Portfolio Risk Management
Let us examine some issues in connection with portfolio risk management from the
perspective of an individual or institutional trader. One of the first pearls of wisdom
learned in finance is that to earn higher expected returns, one has to accept higher risks. This
maxim applies to the risk of a portfolio, called portfolio risk.
This portfolio risk, when applied to a single security, can be broken into
two parts: (1) nondiversifiable risk, which comes from market-wide sources,
and (2) diversifiable risk, which is unique to the security and can be elimi-
nated via diversification. Pioneering works by Markowitz (1952), Sharpe (1964),
Lintner (1965), and Mossin (1966) gave us the capital asset pricing model—the
alphas, the betas—ideas and concepts used by market professionals and finance
academics.
Modern portfolio theory encourages investors to construct a portfolio in a top-
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
down fashion. This involves three steps: (1) do an asset allocation, which means
deciding how to spread your investment across broad asset classes such as cash, bonds,
stocks, and derivatives; (2) do security selection, which means deciding which
securities to hold within each asset class; and (3) periodically revisiting these issues,
rebalancing and hedging the portfolio with derivatives as appropriate. Portfolio risk
management is critical for investment companies such as hedge funds and mutual
funds. These financial intermediaries receive money from the public, invest them in
various financial securities, and pass on the gains and losses to investors after deducting
expenses and fees.
1.7 Corporate Financial Risk Management
Risks That Businesses Face
To understand the risks that businesses face, let’s take a look at a typical company’s
balance sheet, a vital accounting tool that gives a snapshot of its financial condition
by summarizing its assets, liabilities, and ownership equity on a specific date. An
asset provides economic benefits to the firm, whereas a liability is an obligation
that requires payments at some future date. The difference between the two accrues
CORPORATE FINANCIAL RISK MANAGEMENT 15
TABLE 1.1: Risks That a Business Faces
Assets Liabilities
Current assets
■ Cash and cash equivalents (interest rate
■ Accounts payable (interest rate risk,
risk) currency risk)
■ Accounts receivable (interest rate risk,
■ Financial liabilities (interest rate risk,
currency risk) currency risk)
■ Inventories (commodity price risk)
■ Pension fund obligations (interest rate
risk, market risk)
Long-term assets
Equity
■ Financial assets (interest rate risk, market
■ Ownership shares
risk, currency risk)
■ Property, plant, and equipment (interest
rate risk, commodity price risk)
to the owners of the company as shareholder’s equity (hence the identity assets
equals liabilities plus shareholder’s equity). Depending on whether they are for the short
term (one year or less) or for the long haul, assets and liabilities are further classified as
current or noncurrent, respectively. Table 1.1 illustrates some risks that affect different
parts of a firm’s balance sheet.
We see that a typical company faces three kinds of risks: currency risk, interest
rate risk, and commodity price risk. These are the three components of market
risk:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1. If a big chunk of your business involves imports and exports or if you have overseas
operations that send back profits, then exchange rate risk can help or hurt. This
is a risk that you must understand and decide whether to hedge using currency
derivatives.
2. No less important is interest rate risk. It is hard to find companies like Microsoft,
with tens of billions of dollars in cash holdings that can be quickly deployed
for value-enhancing investments. Most companies are cash strapped. Interest rate
fluctuations therefore affect their cost of funds and influence their investment
activities. Interest rate derivatives offer many choices for managing this risk.
3. Unless a financial company, most businesses are exposed to fluctuations in
commodity prices. A rise in commodity prices raises the cost of buying inputs
that may not always be passed on to customers. For example, if crude oil prices
go up, so does the price of jet fuel and an airline’s fuel costs. Sometimes airlines
levy a fuel surcharge, but it is unpopular and often rolled back. Another option is
to hedge such risks by using oil-price derivatives.
16 CHAPTER 1: DERIVATIVES AND RISK MANAGEMENT
Nonhedged Risks
Despite the development of many sophisticated derivatives useful for hedging (and
speculation), some risks are difficult, if not impossible, to hedge. For example, it is
very hard to hedge operational risk. Recall that operational risk is the risk of a loss
owing to events such as human error, fraud, or faulty management. Although a bank
can buy insurance to protect itself from losses due to fire, no insurance company
will insure a bank against the risk that a trader presses the wrong computer button
and enters the wrong bond trade. For examples of such operational risk losses, see
Chapter 26. Other difficult or impossible to hedge risks include losses because of
changes in commodity prices for which there is no futures contract trading.
We will now take a bird’s-eye view of how a blue chip company uses derivatives
for risk management.
Risk Management in a Blue Chip Company
The SFAS (Statement of Financial Accounting Standards) No. 133,
“Accounting for Derivative Instruments and Hedging Activities,” as amended,
requires US companies to report all derivative instruments on the balance sheet at fair
value and establishes criteria for designation and effectiveness of hedging relationships.
Let’s take a look at Form 10-K (annual report) filed with the US Securities and
Exchange Commission by Procter & Gamble (P&G), a giant company that owns
some of the world’s best-known consumer product brands.4
P&G heavily uses derivatives for risk management. A careful reading of this annual
report reveals several interesting characteristics of P&G’s risk management activity:
■ P&G is exposed to the three categories of risk that we just mentioned.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ P&G consolidates these risks and tries to offset them naturally, which means some
risks cancel each other. It then tries to hedge the rest with derivatives.
■ P&G does not hold derivatives for trading purposes.
■ P&G monitors derivatives positions using techniques such as market value, sensi-
tivity analysis, and value at risk. When data are unavailable, P&G uses reasonable
proxies for estimating volatility and correlations of market factors.
■ P&G uses interest rate swaps to hedge its underlying debt obligations and enters into
certain currency interest rate swaps to hedge the company’s foreign net investments.
■ P&G manufactures and sells its products in many countries. It mainly uses forwards
and options to reduce the risk that the company’s financial position will be adversely
affected by short-term changes in exchange rates (corporate policy limits how
much it can hedge).
■ P&G sometimes uses futures, options, and swaps to manage the price volatility of
raw materials.
4
See P&G’s 2015 Annual Report (http://www.pginvestor.com/CustomPage/Index?KeyGenPage=1073748359)
RISK MANAGEMENT PERSPECTIVES IN THIS BOOK 17
■ P&G designates a security as a hedge of a specific underlying exposure and monitors
its effectiveness in an ongoing manner.
■ P&G’s overall currency and interest rate exposures are such that the company is 95
percent confident that fluctuations in these variables (provided they don’t deviate
from their historical behavior) would not materially affect its financial statements.
P&G expects significant risk neither from commodity hedging activity nor from
credit risk exposure.
■ P&G grants stock options and restricted stock awards to key managers and directors
(employee stock options are valued by using a binomial model that you will learn
to implement in Part III of this book).
How does one understand, formulate, and implement hedging and risk manage-
ment strategies like the ones adopted by P&G? Read on, for that is a major purpose
of this book.
1.8 Risk Management Perspectives in
This Book
The Basel Committee’s risk nomenclature has become the standard way that market
participants see, classify, and understand different types of risks in derivatives markets.
In this book, we restrict our focus mainly to market or price risk. For this we usually
wear four separate hats and look at risk management from the perspectives of an
individual trader, a financial institution, a nonfinancial corporation, and a dealer.
In early chapters, we start by examining financial securities and their markets from
an individual trader’s standpoint. This simple and intuitive approach makes it easier to
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
understand the material. It’s also historically correct because many markets, including
those for stocks, futures, and options, were started by individual traders.
But today’s derivatives markets have become playing fields for financial institutions.
This is no surprise because as an economy develops, a greater share of its gross
domestic product comes from services, of which financial institutions are a major
constituent. These institutions generally engage in sophisticated ways of investing.
We discuss derivatives and risk management from a financial institution’s viewpoint
because you may eventually be working for one, or at the very least, you are
likely to have financial dealings with such institutions on a regular basis. The term
institution is a catchall phrase that includes commercial and investment banks,
insurance companies, pension funds, foundations, and finance companies such as
mutual funds and hedge funds. In later chapters, we study swaps and interest
rate derivatives, whose markets are the near-exclusive domain of financial (and
nonfinancial) institutions.
Nonfinancial companies engage in more real activity than financial companies.
Nonfinancials give us food, develop medicines, build homes, manufactures cars,
refine crude oil to create gasoline, generate electricity, provide air travel, create
household chemicals, and make computers to save and express our ideas. They
buy one or more inputs to produce one or more outputs, and different kinds of
risk (including exchange rate risk, interest rate risk, and commodity price risk) can
18 CHAPTER 1: DERIVATIVES AND RISK MANAGEMENT
seriously affect their businesses. This takes us to financial engineering, which applies
engineering tools to develop financial contracts to meet the needs of an enterprise.
This is our third hat: looking at derivatives from a nonfinancial company’s risk management
perspective.
Sometimes, we take a dealer’s perspective. A dealer is a financial intermediary who
posts prices at which she can buy (wholesale or bid price) or sell (retail or ask price)
securities to her customers. Trying to make a living from the spread, or the difference
between these two prices, the dealer focuses on managing books, which means carefully
controlling her inventory of securities to minimize risk. Both an individual trader and
a financial institution can play a dealer’s role in financial markets.
These four perspectives aren’t ironclad, and we flit from one to another as the
discussion demands. In the final analysis, this shifting from one category to another
isn’t bad for the aspiring derivatives expert. In the words of the immortal bard William
Shakespeare, from Hamlet (if we take the liberty of forgetting about spirits and instead
apply this to the mundane), “there are more things in heaven and earth, Horatio,
than are dreamt of in your philosophy.” So stay awake and study derivatives, develop
a sense for risky situations, understand the markets, learn pricing models, and know
their limitations.
1.9 Summary
1. Derivatives are financial securities that derive their value from some underlying
asset price or index. Derivatives are often introduced to hedge risks caused by
increasingly volatile asset prices. For example, during the last three decades of the
twentieth century, we have moved away from a regime of fixed exchange rates to a
world of floating exchange rates. The market-determined foreign exchange rates
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
increased their volatility, creating the need for foreign currency derivatives.
2. In today’s interconnected global economy, risks coming from many different
sources can make or break businesses. Derivatives can both magnify risk (leverage
and gambling) or reduce risk (hedging). Although gambling with derivatives
remains popular in some circles, and derivatives mishaps grab newspaper headlines,
most traders prudently use derivatives to remove unwanted risks affecting their
businesses.
3. There are many types of risk. Besides market risk, the regulators are concerned
about credit risk, liquidity risk, operational risk, and legal risk. Though unglam-
orous, these risks can eat away profits and jolt the running of smooth-functioning
derivatives markets.
4. Derivatives are very useful for managing the risk of a portfolio, which is a
collection of securities. Portfolio risk management is important for both individual
investors and financial companies such as hedge funds and mutual funds. Many
businesses are exposed to exchange rate risk, interest rate risk, and commodity
price risk and use derivatives to hedge them. US companies report derivatives
usage and exposure on Form 10-K filed annually with the Securities and Exchange
Commission. For example, P&G reports that it is exposed to exchange rate risk,
QUESTIONS AND PROBLEMS 19
interest rate risk, commodity price risk, and credit risk. P&G consolidates risks
and attempts to offset them naturally and tries to hedge the remaining risk with
derivatives.
5. We look at risk management from several different perspectives: that of an
individual trader, a financial institution, a nonfinancial corporation, and a dealer.
These perspectives aren’t ironclad, and we flit from one to another as the discussion
demands.
1.10 Cases
Hamilton Financial Investments: A Franchise Built on Trust (Harvard Business
School Case 198089-PDF-ENG). The case discusses various risks faced by a
finance company that manages mutual funds and provides discount brokerage
services.
Grosvenor Group Ltd. (Harvard Business School Case 207064-PDF-ENG). The
case considers whether a global real estate investment firm should enter into a
property derivative transaction to alter its asset allocation and manage its business.
Societe Generally (A and B): The Jerome Kerviel Affair (Harvard Business
School Cases 110029 and 110030-PDF-ENG). The case illustrates the importance
of internal control systems in a business environment that involves a high degree
of risk and complexity in the context of a derivatives trader indulging in massive
directional trades that went undetected for over a year.
1.11 Questions and Problems
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1.1. What is a derivative security? Give an example of a derivative and explain why
it is a derivative.
1.2. List some major applications of derivatives.
1.3. Evaluate the following statement: “Hedging and speculation go hand in hand
in the derivatives market.”
1.4. What risks does a business face?
1.5. Explain why financial futures have replaced agricultural futures as the most
actively traded contracts.
1.6. Explain why derivatives are zero-sum games.
1.7. Explain why all risks cannot be hedged. Give an example of a risk that cannot
be hedged.
1.8. What is a notional variable, and how does it differ from an asset’s price?
1.9. Explain how derivatives give traders high leverage.
1.10. Explain the essence of Merton Miller’s argument explaining what spurs
financial innovation.
20 CHAPTER 1: DERIVATIVES AND RISK MANAGEMENT
1.11. Explain the essence of Ronald Coase’s argument explaining what spurs
financial innovation.
1.12. Does more volatility in a market lead to more use of financial derivatives?
Explain your answer.
1.13. When the international banking regulators defined risk in their 1994 report,
what definition of risk did they have in mind? How does this compare with
the definition of risk from modern portfolio theory?
1.14. What’s the difference between real and financial assets?
1.15. Explain the differences between market risk, credit risk, liquidity risk, and
operational risk.
1.16. Briefly present Warren Buffett’s and Alan Greenspan’s views on derivatives.
1.17. Consider the situation in sunny Southern California in 2005, where house
prices have skyrocketed over the last few years and are at an all-time high.
Nathan, a software engineer, buys a second home for $1.5 million. Five years
back, he bought his first home in the same region for $350,000 and financed
it with a thirty-year mortgage. He has paid off $150,000 of the first loan. His
first home is currently worth $900,000. Nathan plans to rent out his first home
and move into the second. Is Nathan speculating or hedging?
1.18. During the early years of the new millennium, many economists described the
past few decades as the period of the Great Moderation. For example,
■ an empirical study by economists Olivier Blanchard and John Simon found
that “the variability of quarterly growth in real output (as measured by its
standard deviation) had declined by half since the mid-1980s, while the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
variability of quarterly inflation had declined by about two thirds.”
■ an article titled “Upheavals Show End of Volatility Is Just a Myth” in the Wall
Street Journal, dated March 19, 2008, observed that an important measure of
stock market volatility, “the Chicago Board Options Exchange’s volatility
index, had plunged about 75% since October 2002, the end of the latest
bear market, through early 2007”; the article also noted that “in the past 25
years, the economy has spent only 16 months in recession, compared with
more than 60 months for the previous quarter century.”
a. What were the explanations given for the Great Moderation?
b. Does the experience of the US economy during January 2007 to Decem-
ber 2010 still justify characterizing this as a period of Great Moderation?
Report (1) quarterly values for changes in the gross domestic product,
(2) quarterly values for changes in the inflation rate, and (3) the volatility
VIX Index value during this period to support your answer.
1.19. Drawing on your experience, give examples of two risks that one can easily
hedge and two risks that one cannot hedge.
QUESTIONS AND PROBLEMS 21
1.20. Download Form 10-K filed by P&G from the company’s website or the
US Securities and Exchange Commission’s website. Answer the following
questions based on a study of this report:
a. What are the different kinds of risks to which P&G is exposed?
b. How does P&G manage its risks? Identify and state the use of some
derivatives in this regard.
c. Name some techniques that P&G employs for risk management.
d. Does P&G grant employee stock options? If so, briefly discuss this program.
What valuation model does the company use for valuing employee stock
options?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
2
Interest Rates
2.1 Introduction 2.7 Different Ways of Investing in
Treasury Securities
2.2 Rate of Return
The Treasury Auction and Its
2.3 Basic Interest Rates: Simple, Associated Markets
Compound, and Continuously The Repo and the Reverse Repo
Compounded Market
EXTENSION 2.1 Conventions
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
EXTENSION 2.3 A Repurchase
and Rules for Rounding, Agreement
Reporting Numbers, and
Interest Rate Derivatives
Measuring Time
2.4 Discounting (PV) and 2.8 Treasury Bills, Notes, Bonds,
Compounding (FV): Moving and STRIPS
Money across Time 2.9 Libor versus a Libor Rate
EXTENSION 2.2 Moving Index
Multiple Cash Flows across Time
2.10 Summary
2.5 US Treasury Securities
2.11 Cases
2.6 US Federal Debt Auction
2.12 Questions and Problems
Markets
INTRODUCTION 23
2.1 Introduction
Gentlemen prefer bonds. Perhaps you are chuckling at our “misquotation” of the famous
Marilyn Monroe movie title Gentlemen Prefer Blondes. In fact, former US treasury
secretary Andrew Mellon made this observation about bonds many decades before the
movie.1 In all likelihood, he said this because bonds are considered safer than stocks.
Corporations may miss dividend payments on stocks, but bonds are legally bound to
pay promised interest on fixed dates. If the interest payments are not paid, default
occurs, and the corporations are vulnerable to lawsuits and bankruptcy. However, at
this stage in our presentation, we will assume that all bonds considered have no default
(credit) risk. This greatly simplifies the analysis. Consequently, the bonds considered
in this chapter are best viewed as US Treasury securities. Credit risk is considered
only later in the book (in Chapter 26), after we have mastered default-free securities.
Most bonds make interest payments according to a fixed schedule. Hence bonds
are also called fixed-income securities. Bonds are useful for moving money from
one period to another. Does money retain its value over time? Not really. Inflation,
the phenomenon of a rise in the general price level, chips away money’s buying power
little by little, year by year. A popular measure of the US inflation rate is the change
in the consumer price index (CPI). At each date, the CPI measures the price of
a fixed bundle of nearly two hundred goods and services that a typical US resident
consumes. Unless you are buying computers and electronic gizmos whose prices
have drastically fallen because of technological advances, a hundred dollar bill doesn’t
quite buy as much as it did ten years ago. To reduce the cost of inflation, money
earns interest. Instead of stashing cash under a mattress, one can put it into a savings
account and see it grow safely.
What, then, is an interest rate? An interest rate is the rate of return earned on money
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
borrowed or lent. Ponder and you realize that the cost of borrowing a hundred dollar
bill is the interest of perhaps $5 per year that’s paid to the lender. Suppose that the
inflation rate is 3 percent per year, which is close to the US inflation rate in the
new millennium. If the interest rate for risk-free loans is 3.10 percent per year, then
borrowing is incredibly cheap, while 30 percent for such loans would be awfully
expensive. Interest rates partly compensate the lender for inflation and partly reward
her for postponing consumption until a later date.
There are many ways of computing interest rates. Simple, compound, and
continuously compounded rates are the three basic kinds. Each interest rate concept
has its own use in finance theory as well as in practice. There are interest rates for
risky as well as riskless loans. The first interest rates studied in this book are risk-free
rates.
How do you find risk-free interest rates? United States Government Treasury
securities (the “Treasuries”) are considered default-free because of the taxing author-
ity of the mighty federal government. Risk-free interest rates of various maturities
can be easily extracted from their prices. We briefly explain how to do this extraction.
Then we explain how the US Treasury securities market works, and we describe the
1
Andrew William Mellon (1855–1937) was US secretary of the treasury under three presidents (1921–
32). Source: US Treasury’s website, www.ustreas.gov/education/history/secretaries/awmellon.html.
24 CHAPTER 2: INTEREST RATES
bills, notes, bonds, STRIPS, and TIPS that trade in these markets. We conclude the
chapter with a discussion of the London Interbank Offered Rate (libor) and a libor-
based rate index, which is a key rate used in the global interest rate derivatives market.
We begin by introducing the concept of a rate of return. This, in turn, will lead
us to a discussion of the three basic types of interest rates.
2.2 Rate of Return
In finance, we constantly talk about rates of return computed from prices. Rates
of return measure how much an investment has earned. For example, if you buy a
security for $50 today and sell it for $52 after six months, then you have earned $2 on
your investment. The rate of return earned on your investment over this time period is
52 − 50
= 0.04, or 4 percent (when multiplied by 100 to express as a percentage)
50
It is hard to compare rates of return unless we explicitly mention the time interval
over which they are computed. We need to know whether we are computing
these over six months, or thirty-seven weeks, or ninety-one days. Consequently,
it is worthwhile to standardize by computing them over a year. In this case, the
annualized rate of return is
12
( 6 ) × 0.04 = 0.08, or 8 percent per year
A more precise value for the rate of return may be obtained by considering the
actual number of days for which the money was invested using 365 days in a year
(some computations use 366 days in case of a leap year). Assuming that there were
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
181 days in this six-month period, we find the (annualized) rate of return as
365
( 181 ) × 0.04 = 0.0807, or 8.07 percent per year
The capital gain, the difference between the selling (final) and buying (initial)
price, is the $2 earned on this investment. But there could be other cash flows
at intermediate dates. For example, bonds make coupon payments and stocks pay
dividends. If our $50 investment also received $0.50 interest at some intermediate
date, then the (annualized) rate of return is
365 52 + 0.50 − 50
( 181 ) × ( 50 ) = 0.1008, or 10.08 percent per year
A profit of $2.50 is made on this investment. Expenses like brokerage account
fees can cut into your profit. A profit occurs when there’s an overall gain from an
investment or business activity. Businesses define it as total revenue minus total costs
and record a profit if this is a positive number and a loss if it’s negative. In finance, we
define
Profit or loss = (Selling price + Income) − (Buying price + Expenses) (2.1)
BASIC INTEREST RATES: SIMPLE, COMPOUND, AND CONTINUOUSLY COMPOUNDED 25
where the selling price is the final price, the buying price is the initial price, and a
profit or loss happens if the result is positive or negative, respectively.2
We generalize this to develop a formula for computing the annualized rate of
return (which is also called arithmetic return).
RESULT 2.1
Annualized Rate of Return
The annualized arithmetic rate of return is
365
Rate of return = (
T )
Selling price + Income − Expenses − Buying price
×
( Buying price )
365 Profit or loss
=( ) × (2.2)
T ( Buying price )
where T is the time interval, measured by the number of days over which the
investment is held, and Income and Expenses denote positive and negative cash
flows, respectively, from the investment (Rate of return is often multiplied by
100 and expressed as a percentage).
We need some reasonable conventions regarding rounding and reporting numbers
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and for measuring time, which we use throughout the book. They are discussed in
Extension 2.1.
2.3 Basic Interest Rates: Simple,
Compound, and Continuously
Compounded
An interest rate is the rate of return promised (in case the money is yet to be lent)
or realized (in case the money has already been lent) on loans. There seem to be a
zillion ways of computing interest. Financial institutions like banks and credit card
companies have devised clever ways of charging interest that appear less than they
really are. That’s why many governments came up with a standard yardstick to help
customers understand interest costs. For example, in the United States, all lenders have
to disclose the rate they are charging on loans in terms of an Annual Percentage
Rate (APR), a rate that annualizes using a simple interest rate.
2
An exception is short selling (see Chapter 3), in which selling occurs first and buying second, but the
expression retains the positive and negative signs associated with these expressions.
26 CHAPTER 2: INTEREST RATES
EXTENSION 2.1: Conventions and Rules for Rounding, Reporting Numbers, and
Measuring Time
To maintain consistency and to minimize unwanted errors, we follow some rules for rounding, reporting numbers,
and calculations. We usually round to four places after the decimal point when reporting results. In all calculations,
before reporting the numbers in the text, we retain 16 digits for accuracy. Any differences between reported results
based on rounded numbers and calculated results (manipulations of the actual numbers) are due to these rounding
errors:
■ If the number in the fifth place is more than 5, then add 1 to the fourth digit, e. g., 0.23456 becomes 0.2346.
■ If the number in the fifth place is less than 5, then keep it unchanged, e. g., 0.11223 becomes 0.1122.
■ If it’s exactly 5 and there are numbers after it, then add 1 in the fourth place, e. g., 0.123452 becomes 0.1235.
In some contexts, we need to round to more places after the decimal point. The context will indicate when
this is appropriate.
Though we typically report numbers rounded to four places after the decimal point, the final dollar result is
usually rounded to two places after the decimal. The dollar sign “$” is usually attached only in the final answer.
We sometimes omit dollar signs from the prices when the context is understood.
Different markets follow different time conventions. Treasury bill prices use the actual number of days to
maturity under the assumption that there are 360 days in a year. Many banks pay interest compounded daily.
Swaps use simple interest applied to a semiannual period, and many formulas use a continuously compounded rate.
Most models measure time in years. If we start computing at time 0 and the security matures at time T, then
it has a life of T years. If we start our clock at time t, then the life is (T – t) years. We prefer to use the first
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
convention in our formulas. Time periods are converted by the following conventions:
■ If the time period is computed in days, then use the exact number of days and assume 365 days in the year
(unless noted otherwise), e. g., 32 days will be 32/365 = 0.0877 year.
■ If the time period is computed in weeks, use fifty-two weeks in the year, e. g., seven weeks will be 7/52 =
0.1346 year.
■ If the time period is computed in months, use twelve months in the year, e. g., five months will be 5/12 =
0.4167 year.
To summarize, we employ the following rules and conventions.
Rules of Rounding and Conventions of Reporting Numbers and Measuring Time
■ Reported numbers are rounded to four places after the decimal point. All calculations are performed using 16
digit numbers, not the rounded numbers. Any differences between reported results based on rounded numbers
and calculated results (manipulations of the actual numbers) are due to these rounding errors.
■ The final dollar result is rounded to two places after the decimal with the dollar sign attached only in the final
answer.
■ Time is measured in years. If we start computing from time 0 and the security matures at time T, then it has
a life of T years; if the clock starts at time t, then the security has a life of (T – t) years.
BASIC INTEREST RATES: SIMPLE, COMPOUND, AND CONTINUOUSLY COMPOUNDED 27
There are many different ways of computing interest rates. For a certain quoted
rate, the realized interest can vary with (1) the method of compounding, (2) the
frequency of compounding (yearly, monthly, daily, or continuously compounded),
(3) the number of days in the year (52 weeks; 360 days, or 365, or 366 for a leap
year), (4) the number of days of the loan (actual or in increments of months),
and (5) other terms and conditions (collect half of the loan now, the other half
later; keeping a compensating balance). Rather than bombard you with a list of
interest rate conventions for different markets, we focus on the three basic methods:
simple, compound, and continuously compounded interest rates. Understanding
these concepts will lead to discounting and compounding, which will enable us to
transfer cash flows across time.
A simple interest rate is used in some sophisticated derivatives like caps and
swaps. Money is not compounded under simple interest rates. An annual rate is
quoted. When computed over several months, a fraction of the annual rate is used.
Compound interest is when interest is earned on both the original principal
and the accrued interest. Banks offer interest on daily balances kept in your account.
When compounding interest, divide the annual rate by the number of compounding
intervals (daily, weekly, monthly) and multiply these interest components together to
compute the loan value at maturity.
Continuous compounding pays interest on a continuous basis. One can view
continuous compounding as the limit of compound interest when the number of
compounding intervals gets very large and the time between earning interest gets
very small! Example 2.1 demonstrates these three methods of interest computation.
EXAMPLE 2.1: Simple, Compound, and Continuously Compounded Interest
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Rates
Simple Interest Rates
■ Suppose you invest $100 today, and 6 percent is the simple interest rate per year.
- A year from now, it will grow to $106.
- After six months, it will grow to 100 × [1 + 0.06(1/2)] = $103.
- After one day, it will be 100 × [1 + 0.06(1/365)] = 100.0164 = $100.02.
- After 250 days, it will be 100 × [1 + 0.06(250/365)] = $104.11.
■ Summoning the power of algebra, we replace numbers with symbols to develop a formula for money
growth under simple interest. Here L = $100 is the principal, which is the original amount invested.
It is invested at the rate of interest i = 6 percent per year for T years, where T is measured in years
or a fraction of a year. In the preceding expressions, T takes the values 1, 1/2, 1/365, and 250/365,
respectively. Consequently, under simple interest, L dollars invested at i percent per year becomes
after one year L(1 + i) and after T years L(1 + iT). This is illustrated in Figure 2.1. Notice that the
simple interest rate is also the rate of return. This is verified by plugging the values in relation to
Equation 2.2 to get (365/365)[L(1 + iT) − L]/L = i as T is one year.
28 CHAPTER 2: INTEREST RATES
Compound Interest Rates
■ What about interest earning interest? This leads to the notion of compounding, which involves
computing interest on both the principal and the accumulated interest. We split the time period
T over which we are investing our money into m intervals. Within each interval, money earns simple
interest. At the end of each interval, the principal and accumulated interest become the new principal
on which interest is earned.
If the interest is compounded semiannually, one year from now, $100 will grow to3
1 1
100 × [1 + 0.06 ( )] × [1 + 0.06 ( )]
2 2
1 2
= 100 × [1 + 0.06 ( )]
2
= $106.09
Compounding three times a year, one year from now, it will grow to
1 3
100 × [1 + 0.06 ( )] = $106.12
3
What about daily compounding that your local bank offers? With daily compounding, a year from
now, the amount invested will grow to
1 1
100 × [1 + 0.06 ( )] × [1 + 0.06 ( … 365 times
365 365 )]
365
1
= 100 × [1 + 0.06 (
365 )]
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
= $106.18
Now consider a loan for T years. Continuing with our example of daily compounding, after 250
days (or 250/365 year), $100 will grow to
1 1
100 × [1 + 0.06 ( × 1 + 0.06 ( … 250 times
365 )] [ 365 )]
250
1
= 100 × [1 + 0.06 (
365 )]
= $104.19
Rewriting the second line of the preceding expression with an eye toward generalization, we get
250
(365)(
1 365 )
100 × [1 + 0.06 ( )]
365
(The book’s appendix gives the rules underlying these manipulations)
■ As before, we replace numbers with symbols to develop a formula for computing compound interest.
Let L = $100 be the principal, which is invested at i = 6 percent per year for T years; however,
BASIC INTEREST RATES: SIMPLE, COMPOUND, AND CONTINUOUSLY COMPOUNDED 29
the interest is compounded m times every year. In the preceding expressions, m takes the values 2,
3, and 365, respectively. T = 1 year, except in the last example where it takes the value (250/365).
Consequently, under compound interest, L dollars invested for one year becomes L(1 + i/m)m , and
when invested for T years, it becomes (which is also shown in Figure 2.1)
i mT
L (1 +
m)
FIGURE 2.1: Simple, Compound, and Continuously
Compounded Interest Rates
Simple Interest
(Interest does not earn interest)
Time 0 1 year Time T
L L(1 + i) L(1 + iT)
Investment Value after 1 year Value at time T
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Compound Interest
(Interest earns interest)
Time 0 Time 1/m Time 2/m 1 year Time T
L L(1 + i/m) L(1 + i/m)2 L(1 + i/m)m L(1 + i/m)mT
Investment Value after 1 year Value at time T
Continuously Compounded Interest
(Interest earns interest continuously)
Time 0 1 year Time T
L Ler1 LerT
Investment Value after 1 year Value at time T
30 CHAPTER 2: INTEREST RATES
Continuously Compounded Interest Rates
■ But why stop at 365? Why not compound 1 million times a year? Then the investment grows to
$106.1836545 at year’s end. Two things are happening here. As the compounding frequency increases,
so does earned interest, but it grows at a decreasing rate—the money doesn’t grow very fast after a
while. For example, compounding 10 million times will only give $106.1836547 at year’s end.
■ We can talk about “continuous compounding,” where the interest rate r (we now use r instead of i
to denote the interest rate) is continuously compounded. This happens when the annual frequency
of compounding m becomes larger and larger but the interest rate charged r/m becomes smaller and
smaller during each of the compounding intervals. In that case, $1 becomes 1er at year’s end, where
the exponential function e is the base of the natural logarithm (mathematically, we define log x ≡
ln (x) = M as equivalent to eM = x). It is approximated as 2.7183. Calculators have this as ex or exp
( … ), denoting the exponential function.
■ But that was for one year. The dollar investment grows to erT after T years. To summarize, under
continuous compounding, L dollars invested at r percent per year for one year becomes Ler , and
when invested for T years, it becomes LerT . (The appendix shows how this happens, and Figure 2.1
illustrates the result in a diagram.) In our example, this will give at year’s end
100e0.06 = $106.1836547
Notice that the APR remains at 6 percent but the effective annual interest rate = er − 1 = 6.1837
percent.
This is an example of a continuously compounded return (or logarithmic return). Continuously
compounded interest rates are used as an input in the Black–Scholes–Merton model. Our example
suggests that continuous compounding (which gives $106.18 at year’s end) is a much better approximation
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
to daily compounding (which also gives $106.18) than is simple interest (which only gives $106).
3 The appendix discusses how to write expressions involving exponents (or indexes). It also gives rules for manipulating such
expressions. Notice that the APR remains 6 percent, but your investment will earn more than this at year’s end. This is captured
by the concept of an effective annual interest rate (EAR), which expresses the interest rate realized on a yearly basis. This is
given by (1 + i/m)m - 1, where m is the frequency of compounding. Here, because m is 2, the EAR is 0.0609 or 6.09 percent.
We collect these three interest computation methods as Result 2.2.
DISCOUNTING (PV) AND COMPOUNDING (FV): MOVING MONEY ACROSS TIME 31
RESULT 2.2
Simple, Compound, and Continuously Compounded
Interest Rates
Simple Interest
A principal of L dollars when invested for T years becomes
L (1 + iT) (2.3)
where i percent per year is the simple interest rate.
Compound Interest
L dollars invested for T years becomes
L[1 + (i/m)]mT (2.4)
where i percent per year is the compound interest rate and m is the number
of times the interest is compounded every year.
Continuously Compounded Interest
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
L dollars invested for T years becomes
LerT (2.5)
where r percent per year is the continuously compounded interest rate and e
is the exponential function.
2.4 Discounting (PV) and Compounding
(FV): Moving Money across Time
We can use interest rates to find the dollar return or to price a zero-coupon bond.
If we invest a dollar today and it continues to earn interest that we do not withdraw,
then the final amount on the maturity date is the dollar return. The dollar return
measures how an invested dollar grows over time. In real life, you can earn a dollar
return by investing in a money market account (mma), which earns the risk-free
rate in each period.
32 CHAPTER 2: INTEREST RATES
The inverse of a dollar return is the price of a zero-coupon bond. A zero-coupon
bond (or zero) sells at a discount, makes no interest (or coupon) payments over the
bond’s life, and pays back the principal (or the par value) on the maturity date.
A zero-coupon bond and a mma are closely related securities; the price of a zero-
coupon bond (B) and the dollar return (1 + R) obtained from investing in a mma are inverses
of one another, B ≡ 1/(1 + R). Both mmas and zeros help us move funds across time.
Example 2.2 shows how to compute a dollar return and the price of a zero.
EXAMPLE 2.2: Computing a Dollar Return, Pricing a Zero, and Moving Funds
across Time
■ Suppose that the simple interest rate is 6 percent per year. Today is time 0, and the bond matures at
time T = 1/2 year. Then the dollar return after six months of investing in a mma is
1 + R = 1 × [1 + 0.06 (1/2)] = $1.03
This dollar return is the future value of $1 invested today.
■ Today’s price of a zero-coupon bond that pays $1 after six months is
B ≡ 1/ (1 + R) = 1/1.03 = $0.9709 = $0.97
This is the present value of $1 received after six months.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
RESULT 2.3
Compounding and Discounting Cash Flows with Dollar
Returns and Zero-Coupon Bonds
Consider cash flows C(0) available today (time 0) and C(T) available at some
later date (time T). Then multiplication of today’s C(0) by the dollar return
(or equivalently, division by a zero-coupon bond price) gives the future value
C (0) (1 + R) = C (0) /B (2.6)
Multiplication of the future C(T) by a zero-coupon bond price (or division
by the dollar return) gives the present value (or discounted value)
BC (T) = C (T) / (1 + R) (2.7)
DISCOUNTING (PV) AND COMPOUNDING (FV): MOVING MONEY ACROSS TIME 33
where (1 + R) is the dollar return at time T from investing $1 today, B ≡ 1/(1
+ R) is today’s price of a zero-coupon bond that pays $1 at time T, and T is
the time to maturity in years. See Figure 2.2.
In the case of simple interest,
1 + R = (1 + i × T) (2.8)
where i is the simple interest rate per year.
In the case of continuously compounded interest,
1 + R = erT (2.9)
where r is the continuously compounded interest rate per year.
FIGURE 2.2: Compounding and Discounting Cash
Flows using the Dollar Return (1 + R) or
the Zero-Coupon Bond Price (B)
Now (Time 0) Future (Time T)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
C(0) C(0)(1 + R) = C(0)/B
BC(T) = C(T)/(1 + R) C(T)
You can easily extend this result to move multiple cash flows across time, which
may occur at different time periods (see Extension 2.2). Example 2.3 demonstrates
34 CHAPTER 2: INTEREST RATES
how to compound and discount when a continuously compounded risk-free
interest rate (r) is given.
EXAMPLE 2.3: Computing Present Values Using a Continuously Compounded
Rate
■ Suppose that we receive $100 after six months. Let the continuously compounded risk-free interest
rate r be 6 percent per year. Consequently, C(T) = $100, r = 0.06, and T = 0.5 year. Using expressions
(2.4b) and (2.4d), the present value (PV) of $100 is
100 e−0.06 × 0.5 = 97.0446 = $97.04
■ Using the notation defined previously, B = $0.970446. If you invest this at a continuously compounded
rate of 6 percent, you will get $1 in 6 months. Hence we can also write the PV of $100 in six months
as 100B.
■ Notice that (1 + R) = 1/B = 1.0304545. Consequently, expressions (2.4a) and (2.4d) give the future
value (FV) of $100 after six months as
100 (1 + R) = $103.05
Continuous compounding will be useful in powerful options pricing models like
the Black–Scholes–Merton model. If you are trying to compute options prices and
you know the market price of a zero, then, as explained in Example 2.4, you can
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
extract the risk-free continuously compounded rate and plug it into the Black–
Scholes–Merton model.
EXTENSION 2.2: Moving Multiple Cash Flows across Time
Moving cash flows across time is a fundamental tool that we use throughout the book. One can use interest rates,
zero-coupon bonds, and mma values to do this.
EXT. 2.2 EX. 1: Moving Multiple Cash Flows across Time
■ Suppose you graduate a year from now and expect to get a job from which you squirrel away $5,000 at year’s
end. Moreover, you expect a year-end bonus of $10,000. Assuming that we can commit “several sins”—(1)
disregard the basic principle that risky cash flows must be discounted by risky discount rates and (2) operate
under the assumption that the same interest rate applies to loans of different maturities—let us borrow against
these future cash flows and use the funds to help pay your college tuition.
DISCOUNTING (PV) AND COMPOUNDING (FV): MOVING MONEY ACROSS TIME 35
■ Your bank offers you a loan at 6 percent interest, compounded daily. Approximating this by continuous
compounding, we let r = 6 percent be the continuously compounded risk-free interest rate. Then the price
of a zero-coupon bond maturing in t = 2 years is given by B(2) = 1/ert = e-rt = e-0.06 × 2 = $0.8869 (rounded
to 4 decimal places). Writing C1 (2) = $5,000 and C2 (2) = $10,000, discounting the individual cash flows
and adding them up gives B(2)C1 (2) = 0.8869 × 5,000 = $4,434.60 and B(2)C2 (2) = 0.8869 × 10,000 =
$8,869.20, whose sum is
B (2) C1 (2) + B (2) C2 (2) = $13, 303.81 (2.10)
■ Alternatively, you can add up the cash flows and then discount by multiplying by the zero-coupon bond price:
B (2) [C1 (2) + C2 (2)] = 0.8869 × (5, 000 + 10, 000) = $13, 303.81 (2.11)
Both approaches give the same answer. In the actual calculations, B(1) and B(2) are not rounded to 4 decimal
places (see Extension 2.1).
■ The equality of expressions (1) and (2) can be generalized to yield a formula for discounting cash flows
belonging to a particular time period, t = 2:
B (2) C1 (2) + B (2) C2 (2) = B (2) [C1 (2) + C2 (2)] (2.12)
You can express this compactly by using the summation sign (∑). Then expression (3) can be written as
2 2
B (2) Cs (2) = B (2) Cs (2) (2.13)
[∑
s=1 ] ∑s=1
■ What about extending this result to cash flows available at different time periods? Continuing with our previous
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
example, suppose your grandparents have been saving for your college education by investing in a tax-favored
education savings account. Your paternal grandparents will gift you C1 (1) = $12,000 and your mother’s parents
will give you C2 (1) = $10,000 at time t = 1. Assuming that the interest rate r is 6 percent, and the price of
a zero-coupon bond maturing in one year is B(1) = e-0.06 × 1 = $0.9418, the present value of these cash flows is
B (1) [C1 (1) + C2 (1)] = 0.9418 × (12, 000 + 10, 000) = $20, 718.82
■ Suppose you want to determine how much of a loan you can take out today based on these future cash flows.
For this you need to compute the present value of the four cash flows, two after one year and two after two
years. They have the same value, $34,022.63, irrespective of how you add them up:
B (1) C1 (1) + B (1) C2 (1) + B (2) C1 (2) + B (2) C2 (2)
= B (1) [C1 (1) + C2 (1)] + B (2) [C1 (2) + C2 (2)]
■ A generalization yields our next result.
36 CHAPTER 2: INTEREST RATES
RESULT 2.4
The Sum of a Present Value of Cash Flows Is Equal to the Present Value of
the Sum
Suppose that there are S securities, which provide s cash flows (s = 1, 2, … , S ), some of which could be
zero. These cash flows can occur at times t (t = 1, 2, … ,T), so that the cash flow from security s at time t
is Cs (t). Then the sum of the present value of each cash flow is equivalent to (1) adding up the cash flows
from all securities at a particular time t, (2) computing the PV of this sum, and (3) adding up these cash
flows across time. This can be expressed as
T S T S
B (t) Cs (t) = B (t) Cs (t) (2.14)
∑∑ ∑ [∑ ]
t=1 s=1 t=1 s=1
where B(t) is the price of a zero-coupon bond that pays $1 at time t.
EXAMPLE 2.4: Finding the Continuously Compounded Rate from a
Zero-Coupon Bond
■ Suppose a zero-coupon bond that is worth $0.90 today pays $1 after two years. Assuming that money
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
grows at the continuously compounded rate of r percent per year, we have
0.90e2r = 1
or, e2r = 1/0.90
■ Taking natural logarithms and remembering that log(eX ) = X, 2r = log (1/0.90)
or r = (1/2) log (1/0.90) = 0.0527 or 5.27 percent
Replacing numbers with symbols in the preceding expression (2 = T and 0.90 =
B = 1/[1 + R]) gives us Result 2.4, which allows us to move from a notional variable
(a posted interest rate) to a traded asset price (a zero-coupon bond or a dollar return),
and vice versa.
US TREASURY SECURITIES 37
RESULT 2.5
Computing Continuously Compounded Interest Rates
from a Zero-Coupon Bond Price or a Dollar Return
The continuously compounded interest rate per year r can be computed from
a zero-coupon bond price B (or a dollar return 1 + R) by
r = (1/T) log (1/B) = (1/T) log (1 + R) (2.15)
where T is the time to maturity in years.
Now that you have seen how to price zero-coupon bonds (B), move funds across
time (T), find dollar returns (1 + R), and compute continuously compounded interest
(r), you may wonder where zero-coupon bond prices come from. This brings us to
the Treasury securities market, where debts of the US government trade.
2.5 US Treasury Securities
Modern finance has made Treasury securities (Treasuries) interesting and exciting
by developing a whole range of derivatives that depend on them. Although seemingly
ordinary, Treasuries are the bedrock on which the world of finance is built. Following
are five reasons why:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1. Five characteristics make Treasuries particularly important: (1) these debt securities
have no default risk as they are backed by the full faith and credit of the US
government; (2) they trade in a market with some of the smallest bid/ask spreads
in the world, which is the transaction cost of buying and selling Treasuries (only
1 or 2 basis points for the most active issues, a basis point being 1/100th of 1
percent); (3) their interest payments are free from state and local taxes; (4) they
have low minimum denominations starting at $100; and (5) they offer a spectrum
of maturities that range from one day to thirty years.
2. The prices of Treasuries can help us determine borrowing and lending rates for
future dates. What is the forward rate for borrowing money for two years starting
10 years from today? As Chapter 21 shows, the answer can be easily determined
from Treasury prices.
3. We will study options and futures that are written on Treasuries. Remember the
T-bond futures that we mentioned in the first chapter? They are one of the most
actively traded futures contracts. The world has become a more volatile place, and
38 CHAPTER 2: INTEREST RATES
derivatives based on fixed-income securities (bonds) can be very useful for hedging
interest rate risks.
4. Interest rate options pricing models like the Heath–Jarrow–Morton model relax
the constant interest rate assumption of the Black–Scholes–Merton model. They
often use Treasury rates of different maturities as the necessary inputs.
5. Moreover, in part of a growing trend, financial firms convert many individual loans
and debts into a package of securities using a process called securitization and
sell them to third-party investors. Your mortgage loans, car loans, and credit card
loans may have been financed that way. Loan access to wider markets generates
more competitive rates that benefit consumers. The interest rates on these asset-
backed securities are determined by adding a basis point spread to comparable
maturity Treasury rates.
2.6 US Federal Debt Auction Markets
When former US president Ronald Reagan was asked for his thoughts on the
growing federal deficit, he replied, “The deficit is big enough to take care of itself!”
No matter which party held power in Washington, D. C., the last several decades have
seen government spending exceed taxes. Years of budget deficits (= Government
receipts – Government expenditures) have led to a mountain of federal debt,
standing at over $13 trillion at the end of the millennium (Figure 2.3 shows the
level of public debt on an annual basis). This is hardly surprising. There is a natural
tendency toward overspending in a democracy because elected politicians bring
pork-barrel projects to help their constituencies.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Whatever the reason, the US national debt and deficit became alarmingly large
in the 1990s and took center stage on the political arena. US citizens interpreted
continuing deficits as implying higher future taxes. Supply-side tax cuts of Reagan’s
era lost their appeal, and federal income tax rates were raised immediately. A robust
economy also increased tax receipts. Consequently, the US government started paying
off its debts at the turn of the new millennium. But the surpluses soon evaporated,
and the budget deficit is back in the red.
As circumstances shape destiny, a huge federal debt has forced the Treasury to
devise efficient ways of raising funds. Consequently, the United States has developed
an extremely sophisticated system of selling Treasuries through sealed bid auctions,
a model that has been adopted by many other nations. The buyers in such auctions
submit sealed bids. These are promises to buy a fixed number of Treasuries at a fixed
price. The Treasury awards the securities to the highest bidders. In reality, these are
yield auctions, in which the bidders specify a yield and the Treasury translates this
interest rate into a price. Every year, the Treasury finances the public debt through
over 250 auctions, each typically selling $10 to $20 billion worth of securities.
US FEDERAL DEBT AUCTION MARKETS 39
FIGURE 2.3: U. S. Debt Outstanding
20,000
US Federal 18,000
Debt (in billions
of dollars) 16,000
14,000
12,000
10,000
8,000
6,000
4,000
2,000
0
77
79
81
83
85
87
89
91
93
95
97
99
01
03
05
07
09
11
13
15
19
19
19
19
19
19
19
19
19
19
19
19
20
20
20
20
20
20
20
20
Time
Source: http://www.treasurydirect.gov/govt/reports/pd/histdebt/histdebt.htm.
Primary dealers are large securities firms with whom the New York Fed buys
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and sells Treasuries to conduct open-market operations that fine-tune the US money
supply, examples include Goldman Sachs & Co. LLC, UBS Securities LLC, and
Deutsche Bank Securities Inc. These firms also actively bid in Treasury auctions, as
do other direct and indirect bidders. These bidders tread cautiously because bidding
incorrectly, even several basis points, can make a difference between millions of dollars
in profits and losses. So players in the auction submit bids as near to the bidding
deadline as possible using all available information.
Do all investors in Treasuries take part in competitive bidding? No, because the
Treasury allows noncompetitive bids for smaller amounts. These bids are always filled,
and they pay the price determined by the competitive bidders. Historically, 10 to
20 percent of a typical Treasury auction has been awarded to noncompetitive bidders.
Noncompetitive bids encourage the direct participation of regular folks in the auction
process.
The quotes from competitive and noncompetitive bids are tallied, securities are
allocated to the successful bidders, and the results of the auction are announced in
a press conference. The Wall Street Journal and the other news providers carry the
announcement.
40 CHAPTER 2: INTEREST RATES
2.7 Different Ways of Investing in
Treasury Securities
Bidding in auctions is not the only way to purchase Treasury securities. For example,
you can buy Treasuries (1) in the when-issued and secondary markets, (2) through
the repo and reverse repo markets, or (3) via interest rate derivatives. This section
describes each of these alternative ways of buying Treasuries. Barring rare exceptions,
these markets tend to be fairly liquid and are usually dominated by big institutional
investors (institutions such as commercial banks, investment banks, pension funds,
mutual funds, and hedge funds).
The Treasury Auction and Its Associated Markets
A week or so before an auction, the Treasury announces the size of the offering, the
maturities, and the denominations of the auctioned Treasuries. The Treasury permits
forward trading of Treasury securities between the announcement and the auction,
and the to-be-auctioned issue trades when, as, and if issued. Traders take positions
in this when-issued market, and a consensus price emerges. The when-issued market
spreads the demand over seven to ten days, which leads to a smooth absorption of
the securities by the market. The buyer and seller in the when-issued market fulfill
their commitments after the Treasuries become available through the auction.
Newly auctioned Treasuries are called on-the-run and have lower spreads than
off-the-run securities issued in prior auctions. A very active resale (secondary)
market exists for Treasuries, giving investors further chances to invest. The timeline
for the three markets for buying Treasuries is shown in Figure 2.4.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
FIGURE 2.4: Timeline of Treasury Auction and
Associated Markets
Treasury Auction Settlement
announcement
When-issued market Post-auction secondary market
DIFFERENT WAYS OF INVESTING IN TREASURY SECURITIES 41
The Repo and the Reverse Repo Market
Investment banks operate a security dealership business that makes profits by trading
standard securities as well as custom-made products for clients. They do this by highly
leveraging both the asset and liability sides of their balance sheets (see Chapter 1 for a
typical firm’s balance sheet). For example, believing that buying Treasuries will gain
in value and borrowing Treasuries will decline in value, the firm may support $100
billion worth of purchased securities and $95 billion worth of borrowed securities
using only $5 billion of capital. Leverages like this are usually accomplished through
repo and reverse repo markets, examples of which are given in Extension 2.3.
Partially contributing to the immense popularity of repos was the fact that they
initially provided the only method of short selling Treasuries and corporate debt
securities (see Extension 2.3). This short selling technique became antiquated when
interest rate derivatives arrived. Derivatives facilitate the same transaction but far more
efficiently. We will explore how this works in the final part of the book. Although
this particular economic role of repos is now less important, the repo market still
thrives and is an important element of Treasury markets.
EXTENSION 2.3: A Repurchase Agreement
Suppose that Repobank has purchased a large quantity of Treasury securities and it needs an overnight loan to
finance the purchase. Repobank expects to sell the Treasuries on the market the next day, so an overnight loan is
sufficient. To finance its position, it enters into a repurchase agreement (“does a repo”) with RevRepobank
(another fictitious name), who enters into a reverse repo (“does a reverse”) transaction. The next example
outlines the mechanics of this transaction in simple terms.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
EXT. 2.3 EX. 1: Example of a Repo and a Reverse Repo
■ As per standard industry norms, Repobank takes, say, $10 million from RevRepobank and sells RevRepobank
Treasury securities worth a little more. The next day, Repobank repurchases those securities at a slightly higher
price—the extra amount determines an annual interest rate known as the repo rate. Though it involves a
purchase and a repurchase, a repo is basically a short-term loan that is backed by high-quality collateral (see
Ext. 2.3 Fig. 1).
■ If Repobank defaults, then RevRepobank keeps the securities. If RevRepobank fails to deliver the securities
instead, then Repobank keeps the cash longer. In that case, the repo is extended by a day, but the terms remain
the same—Repobank pays the day after tomorrow the amount it was supposed to pay tomorrow, essentially
keeping the funds for an extra day at zero interest. And this is repeated day after day until the cash and the
securities are exchanged. You may find that holders of scarce securities are able to borrow funds in the repo
market at close to a zero percent rate of interest.
■ Repos provide a way of short selling Treasuries and corporate debt securities. After acquiring the securities,
RevRepobank can short sell them to a third party. Later, it can buy these securities in the market, at a lower
price if the bet is successful, and return them to Repobank and close out the repo.
42 CHAPTER 2: INTEREST RATES
■ Entering into a repo agreement is equivalent to borrowing cash and using the short maturity Treasury security
as collateral. The repo rate is the effective borrowing rate for Repobank and the lending rate for RevRepobank.
■ Repos are usually for overnight borrowing and are known as overnight repos. If Repobank wants to borrow
funds for another night, then the whole process must be repeated. Alternatively, term repos are set up for
a fixed period that lasts longer than overnight, whereas open repos have no fixed maturity and may be
terminated by a notice from one of the sides. The market is huge. Major players include heavyweights like
the Federal Reserve Bank, state and local governments, commercial and investment banks, mutual funds, and
large companies.
EXT. 2.3 FIG. 1: An Overnight Repo and a
Reverse Repo Transaction
Starting Date (Today)
$10 million
Repobank RevRepobank
T-securities worth $10 million
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Ending Date (Tomorrow)
$10(1 + interest) million
Repobank RevRepobank
T-securities worth $10 million
TREASURY BILLS, NOTES, BONDS, AND STRIPS 43
Interest Rate Derivatives
Derivatives whose payoffs depend on Treasury (or closely related) securities are called
interest rate derivatives.4 There are derivatives like futures on Treasuries and
options on futures on Treasuries that allow investors to place leveraged bets or set
up hedges on interest rates. We will discuss interest rate derivatives and their pricing
models in Part IV of the book.
2.8 Treasury Bills, Notes, Bonds, and
STRIPS
The Treasury sells three types of marketable debts: (1) coupon bonds that pay
interest (coupons) every six months and a principal amount (par or face value) at
maturity, (2) zero-coupon bonds that don’t pay interest but pay back the principal at
maturity, and (3) floating rate notes that pay interest quarterly. The United States
issues bonds with maturities of one year or less in the form of zero-coupon bonds
and calls them Treasury bills (or T-bills). Coupon bonds have two names: those
with original maturity of two to ten years are called Treasury notes (or T-notes),
whereas those with original maturity of more than ten years up to a maximum of
thirty years are called Treasury bonds (or T-bonds). These notes and bonds make
coupon and principal payments that remain fixed over the security’s life. Treasury
notes of two year maturity, which pay interest quarterly based on the 13 week T-bill
rate also are issued by the Treasury.
There is an exception. In 1997, the Treasury started selling a new class of securities
called inflation-indexed bonds or TIPS (Treasury Inflation Protected Securities).
These bonds guarantee a fixed real rate of return, which is the nominal rate of return
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
in dollar terms minus the inflation rate as measured by the CPI over their life. This
is accomplished by raising the principal of the bond each year by changes in the US
CPI. Each year, the coupon payment is computed by multiplying the adjusted (and
changing) principal by the real rate of return.
The name “coupon” may sound strange to your ears. In days gone by, bonds had
coupons attached to them. The bearer of the bond detached the coupon and sent it
to the issuer, who mailed back interest payments. Coupons are rare these days. In fact,
since 1983, all Treasury securities are kept in book entry form at the Federal Reserve
Bank’s computers, and the owner is given a receipt of ownership and receives deposits
directly from the Fed. But the term coupon survives.
The Treasury does not issue zeros of long maturities. Still, a pure discount bond
with a single payment at maturity appeals to many investors. Wall Street firms figured
out a way of making money by selling what people demanded. Wall Street firms
bought coupon-bearing Treasury securities, put them in a trust to make them safe,
and issued claims against the principal and the different coupon payments. This was
an arbitrage opportunity: the firms paid less for the original Treasury than what they
4
Closely related to Treasury securities are Eurodollar deposits, which are discussed later in this chapter.
44 CHAPTER 2: INTEREST RATES
collected by selling the artificially created zero-coupon bonds. In February 1985, the
Treasury entered this activity by allowing Treasury securities to be STRIPped.
How do STRIPS (Separate Trading of Registered Interests and Principal of
Securities) work? Although the Treasury does not issue or sell STRIPS directly to
investors, it allows financial institutions as well as brokers and dealers of government
securities to use the commercial book-entry system to separate a Treasury note or a
bond’s cash flows into strips and sell them as individual zero-coupon bonds. Claims to
individual cash flows coming out of a Treasury security are synthetically created zero-
coupon bonds of different maturities. Moreover, an investor can buy up the individual
strips from the market and reconstruct the original T-bond or note. Figure 2.5 shows
how thirty-one strips can be created from a newly issued fifteen-year T-bond.
A Treasury security can be stripped at any time from its issue date until its maturity
date. For tracking purposes, each cash flow due on a certain date is assigned a unique
Committee on Uniform Security Identification Procedures (CUSIP) number that
depends on its source—consequently, they are categorized as coupon interest (ci),
note principal (np), and bond principal (bp). Even though they may come from
different Treasury securities, all ci due on a certain date have the same generic CUSIP
and may be combined to create a coupon paying note or a bond. By contrast, np or
bp is unique for a particular security and hence is not interchangeable.
The STRIPS program had been very successful:
1. STRIPS made Treasuries more attractive to investors, leading to greater demand,
higher prices, lower yields, and cheaper financing of the national debt. For
example, not many people would be interested in holding a thirty-year bond with
sixty cash flows. But a newly issued thirty-year bond can be stripped into sixty
semiannual coupon payments and a final principal payment. These cash flows can
be sold to different investors needing zeros of different maturities. For example,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
grandparents of a newborn may gift a strip for future education costs.
2. STRIPS help to identify the term structure of interest rates, a graph that plots
the interest rate on bonds against the time to maturity. Earlier in this chapter, we
talked about the benefits of knowing the risk-free rates of different maturities, an
important input to the pricing of interest rate derivatives studied in Part IV of the
book.
Once we get the price of a Treasury security, we know the cash flows that it
generates. Particularly useful are T-bill and STRIPS prices for moving cash flows
across time.
TREASURY BILLS, NOTES, BONDS, AND STRIPS 45
FIGURE 2.5: Creation of STRIPS from a
15-Year Treasury Bond
Newly Issued 15-Year T-Bond
Payments C C C … … … C C and L
Time (years) ½ 1 1½ 14½ 15
C denotes coupons and L principal. Holder gets all the cash flows.
STRIP Creation
Payments C C C … … … C C and L
Time (years) ½ 1 1½ 14½ 15
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
C and L sold as separate zero-coupon bonds (STRIPs).
Prices of Treasuries can be found in financial newspapers like the Wall Street Journal
and Investor’s Daily. The quotes are collected from the over-the-counter market and
are for transactions of $1 million or more. Notes, bonds, and strips are quoted in
32nds, so a quote of 93:08 (for bid) and 93:09 (for ask) for a strip maturing in one
year means that $93 and 8/32 is the price the dealer will buy and $93 and 9/32 is the
price the dealer will sell. These quotes are based on a par value equal to $100.
T-bill prices follow an entirely different convention. They are quoted in terms of
a banker’s discount yield. Before introducing this notion, it may be useful to see how
the rate of return and the bond-equivalent yield are determined.
When the annualized rate of return (see Result 2.1) is applied to Treasury bills, we get
its bond-equivalent yield. This is reported as the “ask yield” in the Wall Street Journal,
which corresponds to the bond-equivalent yield earned when buying a T-bill from a
46 CHAPTER 2: INTEREST RATES
dealer. The bond-equivalent yield for Treasury bill prices is
365 (1 − B)
Bond equivalent yield = ( (2.16)
T ) B
where B is the bill’s price expressed as a bond that pays $1 at maturity and T is the
number of days from settlement to maturity.
However, Treasury bills are quoted in terms of a banker’s discount yield, which
is typically reported for both ask and bid prices. Ask is the price at which one can
readily buy a security from a dealer, and bid is the price at which one can promptly
sell. They are analogous to prices at which a car dealer will sell or buy a used car.
The banker’s discount yield differs from the bond-equivalent yield in two ways: (1)
the denominator has the face value instead of the price paid (note the 1 in the
denominator) and (2) the interest rate is annualized on a 360-day basis. The banker’s
discount yield is given by
360 (1 − B)
Banker’s discount yield = ( (2.17)
T ) 1
T
or Bill price, B = 1 − (banker’s discount yield) (2.18)
[ ( 360 )]
where 1 is the face value (par value) of the bill, B is bill price, and T is the number
of days from settlement (the date ownership is transferred, time 0) to maturity.
Example 2.5 applies these formulas to data from the financial press.
EXAMPLE 2.5: T-Bill Price Computation
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ The financial press reports the following:
- Days to maturity 24
- Bid 4.71 percent
- Ask 4.67 percent
- Change 0.01
- Ask yield 4.75 percent
■ Using expression (2.7b) and assuming that the bill’s principal is $1, the ask price for the bill is
= 1 − 0.0467 × (24/360)
= $0.996887
In other words, $0.996887 is the ask price of a zero-coupon bond that pays $1 in twenty-four days.
Similarly, $0.996860 is the bid price of a zero that pays $1 in twenty-four days.
LIBOR VERSUS A LIBOR RATE INDEX 47
■ How much are we actually earning on our investment? This is given by the bond-equivalent yield
(reported as ask yield because it is computed from the ask price). Using expression (2.6)
Bond − equivalent yield (for ask yield)
= (365/24) × [(1/0.996887) − 1]
= 0.0475
This verifies the quote.
The inverse relationship between bond prices and interest rates leads to an
inversion of the relation between the ask/offer and bid for interest rates. In most
markets, the ask/offer is higher than the bid, but that’s not the case for the ask and
the bid when we quote them as a banker’s discount yield. Using the numbers from
Example 2.4, this situation is shown in Figure 2.6.
FIGURE 2.6: Inversion of Ask and Bid When
Quoted as Banker’s Discount Yield
(Inverted–bid above ask) (As in most markets–ask above bid)
Bid (as bdy) Ask price
4.71 percent $0.996887
Ask (as bdy) Bid price
4.67 percent $0.996860
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The banker’s discount yield is important because it helps price a T-bill. Using
the face value in the denominator or 360 days in the computation doesn’t have any
economic significance. It’s just the way pricing conventions have developed in this
market. Using 360 days for quoting interest rates and using 1 in the denominator (as
in Equation 2.2) just make it easier to compute.
2.9 Libor versus a Libor Rate Index
Though US Treasury securities are fundamental to the international bond markets,
there are other interest rates that are keenly followed and widely used. In particular,
interest rates earned on dollar deposits in non-US banks are such an example. These
rates are related to the London Interbank Offered Rate (libor, pronounced lie-bore),
which is the interest rate used to determine cash payments on most swap contracts in
the swap market. Swaps are discussed in Chapters 7 and 22.
To understand how libor gets created, let us peek at the London interbank market.
Owing to their business activities, banks and other financial institutions face deficit
or surplus fund situations on a regular basis. Banks in the United States manage
48 CHAPTER 2: INTEREST RATES
their cash by borrowing or lending in the interbank federal funds market. Major
London banks handle those imbalances by borrowing or lending deposits of different
maturities in the London interbank market. The most important of these deposits are
Eurodollars, which, as you may recall from Chapter 1, are US dollars held outside
the United States in a foreign bank or a subsidiary of a US bank. Owing to the dual
benefit of being dollar deposits free from US jurisdiction, Eurodollars have gained
huge popularity among a whole range of holders, including central banks, financial
institutions, companies, and retail individual investors.
Now a bank with surplus funds lends the funds to another bank for a fixed time
period at the libor valid for that period. This is a Eurodollar rate (say, 5.02 percent
per year for a three-month deposit) in case of Eurodollars. Alternatively, a bank
pays the London Interbank Bid Rate (libid, pronounced lie-bid) for the privilege of
borrowing money: three-month libid could be 5.00 percent for Eurodollars. These
rates may change minute by minute, and they may vary from bank to bank, but
competition ensures that they are almost nearly the same at any given point in time.
The Intercontinental Exchange (ICE) collects libor quotes from its panel of major
banks for deposit maturities ranging from overnight to a year and computes a Libor
rate index that is widely reported. For Eurodollar deposits, the ICE collects libor
quotes from numerous banks, truncates some of the largest and smallest values, and,
averages the rest to compute a libor rate index. It does a similar exercise for deposits
denominated in many other currencies, announcing a term structure of Libor rate
indexes for each different currency.
This libor rate index is the most popular global benchmark for short-term interest
rates, and it enters into numerous derivative contracts. An index is used instead of a
particular bank’s Eurodollar rate because an index is less prone to manipulation. As
the libor rate index has some credit risk, its values are larger than a similar maturity
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Treasury security, the difference being known as the Treasury/Eurodollar (TED)
spread. Instead of Treasuries, many large financial institutions use the libor rate index
as a proxy for the risk-free interest rate in derivative pricing models. Like Treasury
securities, Eurodollar interest computations assume 360 days in a year.
The Eurodollar market is now measured in trillions of dollars. In later chapters,
we will study the vast interest rate derivatives market that uses Eurodollars as the
underlying commodity.
2.10 Summary
1. An interest rate is the rate of return earned on money borrowed or lent. The
annualized rate of return on an investment is given by
365
Rate of return = (
T )
Selling price + Income − Expenses − Buying price
×
( Buying price )
365 Profit or loss
=( ×
T ) ( Buying price )
SUMMARY 49
where T is the time interval, measured by the number of days over which the
investment is held, and Income and Expenses denotes any positive or negative
cash flows, respectively, from the investment during this period.
2. There are many ways of computing interest. For a certain quoted rate, the realized
interest can vary with (1) the method of compounding, (2) the frequency (yearly,
monthly, daily, or continuously compounded), (3) the number of days in the year
(52 weeks; 360 days, or 365, or 366 for a leap year), and (4) the number of days
of the loan (actual or in increments of months).
3. The three basic methods of computing interest are simple, compound, and
continuous compounding. In simple interest, an annual rate is used for the loan
duration. When computed over several months, a fraction of that annual rate is
used. In case of compound interest, interest is computed on both the original
principal as well as the accrued interest. Continuous compounding involves
paying interest on the principal and accrued interest on a continuous basis.
4. Interest rates help us find the price of a zero-coupon bond (zeros). These bonds
sell at a discount and pay back the face value at maturity. They pay no interest
otherwise. Multiply any dollar amount by a zero-coupon bond’s price to get its
present value, or divide the amount by it to get its future value.
5. US Treasury securities, whose payments are backed by the taxing power of the US
government, are considered default-free. They help us determine the risk-free
interest rate. They are sold through a sealed-bid auction, in which all successful
bidders pay the same price as the lowest successful bid. One purchases a Treasury
security by (1) trading in the auction and its associated markets, (2) entering
into a repurchase (repos are short-term debts collateralized by high-quality debt
securities) or a reverse repo transaction, or (3) investing in interest rate derivatives.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
6. Marketable Treasury securities (Treasuries) are classified as bills, notes, and bonds.
Treasury bills are zero-coupon bonds with original maturity of one year or less.
Coupon bonds have two names: those with original maturity of two to ten years
are called Treasury notes (or T-notes), whereas those with original maturity of
more than ten years up to a maximum of thirty years are called Treasury bonds
(or T-bonds).
7. Financial institutions and brokers and dealers of government securities can create
zero-coupon bonds by segregating cash flows of Treasury notes and bonds and
sell them to the public. These artificially created zeros are called STRIPS.
8. In the bond markets, the annualized rate of return is called the bond-equivalent
yield. In case of Treasury bills,
365 (1 − B)
Bond equivalent yield = (
T ) B
where B is the price of a T-bill that pays $1 at maturity and T is the number of
days from settlement (the date ownership is transferred, time 0) to maturity. But
50 CHAPTER 2: INTEREST RATES
Treasury bills are quoted in terms of a banker’s discount yield, which is given by
360 (1 − B)
Banker’s discount yield = (
T ) 1
9. Besides Treasuries, major interest rates in the global financial markets include
libor and repo rates. Libor and libid are the respective rates for the banks to
lend and borrow surplus funds in the London interbank market. The ICE (1)
collects libor quotes from major banks for deposits denominated in many different
currencies, (2) for each currency it collects quotes with maturities ranging from
one day to one year, (3) computes a trimmed mean by dropping some of the
highest and lowest rates, and (4) releases these rates to the market, creating a
Libor rate index.
10. A libor rate index for Eurodollar deposit constructed by the ICE (which are
dollar-denominated deposits held outside the US and free from US banking laws
and regulations) has emerged as the most popular global benchmark for short-
term interest rates and enters into numerous derivatives contracts. Many large
financial institutions are replacing Treasuries with bbalibor as a proxy for the
risk-free interest rate. Many floating interest rates are set at a spread above the
relevant Treasury security rate or the Libor rate index.
2.11 Cases
Breaking the Buck (Harvard Business School Case 310135-PDF-ENG). The case
educates students about how money market funds work and the challenges faced
in managing these funds during the financial crisis of 2008–9.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Foreign Ownership of US treasury Securities (Darden School of Business Case
UV1366-PDF-ENG, Harvard Business Publishing). The case describes the work-
ings of the US Treasury securities market and explores the implications of
significant foreign holdings of US debt.
Innovation at the Treasury: Treasury Inflation-Protection Securities (A)
(Harvard Business School Case 204112-HCB-ENG). The case explores the
various risks as well as policy issues surrounding the introduction of a new
financial security product.
2.12 Questions and Problems
2.1. The interest rate is 5 percent per year. Compute the six-month zero-coupon
bond price using a simple interest rate.
2.2. The interest rate is 5 percent per year. Compute the six-month zero-coupon
bond price using a compound interest rate with monthly compounding.
2.3. The interest rate is 5 percent per year. Compute the six-month zero-coupon
bond price using a continuously compounded interest rate.
QUESTIONS AND PROBLEMS 51
Time (in years) Cash Flows (in dollars)
0 (today) 105
1 7
2 9
3 108
2.4. The interest rate is 5 percent per year. Compute the six-month zero-coupon
bond price using a banker’s discount yield (the zero-coupon bond is a US T-bill
with 180 days to maturity).
2.5. What is a fixed-income security? The next three questions are based on the
following table, where the interest rate is 4 percent per year, compounded
once a year.
2.6. Compute the present value of the preceding cash flows.
2.7. Compute the future value of the preceding cash flows after three years.
2.8. What would be the fair value of the preceding cash flows after two years?
2.9. If the price of a zero-coupon bond maturing in three years is $0.88, what is
the continuous compounded rate of return?
2.10. What are the roles of the primary dealers in the US Treasury market?
2.11. What is the when-issued market with respect to US Treasuries? What role does
this market play in helping the US Treasury auction securities?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
2.12. What is the difference between on-the-run and off-the-run Treasuries?
2.13. What is a repurchase agreement? Explain your answer with a diagram of the
transaction.
2.14. What is a Treasury STRIPS? What benefits do the trading of Treasury STRIPS
provide?
2.15. Explain how a libor rate index is computed by the ICE.
2.16. What is a Eurodollar deposit, and what is a TED spread?
2.17. What is the difference between Treasury bills, notes, and bonds? What are
TIPS, and how do they differ from Treasury bills, notes, and bonds?
2.18. You bought a stock for $40, received a dividend of $1, and sold it for $41 after
five months. What is your annualized arithmetic rate of return?
2.19. Suppose that you are planning to enroll in a master’s degree program two years
in the future. Its cost will be the equivalent of $160,000 to enroll. You expect
to have the following funds:
■ From your current job, you can save $5,000 after one year and $7,000 after
two years.
52 CHAPTER 2: INTEREST RATES
■ You expect a year-end bonus of $10,000 after one year and $12,000 after
two years.
■ Your grandparents have saved money for your education in a tax-favored
savings account, which will give you $18,000 after one year.
■ Your parents offer you the choice of taking $50,000 at any time, but you
will get that amount deducted from your inheritance. They are risk-averse
investors and put money in ultrasafe government bonds that give 2 percent
per year.The borrowing and the lending rate at the bank is 4 percent per
year, daily compounded. Approximating this by continuous compounding,
how much money will you need to borrow when you start your master’s
degree education two years from today?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
3
Stocks
3.1 Introduction Alternative Trading Systems:
Dark Pools and Electronic
3.2 Primary and Secondary Communications Networks
Markets, Exchanges, and
Over-the-Counter Markets 3.7 Dollar Dividends and
Dividend Yields
3.3 Brokers, Dealers, and Traders
in Securities Markets 3.8 Short Selling Stocks
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
3.4 Automation of Trading 3.9 Margin—Security Deposits
That Facilitate Trading
3.5 The Three-Step Process
EXTENSION 3.1 Margin and
of Transacting Exchange-
Stock Trading
Traded Securities
3.10 Summary
3.6 Buying and Selling Stocks
Trading at the New York Stock 3.11 Cases
Exchange 3.12 Questions and Problems
Over-the-Counter Trading
54 CHAPTER 3: STOCKS
3.1 Introduction
Institutions that play similar economic roles can differ widely in appearance. A
farmer’s market in a remote African village and a fancy shopping mall in a North
American suburb are both markets. MBA degree–holding sellers of customized
derivatives products and peddlers of Oriental rugs are both dealers. The man at
the airport who helps you find a cheap hotel near an exotic beach and the realtor
who arranges the purchase of a vacation home are both brokers. As these examples
illustrate, seemingly diverse economic phenomena often have commonalities.
This leads to this chapter’s theme—“unity in diversity”; here we try to understand
the features common to different securities markets and their traders.1 We introduce
primary and secondary markets, exchanges and over-the-counter markets, brokers
and dealers, and the bid and ask prices that dealers post. We also discuss market
microstructure, a subfield of finance that studies how market organization and traders’
incentives affect bid/ask spreads. We classify traders into different categories on the
basis of their trading strategies. We describe how automation is transforming trading
and the three-step process of execution, clearing, and settlement that is common
when transacting exchange-traded securities. Next, we study the effect of dividends
on stock prices and portfolios, paving the way for managing a portfolio that replicates
a stock index and pricing derivatives written on it. Finally, we illustrate the process
of short selling stock and margins.
The process of trading securities, the players who trade, and the facilitators who
make trading possible are also present in the bond markets discussed in Chapter 2.
However, we discuss them with respect to stock markets first because these are the
easiest markets to understand—trading of different derivatives can also be seen as
extensions of this process.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
3.2 Primary and Secondary Markets,
Exchanges, and Over-the-Counter
Markets
As a small company grows, it needs more and more cash to finance its operations.
After initial financing from founders and friends, and perhaps private equity from
venture capitalists, the company becomes ready to tap the capital markets. A privately
held company sells shares to the public for the first time in an initial public
offering (IPO), which is the primary market for stocks. To issue stocks, you hire
an investment banker who manages the process. Through cumbersome paperwork,
publishing information brochures, and getting approval from government agencies,
the investment banker burns a lot of your future proceeds in the process.
Soon after their birth in primary markets, stocks move on to secondary markets,
where they change ownership through secondhand transactions. Exchanges and
the over-the-counter (OTC) market are the two basic types of secondary markets.
1
The phrase “Unity in Diversity” is the official motto of the European Union (“United in Diversity”),
Ghana, Indonesia, South Africa, and some other nations; it has also been used to describe India.
PRIMARY AND SECONDARY MARKETS, EXCHANGES, AND OVER-THE-COUNTER MARKETS 55
Traditionally, the main characteristic of an exchange was a central physical location
where buyers and sellers gathered to trade standardized securities under a set of
rules and regulations, the New York Stock Exchange (NYSE) being a prime
example. However, the information revolution of the 1990s changed that forever—
most exchanges in the new millennium, especially those outside of the US, do not
have floor trading; instead, they have a centralized computer network that does trade
matching.
By contrast, any trade away from an exchange is called an OTC transaction.2
The expression over-the-counter goes back to the early days of the US, when banks
primarily acted as dealers for stocks and bonds, which were sold at counters in their
offices. There are organized markets for OTC transactions, where commercial and
investment banks, institutional investors, brokers, and dealers participate. Such OTC
markets (also called interbank markets) have no central location. Telephone, telex,
and computer networks connect geographically dispersed traders and make trades
possible in OTC markets. Exchanges used to handle the bulk of stock trading in the
US, but in recent times, organized OTC markets have taken on increased importance.
Exchanges operate under the basic assumption that if all trades take place in a
central location, this provides a level playing field that leads to fair price formation.
US stock exchanges are regulated by the Securities and Exchange Commission
(SEC), a federal government agency whose job is to protect investors, punish
violators, and prevent fraud. Before the SEC’s formation in 1934, the stock exchanges
were considered a hotbed of shady activity that routinely hurt the ordinary investor.
Mark Twain famously said in Pudd’nhead Wilson’s tale (1894), “October. This is
one of the peculiarly dangerous months to speculate in stocks. The others are July,
January, September, April, November, May, March, June, December, August and
February.” Interestingly, many folks no longer consider stock exchanges as bad or
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
risky; rather, they are now viewed as venues where respectable investors pursue
wealth by taking acceptable levels of risk. The derivatives exchanges, riding a wave
of mysticism and intrigue, have easily snatched the disrepute.
All exchange-traded securities, including derivatives, must satisfy government
requirements before they can be sold to the public or traded on an exchange. For
example, a federal government regulatory agency called the Commodity Futures
Trading Commission (CFTC) approves every kind of futures contract before it
sees daylight. Moreover, for a security to trade on an exchange, the issuing company
must satisfy some listing requirements in terms of the company’s assets, annual
earnings, shareholder interests, and audit requirements, among other factors. The
NYSE has the steepest requirements among the American exchanges, and the US
listing requirements are generally stricter than those in Asia or Europe.
Exchanges tend to have strict rules and regulations, codes of conduct for members,
and self-governance procedures. By lowering the likelihood of market manipulation
and fraud, exchanges make traders comfortable, which increases trade volume
(business). No exchange likes to have a government regulator at close heels, and a
good self-governance program keeps them away. Moreover, self-regulation can stave
2
Most OTC transactions involve a dealer who stands ready to buy and sell securities from an inventory
that she maintains. The next section discusses dealers in greater detail.
56 CHAPTER 3: STOCKS
off legal actions by angry customers who may otherwise feel cheated. Self-regulation
is costly, but like a vaccination, its benefits outweigh the costs.
Most exchanges also have a system of mediating disputes (arbitration) that handles
problems early in the process and lowers the chances of lawsuits. Although small
offenses may avoid detection, the big ones tend to get punished. Regulators and
exchanges regularly report names of guilty individuals and their punishments. The
press disseminates this information to the general public.
OTC contracts require no regulatory approval. Organized OTC markets, being an
electronically connected network of spatially separated traders, have fewer restrictions
than an exchange. They offer investors greater investment choices but little transaction
safety. For many OTC markets, when the going gets tough, the tough may skip
town—there is risk of a counterparty failing to honor his side of the contract. In
OTC markets, as in life, you need to remember the adage “know your customer.”
3.3 Brokers, Dealers, and Traders in
Securities Markets
Want to buy a house or get insurance protection? Call a broker. Need a new car or
planning to purchase a fancy electronic gadget? See a dealer. Because people trade at
different times, and they desire to see a collection of goods to make their pick, dealers
are useful. Dealers intermediate across time by maintaining an inventory of goods.
Buyers can buy from and sellers can sell to a dealer. The dealer sells a commodity
at the offer or the ask price and buys the commodity at the bid price. Ask and bid
prices were defined in Chapter 1. (Write a [for ask] above and b [for bid] below; Ask
is the Buying price for the Customer [abc].) For example, a used car dealer sets two
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
prices: the retail (or the ask) price at which he sells cars and a wholesale (or the bid)
price at which he buys cars.
The dealer earns a living from the difference between these two prices, which is
known as the bid/ask (bid/offer) spread. In finance, spread has four different uses:
(1) the gap between bid and ask prices of a stock or other security, (2) the simultaneous
purchase and sale of separate futures or options contracts for the same commodity
for delivery in different months, (3) the difference between the price at which an
underwriter buys an issue from a firm and the price at which the underwriter sells it
to the public, and (4) the price an issuer pays above a benchmark fixed-income yield
to borrow money. These spread definitions (except for the third definition, which is
more relevant in a corporate finance course) will be used throughout the book.
Paucity of information concerning trading opportunities is an impediment to
transactions. Brokers are intermediaries who help to overcome this hurdle and
facilitate transactions. They match buyers and sellers and earn commissions for this
service. Brokers have no price risk because they carry no inventory and do not trade
on their own accounts. Dealers face price risks because they hold inventories.
In both exchanges and OTC markets, brokers and dealers play significant roles.
They earn a living by maintaining smoothly functioning, orderly markets. Many
exchanges designate specialized dealers as market makers. They make markets
BROKERS, DEALERS, AND TRADERS IN SECURITIES MARKETS 57
by posting ask and bid prices and stand ready to trade at those prices throughout
the trading day—and enjoy enhanced trading privileges for their services. In OTC
markets, some dealers take up market-making functions. Brokerage and dealership
are risky business—charging too much will drive away customers, whereas charging
too little will wipe the business out. Risks originate from many sources:
■ Set security/margin deposit levels. Higher margins make brokers safer, but they are
costlier to customers.
■ Inventory to maintain. A larger inventory provides more to sell, but it’s costlier to
maintain and makes the dealer more vulnerable to price declines.
■ Bid and ask prices to post. A narrower spread means more transactions but fewer
profits per trade, while a widening of the spread has the opposite effect.
■ The amount of securities to bid or offer at those prices. A higher amount at any price
exposes the dealer to a greater risk of sharp price movements.
Brokers and the dealers understand these trade-offs and finely balance risks and
expected returns to survive in ruthlessly competitive markets. The setting of bid and
ask prices is the most important decision that a dealer faces.
There are many ways of characterizing traders in security markets. One common
practice is to distinguish them as individual investors or institutional traders. For our
purpose, it is more useful to categorize them in terms of their trading strategies either
as arbitrageurs, hedgers, or speculators. Arbitrageurs seek price discrepancies among
securities and attempt to extract riskless arbitrage profits. Hedgers try to reduce
risk by trading securities and are often cited as the chief reason for the existence of
derivatives markets. Speculators take calculated risks in their pursuit of profits; they
may be classified as scalpers, day traders, or position traders on the basis of how long
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
they hold their trades:
1. Scalpers trade many times a day with the hope of picking up small profits from
each transaction. They are physically present on an exchange’s trading floor and
pay very low trading costs. Scalpers’ holding periods are often measured in minutes
and seconds. Market makers are scalpers. Scalpers make markets more liquid by
standing ready to buy and sell at posted prices. Liquidity is a desirable feature of
securities markets. It’s the ease with which an asset can be converted to cash, and
vice versa.
2. Day traders try to profit from price movements within the day. They open their
trading positions in the morning and close them out before going home at night.
3. Position traders (also called trend followers) maintain speculative trading
positions for longer periods of time. Position traders often try to identify and
capture the abnormal price differences between two assets (a spread), with the
hope of reaping a profit when the spread is restored to its historical values.
These distinctions aren’t ironclad, and traders sometimes flit from one role to
another.
58 CHAPTER 3: STOCKS
3.4 Automation of Trading
Toward the turn of the century, rapid advancements in IT gave us revolutionary
products and services, for example, cell phones, the Internet, and personal computers.
Exchanges, brokers, and dealers exploiting these technologies have lowered trading
costs and in many cases eliminated human labor from the trading process. IT has
also revolutionized trading: innovations like electronic exchanges, online brokerage
accounts, and network linkages among exchanges are now commonplace.
3.5 The Three-Step Process of Transacting
Exchange- Traded Securities
Security trading on an organized exchange generally involves three steps: execution,
clearing, and settlement. OTC transactions follow a similar approach, except that
there is no clearing done by a third party (Figure 3.1 shows these three operations
for a typical stock trade):
FIGURE 3.1: Execution, Clearing, and Settlement in
the Stock Market
A) Execution
Buyer Seller
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Broker Broker
Agree on trade
(Alternately, a dealer may replace a trader and her broker)
B) Clearing (after market closing but before market opening on next trading day)
Brokerage firm representatives meet with clearinghouse
officials to “clear” (recognize and record) a trade.
C) Settlement (one or more business days after execution)
Buyer pays seller and gets ownership of the securities.
BUYING AND SELLING STOCKS 59
■ Execution. A trade is executed when the buyer (or her representative broker) and the
seller (or his rep) “meet” on an exchange, physical or electronic; agree on price
and quantity; and commit to trade.
■ Clearing. Before the market reopens, an executed trade must clear through the
exchange’s clearinghouse. Some brokers and dealers become clearing members,
who, besides clearing their own trades, clear trades for their clients and for
nonclearing brokers and dealers. The clearinghouse clears a trade by matching
the buy and sell orders, recognizing and recording the trade. The clearinghouse
in a derivatives exchange performs the additional role of guaranteeing contract
performance by becoming a seller to each buyer and a buyer to every seller.
■ Settlement. Finally, a trade ends with a cash settlement when the buyer pays for
and gets the securities from the seller. In the old days, this involved the exchange
of an ownership certificate for cash or check. Nowadays, it is usually done through
transfer of electronic funds for ownership rights between brokerage accounts.
IT advances have drastically reduced the time lag between these three steps.
Trades can be executed in the blink of an eye and even settled on a real-time
basis. While the NYSE and NASDAQ remain dominant venues for stock trading,
there are more execution choices than ever before. Your broker has a “duty to
seek the best execution that is reasonably available for its customers’ orders” (see
www.sec.gov/investor/pubs/tradexec.htm). For example, a broker may send your
order:
■ to trade an exchange-listed stock. to a national exchange (like the NYSE), to a
smaller regional exchange, or to a firm called a third market maker, which
stands ready to trade the stock at publicly announced prices
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ to trade an OTC stock to a market maker in the OTC markets
■ to an electronic communications network (ECN), which is an electronic
system that automatically matches buy and sell orders at specified prices
■ to another division of his firm to internalize the order by filling it from the firm’s
own inventory
You can also instruct the broker to route your order to a specific exchange, a particular
dealer, or a predetermined ECN.
3.6 Buying and Selling Stocks
Stocks have been trading in Europe for centuries. The oldest stock exchange opened
in Amsterdam in 1602 with printed shares of the United East India Company of the
Netherlands. The London Stock Exchange traces its origins back to 1698, when a list
of stock and commodity prices called “The Course of the Exchange and other things”
was issued, and stock dealers, expelled from the Royal Exchange for rowdiness, started
to operate in the streets and coffeehouses nearby (see the London Stock Exchange’s
website at www. londonstockexchange. com). The nineteenth century belonged to
60 CHAPTER 3: STOCKS
England—Britannia ruled the waves, the sun never set on the British Empire, and
London was the commercial capital of the world. After World War I, the hub of
commercial and trading activity shifted from London to New York and to the NYSE.
Trading at the New York Stock Exchange
The NYSE began in 1792, when 24 prominent stockbrokers and merchants assem-
bled under a buttonwood tree on Wall Street in Lower Manhattan and signed an
agreement to trade charging a uniform commission rate. Eventually, the Big Board
(a popular name for the NYSE) became the world’s best known and one of the
largest stock markets. The exchange heavily invests in technology and has pioneered
many innovations such as the stock ticker, an automated quotation service, electronic
ticker display boards, and electronic order display books. Example 3.1 discusses a
hypothetical trade on a typical exchange.
EXAMPLE 3.1: Exchange Trading of Stocks
Trade Orders
■ Suppose Ms. Longina Long wants to buy a round lot (one hundred shares) of YBM, which is the
abbreviation for Your Beloved Machines Inc., a fictitious company that we use throughout this book.
Assume that like many major US companies, YBM trades on a stock exchange. Ms. Long has opened
a brokerage account and placed enough funds into it for trading. She places an order by calling up her
broker or submitting the order through an Internet account.
■ The broker’s representative records and time stamps her order; web orders automatically record this
information. Smaller orders are usually executed quickly. Block trades that involve ten thousand
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
shares or more are often negotiated away from the trading floor, with or without the help of brokers.
■ Various types of orders can be submitted. A limit order must be filled at the stated or a better price
or not traded at all. By contrast, a market order must be immediately transacted at the best available
price. A liquid stock is actively traded and has many limit buy and sell orders around the market price.
As such, reasonably sized stock positions can easily be converted into cash with a minimum loss of
value.
Trade Execution
■ Suppose Long submits a market order and a counterparty is sought to take the other side of the trade.
This could be another investor or a professional. Long sells 100 YBM stocks at the ask price of $100.10
per share, and the trade gets executed. Ms. Long pays $10,010 for shares plus brokerage commissions.
A discount broker may charge her as little as $5, while a full-service broker will cost more.
Clearing and Settlement
■ After execution, most exchanges transmit trade records via computer to a little known but powerful
organization called the National Securities Clearing Corporation (NSCC) for clearance and
settlement. They are sent as locked-in transactions, which already have trade details from buyers and
sellers prematched by a computer. The NSCC also “nets trades and payments among its participants,
BUYING AND SELLING STOCKS 61
reducing the value of securities and payments that need to be exchanged by an average of 98% each
day.”3 It generally clears and settles trades on a two-business-day cycle (T + 2 basis).
3 www.dtcc.com/about/subs/nscc.php.
Over-the-Counter Trading
Over-the-counter markets, along with regional exchanges, have always provided the
major venue for trading small stocks and other securities that could not be listed
on the Big Board. Since the 1970s, NASDAQ has been considered the major OTC
market in the US.
Following the Penny Stock Reform Act of 1990, the OTC Bulletin Board
(OTCBB) began operations as “a regulated quotation service that displays real-time
quotes, last-sale prices, and volume information in over-the-counter (OTC) equity
securities.”4 Subsequently, the OTC market in the US has shifted to the OTCBB
and the Pink Quote OTCBB, which has expanded over the years, bringing greater
transparency to the OTC equities market. Only market makers can quote securities
in the OTCBB, and FINRA rules prevent them from collecting any fees for this
service. OTCBB displays market data through vendor terminals and websites. The
Pink Quote (formerly known as Pink Sheets owing to the color of the paper on
which the quotes used to be printed) is an electronic system that displays quotes from
broker-dealers for some of the riskiest stocks trading in the OTC markets. The SEC
website cautions potential traders, “with the exception of a few foreign issuers, the
companies quoted in Pink Quote tend to be closely held, extremely small and/or
thinly traded. Most do not meet the minimum listing requirements for trading on a
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
national securities exchange. . . . Many of these companies do not file periodic reports
or audited financial statements with the SEC, making it very difficult for investors
to find reliable, unbiased information about those companies.”5 Caveat emptor, or
“buyer beware,” is the guiding principle in this market.
Alternative Trading Systems: Dark Pools and
Electronic Communications Networks
The technology revolution has introduced securities trading in venues that are
opaque. These new trading venues are called alternative trading systems (ATS).
ATS such as dark pools and electronic communications networks (ECNs) are SEC-
approved non-exchange-based markets for trading securities. Such nonexchange
trading venues are a result of the SEC’s 1998 Regulation ATS (Regulation of
Exchanges and Alternative Trading Systems).
Dark pools (or dark pools of liquidity) have several features. First, they are
secretive trading networks that do not send an order directly to an exchange or display
4
www.otcbb.com/aboutOTCBB/overview.stm.
5
www.sec.gov/answers/pink.htm.
62 CHAPTER 3: STOCKS
it in a limit order book. Second, an interested trader either negotiates with a potential
counterparty or gets matched with one by the dark pool. Third, the market is a near-
exclusive domain for institutional players with large orders. National exchanges like
NYSE and the NASDAQ Stock Market also route some of their orders to dark pools.
ECNs are alternate trading systems that are registered with the SEC as broker-
dealers. An ECN’s participants are subscribers, which include institutional investors,
broker-dealers, and market makers. Individual traders can indirectly participate by
opening an account and submitting trades through a broker-dealer subscriber. ECNs
primarily trade stocks and currencies. Trades are usually submitted as limit orders,
which get executed when the ECN matches buy and sell orders according to some
protocol. If no matching order is found, an ECN may send it to another market center
for execution. Unlike a dark pool, the orders are publicly displayed to all subscribers.
Founded in 1969 to provide electronic trading for institutions, Institutional
Networks (Instinet) has been a pioneer in electronic trading and it became the first
ECN in 1997. The NASDAQ (founded in 1971) may be considered an early ECN.
ECNs were formally recognized and approved in 1998 when the SEC introduced
Regulation ATS. By slashing trading fees and providing better execution, ECNs have
enjoyed mind-boggling growth.
3.7 Dollar Dividends and Dividend Yields
“All men are ready to invest their money, but most expect dividends,” observed T.
S. Eliot in Choruses from the Rock in 1934. As Eliot observed, dividend payments in
the form of cash (cash dividend) or additional shares (stock dividend) are a time-
honored way for companies to reward their shareholders. For dividend payments, the
ex-dividend date is an important day. If you buy the stock before the exdividend
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
date, you get the share and the dividend (buying stock cum-dividend); however, if
you buy it on or after this date, then you get the stock without the dividend (buying
stock ex-dividend).
Let us analyze the behavior of the stock price on the ex-dividend date. Suppose
the cum-dividend stock price is $100 and the company pays a $2 dividend. Then,
we claim that the ex-dividend stock price should be $98. Why? If it were not, profit-
hungry traders would take advantage of any deviation, and as profits and losses occur,
this trading activity eventually eliminates the discrepancy from the market price.
To see how this works, suppose the ex-dividend stock price is $99 instead.
Then, clever traders will buy this stock just before it goes ex-dividend. Their $100
investment will immediately become $101 (ex-dividend stock price $99 plus dividend
worth $2), which they can sell and make $1 in instant profit. Conversely, if the stock
price falls by more than $2, say, it falls to $97 after the stock goes ex-dividend, then
the traders can “go short” (borrow and sell the stock) and buy back the stock after
it goes ex-dividend, locking in $1 as instant profits. In this way, $98 is the only ex-
dividend stock price consistent with no riskless profit opportunities.
We formalize this observation as a result. (see Result 3.1).
DOLLAR DIVIDENDS AND DIVIDEND YIELDS 63
RESULT 3.1
The Relation between a Stock’s Price Just Before and
After It Goes Ex-Dividend
In a market free of riskless profit opportunities,a
SCDIV = SXDIV + div (3.1)
where SCDIV is stock price cum-dividend, SXDIV is stock price ex-dividend,
and div is the amount of the dividend.
a
Later, we will call this an arbitrage opportunity. Formally, this result also depends on the stock
price process having a continuous sample path (see Heath and Jarrow 1988). Of course, this
result also depends on the market satisfying the standard frictionless and competitive market
assumptions that are introduced in later chapters.
Dividends are paid in dollars. However, to compare dollar dividends across stocks
with different valuations, one computes dividend yields. A dividend yield is the
dividend expressed as a fraction of stock price. In this example, the dividend yield is
(2/98) = 0.0204 or 2.04 percent. Dividing by the stock price normalizes the dollar
dividend and makes a comparison possible across stocks on an apples-to-apples basis.
The dividend yield is also useful when dealing with stock indexes like the
Standard and Poor’s 500 Index (S&P 500), which is a weighted-price average of
five hundred major US stocks (indexes are discussed in more detail in Chapter 6).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Accounting for hundreds of dividends (near two thousand dividends in the case
of the S&P 500) and adjusting the index’s value accordingly are daunting tasks. A
simple technique for making this adjustment using the dividend yield involves the
following steps: (1) add up all the dividends paid over the year and express this sum as
a percentage of the index value generating a dividend yield, and call it 𝛿; (2) assume
that 𝛿 is paid out over the year on a continuous basis, proportional to the level of the
index; and (3) assume that the dividend gets continuously reinvested in the index so
that the index grows in terms of the number of units. An initial investment of one
unit in this index then becomes eᄕ units after one year and eᄕT units after T years.
We illustrate this computation with Example 3.2.
EXAMPLE 3.2: Dividend Yields
■ Suppose that a fictitious stock index named “INDY Index” stands at 1,000 on January 1 and has a
dividend yield of 𝛿 = 0.06 or 6 percent per year. Assume that dividends are paid m = 3 times a year,
where each payment is 𝛿/m = 0.02 times INDY’s price at the dividend payment date. Let us buy one
unit of the index “INDY Spot.”
64 CHAPTER 3: STOCKS
■ Four months later, INDY stands at 1,100. The dividend payment is 0.02 times this amount, which
equals 22. As we reinvest this dividend, the amount of the dividend is irrelevant—all we need to know
is that we have 1 × 1.02 = 1.02 units of the index from now onward.
■ Four more months later, we will get another 2.00 percent of the index level as a dividend. Because
we were previously holding 1.02 units of INDY Spot, we now have 1.022 = 1.0404 units of INDY
Spot. Similarly, at the end of the year, we have 1.023 = 1.061208 units of the index.
■ Next, we modify the payout frequency and assume that the dividend is paid continuously and plowed
back into the portfolio after each payment. Here an investment of one unit of INDY Spot will be
worth eᄕT units after T years (see Appendix A). Eight months later, we have
eᄕT = e[0.06×(2/3)] = 1.040811
units of the index.
This property is restated as Result 3.2.
RESULT 3.2
Valuing an Asset Paying a Continuously Compounded
Dividend Yield
Consider an asset that pays a continuously compounded dividend yield of
𝛿 per year, which is reinvested back in the asset. Then, a unit investment in
the asset today grows to eᄕT units after T years.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Note that Result 3.2 gives a quantity adjustment to the stock price to incorporate
a dividend payment, while Result 3.1 is a price adjustment. These two modifications
will be used later in the book when considering stocks paying dividends.
3.8 Short Selling Stocks
There are two ways of selling shares: selling a stock from one’s existing portfolio
and selling short a share that one does not own. The first transaction is as straight-
forward as stock buying, whereas the second involves a more complex procedure. A
short seller borrows shares from an existing stockholder and then sells them. He has
the obligation to buy back the shares in the future and return them to the lender.
Moreover, he owes the lender any dividends that the company pays.
Interestingly, the act of short selling artificially creates more long shares than were
originally issued in the primary market. Later in the book, you will see that this can
become a critical factor in market-manipulation trading schemes.
SHORT SELLING STOCKS 65
A stock buyer is bullish because she expects that the stock will go up in value.
A stock seller is bearish because he expects the stock to fall in value. A short seller
is much more aggressive in his bearish stance, unlike a seller, who simply gets rid
of a stock. A short seller bets on his negative view by undertaking a fairly risky
trade, which could be quite dangerous if the bet goes wrong. Chapter 10 discusses
an example in which a manipulator corners the market and squeezes the shorts.
To see how short selling works, let’s walk through a short sell example (see
Example 3.3).
EXAMPLE 3.3: Short Selling
■ “Typically when you sell short, your brokerage firm loans you the stock. The stock you borrow comes
from either the firm’s own inventory, the margin account of other brokerage firm clients, or another
lender,” informs an article on the SEC’s website.7
■ Suppose Mr. Oldham owns one hundred shares of BUG (Boring Unreliable Gadgets, a fictitious
company name), which are held in his account at Mr. Brokerman’s firm. To make them available for
short selling, the shares are held in “street name.” Oldham doesn’t gain anything from this—he is just
helping Mr. Brokerman earn some commissions. For simplicity, let’s assume that Mr. Brokerman is
the only broker in town.
■ Brokerman borrows one hundred shares from Oldham, lends them to Shorty Stock, and helps Shorty
sell them short to Mr. Newman (see Figure 3.2). In the future, Shorty has to buy one hundred BUG
shares, no matter what the price is, and give them back to Oldham. He owes any dividends paid over
this period to Brokerman, who credits them to Oldham’s account.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
FIGURE 3.2: Short Selling Example
Newman
Owns 100 BUG shares
Brokerman Dividends from
BUG
Short-sells
Borrows
Oldham
Shorty Stock
Owns 100 BUG shares
Shorty sold 100 BUG
in street name; Pays dividends shares short
loses voting rights
66 CHAPTER 3: STOCKS
■ The market is now long two hundred and short one hundred BUG stocks. Still, from BUG’s
perspective, there are only one hundred outstanding shares owned by Newman. The other one
hundred shares held long by Oldham and the one hundred shares short sold by Shorty Stock are
artificial creations, for they cancel one another. Newman gets dividends from BUG and has the voting
rights.
■ Strangely, Oldham’s original shares became artificial shares during the short selling process. They are
Shorty Stock’s babies, which he has to look after. Thus Shorty has to match any cash dividends, stock
dividends, and so on, that Oldham would have rightfully received.
■ Dividends lower a stock’s value and benefit Shorty Stock because he can now buy them more cheaply.
A dollar dividend on a $50 stock will lower its value to $49, which will help Shorty earn a dollar if he
now buys back the stock. Paying Oldham this dollar will keep everyone happy by making things fair.
■ The only thing that Oldham really loses is voting rights because BUG will not recognize an extra
owner. He doesn’t mind that, unless he is planning to thrust major changes on BUG through proxy
fights.
7 “Division of Market Regulation: Key Points about Regulation SHO” describes the short selling process for retail customers
(see www.sec.gov).
3.9 Margin—Security Deposits That
Facilitate Trading
The word margin has many meanings. In economics, the terms margin and marginal are
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
often used to denote what happens as a result of a small or unit change. In accounting,
the term profit margin (of a business) refers to the amount by which revenues exceed
costs. In financial markets, margin means “a sum deposited by a speculator with
a broker to cover the risk of loss on a transaction on account; now esp. in to buy
(also trade, etc.) on margin” (according to the Oxford English Dictionary Online).
Margins and collaterals are fundamental to any security market transaction involving
explicit or implied borrowing.
Let’s consider some examples:
■ Buying stocks on borrowed funds. You decide to leverage (usually called gearing in
Europe) your portfolio to increase your potential returns. Suppose you invest
$10,000 of your own funds and take a loan of $8,000 from your broker to buy
securities worth $18,000. You must open a margin account (also known as a
cash and margin account) with your broker. Your investment of $10,000 is the
margin or collateral, which acts like a security deposit. Now, a 10 percent return on
your leveraged portfolio will be $1,800 earned on your investment, an 18 percent
return. But leverage cuts both ways because losses also get magnified. Your actual
return is lower because the broker charges you interest on the daily balance of your
loan until you pay him back.
MARGIN—SECURITY DEPOSITS THAT FACILITATE TRADING 67
■ Short selling stocks. A short seller must also open a margin account. The broker
requires you to keep the proceeds from the short selling plus additional funds as
margin in this account. Remember that to short, you borrowed the stock first.
Before initiating a trade that requires margin, you must put up the initial margin.
This could be in the form of cash or high-quality, low-risk securities. For the broker’s
protection, there is a maintenance margin requirement: the minimum amount
that you must maintain in the account to keep it open. If your position loses value
and the margin account balance declines to this threshold level, the broker will issue a
margin call that requires you to promptly provide enough cash (maintenance margin
is always in the form of cash) to restore your account balance to the maintenance
margin level. If you fail to do this, the broker liquidates your portfolio. Extension 3.1
shows how margin account adjustments work.
EXTENSION 3.1: Margin and Stock Trading
As a result of the great stock market crash of 1929, the US government would like to control the level of stock
speculation and has entrusted the Federal Reserve Board with the job of setting minimum margin requirements
for securities trading. The exchanges and brokerage firms may set them at even higher (but not lower) levels.
However, the real role of margin in financial transactions is to minimize counterparty credit risk.
Currently the Fed sets the initial margin requirement at 50 percent and the maintenance margin at 25 percent
for stocks (and convertible bonds). This means that when initiating a margin trade, you must come up with at
least 50 percent of your own funds and finance the rest with a loan from your broker. Your purchased securities
will appear with a credit sign and the loan with a debit mark in your brokerage account, and the broker will
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
charge interest on it. As it is a secured loan (a low-risk loan backed by collateral; the market value of the stock is
much more than the amount borrowed), your broker will probably charge the call loan rate (or broker’s call)
and add a slight premium. Adding a spread to the benchmark libor index rate usually determines the broker
loan rate. It is regularly published (along with other interest rates) in the Wall Street Journal and in the business
sections of many other newspapers. The following formula is used for margin computations.
Formula 1: Margin Computation for Stock Trading
The margin (or security deposit) computation in connection with buying stocks on margin (by taking a loan
from a broker) or short selling is given by the formula
Margin = [(Market value of assets –Loan) /Market value of assets]
If maintenance margin is 25 percent, then your equity must be at least 25 percent of the portfolio value. If
your equity has dropped below this level, you must respond to a margin call, deposit additional funds, and bring
the equity back to the 25 percent level. Your firm may set its own margin requirements (often called “house
requirements”) at a higher level, say, 40 percent, for maintenance margin.
Margin adjustments can get very complicated (see Sharpe, Alexander, and Bailey 1999). The following example
illustrates the computations involved (see Ext. 3.1 Ex. 1).
68 CHAPTER 3: STOCKS
EXT. 3.1 EX. 1: Margin Account Adjustments for Stock Trading
Buying Stocks on Margin
■ Suppose you buy two hundred YBM stocks at $100 each. The purchase price is $20,000, and you decide to
borrow 50 percent of this amount, $10,000, from your broker. Your investment is $10,000. By our formula,
initial margin is given by
(20, 000 − 10, 000) /20, 000 = 0.5 or 50 percent
■ Suppose YBM goes up to $110. Then
- the market value of your assets is 200 × 110 = $22,000
- the loan amount is still $10,000
- the investment in your account is (Market value of assets – Loan) = 22,000 – 10,000 = $12,000
- the margin in your account = 12,000/22,000 = 0.5455
■ Suppose YBM goes down to $60. Then
- the market value of the portfolio is 200 × 60 = $12,000, and your investment is 12,000 – 10,000 = $2,000
- the margin = 2,000/12,000 = 0.1667
- you will receive a margin call to come up with more cash, an extra $1,000, so that your margin meets the
25 percent requirement
Margin Account for a Short Seller
■ Suppose you decide to sell YBM short instead and hope to use the $20,000 in proceeds to buy a new red car.
However, the broker not only refuses to let you touch any of this money but also requires you to deposit an
additional 50 percent, or $10,000, into your margin account. Isn’t this outrageous?
If you think carefully, the broker’s actions are entirely justified. Consider what would happen if he keeps just
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
$20,000 and the stock price goes up by just 1 cent. Your account now has a negative value of – 200 × 100.01
+ 20,000 = –$2. The 50 percent initial margin requirement provides the cushion he needs. Whereas a stock
buyer may or may not open a margin account, a short seller is always required to do so.
■ In contrast to a regular stock purchase, a short sale involves an initial sell and a subsequent buy. Consequently,
any increase in the stock price will increase the liability and eat up margin in your account.
3.10 Summary
1. This chapter has two themes: a study of securities trading in general and stock
trading in particular. Institutional details matter in finance. Stock trading is the
easiest type to understand, and the trading of different derivatives can be seen as
extensions to this process.
2. Stocks first trade in primary markets and then in secondary markets.
3. Exchanges were characterized by central physical locations where buyers and
sellers gathered to trade. Nowadays, many exchanges are completely electronic
and have a central computer executing the trades.
SUMMARY 69
4. Any trade away from an exchange is called an over-the-counter transaction. There
are organized OTC markets (interbank markets), which are diffused networks of
buyers and sellers, brought together by telecommunication connections.
5. Brokers match buyers and sellers and earn commissions for this service. Dealers
trade on their own account and survive by posting a bid/ask spread.
6. Traders may be classified into three categories: hedgers who use securities for
risk reduction, speculators who accept risky trades with profit expectations, and
arbitrageurs who try to find price discrepancies and extract riskless arbitrage
profits with strategies like buying low and selling high. Speculators can be further
distinguished as scalpers, day traders, and position traders on the basis of how long
they hold their trades.
7. The IT revolution has given us online brokerage accounts, electronic exchanges,
and network links among exchanges and has equipped individual investors with
unprecedented resources. Electronic exchanges minimize human involvement in
the trading process and have eliminated floor trading altogether.
8. Security trading in an organized exchange generally involves three fundamental
steps: execution, clearing, and settlement. There are more execution choices
today than ever before. A trade may be sent to a national exchange, to a smaller
regional exchange, to a third market maker, to a market maker in the OTC
market, to an alternative trading system such as an electronic communications
network, or to another division of the firm.
9. The NYSE is a prototypical securities market. It merged with an electronic
exchange and demutualized and restructured itself as a for-profit company.
10. Dividend payments are a time-honored way of rewarding stock investors. When
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the stock goes ex-dividend, the cum-dividend stock price equals the ex-dividend
stock price plus the dividend.
11. The concept of continuously compounded interest can be used to study an
investment in a portfolio that mimics the cash flows from a stock index. Consider
one unit investment in an asset that pays dividends according to a continuously
compounded yield 𝛿 per year, which is reinvested back in the asset (assume
dividend 𝛿/m is paid and compounded m times per year, where m becomes
infinitely large). Then, after T years, the portfolio will contain eᄕT units of the
asset.
12. A short seller borrows shares held in street name and sells them short. The
original owner loses voting rights, and the short seller compensates the owner
for dividends.
13. Although the concept of margin is similar for trading different securities, the
transactions vary. Traders buying stocks open a margin account only if they would
like to finance a part of their purchase with a loan from the broker. Margins are
security deposits that can be in the form of cash or some high-quality security.
Margin account holders must start with an initial margin. When the account
balance falls below the maintenance margin level, the broker issues a margin call,
at which time the account holder must come up with enough cash to restore
70 CHAPTER 3: STOCKS
the account balance to the maintenance margin level. Short sellers are always
required to open margin accounts and must keep the entire sale proceeds plus
some additional margin with the broker.
3.11 Cases
Martingale Asset Management L. p. in 2008: 130/30 Funds and a Low Volatility
Strategy (Harvard Business School Case 209047-PDF-ENG). The case discusses
the mechanics and the economic implications of leverage and short selling for
investment strategies and evaluates minimum volatility stock investment strategies
and quantitative investing in general.
Deutsche Borse: (Harvard Business School Case 204008-PDF-ENG). The case
explores the implications of Deutsche Borse’s acquisition of a stake in a company
specializing in clearing, settlement, and custody of securities across borders.
ICEX: Making a Market in Iceland (Harvard Business School Case 106038-
PDFENG). The case examines the impact of increased performance on the
international visibility and positioning of the Icelandic Stock Exchange and
considers various options for stock exchange growth in the backdrop of the
country’s strong economic performance during the period.
3.12 Questions and Problems
3.1. Explain the difference between a stock and a bond.
3.2. Explain the difference between an exchange and an OTC market.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
3.3. Explain the difference between a broker and a dealer.
3.4. Explain the difference between a bid and an ask price.
3.5. Explain the difference between a market order and a limit order.
3.6. A forward market is trading for future purchase of a commodity, while a spot
market is trading for immediate purchase. Is the stock market a spot or forward
market?
3.7. For the stock market, can there be a difference between the total number of
shares issued by a company and the total number of shares held by investors in
the market? If yes, explain why.
3.8. Explain the difference between execution and settlement.
3.9. Explain the difference between bull and bear markets.
3.10. The SEC regulates American stock markets. However, NYSE members have
committees that carry out a host of self-regulatory activities. NYSE members
are profit seeking—why would they self-regulate themselves, when regulations
only raise their cost of doing business?
3.11. Why does a dealer offer to trade only a fixed amount at the bid and ask prices?
QUESTIONS AND PROBLEMS 71
3.12. You are a dealer and post a price of $50.00 to $50.50 for a stock. The buy
orders outweigh sell orders, and your inventory is dwindling. How should you
adjust the bid and the ask prices, and why?
3.13. What is the difference between an arbitrageur, a hedger, and a speculator?
3.14. What does trading on the OTC mean?
3.15. What is the difference between a stock trading ex-dividend and cum-dividend?
3.16. Suppose that a stock pay a $5 dividend at time t. The dividend is announced
at time (t – 1), when the stock trades cum-dividend at a price of $100. What
should be the ex-dividend price at time t? Explain your answer.
3.17. Explain how to sell a stock short, assuming that you do not own the underlying
stock. Why would one short sell a stock?
3.18. Consider the following data: YBM’s stock price is $100. The initial margin is
50 percent, and the maintenance margin is 25 percent. If you buy two hundred
shares borrowing 50 percent ($10,000) from the broker, at what stock price will
you receive a margin call?
3.19. If the stock price is $105 and the company had paid in the previous year two
quarterly dividends of $0.50 each and two more of $0.55 each, then what is
the dividend yield?
3.20. Consider an asset that pays a continuously compounded dividend yield of
𝛿 = 0.05 per year, which is reinvested back in the asset. If you invest one
unit in the asset, how many units would you have after 1.5 years?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
4
Forwards and Futures
4.1 Introduction EXTENSION 4.1 Futures
Exchanges in China and India
4.2 Forward Contracts
4.6 Hedging with Forwards and
4.3 The Over-the-Counter Market Futures
for Trading Forwards
4.7 Summary
4.4 Futures Contracts
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
4.8 Cases
4.5 Exchange Trading of a
Futures Contract 4.9 Questions and Problems
FORWARD CONTRACTS 73
4.1 Introduction
Surprisingly although many think of derivatives as new and sophisticated securities,
this is hardly the case. Forward contracts are the oldest known derivatives, tracing
their origins back to India (2000 BC), to ancient Babylonia (1894–1595 BC), and
even to Roman merchants trading grains with Egypt.1 Futures contracts are “the new
kids on the block.” The oldest futures contracts traded in Amsterdam, Netherlands,
in the middle of the sixteenth century and on the first futures exchange, the Dojima
Rice Exchange in Osaka, Japan, in 1688. As such, there must be good reason why
these contracts have existed for millennia to facilitate trade across time. With respect
to those who believe that derivatives are new innovations, we see that their timing is
only off by three or four thousand years—but what are a few thousand years among
friends?
We start our study with the simplest and the most basic derivatives, forwards
and futures. First we introduce forward contracts that trade in the over-the-counter
market. Next we discuss futures contracts. Forward and futures contracts are fraternal
twins, for the two derivatives are very similar but not identical. We explore their
similarities and differences, then we take a bird’s-eye view of how a company uses
forward (and futures) contracts for hedging input and output price risk, a topic
discussed in greater detail in Chapter 13.
4.2 Forward Contracts
Intuitively, a forward contract locks in a price today for a future transaction—contract
now, transact later is the mantra. More formally, a forward contract or a forward is a
binding agreement between a buyer and a seller to trade some commodity at a fixed
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
price at a later date. This fixed price is called the forward price (or the delivery
price, usually denoted by F), and the later date is the delivery date (or the maturity
date, usually referred to as time T). Forward contracts are derivatives as their values
are derived from the spot price of some underlying commodity.
By market convention, no money changes hands when these contracts are created.
Such an exchange can only happen if both sides are happy with the terms of the
contract and believe the contract is fair, that is, it has a zero value. To understand why,
suppose the forward had a positive value to the buyer of the commodity. Then it must
have a negative value to the seller. This means that when the seller enters the forward
contract, his wealth immediately declines. He has been hoodwinked by the buyer! A
rational seller would not freely enter into such a contract. A similar argument holds
in reverse if the forward contract has a negative value. Consequently, if no money
changes hands when the contract is created, it must have a zero value. As such, the
delivery price written into the contract must be a fair price for future trading of the
asset. How do we find this delivery price? It’s actually quite straight-forward. Chapters
11 and 12 will show how to find this price using some basic economic principles.
1
See Chapter 8 for an in-depth history.
74 CHAPTER 4: FORWARDS AND FUTURES
Derivatives trade on an exchange or in an over-the-counter (OTC) market. Like
a magician pulling a hare out of his hat, derivatives also get created out of thin
air! Indeed, they are created when one party wants to trade a particular derivative,
finds another party to take the other side of the transaction, negotiates a price, and
completes the deal. This observation highlights a classic characteristic of derivatives—
they trade in zero net supply markets. A derivative does not exist until a trade takes
place—unlike stocks, the buyer’s and seller’s positions net out, leaving the net supply
of derivatives at zero. As such, trading derivatives is also a zero sum game. What
one side of the contract gains, the other side loses, and vice versa.
Some market jargon is useful in understanding derivative transactions. If you agree
to buy at a future date, then you take a long position or you are going long. If
you agree to sell at a future date, then you are taking a short position or you are
going short. A forward buyer is bullish while a forward seller is bearish about the
direction of the price of the underlying commodity. Consider the following example
of a forward contract.
EXAMPLE 4.1: The Life of a Forward Contract
■ Today is January 1. Ms. Longina Long longs to buy some gold at a fixed price in the middle of the
year. Forwards trade in the OTC (interbank) market, where Long’s broker finds Mr. Shorty Short.
Long and Short agree to trade fifty ounces of gold at a forward/delivery price of $1,000 per ounce
on June 30 at a mutually convenient place (see Figure 4.1).
■ When the contract begins, its delivery price is adjusted so that the contract is executed without
exchanging cash. Consequently, the delivery price is fair, and the market value of the contract is zero
at the start.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Thus a derivative (forward contract) gets created through mutual agreement. The brokers collect commissions
from the traders for their matching service.
■ If Long and Short want, they can close out their positions early through mutual agreement.
Theoretically speaking, they can sell their respective sides of the trade to other investors in the
secondary market. Practically speaking, it’s a hard task. Surely they can trade with others to stop
price risk, but credit risk will remain. If they don’t close out their positions early, Long and Short
meet on the maturity date of June 30.
■ What happens on June 30 depends on where the spot settles on that fateful day. Because they agreed
on physical delivery, Short sells fifty ounces of gold to Long at $1,000 per ounce. Had they decided
on cash settlement instead, the loser would have paid the price difference to the winner. If the spot
price of gold is higher than the delivery price of $1,000 per ounce, then Long wins and Short loses—
Long pays less for gold than it’s worth in the spot market. Conversely, if the spot price of gold is lower
than $1,000 per ounce on June 30, then Long loses and Short wins—she pays more for that gold than
it’s worth in the spot market. The forward contract terminates after delivery.
FORWARD CONTRACTS 75
FIGURE 4.1: Timeline for a Forward Contract
Now (start date) Intermediate dates Delivery/maturity date
(Time 0 = January 1) (Time T = June 30)
Long agrees to buy 50 Traders can close out Long buys gold for $1,000
ounces of gold on June 30, positions by making a from Short.
which Short agrees to sell. reverse trade (sell if long, If gold price >$1,000, Long
buy if short). wins. If gold price <$1,000,
Short wins.
No money is paid now. If they don’t close out Zero–sum game-one’s gain
positions, they meet on is the other’s loss.
June 30.
They negotiate the forward Long’s payoff S(T) – 1,000.
price F = $1,000 and Short’s payoff
contract terms.* – [S(T) – 1,000].
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
* Prices are per ounce
We will now summon the power of algebra to understand these concepts and
develop formulas for Long’s and Short’s payoffs on the maturity date. Before doing
so, though, let’s adopt some conventions that will be useful (see Table 4.1).
TABLE 4.1: Conventions
(used unless otherwise noted)
■ Prices are for one unit of an asset or a commodity. Multiply by the
contract size to get the total value of the transaction.
■ We start at time 0.
■ We will often suppress arguments like t and T to reduce clutter.
For example, we denote F(t, T) by F(t) or F.
76 CHAPTER 4: FORWARDS AND FUTURES
Figure 4.1 gives the key dates in a forward contract. Because there are no
intermediate cash inflows or outflows to a forward contract, we can just focus on
the beginning and the end. Suppose the forward starts today. Remember our mantra
contract now, transact later—we need two symbols to capture “now” and “later.” Let
F(t, T) be the forward price (or delivery price) that we decide now (time t) for a
transaction later on the delivery date (T). The first symbol, t, indicates the date the
forward price is quoted, and the second argument, T, gives the delivery date. To
keep things simple, we will usually start our clock at time t = 0. The forward price is
F(0, T). To reduce clutter, we will often write F(t, T) or F(0, T) as F.2
How is the delivery price linked to the forward price? When the contract starts, by
definition, the forward price equals the delivery price, denoted by F. The delivery price
remains fixed over the life of the contract. The forward price, which is the delivery price of
newly written contracts, can, of course, change. If we had waited a day and entered
into a new forward contract tomorrow (date t + 1 or date 1) for buying gold at the
same delivery date T, the price might be different. The new delivery price or the
forward price is F(t + 1, T) or F(1, T). But that’s not our price. We are committed
to pay F(0, T) at time T. Fair or foul, we have to live with our agreed on delivery
price until the contract matures.
Let S(.) be the spot price (or cash price) of the commodity for an immediate (spot)
transaction. Generically, S(t) is the spot price at any time t; it takes values S(0) at date 0
and S(T) on the delivery date T. We can call up a broker and easily get S(0), but who
can foretell what S(T) will be on the delivery date? No one can exactly because the
spot price on the delivery date is random and unknown. In fact, if you could predict
stock prices correctly, using derivatives, you’d become a multimillionaire quickly.
As a result, standing in the present, we don’t know the value of the forward contract
to the long position holder at the delivery date (T). So we write
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Long’s payoff at the delivery date T
(4.1)
= [S (T) − F]
If there’s physical delivery, the long gets a commodity worth S(T), which she can
immediately sell. This is shown with a positive sign. She pays the delivery price F.
This payment enters with a negative sign.
If there’s cash settlement, the long gets S(T) – F from the short when S(T) > F
and pays F – S(T) if S(T) < F:
Short’s payoff at the delivery date T
= − [S (T) − F] (4.2)
= F − S (T)
With physical delivery, the short gets paid the delivery price F, a cash inflow. He
surrenders to the long an asset that’s worth S(T) in the market.
2
For simplicity, we ignore market imperfections like transaction costs, taxes, and convenience yields
from holding a long position in the underlying asset. We will include these market imperfections in later
chapters.
FORWARD CONTRACTS 77
If the contract is cash settled, then the short gets F – S(T) if F > S(T); otherwise,
she pays S(T) – F to the long.
Notice that the long’s and short’s payoffs are exactly equal and opposite. As
mentioned previously, forward contracts are zero-sum games—if you add up the two
payoffs, the net result is zero.
EXAMPLE 4.2: Payoffs to a Long and Short Forward Contract
■ We revisit Example 4.1 again, but this time using notation. Today is January 1 (date 0). The forward
price on which Ms. Long and Mr. Short agree for delivery on June 30 (date T) is F(0, T) = F =
$1,000 (all prices are for an ounce of gold).
■ Suppose the spot price on the delivery date T is S(T) = $1,005. Long’s forward payoff on delivery
is (1,005 – 1,000) = $5 or a gain of $5. Long pays $1,000 for gold worth $1,005 in the spot market.
Short’s forward payoff at time T is –(1,005 – 1,000) = − $5 or a loss of $5. The contract forces short
to accept this below-market price for gold on the delivery date.
■ If, instead, S(T) is $990 on the delivery date, then Long’s payoff at delivery is (990 – 1,000) = – $10
or a loss of $10. She pays $1,000 for gold worth only $990 in the market. Short’s payoff at delivery is
– (990 – 1,000) = $10 or a gain of $10. Observe that for each possible spot price S(T), the long’s and
the short’s payoffs add to zero, confirming the zero-sum nature of this trade.
■ The profit and loss for long’s and short’s forward positions on the delivery date (time T) are graphed
on the profit diagram in Figure 4.2. The x axis plots the spot price of gold S(T) on the delivery
date, while the y axis plots the profit or loss from the forward contract.
■ The payoff to Long is a dashed straight line, making a 45 degree angle with the horizontal axis and
cutting the vertical axis at a negative $1,000, which is the maximum loss. This happens when the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
spot (gold) is worthless, but Long still has to pay $1,000 for it. Above this, each dollar increase in the
spot price cuts Long’s loss by a dollar. If gold is worth $1, then Long’s loss is 1,000 – 1 = $999. If it’s
worth $800, then the loss is $200. Ms. Long and Mr. Short break even at $1,000. If the gold price
ends up higher than $1,000 on the delivery date, then Long sees profits. Figure 4.2 shows that the
profit potential for a long forward position is unbounded above.
■ The payoff to Short is the downward sloping dashed straight line, making a negative 45 degree angle
with the horizontal axis. Considering the horizontal axis as a mirror, Short’s payoff is the mirror image
of Long’s payoff. When gold is worth 0, Short makes $1,000, as worthless gold is sold for $1,000. Thus
Short’s payoff line touches the vertical axis at $1,000. For each dollar increase in gold’s price, Short’s
profit declines by a dollar. If gold is worth $1, then Short’s profit will be 1,000 – 1 = $999, and so on. If
gold increases beyond $1,000, then Short starts losing money. Short’s maximum loss is unbounded—
if gold soars, Short plunges into the depths of despair! This shows the risks of naked short selling
(short selling without holding an offsetting position in the underlying asset).
78 CHAPTER 4: FORWARDS AND FUTURES
FIGURE 4.2: Profit Diagram for Forward
Contract on the Delivery Date
Profit
$1,000 Long forward
45°
0
S(T), spot price on
F = $1,000 the delivery date
Short forward
-$1,000
4.3 The Over-the-Counter Market for
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Trading Forwards
The OTC derivatives market has seen tremendous growth over the last four decades.
Gigantic markets exist for foreign exchange (“forex”) forwards, money market
instruments, and swaps. This is a market for the “big guys”—the banks—with
excellent credit ratings. As such, it is often called the interbank market. The participants
trade by telephone, telex, and computer, and they tend to be located in the world’s
major financial centers such as London, New York, and Tokyo. Example 4.3 shows
a typical transaction in the forex (foreign exchange) forward market.
EXAMPLE 4.3: Trading Currency Forward
The Contract
■ Suppose a US company has bought a machine worth €3 million from a German manufacturer with
payment due in three months. The treasurer of the US company feels that at $1.4900, the euro is
attractively priced in the spot market.3 But he doesn’t know what will happen in three months time;
for example, if a euro costs $1.6000, his company will pay an extra $330,000.
THE OVER-THE-COUNTER MARKET FOR TRADING FORWARDS 79
■ The treasurer would like to pay today, but the company is short of cash. He checks the forward market
and finds that DeutscheUSA (a fictitious name), a large commercial bank, bids euros for $1.5000 and
offers euros for $1.5010 in three months’ time (see Figure 4.3). He readily agrees and locks in a
price of $1.5010 × 3,000,000 = $4,503,000 for the machine. This forward market trade allows the
treasurer to eliminate exchange rate risk from the transaction so that the business can focus on its core
competency.
The Dealer’s Hedge
■ How does DeutscheUSA protect itself? Having far-flung operations and diverse customer bases, big
banks can often find another firm to take the other side of the transaction. Suppose DeutscheUSA has
branches in Germany. Contacting its customer base, DeutscheUSA finds a German importer hoping
to buy €3 million worth of computer parts from the United States in three months’ time. To shed
foreign exchange risk, the German importer agrees to buy €3 million for $1.5000 × 3,000,000 =
$4,500,000.
■ DeutscheUSA has done something cool. It has removed price risk for both the US and German
importers and managed to earn a riskless spread of (1.5010 – 1.5000) × 3,000,000 = $3,000 in the
process! Skillful dealers try to perfectly offset such OTC trades.
■ In reality, such perfect offsets are unlikely, and DeutscheUSA may have residual price risk after
netting out all of its foreign currency transactions. Moreover, DeutscheUSA has some credit risk
if any counterparty fails, and it will have to keep tab of the transactions until the deal is done. If it
desires, DeutscheUSA may hedge any residual net exposure with forex futures or by entering into a
forward transaction with another dealer in the forex market.
3 The International Organization for Standardization, or the ISO, has established three-letter codes for representation of
currencies and funds. ISO’s currency codes are USD for the US dollar, AUD for the Australian dollar, EUR for the euro, JPY
for the Japanese yen, GBP for the UK pound sterling, etc. When no confusion arises, we will use commonly used expressions
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
(and symbols) like dollar ($), euro (€), yen (¥), and pound sterling or pound (£).
OTC market participants may be classified as brokers, dealers, and their clients.
Usual clients are banks, corporations, mutual funds, hedge funds, insurance compa-
nies, and other institutions. Many banks act as dealers and make markets in a variety
of derivatives. Sometimes clients call up dealers for quotes. Typically a dealer’s sales
force makes regular cold calls to potential and existing clients, trying to sell their
derivatives products. Owing to competition, simple derivatives like currency forwards
aren’t very lucrative. Banks’ proprietary trading desks make more profits when they
can identify some special client need and create a customized product. Big players
have trading or dealing rooms, where trading desks dedicated to spot, forward,
and options trading are located. Forward prices are quoted in terms of a bid and ask,
and because there are no specified delivery months like June and September, forward
prices get quoted for delivery in one, two, three, six, or more months from the current
date. Recently, owing to new financial regulations, many banks are deemphasizing
proprietary trading and focusing more on their dealership businesses.
80 CHAPTER 4: FORWARDS AND FUTURES
FIGURE 4.3: Intermediary’s Role in the Forward
Transaction
Initially
U.S. Co. needs to buy
German machines worth Faces currency risk
€3 million in three months German importer needs to
Faces currency risk buy computer parts from
U.S. in three months
After DeutscheUSA Enters
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Deutsche USA
(dealer)
U.S. Co. agrees to buy €3 $1.5010 per euro
Seller
millions in three months (Short)
German importer needs Buyer
dollars and agrees to sell (Long)
$1.5000 per euro
€3 million in three months
Deutsche USA has perfectly hedged its book and earns $3,000 spread. It has credit risk
but no price risk.
FUTURES CONTRACTS 81
4.4 Futures Contracts
Forward and futures contracts are fraternal twins. Both involve an exchange of cash
for an asset purchased at a later date but at prices negotiated at the start of the contract.
Yet their differences (explored in Table 4.2 and Figure 4.4) make it hard to label them
as identical. Just as a variety of features like air bags, antilock brakes, power steering,
and seat belts make a car easier and safer to drive, a number of added features make
futures easier and safer to use. You may view futures as a standardized forward or a
forward as a customized futures.
Unlike a forward contract, a futures trades on an organized exchange, which is
a regulated marketplace where buyers and sellers gather to trade a “homogeneous
product.” The federal Commodity Futures Trading Commission and the industry’s
National Futures Association regulate commodity exchanges in the United States.
In contrast, a forward trades in the OTC market, where no regulator tells them the
dos and don’ts.
An exchange standardizes futures contracts for easy trading. A trader only has to
tell how many contracts to buy or sell and at what price. Standardization helps create a
secondary market. It reduces transaction costs and makes futures markets more liquid.
A short can easily close out her position before delivery by going long the same
contract with another trader, who becomes the new short, while the original long
TABLE 4.2: Comparison of Futures and Forward Contracts
Futures Forward
a. Regulated Largely unregulated
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
b. Trades in an organized exchange Trades in an OTC market
c. Standardized Customized
d. Usually liquid and has a secondary market Illiquid and has virtually no secondary market
e. A range of delivery dates Usually has a single delivery date
f. Usually closed out before maturity to avoid taking Usually ends in physical delivery or cash settlement
physical delivery
g. Guaranteed by a clearinghouse and has no credit No such guarantee and has counterparty credit risk
(counterparty) risk
h. Trading among strangers; individual’s Forward traders know each other and usually have
creditworthiness is irrelevant high credit rating or require collateral
i. Requires margins (security deposits) No margin requirements (but collateral may be
required)
j. Small transaction costs Transaction costs are high
k. Settled daily Settled at maturity
82 CHAPTER 4: FORWARDS AND FUTURES
FIGURE 4.4: A Diagram of a Futures Contract
CFTC, NFA regulate
Broker, Exchange, Clearinghouse monitor
… …
Daily margin payments throughout contract life
… … … … … … … …
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Starting date Intermediate dates Delivery/maturity
period
position remains undisturbed. In fact, most futures are closed out before delivery. This
happens because making physical delivery is a costly exercise, and closing out futures
early avoids this expense. If, for some reason, a position remains open until maturity,
then delivery must take place on one of several dates during the delivery period. The
details of this delivery procedure are discussed in Chapter 9.
Contrast this with a forward contract, which is privately negotiated between two
traders. As such, the costs of transacting are high. Early termination requires the
consent of both sides—you cannot unilaterally close out a position before delivery.
Consequently, forwards are highly illiquid, have virtually no secondary market,
and are usually held until maturity. Conversely, forwards are tailor-made to suit
counterparties’ needs and objectives. Forwards can be designed and traded even in
situations when no futures contracts are available.
EXCHANGE TRADING OF A FUTURES CONTRACT 83
A futures exchange has associated with it a clearinghouse that clears executed
trades and guarantees contract performance by becoming the counterparty to every
trader. As such, it eliminates counterparty risk in the transaction. The clearinghouse
controls its counterparty risk by requiring traders to keep a margin (security) deposit
in their brokerage accounts, which are marked to market and settled at the end of
each trading day. We discuss this in greater detail in Chapter 9. By contrast, a forward
contract fixes a price for a future transaction that remains locked in over the contract’s
life, and settlement only takes place on the delivery date.
Futures are primarily designed as risk management tools. They are poor instru-
ments for buying and selling the commodity owing to the hassle of shipping to and
from delivery points in specific locations like Chicago and New York, which may ill
suit most traders. The US law takes this into consideration and distinguishes contracts
on how they end. If delivery is planned and regularly happens, then the contract is
likely to be classified as a forward; otherwise, it is a futures. During the 1960s and the
early 1970s, some brokerage firms were offering off-exchange forward contracts that
were standardized like futures (see Edwards and Ma 1992). Regulators tried to bring
them under their jurisdiction, but in 1974, the US Congress denied such incursions
by defining forwards as deferred delivery contracts that naturally end in delivery,
thereby permitting them to trade away from the exchange floor.
Although institutional differences are important, forwards and futures are also
different from an economic perspective because a futures contract has daily cash
flows, whereas a forward only has a final cash flow. This payout feature implies that
futures and forward prices are usually different, but more important, it implies that the
risks from holding the contracts differ. Futures face cash flow reinvestment risk with
changing interest rates, whereas forwards do not. Because interest rates are random
and constantly changing, this reinvestment risk is important, and it affects both pricing
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and hedging considerations.
4.5 Exchange Trading of a Futures
Contract
Next we sketch the trading of a futures contract, which has evolved over time to
become more electronic, which efficiently serves user needs. Moreover, this will also
help us to understand options exchanges, which evolved from futures exchanges.
EXAMPLE 4.4: Exchange Trading of Futures Contracts
Order Placement
■ Today is January 1. Ms. Longina Long and Mr. Shorty Short are trying to do the transaction of
Example 4.1 with exchange-traded futures contracts. They both want to trade fifty ounces of gold at
a fixed price on June 30.
84 CHAPTER 4: FORWARDS AND FUTURES
■ Gold futures trade on the NYMEX and the CBOT, both divisions of the CME Group, and on other
exchanges. Each regular contract is for one hundred troy ounces of gold, but there are also some
mini-contracts of smaller size. No gold futures contract matures in June. There are contracts for April
and July, but these months may not suit Long’s buying needs or Short’s selling desire. Standardization
allows quick trading, but there is a trade-off in that it reduces available choices.
■ Futures traders must first open margin accounts (security deposits in the form of cash or some
acceptable securities) with a Futures Commission Merchant (FCM).4 Old-fashioned Shorty calls
up his FCM and dictates a market order to sell one July gold futures at the best available price.
■ Technologically savvy Longina submits her market order through an FCM’s website. Futures traders
can place their orders in many other ways besides market orders (see Example 5.8 of Chapter 5 for
different order placement strategies).
Trade Execution
■ Shorty’s broker watching his computer screen sees that most trades are occurring at a little over $1,000.
He senses the market and submits a limit order to sell one contract at $1,001 or higher trade. The
trade is executed.
■ Long and Short pay their respective broker commissions on a round-trip basis. There is no charge
when entering a position, but the full charge is due at the time the futures transaction is closed. The
amount of commission varies from trader to trader. It could be as low as $10 per round trip trade for
a discount broker but far more for a full-service brokerage firm.
Clearing a Futures Trade
■ In traditional exchanges, before the market reopens, representatives of the clearing members take the
trade records to the exchange’s clearinghouse. The clearinghouse checks the records, and once the
records match, the clearinghouse clears the trade by recognizing and recording it.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ The clearinghouse also guarantees contract performance by acting as a seller to every buyer and as a
buyer to every seller (see Figure 4.5). The clearinghouse minimizes default risk by keeping margins
for exchange members, who, in turn, keep margin for their customers.
Settling a Futures Trade
■ A stock trade is settled when a buyer pays cash and gets the stock’s ownership rights from the seller.
In contrast, a futures trade has a special feature called daily settlement, by which profits and losses
are paid out daily. At the end of each trading day, a futures position is marked to market by crediting
or debiting the day’s gain or loss, respectively, to a trader’s margin account, an amount that equals the
difference between yesterday’s and today’s futures prices. If the futures price goes up, Long wins; if it
goes down, Short gains.
■ Marking-to-market resets the delivery price at maturity for outstanding futures contracts to the current
futures price in freshly minted futures. This process of marking-to-market also resets the value of a
futures contract to zero at the end of each trading day.
■ To understand marking-to-market, suppose Long buys one hundred ounces of gold from Short in July
for $1,001 per ounce. Suppose the next day’s settlement price is $1,004. Long would screech if you tell
EXCHANGE TRADING OF A FUTURES CONTRACT 85
her that she now has to buy gold in July at $1,004 because she thought she locked a price of $1,001!
However, she would agree to buy at $1,004 if you were to pay her the price difference between her
earlier price and the new one: ($1,004 − $1,001) = $3 per ounce. To handle this payment, $3 × 100
= $300 is taken from Short’s margin account and deposited into Long’s margin account. Wouldn’t
$300 earn daily interest in a bank account? And wouldn’t such daily receipts or payments happen in
an unpredictable fashion? The answer to both questions is yes. As noted earlier, these observations are
what differentiate futures from forward contracts. Chapter 9 explains how marking-to-market works
in greater detail.
FIGURE 4.6: Role of a Clearinghouse in a Derivative
Transaction
301.56........301.
.......103.71.....
99.02.......95.01.
Clearinghouse
Buyer Seller Buyer Seller
(Long) (Short) (Long) (Short)
Derivative trades are executed in an exchange where the buyer and seller meet
through brokers. After a trade is cleared, the clearinghouse becomes a seller to the
buyer and a buyer to the seller. It has counterparty credit risk but no price risk.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Closing the Contract
■ A month later, Longina sells one July gold futures to Miss Tallmadge, closes out her position, and
clears her slate. Her effective selling price after all the marking-to-markets is close to $1,001. Taxes,
interest charges on margin balances, and other market imperfections will make the realized price a bit
different from $1,001. Tallmadge is the new long vis-à-vis Shorty’s short.
■ Shorty decides to physically deliver the gold during the contract’s delivery period in July. He delivers
one hundred ounces of gold as per exchange requirements and collects an amount close to $100,100
from Tallmadge.
4 A FCM provides a one-stop service for all aspects of futures trading: it solicits trades, takes futures orders, accepts payments from
customers, extends credit to clients, holds margin deposits, documents trades, and keeps track of accounts and trading records.
Futures trading is becoming increasingly popular in different parts of the world,
including in countries that were previously hostile to it. Extension 4.1 discusses
futures trading in some developing countries, with particular focus on China and
India.
86 CHAPTER 4: FORWARDS AND FUTURES
EXTENSION 4.1: Futures Exchanges in China and India
Moving Commodities from Farms to Consumers
When you read Grimm’s fairy tales about the old days, you come across farmers who keep a portion of the
crop for family consumption and sell the rest in the market. Even today, this is true in many parts of the world,
where farmers’ markets are held on a regular basis. But when food moves to towns and cities, you need a supply
chain—a network of storage facilities, transporters, distributors, and retailers that take food from the farmer and
deliver it to the consumer.
Advanced economies use sophisticated technology at each stage of the supply chain to minimize costs and
maintain product quality. For example, fruits and vegetables are flash frozen, sent in refrigerated trucks, and kept
in temperature-controlled warehouses to extend their shelf life. Owing to this process, restaurants can offer an
incredible variety of “fish, flesh, and fowl” such as Maine lobster, Alaskan king crab, Hawaiian mahimahi, and
Russian caviar. And as Walmart and many other companies have demonstrated, sophisticated supply chains can
significantly lower costs that in competitive markets get passed on to consumers.
However, the story is different when commodities move from rural to urban areas in some of the poorest
countries. Farmers sell their excess crops to a middleman, who acts as a broker or dealer. The produce goes
through several middlemen before it reaches retail outlets. Significant waste can occur during repackaging and
reinspection by the middlemen, and poor handling and lack of refrigeration can further diminish the quantity. The
farmers get only a small fraction of the final price. True, these intermediaries serve a useful economic function, but
they are relevant only because the infrastructure is primitive and sophisticated risk management tools are absent.
Futures Trading in China
Consider China, a country home to over 1.3 billion people. During the 1980s and 1990s, China became
interested in developing sophisticated financial markets and institutions. Futures trading began in China in
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1993. However, during the first few years, “new exchanges opened with wild abandon, and speculative volume
ballooned” (Qin and Ronalds, 2005). Eventually, the Chinese authorities closed more than forty of these
exchanges. The Dalian Commodity Exchange, the Shanghai Futures Exchange, the Zhengzhou
Commodity Exchange, and the China Financial Futures Exchange (founded in 2006) are the only
operating futures exchanges remaining in China today. These exchanges are either fully electronic or use a
combination of open-outcry and electronic trading.
Interestingly, an October 12, 2009, article in the Wall Street Journal titled “China Targets Commodity Prices
by Stepping Into Futures Markets” reported that the leaders of Communist China are planning to use futures
exchanges “to fight back” foreign suppliers who “inflate [China’s] commodity prices.” By developing the three
commodities exchanges as “major players in setting world prices for metal, energy and farm commodities,” China
expects to be less susceptible to exchange prices elsewhere. Futures traders joke that “China is second only to the
weather in driving some commodity prices—but less predictable.” Chinese futures prices have begun affecting
global prices for many key commodities. For example, Chinese demand was a major contributor to huge swings
in crude oil prices during 2008–9.
Futures Trading in India
At the start of the new millennium, India removed a four-decade-long ban on commodity futures trading,
allowed resuscitation of moribund exchanges, and approved the opening of new ones. Soon afterward, during
2002–3, three major electronic multicommodity exchanges, the National Multi Commodity Exchange of
EXCHANGE TRADING OF A FUTURES CONTRACT 87
India Ltd., the Multi Commodity Exchange (MCX), and the National Commodity and Derivatives
Exchange Limited were formed.
In yet another development, September 2010 saw the opening of the United Stock Exchange (USE) of India.
Backed by government and private banks as well as corporate houses, the USE currently offers trading in currency
futures (on-the-spot exchange rate of Indian rupees against the dollar, euro, pound sterling, and yen) and options
(on the US dollar Indian rupee spot rate) but plans to expand its offerings to include interest rate derivatives.5
Commodity futures exchanges have also been founded in many other countries, including those in Latin America,
Eastern Europe, and Asia. However, Africa has been slow in adopting futures trading.
5
See www.useindia.com/genindex.php#.
In futures trading, the terms open interest and trading volume need careful distinction:
trading volume is the total number of contracts traded, whereas open interest is
the total number of all outstanding contracts, which may also be counted as the
total number of long (or short) positions. Trading volume is a measure of a market’s
liquidity: the more the volume is traded, the more active the market is. In contrast,
the open interest is a measure of the market’s outstanding demand for or exposure
to a particular commodity at the delivery date. Example 4.5 further explores the
distinction between these two concepts.
EXAMPLE 4.5: Volume versus Open Interest
■ Suppose that a July gold futures just becomes eligible for trading. Heather buys twenty of those
contracts from Kyle. Trade records will show the following (for convenience, they are also shown in
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Table 4.3):
- Heather is long twenty contracts.
- Kyle is short twenty contracts.
- Trading volume is twenty contracts.
- Open interest is twenty contracts.
■ Heather decides to reduce her exposure. She sells ten contracts to Tate. As a result,
- Heather is now long ten contracts.
- Tate is long ten contracts.
- Kyle is short twenty contracts, as before.
- Trading volume rises to thirty contracts.
- Open interest is twenty contracts, as before.
■ Kyle reduces his short position. He buys five contracts from Heather. Consequently,
- After selling five contracts, Heather is long five contracts.
- Tate is long ten contracts, as before.
- Kyle is short fifteen contracts after this trade.
- Trading volume, which adds up the number of trades, is thirty-five contracts.
- Open interest, which is the sum total of all outstanding long positions (or short positions), is now
fifteen contracts.
88 CHAPTER 4: FORWARDS AND FUTURES
TABLE 4.3: Trade Records for Gold Futures
Contracts
Kyle sells twenty contracts to Heather
Trader Long Short Trading Volume Open Interest
Heather 20 20 20
Kyle 20
Heather sells ten contracts to Tate
Trader Long Short Trading Volume Open Interest
Heather 10 30 20
Kyle 20
Tate 10
Heather sells five contracts to Kyle
Trader Long Short Trading Volume Open Interest
Heather 5 35 15
Kyle 15
Tate 10
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
4.6 Hedging with Forwards and Futures
Forwards and futures are widely used for hedging commodity price risks at future
dates. Consider a typical firm, which uses inputs to produce one or more outputs.
Input and output prices are susceptible to price fluctuations. Usually, a firm tries to
shed price risk for both its inputs and outputs, provided it is not costly to do so.
Then it can concentrate on production; in other words, it can focus on what it does
best. Obviously, it is a judgment call to decide which risks to cure and which to
endure.
Suppose that to reduce input price risk, a firm sets up a long hedge (or a buying
hedge) by taking a long position in a forward contract. For example, a company that
mines and refines gold in some remote area and uses natural gas to power its electricity
generators may find it prudent to buy a forward contract on natural gas. If natural
gas prices go up in the future, then the company pays more in the spot market, but
the spot market loss is largely offset by the gain on the forward contract. On the
flip side, if natural gas prices go down, the company will surrender some of its spot
market gains through losses on the forward position. This is always the case when one
employs a hedging strategy. Hedging with forwards (and futures) cuts both ways: it
SUMMARY 89
reduces the risk of financial loss from adverse price movements, but also takes away
potential gains from favorable price changes.
Alternatively, to reduce output price risk, the company can establish a short hedge
(or selling hedge) by taking a short position in a forward contract. If the company sells
gold forward, then it will remove or lessen the potential for loss (as well as gains) from
future spot price fluctuations. If the spot goes down, then the company loses money
when it sells gold in the cash market, but it profits from the forward contract. The
losses and gains are reversed when the spot price goes up.
Hedging is analogous to purchasing an insurance policy on the commodity’s price.
It pays off when prices move in an adverse direction. But as with all insurance policies,
there is a cost—the premium. If the price does not move in an adverse way, you paid
for insurance that you didn’t use. This is the cost. Risk-averse individuals will often
buy insurance despite this cost, and analogously, firms will often hedge their input or
output price risk.
4.7 Summary
1. The buyer and seller in a forward contract agree to trade a commodity on some
later delivery date at a fixed delivery (forward) price. Forwards are zero net supply
contracts. Forward trading is a zero-sum game. On the maturity date, the buyer’s
(or the long’s) payoff is the spot price minus the delivery price. The seller’s (or the
short’s) payoff is equal in magnitude but opposite in sign to the long’s payoff.
2. There are organized OTC (interbank) markets for trading forwards, which are
dominated by banks and other institutional traders. These markets are huge. For
example, foreign currency forwards attract billions of dollars of trade every day
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and have outstanding obligations measured in trillions of dollars.
3. The buyer and seller of a futures contract agree to trade a commodity at a fixed
delivery price on the maturity date. A futures contract is similar to a forward. The
long position holder in both these contracts agrees to buy the commodity, while
the short agrees to sell. The forward or futures price is set so that no cash changes
hands when the contract is created. This implies that the contracts have zero initial
value.
4. Exchange-traded derivatives are highly regulated contracts that trade in organized
exchanges. A clearinghouse clears trades and guarantees contract performance.
Traders are required to keep margins (security deposits) that are adjusted daily to
minimize default risk. Exchanges also act as secondary markets where traders can
take their profits or cut their losses and exit the market.
5. Many firms reduce price risk by trading futures and forward contracts. A firm can
lower input price risk by setting up a long hedge buying a forward and reduce
output price risk by establishing a short hedge selling a forward.
90 CHAPTER 4: FORWARDS AND FUTURES
4.8 Cases
The Dojima Rice Market and the Origins of Futures trading (Harvard Business
School Case 709044-PDF-ENG). The case blends business history with policy
issues surrounding the introduction of rice futures at the Dojima Exchange, the
world’s first organized (but unsanctioned) futures market.
United Grain Growers Ltd. (A) (Harvard Business School Case 201015-PDFENG).
The case considers how a Canadian grain distributor can identify and manage
various risks.
ITC eChoupal Initiative (Harvard Business School Case 297014-PDF-ENG). The
case discusses the use of Internet technologies and derivative contracts to help poor
farmers in rural India.
4.9 Questions and Problems
4.1. Define a forward contract. If a forward contract on gold is negotiated at a
forward price of $1,487 per ounce, what would be the payoff on the maturity
date to the buyer if the gold price is $1,518 per ounce and to the seller if the
gold price is $1,612 per ounce?
4.2. Define a futures contract.
4.3. Discuss the similarities and differences between forward and futures contracts.
4.4. What are the costs and benefits to a corn grower trading a forward contract? If
she is expecting a harvest in three months, should she buy or sell the derivative?
4.5. Which contract has more counterparty risk, a forward contract or a futures
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
contract? Explain your answer.
4.6. Discuss the benefits of standardization of a futures contract.
4.7. What are the two roles that a clearinghouse plays in the case of a futures
contract?
4.8. Ang, Bong, Chong, and Dong are trading futures contracts. Carefully identify
the trading volume and open interest for each trader from the following
transactions:
a. Ang buys five October silver futures from Bong.
b. Chong sells nine December silver futures to Bong.
c. Dong buys ten October silver futures from Chong.
d. Bong sells five December silver futures to Ang.
e. Ang sells three October silver futures to Chong.
4.9. Suppose you are a trader specializing in futures on corn, wheat, oats, barley,
and other agricultural commodities. From the following list, which risks do
you face (mark each with a yes or no): credit risk, market risk, liquidity risk,
settlement risk, operational risk, legal risk.
QUESTIONS AND PROBLEMS 91
4.10. For forward and futures contracts, what is the difference between physical
delivery and cash settlement?
4.11. Why does a futures contract have zero value when it is first written?
4.12. What is marking-to-market for a futures? Why is this marking-to-market
important for reducing counterparty risk?
4.13. Explain why a futures contract is a zero-sum game between the long and short
positions.
4.14. Are forward contracts new to financial markets? Explain.
4.15. Can you think of a reason why forward prices and futures prices on otherwise
identical forward and futures contracts might not be equal?
4.16. What is the OTC market for trading derivatives? How do OTC markets differ
from exchanges?
4.17. When holding a futures contract long, if you do not want to take delivery of
the underlying asset, what transaction must you perform? Explain.
4.18. If you are short a futures contract, why do you not have to borrow the futures
contract from a third party to do the short sale?
4.19. Is the futures price equal to the value of the futures contract? If not, then what
is the value of a forward contract when it is written? Explain.
4.20. Is the forward price equal to the value of a forward contract? If not, then what
is the value of a forward contract when it is written? Explain.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
5
Options
5.1 Introduction Price Bounds for American Puts
5.2 Options 5.5 Exchange-Traded Options
5.3 Call Options 5.6 Longs and Shorts in Different
Call Payoffs and Profit Diagrams Markets
The Call’s Intrinsic and Time Values 5.7 Order Placement Strategies
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Price Bounds for American Calls 5.8 Summary
5.4 Put Options 5.9 Cases
Put Payoffs and Profit Diagrams
5.10 Questions and Problems
The Put’s Intrinsic and Time Values
OPTIONS 93
5.1 Introduction
After introducing options in an intermediate finance class, a professor asked, “Is auto
insurance an option?” A serious student from the front row replied, “It is not an
option because state law requires it, but you have the option to choose from different
insurance companies.” Though the student was entirely correct, this was not the
answer for which the professor was hoping. He took a deep breath and went on to
explain that put options are very similar to insurance contracts because they restore an
asset’s value after a decline and that much of option jargon comes from the insurance
industry—and he resolved to frame his questions more carefully in the future.
As the example illustrates, the word option has many meanings in daily life.
Exchange-traded options are well-defined contracts with specific features. This
chapter describes these option features and discusses how they trade. At the chapter’s
end, we discuss how traders buy or sell options for various strategic objectives.
5.2 Options
Options come in two basic types: calls and puts, names you have heard before.
A call option gives the buyer:
- the right, but not the obligation
- to buy a specified quantity of a financial or real asset from the seller
- on or before a fixed future expiration date
- by paying an exercise price agreed on today.
The buyer is also called the holder or the owner. The seller is known as the grantor,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the issuer, or the writer. The known future expiration date is also called the maturity
date, and the fixed price is known as the exercise price or strike price. The holder
exercises the call when she exercises her right and buys the underlying asset by
paying the exercise price; otherwise, she lets the call expire worthless.
By contrast, a put option gives the right to sell. Specifically, a put option gives the
owner the right to sell the asset to the writer at the strike price until the expiration
date. A buyer exercises the put when he sells the underlying asset and receives
the strike price from the put writer; otherwise, he can let the put expire worthless.
Options do not come free—the option buyer must pay the writer a fee (option’s
price or premium) for selling those rights.
The terms call and put come from what the buyer can do with these options. A
call gives the option to buy, that is, to call the asset away from the writer. Conversely,
a put gives the option to sell or to put the asset to the writer. In each case, the writer
stands ready to take the other side of the buyer’s decision. We say that a call buyer is
bullish because she expects the underlying asset to go up in value, and a put buyer is
bearish because he expects the asset price to decline. Naturally, their counterparties
hold opposite market views: the call writer is bearish and the put writer is bullish.
We also say that the option (call or put) buyer is long the option, so the writer ends
up with a short position.
94 CHAPTER 5: OPTIONS
Notice that calls and puts have a lot in common with insurance policies and may
be viewed as insurance contracts for hedging or risk reduction. If someone hits a car,
the insurance company pays the owner enough money (subject to some deductibles)
to restore the car to its original value. Likewise, a put pays the holder in case of a
meltdown in the asset’s value. A call buyer benefits from an asset price rise but does
not suffer from its decline. The call buyer pays for insurance to avoid losses if the
stock falls. The call writer provides this insurance. For these reasons, an option price
is called a premium. A premium is what you pay for the insurance.
Calls and puts are examples of plain vanilla or standard derivatives. Later
we will consider nonstandard or exotic derivatives, which are derivatives with
more complex payoffs. Plain vanilla options are usually defined in one of two ways:
American or European. A European option can only be exercised at the maturity
date of the option, whereas an American option can be exercised at any time up
to and including its expiration date. What we defined earlier were calls and puts of
the American type. The adjectives European and American tell how the options differ
in their exercise choices and have nothing to do with where they trade—in Europe,
America, Africa, Asia, Australia, or even in Antarctica. European-style options trade
in the US and across the Atlantic, and American-style options transact in Europe.
Both American and European options have the same value on the expiration date,
if the American option remains unexercised. Before expiration, however, American
options are at least as valuable as their European counterparts. They duplicate what
European options can do, but they also offer an early exercise feature, which is more.
You don’t pay less when you get more. Chapter 6 will show how the assumption
of no arbitrage makes this happen. Such simple insights are useful when trading
options.
The important question studied in later chapters is, how do you find an
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
option’s premium? This has bewildered academics for a long time. A break-
through came in 1973 with the development of the seminal Black–Scholes–
Merton (BSM) option pricing model (see Chapter 19). The importance of the
BSM model can hardly be overstated—it helped spawn the entire field of
derivative pricing that we see today. However, option pricing still remains
a Herculean task. For example, there is no generalized closed-form solution
for finding an American option’s premium.1 This is important because most
exchange-traded options are of the American type, and we need to price
them. When closed-form solutions are not available, numerical procedures (com-
puter programs) are used instead (see Wilmott 1998; Duffie 2001; Glasser-
man 2003). There remain many open questions, and option pricing remains
an intellectually challenging field to which interesting contributions can still
be made.
1
A closed-form solution is an exact mathematical formula like x = (a + b)2 . If you know a and b, you
can easily solve for x. The BSM model solves an option’s price in terms of known variables and functions
like the cumulative normal distribution.
CALL OPTIONS 95
5.3 Call Options
Call Payoffs and Profit Diagrams
Option payoffs on the expiration date or at early exercise are easier to understand
in a perfect market, where market imperfections like transaction costs, taxes,
and trading restrictions are assumed away. In reality, the market contains such
imperfections. The brokers who match option trades charge commissions for their
services. A dealer will quote an ask price (say, $4) and a bid (say, $3.90) for the
option premium. Large trades tend to have market impact, and the entire order may
not be filled at the price posted by a dealer. Unfortunately, these real-world features
detract from understanding the basic payoffs. Hence we study option payoffs under a
perfect market structure. Once you know how these contracts work, you can adjust
for these imperfections.
Example 5.1 considers a call buyer’s and writer’s payoffs and profit diagrams on the
expiration date.
EXAMPLE 5.1: European Call Option Payoff and Profit Diagrams
Call Buyer’s Payoff
■ Today is January 1 (time 0). An April 100 European call option on the stock of Your Beloved Machines
Inc. (YBM) is created when Ms. Longina Long pays the premium c = $4 to Mr. Shorty Short, who
becomes the writer. This call has a strike price of K = $100, and it ceases trading on the third Friday
of April. This is the only day this European call can be exercised (time T); see Figure 5.1 for this
option’s timeline.2
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Long’s exercise decision depends on the stock price S(T) on the fateful expiration date. If S(T) is
$106, then Long exercises the call because the stock is worth more than the strike (traders say that
the option is in-the-money). She pays the strike, receives the stock worth S(T), and makes [S(T) –
K] = $6 in the process. The $4 premium is a sunk cost and does not affect the exercise decision.
■ Conversely, if the stock is worth less than the strike, then the option expires worthless—out-of-the-
money, in traders’ parlance. If Long is keen to acquire the stock, she should directly tap the stock
market. It’s imprudent to exercise out-of-the-money options. For example, if S(T) is $92, she gets 92 –
100 = –$8 by exercising the call, an outcome she should avoid. And when S(T) equals the strike price,
the option is at-the-money, and the long is indifferent regarding exercise. In reality, a comparison of
transaction costs for purchasing the stock is likely to influence her decision. For example, suppose Long
is planning to purchase a thinly traded stock. If the stock’s spread is too high, Long may exercise an
at-the-money call (or even a call that is slightly out-of-the-money) to more cheaply acquire the stock.
■ True to the proverb “a picture says a thousand words,” patterns emerge when we graph these payoffs in
a payoff diagram. These diagrams plot the stock price at expiration, S(T), along the horizontal axis
and the option payoff (or gross payoff ) along the vertical axis. Long call’s payoff starts at 0 value when
the stock price is zero and lies flat along the axis until the stock price reaches K = $100 (see Figure 5.2).
Beyond this, the payoff increases at a 45 degree angle, rising by a dollar for each dollar increase in S(T).
96 CHAPTER 5: OPTIONS
Call Writer’s Payoff
■ Because option trading is a zero-sum game, Short’s call’s payoff is the mirror image across the horizontal
axis of Long’s payoff. The writer loses nothing at expiration until S(T) reaches $100, beyond which
his payoff declines at a negative 45 degree angle. When S(T) = $106, Short loses $6, which is Long’s
gain. When S(T) = $92, Short loses nothing.
Call Buyer’s Profit
■ Payoff diagrams fail to reveal why Short, who has a zero or negative payoff on the expiration date,
will get into this venture. This limitation is overcome by a profit diagram, which adds the option’s
premium to the trader’s payoff. The profit diagram plots S(T) along the horizontal axis but has profit
(or loss) from option trading along the vertical axis. For convenience, we ignore any interest that the
option premiums may earn in the margin account. This allows us to combine the cash flows at the
option’s initiation and expiration dates on the same diagram.
FIGURE 5.1: Timeline for a Call Option
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Today (start date) Intermediate dates Expiration/maturity date
(Time 0) (Time T)
Buyer (Long) pays the Long can exercise an If S(T) > K, long exercises:
call premium and gets American option. Long gives up the call,
the call from the writer pays the strike price, and
(Short) (received as an gets the stock from the
electronic account). writer. Long nets
S(T) – K.
Traders can close out If S(T) <
– K, call expires
their positions. worthless.
(If none of these
happens, traders meet
at time T).
CALL OPTIONS 97
■ If S(T) does not exceed the strike price K = $100, Long loses c = $4, the premium she paid for the
call. This is her maximum loss. When S(T) exceeds $100, Long will definitely exercise the call. A
dollar increase in S(T) raises the value of her position by a dollar and cuts her losses. Consequently,
the profit graph is a horizontal straight line at –$4 until it reaches the strike price, where it increases
at a 45 degree angle. When S(T) = $104, Long’s $4 gain exactly offsets her initial $4 investment and
gives “zero profit.” When S(T) goes beyond $104, it’s even sweeter. Technically speaking, the call
holder has limitless profit potential (see Figure 5.2). Focusing on our familiar points, when S(T) =
$106, Long has a net gain of $2 because she makes $6 but she paid $4 for the call. Conversely, when
S(T) = $92, Long does not exercise and loses only the $4 premium.
FIGURE 5.2: Payoff and Profit Diagrams for an April 100
Call Option on the Expiration Date
Payoff Payoff
Long call
45˚
K
0 0
100 S(T), stock price –45˚ S(T)
at expiration
Short call
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Profit Profit
Buy call
45˚ c
100 104
0 0
–4 c+K S(T) K –45˚ S(T)
Sell call
Call Writer’s Profit
■ When S(T) does not exceed the strike, the call writer keeps the $4 premium collected up front. This is
also the maximum profit. When S(T) increases above $100, Long’s exercise reduces this amount. Zero
profits occur when S(T) = $104. Below this level, Short plunges deeper into losses (see Figure 5.2).
98 CHAPTER 5: OPTIONS
In the example considered, when S(T) = $106, the writer has a loss of $2. When S(T) = $92, Long
does not exercise, and Short’s profit is $4. Because option trading is a zero-sum game, the total gain
of both traders is zero. In reality, brokerage commission lowers the profits accruing to each trader.
■ Notice what the diagrams reveal: call writing can generate far larger losses than call buying. The
maximum loss to the buyer is the premium. In contrast, the writer can literally fall into a bottomless
abyss where the maximum loss is unbounded. Recognizing this asymmetry, the options exchange
requires a call writer to hold more funds in a margin account than a buyer.
2 The OCC (www.optionsclearing.com/about/publications/expiration-calendar-2010.jsp) and exchange websites label the
following Saturday as the expiration date. This is a technicality. The third Friday, which we label as time (date) T, is the relevant
economic date for our purpose because it is (1) the day when the option stops trading, (2) the only day a European option can
be exercised, and (3) the last day an American option may be exercised. For simplicity, we will use the terms exercise date and
expiration date interchangeably.
The Call’s Intrinsic and Time Values
An American call is identical to a European call, except that it can be exercised any
time until expiration. Remember that if an American call is not exercised early, then
it has the same value as a European call on the expiration date so that the payoff
diagrams are identical. And prior to expiration, the exercise strategy for an American
option is guided by two facts: (1) one should never exercise an out-of-the-money
option, and (2) when exercising an in-the-money call early, one receives the stock
price minus the strike price. In other words, if an American call is exercised early,
then its payoff looks like that of a European call on the expiration date.
We summarize these insights in Result 5.1.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
RESULT 5.1
Intrinsic (or Exercise) Value of a Call Option
A call holder’s payoff at expiration or if exercised early is
S − K for S > K
Call’s intrinsic value ≡ (5.1)
{0 otherwise
which may be written as
Call’s intrinsic value = max (S − K, 0) (5.2)
where S is the stock price on the exercise date, K is the strike price, and max
means the maximum of the two arguments that follow.
This is a call option’s boundary condition. The payoff is also referred to
as the call’s intrinsic or exercise value.
CALL OPTIONS 99
What does this mean? For S < K, the term (S – K) is negative, the call is not
exercised, and 0 is the value. When S ≥ K, the call’s value equals S – K. The profit
is obtained by subtracting the call’s premium from the preceding payoff.
This result has a simple corollary. It implies that if the stock price hits zero, then
the call price is also zero. Indeed, if the stock hits zero, it has no future, and it is never
going to take a positive value again. Consequently, a call, which gives the right to
purchase the stock in the future, must also be zero.
The intrinsic value is only part of an option’s premium. The remainder is the
additional value accruing to an option because the stock price may increase further
before maturity. Consequently, we define the time value of a European or American
option as
Time value of an option = Option price − Option’s intrinsic value (5.3)
Time value declines as one approaches the expiration date and (by its definition)
becomes zero when the option matures. Example 5.2 computes the time value of a
call option.
EXAMPLE 5.2: Time Value of a Call Option
■ On some date t before expiration, suppose that YBM’s stock price S(t) = $101 and the call price c(t)
= $2.50. If the strike price K = $100, then
Call option’s time value
= Call price − Call’s intrinsic value
= c (t) − [S (t) − K]
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
= 2.50 − (101 − 100)
= $1.50
Price Bounds for American Calls
Although we are yet to price options, the boundary condition and some simple
reasoning can determine what call option prices are admissible and what prices aren’t
(see Example 5.3).
EXAMPLE 5.3: Price Bounds for an American Call Option
■ As in Example 5.1, consider an American call option with a strike price K = $100. Its price can never
exceed the underlying stock’s price. Indeed, the call is a security that gives the holder the right to buy
the underlying stock by paying the strike price. If you have to choose between the two—a gift of the
stock versus getting the stock by paying some money—which one would you prefer? Getting the gift,
of course, which implies that the stock is worth more than the option.
100 CHAPTER 5: OPTIONS
■ Consequently, a call’s price can never exceed the underlying stock’s price. Figure 5.3 plots today’s
stock price S(t), written as S for simplicity, along the horizontal axis and has both the stock and the
call price along the vertical axis. The upward sloping 45 degree line from the origin represents the
stock price. The call price can never exceed this value, and the 45 degree line forms an upper bound.
A rational investor will never let a call’s price fall below the boundary condition given by Result 5.1,
max(S – 100, 0)max(S – 100, 0), which forms a lower bound. This is the kinked line that starts at the
origin and increases at a 45 degree angle when the stock price is $100.
FIGURE 5.3: Price Bounds for an
American Call
Prices
Upper bound
Lower bound
Call prices
lie here
45˚
0
K = $100 S, stock price
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Let us illustrate these bounds with some numbers:
- When the call is out-of-the-money, say, the stock price is $90, the time value prevents the call price
from being zero. So it will lie above the horizontal axis. However, the call price can never exceed
the stock price of $90, and hence it will lie below the 45 degree line emanating from the origin.
- When the call is in-the-money, say, the stock price is $115, the call is worth at least S – K = 115 –
100 = $15 if exercised. So the call price can never fall below the lower bound, and of course, it can
never exceed $115, which lies on the upper bound.
5.4 Put Options
Put Payoffs and Profit Diagrams
We illustrate a put option’s payoff and profit diagrams with Example 5.4.
PUT OPTIONS 101
EXAMPLE 5.4: European Put Option Payoff and Profit Diagrams
Put Option Payoffs
■ Today is January 1 (time 0). Ms. Long pays Mr. Short p = $6 and buys a May 50 European put option
on Boring Unreliable Gadgets Corp (BUG). This put gives Long the right to sell BUG stocks on the
exercise date, the third Friday of May (time T), at the strike price K = $50.
■ A put’s timeline is similar to that for the call given in Figure 5.1. If the put is not closed out early, the
traders meet on the expiration date T. If the stock price is less than the strike price (the put is in-
the-money), Long exercises the put, receives K dollars, surrenders the stock worth S(T), and makes
[K – S(T)] in the process. For example, if S(T) = $40, then Long sells the stock for $50 and makes
$10. Conversely, if S(T) is greater than K (the put is out-of-the-money), it’s unwise to exercise. For
example, if S(T) = $55, Long gets (50 – 55) = – $5 if she exercises and zero if she doesn’t. Exercise
is irrelevant if BUG closes at $50 (the put is at-the-money).3
■ These considerations enable us to draw the payoff diagrams. For large stock prices, the put holder’s
payoff has zero value, and it lies on the horizontal axis. As the stock price declines and dips below the
strike, the payoff forms a kink at K and increases at a 45 degree angle, increasing by a dollar for each
dollar decrease in the stock price (see Figure 5.4). The profit peaks when the stock is worthless, but
Long can sell it for K dollars as per contract terms.
FIGURE 5.4: Payoff and Profit
Diagrams for a May
50 Put Option on
the Expiration Date
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Payoff Payoff
50
Long put
45˚ K
0 0
50 S(T), stock price –45˚ S(T)
at expiration
Short put
–50
Profit Profit
44
Buy put
45˚ 6
0 K 0 44
–6 –45˚ S(T)
S(T)
Sell put
–44
102 CHAPTER 5: OPTIONS
■ Because option trading is a zero-sum game, the put writer’s payoff is the mirror image across the x-
axis of the buyer’s payoff. When S(T) is greater than $50, the put seller loses nothing, and the payoff
coincides with the horizontal axis. When the stock goes below the strike, the payoff decreases below
the horizon line at a 45 degree angle starting at K, and it reaches the nadir when the stock becomes
worthless.
Put Option Profit
■ The profit diagram adjusts the payoff diagram for an option’s premium. If S(T) is greater than the
strike price K = $50, Long loses the premium p = $6, which is the most she can lose. When S(T)
falls below $50, Long exercises the put—each dollar decrease in S(T) raises the value of her position
by a dollar and cuts her losses.
■ Consequently, the profit graph is a horizontal line at – $6 for high values of S(T). It increases upward
at a 45 degree angle when S(T) touches K = $50 on its downward journey, and it cuts the horizontal
axis at the zero-profit point when S(T) = $44. It hits the vertical axis and makes an intercept of
$44 when S(T) = 0 (see Figure 5.4). The profit graph for a put writer is the mirror image of the
buyer’s graph across the horizontal axis. Notice that both the maximum profit and the biggest loss are
bounded for a put option.
■ Next we illustrate these profit (or loss) computations. For S(T) = $40, Long makes (50 – 40) = $10
by exercising the put. However, he paid $6 for the option, netting a profit of 10 – 6 = $4, which is
also Short’s loss. For S(T) = $55, Long does not exercise the put, and her loss is the $6 premium,
which is also Short’s profit.
3 Transaction and liquidity costs may slightly modify this decision. If the long has a substantial holding in the underlying stock
that is thinly traded and wants to sell, then she may exercise an at-the-money or slightly out-of-the-money put option to sell the
stock at minimum transaction cost.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The Put’s Intrinsic and Time Values
An American put may also be exercised before expiration. Of course, if an American
put is not exercised early, then the American and European puts have identical value
on the expiration date so that their payoff diagrams are identical. Before expiration,
the exercise strategy for an American option is determined by two facts: (1) one
should never exercise an out-of-the-money option, and (2) in case of early exercise
of an in-the-money put, the value is the strike price minus the stock price. When
exercised early, an American put’s payoff looks like that of a European put on the
expiration date.
PUT OPTIONS 103
These observations are summarized as Result 5.2.
RESULT 5.2
Intrinsic (or Exercise) Value of a Put Option
A put holder’s payoff at expiration or if exercised is
K − S for S < K
Put’s intrinsic value ≡ (5.4)
{0 otherwise
which may be written as
Put’s intrinsic value = max (K − S, 0) (5.5)
where K is the strike price, S is the stock price on the exercise or expiration
date, and max means the maximum of the two arguments that follow.
This is a put option’s boundary condition. The payoff is also referred to as
the put’s intrinsic (or exercise) value.
For example, consider a put with a strike price of K = $50. When S(T) = $55,
K – S(T) = – $5. As the maximum of –$5 and 0 is clearly 0, the put value p(T) = 0
at expiration. When S(T) = $40, then K – S(T) = $10 is the intrinsic value. For
S(T) < K, the positive term [K – S(T)] is the put’s payoff. When S(T) ≥ K, the term
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
K – S(T) is no longer positive, and zero becomes the put’s intrinsic value.
Example 5.5 illustrates the computation of a put’s time value.
EXAMPLE 5.5: Time Value of a Put Option
■ At time t, let BUG’s stock price S(t) = $48 and its put price p(t) = $3 for a strike price K = $50. Then
Put option’s time value
= Put premium − Put’s intrinsic value
= p (t) − [K − S (t)]
= 3 − (50 − 48)
= $1
Price Bounds for American Puts
As with calls, we can use the boundary condition to determine price bounds for put
options (see Example 5.6).
104 CHAPTER 5: OPTIONS
EXAMPLE 5.6: Price Bounds for an American Put Option
■ As in Example 5.4, consider an American put option with a strike price of K = $50. Its value can
never exceed the strike price because the put gives the holder the right to sell the underlying stock at
the strike price. Thus the maximum payoff from the put is the exercise price K = $50. This forms an
upper bound for the put price and is depicted by the horizontal line at $50 in Figure 5.5.
■ The put’s price can never fall below max(0, 50 – S), the boundary condition given by (5.3), which
forms a lower bound. This is the line that originates at $50 on the vertical axis, decreases at a 45 degree
angle until it hits $50 on the horizontal axis, and then continues along the horizontal axis.
■ We illustrate these bounds with an example (see Figure 5.5):
- When the put is in-the-money, say, the stock price is $10, the put price cannot fall below its exercise
value 50 – 10 = $40. Because the put price cannot rise above its upper bound of $50, it lies between
$40 and $50.
- When the put is out-of-the-money, say, the stock price is $60, the put price cannot fall below the
horizontal axis because no one would exercise and lose money. Of course, it cannot exceed its upper
bound: the flat line at $50 above the horizontal axis.
FIGURE 5.5: PRICE BOUNDS FOR AN
AMERICAN PUT
Prices
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
$50
Upper bound
Put prices
lie here
Lower bound
0
K = $50 S, stock price
Put option prices cannot lie in the shaded regions above
the upper bound or below the lower bound.
EXCHANGE-TRADED OPTIONS 105
5.5 Exchange-Traded Options
Relative to futures contracts, exchange-traded options are newcomers. They started
trading in 1973 with the opening of the Chicago Board Options Exchange
(CBOE). Today options trading is dominated by big exchange groups, which offer
diverse product lines (stocks, bonds, options, futures, etc.). The Options Clearing
Corporation (OCC) clears the bulk of exchange-traded options in the US. As such,
the OCC becomes the guarantor of contract performance to each buyer and seller.
As discussed before, a clearinghouse makes derivatives safer to trade, helps develop a
secondary market, allows anonymous trading, and eliminates the need to track one’s
counterparty. These benefits have greatly contributed to option market growth.
Example 5.7 considers a hypothetical options trade.
EXAMPLE 5.7: Exchange Trading of a Call Option
Order Placement
■ Today is January 1 (time 0). Ms. Longina Long and Mr. Shorty Short have opened margin accounts and
have been approved for trading options with their respective brokers. Unbeknownst to each another,
they simultaneously decide to trade ten April 100 call options on Your Beloved Machines Inc. (YBM).
These American calls have a strike price of K = $100 and an exercise date of the third Friday of April.
Because each option contract is on one hundred shares, ten contracts reflect gains or losses from one
thousand shares.
■ Tech-savvy Longina flips open her laptop and logs onto her brokerage account. Finding that YBM
April calls have a bid price of $4.00 and an ask–offer price of $4.05, she types in a limit order that
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
instructs her broker to buy ten contracts at $4.00 or a lower price. Old-fashioned Shorty calls up his
broker and places a market order that is immediately executed. Small orders are matched via computers
and are filled and confirmed within a few seconds. Most option trades are executed on the exchange
floor.
■ Suppose Long’s and Short’s orders are sent via computer to their respective brokerage firms.
Trade Execution, Clearing, and Settlement
■ Shorty’s broker accepts Longina’s limit order, and ten contracts are executed at c = $4.00. Longina sees
the trade confirmation on her web page, and Shorty gets a confirmatory phone call from the broker’s
representative. Vital transaction data also go into the exchange’s main computer system.
■ Commission rates vary. A discount broker typically charges a fixed rate, which may be as small as $5 to
$20 per trade, and a variable rate, which typically ranges from 50 cents to $2 for each option on one
hundred shares. Many firms charge lower commissions for (1) internet orders; (2) active traders; (3)
the simplest orders like market orders; (4) small-sized orders, because they are unlikely to be based on
privileged information; and paradoxically, (5) larger orders, to give customers a break. Brokers charge
two commissions: one at the time of buying (or selling) and a second commission when selling (or
buying back) or exercising the option.
106 CHAPTER 5: OPTIONS
■ Suppose Long’s discount broker charges her $10 for the trade plus a dollar per contract—$20 in
all. Consequently, Long’s brokerage account is debited an amount equal to the buying price plus
commission:
Price × 100 × (Number of contracts) + Commission
= 4 × 100 × 10 + 20
= $4, 020
■ Shorty has a full-service broker who charges $50 for the trade plus $3 per contract—$80 in all. Short’s
margin account is credited an amount equal to the selling price less the commission:
Price × 100 × (Number of contracts) − Commission
= 4 × 100 × 10 − 80
= $3, 920
Because selling options is a risky transaction, Short’s broker requires him to keep the entire proceeds
as well as some extra funds as a security deposit in the margin account.
Closing the Contract
■ Being exchange-traded contracts, options have secondary markets. Traders retain the flexibility of
taking their profits or cutting their losses and walking away. Long may sell the option to a third party
and close out her position—she may win or lose in the process. Quite independently, Short can buy
the option in a closing transaction, and a third trader then becomes the writer. Because this is an
American call, Long can exercise at any time until the maturity date.
■ Fluctuations in option prices affect the margin account’s value. Suppose that YBM rallies tomorrow
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and the call price increases by $3. Then Long’s margin account gains $3,000, and Short’s margin
account declines by the same amount. If his maintenance margin is breeched, Shorty has to add funds
to his margin account.
■ Suppose that the option is neither exercised nor closed out early so that the counterparties meet again
on the expiration date. To avoid the hassle of paying the broker a fee for exercising and yet another
fee for trading YBM shares, they decide to close out their positions just before the option’s expiration
and trade at a $7 premium per option. Then Long’s profits are
Sale proceeds − Cost
= (7 × 100 × 10 − 20) − 4, 020
= $2, 960
and Short’s profit (remember, his brokerage costs are higher) is
Sale proceeds − Cost
= 3, 920 − (7 × 100 × 10 + 80)
= − $3, 160
ORDER PLACEMENT STRATEGIES 107
5.6 Longs and Shorts in Different Markets
Market practitioners often use the terms long and short to convey the notions of
buying and selling. But these terms have different meanings in different markets.
For derivatives, long means “buy,” and short means “sell.” Derivatives trade in zero
net supply markets, where each buyer has a matching seller. Long and short in the
derivatives markets are just opposite sides of the same trade. Consequently, it is a
zero-sum game because one trader’s gain is the other’s loss.
The short seller in the stock market does something more complex. She borrows
shares, sells them, and buys back those shares at a later date to return them to the
lender. For futures and forwards, both a long and a short have obligations to transact
at a future date. The long agrees to buy the asset, and the short agrees to sell.
Options give the buyer the right to buy (call) or sell (put) at a fixed price until a
fixed date. No matter whether it’s a call or a put, the buyer who is long has an option.
They decide to exercise or not. In contrast, the writer, who holds a short position,
has an obligation. The short needs to fulfill the option’s terms as chosen by the long.
For this situation, the buyer pays the writer a premium at the start.
5.7 Order Placement Strategies
All traders face the same dilemma: to trade immediately or to wait to see if one can
get a better price. If you want to transact immediately, submit a market order; however,
if you can wait for a better price, then submit contingent orders that depend on
the price, time, and other relevant factors. But there is a trade-off: market orders may
get traded at a bad price, whereas price-contingent orders may never get filled.
Example 5.8 considers different types of order placement strategies for trading
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
stocks in fairy-tale (hypothetical) situations. These same order placement strategies
also apply to futures and options. Let’s start our discussion with stocks for easy
illustration and then show how the argument extends to derivatives like options.
EXAMPLE 5.8: Order Placement Strategies
A Market Order
■ A dealer quotes an ask–offer price of $100.10 and a bid price of $100.00 per share for YBM stock.
Hoping that it will rally, you want to buy YBM in a hurry. You can place a market buy order,
which gets executed immediately at the best price your broker can find—very likely at $100.10. You
may be forced to pay a bit more or enjoy paying slightly less if a spate of buy or sell orders suddenly
comes in. Unless there is a severe market breakdown, market orders are always filled.
■ A market order is also the simplest way to buy options and futures. Recall in the previous example
that Shorty placed a market sell order and sold the option for $4, which was the bid price. If gold
futures prices are quoted at $1,500.10 to sell and $1,500.00 to buy, then a market buy order is likely
to get executed at $1,500.10.
108 CHAPTER 5: OPTIONS
A Price-Contingent Order
■ Sometimes you may get a better price by waiting. You can place a limit order that must be executed
at the stated or a better price or not executed at all. One must specify whether this is a day order or
a good-until-canceled order. An unfilled day order automatically cancels at the day’s end, while a
good-until-canceled (GTC) order stays open almost indefinitely in the specialist’s order book until
it is closed. To eliminate stale orders, many brokers automatically cancel GTC orders after a significant
time period has passed, for example, 60 days.
■ Suppose you place a limit buy order at $99.00. If the stock price decreases from $100.10 and your
trade gets executed, then you obtained a $1.10 price improvement. However, you also run the risk
that the stock may rally and the order may never be filled. Similarly for an option on YBM that is
being quoted at a bid price of $4.00 and an ask price of $4.05, a trader hoping to benefit from market
fluctuations may place a limit sell order at $4.50.
A Time-Contingent Order
■ A time-of-day order specifies execution at a particular time or interval of time. For example, you
may request execution at 2:00 pm in the afternoon, or you may request execution between 2:30pm
and 3:00pm.
Combinations of Related Orders
■ Sometimes traders place orders in combination. For example, a spread order may instruct the broker
to execute the order when the price difference between two securities reaches a certain value. A
cancel order or a straight cancel order completely cancels a previous order. A cancel replace
order replaces an existing order with a new one.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Our discussion also applies to options and futures trading, and we can similarly
place such orders for bonds, the Treasury securities introduced in Chapter 2.
5.8 Summary
1. A call option gives the buyer (the “long”) the right but not the obligation to buy
a specified quantity of a financial or real asset from the seller (the “short”) on or
before a fixed future expiration date by paying an exercise price agreed today. A
put gives the holder the right to sell. A put is similar to an insurance contract that
pays in case of a decline in an asset’s value. The writer is paid a premium for selling
such rights.
2. A European option can only be exercised at the maturity date, whereas an
American option can be exercised at any time up to and including the expiration
date. The names “American” and “European” do not depend on where the option
trades. Most exchange-traded options in the US are of the American type.
3. A call holder’s payoff at expiration or if exercised early, its intrinsic value, is
the stock price minus the strike price when the stock price is greater than the
strike (call is in-the-money) and zero otherwise (call is out-of-the-money). A put’s
QUESTIONS AND PROBLEMS 109
intrinsic–exercise value is the strike price minus the stock price when the stock
price is less than the strike and zero when the stock price is greater than or equal to
the strike. Intrinsic value is only part of an option’s value. The rest is the time value,
which takes into account that the option’s value may increase before it expires.
4. Payoffs to different option positions as a function of the stock price at expiration
generate well-known patterns, which can be depicted in a payoff diagram. Profit
diagrams adjust these diagrams for the option’s premium.
5. Traders in a hurry place market orders that are immediately executed. Traders who
can wait in the hope of a better price strategically submit limit orders that may be
contingent on price, time, and other relevant factors or submitted in combination
with other orders. There is a trade-off: market orders may get traded at a bad
price, while price improvement orders may never be filled. Contingent orders are
possible for trading stocks, options, and futures.
5.9 Cases
Goldman, Sachs, and Co.: Nikkei Put Warrants—1989 (Harvard Business School
Case 292113-PDF-ENG). The case illustrates how investment banks can design,
produce (hedge), and price put warrants on the Nikkei Stock Average.
Nextel Partners: Put Option (Harvard Business School Case 207128-PDF-ENG).
The case examines issues surrounding Nextel’s shareholders’ vote to exercise a put
option that requires the company’s largest shareholder, Sprint Nextel Corp., to
purchase all the shares it does not already own.
Pixonix Inc.—Addressing Currency Exposure (Richard Ivey School of Business
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Foundation Case 908N13-PDF-ENG. Harvard Business Publishing). The case
discusses the impact of exchange rate fluctuations on a firm’s cash flows and how
derivatives can be used to manage this risk.
5.10 Questions and Problems
5.1. Explain the differences between an option’s delivery and expiration dates.
5.2. Explain the differences between European and American options. How do the
names relate to the geographical continents?
5.3. Explain the differences between call and put options. Are they just the exact
opposites of each other?
5.4. Explain the differences between long and short in the stock market. Explain
how one shorts a stock.
5.5. Explain the differences between long and short in the option market. What is
the difference between an option buyer and writer?
5.6. Explain the differences between an option’s intrinsic and time values. How do
these relate to the option’s price?
110 CHAPTER 5: OPTIONS
5.7. Explain the differences between an option’s maturity and exercise date. When
does exercise take place for a European option? For an American option?
5.8. Explain and carefully describe the following four security positions, drawing
payoff diagrams wherever necessary to support your answer:
a. short a forward contract with a delivery price of $100
b. short selling a stock at $100
c. going short on an option with a strike price of $100
5.9. Suppose that the current price of platinum is $400 per ounce. Suppose you
expect that in three months the price will increase to $425. You are worried,
however, that there is a small chance that platinum may fall below $390 or even
lower. What securities can you use to speculate on the price of platinum?
5.10. Why is an American option worth at least as much an otherwise similar
European option? Is there an exercise strategy that one can use to turn an
American option into a European option? Explain.
5.11. You ask your broker for a price quote for (a fictional company) Sunstar Inc.’s
March 110 calls. She replies that these options are trading at $7.00. If you want
to buy three contracts, what would be your total cost, including commission?
Assume that the broker charges a commission equal to a flat fee of $17 plus $2
per contract.
5.12. “Because a call is the right to buy and a put is the right to sell, a long call
position will be canceled out by a long put position.” Do you agree with this
statement? Explain your answer.
5.13. The short position in a put option has the obligation to buy if the option is
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
exercised. Isn’t this counter to the notion that a short position indicates a “sell”?
Explain your answer.
5.14. The following option prices are given for Sunstar Inc., whose stock price equals
$50.00:
Strike Price Call Price Put Price
45 5.50 1.00
50 1.50 1.50
55 1.00 5.50
Compute intrinsic values for each of these options and identify whether they
are in-the-money, at-the-money, or out-of-the-money.
5.15. Do you agree with the following statement: “An out-of-the-money option has
an intrinsic value of zero and vice versa”? Explain your answer.
QUESTIONS AND PROBLEMS 111
5.16. Compute the profit or loss on the maturity date for a short forward position
with a forward price of $303 and a spot price at maturity of $297.
5.17. Compute the profit or loss on the maturity date for a December 45 call for
which the buyer paid a premium of $3 and a spot price at maturity of $47.
5.18. Compute the profit or loss on the maturity date for a November 100 put for
which the seller received a premium of $7 and a spot price at maturity of $96.
5.19. Because options are zero-sum games, the writer’s payoffs are just the negatives
of the sellers. Demonstrate the following (by algebraic arguments or by using
numbers):
a. the call seller’s payoff at expiration is –max(0, S – K) = min(0, K – S)
b. the put seller’s payoff is –max(0, K – S) = min(0, S – K)
5.20. What is the difference between a market buy order and a limit order? When
would you use a market versus a limit order?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
6
Arbitrage and Trading
6.1 Introduction Spread Trading
6.2 The Concept of Arbitrage Index Arbitrage
EXTENSION 6.1 Stock Indexes
6.3 The No-Arbitrage Principle
EXTENSION 6.2 Index Arbitrage
for Derivative Pricing
The Law of One Price 6.6 Illegal Arbitrage
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Nothing Comes from Nothing Opportunities
6.4 Efficient Markets 6.7 Summary
6.5 In Pursuit of Arbitrage 6.8 Cases
Opportunities 6.9 Questions and Problems
The Closed-End Fund Puzzle
THE CONCEPT OF ARBITRAGE 113
6.1 Introduction
Once upon a time in days not too long ago, a venerable finance professor went with
two of his graduate students to a McDonald’s restaurant for a working lunch. One
student noticed that the Big Mac sandwiches were on a buy-one-get-one-free special.
Bubbling with excitement, the student ran to the mentor to share the news. His jaws
dropped as he approached, for the master himself was sitting with two Big Macs,
discussing the selling of one to the second student at half price: “You also know
about this?” “Yes,” the professor replied, “it’s an arbitrage.” Buying two for the price
of one and selling the second free sandwich at half price would indeed be an arbitrage
opportunity!
Taking a cue from this embellished story, we define arbitrage (arb) as a chance
to make riskless profits with no investment. The concept of arbitrage is invaluable for
building pricing models, which are mathematical formulas that price derivatives in
terms of related securities. Since the 1970s, financial markets have been generating
innovative derivatives by attaching extra features to simple financial securities, by
designing entirely new derivatives with complicated payoffs, or by combining several
derivatives into a single composite security. For this last example, the composite
security’s price must equal the sum of the prices of these simpler constituent
derivatives. Otherwise, eagle-eyed traders can make arbitrage profits by buying the
relatively cheaper of the two and selling the more expensive one. This observation
leads one to recognize that a more general truth holds. In the absence of arbitrage, all
portfolios or securities with identical future payoffs must have identical values today. This is
the essence of the no-arbitrage principle that we discuss later in this chapter.
Easy arbitrage opportunities are hard to find. Still we start with them because
they form the foundations of our study and help us understand the more complex
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
arbitrages that appear in today’s marketplace. The existence of arbitrage opportunities
implies that the markets are somehow inefficient. We introduce the various notions
of market efficiency and discuss what happens when they are violated. Next we talk
about some pricing anomalies and the trading strategies that profit from them.
6.2 The Concept of Arbitrage
Financial markets exert a ruthless discipline on their players. No one advises a
careless trader to modify bad quotes; instead, people grab such mispricings, and
the misquoting trader sees his wealth disappear. Earlier we defined arbitrage as a
chance to make riskless profits with no investments. This trivial description masks an
activity that makes and takes millions of dollars every day. Eager traders pour over
computer screens displaying prices from diverse markets and scan them for arbitrage
opportunities. The act of arbitrage supports the livelihood of many. But gone are
the days when one could sit in an armchair reading the financial press and find easy
arbitrages. Today’s arbitrageurs work very hard for their money.
114 CHAPTER 6: ARBITRAGE AND TRADING
Traders searching for arbitrage is rational and predictable behavior. Picking up a
hundred dollar bill from the ground is normal behavior.1 It is hard, if not impossible,
to model behavior if one oftentimes picks up the bill, sometimes doesn’t, and yet at
other times picks up the bill but leaves a few bucks in its place for the person coming
next! Therefore, for the subsequent theory, we assume that one never leaves money
behind and relentlessly picks up even the tiniest of arbitrage opportunities. As we
later see, this assumption has very powerful implications; indeed, it lies at the heart
of the derivative pricing models presented in this book.
There are many types of arbitrage. Some arbitrage operations are easy to under-
stand and implement but hard to find in practice, whereas others are just the opposite.
All such arbitrages can be decomposed into one of two types involving mispricings
across space and mispricings across time. An example of mispricing across space is
when the sum of the parts’ prices differ from the price of the whole. Example 6.1
discusses these two types of arbitrage opportunities.
EXAMPLE 6.1: The Two Types of Arbitrage Opportunities
■ Arbitrage across space (see Figure 6.1). Suppose Your Beloved Machines Inc.’s (YBM) stock has an ask
price of $100.00 on the Big Board and a bid price of $101.00 in Tokyo. If the brokerage cost is $0.10
per share, then one can buy stocks at a lower price in New York, simultaneously sell them at a higher
price in Tokyo, and make 80 cents in arbitrage profits.
■ Example (arbitrage across space). The sum of the parts’ prices differs from the price of the whole. Some people see
arbitrage opportunities in conglomerates, which are huge companies with diverse lines of business.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
They claim that the share prices of conglomerates often fail to reflect the true value of the underlying
assets and trade too low. Corporate raiders buy conglomerates, sell the constituent businesses or their
assets, and more than recover their initial investments. “I saw the company and thought that the sum
of the parts was worth more than what it was trading for,” observed a notorious corporate raider Asher
Edelman as he targeted a French conglomerate.2 Investment bankers doing mergers and acquisitions
believe the opposite. They believe that mergers create arbitrage opportunities where the sum of the
parts is less than the whole. The idea that the same payoffs no matter how they are created, trade at
the same price is known as the law of one price.
■ Arbitrage across time (see Figure 6.1). Free lottery tickets are an example of arbitrage across time.
Result 3.1 of Chapter 3, which states that a stock’s price falls exactly by the amount of the dividend
on the ex-dividend date, if violated, also represents an example of an arbitrage across time.
1
In an often-told tale, two finance professors are walking together, and they see a $100 bill lying on the
ground. When one ponders whether he should pick it up, the other replies not to bother because the
bill is not really there: “As markets are efficient, someone must have already picked it up before.” This
story is usually told to clarify the “ideal” notion of an efficient market, while teasing finance academics.
THE CONCEPT OF ARBITRAGE 115
FIGURE 6.1: The Two Types of Arbitrage
Opportunities
Security trades
at some price
Buy cheap, sell expensive; make
net profit after paying
brokerage costs
Same security trades
elsewhere at a lower price
Now Later
Nonnegative payoff always
i)
Zero net investment and strictly positive payoff
with positive probability
ii)
Positive cash flows Nonnegative payoff always
2 Corporate raider Asher B. Edelman once taught a business course at Columbia University, in which he offered $100,000 to any
student who would find a takeover target for him. However, the university banned the offer. In the late 1990s, he targeted Societe
du Louvre, the French holding company to which the above comment applies (see “Champagne Taste: An American Raider in
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Paris Challenges an Old World Dynasty—Edelman Presses Taittingers to Choose Family Values or Shareholder Value—Free to
Paint a Hotel Pink,” Wall Street Journal, November 11, 1998).
Are such arbitrage opportunities likely to exist in today’s financial market, where
technological advances have lowered brokerage costs and made it possible to move
millions of dollars in a matter of seconds? Most economists believe that the answer
to this question is no—and if they exist, it is especially hard to find them in well-
functioning and efficient markets. Perhaps they can be found in some emerging mar-
kets, but they fade away quickly as traders start grabbing them. Sometimes what may
seem like arbitrage opportunities in a perfect world (where market imperfections
or frictions like transaction costs, bid/ask prices, taxes, and restrictions on asset
divisibility are assumed away) cease to be so when market imperfections are included.
In actual markets, market frictions are relevant and should always be taken into
account when considering arbitrage opportunities.
Let’s discuss a few real-life examples of arbitrages.
116 CHAPTER 6: ARBITRAGE AND TRADING
EXAMPLE 6.2: Real-Life Arbitrages
■ Sports arbitrage. Not too long ago the soccer World Cup finals had two powerhouse teams, Brazil and
Germany, playing. Uncle Sanjoy relishes the story of how he visited a Wall Street firm and entered
into two bets with acquaintances. First, he agreed to pay one person $10,000 if Brazil won but receive
$5,000 otherwise. Then, uncle slyly adds that he visited another office, and bet with a second person
to receive $11,000 if Brazil won but pay $4,000 otherwise. No matter which team won, uncle locked
in $1,000. Although gambling exasperates his wife, uncle quietly commented that he had locked in
arbitrage profits by exploiting inefficiencies in the system and that he truly enjoyed watching the game!
■ Airline miles arbitrage. Airlines put all kinds of restrictions on tickets, serve scant morsels of food,
manhandle baggage, make passengers endure long delays, and cancel flights. Those who have suffered
in the sky at the hands of airlines will be amused by a story in the Wall Street Journal, “How Savvy
Fliers Make the Most of Their Miles” (December 16, 2008), which related how some travelers got
the best out of their frequent flier miles. Normally a ticket obtained by redeeming frequent flier miles
would yield the value of approximately one cent per mile. Some folks figured out how to transform
this once cent per mile into more. One gentleman paid about 1.3 cents per mile to “friends” through
the Internet and then used the miles for business class tickets between China and the United States.
“I call it airline miles arbitrage,” said the savvy traveler, who created business class tickets for coach
ticket prices and in the process, earned about six to nine cents per mile.
■ Uncle Sam’s generosity. The idea of arbitraging Uncle Sam (the US government) appeals to most people,
but few succeed in doing so. However, an article in the Wall Street Journal, “Miles for Nothing:
How the Government Helped Frequent Fliers Make a Mint—Free Shipping of Coins, Put on Credit
Cards, Funds Trip to Tahiti; ‘Mr. Pickles’ Cleans Up” (December 7, 2009), reported a story where this
happened. What happened? Because a paper dollar lasts only about twenty-one months in circulation,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
dollar coins save a country money as they can last for thirty years or more. To promote their use, the
US Mint offered to sell presidential and Native American $1 coins at face value, ship them for free,
and buyers could charge the purchase to their credit cards without handling fees. However, well-
intentioned plans sometimes have unintended consequences. Arbitrageurs bought large quantities of
coins from the mint, paid for them with credit cards that offered frequent flier miles for free air travel
(typically, one cent for each dollar charged in purchases), and immediately deposited the coins in the
bank to pay for the credit card coin purchase. The story states that one software consultant ordered
$15,000 worth of coins and even had the delivery person deposit the coins directly in his car’s trunk
so that he could take them to the bank. Discovering such activities, US Mint officials sent letters to
prospective buyers seeking the reasons for their purchase and denied further access to those who did
not respond.
THE NO-ARBITRAGE PRINCIPLE FOR DERIVATIVE PRICING 117
6.3 The No-Arbitrage Principle for
Derivative Pricing
Modern finance theories use the idea that arbitrage opportunities do not exist in
well-functioning markets to derive many important results. This was pioneered by
the Miller and Modigliani (M&M) propositions in corporate finance, which state
that if market imperfections like taxes are assumed away, then the choice of debt or
dividend policy does not affect the value of a firm. (Taxes are imperfections! Who
says finance jargon is boring?) In our case, the absence of arbitrage opportunities lies
at the heart of derivative pricing models, including the celebrated Black–Scholes–
Merton and the Heath–Jarrow–Morton models included in this text. This section
explains how.
The Law of One Price
Suppose one wants to price a market-traded derivative, which we label as “portfolio
A.” Assume that there are no cash flows on this traded derivative before maturity so
that there are only two dates of concern—today and when the derivative matures.
Now let us create portfolio B, a collection of different traded securities with known
prices and no intermediate cash flows, to match portfolio A’s payoffs on the maturity
date (see Figure 6.2). The values of these two portfolios are linked by a no arbitrage
principle: to prevent arbitrage, two portfolios with identical future payoffs must have the
same value today.3 Using this principle, the cost of creating portfolio B is the “fair” or
“arbitrage-free” value of the traded derivative portfolio A.4 This is known as the law
of one price.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Nothing Comes from Nothing
Continuing with the previous discussion, if we subtract portfolio A from portfolio B,
then we get the second representation, which will be one of our main methods for
proving results (see Figure 6.2). By construction, the value of this new portfolio is
zero for sure on the maturity date. To prevent arbitrage profits, it must also have zero
value today. The derivative’s price can be solved from this condition. This insight can
be paraphrased as nothing comes from nothing.5 When identical cash flow securities are
subtracted, yielding zero at a future date, the result is emptiness, nothing, nada, zero
now!
These are the two standard techniques used for pricing derivatives from Chapter
11 onward in this book.
3
Suppose this principle does not hold. If portfolio A is cheaper than portfolio B today, “buy low, sell
high”: buy A and sell B. You will have a zero cash flow in the future because the cash inflow will cancel
the outflow, and the price difference you capture today becomes arbitrage profit. If B is cheaper than A,
just reverse the trades to capture the arbitrage profit.
4
The same argument works when the price of the derivative is known but there is an unknown variable
in the payoff. For example, a forward contract has a zero value on the starting date, but the unknown
forward price is in the payoff. The solution to this equation is the arbitrage-free forward price.
5
“Nothing comes from nothing” (or ex nihilo nihil fit in Latin) is a philosophical expression of a thesis
first advanced by the Greek philosopher Parmenides of Elea, who lived in the fifth century bc.
118 CHAPTER 6: ARBITRAGE AND TRADING
FIGURE 6.2: Pricing Derivatives
The law of one price
Today Later date
Portfolio A
Must be equal; Portfolios have the
No intermediate cash flows
otherwise arbitrage same final value
Portfolio B
Nothing comes from nothing
Now Later date
Portfolio A No intermediate cash flows
Zero value (for sure)
– (Portfolio B)
Must be zero;
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
otherwise arbitrage
As we have often observed, arbitrage opportunities are likely to be absent in well-
functioning and efficient markets—but what do we really mean by an “efficient
market”? Well, read on.
6.4 Efficient Markets
When a finance academic states that “markets are efficient,” she means an informa-
tionally efficient market: a market where asset prices quickly absorb and “fully”
reflect all relevant information. But what relevant information is this referring to?
The efficient markets hypothesis (EMH) is presented with respect to three
different information sets:
1. Weak-form efficiency asserts that stock prices reflect all relevant information
that can be gathered by examining current and past prices. If this is true, then
technical analysis (which involves looking at price patterns to predict whether
a stock’s price will go up or down) will not generate returns in excess of the
IN PURSUIT OF ARBITRAGE OPPORTUNITIES 119
risk involved. In addition, and more relevant to us, if the market is weak-form
efficient, then there are no arbitrage opportunities. Why? Because finding arbitrage
opportunities only depends on the information contained in current and past
prices. It is this connection that relates an efficient market to the arbitrage-free
pricing of derivatives.
2. Semistrong-form efficiency asserts that stock prices reflect not only historical
price information but also all publicly available information relevant to those
particular stocks. If this holds, then fundamental analysts (fundamental analysis
involves reading accounting and financial information about a company to
determine whether a share price is overvalued or undervalued) would be out of
business.
3. Strong-form efficiency asserts that stock prices reflect all relevant information,
both private and public, that may be known to any market participant. If this is
true, then even insiders (who have privileged, yet to be made public information
about the company) will not make trading profits in excess of the risk involved.
There is a never-ending debate on whether markets are efficient with respect to
these three different information sets. The preponderance of evidence tends to accept
weak-form efficiency, and strongly reject strong-form efficiency, especially given the
numerous insider-trading lawsuit convictions historically obtained by the Securities
and Exchange Commission (SEC). The evidence with respect to semistrong-form
efficiency is mixed.
The assumption of no arbitrage underlying the models developed in this book is
very robust. Indeed, if markets are weak-form efficient, then no arbitrage opportunities exist.
Hence the weakest form of market efficiency justifies the assumption of no arbitrage.
But don’t get confused. Assuming there are no arbitrage opportunities in subsequent
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
modeling does not imply that we are assuming that the market is weak-form efficient.
As just discussed, the assumption of no arbitrage is a much weaker notion than even a
weak-form efficient market. Hence it is more likely to be satisfied in actual markets,
and thus it provides a more robust model than market efficiency does.
6.5 In Pursuit of Arbitrage Opportunities
Well-resourced, astute players trade for the purpose of exploiting arbitrage oppor-
tunities. As the next section illustrates, academics have uncovered phenomena that
seem to suggest the existence of such arbitrages.
The Closed-End Fund Puzzle
Closed-end fund share prices are often claimed to represent arbitrage opportunities.
To understand this claim, we must first understand closed-end funds, and to
understand closed-end funds, we need to first understand investment companies.
An investment company sells shares to investors, invests the funds received in various
securities according to a stated objective, and passes the profits and losses back to
the investors after deducting expenses and fees. Federal securities law classifies invest-
ment companies into three basic types: mutual funds (legally known as open-end
120 CHAPTER 6: ARBITRAGE AND TRADING
companies), closed-end funds (legally known as closed-end companies), and
UITs (legally known as unit investment trusts). Mutual funds, like those offered
by Fidelity and Vanguard, are open ended and take new investments at any time.
A shareholder can redeem his shares directly from the fund whenever he desires.
Closed-end funds issue shares at their creation but none thereafter. Just like regular
stocks, the shares of closed-end funds trade on organized exchanges. Closed-end
funds are not redeemed by the fund for cash. To close out an investment, an investor
must sell her shares on the secondary market. Finally, unit investment trust shares can
be redeemed like mutual funds; moreover, they have a termination date when the
fund’s assets are sold and the proceeds are paid out to investors.
Redemptions in mutual funds and UITs pay the company’s net asset value per
share. Net asset value (NAV) is the value of the company’s total assets minus its
total liabilities, often expressed on a per share basis. For example, if an investment
company has securities and other assets worth $110 million and has liabilities of $5
million, its NAV would be $105 million. If the investment company has issued 1
million shares, then $105 would be its NAV per share. NAV changes daily due to
market fluctuations. Mutual funds and UITs are usually required to compute their
NAV at least once per business day (which they usually do after the market closes).
Closed-end funds, however, are exempted from this requirement.
For reasons that are doubted and debated, the shares in most closed-end funds trade
at a discount from their asset value. This is known as the closed-end fund puzzle.
These discounts represent potential arbitrage opportunities. To see this, suppose that
a closed end fund is worth $100 million in the market, but the underlying securities
have a net asset value of $105 million. What can one do? One can buy all the
shares, open up the fund, and sell the underlying securities to make $5 million in
arbitrage profits. However, not all closed-end funds trade at a discount. Some trade
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
at premiums over their underlying asset values, like the legendary Warren Buffett’s
Berkshire Hathaway Company at the end of the twentieth century. Why? Again, a
puzzle.
Spread Trading
At the heart of many arbitrage strategies is spread trading. Spread trading involves
buying and selling one or more similar securities when their price differences are large,
with the aim of generating arbitrage profits as their price differences later converge
to some target value.
For example, suppose that Your Beloved Machines Inc. (YBM) has two classes of
shares with different voting rights. You find that class A shares are trading at $100 and
class B shares are trading at $85. This is a spread of 100 – 85 = $15. However, you’ve
discovered that the spread has historically been $10, and you expect it to return to
that level.
To take advantage of this price discrepancy, you can buy the underpriced class
B shares and sell the overpriced class A shares. If your bet is correct and the spread
returns to $10, you can reverse the trades and capture $5 in profits. Notice that you
are betting on the movement of a spread and not on the direction of the stock prices
IN PURSUIT OF ARBITRAGE OPPORTUNITIES 121
themselves. Spread trading is one of the most common arbitrage trading strategies
used in financial markets. The next section discusses one such widely used strategy,
called index arbitrage.
Index Arbitrage
A stock index is an average of stock prices that are selected by some predetermined
criteria. For example, the Dow Jones Industrial Average (also called the DJIA,
Dow 30, the Dow Industrials, the Dow Jones, or simply the Dow) and the
Standard and Poor’s 500 Index (S&P 500) are stock indexes created by using a
price-weighted average of thirty major US stocks and the market value–weighted average
of five hundred major US companies, respectively (see Extension 6.1 for a discussion
of indexes).
Sometimes the value of a stock market index gets disjointed from the arbitrage-free
price of a futures contract written on the index. In such situations, the potential for
index arbitrage exists. Suppose the traded futures contract on an index is overvalued.
The arbitrage is to short the traded futures contract and go long a synthetic futures
contract constructed to have the same cash flows as the traded futures contract. The
synthetic futures contract is constructed using the portfolio of the traded stocks
underlying the index. This strategy will lock in the price difference at time 0 and
have zero cash flows thereafter. The arbitrage is a violation of the nothing comes from
nothing principle. Extension 6.2 illustrates index arbitrage via a simple example.
EXTENSION 6.1: Stock Indexes
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
There are two basic types of stock indexes: price-weighted and value-weighted indexes. We illustrate both using
examples.
EXT. 6.1 EX. 1: A Price-Weighted Index (PWI)
■ Holding one share of each stock, summing the share prices and dividing by the total number of shares creates
a price-weighted index. Consider the hypothetical stock price data in Ext. 6.1, Table 1. Here YBM’s share
price is $100 today and $110 tomorrow. BUG’s price is $50 today and $45 tomorrow. The equal average PWI
is equal to
Sum of stock prices / Number of stocks
= (100 + 50) /2 = 75 today and (110 + 45) /2 = 77.5 tomorrow
■ While one stock went up and the second went down, the PWI index registered a 2.5 point gain or
(Tomorrow’s index value – Today’s index value) / Today’s index value
= (77.5 – 75) / 75 = 0.0333 or a 3.33 percent increase
122 CHAPTER 6: ARBITRAGE AND TRADING
■ This is called a price-weighted average because when viewed as a portfolio, each stock’s return is weighted
by a percentage proportionate to the price of the stock relative to the value of the portfolio.
The DJIA is computed by creating a price-weighted average of thirty high-quality blue chip stocks from
diverse industries. The component stocks are selected by the editors of the Wall Street Journal. The membership
list is periodically modified so that the index continues to accurately reflect the general US stock market. Such
revisions are imperative because companies go through mergers, acquisitions, and bankruptcy.
EXT. 6.1 TABLE 6.1: Stock Price Data
Stock Today’s Tomorrow’s Shares Today’s Market Tomorrow’s
Price Price Outstanding Value (millions) Market Value
(millions) (millions)
YBM $100 $110 50 $5,000 $5,500
BUG $50 $45 20 $1,000 $900
Total $6,000 $6,400
EXT. 6.1 EX. 2: A Value-Weighted Index (VWI)
■ Multiplying the stock price by the shares outstanding (which gives the stock’s total market capitalization)
and then summing across all stocks creates a value-weighted index. Using the data from Ext. 6.1 Tab. 1, we
get a VWI equal to
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Sum of stock price times shares
= (100 × 50 + 50 × 20) = 6, 000 today and (110 × 50 + 45 × 20) = 6, 400 tomorrow
The weights (percentage holdings of each stock) in the index are
for YBM = $5, 000/$6, 000 = 0.8333
for BUG = $1, 000/$6, 000 = 0.1667
■ The percentage change in the VWI is
(Tomorrow’s index value – Today’s index value) /Today’s index value
= (6, 400 – 6, 000) /6, 000 = 0.0667 or a 6.67 percent increase
■ In practice, the index is often normalized to make its value equal to some arbitrary number at a given date.
This is for convenience in reporting the index. For example, the VWI index’s value could be normalized to
be 100 today by dividing by 60. The number 60 must also divide the index’s value at all future dates. In this
case, the index value tomorrow would be 6,400/60 = 106.67.
IN PURSUIT OF ARBITRAGE OPPORTUNITIES 123
■ This is called a value-weighted index because when viewed as a portfolio, each stock’s return is weighted by
its market value relative to the total value of the stock market.
The S&P 500 is a market-value-weighted index of five hundred large-sized US companies. A value-weighted
index is sometimes called a market-cap weighted index.
EXTENSION 6.2: Index Arbitrage
This extension explains index arbitrage. To simplify the presentation, the discussion replaces futures with forward
contracts. For the purposes of this demonstration, this difference is not important. The differences between
forward and futures contracts, which were discussed in Chapter 4, will be further clarified in subsequent chapters.
Index arbitrage generates riskless profits by putting the no-arbitrage principle to work. We illustrate it with a
simple example.
EXT. 6.2 EX. 1: The No-Arbitrage Principle
■ Let a PWI be created by computing an average of the stock prices of YBM and BUG worth $100 and $50
today, respectively. This gives an index value of $75 today.
- Suppose that a dealer quotes a forward price of $80 on a forward contract on PWI that matures in one year.
- Let the simple interest rate be 6 percent per year. And suppose that neither stock will pay any dividends
over the next year.
■ Next, suppose that a trader believes the forward price is too high. Intuitively, she wants to sell the traded forward
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
contract and create a synthetic forward contract by buying the PWI index portfolio, borrowing the funds to
do so, and holding the portfolio until the traded forward contract matures. She performs the following steps:
- She sells two forward contracts to the dealer, whereby she locks in a selling price of $80 per unit of the
index: $160 in all.
- Simultaneously, she buys one share of each stock for a total cost of $150 and finances this purchase by
borrowing. Notice that she has a zero cash flow today because the traded forward requires no cash flows
on its initiation date.
- One year later, she gives two units of the index to the dealer who sold her the forward. This costs nothing
because she made allowance for this by purchasing the stocks earlier. She gets $160 from him in return.
- But she has to repay the loan with interest. This entails an outflow that equals
Today’s index purchase price × (1 + interest cost)
= 150 × 1.06
= $159
- She pays out $159 but receives $160, and makes $1 in arbitrage profits. She makes this profit no matter
where the index moves a year from today.
We make extensive use of this technique to develop cost-of-carry models in Chapters 11 and 12.
124 CHAPTER 6: ARBITRAGE AND TRADING
6.6 Illegal Arbitrage Opportunities
Arbitrage is a broad concept. You can bring under its umbrella dubious, socially
unfavorable, and illegal trading activities. Medical students first learn how a healthy
body functions before they study abnormalities and diseases. We follow a similar
approach. Previously we introduced legal arbitrage opportunities that facilitate market
efficiency. Now we shift gears and study nefarious arbitrage opportunities. These
illegal arbitrage opportunities occur in malfunctioning securities markets and remain
a major problem that regularly appears in the national, local, and financial press.
Trading abuses and manipulations are forms of illegal arbitrage opportunities that we
now discuss.
One such abuse is front running, which is trading based on an impending
transaction by another person. For example, a broker may buy on his own account in
front of his customer’s buy order. Then, he makes unscrupulous gains by selling the
shares he bought to the customer at a higher price when executing the customer’s
buy order.
A chronic problem facing free markets is that individuals often try to manipulate
prices to their advantage. Manipulation has long been a concern in the securities
market. For example, in its report on the [Securities] Exchange Act of 1934, the
Senate Committee wrote, “The purpose of the act is... to purge the securities
exchange of these practices which have prevented them from fulfilling their primary
function of furnishing open markets for securities where supply and demand may
fully meet at prices uninfluenced by manipulation or control.”6 The Commodity
Exchange Act states that it is illegal to manipulate or attempt to manipulate a
commodity’s future delivery price, yet proving manipulation in a court of law is
not easy. Next we discuss some well-known examples of market manipulation.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
A bear raid is a market manipulation strategy generating arbitrage profits that was
widely prevalent in US stock markets in the nineteenth and early twentieth centuries.
Remember from Chapter 2 how short selling works—a short seller borrows shares,
accepts the obligation to pay dividends to the lender, and agrees to return the security
to the lender at a future date. Traders organizing a bear raid take large short positions
in a company’s stock, spread unfavorable rumors that depress the stock further, and
buy back those shares when other shareholders panic and unload their stock holdings
at depressed prices.
Short sellers can get severely burned if they get caught in a market corner and a
short squeeze, which was also a pervasive problem in the early days of stock markets.
The next example shows how a market corner and short squeeze works.
6
Report of the Committee on Banking and Currency, “Stock Exchange Practices,” U.S. Senate Report
No. 1455, 73d Cong., 2d Sess., 1934 (the “Fletcher Report”). The quote appears in Chapter I “Securities
Exchange Practices,” Section 11 “The government of the exchanges,” Subsection (b) “Necessity for
regulation under the Securities Exchange Act of 1934.”
ILLEGAL ARBITRAGE OPPORTUNITIES 125
EXAMPLE 6.3: A Market Corner and Short Squeeze
■ Suppose a company has ten thousand outstanding shares, which sell for $50 in the market. This is the
deliverable supply, of which Alexander owns three thousand shares, Bruce owns three thousand
shares, and Caesar owns four thousand shares. The shares are held in street name with their common
broker, Mr. Brokerman, so that short sellers can borrow (see Figure 6.3).
FIGURE 6.3: Example of a Market Corner and Short
Squeeze
Shorty borrows shares from Alexander
Alexander Bruce Caesar Shorty
3,000 shares 3,000 shares 4,000 shares Short 3,000
shares
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Sells shares to Ruth
Ruth owns 13,000 shares
Short sells to Ruth
Ruth controls 13,000 shares, though the deliverable supply is 10,000 shares.
Ruth has cornered the market.
She will squeeze Shorty by charging him a high price for 3,000 shares that he
has to buy and return to Alexander.
126 CHAPTER 6: ARBITRAGE AND TRADING
■ Shorty is bearish on the company and expects the share price to decline from $50 to $40. He short
sells three thousand shares, which his broker, Mr. Brokerman, borrows from Alexander. They are
sold at $50 per share to Ruth, who happens to be a “ruthless manipulator.”
■ Ruth now buys all the shares from Alexander, Bruce, and Caesar. She now has a monopoly position
and has bought three thousand more shares than the deliverable supply of ten thousand shares, making
it thirteen thousand in all. Ruth has cornered the market.
■ At the time of settlement, Alexander does not have shares to give to Ruth. He asks Mr. Brokerman to
return his three thousand shares. Since Ruth owns them all, nobody is willing to lend shares. Normally,
Shorty can keep his short position open as long as he likes. As scarcity is preventing Alexander’s trade
from getting settled, Brokerman informs Shorty that he has to return the borrowed shares at once.
■ Shorty has nowhere to go but Ruth to buy shares. This is a case of congestion: Shorty, trying to
cover his short position, finds that there is an inadequate supply of shares to borrow, and Ruth is
unwilling to sell her shares, except at sharply higher prices. In other words, Ruth is squeezing the
short. She can easily bankrupt Shorty.
This example easily extends to futures market. First the trader goes long the
futures contract in excess of the immediately deliverable supply. Then she keeps
her long position open, eventually acquires all the deliverable supply, and ends up
in a monopoly position. After she has cornered the market, a short squeeze develops.
She demands delivery, but the short cannot find any supply to cover their positions,
except from her—and she extracts her price.
These undesirable situations are in violation of the Commodity Exchange Act
statute. To stop short squeezes, the CFTC can suspend trading and force settlement
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
at a “fair” price set by the exchange issuing the contract. This happened, for example,
with the Hunt brothers’ holdings of silver futures in 1979 and 1980. Contract
provisions are also designed to increase the deliverable supply, which minimizes the
likelihood of manipulation, by allowing variation in the quality of the asset delivered
(with appropriate price adjustments) and a longer delivery period. Still, manipulations
happen. We discuss additional futures market manipulations in Chapter 10.
6.7 Summary
1. Arbitrage is a chance to make riskless profits with no investment. Arbitrage profits
can be generated in several ways—across time and across space. An example of an
arbitrage across space is when the sum of the parts’ prices differ from the price of
the whole.
2. The no-arbitrage pricing principle involves skillfully creating a portfolio of assets
that replicates the payoff to a traded derivative and then arguing that the cost of
this portfolio is the “fair” or “arbitrage-free” value of the traded derivative. This
is the standard technique used for pricing derivatives.
QUESTIONS AND PROBLEMS 127
3. Arbitrage profits do not exist in an efficient market where market prices reflect all
past and current price information. However, markets aren’t always efficient, and
sophisticated traders use advanced tools and techniques to make arbitrage profits.
4. A chronic difficulty with free markets is that traders try to manipulate prices
to their advantage. Market manipulation is an example of an illegal arbitrage
opportunity.
5. A market corner and short squeeze is a common manipulation technique that
has a checkered past in futures markets. It involves a trader going long in the
futures market and squeezing the shorts subsequently in the cash market, when
they scramble to cover their short positions.
6.8 Cases
Arbitrage in the Government Bond Market? (Harvard Business School Case 29-
3093-PDF-ENG). The case examines a pricing anomaly in the large and liquid
Treasury bond market, where the prices of callable Treasury bonds seem to be
inconsistent with the prices of noncallable Treasuries and an arbitrage opportunity
appears to exist.
RJR Nabisco Holdings Capital Corp.—1991 (Harvard Business School Case 29-
2129-PDF-ENG). The case explores a large discrepancy in the prices of two nearly
identical bonds issued in conjunction with a major leveraged buyout and considers
how to capture arbitrage profits from the temporary anomaly.
Nikkei 225 Reconstitution (HBS Premier Case Collection Case 207109-PDFENG).
The case considers how an institutional trader who may receive several billion
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
dollars of customer orders in connection with a redefinition of the Nikkei 225
index can provide liquidity for the event and pick up arbitrage profits.
6.9 Questions and Problems
6.1. What is an arbitrage opportunity across space? Give an example.
6.2. What is an arbitrage opportunity across time? Give an example.
6.3. What is the law of one price?
6.4. Explain the pricing principle “nothing comes from nothing.”
6.5. Suppose a two-year Treasury note is trading at its par value $1,000. You
examine the cash flows, and if you sell them individually in the market, you get
$47.85 for the six-month coupon, $45.79 for the one-year coupon, $43.81 for
the one-and-a-half-year coupon, $41.93 for the two-year coupon, and $838.56
for the principal.
a. Are these prices correct?
b. If not, show how you can capture arbitrage profit in this case.
128 CHAPTER 6: ARBITRAGE AND TRADING
6.6. Suppose a two-year zero-coupon bond has a price of $0.90 and a three-year
zero has a price of $0.85. A bank allows you to borrow or lend at 4 percent,
compounded once a year. Show two ways that you can make arbitrage profits
from these prices.
6.7. You can trade Boring Unreliable Gadget Inc.’s stock for $77 per share in the
United States and for €50 in Europe. Assume a brokerage commission of $0.10
per share in the United States and €0.10 in Europe. A foreign exchange dealer
quotes a bid price of $1.5000 for each euro and offers them for $1.5010.
a. Are these prices correct?
b. If not, show how you can capture arbitrage profit by trading BUG stock.
6.8. What is an efficient market, and what does it mean for a market to be weak-
form, semistrong-form, and strong-form efficient?
6.9. If the market is weak-form efficient, do arbitrage opportunities exist? Explain
your answer. Do you think arbitrage opportunities exist? Explain your answer.
6.10. What is a closed-end fund, and what is the “closed-end fund puzzle”?
6.11. Suppose that Boring Unreliable Gadget Inc. has two classes of shares with
different voting rights. You find that class A and class B shares are trading at
$49 and $37, respectively. However, historically, the spread has been $15, and
you expect the price difference to reach that level.
a. Explain how you would set up a spread trade and how much profit you
expect to make once the prices correct themselves.
b. Would the preceding strategy work if class A stock goes up to $75 per share?
6.12. What is a stock index? Describe the differences between the Dow Jones
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Industrial Average and the S&P 500 stock price indexes.
6.13. What is algorithmic trading, and how can it be used to make profits? How does
this relate to program trading?
6.14. What is front running? Explain why an investor should be concerned if his
stockbroker front runs his trades.
6.15. What is a short squeeze, and under what circumstances does it occur?
6.16. What is a bear raid? How does a bear raid relate to trading on inside
information?
7
Financial Engineering and
Swaps
7.1 Introduction Interest Rate Swaps
Forex Swaps
7.2 The Build and Break Approach
Currency Swaps
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
7.3 Financial Engineering
EXTENSION 7.1 Valuing
Cash Flows versus Asset Values Fixed-for-Fixed Currency Swaps
Examples Commodity and Equity Swaps
7.4 An Introduction to Swaps Credit Default Swaps
7.5 Applications and Uses of 7.7 Summary
Swaps
7.8 Cases
7.6 Types of Swaps
7.9 Questions and Problems
130 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
7.1 Introduction
In January 1993, the French chemical and pharmaceutical giant Rhône-Poulenc
offered its employees an opportunity to buy discounted shares as part of its forthcom-
ing privatization plan. With an eye toward “wide employee participation,” the scheme
involved a dizzying array of features, including minimum investment incentives, price
discounts, free shares, interest-free loans, tax benefits, an option to opt in or out
of dividends, and an installment plan to pay for the shares over time. Despite the
best intentions of this intricate scheme, the plan badly flopped. Then came Bankers
Trust, an American bank famous for its derivatives and trading prowess (and infamous
for widely publicized losses incurred by its clients Procter & Gamble and Gibson
Greetings). Bankers Trust streamlined the terms of the Rhône-Poulenc plan and also
added a critical safety feature—the investment would earn a guaranteed minimum
return but surrender some gains if the stock price increased. The modified plan was
a success.
How could they guarantee a minimum return? Armed with your derivatives
knowledge, you can now understand what happened. The plan protected Rhône-
Poulenc’s employees from the downside (by buying a put option), and it financed this
plan by removing some of the upside (by selling a call option). This is an example of
financial engineering.
Financial engineering studies how firms design derivatives to solve practical
problems and exploit economic opportunities. This vast subject is taught at varying
depths in many universities around the world. After presenting some simple examples,
we introduce swaps, which were among the first uses of financial engineering. We
briefly discuss applications and uses of swaps and illustrate the basic workings of
interest rate swaps, forex swaps, currency swaps (including a valuation formula),
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
commodity swaps, equity swaps, and the highly useful but much maligned credit
default swaps. Chapter 22 discusses swap markets and interest rate swaps again in
greater detail.
7.2 The Build and Break Approach
Sometimes children snap together plastic blocks to build beautiful toys. At other
times, they rip apart their creations into simpler structures. We adopt a similar build
and break approach. Sometimes we combine derivatives and financial securities
to create new derivatives tailored to meet investment needs, and at other times, we
break down a complex derivative into its simpler components. This is important for
two reasons.
The first reason is that it provides an approach for pricing derivatives—break a
derivative into simpler parts, price the simpler parts, and sum to get the original derivative’s
price. Arbitrage, like adhesive, binds the parts’ prices together to equal the price of
the whole and makes sure that the technique works. For example, you can price a
convertible bond as the sum of a regular bond and an option to convert the bond
into a fixed number of stocks. An extendible bond may be valued as the price of
FINANCIAL ENGINEERING 131
a straight bond plus an option to extend the bond’s life. So when a rocket scientist
tries to sell you a complex derivative at an exorbitant price, you can value it by this
method and prevent yourself from being gouged.
The second reason is that it enables the use of derivatives to modify a portfolio’s
return to meet various investment objectives. For example, suppose a company
is considering issuing fixed rate debt when interest rates are high to finance an
investment project. The company is worried about getting stuck with an expensive
debt issue if interest rates go down. To avoid this problem, the company can issue
callable bonds. In the case that interest rates decline, the company can exercise the
call provision in the callable bonds and buy back the bonds at a predetermined price.
Where does the company get the funds to buy back the bonds? It issues new fixed
rate bonds at the lower interest rate.
7.3 Financial Engineering
Toward the end of the 20th century, financial engineering emerged as a vibrant,
new subfield of finance and operations research.1 Financial engineering applies
engineering methods to finance. At the simplest level, financial engineering and risk
management puts our build and break approach to work. The best way to understand
this approach is through some examples of financial engineering in practice.
Cash Flows versus Asset Values
To a beginning student, when forming and liquidating portfolios, the signs of the
cash flows and asset values may be confusing. To preempt such confusion, we make
the following observations. When constructing a portfolio, a positive value means
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
that you are purchasing an asset, hence there is a negative cash flow. For example, if
you are buying Your Beloved Machine stock, it shows up as a positive value in the
portfolio, but you spend $100, a negative cash flow, in acquiring it. If the value is
negative, you are creating a liability, and there is a positive cash flow. For example,
borrowing $90 gives a positive cash inflow of $90. The $90 is a liability because you
must pay it back in the future. Note that when a portfolio is constructed, the signs of
the cash flows and asset values are opposite.
At liquidation, this relationship changes. When liquidating a positive value
position, you are selling an asset; hence there is a positive cash flow. For example,
YBM has a positive value in your portfolio, and selling it for $120 will generate
a positive cash flow. If the value is negative before liquidation, then the cash flow
is negative because you are closing a liability. For example, your $90 loan, being a
liability, has a negative value in the portfolio. Repaying it on the liquidation date,
say, for $94, incurs a negative cash flow. Note that at liquidation, the signs of the cash
1
Cornell University offered the first financial engineering degrees. It states on the Cornell University
website (www.orie.cornell.edu/orie/fineng/index.cfm) that “Robert Jarrow and David Heath advised
students for several years before formalizing the program in 1995, making Cornell one of the very first
universities to have a graduate program in Financial Engineering, and arguably the oldest such program
in the world.”
132 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
flows and asset values are the same. We will often call both the cash flows and values at
liquidation the payoffs. As they have the same sign, no confusion should result.
When proving results, for some arguments we will consider cash flows and for
others asset values. As long as you keep these relationships in mind, there should be
no confusion. The key is to “follow the money!”
Examples
Example 7.1 considers a commodity derivative—a bond whose payoff is indexed to
gold prices.
EXAMPLE 7.1: A Bond Indexed to Gold Prices
■ Goldmines Inc. (fictitious name) is a gold mining company. It sells a highly standardized product,
pure gold, in the world market. It needs money for exploring, for setting up a mine, for running a
mine, and for refining operations. The company’s fortune ebbs and flows with the price of gold.
■ The company decides to raise cash by selling bonds. This is attractive because interest payments are
tax deductible. But debt makes a company vulnerable to financial distress—a debt-free company, by
definition, can never go bankrupt!
■ An investment banking firm designs a bond with the following payoff on the maturity date T:
- A payment of $1,000
- An additional amount tied to gold’s price per ounce, S(T), which is
0 if S (T) < $950
$10 [S (T) − 950] if $950 ≤ S (T)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ This payoff is given in Table 7.1 and graphed in Figure 7.1. You can generate the first payment
synthetically by investing in zero-coupon bonds. The second payoff looks similar to a long call option
position: buying ten calls with a strike price of $950 does the trick. Consequently, the cash flows from
the traded bond may be obtained synthetically by forming a portfolio consisting of (1) a zero-coupon
bond worth 1,000B, where B is today’s price of a zero-coupon bond that pays $1 at maturity, and (2)
buying ten European calls, each with a price of c and a strike price of $950 (see Table 7.2).
■ The no-arbitrage principle assures that the replicating portfolio must have the same value as the traded
bond. Therefore the bond’s value is (1,000B + 10c). In later chapters, you will learn how to price
this call option with the binomial or the Black–Scholes–Merton model.
■ Why sell a security like this? First, it will raise more cash up front. Second, it hedges some of the
output price risk that Goldmines Inc. faces. If gold prices are high at maturity, it shares some of its
profits—that’s not too bad. And if gold prices decline, only the principal is repaid.
FINANCIAL ENGINEERING 133
TABLE 7.1: Payoffs for a Traded Bond
Indexed to Gold
Bond Payoff Time T (Maturity Date) Cash Flow
S(T) < 950 950 ≤ S(T)
1,000 1,000
0 10[S(T) – 950]
Net cash flow 1,000 1,000 + 10[S(T) – 950]
FIGURE 7.1: Payoff Diagram for a Bond
Indexed to Gold
Payoff
Slope = 10
1,000
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
0
950 S(T), spot price of gold at maturity
TABLE 7.2: Cash Flows for the Synthetic Bond
Indexed to Gold
Portfolio Time T Cash Flow
S(T) < 950 950 ≤ S(T)
Long zeros (face value 1,000) 1,000 1,000
Long 10 calls (strike price 950) 0 10[S(T) – 950]
Net cash flow 1,000 1,000 + 10[S(T) – 950]
134 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
Example 7.1 is a commodity-indexed note, which has its return tied to the
performance of a commodity (like platinum, gold, silver, or oil) or a commodity
index (like Bridge/CRB Futures Index, a commodity futures index that goes back
to 1957). Sometimes, the holder can get unlimited benefits from price rises (as in this
example), but at other times, the payoffs are capped.
Close cousins of these derivatives are equity-linked notes (ELN). An ELN is a
combination of a zero-coupon (or a small coupon) bond whose return is based on the
performance of a single stock, a basket of stocks, or a stock index (see Gastineau and
Kritzman [1996] for definitions and discussions). These come in numerous variations
and have fancy trade names like ASPRINs, EPICs, GRIP, SIR, and SUPER. This
extra kicker from the equity component makes these securities attractive to buyers,
helps the issuers raise more cash up front, and also hedges some of the inherent risks
in their business.
Example 7.2 shows a hybrid security created to achieve a tax-efficient disposal of
a stock that has appreciated in value.
EXAMPLE 7.2: Hybrid Securities
■ Suppose you are the chief executive officer of Venturecap Co. (a fictitious name) that has bought
1 million shares in a start-up company Starttofly Inc. at $3 per share. The start-up has done very well
and is about to go public. Because Venturecap bought these shares through a private placement,
securities laws prevent selling those shares for two years. However, you want to dispose of these shares
because finance academics report that high prices achieved after an initial public offering usually
disappear after six months.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ What should you do? One alternative is for Venturecap to issue hybrid debt (also called a structured
note), which is a combination of a bond and a stock. Your investment banker can design premium
equity participating securities (PEPS) that allow you to get the benefits of a sale without actually
selling Starttofly’s stock.
■ How does this work? Suppose the investment banker designs a three-year bond that pays interest at an
annual coupon rate of 4 percent per year on the PEPS selling price of $20. The buyer of each PEPS
also gets a share of Starttofly after three years. Many investors like the idea of buying Starttofly for
$20 in the future and getting the coupon payments over the intermediate years, and the no-arbitrage
principle ensures that this security is equivalent to buying a bond and buying the stock.
■ Venturecap sells 1 million PEPS and raises $20 million up front. Moreover, it pays capital gains taxes
only when it sells the Starttofly stock, which is three years from now, and it gets to deduct the interest
payments of $800,000 from its taxes every year.
■ Hybrids have been sold in the US, Asia, Europe, and Australia.
APPLICATIONS AND USES OF SWAPS 135
7.4 An Introduction to Swaps
Swaps are one of the most successful uses of financial engineering. A swap is an
agreement between two counterparties to exchange (“swap”) a series of (usually)
semiannual cash flows over the life (tenor or term) of the contract. Salomon Brothers
arranged the first swap between International Business Machines Corp. and the World
Bank in 1981.
A swap contract specifies the underlying currencies, applicable interest rates,
payment timetable, and default contingencies. Being over-the-counter (OTC) instru-
ments, swaps do not enjoy the same level of protection as exchange-traded derivatives.
You must carefully document them and, if necessary, add collateral provisions to
mitigate credit risk (see Chapter 22 for a detailed discussion of swap documentation).
As swaps are customized, they come in many flavors and varieties. The simplest
swaps go by the fancy name of plain vanilla, whereas more complex varieties are
called exotics. Parties in a plain vanilla interest rate swap exchange fixed for
floating interest payments, while those in a plain vanilla currency swap exchange
fixed payments in one currency for fixed payments in another currency. The libor
rate index (see Chapter 2) has traditionally been used as a reference rate for the
cash payments on most swap contracts. Nowadays, a large number of swaps are
denominated in the European Union’s euro as well as other major currencies like
the Japanese yen and the pound sterling. Swaps have flourished because they provide
an efficient way of transforming one cash flow to another with very low transaction costs.
Our discussion of how the pursuit of lower transaction costs is shaping the global
markets is relevant here (see Extension 1.1).
7.5 Applications and Uses of Swaps
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Swaps have many financial engineering uses:
■ Transforming loans. A company can transform a fixed interest rate loan into a floating
interest rate loan and vice versa through a plain vanilla interest rate swap.
■ Hedging currency risk. A Japanese manufacturer repatriating profits from its US
subsidiary can convert dollars into yen through a cross-currency swap and thereby
protect itself from currency risk.
■ Altering asset/liability mix. The 1970s showed that a sharp rise in interest rates can
bankrupt a savings and loan bank (S&L) specializing in home mortgage lending.
A rate hike increases the value of the liabilities (because the deposits must be paid
higher interest) but has little immediate impact on the assets (which primarily
consist of fixed rate mortgage loans). An S&L can hedge this risk via an interest
swap, in which the S&L pays a fixed rate but receives a floating rate. This swap
lowers the mismatch in cash flows from both the S&L’s floating rate liabilities
(deposits) and fixed rate assets (mortgage loans).
■ Creating payoffs that are hard to attain. As there are no exchange-traded derivatives
on jet fuel, an airline can hedge its fuel costs by entering into a commodity swap
136 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
that pays the average aviation fuel price computed over a month in exchange for a
fixed payment.
■ Avoiding market restrictions. An international investor who is prevented from invest-
ing in a country’s stocks by local governmental rules can go around this restriction
by entering an equity swap that pays her an amount tied to the return on the local
stock index.
7.6 Types of Swaps
As in other financial markets, the swap market has its own financial intermediaries.
These swap facilitators (also called swap banks or just banks) either act as a broker
(which is becoming less common these days) or perform the role of a dealer. For
simplicity, we omit these dealers from our examples. This hardly alters the story. Just
think of these dealers as intermediaries who siphon off a few basis points from the
cash flows without affecting the basic nature of the transactions.
Interest Rate Swaps
A plain vanilla interest rate swap exchanges fixed for floating interest rate
payments at future dates. As all payments are denominated in the same currency,
a notional principal is used to determine the payments, and cash flows are netted.
Example 7.3 shows the setup and cash flows from a plain vanilla interest rate swap
(see Chapter 22 for a valuation formula).
EXAMPLE 7.3: A Plain Vanilla Interest Rate Swap (Fixed-for-Floating)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Contract Setup
■ Suppose that Fixed Towers Inc. borrowed $100 million for three years at a floating rate of the one-year
libor index rate, but it now wants to switch to a fixed interest rate loan. Floating Cruisers Co. raised
the same sum and for the same period at a fixed rate of 6 percent but now wants to switch to a floating
rate, the one-year libor rate index.
■ They swap. They work out a deal in which Fixed agrees to pay Floating 6 percent and receive the
libor rate index on the $100 million notional. This will go on for three years, which is the swap’s
tenor or term. The swap is shown in Figure 7.2.
■ Though swaps often have quarterly payments, let’s assume for simplicity that they exchange cash flows
at year’s end. As a result of this agreement, Fixed’s cash flows are (– libor + libor – 6 percent =)
– 6 percent times the notional, whereas Floating’s cash flows are (– 6 percent + 6 percent – libor =)
– libor times the notional. The swap switched each of their existing loans to the other type.
■ It makes no sense to simultaneously exchange $100 million for $100 million. That’s why $100 million
is called the notional principal—it exists as a notion, never changes hands, and is only used for computing
payments.
TYPES OF SWAPS 137
The Cash Flows
■ We compute the net payments assuming that the libor index rate takes the values 5, 6, and 7.50
percent, respectively, over years 1, 2, and 3.
(1st year, libor = 5 percent)
Fixed owes Floating 6 percent times the notional.
Floating owes Fixed 5 percent times the notional.
Fixed will pay Floating 1 percent of $100 million or $1,000,000 at the end of the first year.
(2nd year, libor = 6 percent)
Net payment is 0.
FIGURE 7.2: Example of a Plain Vanilla Interest
Rate Swap Transforming a Fixed
Rate Loan Into a Floating Rate
Loan
Before Swap
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Fixed Towers Floating Cruisers
Raised $100 million Raised $100 million
Cost libor per year Cost 6 percent per year
libor 6 percent
After Swap (for three years)
bbalibor
Fixed Towers Floating Cruisers
Net cost 6 percent 6 percent Net cost libor per
per year year
libor 6 percent
138 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
(3rd year, libor = 7.50 percent)
Fixed owes Floating 6 percent times the notional.
Floating owes Fixed 7.50 percent times the notional.
Fixed receives from Floating 1.50 percent of $100 million or $1.5 million at the end of the third year.
Then the contract ends.
■ In reality, counterparties in swaps and other OTC derivatives contracts sign detailed bilateral
agreements. Since its founding in 1985, the International Swap Dealers Association (later
renamed the International Swaps and Derivatives Association) has been instrumental in developing
convenient documents for these contracts.
Forex Swaps
“Rule, Britannia, rule the waves.”2 After attaining supremacy in the early eighteenth
century, the Royal Navy was the world’s most powerful navy for over two hundred
years, and the British pound reigned supreme. However, the pound lost some of its
luster in the mid-twentieth century when the Bretton Woods system established the
dollar as the standard currency.
The dollar’s dominance is seen in the foreign exchange (forex) market, where
most transactions involve exchange of foreign currencies for dollars, and vice versa.
The term exchange rate denotes the price of one currency in terms of another.
The currencies are quoted either in direct or American terms as dollars per unit
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
of foreign currency (say, $2 per pound sterling) or in indirect or European terms as
foreign currencies per unit of dollar (say, £0.5 per dollar). An exchange rate between
two currencies other than US dollars is referred to as a cross-rate.
Just as you buy an ounce of gold by paying dollars, think of foreign currencies as
tradable goods with dollar prices. If the price goes up and you have to spend more
dollars to buy one unit of a foreign currency (say, £1 = $3), then we say that the
dollar has depreciated. Fewer dollars for a unit of foreign currency (say, £1 = $1)
means that the dollar has appreciated.
Forex swaps are widely used in forex markets for managing currency risks. A forex
swap (or FX swap) is a single transaction with two legs: (1) the contract starts with
a spot exchange of one currency for another and (2) it ends with a reverse exchange
of these currencies at a subsequent date. Example 7.4 shows the basic idea behind a
forex swap.
2
“Rule, Britannia!” is a patriotic British national song, which originated from the poem with the same
name written by Scottish poet James Thomson in the mid-eighteenth century.
TYPES OF SWAPS 139
EXAMPLE 7.4: A Forex (FX) Swap
■ Suppose that Americana Bank has $200 million in excess funds for which Britannia Bank (both
fictitious names) has an immediate need. They enter into a forex swap with a tenor of one month.
■ The spot exchange rate SA is $2 per pound sterling in American terms, and its inverse SE = 1/2 =
£0.50 per dollar in European terms. Americana gives $200 million to Britannia and receives (Dollar
amount × Spot exchange rate in European terms) = $200 × £0.50 per dollar = £100 million today.
■ The annual simple interest rates are i = 6 percent in the US and iE = 4 percent in the United
Kingdom. After one month (using a monthly interest rate computation for simplicity), Britannia
repays Americana,
Dollar amount × Dollar return
= $200 × [1 + (0.06/12)]
= $201 million
and Americana repays Britannia,
Foreign currency amount × Value of one pound invested for one month
= £100 × [1 + (0.04/12)]
= £100.3333 million
Instead of beginning immediately, a forex swap may start at a future date. A
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
majority of forex swaps have lives that are measured in weeks rather than months.
These versatile instruments have many uses such as (1) managing cash flows in
connection with imports and exports, (2) managing borrowings and lendings in
foreign currencies, (3) handling foreign exchange balances, and (4) temporarily
exchanging excess amounts of one currency for another.
Currency Swaps
A plain vanilla currency swap (or a cross-currency swap) is an arrangement
between two counterparties involving (1) an exchange of equivalent amounts in two
different currencies on the start date, (2) an exchange of interest payments on these
two currency loans on intermediate dates, and (3) repayment of the principal amounts
on the ending date. Because it’s an OTC contract, the counterparties can modify the
terms and conditions at mutual convenience. Some currency swaps skip the exchange
of principal and just use them as notional amounts for computing interest. Interest
payments are made in two different currencies and are rarely netted.
140 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
Corporations use currency swaps to exchange one currency into another. They
do this to repatriate profits and to set up overseas factories. Example 7.5 shows how
such a swap works.
EXAMPLE 7.5: A Plain Vanilla Currency Swap (Fixed-for-Fixed)
■ Suppose that Americana Auto Company wants to build an auto plant in the UK and Britannia Bus
Corporation (both fictitious names) wants to do the same in the US. To open a factory, you need local
currency. Often a manufacturer can more easily raise cash at home because it has relationships with
local banks. Americana and Britannia both plan to do this, and seeing a large spread between foreign
exchange buying and selling rates, they decide to do the currency conversion via a currency swap.
This is a generic fixed-for-fixed currency swap that involves regular exchange of fixed payments
over the swap’s life.
The Swap Contract
■ The automakers enter into a swap with a three-year term on a principal of $200 million. The spot
exchange rate SA is $2 per pound. Americana raises 100 × 2 = $200 million and gives it to Britannia,
which, in turn, raises £100 million and gives it to Americana. Assuming annual payments, Americana
pays Britannia 4 percent on £100 million and Britannia pays Americana 6 percent on $200 million
at the end of each year for three years. The companies are basically exchanging their borrowings.
The Cash Flows
■ The swap ends after three yearly payments, and the principals are handed back. Figure 7.3 shows these
cash payments.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Extension 7.1 provides a valuation formula for fixed-for-fixed currency swaps.
Currency swaps have several interesting features. First, a swap’s value depends
on two sets of interest rates: domestic as well as foreign interest rates of different
maturities.3 Second, although our example assumes annual payments, most swaps
involve semiannual payments. You can easily add this feature. Notice that Result 7.1
of Extension 7.1 is quite general and does not depend on whether we have annual
or semiannual cash flows. Third, the swap value is arbitrage-free. If an errant dealer
quotes a different price for the same swap, then you can buy “the cheaper and sell the
more expensive swap” and create arbitrage profits. Fourth, anything goes in the swaps
markets—these OTC contracts can be designed to satisfy the whims and fancies of
the counterparties.
3
Using the covered interest rate parity formula (see Result 12.3), you can eliminate the foreign interest
rate and rewrite the formula in terms of the spot and the currency forward rate.
TYPES OF SWAPS 141
FIGURE 7.3: Plain Vanilla Currency Swap
Today: Swap Begins with Exchange of Principal
$200 million
Americana Britannia
Raises $200 million £100 million Raises £100 million
in US at 6 percent in UK at 4 percent
$200 million £100 million
US investors UK investors
Intermediate Dates: Interest Payment at the End of First and Second Years
$12 million
Americana £4 million Britannia
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
$12 million £4 million
US investors UK investors
Maturity Date: Interest Payment and Principal Repayment
$212 million
Americana £104 million Britannia
$212 million £104 million
US lenders UK lenders
142 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
EXTENSION 7.1: Valuing Fixed-for-Fixed Currency Swaps
We expand upon Example 7.5 to develop a valuation formula for currency swaps.
EXT. 7.1 EX. 1: A Valuation Formula for a Currency Swap
Contract Setup
■ The situation and the terms guiding this swap are as follows:
- Consider the US dollar as the domestic currency and pounds sterling as the foreign currency.
- The swap’s tenor T is three years, and it involves annual interest payments.
- The spot exchange rate SA is $2 per pound in American terms, which is SE = 1/SA = £0.50 per dollar in
European terms.
- Americana raises the principal amount L = $200 million at the coupon rate i of 0.06 or 6 percent per year.
- Britannia raises an equivalent sum at the coupon rate iE of 0.04 or 4 percent per year. The principal amount
LE is calculated by multiplying the dollar principal by the spot exchange rate: L × SE = LE = 200 × 0.50
= £100 million.
■ Americana pays US investors:
C = Principal (in domestic currency) × Interest rate (on dollar principal)
=L×i
= 200 × 0.06
= $12 million
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
at the end of each year for three years.
- Britannia pays UK investors an annual amount:
CE = Principal (in “European” currency) × Interest rate (on foreign currency principal)
= LE × iE
= 100 × 0.04
= $4 million
The Cash Flows
■ The swap generates the following cash flows (see Figure 7.3):
Time 0: Today
The swap begins with an exchange of the principals. Americana pays L = $200 million (shown with a solid
line arrow) and gets LE = £100 million from Britannia (shown with a broken line arrow).
TYPES OF SWAPS 143
Time 1 (after one year) and Time 2 (at the end of the second year)
Interests are paid at year’s end. Americana pays CE of £4 million to Britannia (see solid line arrow), which
passes it on to the original British investors (see broken line arrow). Britannia pays C of $12 million to
Americana (broken line arrow), which passes it on to the original American investors (solid line arrow).
The companies have exchanged their borrowings. Americana has transformed a dollar liability into a sterling
liability, while Britannia has done the opposite and accepted a dollar loan.
Time 3 (after three years)
The swap ends when the companies return their principals and final interest payments. Americana pays
£4 million in£4 million in interest and repays the principal of £100 million to Britannia and receives
$212 million.
A Pricing Model
■ Today’s cash flows cancel and do not affect the swap valuation. As Americana will be receiving dollars and
paying sterling, we can value this swap to Americana by adding up the present value of all the dollar payments
that the company receives and subtracting from this sum the present value (in dollar terms) of all the sterling
payments that the company makes.
■ To compute present values, first find the zero-coupon bond (or “zero”) prices in the two countries. They
come from the government securities in the US and the UK. The UK government–issued bonds are called
gilt securities. Like their American cousins, gilts have a whole range of maturities, semiannual coupon
payments, STRIP features, callable issues, and the presence of inflation-protected securities. The UK has even
issued consols, which are perpetual bonds that never mature.
■ Denote US zero prices by B(t) and UK zero prices by B(t)E , where t = 1, 2, and 3 years are the times to
maturity (see Ext 7.1 Tab. 1).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Discount the cash flows by multiplying them by the respective zero-coupon bond prices (Ext. 7.1 Tab. 1 shows
how this works). Multiplying the dollar receipts by the US zero prices, we get the following:
Present value of Americana’s dollar receipts
= PV of dollar interest payments + PV of dollar principal
= [(Price of a US zero − coupon bond maturing in one year × Interest payment) +
(Price of a zero − coupon bond maturing at time 2 × Cash flow)
+ (Price of a zero − coupon bond maturing in year 3 × Interest)]
(7.1)
+ (Price of a zero − coupon bond maturing in three years × Principal)
= [B (1) C + B (2) C + B (3) C] + [B (3) L]
= (0.95 + 0.89 + 0.82) × 2 + (0.82 × 200)
= $195.92 million
= PV
144 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
EXT. 7.1 TABLE 7.1: Zero-Coupon Bond Prices in the United States
(Domestic Country) and the United Kingdom (Foreign
Country)
Time to Maturity (in years) US (Domestic) Zero-Coupon UK (Foreign) Zero-Coupon
Bond Prices (in dollars) Bond Prices (in pounds
sterling)
1 B(1) = $0.95 B(1)E = £0.96
2 B(2) = 0.89 B(2)E = 0.91
3 B(3) = 0.82 B(3)E = 0.85
Multiplying the sterling payments by UK zero prices, we get the present value of Britannia’s sterling receipts,
which equals Americana’s payments. Consequently,
Present value of Americana’s sterling payments
= PV of interest payments in pounds sterling + PV of sterling principal
= [B (1)E CE + B (2)E CE + B (3)E CE ] + [B (3)E LE ]
(7.2)
= (0.96 + 0.91 + 0.85) × 4.00 + 0.85 × 100
= £95.88 million
= PVE
■ Convert this into US dollars by multiplying it by the spot exchange rate:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Present value of Americana’s sterling payments in terms of today’s dollars
= PVE × SA
(7.3)
= 95.88 × 2
= $191.76 million
■ The dollar value of the foreign-currency swap to Americana Auto Company is
Value of the currency swap to the domestic investor
= (Present value of Americana’s dollar receipts)
− [Present value of Americana’s sterling payments (in dollar terms)]
(7.4)
= PV − PVE × SA
= 195.92 − 191.76
= $4.16 million
This means Americana must pay Britannia $4.16 million to enter into the swap.
TYPES OF SWAPS 145
■ Owing to credit risk, it is prudent to avoid any upfront payment and make this a par swap, which has a zero
value. You can do this by tweaking the interest rates until this happens. For example, if you keep i fixed, you
will find that iE = 0.047647 makes this a par swap.
We restate this as a result (see Ext 7.1 Result 1).
EXT. 7.1 RESULT 1
Valuing a Plain Vanilla Currency Swap
A currency swap begins at time 0, when
1. a domestic counterparty (“American”) pays the principal L dollars to a foreign counterparty
(“European” or “Foreigner”) and receives an equivalent sum LE in foreign currency
2. American agrees to pay a fixed interest on LE to Foreigner in exchange for a fixed dollar interest on L
at the end of each period (times t = 1, 2, . . . , T)
3. the counterparties repay the principals on the maturity date T
Then,
The value of the swap to American, VSWAP
= Present value of dollar receipts − Present value of foreign currency payments (converted
(7.5)
dollars into US at the spot rate)
= PV − PVE × SA
where PV is the present value of future dollar cash flows received by American, PVE is the present
value of future foreign currency payments, and SA is today’s spot exchange rate in American (dollar)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
terms.The present value of the cash flows is computed as
T
PV = B (t) C + B (T) L (7.6)
∑
t=1
T
= B (t) i + B (T) L (7.7)
[∑
t=1 ]
where B(t) is today’s price of a US zero-coupon bond that pays one dollar at time t and C is the coupon
or the interest on the dollar principal that is paid at times t = 1, 2, . . . , T (computed as dollar principal
times the simple interest rate or L × i).
PVE is similarly computed by attaching a subscript E and by replacing domestic by foreign currency
cash flows and zero-coupon bond prices in the preceding formula.
146 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
Commodity and Equity Swaps
Example 7.6 illustrates a simple commodity swap.
EXAMPLE 7.6: Hedging Airline Fuel Costs with a Commodity Swap
■ Aviation fuel is a major cost for airlines. Some airlines hedge this price risk by entering into a commodity
swap in which they pay a fixed price and receive the average aviation fuel price computed over the
previous month.
■ Suppose HyFly Airlines (fictitious name) needs 10 million gallons of aviation fuel per month, which
it buys on a regular basis from the spot market. Seeing a huge rise in oil (and aviation fuel prices) in
recent years, HyFly is scared that a continued run-up will ruin the airline, and it decides to hedge its
exposure by entering into a commodity swap.
■ The swap is structured as follows (see Figure 7.4 for a diagram of this swap):
- The notional is 10 million gallons of aviation fuel.
- The swap has a two-year term, and the payments are made at the end of each month.
- HyFly pays the dealer a fixed price of $1.50 per gallon or 10,000,000 × 1.50 = $15 million each
month.
- The dealer pays a floating price that is the average spot price of fuel during the previous month.
- The payments are net.
■ Suppose last month’s average price is $1.60 per gallon.
- Then the dealer pays 10,000,000 × 1.60 = $16 million at the end of the this month.
- As the payments are netted, the airline receives $1 million.
- This extra payment helps alleviate the airline’s higher aviation fuel price paid in the spot market.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Suppose oil prices fall unexpectedly and last month’s average fuel price was $1.30. The savings that
HyFly generates from the lower price have to be surrendered to the dealer, who must be paid
$2 million$2 million by the terms of the contract.
This commodity contract is hard to price because it involves an average price (see
Jarrow and Turnbull [2000] for a valuation formula for this swap).
Equity swaps are similar to the swaps previously discussed. The simplest equity
swaps involve counterparties exchanging fixed rate interest payments for a floating
payment that is tied to the return on a stock or a stock index. The payments are
computed on a notional and exchanged at regular intervals over the swap’s term. The
index may be a broad-based index (such as the Standard and Poor’s 500) or a narrowly
defined index (such as for a specific industry group like biotechnology). Equity swaps
allow money managers to temporarily change the nature of their portfolios and bet
on the market. Suppose that we find two portfolio managers who hold completely
opposite views about the direction of the stock market. They can enter into an equity
swap in which the bullish manager receives a floating rate tied to the market index
and pays the bearish manager a fixed interest rate.
TYPES OF SWAPS 147
FIGURE 7.4: Hyfly Airline Hedging Fuel Price
Risk
For 2 Years
$1.50 per gallon
Hyfly Airline Swap Dealer
Average spot price
Aviation fuel Spot price of
aviation fuel
Spot
Market
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Credit Default Swaps
Credit default swaps (CDS) have been hailed during good times as a “wonder
of modern finance” and vilified during the financial crisis of 2007 as a “‘Ponzi
scheme’ that no self-respecting firm should touch” (see “Credit Derivatives: The
Great Entangling,” Economist, November 6, 2008). Designed during the 1990s
(by J. P. Morgan & Co.), CDS combine the concepts of insurance and swaps to create
a contract that plays a useful economic role. It enables the holder of a risky bond to
sell the credit risk embedded in the bond to a counterparty willing to bear it. Let’s
consider how a CDS works.
A CDS is a term insurance policy on an existing corporate or government bond.
A typical CDS matures in one to five years and has a notional amount equal to
the face value of the bond being insured, with most CDS being in the $10 to
$20 million range. The seller of a CDS (the insurer) receives a periodic premium
payment (usually quarterly) for selling the insurance. If a credit event occurs—the
bond defaults—then the buyer of the CDS (the insured) receives the face value of
148 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
the debt from the seller but hands over the bond in return.4 The CDS contract ends
if a credit event occurs.
The premium payment is quoted as a CDS spread (say, 100 basis points per
annum). The actual dollar payment is the prorated spread (adjusted for the length of
the payment period) times the notional. The CDS spread is quoted in the market so
that no cash is exchanged when the CDS is written. This implies that the CDS has
zero value at initiation. Example 7.7 illustrates the working of a CDS.
EXAMPLE 7.7: Hedging Default Risk
■ A pension fund, the Worried Lenders Fund (a fictitious name) has purchased a bond issued by
Candyfault Enterprises (CE) (another fictitious name). The bond has five years until it matures,
pays a 5.5 percent annual coupon, and has a notional of $5 million. The pension fund manager is
concerned that Candyfault might default on its interest payments over the next year owing to the
current recession. She wants to hedge its default risk but does not want to sell the bond.
■ To hedge the default risk on CE over the next year, it buys a one-year CDS on CE with a notional of
$5 million. The CDS quoted spread is 400 basis points per year, paid quarterly. The Worried Lenders
Fund buys this CDS.
■ The quarterly premium payment made by the Worried Lenders Fund is
0.04 × (1/4) × $5 million = $50, 000
■ The CDS seller gets this payment quarterly until the CDS expires or until CE defaults on its bond,
whichever comes first. In the event of default, the CDS seller pays the Worried Lenders Fund $5
million, and it takes possession of the CE bond.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Pricing of this simple contract, however, is a difficult task. CDS are also written on
asset backed securities, called ABS CDS, but their cash flows are somewhat different
(see Jarrow 2011).
7.7 Summary
1. We emphasize a build and break approach, in which derivatives can be broken
down into component parts or combined with other derivatives to create new
ones.
2. Once the building blocks are understood, one can venture into financial
engineering, a new discipline that applies engineering methods to financial
economics.
3. A swap is an agreement between two counterparties to exchange a series of cash
payments over its life. Swaps are successful vehicles for transforming one kind of
4
This is the case of physical delivery. If cash delivery occurs instead, the swap buyer receives the difference
between the face value of the debt and the market price of the defaulted debt issue (usually determined
through an auction).
QUESTIONS AND PROBLEMS 149
cash flow to another. Swaps are OTC contracts that can be customized to the
counterparties. The simplest kinds of swaps have the following structure:
a. A plain vanilla interest rate swap exchanges fixed for floating rate payments.
The principal, called the notional principal, never changes hands but is used
only for computing cash flows. Payments are netted because the cash flows are
denominated in the same currency.
b. A forex swap involves a spot exchange of foreign currencies that is followed by
a reverse exchange of equivalent amounts (principal plus relevant interest) at a
future date.
c. A plain vanilla currency swap (or a cross-currency swap) is an arrangement
between two counterparties involving (1) an exchange of equivalent amounts
of two different currencies on the starting date, (2) an exchange of interest
payments on these two currency loans on intermediate dates, and (3) repayment
of the principal on the ending date.
d. A commodity swap involves an exchange of an average price of some notional
amount of a commodity in exchange for a fixed payment.
e. An equity swap exchanges fixed rate interest payments for a floating payment
that is tied to the return on a stock or a stock index.
f. A credit default swap is a term insurance policy on a bond.
7.8 Cases
Times Mirror Company PEPS Proposal Review (Harvard Business School Case
296089-PDF-ENG). The case examines the design of a premium equity partici-
pating security for tax-efficient disposal of an appreciated common stock.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Privatization of Rhone-Poulenc—1993 (Harvard Business School Case 295049-
PDF-ENG). The case illustrates how a recently privatized French company can
use financial engineering to design an investment plan for employees that would
enhance their ownership of company shares.
Advising on Currency Risk at ICICI Bank (Harvard Business School Case 205074-
PDF-ENG). The case studies how a large Indian bank can hedge risks from
multiple interest rates and currencies by using a variety of derivatives like swaps,
options, and futures contracts on interest rates and currencies.
7.9 Questions and Problems
7.1. Briefly describe a plain vanilla interest rate swap. Is the notional principal paid
out at the swap’s beginning and maturity?
7.2. Briefly describe a currency swap. Is the notional principal paid out at the swap’s
beginning and maturity?
7.3. How does a plain vanilla interest rate swap differ from a currency swap?
The next two questions use the following information:
American Auto enters into a plain vanilla currency swap deal with European
Auto. American raises $30 million at an 8 percent fixed interest rate, while
European raises an equivalent amount of €33 million at 9 percent. They swap
150 CHAPTER 7: FINANCIAL ENGINEERING AND SWAPS
these two loans. The swap lasts for three years, and the payments take place at
the end of each year.
7.4. What is the principal? Does it change hands at the beginning and end of the
swap?
7.5. Calculate the gross payments involved and indicate who pays what in this swap
deal.
The next two questions use the following information:
Esandel Bank enters into a plain vanilla interest rate swap with a swap facilitator
Londoner Inc. Esandel pays a fixed amount of 8 percent per year to Londoner,
which in turn pays a floating amount libor + 1.50 percent. The principal is
$15 million. The swap lasts for five years, and the payments take place at the
end of each year.
7.6. What is the notional principal? Does it change hands at the beginning and end
of the swap?
7.7. Who is in the “receive fixed” situation? Who is in the “pay-fixed” situation?
7.8. Calculate the net payments involved and indicate who pays what in this swap
deal if the libor takes on the values 7.00 percent, 6.50 percent, 7.00 percent,
7.50 percent, and 6.00 percent at the end of the first, second, third, fourth, and
fifth year, respectively.
Consider the swap in Extension 7.1:
- The automakers enter into a swap with a three-year term on a principal of
$200 million.
- The spot exchange rate SA is $2 per pound. Americana raises 100 × 2 =
$200 million and gives it to Britannia, which in turn raises £100 million
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and gives it to Americana.
- Americana pays Britannia 4 percent per year on £100 million and Britannia
pays Americana 6 percent per year on $200 million at the end of each year
for three years.
Now assume that the companies make payments every six months: the swap ends
after six semiannual payments, and the principals are handed back after three
years.
7.9. Using zero-coupon bond prices (maturing every six months) given in
Table 7.3, compute the value of this swap.
7.10. How can one use a currency swap to hedge currency risk?
7.11. How can one use an interest rate swap to change a fixed rate loan into a floating
rate loan?
7.12. How can a savings and loan bank use an interest rate swap to match its long-
term fixed rate investments with the risks of its short-term floating rate demand
deposit obligations?
QUESTIONS AND PROBLEMS 151
TABLE 7.3: Zero-Coupon Bond Prices in the
United States (Domestic Country)
and the United Kingdom (Foreign
Country)
Time to Maturity US (Domestic) UK (Foreign)
(in years) Zero-Coupon Bond Zero-Coupon Bond
Prices (in dollars) Prices (in pounds
sterling)
0.5 B(0.5) = $0.99 B(0.5)E = £ 0.98
1 B(1) = 0.97 B(1)E = 0.96
1.5 B(1.5) = 0.95 B(1.5)E = 0.93
2 B(2) = 0.93 B(2)E = 0.91
2.5 B(2.5) = 0.91 B(1.5)E = 0.88
3 B(3) = 0.88 B(3)E = 0.85
7.13. What is an equity swap? What is a commodity swap?
7.14. Suppose an investor is precluded from investing in a country’s stocks by
government regulations but can invest in the country’s government bonds.
Design a swap to overcome these regulations.
7.15. What is a forex swap? Explain how it works.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
7.16. In Chapter 2, we described how the US Treasury STRIPS worked. Explain
how Wall Street firms had financially engineered a product similar to STRIPS.
7.17. What is a callable bond? When would it be used?
7.18. What is a putable bond? When would it be used?
7.19. What is a convertible bond? How would you break it down into simpler parts
and price it?
7.20. Check the Internet or other sources and answer the following questions:
a. Briefly describe the Troubled Asset Relief Program (TARP) program of the
US government.
b. What was the derivative used in the program? Why would the US
government trade such a “toxic thing” as a derivatives contract?
c. How would you approach the pricing of the security used in the TARP?
d. Discuss the success of the program to date.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
II
Forwards and Futures
CHAPTER 8
Forwards and
Futures Markets
CHAPTER 9
Futures Trading
CHAPTER 11
CHAPTER 10
The Cost-of-Carry
Futures Regulations
Model
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
CHAPTER 12
The Extended
Cost-of-Carry Model
CHAPTER 13
Futures Hedging
8
Forwards and Futures
Markets
8.1 Introduction 8.4 Futures Contract Features
and Price Quotes
8.2 Applications and Uses of
Forwards and Futures Commodity and Financial Futures
Contracts
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
8.3 A Brief History of Forwards The Gold Futures Contract
and Futures
Gold Futures Price Quotes
Early Trading of Forward and
The Exchange and Clearinghouse
Futures-Type Contracts
US Futures Exchanges and the 8.5 Commodity Price Indexes
Evolution of the Modern Futures EXTENSION 8.1 Doomsters and
Contract, 1848–1926 Boomsters
Recent Developments, 1970
8.6 Summary
Onward
8.7 Cases
8.8 Questions and Problems
APPLICATIONS AND USES OF FORWARDS AND FUTURES 155
8.1 Introduction
In 1688, the world’s first futures exchange, the Dojima Rice Exchange, opened its
doors. It was located in Osaka, which was Japan’s commercial center at the time
known as the “country’s kitchen.” In those days, if you controlled rice, then you
controlled the Japanese economy. The Dojima Exchange devised a fairly advanced
system of trading that has many commonalities with the way futures trade today. A
day’s trading time was determined by a firebox system. The market would open in the
morning with the lighting of a wick in a hanging wooden box, and it would close
when the wick was completely burned down. If the wick was extinguished early,
then all trades of that day were canceled. You can imagine how losing traders tried
to puff out the fire early, while winning traders and exchange officials attempted to
foil such plans. Raw emotions would sometimes get expressed through brawls and
fistfights (see Alletzhauser 1990; West 2000).
As this story relates, emotions (and stakes) can run high at futures exchanges, the
workings of which are described in this chapter. However, as floor trading is being
phased out, action these days takes place behind electronic terminals in airconditioned
rooms. We take up forward and futures contracts from where we left off in Chapter 4.
We discuss the usefulness of forward and futures contracts and present a brief history.
Although futures-type contracts have traded in many times and places, the evolution
of the modern futures contract really began in mid-nineteenth-century Chicago and
was completed in about seventy-five years. After decades of relative calm, the 1970s
saw a period of astonishing changes in futures markets. These developments can be
classified into two categories: (1) the introduction of new futures contracts, and (2)
the opening of new exchange and the automation of trading. Finally, we examine a
gold futures contract and study how futures prices get reported in the financial press
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and quoted by vendors.
8.2 Applications and Uses of Forwards
and Futures
As forward and futures markets are growing by leaps and bounds (and huge securities
markets don’t exist without good reason), they must be serving a useful role in
the economy. And this growth is not just in futures on commodities like corn and
gold. The 1970s saw the introduction of financial futures (“financials”), which
are futures contracts based on financial assets (such as foreign currencies) and on
economic or financial variables (such as interest rates). These contracts have become
very popular.
What is the utility of forward and futures contracts including those written on
financial assets and financial variables? Futures and forwards take investors far beyond
the realm of ordinary stocks and bonds and, in the language of Star Trek, “where no
man has gone before.” Consider the following observations:
1. By purchasing forwards and futures in advance, a trader can smooth unexpected
future price fluctuations that can come from demand–supply mismatches.
156 CHAPTER 8: FORWARDS AND FUTURES MARKETS
2. Forwards are useful for acquiring a commodity at a fixed price at a later date.
Futures are seldom used for trading assets because of the costs involved with
physical delivery. However, their features make them excellent tools for managing
price risk.
3. Forwards and futures help make the market more “complete.” In a complete
market, sufficient securities trade such that investors can construct portfolios to
obtain all possible probability distributions over future payoffs.
4. They help traders to speculate. Speculators have a beneficial presence in most
futures markets. Speculators increase liquidity and enhance market efficiency, but
destabilizing speculation can also take place, and we look at several manipulation
cases in Chapter 10.
5. They allow investors to leverage their capital and hold large trading positions
without tying up cash. However, like speculation, leverage cuts both ways and can
sometimes do more harm than good. For example, futures bets destroyed Barings
PLC, a venerable British bank that had existed for over two hundred years.
6. The process of trading forwards and futures generates useful information about
future price expectations called price discovery, an important function discussed
in greater detail in chapter 10.
8.3 A Brief History of Forwards and
Futures
Forward trading began early in the dawn of civilization (see Table 8.1) and thrives to
this day. The history of forwards and exchange-traded futures may be divided into
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
three phases:
1. Early trading of forward and futures-type contracts. Forward contracts evolved separately
in many parts of the world. Sometimes they traded in centrally located markets
with standardized trading rules and contract terms that prompted some scholars to
label them as futures contracts.
2. Evolution of the modern futures contract. Today’s futures contracts started taking shape
in the United States during the nineteenth century and arrived at its present form
over the next seventy-five years.
3. Developments since the 1970s. An astonishing variety of futures were conceived and
marketed during the last three decades of the twentieth century.
Early Trading of Forward and Futures-Type Contracts
Forward contracts have been used to reduce risk for eons. Forwards traded in India
around 2000 BC, in ancient Greece and Rome, in medieval Europe, and in Japan
during the seventeenth century (see Table 8.1). Records indicate that Amsterdam
grain dealers used both futures and options in the 1550s and the decades that followed
to hedge trade goods shipped far distances.
A BRIEF HISTORY OF FORWARDS AND FUTURES 157
TABLE 8.1: Early History of Trading Forward and
Futures-Type Contracts
Year Development
2000 BC and onward Forward contracts traded in Greece,
India, and the Roman Empire
Eleventh and twelfth centuries AD Medieval fairs
and onward
1531 The first commodities exchange in
Antwerp, Belgium
1550s onward Futures and options contracts traded in
Amsterdam, Netherlands
1571 Royal Exchange for trading
commodities in London, England
1688 Dojima Rice Exchange in Osaka, Japan
One of the earliest commodities exchanges was the Royal Exchange of London,
which was inaugurated by Queen Elizabeth I in 1571. Today the original location
houses an upscale shopping mall. The renowned London Metal Exchange and the
London Stock Exchange both trace their ancestry back to the Royal Exchange.
The world’s oldest stock exchange, as noted in Chapter 2, opened in Amsterdam
in 1602. Soon futures and options contracts started trading. This was a time when
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the tulip was gaining popularity in Holland. A few decades later, the Dutch got
involved in an extraordinary speculative fervor over rare tulip bulbs, and futures and
options were devised to help the speculating public. As with other price bubbles,
the famous tulip bulb price bubble ended when prices came crashing down and
derivatives trading dried up.
As noted at the beginning of this chapter, Osaka’s Dojima Rice Exchange opened
its doors in 1688. It is considered the world’s first organized futures market because
the contracts were standardized, had fixed life, required margin payments, and
involved clearinghouses. The exchange operated for centuries and was closed down
in 1939 because of wartime controls.
US Futures Exchanges and the Evolution of the
Modern Futures Contract, 1848–1926
Despite interesting developments elsewhere, modern futures trading really began
in Chicago in the mid-nineteenth century with the establishment of the Chicago
Board of Trade (CBOT or CBT). Why Chicago? Located on the windswept shores
of Lake Michigan, the “Windy City” is close to the fertile farmlands of the great
Midwest. Nineteenth-century Chicago became America’s major transportation and
158 CHAPTER 8: FORWARDS AND FUTURES MARKETS
TABLE 8.2: Evolution of Today’s Futures Contract (1848–1926)
Year Development
1848 The CBOT was established to standardize grains trading
1849–50 To arrive contracts started trading at the CBOT
1865 General rules were developed at the CBOT that standardized futures contracts; traders were
required to post margins
1874 The Chicago Produce Exchange was formed to trade farm products; it became the Chicago
Butter and Egg Board in 1898, and in 1919, it became the CME, with its own clearinghouse
1877 Speculators were allowed to trade at the CBOT
1882 NYMEX was established
1925 A clearinghouse was established at the CBOT
distribution center for agricultural produce. Farmers shipped their grains to Chicago,
much of which then moved eastward to where most Americans lived.
But agriculture is a seasonal business. Prices crashed at harvest times as farmers
flooded Chicago with grains and rose again as the grain was used and became scarce.
A need was felt for a central marketplace to smooth such demand–supply imbalances.
The futures market was devised to fill this void.
To standardize the quality and quantity of grains traded, eighty-two merchants
founded the CBOT in 1848. Spot market trading began immediately, and more
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
developments followed. Table 8.2 shows how US futures trading evolved to reach
its present-day perfection. In the mid-1860s, the CBOT streamlined trading and
further standardized contracts in terms of size, quality of commodities, delivery dates,
and places. The contracts came to be called futures. Most important, the CBOT
also required performance bonds called margins from both buyers and sellers. This
eliminated counterparty risk. In 1877, the CBOT allowed speculators to trade. Prior
to this time, only producers or purchasers could trade. In contrast, speculators trade
for profits and not for hedging purposes.
Meanwhile there was a need to organize the trading of butter, eggs, poultry, and
other farm products. Merchants trading these commodities founded the Chicago
Produce Exchange in 1874; this became the Chicago Butter and Egg Board in 1898
and emerged as the Chicago Mercantile Exchange (CME or the Merc) in 1919.
It became widely known for trading traditional futures contracts on dairy, meat, and
poultry products.
A group of merchants founded the Butter and Cheese Exchange of New York,
renamed the New York Mercantile Exchange (NYMEX) in 1882. Another
innovation came when clearinghouses were established by the CME after its reor-
ganization in 1919, and by the CBOT in 1925. As noted earlier, in exchange-
traded derivatives markets, a clearinghouse plays the crucial role of clearing a trade
by matching the buyer and seller, recognizing and recording trades, and (nearly)
A BRIEF HISTORY OF FORWARDS AND FUTURES 159
eliminating counterparty default risk. These developments almost created the futures
contract as we know it today.
Recent Developments, 1970 Onward
Futures markets have undergone significant changes since 1970 (see Table 8.3).
We classify these changes into two categories: (1) the introduction of new futures
contracts, and (2) the opening of new exchanges and the automation of trading.
NEW FUTURES CONTRACTS In 1972, the CME created the International
Monetary Market (IMM) division for trading forex futures. The seven foreign currency
futures that started trading at the IMM became the first successful financial futures traded.
But exchanges don’t always succeed. Credit for the first currency futures exchange
goes to the International Commercial Exchange, an exchange founded by the
members of the New York Produce Exchange in 1970 but that closed its doors in
1973. Its failure has been attributed to several causes: (1) bad timing (as it was founded
when the Bretton Woods agreement was still in force), (2) limitations of a small
exchange, (3) small contract size, (4) a failure to adequately promote itself, (5) high
margin requirements, and (6) high brokerage fees (see “New Game in Town,” Wall
Street Journal, May 16, 1972).
TABLE 8.3: Milestones in the History of Futures Trading since 1970
Year Development
1972 The first successful financial futures, foreign currency futures, started trading at the newly created
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
IMM at the CME
1973 The CBOE was founded by members of the CBOT
1974 US Congress passed the Commodity Futures Trading Commission Act that created the CFTC
1975 The first interest rate futures, GNMA futures contracts, started trading at the CBOT
1977 Treasury bond futures began trading at the CBOT
1981 The first cash-settled contract, the eurodollar futures, started trading at the CME
1982 The NFA, a self-regulatory organization made up of firms and people who work in the futures
industry, was established
1982 Options on US Treasury bond futures started trading at the CBOT
1982 The first index futures contract, the Value Line index futures, started trading at the Kansas City
Board of Trade; within months, the S&P 500 stock index futures started trading at the CME
1984 The first international futures link, the CME/SIMEX mutual offset trading link, was established
1992 A postmarket global electronic transaction system, Globex, was launched by the CME and Reuters
160 CHAPTER 8: FORWARDS AND FUTURES MARKETS
Novel products and new markets generate a need for greater regulatory oversight.
In 1974, the US Congress passed the Commodity Futures Trading Commission
(CFTC) Act, which created the CFTC as the “federal regulatory agency for futures
trading.” In 1982, the National Futures Association (NFA), an industry body,
was established. NFA does a slew of self-regulatory activities, which (remember from
Chapter 1) enhances the reputation of the futures markets and reduces the need for
greater federal oversight.
In the 1970s, oil shocks and other supply-side disturbances hiked up the inflation
rate, causing interest rates to be more volatile. In 1975, the CBOT launched the
Government National Mortgage Association (GNMA) futures, which protected
mortgage holders from interest rate risk. This was the first interest rate derivative.
The Ginnie Mae futures contract was the spark that ignited the creation of the vast
global market for interest rate derivatives, one of which was the eurodollar futures
introduced by the CME in 1981.
Eurodollar futures were the world’s first cash-settled futures contracts. Normally,
when a futures contract is carried to delivery, the short collects the futures price
and delivers the underlying commodity. A cash-settled futures contract has no
such physical delivery provision. If taken to maturity, a cash payment equal to the
difference between the settlement price at the contract’s end and the previous day’s
settlement price closes out the contract. Without cumbersome delivery, it became
possible to design an extraordinary range of derivatives based on notional variables
like indexes, interest rates, and other intangibles.
The year 1982 saw the introduction of options both on futures and on stock index
futures. In Chapter 5, we introduced options on spot, which are options contracts
based on an underlying asset or a notional variable. Equity options on blue chip stocks
like Ford and International Business Machines are its simplest examples. By contrast,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
options on futures or futures options have a futures contract as the underlying.
This market started in 1982 with US Treasury bond futures. Most futures options
now trade side by side with their underlying futures contracts on their respective
exchanges.
Index futures came as a natural next development. These are cash-settled futures
contracts that have an index as the underlying variable. In 1982, the Kansas City Board
of Trade it got regulatory approval to trade futures on the Value Line Index. This
is an average of over 1,700 stocks tracked by Value Line Incorporated, a well-known
provider of investment advice. Months later, the CME introduced the Standard and
Poor’s 500 index futures (S&P 500), which soon overtook the former. Index futures
are very popular for implementing different trading strategies, including hedging,
speculation, and index arbitrage.
Subsequent years saw the introduction of an assortment of new contracts. Futures
contracts on interest rates in major currencies like dollars (eurodollars), the euro, the
Japanese yen, and the pound sterling are offered by major exchanges including the
CME Group. Moreover, they trade on short-term interest rate benchmarks, which
are often determined by a method similar to the libor index rate (see Chapter 2)
and use the suffix IBOR or BOR (for “interbank offered rate”). Examples include
interest rate futures on BUBOR in Budapest, HIBOR in Hong Kong, KORIBOR
FUTURES CONTRACT FEATURES AND PRICE QUOTES 161
in Korea, MOSIBOR in Moscow, and STIBOR in Stockholm. The emergence of
the European Community has also created the need for euro-based futures. Just open
the Wall Street Journal or go to the website of one of the futures exchanges to get an
idea of this diversity.
NEW EXCHANGES AND THE AUTOMATION OF TRADING In terms
of trading volume, some of the world’s largest futures and options exchanges have
opened in recent years. There are futures exchanges even in countries with emerging
financial markets. In all, nearly a hundred futures exchanges operate today. Many
of them are modeled after US futures exchanges, and they even relied on American
technical expertise to be set up. The European Union’s motto “United in Diversity”
is an apt description of today’s financial world: numerous exchanges with diverse
product lines in distant lands have proliferated with customers linked through a
single electronic terminal. Recent years have seen a trend toward extended trading
hours. As discussed in Chapter 3, IT advances have revolutionized securities trading
by making innovations like electronic exchanges, online brokerage accounts, and
network linkage fairly commonplace.
8.4 Futures Contract Features and Price
Quotes
This section discusses a futures contract’s specifications and how prices and trades are
recorded.
Commodity and Financial Futures Contracts
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Futures contracts are divided into those on commodities or financial futures.
Commodity futures usually refer to futures contracts on commodities like cattle,
coffee, copper, and corn. Financial futures refer to futures contracts on financial
assets like bonds, currencies, or even on notional variables like stock indexes and
interest rates (see Table 8.4).1
Greater connectivity among exchanges has considerably expanded the menu of
choices available for trading futures. Most futures have similar features, especially those
trading on US exchanges. As “seeing one is seeing them all,” we select a popular
contract and examine its specifications. Gold is the natural choice since gold futures
trade on several US exchanges and on at least ten foreign ones.
The Gold Futures Contract
Let us examine some features of the gold futures contract that trades on the COMEX
division of the CME Group (see Example 8.1).2
1
Legally the term commodity has a broader definition that includes nearly all goods and services (see
Title 7 [Agriculture], Chapter 1 [Commodity Exchanges], Section 1a of the US Code), and under this
legal definition, financial futures are considered commodity futures.
2
June 11, 2018, www.cmegroup.com/trading/why-futures/welcome-to-comex-gold-futures.html.
162 CHAPTER 8: FORWARDS AND FUTURES MARKETS
EXAMPLE 8.1: The Gold Futures Contract Traded on the COMEX Division of
the CME Group
■ A trading or ticker symbol identifies the futures contract. The gold futures contract has the ticker
symbol GC.
■ A trading month identifies the contract’s delivery month. At any time, GC contracts trade with
“delivery during the current calendar month; the next two calendar months; any February, April,
August, and October falling within a 23-month period; and any June and December falling within
a 72-month period beginning with the current month.”3 Table 8.5 gives actual price quotes for
these contracts and shows some months that trade. The choice of trading months is crucial for
many commodities. As gold is produced steadily throughout the year, it makes sense to have delivery
months spaced at regular intervals throughout the year. Futures on agricultural commodities often
have delivery months clustered around harvest times.
■ The trading unit or contract size (also called even lot) is one hundred troy ounces per contract.
Some commodities including gold have futures contracts of multiple sizes.
■ Trading hours In both the CME Globex (electronic trading system) and CME ClearPort (system
for clearing over-the-counter trades), trading takes place from 24 hours a day (with a daily break from
4 PM to 5 PM Chicago time), seven days a week. Floor trading for gold futures has been phased out.
■ The price quotation is in US dollars and cents per troy ounce. If the quoted price is $500, then the
position size for one contract is 500 × 100 = $50,000.
■ The minimum price fluctuation or tick size measures the minimum price jump. GC has a tick
size of 10 cents ($0.10) per troy ounce or $10 per contract. For example, if the gold futures price
is $500.00, the next higher price would be $500.10, the next would be $500.20, and so on. The
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
long position gains and the short position loses by $10 for each of these jumps. If the price falls from
$500.00 to $498.50, then the long loses $150 per contract.
■ The maximum daily price fluctuation or daily price limit is the maximum price change allowed
in a day. Some commodities have price limits (see the relevant exchange websites for details).
■ The last trading day is the third to last business day of the delivery month. This is the last day
on which a GC contract maturing in that month can trade.
■ Delivery is the traditional way of ending a futures contract. Gold delivered against a GC contract
must bear a serial number and an identifying stamp of a refiner approved and listed by the exchange.
Delivery must be made from a depository that has been licensed by the exchange.
■ Besides delivery, a contract may end with an exchange of physicals (EFP) for the futures contract.
In this case, the buyer and the seller privately negotiate the exchange of a futures position for a physical
position of equal quantity. EFPs take place at locations that are away from the exchange floor.
■ The delivery period is the time period during which delivery can occur. The first delivery day
is the first business day of the delivery month, while the last delivery day is the last such day. Like
most futures, the GC seller decides if a delivery should take place, and if so, when to start the process
and what grade to deliver. This makes sense because historically, sellers were hedgers who held the
physical commodity for sale.
FUTURES CONTRACT FEATURES AND PRICE QUOTES 163
TABLE 8.4: Examples of Futures Contracts That Trade in the United
States
Category Examples of Commodities
COMMODITY FUTURES
Grains and oilseeds Corn, soybean, wheat, canola, . . .
Livestock and meat Cattle, pork belly, hogs, . . .
Food and fiber Coffee, orange juice, sugar, cotton, . . .
Metals Copper, gold, platinum, silver, . . .
Energy Gasoline, heating oil, natural gas, electricity, . . .
Other products Ammonia, lumber, plywood, . . .
FINANCIAL FUTURES
Interest rates Canadian bonds, euribor, eurodollars, Treasury securities, . . .
Currencies Euro, peso, pound, Swiss franc, . . .
Indexes Dow, Nasdaq, Nikkei, S&P 500, . . .
■ Trading at Settlement allows traders to trade any time during trading hours but at a delivery price
that is the settlement price determined at the day’s close. This is only allowed for certain months.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Regarding grade and quality specifications, in case of delivery, the seller must deliver 100 troy
ounces of gold, subject to various specifications regarding assaying fineness, the form in which it can
be delivered, and acceptable refiners (details are available from the exchange on request.)
■ The position limit restricts the number of contracts that a speculator can hold in particular
commodities (see the relevant exchange websites for details). Position limits try to prevent traders
from manipulating by accumulating huge long positions and then squeezing short sellers. Several
forward and futures market manipulation cases discussed in chapter 10 show that this risk is real.
■ Margins are security deposits required to open and maintain futures positions. A nonmember
opening a speculative position has to keep an initial margin (security deposit needed to open a
futures position) per contract. The maintenance margin is the amount that if the account value
falls to this level, the trader has to supply more funds to bring the margin back to the initial margin.
For the exact dollar margins required, see the relevant exchange websites.
3 June 11, 2018, www.cmegroup.com/trading/why-futures/welcome-to-comex-gold-futures.html.
164 CHAPTER 8: FORWARDS AND FUTURES MARKETS
TABLE 8.5: COMEX Gold Futures Prices at Market Close on Thursday,
August 24, 2007
GOLD (Comex Division, NYMEX) 100 troy oz.; $ per troy oz.
Last Open High Open Low High Low
Aug 2007 668.10 n/a 662.10 668.80 662.10
Sep 2007 667.00 659.60 659.60 667.00 659.60
Oct 2007 671.20 665.00 664.00 673.10 661.00
Nov 2007 n/a n/a n/a n/a n/a
Dec 2007 678.00 671.20 670.90 679.00 666.70
Feb 2008 683.00 n/a 673.90 683.10 673.00
April 2008 679.40 679.40 679.40 679.40 679.40
June 2008 694.00 685.00 685.00 694.00 685.00
Aug 2008 n/a n/a n/a n/a n/a
Oct 2008 n/a n/a n/a n/a n/a
Dec 2008 710.00 n/a 701.30 710.00 701.30
Feb 2009 n/a n/a n/a n/a n/a
April 2009 n/a n/a n/a n/a n/a
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
June 2009 n/a n/a n/a n/a n/a
Dec 2009 n/a n/a n/a n/a n/a
June 2010 n/a n/a n/a n/a n/a
Dec 2010 n/a n/a n/a n/a n/a
June 2011 n/a n/a n/a n/a n/a
Dec 2011 n/a n/a n/a n/a n/a
June 2012 n/a n/a n/a n/a n/a
Gold Futures Price Quotes
Wire news services, websites, and many major daily newspapers carry futures price
quotes. As there are some differences in how they report, we discuss price quotes
for the COMEX gold futures. Table 8.5 gives these prices for Thursday, August
24, 2007, from the NYMEX website (this contract currently trades on the CME
Group’s COMEX division; quotes are available online4 ). Newspapers provide basic
FUTURES CONTRACT FEATURES AND PRICE QUOTES 165
8.6: COMEX Gold Futures Prices at Market Close on Thursday,
August 24, 2007
mex Division, NYMEX) 100 troy oz.; $ per troy oz.
Most Recent Settle Change Open Interest Estimated Volume Last Updated
668.00s +0.10 261 204 8/24/2007 2:57:39 pm
668.70s −1.70 234 41 8/24/2007 2:58:42 pm
671.50s −0.30 35815 2069 8/24/2007 3:32:29 pm
n/a 0 n/a n/a n/a
677.50s +0.50 188352 59490 8/24/2007 3:41:59 pm
683.40s −0.40 16264 659 8/24/2007 1:49:42 pm
689.10s −9.70 17710 9 8/24/2007 1:41:15 pm
694.70s −0.70 14606 37 8/24/2007 1:42:53 pm
700.10s 0 7541 47 8/24/2007 1:41:15 pm
705.50s 0 1469 n/a 8/24/2007 1:41:15 pm
710.90s −0.90 15978 181 8/24/2007 1:41:15 pm
716.60s 0 11076 n/a 8/24/2007 1:41:15 pm
722.20s 0 1630 n/a 8/24/2007 1:41:15 pm
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
727.80s 0 10983 n/a 8/24/2007 1:41:15 pm
744.90s 0 2566 n/a 8/24/2007 1:41:15 pm
762.30s 0 2569 n/a 8/24/2007 1:41:15 pm
780.10s 0 2585 n/a 8/24/2007 1:41:15 pm
798.10s 0 1724 n/a 8/24/2007 1:41:15 pm
816.50s 0 1035 n/a 8/24/2007 1:41:15 pm
835.90s 0 0 n/a 8/24/2007 1:41:15 pm
information like where the contract trades, the size of the contract, and how prices
are quoted. This redundant information is not reported on the NYMEX website.
The first column lists the delivery months. The last trade price is reported next:
this was $678.00 for the December 2007 contract. Many exchanges begin each day’s
trading with an opening call period for each contract month. The first few bids,
offers, and traded prices during this initial time period establish an opening price or
an opening range. This can give rise to different open high and open low prices,
166 CHAPTER 8: FORWARDS AND FUTURES MARKETS
which were $671.20 and $670.90, respectively, for December 2007. Regular trading
begins after everyone gets a chance to execute trades at the opening call.
High and low are the highest and the lowest traded prices per contract per day.
The high price for December 2007 gold was $679.00, while the low was $666.70.
The difference between the high and the low prices determines the trading range
for the trading session.
Like the opening, trading ends in a special way. The closing minutes are usually
the busiest because this is when many traders close out their open trades to avoid
margins for overnight positions. During the brief ending period, called the market
close, or the last minute of trading, known as the closing call, a closing price or
a closing range is established.
For actively traded contracts with few price fluctuations, the exchange’s Settlement
Committee picks a settlement price from this closing range, often the last traded
price. The settlement price (abbreviated settle) is the fair value of the contract at
market close. For December 2007, the most recent settlement price was $677.50.
The task is harder for thinly traded contracts. For example, April 2008 traded nine
contracts, which may have traded hours before the close. Moreover, many contracts,
like February, April, and June for the year 2009, did not trade at all. In this case,
the Settlement Committee considers the spreads relative to other futures prices.
The spreads between different maturity futures prices are generally quite stable. For
example, the spread between the April and February 2008 settlement prices was
$5.70, between June and April $5.60, and between August and June $5.40. Here the
spreads between the February and April 2009 and the April and June 2009 settlement
prices (contracts with no current trading volume) were set at $5.60.
The next column reports the change in the settlement price from the day before.
For example, the December 2007 settlement price rose by $0.50 to end at $677.50.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Some tables report lifetime highs and lows, which refer to the highest and
the lowest traded prices recorded for a contract month since it began trading. This
information is absent here.
Open interest is the number of outstanding contracts for a particular maturity
month. December 2007 has 188,352 open contracts. Each of these contracts must
end through a closing trade, a delivery, or a physical exchange.
The estimated trading volume during the current session is 59,490 contracts
for December 2007, and the last updated gives the time of the last trade.
The Exchange and Clearinghouse
US securities exchanges are governed by a tight set of rules and regulations. Tradition-
ally, they have been organized as voluntary, nonprofit institutions. Futures exchanges
in the United States have clearinghouses affiliated with them. A clearinghouse forms
the foundation on which a futures exchange is built. As noted before, it serves
several useful functions: it clears trade (and collects a clearance fee for each contract
cleared), it guarantees contract performance to all traders, and it collects margin from
COMMODITY PRICE INDEXES 167
clearing members. A clearinghouse has matched books; therefore it has no market
risk, but it does have counterparty risk. To alleviate this credit risk, it has several
layers of financial protection. On top of margins, a member must keep guarantee
funds with the clearinghouse. In case a member defaults, the clearinghouse first pays
from his margin, then from the member’s guarantee fund, subsequently from the
surplus fund of clearance fees, and finally from the guarantee fund of all members
according to some prespecified formula.
8.5 Commodity Price Indexes
We have used equity market indexes, the Dow Jones Industrial Average, and the S&P
500 index to understand the history of stock markets, to gauge market sentiments,
to evaluate portfolio performance, and even as the underlying for derivatives. Can
indexes also play such roles in commodity markets? The answer is yes.
During the 1950s and 1960s, inflation was low, commodity prices were relatively
stable, and volatility was not a matter of concern. The 1970s, 1980s, and 1990s
exhibited more volatility but no discernable trend in the index. A dramatic upward
trend began in 2003 with increased volatility. Not surprisingly, this increased volatility
has been accompanied by an increase in the trading of commodity futures. Extension
8.1 presents an interesting debate concerning the direction of future commodity
prices between the doomsters and the boomsters.
Several commodity price indexes have been created. The most popular com-
modity price index, the Reuters/Jeffries CRB index (RJ/CRB), is computed
by averaging the futures prices of nineteen different commodities. Several other
commodity price indexes are popular in the United States. Standard and Poor’s
publishes world production–weighted S&P GSCI indexes based on twenty-four
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
commodities, which consist of four separate but related S&P GSCI indexes: the
spot index, the excess return index, the total return index, and a futures price
index. Introduced in 1998, the Dow Jones–AIG Commodity Index Family
(DJ-AIGCI) is composed of indexes based on futures contracts on nineteen physical
commodities. DJ-AIGCI indexes are computed on an excess and total return basis,
report spot as well as forward indexes, and are available in US dollar and several
foreign currency versions. The Rogers International Commodity index (RICI)
was developed in 1998 by commodity investor James B. Rogers (see RICI Handbook
2008). Covering more commodities than other indexes, RICI is a dollar-based total
return index that corresponds to a collection of commodities representing the global
economy including futures traded in different exchanges, in different countries, and
quoted in different currencies.
Notice the following about these indexes. First, as these indexes are weighted
averages, the booming energy sector gets high weights, ranging from 30 plus percent
for RJ/CRB and DJ-AIGCI to over 70 percent for the S&P GSCI index in 2008.
Second, because these indexes are based on short-lived futures contracts that mature
on a regular basis, the managers of the index must periodically roll the futures index
168 CHAPTER 8: FORWARDS AND FUTURES MARKETS
by replacing the expiring futures with similar new contracts. Third, many derivatives
have been created based on these indexes. For example, there are exchange-traded
funds or exchange-traded notes based on these indexes, and futures or options on
futures trade on these indexes. Exchange-traded funds (ETFs) are similar to mutual
funds but trade in an exchange on a real-time basis. Exchange-traded notes are
bonds issued by an underwriting bank, which promises to pay a return (minus any
fees) based on the performance of a market benchmark (like a commodity price
index) or some investment strategy.
EXTENSION 8.1: Doomsters and Boomsters
Two Opposing Views on the Future Direction of Commodity Prices
The first camp, popularly known as doomsters, owes its intellectual origins to the great English economist
Thomas Malthus, whose ideas on the economics of population growth came to be known as the Malthusian
doctrine. It argues that as productivity increases and new lands are discovered, the food supply increases in
arithmetic proportions, such as 1, 2, 3, . . . . However, due to the population’s propensity to reproduce, it increases
geometrically, such as 1, 2, 4, 8, . . . . Consequently, events like war, famine, and pestilence check the population
growth relative to the food supply so that people always stay just at the subsistence level. No wonder economics
has been called the dismal science!
Building on this intellectual foundation, the doomster school argues that a growing number of humans rely on
a limited supply of natural resources, and scarcity will force commodity prices to rise through time. A prominent
proponent of this neo-Malthusian view is Stanford University professor Paul R. Ehrlich. His 1968 book The
Population Bomb warned, “The battle to feed all of humanity is over. In the 1970s and 1980s hundreds of millions
of people will starve to death in spite of any crash programs.”
The opposite camp, popularly referred to as boomsters, espouses free-market environmentalism, which
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
argues that free markets, property rights, and a good legal system will take care of the environment. They dismiss
the other side’s gloom and doom. They cite that after the high-inflation days of the 1970s, commodity prices
failed to show a pronounced upward trend implied by increasing scarcity. The champion of the second group
was Julian L. Simon (b. 1932), a professor at the University of Maryland when he died in 1998. His 1981 book
The Ultimate Resource challenged the notion of Malthusian catastrophe and offered an alternate explanation. He
argued that the power of human beings to invent and adapt is the ultimate resource. Simon showed that after
adjusting for inflation and wage increases, most raw material prices fell over the past decades. To quote from a
song by the Beatles, “It’s getting better all the time.”
A Bet between Professor Doomster and Professor Boomster
Boomster Simon and doomster Ehrlich entered into a wager based on their views (see “When the Boomster
Slams the Doomster, Bet on a New Wager—Clash of Eco-Titans Rekindles a Rancorous Core Debate over
Future of the Earth,” Wall Street Journal, June 5, 1995). The bet’s conditions were as follows: on September 29,
1980, Ehrlich would buy on paper $200 worth of each of five strategically important metals (chromium, copper,
nickel, tin, and tungsten). If, ten years later, their inflation-adjusted prices went up, then Simon would pay the
inflation-adjusted value increase for each of these portfolios. If they declined, Simon would receive a payment
instead. Note that the professors were trading a derivative that was essentially a combination of five forward
contracts with some price adjustments.
SUMMARY 169
Ehrlich and his colleagues saw an arbitrage opportunity: “I and my colleagues, John Holdren (University of
California, Berkeley) and John Harte (Lawrence Berkeley Laboratory), jointly accept Simon’s astonishing offer
before other greedy people jump in.” Ten years later, they were in for a rude surprise—for a variety of reasons,
the real prices for all these metals and the nominal prices for three declined. Wired magazine (“The Doomslayer,”
www.wired.com/wired/archive/5.02/ffsimon_pr.html) reports that the drubbing was particularly hurtful because
Simon had given Ehrlich and his colleagues a priori advantage by letting them select the five metals. A month
after the bet ended, Professor Ehrlich quietly mailed Professor Simon a check for $576.07.
8.6 Summary
1. Both a forward and a futures contract fix a price for a later transaction. They have
many differences: unlike a forward, a futures contract is (1) regulated, (2) exchange
traded, (3) standardized, (4) liquid, (5) guaranteed by a clearinghouse, (6) margin
adjusted, (7) daily settled, (8) usually closed out before maturity, and (9) has a range
of delivery dates.
2. Futures and forwards have many uses: (1) they help smooth out price fluctuations
that can come from demand–supply mismatches, (2) they help create a complete
market, (3) they help people to speculate, (4) they allow traders to leverage
their capital, (5) they make efficient trading possible, and (6) they help generate
information and aid in price discovery.
3. Forward trading is as old as antiquity. Today’s futures contract started in nineteenth-
century Chicago and more or less arrived at its present-day form over the next
eighty years. An astonishing variety of futures were introduced since the 1970s—
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
financial futures being a major innovation. Moreover, exchanges merged, new
exchanges emerged, and electronic trading replaced the traditional open-outcry
method of pit trading throughout the world.
4. As futures are standardized, they allow quicker transactions. Futures contract
specifications are available on an exchange’s website.
5. Many vendors carry futures price quotes. A typical quote table gives the opening
price, the day’s high and low, the settlement price, the change in settlement price
from the previous day, lifetime high and low, open interest for each contract, and
other information like the estimated volume and open interest for all contracts on
that particular commodity.
6. Several commodity price indexes like the Reuters/Jeffries CRB index, S&P GSCI
indexes, Dow Jones–AIG Commodity Index Family, and Rogers International
Commodity index are used to gauge the direction of commodity prices and to
create derivatives that allow investors to add a commodity price exposure to their
portfolios.
170 CHAPTER 8: FORWARDS AND FUTURES MARKETS
8.7 Cases
CME Group (Harvard Business School Case 711005-PDF-ENG). The case describes
the CME Group, the world’s largest commodities exchange, futures and options
on futures contracts, history, regulation, and the strategic choices the company
faced.
Bringing OTC Back to the Exchange: Euronext. liffe’s Launch of ABC
(Harvard Business School Case 706489-PDF-ENG). The case examines value
creation, market design, and competitive positioning issues for a derivatives
trading exchange in the context of launching matching, clearing, and confirmation
services for the over-the-counter market.
Lessons Learned? Brooksley Born and the OTC Derivatives Market (A)
(Harvard Business School Case 311044-PDF-ENG). The case studies a proposal to
regulate the over-the-counter derivatives market, whose lack of implementation
might have been one of the factors that contributed to the financial crisis of
2007–9.
8.8 Questions and Problems
8.1. Where and when was the world’s first modern futures exchange founded?
Explain how trading took place on this exchange.
8.2. Discuss the four different categories into which the history of futures contracts
since 1970 has been classified, giving two examples of some of the major
developments under each of the four categories.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
8.3. Consider a fairly illiquid futures contract that has not traded for days. Do you
still need a settlement price for this contract? If so, how would the exchange
go about determining this settlement price?
8.4. Explain the difference between closing price and settlement price. Are they
the same or different? Which price is used for marking-to-market?
8.5. If regulation is bad for business, why does the National Futures Association put
a significant amount of regulation on its members?
8.6. Consider the gold futures contract traded in the COMEX division of the CME
group. What is the trading unit size and minimum tick size?
8.7. Using Table 8.5, what is the last price on the June 2008 futures contract? What
does the last price mean?
8.8. Using Table 8.5, what is the open interest on the June 2008 futures contract?
What does the open interest mean?
8.9. Using Table 8.5, looking at the open, high, and low prices on the June futures
contract, what can you tell about the trend of futures prices on this contract
during the day?
8.10. Using Table 8.5, which three futures contracts have the most trading activity?
QUESTIONS AND PROBLEMS 171
8.11. Today, are futures contracts only traded on US exchanges? Explain your answer.
8.12. Is trading in futures contracts mainly for hedging, or speculation, or both?
Explain your answer.
8.13. Explain why forward contracts have been trading for centuries. What economic
function do they perform? What improvement did futures contracts provide
over forward contracts? Explain your answer.
8.14. The following questions concern the doomsters and the boomsters.
a. What were the opposing views promoted by Professors Ehrlich and Simon?
b. Describe the wager between Professors Doomster and Boomster and the
outcome of this bet.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
9
Futures Trading
9.1 Introduction Equality of Futures and Spot Price
at Maturity
9.2 Brokers, Dealers, and the
Equality of Forward and Futures
Futures Industry
Prices before Maturity
9.3 Trading Futures Convergence of the Basis
9.4 Margin Accounts and Daily
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
9.6 Trading Spreads
Settlement
9.7 Summary
Daily Settlement
9.8 Cases
9.5 Futures and Forward Price
Relations 9.9 Questions and Problems
BROKERS, DEALERS, AND THE FUTURES INDUSTRY 173
9.1 Introduction
The movie Trading Places made a caricature of futures trading. In a particular sequence,
Dan Aykroyd and Eddie Murphy entered the trading pit, visibly nervous and yet
pumped up to speculate by shorting orange juice futures. They suddenly realized
that the powerful Duke brothers were trying to corner the OJ futures market. This
was in anticipation of crop damage from a frost in the orange-growing states. Some
traders thought that the Duke brothers had inside information on the extent of the
losses and joined the fray. Futures prices increased. Dan, Eddie, and other sellers
soon felt that they were sliding down a bottomless pit. Suddenly it all changed. The
commerce secretary announced on television that the frost had bypassed the orange-
growing regions. This implied a rich harvest with soft prices. Dan and Eddie were
ecstatic: OJ futures prices rapidly fell, and they made a ton of cash; the Duke brothers
went completely broke—and all of this happened in an hour! Although the movie is
fictionalized, it does capture the dark side of futures trading (it is suspiciously similar
to the Hunts brothers’ silver manipulation story in the next chapter).
To put this movie in context, this chapter describes the mechanics of trading
futures contracts. We discuss brokers, dealers, and others who work in the industry,
then we list various ways of moving in and out of futures. This is followed by a
description of margins and daily settlement. Next we explore various properties
of futures and forward prices. We end with a discussion of futures spread trading
strategies.
9.2 Brokers, Dealers, and the Futures
Industry
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Brokers match buyers and sellers in futures and earn commissions for this service.
They also represent their customers to the exchanges and clearinghouse. Dealers step
in and take the other side of trade when no one else offers a better price. Like a
used car dealer who maintains an inventory and posts a retail price for trading cars,
dealers in futures markets keep an inventory of futures contracts and quote bid and
ask prices to maximize income. Moreover, brokers and dealers also act as first-level
regulators because they have a duty to report irregularities and fraud to the exchange
authorities and government regulators. Brokers and dealers in futures markets may
do business as individuals, or they may be organized as associations, partnerships,
corporations, or trusts. They must register with the Commodity Futures Trading
Commission (CFTC).
To open a futures trading account, one must go through a futures com-
mission merchant (FCM) or an introducing broker (IB) registered with an
FCM. An IB’s role and responsibilities are limited: an IB may seek and accept
orders but must pass them on to a carrying broker (usually an FCM) for exe-
cution, and an IB cannot accept funds—funds must be directly deposited with
the carrying broker. In contrast, an FCM provides a one-stop service for all
aspects of futures trading: solicit trades, take futures orders, accept payments from
customers, extend credit to clients, hold margin deposits, document trades, and
maintain accounts and trading records. Numerous Associated Persons (APs)
174 CHAPTER 9: FUTURES TRADING
work for FCMs, IBs, and others, and help people trade (see CFTC’s website
https://www.cftc.gov/ConsumerProtection/EducationCenter/registration.html for
a description of the intermediaries who facilitate the trading process).
Where does one go for trading advice? There are commodity trading advisors
who dispense wisdom through newsletters or advise clients on an individual basis.
Most of them are technical analysts who use past price patterns to predict future
price movements. To invest with professional managers, one can go to commodity
pool operators (CPOs). They run commodity pools (funds), which are mutual
fund–type operations speculating in futures. Rarely are these pools stellar performers:
the returns are often negative, and when positive, they often underperform the market
on a risk-adjusted basis. Moreover, they charge large management fees and require
big margin deposits.
9.3 Trading Futures
Entering into a long or a futures position is easy: one calls a broker and places
an opening order. But, there are four ways of ending a futures position: a clos-
ing transaction, physical delivery, cash settlement, or an exchange for physicals
(see Figure 9.1):
1. A closing transaction. Suppose you go long a December gold futures. You are obliged
to buy gold at the futures price during the December delivery period. A futures
position can be ended earlier. You can do this by an offsetting transaction called
a closing transaction or a reversing trade. Just take a short position in the
same December gold futures. You are now obliged to sell gold at that day’s futures
price during December. This cancels the initial position. Tell your broker that the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
second trade was a closing transaction, and he will wipe your slate clean. Over 95
percent of futures contracts close before maturity.
2. Physical delivery. This was the only way of closing futures in the past. In a physical
delivery, the short gets cash from the long and obtains a warehouse receipt that
gives ownership of goods held in a licensed warehouse or a shipping certificate
that is a promise by an exchange-approved facility to deliver the commodity
under specified terms. A futures contract specifies (1) acceptable deliverable grades
(some contracts allow substitution of inferior grades with price discounts and
superior grades with price premiums), (2) acceptable delivery dates (usually within
the last trading month, i.e., the month that includes the last trading day), and
(3) acceptable delivery places (with price adjustments for different locations).
Though conceptually similar, the delivery methods vary slightly from exchange
to exchange. For most contracts, the short has the privilege of initiating a delivery
notice. The exchange usually matches him with the oldest existing long position.
In a typical delivery process, there would be a first notice day, a last notice day,
and a last trading day. The first notice day is the first day on which a short can
submit a notice of intention to make delivery, and the last notice day is the last
such day. The last trading day is usually one or more days before the last notice
day. Exchanges generally do not make or take delivery of the actual commodity;
they only specify how the delivery process occurs.
TRADING FUTURES 175
FIGURE 9.1: Four Methods of Ending a Futures Position
Closing Futures before the Delivery Period
Start Delivery Period
Last Trading Day
(a) Closing transaction (preferred way of ending most futures contracts)
Ending Futures during the Delivery Period
Chicago
Corn moves
Colorado
Farmer Kansas City
Corn Processor
(b) Physical delivery (inconvenient, large transaction costs, e.g., farmer ships corn to Chicago
or Kansas City and delivers corn for cash; processor has to transport corn back to Colorado)
(c) Cash settlement (no need to ship the underlying commodity; the contract ends with a
final margin adjustment on the last trading day)
Traders inform exchange Chicago
Colorado
Farmer
Corn
moves
Corn
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Processor
(d) Exchange for physicals (hedgers avoid needless transaction costs, e.g., a farmer and a corn
processor arrange delivery for payment and inform the exchange to close positions)
3. Cash settlement. In case of a cash settlement, a final adjustment is made to the margin
accounts of the buyer and seller equaling the difference between the last trading
day’s settlement price and the settlement price of the previous day—and then the
contract ends.
4. Exchange for physicals. Another possibility is a spot market trade between two parties
who have genuine buying and selling needs for the underlying commodity and
have already hedged their prices by establishing futures positions. This is called an
exchange for physicals (EFP) transaction. Afterward, they notify the exchange
and the clearinghouse about the transaction, and their slates are rubbed clean.
Example 9.1 shows how this works.
176 CHAPTER 9: FUTURES TRADING
EXAMPLE 9.1: An Exchange for Physicals
■ Suppose a farmer grows corn in Colorado. She sells short fifty corn futures to fix a selling price for her
produce. A corn processing plant is located five miles up the road from the farmer. It buys, processes,
and sells corn to the cereal maker who makes cornflakes. The corn processor goes long fifty corn
futures and fixes a buying price for the input. Both traders are hedgers: the farmer genuinely needs
to sell corn, while the processing plant legitimately needs to buy it.
■ At delivery time, why ship 250,000 (= 50 contracts × 5,000 bushels/contract) bushels of corn all
the way to Chicago or Kansas City and then ship it back to Colorado as per CME Group’s contract
stipulations? This will incur needless transportation costs and time delays.
■ It would be much better to do the transaction directly: the farmer ships corn to the processing plant
and receives the day’s settlement price (although provisions allow EFPs at a mutually agreed price).
Both report the price and the quantity traded to the Chicago Board of Trade, which then checks that
these reports match and closes the fifty long and short positions.
EFPs have grown in popularity in recent years. Most oil futures are settled by EFPs,
and their use is increasing in both gold and silver futures markets.
9.4 Margin Accounts and Daily Settlement
To buy and sell futures, one must open a futures account with a broker. To open a
futures position, one must post initial margin (or performance margin), which
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
acts as a performance bond. Most exchanges allow initial margins to be paid with
cash, bank letters of credit, or short-term Treasuries. The broker, if she is a clearing
member, keeps margin deposits with the exchange’s clearinghouse; if she is not, she
keeps them with a clearing member, who in turn keeps margin with the exchange.
This earnest money ensures that when the going gets tough, the tough do not leave
town!
Daily Settlement
When futures prices increase, a long position gains value. This makes sense: if you buy
something at a fixed price and the new purchase price increases, then you must make
a profit. The seller’s situation is just the opposite. So when the futures price increases,
the long’s account is credited with variation margin, which is the change in the
settlement price from the previous trading day, and the short’s account is debited the
same. These additions (collects) or subtractions ( pays) of variation margins at the end of
each trading day are known as daily settlement. When this is done, the margin account
is said to be marked-to-market. Traders can keep only a small amount as margin, often
5 percent or less of the position size (e.g., if the gold futures price is $1,000 and the
initial margin is $5,000, then it is only 5,000/(1,000 × 100) = 0.05 or 5 percent),
and still trade.
MARGIN ACCOUNTS AND DAILY SETTLEMENT 177
Marking-to-market substantially lowers credit risk and makes futures safer to use
than forward contracts. The reason is simple. With futures, the largest loss to the
contract is a one-day movement in the futures price. In contrast, with forwards, the
largest loss is the movement in the forward price over the entire life of the contract.
Margins buffer this smaller loss for futures, whereas collateral buffers this larger loss
for forwards.
Example 9.2 shows how daily settlement works for a futures contract.
EXAMPLE 9.2: Daily Settlement of a Gold Futures Contract
■ On Monday, Ms. Longina Long buys three gold futures contracts from Mr. Shorty Short. Both traders
have margin accounts. We refer to Monday as time 0, Tuesday as time 1, and so on. The futures contract
matures on Friday (time T or 4) when the day’s trading ends. For convenience, we use the end of the
day’s spot and settlement prices. All prices and computations are on a per ounce basis. The futures
prices are obtained from the cash-and-carry model (see Chapter 11).
■ Monday (time 0)
- The spot price on Monday, S(0), is $1,000.00.
- The futures price on Monday for the contract maturing on Friday is F(0,4) = $1,000.60. For
simplicity, we suppress T from the notation for the futures price F(0,T) and write it as F(0). We
will reintroduce it later when considering contracts maturing at different dates. Table 9.1 records
the trading day as the first column and futures prices as the second column.
- Because futures contracts always clear at fair prices, no cash changes hands when the position is
entered at the close of Monday’s trading.
■ Tuesday (time 1)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- S(1) = $1,006.00
- F(1) = $1,006.40
- Yesterday, Ms. Long locked in a price of $1,000.60 for buying gold on Friday. Now she is asked to
pay $1,006.40 instead. She is understandably upset. However, she will be indifferent if Mr. Short
pays her the price difference. By daily settlement, this amount, F(1) – F(0) = 1,006.40 – 1,000.60
= $5.80, is credited to Long’s margin account (and it is debited to Short’s margin account) after
Tuesday’s close. This is reported in the third column of Table 9.1. Long’s contract now has a futures
price of F(1) = $1,006.40.
■ Wednesday (time 2)
- S(2) = $996.00
- F(2) = $996.20
- Long would be happy to buy gold for $996.20 on Friday, while Short would be unhappy to sell
at that price. To mark-to-market the contract, Long’s account is debited by a variation margin of
F(2) – F(1) = 996.20 – 1,006.40 = – $10.20. Short’s account is credited the same amount. Values
become fair again.
■ Thursday (time 3)
- S(3) = $988.00
- F(3) = $988.20
- Long’s cash flow is F(3) – F(2) = $988.20 – 996.20 = – $8.
178 CHAPTER 9: FUTURES TRADING
■ Friday (time T or time 4)
- S(4) = $995.00 = F(4)
- The spot and futures prices equal each other at the end of Friday’s trading. They should because
a futures contract maturing in an instant is the same as a spot market transaction and therefore the
futures price must be the same as the spot price.
- Long gets F(4) – F(3) = 995.00 – 988.20 = $6.80 from Short.
TABLE 9.1: Futures Prices and Margin Account for Longina Long
Day Futures Change Daily Interest Margin Margin Margin
Settle- in Futures Gain Earned Account Call Account
ment Futures or Loss (Last on Balance Balance
Price Settle- Column × Previous (After
ment 300) Balance Adjust-
Price ment)
Mon F(0) = $15,000.00 $15,000.00
$1,000.60
Tues F(1) = 5.80 1,740 1.50 16,741.50 16,741.50
1,006.40
Wed F(2) = –10.20 –3,060 1.84 13,683.34 13,683.34
996.20
Thurs F(3) = –8.00 –2,400 1.64 11,284.98 3,715.02 15,000.00
988.20
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Fri F(4) = 6.80 2,040 1.50 17,041.50 17,041.50
995.00
How are these adjustments actually made? If your futures position gains value,
you may remove funds that are in excess of the initial margin, but if it loses value,
the broker gets uncomfortable. So there is a mandatory amount called a maintenance
margin (often set at 75 percent of the initial margin) that must be maintained in the
account at all times. If the account value touches or drops below this level, then the
broker places a margin call and requests that you come up with enough variation
margin in cash to bring your account to the initial margin level. If you fail to meet
a margin call, your broker can close out your positions. You may be liable for more
cash if your position is liquidated at a loss. To avoid this hassle, traders often keep
more funds than the required initial margin.
In reality, other issues must also be reckoned with. For example, there are
transaction costs, and the margin account earns interest. As this discussion can get
complicated quickly, to simplify the presentation we embrace the old maxim “follow
the money!” Following this maxim, we take an accountant’s perspective and keep
track of the margin account’s cash flows, including daily interest (see Example 9.3).
MARGIN ACCOUNTS AND DAILY SETTLEMENT 179
EXAMPLE 9.3: A Margin Account’s Cash Flows Including Interest Earned
■ Continuing with the previous example, let us track the margin account of Ms. Longina Long, who
buys three gold futures contracts on Monday at a futures price of F(0) = $1,000.60 per ounce. We
write the number of contracts as n = 3, the contract size as 𝜅 = 100 ounces, and the margin account’s
balance as Bal(t), where t = 0, 1, 2, 3, 4 are the days over which the futures contract lasts.
■ As a speculator who is not an exchange member, Long must keep the commodity exchange mandated
initial margin of $5,000 per contract or $15,000 in total, which we write as Bal(0). Long’s margin
account balances are plotted in Figure 9.2. Her maintenance margin is $4,000 per contract. The
margin account balance earns interest. Assume that the interest rate i(0) is 1 basis point for Monday.
This overnight interest rate changes randomly across time and gets declared at the start of each trading
day.
■ On Tuesday, the gold futures settlement price goes up to F(1) = 1,006.40. Long’s margin account
gets credited for the increase in the position’s value and earns interest on the previous day’s balance,
values reported in the fourth and fifth columns of Table 9.1, respectively. Tuesday’s margin account
balance is
Bal(1) = [Futures (Tues) − Futures (Mon)] × Number of contracts × Contract size
+ (1 + Daily interest) × Margin account balance (Mon)
= [F (1) − F (0)] × n × 𝜅 + [1 + i (0)] × Bal (0)
= 5.80 × 3 × 100 + (1 + 0.0001) × 15, 000
= 1, 740 + 15, 000 + 1.50 = $16, 741.50
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ This amount is noted in the sixth column. As there is no margin call, we repeat this in the eighth
column. Long can remove the excess margin of $1,741.50 if she wants. We assume that she keeps
it in the margin account.
■ On Wednesday, the gold futures price falls to F(2) = $996.20 and Long’s futures position loses value.
Assuming Tuesday’s interest rate i(1) is 0.00011 for the day, the new margin balance is
Bal (2) = [F (2) − F (1)] × n × 𝜅 + [1 + i (1)] Bal(1)
= −10.20 × 300 + 1.00011 × 16, 741.50
= $13, 683.34
■ Her account is now under-margined as it has fallen below the initial margin level. The broker is
uncomfortable but does nothing.
■ On Thursday, the gold futures price falls by another $8. Assuming i(2) is 0.00012 for the day, the new
margin balance is Bal(3) = $11,284.98. As this is below the maintenance margin level of 3 × 4,000 =
$12,000, the broker issues a margin call instructing her to come up with $3,715.02 in cash (which
is reported in the sixth column of Table 9.1 and also shown in Figure 9.2). She does this and her
180 CHAPTER 9: FUTURES TRADING
account balance is restored to the initial margin account level of $15,000, and this figure is reported
in the last column. Notice that futures traders are required to bring the margin account balance up
to the initial margin level and not to the maintenance margin level as required in stock markets.
■ On Friday, the gold futures price goes up by $6.80. If i(3) = 0.0001, her margin account balance is
$17,041.50.
■ Suppose Mr. Shorty Short took the other side of the transaction. His daily futures gains and losses
would be the opposite of Long’s, but his margin account balance would earn different interest. Also,
as the futures price did not move too much against Short, he will not get any margin calls.
■ As noted in Table 9.1, for the magnitude of these investments, the interest earned is small
($1.50 to $1.84) and perhaps unimportant. However, consider a futures position with a notational
in the tens of millions of dollars. In that case, the interest earned per day is considerable and not so
easily ignored.
FIGURE 9.2: Long’s Margin Account
Balance
Margin
Account
Balance
$17,000
16,000
15,000 Initial Margin
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
14,000
Margin Call
13,000 Maintenance
Margin
12,000
11,000
Mon Tue Wed Thu Fri Days
9.5 Futures and Forward Price Relations
What is the relation between futures and forward prices? Before addressing that
question, let’s consider an example that computes a futures contract’s payoffs when
positions may be entered and exited early (see Example 9.4).
FUTURES AND FORWARD PRICE RELATIONS 181
EXAMPLE 9.4: Futures Payoffs When Positions Are Closed Out Early
■ Consider the data from Example 9.3. Ms. Long bought gold futures on Monday. Her payoff on Tuesday
due to daily settlement is
Futures (Tues) − Futures (Mon)
+ One day’s interest on Monday’s margin account balance
= F (1) − F (0) + Interest
= $5.80 + Interest
■ Instead of on Monday, if Ms. Long buys futures on Tuesday and closes out her position with a
reversing trade on Thursday, then her payoff is
[Futures (Wed) − Futures (Tues) + One day’s interest on Tuesday’s margin account balance] +
[Futures (Thurs) − Futures (Wed) + One day’s interest on Wednesday’s margin account balance]
= F (3) − F (1) + Interest
= −$18.20 + Interest
on a per ounce basis. The interest earned on the margin account balance is unknown in advance and
hence random.
As indicated in Example 9.4, the interest earnings are dependent on the futures
price changes. Furthermore, interest rate changes affect the interest earnings on the
margin account as well. In conjunction, these two effects modify the risk of holding
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
a futures contract relative to a forward contract that has no such cash flows. Indeed,
if interest rate changes are positively correlated with futures price changes, then
the futures position benefits from interest rate movements because as cash flows are
received, more interest is earned. In this case, the risk of a futures position is reduced
slightly relative to an otherwise identical forward contract position. Conversely, if
interest rate changes are negatively correlated with futures price changes, then the
futures position suffers from interest rate movements relative to a forward position.
Here the risk is increased. It is important to note that this interest rate risk is priced into the
market clearing futures prices, driving a wedge between futures and forward prices. Recall that
forward contracts have no intermediate cash flows and therefore no explicit interest
earnings risk.
A generalization of this example gives a useful result for computing a futures trader’s
profits and losses. Let date 0 be the initiation date of a futures contract and date T be
its delivery date. Suppose a trader enters a long futures position at time t1 (which can
be as early as date 0) when the futures price is F(t1 ) and closes out her position at a
later date t2 (which can be as late as date T) when the futures price is F(t2 ). A futures
trader’s profit or loss is the sum of daily variation margins and interest earned on the
margin account balance for the time period over which the futures position is held:
Long’s payoff = Closing futures price − Initial futures price + Interest
(9.1a)
= F (t2 ) − F (t1 ) + Interest
182 CHAPTER 9: FUTURES TRADING
and
Short’s payoff = − (Closing futures price − Initial futures price) + Interest
(9.1b)
= − [F (t2 ) − F (t1 )] + Interest
on a per ounce basis. Of course, the interest earned on the long and short positions
differ.
These results make sense and ease the task of computing margin account balances
(see Example 9.5). Remember that futures prices are determined in the market so that
a trader entering a position pays nothing at the start. If the underlying commodity
increases in value, then futures prices also move up. Long benefits because the fixed
price at which she agreed to buy has increased in value, and Short loses because what
he agreed to sell at a fixed price has become more expensive.
EXAMPLE 9.5: Margin Payments for a Portfolio of Futures Contracts
■ Pearl is a speculator. Trading at the prices given in Table 9.2a, she went long two contracts of gold on
January 20, short three platinum contracts on January 21, and long five silver contracts on January 23.
She then liquidated all her futures positions on January 26. Her gains and losses are computed in Table
9.2b. Her net gain is
Long gold futures payoff + Short platinum futures payoff + Long silver futures payoff
= 2, 380 − 1, 785 + 4, 975
= $5, 570
where we have ignored the interest earned on the margin account balances for simplicity.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Let us continue with Examples 9.2 and 9.3 to explore several important properties
of forward and futures prices.
TABLE 9.2A: Futures Prices
Trading June Gold Futures, April Platinum March Silver
Date 100 troy Futures, 50 troy Futures, 5,000 troy
oz/Contract oz/Contract oz/Contract
(dollars/ounce) (dollars/ounce) (cents/ounce)
Jan 20 $989.1 $1,149.0 $2,542.1
Jan 21 992.8 1,153.3 2,600.0
Jan 22 989.4 1,154.1 2,553.3
Jan 23 1,007.0 1,169.5 2,623.3
Jan 26 1,001.0 1,165.2 2,643.2
Jan 27 1,008.4 1,175.0 2,651.8
FUTURES AND FORWARD PRICE RELATIONS 183
TABLE 9.2B: Margin Account for Pearl
Portfolio Computation of Profits/Losses Dollar Amount
Long 2 June gold (1,001 – 989.1) per ounce × 100 troy $2,380
futures on Jan 20 ounce per contract × 2 contracts
Short 3 April –(1,165.2 – 1,153.3) × 50 × 3 –1,785
platinum futures
on Jan 21
Long 5 March (2,643.2 – 2,623.3) × 5,000 × 5 ÷ 100 4,975
silver on Jan 23
Net profit 5,570
Equality of Futures and Spot Price at Maturity
In Example 9.2, suppose that at Friday’s close, the futures price is $995 but the spot
price is $997. Then anyone can buy the futures, acquire gold for $995, and sell it
immediately in the open market for $997 to pick up an arbitrage profit of $2. The
buying pressure raises the futures price, the selling pressure lowers the spot price, and
the two converge. If instead the futures price is higher than the spot price, then an
arbitrageur can sell futures, make a delivery, and take an arbitrage profit. To prevent
arbitrage opportunities, the futures price must equal the spot price at the maturity of the
contract. Of course, we need to make assumptions such as that there are no market
frictions (such as transaction costs [brokerage commissions, bid/ask spread], margin
requirements, short sales restrictions, taxes that may be levied at different rates, and
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
assets that are not perfectly divisible), that spot and futures can be simultaneously
traded, and that the futures mature on a single day.
Equality of Forward and Futures Prices before
Maturity
Are forward and futures prices equal before maturity? They are equal only under a
restrictive set of assumptions such as constant interest rates, no market frictions, and
no credit risks, but not otherwise. Example 9.6 explores this and related issues.
EXAMPLE 9.6: Equality of Forward and Futures Prices
■ Consider the gold futures contract introduced in Example 9.2. Next, consider a gold forward contract
that is similar to this futures contract. It begins on Monday, matures on Friday, and has one hundred
ounces of pure gold as the underlying. We compare two trades by Ms. Longina Long entered at
Monday’s close: buying one forward contract versus buying one futures contract.
184 CHAPTER 9: FUTURES TRADING
■ First, assume an idealistic setting in which interest rates are assumed to be zero, market frictions are
assumed away, and counterparties have no credit risk. Now let us compare the payoffs on a per ounce
basis from forward and futures trades.
■ At Friday’s close, Ms. Long’s payoff from a long forward trade is
= Spot (Fri) − Forward (Mon)
(9.2a)
= 995 − F (Mon)
where F(Mon) is the forward price for the contract maturing on Friday.
■ At Friday’s close, Ms. Long’s payoff from a long futures trade is
= [Futures (Tues) − Futures (Mon)] + [Futures (Wed) − Futures (Tues)]
+ [Futures (Thurs) − Futures (Wed)] + [Futures (Fri) − Futures (Thurs)] (9.2b)
= 995 − F (Mon)
where F(Mon) is the futures price on Monday. This happens because Futures(Fri) = Spot(Fri) = $995
at Friday’s close.
■ The two payoffs must be equal because they acquire the same asset on the same date while requiring
no cash when initiated. Hence, comparing expressions (9.2a) and (9.2b), we get
F (Mon) = F (Mon) = $995 (9.3)
■ Let us allow interest rates to be random, while holding the other assumptions the same. Again, let’s
compare payoffs from a forward and futures trade.
- At Friday’s close, Ms. Long’s payoff from a long forward trade is
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
= 995 − F (Mon) (9.4a)
- At Friday’s close, Ms. Long’s payoff from a long futures trade is
= 995–F (Mon) + Interest earned on margin account balance over the time
(9.4b)
period during which futures position is held
on a per ounce basis.
- Equate the payoffs as before and rearrange terms to get
F (Mon) = F (Mon) + Interest (9.5)
■ Expression (9.5) shows that futures and forward prices are structurally different. This is because a
futures trader earns interest on her margin account balances but a forward trader does not.
■ Introducing additional market imperfections such as transaction costs serves to complicate the relation
between forward and futures prices even more.
TRADING SPREADS 185
■ In addition, forward and futures contracts face different risks, on top of the market risks discussed
earlier. Let us consider the three additional risks stated in Chapter 1 (credit, liquidity, operational) and
see how each affects forward and futures contracts differently.
- Being exchange traded, a futures contract is essentially free from credit risk; a forward contract may
not be. It depends on the collateral relationships prearranged between the relevant counterparties.
- Over-the-counter traded forwards are more illiquid than are the exchange-traded futures, introduc-
ing greater liquidity risk in forward contracts.
- Being an over-the-counter instrument, a forward also has more operational risk than does an
exchange-traded futures contract.
This example demonstrates that under the assumption of zero interest rates,
forward and futures prices are equal. But what happens if interest rates are nonzero?
In this case, it can be shown (see Jarrow and Turnbull [2000] for a proof) that forward
and futures prices are equal, but only if interest rates are nonrandom. This is an
improvement over the first assumption, but it is still an unreasonable one. Interest
rates are random and wiggle unpredictably across time.
If one allows interest rates to be random, which is the truth, then the equality
between forward and futures prices no longer holds. The primary reason is that
futures earn interest on the margin account, while forward contracts do not. The
equality argument that gave us expression (9.3) does not apply. This is demonstrated
by continuing the previous example.
The moral of the story is that forward and futures contracts are different securities
and, consequently, have different prices. They are fraternal twins, not identical!
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Convergence of the Basis
The spot price minus the futures price is called the basis. Although it randomly
changes across time, it eventually becomes zero at maturity. This well-known property
of futures prices is known as convergence of the basis.
We can easily see from Example 9.4 that the basis takes on the values –0.60
(Monday), –0.40 (Tuesday), –0.20 (Wednesday), –0.20 (Thursday), and 0 (Friday).
Figure 9.3a plots the spot and futures prices from Examples 9.2 and 9.3, and
Figure 9.3b shows the behavior of the basis over time.
9.6 Trading Spreads
Futures traders trade naked futures or establish various hedge positions. They also
trade spreads whereby they try to exploit relative movements between two futures
prices that usually track one another for economic, historical, and other reasons.
One can set up an intracommodity spread in the futures market by simulta-
neously buying and selling futures contracts on the same commodity but with different
maturity months or an intercommodity spread by simultaneously buying and selling
futures contracts on different commodities with the same or different delivery months.
When delivery dates differ, it is called a time spread or a calendar spread.
186 CHAPTER 9: FUTURES TRADING
Why trade spreads? If you are placing directional bets with futures, you need to
have expectations about future spot prices. Alternatively, market forces sometimes
create abnormal differences between two futures prices that should closely track one
another. People speculate with spreads in such situations, hoping to profit when the
price difference corrects itself. Notice that when trading spreads, there is no risk from
the direction in which the futures prices move.
Spread trading can be tricky because spreads may vary with the seasons. For
example, although since 1970 wheat has on average cost a dollar more per bushel
than corn, an article titled “Corn, Wheat Swap Roles as Prices Surge” (Wall Street
Journal, August 9, 2011) noted the opposite: corn was trading higher than wheat, a
gap that grew to 66.75 cents in July 2011. This reversal was caused by low supply as
well as rising demand for corn from China, while wheat prices fell as many countries
had bumper crops. If futures prices move together with spot prices, this is an ideal
setting for an intercommodity spread trade that sells corn and buys wheat futures.
This strategy was adopted by many traders at the time hoping for “wheat gaining
on corn, since the main wheat harvest is over and more corn supplies are poised to hit
the market.” But many traders also avoided this strategy in 2011 because, as observed
an agricultural consultant, “the seasonal reliability is not there.”
Still, spreads are usually safer than an outright long or short position. For this
reason, brokers tend to require less margin for spread positions than for naked
positions. When closing out a futures or an option position, they regularly ask, Is this
part of a spread position? The next two examples illustrate hypothetical spread trading
when the futures prices are narrower (see Example 9.7) or wider (see Example 9.8)
than expected.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
FIGURE 9.3A: Spot and Futures Prices
Dollars
1,010
1,005
Spot Prices
1,000
Futures
Settlement
995
Prices
990
985
980
975 Time
1 2 3 4 5
TRADING SPREADS 187
FIGURE 9.3B: Behavior of Basis
Dollars
0 Time
1 2 3 4 5
–0.1
–0.2
–0.3
–0.4
–0.5
–0.6
–0.7
EXAMPLE 9.7: An Intracommodity Spread
■ Suppose you notice that the April gold futures price is $1,008, whereas that for December is $1,014 (all
prices are per ounce). In traders’ jargon, the spread is $6 to the December side. Sensing a mispricing,
you consult your extensive collection of past price data and find that this intracommodity spread has
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
historically hovered around $8.
■ Set up a spread trading strategy: buy the relatively underpriced December futures contract and
simultaneously sell the relatively overpriced April futures contract (see Figure 9.4). Traders say that
you are buying the spread for $6. No money changes hands because you are trading futures, but
you have to keep margin deposits with your broker.
■ Suppose that next month, the April futures price stays the same, but the December futures price rises
to $1,016, as predicted. Close out your positions by selling the spread for $8: sell December and buy
the April gold futures contract. You make no profit on the April contract but $2 on the December.
Stated differently, you make $2 by buying the spread for $6 and selling it for $8.
188 CHAPTER 9: FUTURES TRADING
Example 9.8 considers an intercommodity spread.
EXAMPLE 9.8: An Intercommodity Spread
■ Suppose that the platinum futures price is $1,080 per ounce and the gold futures price for a contract
expiring in the same month is $1,000 per ounce.
■ Suppose you become a technical analyst, put on green goggles, and gaze intently at price charts on
the computer screen. You note that the charts reveal that the intercommodity spread should narrow
to $50. You can speculate on this insight by selling the spread for $80: sell the relatively overpriced
platinum futures and buy the relatively underpriced gold futures contract.
■ Gold futures prices can rise, stay level, or fall, and so can platinum futures prices. Suppose that a month
from now, the spread magically narrows to $50. You earn a profit of $30
FIGURE 9.4: Trading Intra-Commodity Spread
Today Next Month
Dec futures price $1,016
December futures price $1,014
Spread $6 Spread $8
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
April futures price $1,008 April futures price $1,008
Buy Dec and sell April Sell Dec and buy April
(“buy spread for $6”) (“sell spread for $8”)
When trading spreads, remember the wry warning (attributed to renowned
economist John Maynard Keynes) that “the market can remain irrational longer than
you can remain solvent.”
9.7 Summary
1. One can initiate a futures position by calling a broker and placing an opening
order. There are many order placement strategies, the simplest being a market
order that is immediately executed. One can move out of a futures position by
(1) a closing transaction, (2) physical delivery, (3) cash settlement, or (4) an
exchange for physicals.
2. To open a futures position, one must keep an initial margin to guarantee contract
performance. When a futures price goes up, a long position gains and a short
QUESTIONS AND PROBLEMS 189
position loses profits. Through a process called daily settlement, the long’s account
is credited with variation margin (which is the change in the settlement price
from the previous trading day) and the short’s account is debited the same. The
opposite happens when a futures price goes down—long loses and short makes a
profit. After daily settlement adjustments are made, the brokerage account is said
to be marked-to-market. This substantially lowers credit risk and makes futures
safer to use.
3. Futures prices have a number of useful properties. For example, the futures price
must equal the spot price at the contract’s maturity. As the contract maturity
approaches, the basis (= Spot – Futures price) converges to zero. In an ideal setting
(when market frictions are zero), forward and futures prices are equal when interest
rates are constant, but unequal when interest rates are random.
9.8 Cases
Investment Linked to Commodity Futures (Harvard Business School Case 2930-
17-PDF-ENG). The case examines an investment linked to an index of com-
modity futures prices and explores how the index is constructed, how commodity
futures behave, and what the portfolio impacts of such an investment might be.
Futures on the Mexican Peso (Harvard Business School Case 296004-PDF-ENG).
The case considers the issues that the Chicago Mercantile Exchange faces regarding
how to design, and whether and when to introduce, a futures contract on the
Mexican peso.
Alcoma: The Strategic Use of Frozen Concentrated Orange Juice Futures
(Harvard Business School Case 595029-PDF-ENG). The case explores price
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
risk management in the orange juice industry when an increase in orange tree
production led to a surplus production of orange juice.
9.9 Questions and Problems
9.1. What is a closing transaction on a futures contract? Do you need to take delivery
to terminate a futures contract position?
9.2. What does marking-to-market mean?
9.3. What risk do margin accounts on futures contracts and marking-to-market
minimize? Explain your answer.
9.4. What is the difference between initial margin and variation margin of a futures
contract? Provide a simple example to explain your answer.
9.5. Explain what a margin call is and when it occurs on a futures contract.
9.6. Ignoring credit risk, if interest rates are constant and equal to zero, do forward
prices equal futures prices? Why?
190 CHAPTER 9: FUTURES TRADING
9.7. What is the basis of a futures contract? Give a simple example to explain your
answer. What does convergence of the basis mean with respect to a futures
contract?
9.8. Ignoring credit risk, when interest rates are random, must forward prices and
futures prices be equal? Explain why or why not.
9.9. Suppose you go short one contract of gold at today’s closing futures price
of $1,300of $1,300 per ounce. Suppose that your brokerage firm requires an
initial margin of 5 percent of the position size ($130,000 in this example) and sets
the maintenance margin at 80 percent of the initial margin. Contract size is one
hundred ounces. Keep the margins constant throughout this example. Track the
value of your margin account if the closing futures prices are as follows. Clearly
identify any margin call received and the amount of variation margin that you
have to produce.
Day Closing Futures Price (dollars/ounce)
0 (today) $1,300
1 1,303
2 1,297
3 1,290
4 1,297
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The next two questions are based on the following data:
The euro FX (currency) futures contract that trades in the Merc has the following
features:
Euro FX Contract Highlights
Ticker symbol EC
Trading unit (underlying) Euro 125,000
Quotation US$ per euro
Minimum price fluctuation (tick) $.0001/euro = $12.50/contract
Contract months March, June, September, December
Assume initial margin is $2,511 and maintenance margin is $1,860 per contract
(the exchange periodically revises these numbers).
9.10. Track margin account payments to a trader holding long position in two
contracts when the euro takes the following values (in terms of US dollars):
0.8450, 0.8485, 0.8555, 0.8510, 0.8480, 0.8423, 0.8370, 0.8300, 0.8355.
QUESTIONS AND PROBLEMS 191
9.11. Track the margin account payments to a trader holding short position in one
contract when the euro takes the following values (in terms of US dollars):
0.8450, 0.8485, 0.8555, 0.8510, 0.8480, 0.8423, 0.8370, 0.8300, 0.8355.
9.12. Suppose that the platinum futures price is $1,580 per ounce and the gold futures
price for a contract expiring in the same month as platinum is $1,500 per ounce
(see Example 9.8). Hoping that the spread will narrow to $50 in a month’s time,
set up a spread trading strategy, discuss all possible outcomes for the futures
prices after a month, and illustrate these outcomes in a diagram.
9.13. Explain the difference between cash-settled and physical-settled futures
contracts.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
10
Futures Regulations
10.1 Introduction Enforcement Actions by the CFTC
and the NFA
10.2 Which Markets Have Futures?
Price Discovery 10.4 Manipulation in Futures
Markets
Hedging
The Hunts Silver Case
Speculation
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Salomon Brothers and US
EXTENSION 10.1 When Sellers
Treasury Securities
Have Market Power
Banging the Close
10.3 Regulation of US Futures
Markets 10.5 Managing Commodity
Markets
Major Regulatory Acts
The CFTC, the NFA, and the 10.6 Summary
Regulatory Role of Exchanges 10.7 Cases
10.8 Questions and Problems
WHICH MARKETS HAVE FUTURES? 193
10.1 Introduction
Price discovery is an important function of futures markets. Even so, US citizens
became upset in July 2003 when the Defense Advanced Research Projects Agency
(DARPA), a research wing of the US Department of Defense, announced plans for
a “Policy Analysis Market,” where anonymous individuals could trade futures whose
payoffs would depend on a terrorist indicator. The proposal that the US government
set up an online betting parlor for terrorist activities enraged both legislators and the
public. The project was killed soon after it was unveiled. Despite its ethical problems,
it was hailed as a good idea by some economists, who noted that futures prices have
done an excellent job of predicting conventional as well as nonconventional events.
DARPA argued that futures markets usually fare better than experts in predicting
events like elections, and this project’s mission was to tap the collective wisdom in
market prices—price discovery—to understand the probabilities of terrorist activities.
But this argument from the Department of Defense didn’t quite fly.
This chapter discusses the price discovery role of futures markets, regulations, and
market manipulation. First, we explore the market conditions needed for futures to
trade. Next, we look at futures market regulations. Finally, we discuss futures trading
abuses and market manipulation: colorful stories of powerful players who sought
quick gains by cornering the market and squeezing the shorts until regulators foiled
their ploys.
10.2 Which Markets Have Futures?
When is a futures market for a particular commodity likely to succeed? Three
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
conditions—competitive markets, standardization, and volatile prices—seem to be among
the key ingredients needed.
A competitive market is one in which each investor’s trades are too insignificant
to influence prices and traders act as price takers (see Varian 2003). Economists
view competitive markets as the ideal setting. Markets cease to be competitive when
some participants gain market power and manipulate prices to their benefit. When
markets can be manipulated, traders are reluctant to trade knowing that they can lose
significant wealth to the manipulators. For example, consider the following:
■ When a monopoly (a single seller), an oligopoly (a small band of sellers), or a
cartel fixes commodity prices, futures trading may not develop. This is the case for
diamonds but not for oil. Extension 10.1 tells a tale of the spectacularly successful
diamond cartel.
■ When a sole buyer or a small number of buyers drive the market, futures markets
for the commodity are unlikely. There are no futures markets for missiles and
weapons where the government is the sole buyer.
Standardization refers to the ability of a commodity to be sold in standardized
units. For example, gold and silver are easy to standardize in terms of quantity
194 CHAPTER 10: FUTURES REGULATIONS
and quality. As such, one can easily write contracts on future purchases or sales
of these commodities in large quantities. In contrast, automobiles are not so easily
standardized. There are hundreds of different car models available, distinguished by
manufacturer and other features like manual or automatic, electric, gas, or diesel, and
so on. Consequently, it is no surprise that automobile futures do not exist.
Volatile prices are commodity prices that change randomly across time in an
erratic (up and down) fashion. Volatile prices create risk for producers in determining
both their input costs and output sales. Volatile prices therefore generate a demand to
hedge these price risks and also a need to forecast future prices. Futures markets exist
to fulfill these two needs. Let us explore both these roles filled by futures markets,
price discovery, and hedging in more detail.
Price Discovery
Price discovery is related to the notion of market efficiency (see Chapter 6). In
efficient markets, prices accurately reflect available information. Such information
includes past and current prices (weak-form efficiency), publicly available information
(semi-strong-form efficiency), and private information (strong-form efficiency). This
concept is not only true for spot markets but for futures markets as well. Because
futures market traders include producers who have private knowledge of supply
and demand conditions, futures markets are widely believed to be (semistrong-form)
efficient, thereby generating prices that reflect more information and provide better
price forecasts—price discovery—than any ordinary trader’s information could. This is
the motivation underlying the US Department of Defense’s July 2003 proposal to
set up a futures market in terrorism indicators. Although well intentioned, the plan
was shelved following heavy criticism concerning its ethical considerations. Price
discovery helps futures buyers and sellers of commodities. For example, firms that
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
process and distribute corn can fix the buying and/or selling price by trading futures,
or the retail seller, which may be your neighborhood supermarket chain, can do the
same.
Hedging
Hedging in the New Oxford Dictionary of English is defined as to “protect (one’s
investment or an investor) against loss by making balancing or compensating contracts
or transactions.” If you hedge with futures (for which you pay nothing to start), then
you get protection from price swings. You benefit when the spot adversely moves but
surrender gains when the spot moves in your favor. Hedging with futures is beneficial
to firms for minimizing input and output price risks, thereby smoothing costs and
profits. This is true for both producers (e.g., farmers, manufacturers) and suppliers
(e.g., gold and copper mines). Hedging is the motivation for the existence of oil
futures markets, despite that the oil producers form a pricing cartel known as the
Organization of the Petroleum Exporting Countries (OPEC).
WHICH MARKETS HAVE FUTURES? 195
Speculation
In contrast to hedging, the term speculation carries a negative connotation. The
humorist Mark Twain said, “There are two times in a man’s life when he should not
speculate: when he can’t afford it, and when he can.” But speculation is really the
flip side of hedging. Hedgers are buying insurance against adverse price movements.
Speculators are selling the insurance: gambling that the adverse price movements will
not occur.
For futures markets to work, both hedgers and speculators are usually needed, and
both play a crucial role. The reason is that it is unusual for hedgers on both sides of
the market (suppliers and users) to have exactly offsetting demands. As such, hedging
demands for a commodity are often one-sided. In such circumstances, speculators
provide the liquidity needed for a successful futures market to exist.
In addition to their role in providing market liquidity—being the insurance
sellers—speculators also play a second important role in futures markets: they facilitate
the price discovery role because they often trade based on private information,
obtained by costly investigation. Speculators trading on the basis of their private
information increase market efficiency because as they trade, their information gets
reflected in market prices.
Without speculators, futures markets would be both less liquid and less efficient,
and sometimes, without speculators, futures markets would not exist. An example
is the market for precious diamonds, in which the Diamond Cartel removed the
speculators’ incentives to participate (see Extension 10.1).
EXTENSION 10.1: When Sellers Have Market Power
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Cartels
An oligopoly is a market that is dominated by a handful of sellers, more than one but not too many. For
example, Boeing of the United States, Airbus of Europe, and some fringe suppliers like Ilyushin and Tupolev
(now divisions of Russia’s United Aircraft Corporation) dominate the production of large passenger airplanes.
Members of an oligopoly sometimes join to form a cartel. A cartel is a coalition of producers that monopolizes
a commodity’s production and sale. A cartel tries to maximize joint profits by controlling supply—it allocates
a share of the common market to each member, sets sales quotas, regulates production, and fixes prices. By
eliminating competition, a cartel raises prices for consumers and profits for the producers. In the process, it
may keep inefficient companies in business. Cartels are banned in the United Kingdom and the United States.
However, they are legal in many parts of the world, including highly industrialized nations like France, Germany,
Italy, and Japan.
By controlling the supply of a commodity, and therefore prices, cartels can potentially eliminate the benefits
of both hedging and price discovery in futures markets. In such circumstances, futures markets will not exist.
This is the case with the Diamond Cartel.
196 CHAPTER 10: FUTURES REGULATIONS
De Beers and the Diamond Cartel
Cecil Rhodes founded this enterprise during the 1880s by consolidating several diamond mines in Southern
Africa. Around the same time, he formed the London Diamond Syndicate to control the world’s diamond
trade. This was later transformed by Sir Ernest Oppenheimer into the Central Selling Organization, a group of
marketing organizations that controlled much of the world’s diamond trade until the last decade of the twentieth
century.
A February 2001 Fortune magazine article (“The De Beers Story: A New Cut on an Old Monopoly”) reported
that De Beers entered agreements with most major diamond producers (including the former Soviet Union) and
led a very effective cartel. In the 1990s, about 45 percent of the global diamond production was mined by De Beers
and another 25 percent went through its hands. In 2000, De Beers announced that it would stop manipulating
the supply by buying and storing diamonds. Afterward, De Beers focused on selling diamonds from its own mines
after branding them with a “Forevermark” to guarantee their integrity. These steps strengthened the demand for
diamonds à la De Beers. Despite this announcement, the effectiveness of the diamond cartel did not disappear.
A New York Times article dated August 9, 2008, titled “Talking Business: Diamonds Are Forever in Botswana,”
reported that “today, De Beers has about 40 percent of the diamond market—but it is far more profitable than
under the old regime, when it controlled 80 percent of the market.”
The strong control of the supply of diamonds by the De Beers Group destroyed the price discovery role
of a diamond futures market as documented because diamond prices do not reflect supply and demand but the
activities of De Beers. Although hedging demands certainly exist for firms using diamonds in production, there are
no offsetting hedging demands from the suppliers (the cartel). For a diamond futures market to work, speculators
need to provide the missing liquidity, but the strong control of the market supply by De Beers removed the desire
for speculators to participate in such a market. Consequently, no diamond futures markets exist today.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
10.3 Regulation of US Futures Markets
Futures markets are not for the fainthearted. Wild price swings have been common
since inception, and there have been numerous attempts to manipulate futures prices.
Competitive markets can lead to economic growth as well as efficiency. However,
sometimes they fail to function, creating a need for regulation. As is often the case
with regulations, many of the rules and regulations of the US securities market have
arisen to fix problems evidenced within financial markets during a financial crisis.
Major Regulatory Acts
Regulation of the US futures markets began on a cold winter’s day in February 1859,
when the governor of Illinois signed a legislative act that granted a corporate charter to
the Chicago Board of Trade (CBOT). The charter, among other things, granted the
exchange self-regulatory authority over its members, standardized commodity grades,
REGULATION OF US FUTURES MARKETS 197
and provided for CBOT-appointed grain inspectors whose decisions were binding on
members. This was a critical early step in the evolution of futures contracts.1
Another major step was taken in October 1865, when formal trading rules were
introduced at the CBOT, including margin requirements and delivery procedures.
Three years later, the exchange adopted a rule to deter manipulation by banning
“‘corners’ (defined as ‘making contracts for the purchase of a commodity, and then
taking measures to render it impossible for the seller to fill his contract, for the purpose
of extorting money from him’).”2 In 1877, the CBOT began publishing futures prices
on a regular basis, and in 1883, the first organization was formed to clear CBOT
contracts on a voluntary basis.
US federal regulations with respect to futures markets began in the 1880s when
the first bills “to regulate, ban, or tax futures trading in the U.S.”3 were introduced in
Congress. About two hundred such bills were introduced over the next forty years
(see Table 10.1 for a list of some major US futures regulations). As agricultural futures
dominated the markets in the early days, regulations primarily focused on grain (such
as corn, wheat, oats, and rye) trading, and the regulatory controls were placed in the
hands of the US Department of Agriculture (USDA).
In 1922, the Grain Futures Act was passed. It regulated grain futures trading,
banned “off-contract-market futures” trading, created (an agency of the USDA) the
Grain Futures Administration for administering the act, and established the Grain
Futures Commission (consisting of the secretaries of agriculture and commerce and
the attorney general), which had the broad power of suspending or revoking a contract
market’s designation.
The Commodity Exchange Act of 1936 amended and extended the 1922 act.
It replaced references to “grains” with the term “commodities,” and it expanded the
approved list to include commodities like butter, cotton, and eggs. The Grain Futures
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Commission became the Commodity Exchange Commission and was granted the
authority to establish federal speculative position limits. In 1947, this commission
evolved into the Commodity Exchange Authority (also an agency of the USDA). It
continued to administer the Commodity Exchange Act until the mid-1970s. But all
this changed with the passing of the Commodity Futures Trading Commission
(CFTC) Act of 1974.
This created the CFTC as the independent, ultimate “federal regulatory agency
for futures trading,” taking responsibility for regulating futures trading in 1975. The
CFTC Act also proposed creation of a self-regulatory organization. Subsequently, an
industry body called the National Futures Association (NFA) was established in
1982, which works with the CFTC to help regulate the markets. More legislation
followed in 1978 and 1982. This whole body of legislation is known as the amended
Commodity Exchange Act (CEA).
1
See “History of the CFTC: US Futures Trading and Regulation Before the Creation of the CFTC”
from www.cftc.gov/About/HistoryoftheCFTC/history_precftc.
2
ibid.
3
ibid.
198 CHAPTER 10: FUTURES REGULATIONS
TABLE 10.1: Some Major Milestones in the Federal
Regulation of US Futures Markets
Year Development
1880s The first bills are introduced in Congress to regulate futures markets.
1922 The Grain Futures Act was passed. The Grain Futures
Administration was created as a department of the US Department
of Agriculture.
1936 The Commodity Exchange Act, an amendment of the 1922 Act,
was passed. The Commodity Exchange Administration was created
as the first federal regulator of futures markets.
1974 The Commodity Futures Trading Commission Act was passed. This
created the Commodity Futures Trading Commission as an
independent “federal regulatory agency for futures trading.”
1978, 1982 The Futures Trading Acts were passed.
1982 The National Futures Association, a self-regulatory agency for the
futures industry, was established.
2000 The Commodity Futures Modernization Act of 2000 was passed.
2010 The Dodd–Frank Wall Street Reform and Consumer Protection
Act was passed. This significantly increased regulatory oversight of
the over-the-counter derivatives market.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
A landmark legislation that substantially altered the CEA, the Commodity
Futures Modernization Act of 2000 (CFMA), was passed by Congress and
signed into a law in December 2000. The major changes included the following:
1. Lifting of a long-standing ban by the CFMA, which made it possible to trade security
futures products, which are futures contracts based on single securities and narrowly
based stock indexes.
2. Easing up on regulations for swaps, other over-the-counter (OTC) derivatives, and trading
facilities for these products.
3. Streamlining of regulation and the creation of three types of markets: (1) designated
contract markets (DCMs, which are boards of trade [or exchanges] that
operate under the regulatory oversight of the CFTC), (2) derivative transaction
execution facilities, which have fewer regulations but more restrictions on who
and what trades, and (3) exempt boards of trade with restrictions similar to
(2) but exempt from CFTC regulation, except for antifraud and antimanipulation
provisions.
The CFMA regulations reflected the deregulatory mood prevalent at the time. The
loosely regulated OTC derivative markets seemingly worked well until the financial
REGULATION OF US FUTURES MARKETS 199
crisis of 2007–9, where it was documented that trading in OTC derivatives helped
cause the crisis. In response to this lack of regulation, in 2010, Congress passed the
Dodd–Frank Wall Street Reform and Consumer Protection Act. This massive
849-page act4 was the largest effort to regulate US financial markets since the 1930s
Great Depression. It introduced numerous significant changes to bank and financial
market regulation. Some key features of the Dodd–Frank Act related to derivatives
are as follows:5
■ Increased regulation. Provides the SEC and CFTC authority to regulate OTC
derivatives so that irresponsible practices and excessive risk taking no longer escape
regulatory oversight.
■ Central clearing and exchange trading. To require central clearing and exchange trading
for more derivatives, this act replaced derivatives transaction execution facilities
with a new type of entity, the swap execution facility.
■ Market transparency. This requires data collection by clearinghouses and swap
repositories to improve market transparency and provide regulators the necessary
information for monitoring and responding to systemic risks.
■ Financial safeguards. These add safeguards by ensuring that dealers and major
swap participants have adequate financial resources to guarantee the execution of
derivatives contracts.
The CFTC, the NFA, and the Regulatory Role of
Exchanges
In establishing the CFTC, Congress recognized that futures markets are impor-
tant to economic growth. The CFTC carries out a variety of activities (see
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
https://www.cftc.gov/About/CFTCOrganization/index.htm):
■ The CFTC must approve new contracts and changes to existing contracts. Each new futures
contract is evaluated with respect to three issues: (1) justification of individual terms
and conditions, which thoroughly examines the details of the contract to ensure
that there will be sufficient supply of the underlying commodity to prevent market
manipulation; (2) an economic purpose test, which requires exchanges to show
that the proposed contract will be useful for price discovery and hedging; and (3)
other public interest requirements.
■ The CFTC develops rules and regulations that govern the NFA and all futures exchanges.
These rules and regulations define the requirements for registration, disclosure,
minimum financial standards, daily settlement, separation of customer funds,
supervision and internal controls, and other activities.
4
See www.gpo.gov/fdsys/pkg/PLAW-111publ203/pdf/PLAW-111publ203.pdf.
5
See http://banking.senate.gov/public/_files/070110_Dodd_Frank_Wall_Street_Reform_comprehen-
sive_summary_Final.pdf.
200 CHAPTER 10: FUTURES REGULATIONS
■ The CFTC requires companies and individuals who handle customer funds or give trading
advice to register with them. Actually, the NFA handles the registration process on the
CFTC’s behalf.
■ The CFTC ensures compliance of its rules and regulations as well as the Commodity
Exchange and Dodd–Frank acts. It conducts trade practice surveillance and audits
selected registrants. It investigates and prosecutes alleged violations of CFTC
regulations. Whenever necessary, the CFTC files court cases.
■ The CFTC detects and prevents manipulation, congestion, and price distortions. It conducts
daily market surveillance. It can intervene and take corrective action if it believes
manipulation is present.
■ The CFTC oversees training of brokers and their representatives and other industry
professionals. It can force brokers and representatives to take competency tests. The
CFTC also makes sure that all registrants complete ethics training.
■ The CFTC does research on futures markets and provides technical assistance. It makes
economic analyses for enforcement investigations and gives expert help and
technical aid with case development and trials to US attorney’s offices, other federal
and state regulators, and international authorities.
■ The CFTC helps coordinate global regulatory efforts. It develops policies and regulations
governing foreign and cross-border transactions.
■ The CFTC handles customer complaints against registrants. It hears and decides
enforcement cases. The CFTC offers a reparations procedure for persons who have
reason to believe that they have suffered a loss due to a violation of the Commodity
Exchange Act or CFTC regulations in their dealings with a CFTC registrant.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The SEC regulates securities and the CFTC regulates futures, but who regulates
options? Following an intense turf battle between these two regulators, the agreement
is that the SEC regulates options on stocks, while the CFTC regulates those on
futures. The fight relapses whenever new derivatives are introduced that carry joint
characteristics. The issue of overlapping jurisdiction remains unresolved at this time.
On the next rung of the regulatory ladder are the commodity exchanges and the
National Futures Association. These self-regulatory organizations (SROs) must,
among other things outlined in the CFTC’s regulations, enforce minimum financial
and reporting requirements for their members.
As a registered futures association under the Commodity Exchange Act, the NFA
is an SRO composed of futures commission merchants, commodity pool operators,
commodity trading advisors, introducing brokers, leverage transaction merchants,
commodity exchanges, commercial firms, and others who work in the futures
industry.6 Banks and exchanges may join the NFA, but they are not required to
do so. The NFA is involved in a range of activities that include the following:
6
www.nfa.futures.org.
REGULATION OF US FUTURES MARKETS 201
■ On behalf of the CFTC, the NFA registers all categories of persons and firms
dealing with futures customers.
■ The NFA screens and tests registration applicants and determines their qualifica-
tions and proficiency.
■ The NFA requires futures commission merchants (FCMs) and introducing brokers
to keep sufficient capital and maintain good trading records.
■ The NFA tracks financial conditions, retail sales practices, and the business conduct
of futures professionals, and it ensures compliance with the requirements.
■ The NFA audits, examines, and conducts financial surveillance to enforce compli-
ance by members.
■ The NFA maintains an arbitration program that helps to resolve trade disputes.
Commodity exchanges complement federal regulations with their own rules.
These rules cover many aspects of futures trading such as the clearance of trades,
trade orders and records, position limits, price limits, disciplinary actions, floor
trading practices, and standards of business conduct. Clearinghouses also take part
in regulatory activities.
Enforcement Actions by the CFTC and the NFA
The NFA has comprehensive compliance rules “covering a wide variety of areas such
as advertising, telephone solicitations, risk disclosure, discretionary trading, disclosure
of fees, minimum capital requirements, reporting and proficiency testing.” The NFA
takes disciplinary action against violators, ranging from issuing warning letters to filing
formal complaints that can result in penalties that include “expulsion, suspension for
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
a fixed period, prohibition from future association with any NFA Member, censure,
reprimand and a fine of up to $250,000 per violation.”
The CFTC’s Division of Enforcement investigates and prosecutes alleged viola-
tions of the Commodity Exchange Act and Commission regulations. The CFTC
takes enforcement action against individuals and firms registered with the commis-
sion, those who are engaged in commodity futures and option trading on designated
domestic exchanges, and those who improperly market futures and options contracts.
The CFTC also vigorously investigates and prosecutes Ponzi schemes. Named
after the notorious swindler Charles Ponzi, a Ponzi scheme is a scam that pays early
investors returns from the cash coming in from subsequent investors. A common
technique is to start a commodity pool and lure unwary, unsophisticated investors by
promising them a generous return. Eventually, the whole scheme unravels and leaves
behind a trail of cheated customers. Ponzi schemes have attracted special attention
since 2008, when it was revealed that prominent stockbroker and financial adviser
Bernard Madoff ran a giant Ponzi scheme that bilked investors for tens of billions of
dollars.
202 CHAPTER 10: FUTURES REGULATIONS
10.4 Manipulation in Futures Markets
In 1869, notorious financiers Jay Gould, James Fisk, and their co-conspirators tried
to corner the gold market. As told in “Gold at 160. Gold at 130” (Harper’s Weekly,
October 16, 1869), the conspirators bought up huge amounts of gold and attempted
a market corner. The result was a two-week period of turmoil in the gold market that
brought America’s foreign trade, which used gold as the medium of exchange, to a
virtual standstill. To defuse a looming economic panic and restore normal conditions
in the gold market, President General Ulysses Grant ordered the sale of gold to
buy bonds. On September 24, 1869, minutes after the Treasury secretary’s telegram
instructing a sale of $4 million worth of gold reached the “Gold Room,” the price
of gold, which had earlier ranged between $160 and $162, fell to $133 per ounce.
The day went down in history as Black Friday, and the speculators got a severe jolt.
This was just one incident. As we noted in Chapter 6, market corners and short
squeezes have a long history in financial markets. Market structure, the creation of
shorts as futures get transacted, and supply variability for many of the underlying
commodities make futures markets especially prone to price manipulation and trading
abuses. We now discuss some well-known cases of manipulation in these markets.
The Hunts Silver Case
A Peruvian government firm Minpeco SA sued the Hunt brothers and “co-
conspirators (including Arab sheiks)” for $150 million, charging that they conspired
to corner the silver market. A jury found them guilty, and the Hunt brothers were
given huge fines. A class-action suit was also filed on behalf of seventeen thousand
investors. Eventually, in the late 1980s, two of the Hunt brothers filed for bankruptcy.
The details were as follows (see Duffie [1989], pp. 335–9 and Edwards and Ma [1992],
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
p. 198–9):
■ Oil magnate H. L. Hunt, one of world’s richest persons, had two wives. One of his
wife’s sons (referred to here as the Hunt brothers or the Hunts) had a mindboggling
net worth of over $5 billion in 1980 but ended bankrupt after the Silver Crisis.
■ Several times during the 1970s, the Hunts took large long positions in both silver
spot and futures markets and rolled their hedge forward (i.e., closed out a position
in an expiring futures and then took an identical position in a distant-month futures
contract). They also invested in a number of silver producers.
■ This pattern was repeated in 1979–80 on a grander scale. The Hunt brothers (1)
held large long positions in silver futures (at times reaching nearly 250 million
ounces—an amount equivalent to fifty thousand COMEX contracts [that traded in
the Commodity Exchange] and much more than US domestic silver consumption),
(2) held dominant long positions in many near-month contracts, (3) demanded
delivery when contracts matured, and (4) made supply scarce by holding huge
quantities of silver bullions and coins. Some twenty futures commission merchants
helped the Hunt brothers, and the firms Bache Halsey Stuart Shields and Merrill
Lynch handled over 80 percent of the Hunt brothers’ spot and futures trading.
MANIPULATION IN FUTURES MARKETS 203
■ Silver prices shot up. Historically, silver prices rarely crossed $10 per ounce, but the
spot and futures price of silver rose from about $9 per ounce in July 1979 to $35
per ounce at year’s end, peaking at over $50 per ounce in January 1980. Apparently,
the Hunts had cornered the market.
■ The exchanges consulted the CFTC and took steps to break the squeeze. Margin
requirements were increased, stricter position limits were imposed, and the Hunts
were forced to implement liquidation-only trading. They could not increase
their silver position unless it was for hedging purposes. They also had to exit the
market as contracts expired. By the end of March, silver prices dropped to $11 per
ounce. It was this decline in silver prices that destroyed the Hunts’ wealth.
Salomon Brothers and US Treasury Securities
The US Treasury securities market is one of the world’s largest and most liquid
securities markets. Legal scholars like Easterbrook (1986) and Fischel and Ross
(1991) have opined that it is nearly impossible to corner the supply of Treasuries.
Unfortunately, the impossible occurred as Salomon Brothers Inc., a major player in
the Treasury securities market, admitted deliberate and repeated violations of Treasury
auction rules beginning in 1990.
Here is how the Salomon short squeeze in the May 1991 auction of the two-year
US Treasury note worked:
■ There are three related markets for buying Treasuries: (1) the sealed bid Treasury
auction, (2) the preauction when-issued market, and (3) the postauction resale
market (see Chapter 2).
■ Salomon took considerable pride in its trading prowess and boasted itself to be the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
greatest bond trading firm in the world. To ensure broad participation (read, stop
Salomon’s dominance!) in its auction of Treasury securities, the Treasury enacted
the 35 percent rule for each auction: a single bidder cannot bid more than 35
percent of the offering at any particular yield or be awarded more than 35 percent
of the offering.
■ Many Wall Street firms sell Treasuries to their clients in the when-issued market,
enter the auction with a net short position, and buy securities in the auction to
fulfill their obligations. This is a profitable business because the price usually goes
down slightly from the when-issued price owing to greater supply in the auction—
but not in this auction. Salomon submitted aggressive bids at a yield of 6.81 percent
when the notes were trading on a when-issued basis at approximately 6.83 percent
directly prior to the auction (which means Salomon paid more for the auctioned
securities).
■ Salomon evaded the 35 percent rule by submitting multibillion dollar bids on behalf
of clients and then transferred those securities to its own account at cost. Salomon
ended up controlling 94 percent of the competitively auctioned securities. Having
cornered the market, Salomon squeezed the shorts from the when-issued market
by charging them a premium price for these notes when they came to cover their
short positions in the post auction resale market.
204 CHAPTER 10: FUTURES REGULATIONS
■ Academics have tried to measure how much Salomon stood to gain from this
squeeze. Studies estimate that this ranged between $5 million and $30 million (for
details, see Jegadeesh 1993; Jordan and Jordan 1996).
■ The penalties to Salomon were enormous. Its reputation was badly bruised. Soon
after the cornering allegations surfaced and federal agencies started investigating,
regulators barred Salomon from making a market in various federal and state
government securities, cutting off a major source of the company’s income.
Salomon stock, which was trading in the $50 to $60 range, soon plunged to $30.
Billionaire Warren Buffett, a major shareholder of Salomon, flew in from Nebraska
and took control. Several top executives lost their jobs, and many talented personnel
left. Salomon admitted no wrongdoing other than violating Treasury auction rules
and paid a staggering $290 million in fines. The once-feared Salomon now belongs
to the past. It went through mergers and acquisitions and is now part of Salomon
Smith Barney, which is a subsidiary of Citigroup. Even the name Salomon has
been phased out.
Banging the Close
Technology has opened up new ways of manipulating prices. On July 24, 2008,
the CFTC filed a complaint against Netherlands-based Optiver Holding, two of its
subsidiaries, and three Optiver employees in a US district court, charging that the
defendants had manipulated or attempted to manipulate NYMEX crude oil, gasoline,
and heating oil futures contracts on nineteen separate instances in March 2007 and
succeeded in creating artificial prices in at least five of those cases.7
The defendants traded three NYMEX energy futures contracts: the Light Sweet
Crude Oil futures contract (Crude Oil, also referred to as West Texas Intermediate
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
[WTI]), the New York Harbor Heating Oil futures contract (Heating Oil), and the
New York Harbor Reformulated Gasoline Blendstock futures contract (New York
Harbor Gasoline).
The CFTC alleged that the defendants employed a practice popularly known as
banging (or marking) the close, which involves manipulating the prices by trading
a large position leading up to the close followed by offsetting the position before the
end of trading. The scheme had the following ingredients:
■ It involved Trading at Settlement (TAS) contracts for crude oil, heating oil, and
gasoline contracts. These are special futures contracts in which the counterparties
decide at the time of trading that the contract price will be the day’s settlement
price plus or minus an agreed differential. If you trade a TAS contract, you can
offset it by trading a futures contract on the other side of the market.
■ Computing settlement prices is tricky in many futures markets because of potential
manipulation. The settlement prices for each of these three energy futures contracts
7
Our discussion is based on “Complaint: Optiver US, LLC, et al.,” “CFTC Charges Optiver Holding BV,
Two Subsidiaries, and High-Ranking Employees with Manipulation of NYMEX Crude Oil, Heating
Oil, and Gasoline Futures Contracts” and “Case Background Information re CFTC v. Optiver US, et
al.” (www.cftc.gov/PressRoom/PressReleases/ pr5521-08).
MANAGING COMMODITY MARKETS 205
are determined using the volume-weighted average prices (VWAP) of futures
trades occurring during the closing period (the close) for the contracts, which lasts
from 2:28 to 2:30 PM. The defendants planned to control the VWAP and wrote
about “bullying the market” in their recorded conversations and e-mails.
■ The traders concocted a three-step system to make manipulative profits:
- First, they would accumulate substantial positions in TAS contracts by trading at
different times of the trading day.
- Second, they would enter closing trades for approximately 20 to 30 percent of
their contracts during the pre-close period, which are the dying minutes before
the market enters the close.
- Third, Optiver would sell 70 to 80 percent of its holdings during the close, which
lasts from 2:28:00 to 2:29:59.
In conjunction, these three steps manipulated prices so that Optiver could obtain
profits from its trades.
■ The CFTC alleged that the defendants had forced the futures prices lower on three
instances and higher in two cases and made approximately $1 million in the process.
10.5 Managing Commodity Markets
The economic definition of a speculator is someone who is trading and not hedging.
For example, a trader who is short gold futures but does not hold gold in inventory is
speculating. In contrast, if this trader holds gold in inventory for production purposes,
then she is a hedger. Futures regulations regard speculators as “evil” and impose size
limits on their positions—to preclude the accumulation of market power for the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
purposes of manipulative trading strategies. In contrast, they view hedgers as “saintly”
and impose no such size limit restrictions. Although the economic definition of a
speculator and hedger is unambiguous, in practice, it is not always so easy to tell
speculators from hedgers.
Prior to the 1990s, the CFTC and exchanges classified futures traders like
farmers, manufacturers, and commodities dealers (with legitimate hedging needs)
as commercial traders, and the CFTC allowed them to trade virtually any number
of exchange-traded futures contracts desired. These were the hedgers. By contrast,
they imposed speculative position limits on all other traders, who were viewed as
noncommercial traders. These were the speculators. Obviously, noncommercial
traders could be hedgers using the economic definition, but they were not considered
as such.
However, beginning in the 1990s, two new kinds of traders, commodity index
traders and swap dealers, entered the markets and changed these traditional classifi-
cations. Commodity index traders include institutional players like endowment
funds and pension funds. They diversify their portfolios by adding commodity
exposures through passive long-term investment in commodity indexes. Swap
dealers play a more complex role—on one hand, they compete with the futures
markets by offering customized derivatives to their clients, but on the other hand, they
206 CHAPTER 10: FUTURES REGULATIONS
also trade in the futures markets to hedge risks in their swap books. Typically affiliated
with a bank or other large financial institution, swap dealers operate as market makers
by being ready to act as the counterparty to both commercial and noncommercial
traders.
A change in the regulatory framework took place in 1991, when the CFTC
granted a Goldman Sachs subsidiary J. Aron the same exemption from speculative
position limits as those allowed to commercial traders (see “A Few Speculators
Dominate Vast Market for Oil Trading,” Washington Post, August 21, 2008). J. Aron
was into commodity merchandising and traded swaps as a part of its business. It
planned to sell to a large pension fund a commodity swap based on an index that
included wheat, corn, and soybeans, all of which fell under federal speculative position
limits. To hedge this short commodities exposure, it planned to buy exchange-
traded futures contracts on those commodities. The CFTC classified this as a
bona fide hedge because the swap dealer had demonstrated that the positions were
“economically appropriate to the reduction of risk exposure attendant to the conduct
and management of a commercial enterprise.” Once the door was opened, more
exemptions followed.
10.6 Summary
1. Futures contract trading in the United States is a highly regulated activity.
It is governed by the Commodity Exchange Act. The Commodity Futures
Modernization Act of 2000 made substantial changes to the existing legislation.
It made it possible to trade security futures products, it eased up on regulations of
OTC derivatives, it allowed new types of futures trading facilities and products,
and it facilitated clearing of OTC derivative trades. In 2010, the Dodd-Frank
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Wall Street Reform and Consumer Protection Act was passed to increase OTC
derivative market regulation in response to the financial crisis of 2007–2009.
2. Market regulation in the United States is implemented at the highest level by the
federal regulatory agency, the Commodity Futures Trading Commission, and then
by a self-regulatory organization, the National Futures Association, the exchanges
themselves, and finally, at the grassroots level, by brokers handling customer
accounts.
3. The Commodity Futures Trading Commission monitors the markets, investigates
and prosecutes alleged violations of the Commodity Exchange Act and commis-
sion regulations, and punishes violators with fines and (temporary or permanent)
banishment from the futures industry.
10.7 Cases
Amaranth Advisors: Burning Six Billion in Thirty Days (Richard Ivey School of
Business Foundation Case 908N03-PDF-ENG, Harvard Business Publishing).
The case provides students with (1) a deeper understanding of commodity futures
markets in general and of natural gas markets in particular, (2) an introduction to
QUESTIONS AND PROBLEMS 207
hedge funds and an insight into the largest hedge fund collapse in history, and (3)
an introduction to such concepts as liquidity risk, value-at-risk, spread trades, and
the use of derivatives.
Mylan Lab’s Proposed Merger with King Pharmaceutical (Harvard Business
School Case 209097-PDF-ENG). The case considers how hedge funds and other
investors may use derivatives to separate votes from shares and the legal, moral,
and economic implications of this ability.
Rogue Trader at Daiwa Bank (A): Management Responsibility under Different
Jurisprudential Systems, Practices, and Cultures (University of Hong Kong
Case HKU442-PDF-ENG, Harvard Business Publishing). The case examines the
importance of complying with regulations in the context of a major foreign bank
operating in the United States.
10.8 Questions and Problems
10.1. Give two reasons why futures markets may not be started on a commodity.
Give an example of such a commodity to support your answer.
10.2. What is price discovery with respect to the trading of a futures contract? Give
an example of its use.
10.3. Explain why hedging and speculation are like “two sides of the same coin.”
10.4. What is the CFTC’s mission?
10.5. What is the National Futures Association and how does it differ from the
CFTC?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
10.6. If regulation is bad for business, why does the National Futures Association
put a significant amount of regulation on its members?
10.7. Suppose that you are an economist working for the CFTC and an exchange
has proposed to introduce futures trading on individual stocks. Would you
accept this proposal? Give reasons for your answer.
10.8. Suppose that it is five minutes to the close of trading for the day and you
are a trader on the exchange floor. Your client gives you a sell order for one
thousand contracts. You are holding one hundred contracts long. How can
you take advantage of this situation?
10.9. Summarize what happened in the Hunts silver case.
10.10. Summarize what happened in the Salomon Brothers manipulation.
10.11. Why is the possibility of manipulation bad for the trading of futures contracts?
Give an example of what can go wrong.
10.12. Is selling an insurance contract speculating? Is buying an insurance contract
hedging? Are insurance contracts a benefit to society? Explain why or why
not.
208 CHAPTER 10: FUTURES REGULATIONS
10.13. Explain why speculation in futures trading is analogous to selling insurance
contracts. Does this imply that speculation with futures is beneficial or
harmful to the economy? Explain your answer.
10.14. Do futures contracts need to be regulated? If yes, what can go wrong without
regulation? Give an example to explain your answer.
10.15. Can you identify common features in the manipulation stories involving
forward or futures markets? How can “authorities” intervene and break
squeezes in a typical situation?
10.16. What is banging the close?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
11
The Cost-of-Carry Model
11.1 Introduction The Arbitrage Table Approach
11.2 A Cost-of-Carry Example 11.5 Valuing a Forward Contract
at Intermediate Dates
11.3 The Assumptions
11.6 Linking Forward Prices of
11.4 The Cost-of-Carry Model
Different Maturities
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The Model Setup
11.7 Summary
Using the Law of One Price
Using Nothing Comes from 11.8 Cases
Nothing
11.9 Questions and Problems
Different Methods for Computing
Interest
210 CHAPTER 11: THE COST-OF-CARRY MODEL
11.1 Introduction
To trade a forward contract on a spot commodity and not get ripped off, one needs
to know the fair forward price. But how can one determine this price? We claim that
if you understand the concepts of present and future value introduced in chapter 2,
then you should already be able to guess the answer. Give up? The answer is that the
forward price should be the future value of the spot commodity today! Why? If all’s
right with the world, then this should be today’s price for buying the spot commodity
in the future. This chapter shows that under a reasonable set of assumptions, including
the assumption of no arbitrage studied in chapter 6, this guess is correct.
The technique we use to prove this guess is called the cost-of-carry model,
and it is the simplest and first application of the arbitrage-free pricing methodology
studied in chapter 6. Later chapters will use the same no-arbitrage approach to price
more complex derivatives such as options.
The cost-of-carry model is derived via a cash-and-carry argument whereby the
underlying commodity is purchased with borrowed cash and held until the forward
contract’s maturity date, thereby it is carried into the future. When held, various
costs-of-carry (or carrying charges) such as interest on the borrowed cash and
storage costs for the commodity are paid. At the forward contract’s maturity, the
borrowing is repaid, yielding unfettered ownership of the commodity. The cost
of purchasing the commodity in the future by the cash-and-carry strategy is easily
determined. All of the ingredients are known at the start (the spot price, interest rates,
storage costs). Because this strategy generates the same payoff as the forward contract,
the forward price can be determined from the cost of constructing the strategy today;
otherwise, an arbitrage opportunity exists. All’s right with the world after all!
The reasonable assumptions underlying the cost-of-carry model are frictionless
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
markets, no cash flows to the underlying commodity, no counterparty credit risk,
and competitive markets. Interest rates and commodity prices are random. Financial
assets that satisfy the no cash flow assumption include stocks that pay no dividends
over the forward’s life and zero-coupon Treasury bonds that mature after the delivery
date. The final results of this chapter use the cash-and-carry argument to value a
forward contract at an intermediate date and to establish a link between two different
maturity forward prices on the same commodity. We adopt a data-model-arbitrage
approach that is used throughout the book. This approach starts with a numerical
example (data). Next, we replace numbers with symbols (model). Finally, we generate
arbitrage profits when securities are mispriced (arbitrage).
11.2 A Cost-of-Carry Example
The cost-of-carry model uses the concept of no arbitrage. We develop the basic
cost-of-carry model via either of two equivalent techniques, which we labeled the
law of one price and nothing comes from nothing in chapter 6. These simple techniques
rest on several key assumptions. We discuss the assumptions after illustrating the ideas
involved.
A COST-OF-CARRY EXAMPLE 211
Example 11.1 uses the law of one price. It considers two different ways of buying a
stock in the future: buying a forward contract or buying the stock in the spot market
and carrying it to the future. We create these two positions so that we can extract the
forward price by equating their values today. We replace numbers with symbols to
obtain the basic cost-of-carry model.
EXAMPLE 11.1: Finding the Forward Price Using the Law of One Price
The Data
■ Today is January 1. Suppose that we go to the forward market and buy a one-year forward contract
on Your Beloved Machine Inc. (YBM) stock. Unless noted otherwise, all computations are on a per
stock basis. No cash changes hands today because the forward price F is the fair price. At maturity,
we receive one YBM stock worth S(T) from the forward’s seller by paying him the forward price F
dollars. Record the cash flows (in Table 11.1a) as 0 today (time 0) and S(T) – F on the delivery date
(time T). This is Portfolio A, which we label as the market traded forward.
■ Alternatively, one can buy one share of YBM in the spot market today and carry it to the future.
Assuming that YBM pays no dividends over the forward’s life, the cash flows (in Table 11.1b) are –100
today and S(T) on December 31.
■ To create a future liability equal to the forward price, borrow its present value by shorting zero-
coupon bonds. This cash inflow reduces today’s net payment. If the interest rate is fixed at 6 percent
per year, the cash flows are F/(1 + interest) = F/1.06 today and –F a year later (see Table 11.1b).
These long stock and short bond positions make up portfolio B, which is our synthetic forward.
■ By the law of one price, to prevent arbitrage, portfolios A and B with identical future payoffs must
have the same value today. Consequently,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
−100 + F/1.06 = 0
(11.1)
or F = 100 × 1.06 = $106
The forward price equals the amount repaid on the loan for the stock purchase.
TABLE 11.1A: Portfolio A: Long Market Traded Forward
Portfolio Today (Time 0) Delivery Date (Time T) Cash Flow
Cash Flow
Long forward (forward price F) 0 S(T) – F
Net cash flow 0 S(T) – F
212 CHAPTER 11: THE COST-OF-CARRY MODEL
TABLE 11.1B: Portfolio B: Long Stock and Short Bonds (Long Syn-
thetic Forward)
Now (January 1) Maturity Date (December 31)
Portfolio Cash Flow Cash Flow
Buy YBM in the spot market (stock –100 S(T)
price S = $100)
Short zero-coupon bonds to borrow the F/1.06 –F
present value of the forward price F
(interest rate is 6 percent per year)
Net cash flow –100 + F/1.06 S(T) – F
The Model
■ Use symbols to generalize. Replace 100 with S and 0.06 with R to get our first result (stated later as
Result 11.1):
F = S (1 + R) (11.2a)
or S = BF (11.2b)
where (1 + R) is the future value of $1 invested today (which is also known as the dollar return) and
B ≡ 1/(1 + R) is today’s price of a zero-coupon bond that pays $1 at maturity.
A Graphical Approach
■ One can also develop the cost-of-carry model using payoff diagrams. Recall from chapter 5 that these
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
diagrams plot “Payoffs” (or gross payoffs) along the vertical axis and the “stock price on the delivery
date,” S(T), along the horizontal axis.
■ First, buy one YBM stock. Its payoff is a straight line starting at the origin and rising at a 45∘ angle.
If the stock is worthless on the forward’s delivery date, then you get 0; if it’s worth $5, then you get
$5, and so on (see Figure 11.1). Next, draw another diagram under this depicting the payoff from a
short bond position with a future liability of F dollars. This is a line parallel to the horizontal axis that
begins at –F on the vertical axis.
■ Now, vertically add up the payoffs from these two graphs. The short bond trade lowers the long stock’s
payoff by a parallel downward shift of F dollars. Consequently, the combined payoff starts at –F
along the vertical axis, rises at a 45∘ angle to the horizontal axis at F, and keeps increasing. The
combined position is a synthetic forward because its payoff on the maturity date is identical to that of
a market-traded forward contract. Because the forward price F is chosen so that the traded forward
contract’s value is zero, to prevent arbitrage, the synthetic forward must also have zero value. Setting
the synthetic forward’s cost of construction (S – BF) to zero yields the cost-of-carry model.
Arbitrage Profits
■ What would you do if the forward price differs from the arbitrage-free price? Suppose that an errant
trader quotes a forward price of $110. As this is higher than $106, sell the overpriced forward, buy
A COST-OF-CARRY EXAMPLE 213
the stock, and borrow the present value of $110. Then – 100 + 11/1.06 = $3.77 is your immediate
arbitrage profit—and you have no future cash flow because the short traded forward and the long
synthetic forward’s cash flows exactly offset each other. If the trader quotes $100 instead, reverse the
trades and make –100/1.06 + 100 = $5.66. Any forward price other than the arbitrage-free price of
$106 creates an arbitrage opportunity.
FIGURE 11.1: Cost-of-Carry Model
Gross
payoff
Long stock
0
Long stock +
S(T), stock price
Short bond
on delivery date
Payoff
0
S(T)
Payoff similar −F Short bond
to long forward
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Payoff
0
F S(T)
−F
This example illustrates that there is nothing magical about the forward price.
The forward price is linked to today’s spot price via the cost-of-carry model, which
only includes the interest cost in this simple setting. This simple model raises several
questions:
1. What are the hidden assumptions underlying this argument? The next section discusses
the assumptions underlying this and most other models in this book.
214 CHAPTER 11: THE COST-OF-CARRY MODEL
2. Is the cost-of-carry model useful in other contexts? In its simple form, it generates
the forward prices for investment assets like currencies and stocks. Forward
prices for consumption or production assets like corn, copper, and oil,
which are primarily held for consumption or as inputs to production, may be
obtained by modifying this model to incorporate additional carrying charges
such as transportation and storage costs as well as convenience yield benefits. This
extension is presented in Chapter 12.
The remainder of this chapter generalizes this example to build the cost-of-carry
model. First, we need to introduce the hidden assumptions underlying our arguments.
11.3 The Assumptions
A model is a simplified or idealized description, often expressed mathematically, that
helps us understand and explain reality. Financial models serve this purpose. Being
abstractions, models rely on simplifying assumptions. To understand a model, one
needs to understand the assumptions. First, they show when to use, when not to use,
and how not to abuse a model. Second, they give a sense of when the model fails
and the biases that may result. Understanding the assumptions changes a model from
a black box to an open box.
We need the following assumptions for derivative pricing models in general and
for forward pricing models in particular.
A1. No market frictions. We assume that there are no market frictions like
transaction costs (brokerage commissions and bid/ask spreads), margin requirements,
short sale restrictions, taxes that may be levied at different rates, and assets that are
not perfectly divisible.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The frictionless market assumption helps us create a benchmark model against
which other models can be compared. Once you understand this ideal setting, you
can add back frictions to better capture reality. In addition, it’s a reasonable first
approximation for many institutional traders. For example, professional traders at
major investment banks have far lower transaction costs than do small investors.
Professionals often trade within the bid/ask spread, they have less stringent margin
requirements, they can avoid short sale restrictions, and their trading profits are taxed
at a uniform rate. Furthermore, frictionless markets may now be a reasonable first
approximation for small traders in some markets because of the accumulated advances
in information technology that have drastically reduced trading costs.
A2. No credit risk. Also known as default risk or counterparty risk, this is
the risk that the counterparty to a transaction will fail to perform on an obligation.
We assume that there is no credit risk.
If credit risk exists, cash flows become less predictable, and we stumble at the
onset. No credit risk is a reasonable approximation for exchange-traded derivatives,
whose performance is guaranteed by a clearinghouse. Vigilant exchange officials,
alert brokers and dealers, the customer’s margin account, and the exchange’s capital
minimize this risk. Over-the-counter derivatives, however, may possess substantial
credit risks. In practice, prudent actions, such as trading only with counterparties
THE ASSUMPTIONS 215
that have high credit ratings, having well-documented legal contracts, and requiring
parties to keep collateral, reduce this risk. Pricing derivatives in the presence of credit
risk is an area of active research. Jarrow and Turnbull (1995) developed a popular
model for pricing credit derivatives that is used in industry, known as the reduced
form approach. This extension is discussed in Chapter 26.
A3. Competitive and well-functioning markets. In a competitive market,
traders’ purchases or sales have no impact on the market price; consequently, traders
act as price takers. In a well-functioning market, price bubbles are absent. We assume
competitive and well-functioning markets.
In a competitive market, a trader is individually too small for his trades to have a
quantity impact on the price and to possess market power. Consequently, he acts as a
price taker vis-à-vis the market. The competitive markets assumption is the workhorse
of modern economics, and many powerful results are derived from it. As illustrated
via the futures market manipulations described in chapter 10, this assumption does
not always hold. Pricing securities when traders have market power and when prices
can be manipulated is a difficult exercise. One must take into account bargaining and
strategic interactions—ideas and concepts that are outside the scope of this book.
Consequently, the cost-of-carry model is a poor model in such a setting. Focusing
on competitive markets avoids these complications.
We also assume that there are no asset price bubbles. A price bubble happens
when an asset’s price deviates from its intrinsic or fundamental value. The funda-
mental or intrinsic value can be defined in two equivalent ways. The first is the
present value of its future cash flows. The second is the price one would pay if, after
purchase, you had to hold the asset forever. If you reflect on these, it becomes clear
that they are equivalent definitions. A difference between the market price and the
fundamental value only occurs if one believes that selling (retrading) can generate a
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
higher value than holding forever—this difference is the bubble. The terminology
reflects the popular belief that such deviations are short-lived and surely burst, like a
soap bubble. It has been shown that many Internet stocks exhibited a price bubble
during the 1990s, when they reached astonishingly high values despite an absence of
dividend payments and continued reports of zero or negative earnings. We rule out
price bubbles because our pricing arguments fail to hold in this context.
Pricing derivatives in noncompetitive markets and in the presence of bubbles is a
daunting task and a hot area of research. We return to this topic again in Chapter 19,
when we discuss pricing options.
A4. No intermediate cash flows. As a first pass, we assume that the underlying
commodity has no cash flows over the forward contract’s life. This is a simplification
that we will need to relax.
When considering stocks, most companies reward their shareholders by pay-
ing dividends. US companies tend to be fairly conservative about their dividend
policies. They pay fixed dividends and dislike tinkering with either the payment
date or the amount, lest the market infers something negative from their actions.
In addition, many commodities receive income or cash flows—coupon-paying
bonds or consumption/production assets that provide convenience yields are prime
216 CHAPTER 11: THE COST-OF-CARRY MODEL
examples. Moreover, consumption/production assets tend to incur transaction,
storage, insurance, and other costs. For the present, we ignore these costs.
A5. No arbitrage opportunities. This assumption’s meaning is self-evident.
These five assumptions form the bedrock of our analysis, and they determine
the validity of our models. Talented traders regularly monitor the market and check
whether the assumptions are satisfied. If the assumptions aren’t satisfied, the pricing
models and trading strategies may fail to work. In such a situation, one should not
rely on the model’s implications. If one is using the model to hedge, the hedge may
no longer work. An optimal strategy may be to close out the position, rethink, and
come back another day. As in Kenny Rogers’s song, “you got to know when to hold
’em, know when to fold ’em, know when to walk away, know when to run.”
11.4 The Cost-of-Carry Model
We now assemble the different components that we have developed to formally derive
our forward pricing model. If the notations seem abstract, remember the data from
our example (S is $100, 1 + R = 1/B = $1.06) to get a numerical feeling for the
argument.
The Model Setup
The cost-of-carry forward pricing model has the following structure:
■ There are two dates. At time 0 (which is today), we start the clock, trade, and create
one or more portfolios of securities. The clock stops at some later time T.
■ Three securities trade: a stock, a forward on the stock, and a zero-coupon bond.
Unless specified otherwise, we use the terms asset, commodity, stock, and spot
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
somewhat interchangeably.
■ The stock price is S(0) ≡ S today and S(T) at time T. The forward contract starts
today and a forward price F(0, T) ≡ F is determined in the market such that no cash
changes hands. This is the fair price. To understand why this price is fair, suppose
the contrary. If it were not fair, then one side of the contract would be entering
into the contract with an immediate loss and with no additional compensation.
■ The contract ends on the delivery or maturity date by cash settlement or physical
delivery. The long’s payoff is S(T) – F and the short’s payoff is –[S(T) – F], which
are equal and opposite.
■ Traders can borrow or lend funds at the risk-free rate by shorting or buying zero-
coupon bonds. Let B(0, T) ≡ B be today’s price of a zero-coupon bond that pays
one dollar at time T. Let 1 + R be the dollar return obtained at time T from
investing one dollar in the zero today.
We utilize the law of one price (from Chapter 6) to formally derive the cost-of-
carry model.
THE COST-OF-CARRY MODEL 217
Using the Law of One Price
To find the arbitrage-free forward price F, let’s consider two different ways of buying
the stock in the future without a cash flow today. First, buy a forward contract. This
cashless trade obliges you to acquire the stock by paying the forward price on the
delivery date (this is our Portfolio A). Second, buy the stock today and finance this
purchase by borrowing. This costless transaction gives you the stock and a liability
equal to today’s stock price with its accrued interest on the same maturity date (this
is our Portfolio B). These transactions are summarized as follows:
Portfolio A (buy forward): Zero-cost portfolio (today) → Get the stock, pay the
forward price F (write this as S[T] – F ) (maturity date)
Portfolio B (buy the stock, sell the zero-coupon bonds to finance the stock purchase):
Zero-cost portfolio (today) → Get the stock, repay the loan (write this as S[T]
– S[1 + R]) (maturity date)
Note that these two final payoffs only differ by the constant [F – S(1 + R)] and
that this constant is known today. These two portfolios with identical initial values
(zero) must also have the same value on the delivery date (see Figure 11.1). If one
of these values differs, it opens up an arbitrage opportunity that will be exploited by
vigilant traders standing ready to pick up money lying on the street—they will trade
until the two prices are equal and no arbitrage remains. This implies that the forward
price must equal to today’s stock price repaid with accrued interest, that is, [F = S(1 + R)].
Using Nothing Comes from Nothing
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Instead of using the law of one price, if we subtract portfolio B from portfolio A, we
could use the nothing comes from nothing principle instead. This technique creates
a new portfolio that has a zero value in the future and therefore must also have zero
value today. We show this trading strategy in Table 11.2, which shows the trades and
arbitrage opportunity. As depicted, this portfolio is costless to create and worth [S(1
+ R) − F] on the delivery date. To prevent arbitrage, a portfolio that is worthless
today with a constant value at delivery must also have zero value on the delivery date.
This is the nothing comes from nothing principle.
The trading strategy is called a reverse cash-and-carry because it involves short
selling the asset and making good on this obligation in the future. Alternatively, we
could have done a cash-and-carry where we purchase the spot and carry it to the
future by paying the necessary carrying costs. We skip the details because you can
readily obtain the result by reversing the trades and changing the signs in Table 11.2.
Figure 11.2 graphically shows the reverse cash-and-carry argument.
218 CHAPTER 11: THE COST-OF-CARRY MODEL
TABLE 11.2: Arbitrage Table for Reverse Cash-and-Carry
Portfolio Today (Time 0) Cash Flow Maturity Date (Time T)
Cash Flow
Buy forward 0 S(T) – F
Short sell stock S –S(T)
Invest proceeds from short –S S(1 + R)
selling by buying risk-free
zero-coupon bonds
Net cash flow 0 S(1 + R) – F
FIGURE 11.2: No-Arbitrage Chain for
Cost-of-Carry Model
Now Maturity
(Cash flow) (Cash flow)
Buy forward Payoff S(T) – F
Cost 0 How to get rid of S(T)?
Short sell stock Liability – S(T)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Proceeds (+S) Asset disappears
Lend proceeds
Payoff
Buy zero-coupon
S(1 + R)
bonds worth S
Zero value The constant net payoff [S(1 + R) – F]
(for sure) must be zero
(otherwise an arbitrage opportunity)
Different Methods for Computing Interest
Remember from chapter 2 that we can compute B and R by using various discrete or
continuous compounding methods. So far, we have computed B by using simple
interest, which is the easiest case. We can also use continuous compounding by
writing the zero-coupon bond price B or dollar return 1 + R as
B = e−rT and 1 + R = 1/B = erT
where r is the continuously compounded interest rate.
THE COST-OF-CARRY MODEL 219
We gather all our results together.
RESULT 11.1
Cost-of-Carry Model
Consider a forward contract on a commodity beginning at time 0 (today) and
maturing at time T. Then
F = S (1 + R) (11.2a)
Equivalently,
S = BF (11.2b)
where F ≡ F(0, T) is today’s forward price for delivery at time T, S ≡ S(0) is
today’s spot price of the underlying commodity, 1 + R is the dollar return at
time T from investing $1 in a zero-coupon bond today, and B ≡ 1/(1 + R) is
today’s price for a zero-coupon bond that pays $1 at time T.
Dollar returns are computed by the formulas
Simple interest ∶ 1 + R = (1 + i × T) (11.2c)
Continuously compounded interest ∶ 1 + R = erT (11.2d)
where i is the simple interest rate per year and r is the continuously
compounded interest rate per year.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Continuously compounded interest rates give a better approximation than do
simple interest rates when compounding is computed daily. Moreover, we will show
in the next chapter that this assumption lays the foundation for developing a variety
of cost-of-carry models, including those written on indexes and commodities that
incur storage costs and have convenience yields. Example 11.2 uses this formula to
compute forward prices and capture arbitrage profits.
EXAMPLE 11.2: Finding Forward Prices and Capturing Arbitrage Profits
The Forward Price
■ Consider the data from Example 11.1. The YBM stock price S is $100 today, a newly written forward
on YBM matures in one year, and the simple interest rate is 6 percent per year. Then Equation 11.2c
gives
Dollar return 1 + R = (1 + 0.06 × 1) = $1.06 and the
zero-coupon bond price B = 1/1.06 = $0.9434
220 CHAPTER 11: THE COST-OF-CARRY MODEL
By expression (11.2a), the forward price is
F = S (1 + R) = $106.00
which is the same price as in Example 11.1.1
Capturing Arbitrage Profits
■ Suppose that an errant trader quotes a forward price of $110. Then the spot and forward prices are
too far apart. The spot is $100, while the discounted forward price BF = $103.77 so that S – BF =
– $3.77.
■ One can profit by doing a reverse cash-and-carry arbitrage. Simultaneously buy the relatively
underpriced stock, sell the relatively overpriced forward, and short zero-coupon bonds with a maturity
value of $110 to balance the portfolio. Immediately collect $3.77. The net payoff will be zero for sure
on the maturity date.
■ Want to take your profits at maturity? Modify the bond position—short only enough zero-coupon
bonds to finance the spot purchase, and make F – S(1 + R) = $4 on December 31.
■ Now, consider a different scenario—use the same data as before, except let the forward price be $100.
Here S – BF = $5.66, suggesting that either the stock price is too high or the forward price is too low
(or both). Do a reverse cash-and-carry arbitrage by selling the stock, buying the forward, and buying
zero-coupon bonds. As before, the bond position can be tinkered with to take the profit now or later.
If you buy zero-coupon bonds with a face value of $100, then you will make –BF + S = $5.66 of
arbitrage profits today.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1 If this is continuously compounded interest, r is 6 percent, and then by (11.2a) and (11.2d), the forward price is F = S(1 + R) =
100 exp(0.06 × 1) = $106.18.
The Arbitrage Table Approach
The previous examples motivate an arbitrage table approach for solving problems and
proving results. By systematically recording cash flows at different dates, an arbitrage
table enables direct comparison, which makes proofs easier. Although there are no
set rules, we suggest three possible methods:
1. Understand the problem, using numbers if necessary. What do the different variables
mean? Which security generates what cash flows? When? Draw a timeline. Plug
dates and variable values into the formula to get a feel for the result to prove or
how to make arbitrage profits. Use your intuition to sense which variables are
overvalued or undervalued, using our “buy low, sell high” dictum.
2. Interpret cash flows. If the cash flows are positive today and zero on the delivery
date, then you have made arbitrage profits. Alternatively, if cash flows are zero
today and have a positive value on the delivery date, then you also have made
arbitrage profits.
VALUING A FORWARD CONTRACT AT INTERMEDIATE DATES 221
3. Gather variables to one side of the equality and set the net cash flows at one of the dates to
zero. It is inconvenient to work with nonzero cash flows at both the starting and
ending dates. Life is easier if you gather variables to one side of an equality so that
you can set today’s or the delivery date’s net cash flows to zero. Then, to prevent
arbitrage, the other net cash flow must also be zero.
11.5 Valuing a Forward Contract at
Intermediate Dates
Suppose you bought a forward contract some time back, but it still has remaining life.
How much it is worth today? If the spot price soars or plunges, how would it affect
the forward’s value?
We use a cash-and-carry argument to develop a formula that answers these
questions. The model may be used for mark-to-model accounting, which requires
the valuation of assets and liabilities on a daily basis, a practice that is increasingly
becoming popular with companies and regulators. Example 11.3 introduces a
numerical example that is then generalized to obtain the relevant result.
EXAMPLE 11.3: Valuing a Forward Contract at an Intermediate Date
The Data
■ Suppose that you bought a forward on January 1. Three months have passed. Meanwhile, the
underlying price has increased. Of course, your long position has gained value because you entered
into a contract to buy the underlying at a fixed price, but the underlying has become more valuable.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
What is the exact value of your long position today? Let us consider an example that uses some of our
previous data:
- On January 1 (time 0), YBM’s stock price S(0) was $100.
- The simple interest rate i was 6 percent per year.
- You purchased a newly written forward contract on YBM that matures in a year on December 31
(time T).
- The forward price was set at F(0) = S(1 + R) = $106.
■ Three months have passed; today is April 1 (time t). YBM has rallied and the new stock price S(t) ≡
S is $120. You are still obliged to buy YBM by paying F(0) in nine more months (T – t is 0.75 years).
But what about the value of your forward, V (t) ≡ V ? If you sell your position, how much should you
get paid?
■ Draw a timeline to keep track of the cash flows (see Figure 11.3). The stock price increase has made
the long forward position more valuable, the exact amount of which is determined by the arbitrage
table (see Table 11.3).
■ Begin the portfolio construction as if you are buying the forward position by paying V. This has a
payoff of S(T) – F(0) = S(T) – 106 on December 31. Next, short sell the stock. This gives $120 today
but creates a liability –S(T) on the delivery date. The stock disappears, but you still have to come up
222 CHAPTER 11: THE COST-OF-CARRY MODEL
with $106. This can be arranged by lending the present value of $106 via buying zero-coupon bonds.
Notice that on April 1, the price of a zero-coupon bond that pays $1 on December 31 is
B ≡ B (t, T) = 1/ (1 + 0.06 × 0.75) = $0.9569
■ The portfolio has a zero value on December 31. To prevent arbitrage,
−V + 120 − 0.9569 × 106 = 0 (11.3)
or V = $18.561
The Model
■ Replace numbers with symbols in Equation 11.3 to get
−V + S − BF (0) = 0
or V = S − BF (0)
1 Instead, if we assume a continuously compounded interest rate r of 6 percent per year fixed, then F(0) = S(1 + R) = 100 ×
exp(0.06 ×exp(0.06 × 1) = $106.18, B(t, T) = exp(– 0.06 × 0.75) = $0.9560, and V = 120 – 0.9560 × 106.18 = $18.49. There
is a 7 cent difference in value of a forward contract because of the different ways of computing interest.
This is Result 11.2, which is formally stated as follows.
RESULT 11.2
Valuing a Forward Contract at an Intermediate Date
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Consider a forward contract created at time 0 and that matures at time T. On
an intermediate date t (today),
V = S − BF (0) (11.4a)
where F(0) is the forward price at time 0 for a contract maturing at time T,
V ≡ V (t) is today’s value of a long forward position, S ≡ S(t) is today’s spot
price of the underlying commodity, and B ≡ B(t, T) is today’s (time t) price
of a zero-coupon bond that pays one dollar at time T.
In the case of simple interest, B is 1/[1 + i(T – t)], where i is the
interest rate per year. In the case of continuously compounded interest, B is
e–r(T − t) where r is the interest rate per year.
This is a more general version of our first result. If we set V = 0, as happens
on a forward contract’s starting date, then Result 11.1 becomes a special case
of Result 11.2. Using Results 11.1 and 11.2 together, we can obtain another insight.
First,
VALUING A FORWARD CONTRACT AT INTERMEDIATE DATES 223
TABLE 11.3: Valuing a Forward Contract That Began
Earlier
Portfolio Today, April 1 (Time t) Delivery Date,
Cash Flow December 31
(Time T) Cash Flow
Buy forward (worth V (t) –V S(T) – 106
or V to long), forward
price F(0) = $106
Short sell spot (S = 120) 120 –S(T)
Lend present value of –(1/1.045)106 106
forward price BF(0)
Net cash flow –V + 120 – 0.9569 × 106 0
FIGURE 11.3: Timeline for a Forward Contract That
Began Earlier
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Time 0 Time t (today) Time T
B = 0.9569 $1
Forward started Today (April 1). Forward matures
on January 1. one year from
starting date.
Forward price Value of forward contract to long is V(t). Value of forward to
F(0) negotiated. long is S(T) – F(0).
Value of forward
to long is 0.
224 CHAPTER 11: THE COST-OF-CARRY MODEL
we note that Result 11.1, when applied to the time t forward price, implies that
S = BF(t). Second, substitution of this fact into expression (11.4a) and some simple
algebra yields an alternative expression for the value of the forward contract at time t:
V = B [F (t) − F (0)] (11.4b)
This expression shows that the value of the forward contract increases by the present
value of the change in the new delivery price, F(t), relative to the original delivery
price, F(0). The present value is needed to reflect the fact that the payments are not
made until the delivery date, time T.
11.6 Linking Forward Prices of Different
Maturities
Consider a market in which the cost-of-carry model determines forward prices with
high accuracy. If we change the delivery date, the model gives us a series of forward
prices corresponding to contracts maturing on those different dates. We can also use
the cash-and-carry argument to develop a formula linking these prices. This enables
us to find one forward price when only the carrying cost and another forward price
are known. When several forward contracts exist on the same commodity, this model
also enables us to determine whether all the forward prices are correct or some of
them are out of line. The model is given as our Result 11.3. A simple proof follows
the result.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
RESULT 11.3
Link between Near- and Distant-Maturity Forward Prices
Consider two forward contracts that begin at time 0 (today), mature on the
near-maturity date (time n) and the distant-maturity date (time d), respectively.
Then, standing at time n,
B (0, d)
F (0, n) = × F (0, d) (11.5)
B (0, n)
where F(0, n) and F(0, d) are the forward prices negotiated at time 0 for
forward contracts maturing at times n and d, respectively, and B(0, t) is the
price of a zero-coupon bond at time 0 that pays $1 at times t = n and t = d.
The zero-coupon bond price B is 1/(1 + it) for simple interest and e–rt for
continuously compounded interest, where i and r are the respective annual
interest rates.
SUMMARY 225
Although we view this result as an application, it’s also a generalization. If you set
the near-maturity forward to be at its delivery date n, then by convergence of the
basis, F(0, n) becomes the spot price S(n), B(n, n) = 1, and expression (11.5) gives
back Result 11.1 at time n.
To prove Result 11.3, we just use Result 11.1 twice, once for time n and once for
time d, getting the two equations
S (0) = B (0, d) × F (0, d)
S (0) = B (0, n) × F (0, n)
Setting these two equations equal and solving for F(0, n) gives the result.
11.7 Summary
1. We use no arbitrage via the law of one price and the principle that nothing comes
from nothing introduced in chapter 6 to develop a forward pricing model. First, we
create a securities portfolio to replicate a market-traded forward’s payoff. To avoid
arbitrage, the cost of the replicating portfolio must equal to the forward contract’s
arbitrage-free value. Equivalently, an alternative portfolio going long the forward
contract and selling short the replicating portfolio (or vice versa) will have a zero
value both now and in the future. Both approaches give the same price.
2. A portfolio consisting of a long stock, a short forward, and short some zero-coupon
bonds can be created so that it has zero value on the forward’s maturity date. To
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
prevent arbitrage, it must also have a zero value on the starting date. The forward
price is determined from this condition. This gives Result 11.1, which is the basic
cost-of-carry model. It states that the forward price is the future value of the spot
price or F = S(1 + R).
3. The cost-of-carry model makes a number of assumptions: no market frictions,
no credit risk, competitive and well-functioning markets, no dividends over the
contract’s life, and no arbitrage opportunities.
4. There are several ways of developing cost-of-carry models and solving arbitrage
problems. We primarily use arbitrage tables, which systematically record cash flows
at different dates.
5. We use the cash-and-carry argument to build more complex models. Result 11.2
states that the value of a forward contract at an intermediate date equals the spot
price minus the discounted value of the old forward price or V = S – BF(0).
Result 11.3 states that when two different forward contracts are written on the
same commodity but with different maturity dates, the near-maturity forward
price is a fraction of the distant-maturity forward price or F(0, n) = [B(0, d)/
B(0, n)] × F(0, d).
226 CHAPTER 11: THE COST-OF-CARRY MODEL
11.8 Cases
American Barrick Resources Corp.: Managing Gold Price Risk (Harvard Busi-
ness School Case 293128-PDF-ENG). The case discusses derivative usage for
hedging by a gold mining company.
Banque Paribas: Paribas Derives Garantis (Harvard Business School Case
295008-PDF-ENG). The case explores issues connected with a broker-dealer
setting up a derivatives subsidiary to achieve a credit rating.
Hedging Currency Risks at AIFS (Harvard Business School Case 205026-PDF-
ENG). In this case a company considers managing foreign currency risks with
derivatives.
11.9 Questions and Problems
11.1. Today’s price of gold in the spot market is $1,500 per ounce. The price of a
zero-coupon bond maturing in one year is $0.95. What will be the one-year
forward price for gold?
11.2. The current price of Your Beloved Machine’s stock is $109. The continuously
compounded interest rate is 5.25 percent per year. What will be the five-
month forward price for YBM stock?
11.3. The spot price of silver is $30 per ounce. The simple interest rate is 6 percent
per year. The quoted six-month forward price for silver is $31.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
a. What should be the arbitrage-free forward price for silver for a forward
contract maturing in six months?
b. Demonstrate how you can make arbitrage profits in this market.
11.4. Today’s spot price of silver is $30 per ounce. The simple interest rate is
6 percent per year. The quoted six-month forward price for silver is $32.
Transaction costs are $0.10 per ounce whenever spot silver is traded and a
$0.25 per ounce one-time fee for trading forward contracts but no charges
for trading bonds.
a. Demonstrate how you can make arbitrage profits in this market if there are
no transaction costs.
b. If you have to pay transaction costs, demonstrate how you will make
arbitrage profits or explain why you cannot make any such profits.
11.5. Suppose that the continuously compounded interest rate is 6 percent per year,
and the nine-month forward price for platinum is $1,750. What is today’s spot
price of platinum?
11.6. What does the assumption of no market frictions mean? Is this assumption
true in current commodity markets?
QUESTIONS AND PROBLEMS 227
11.7. If the assumption of no market frictions is not true, then why should we study
models using this assumption?
11.8. What is the competitive market assumption, and what adverse market
behavior does it exclude?
11.9. What are the five assumptions underlying the cost-of-carry model for pricing
forward contracts? Which of these assumptions are most likely to be satisfied
in current commodity markets?
11.10. State the key result of the cost-of-carry model in your own words.
11.11. Today is January 1. Forward prices for gold forward maturing on April 1 is
$1,500 per ounce. The simple interest rate is 6 percent per year. What would
be the forward price for a forward contract on gold maturing on August 1?
11.12. Today is January 1. Forward prices for contracts maturing on April 1 and
on October 1 are $103 and $109, respectively. The simple interest rate is
8 percent per year. Assuming the spot price is $100 today, demonstrate two
ways in which you can make arbitrage profits from these prices.
11.13. Today is January 1. Forward prices for contracts maturing on April 1 and
on October 1 are $103 and $109, respectively. On April 1, the price of
a zero-coupon bond maturing on October 1 is $0.97. Assuming that the
underlying interest rate is a constant interest rate, demonstrate one way of
making arbitrage profits from these prices.
11.14. State the equation for the valuation of a forward contract in your own words.
11.15. Suppose you bought a forward on January 1 that matures a year later. The
forward price was $214 at that time, and the simple interest rate was 7 percent
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
per year. Six months have passed, and the spot price is now $150. What is the
value of your forward contract today?
11.16. Suppose you trade a forward contract today that matures after one year. The
forward price is $105, and the simple interest rate is 7 percent per year. If, after
six months from today, the spot price is going to be $150 and the value of the
forward contract is $20, demonstrate how you can make arbitrage profit from
these prices.
11.17. The value of a forward contract that you have been holding for the last six
months is $50 today. It matures in 3 more months. If today’s spot price is $100
and the underlying interest rate is 5 percent, what was the forward price that
you had negotiated when you purchased the contract six months back?
11.18. Explain the difference between the forward price and the value of a forward
contract. How are they related?
11.19. What is the relation between forward prices of different maturities on the
same underlying?
11.20. (Excel) For gold and silver, collect from the internet today’s spot prices as well
as futures prices for futures contracts maturing up to one year. Also collect
Treasury bill prices of different maturities. Using the above data, perform the
228 CHAPTER 11: THE COST-OF-CARRY MODEL
following for three of the most actively traded futures contracts for gold as
well as for silver:
a. Assuming that interest is the only cost of carry, use Result 11.1 (Cost-of-
Carry Model) of Chapter 11 to compute forward prices that correspond
to the maturities of three of the most actively traded contracts for the two
metals.
b. Compare these prices to the traded futures prices. Do they seem signifi-
cantly different? If so, explain why.
Hints and Suggestions:
■ New York spot price for gold and silver can be found on a metal dealer’s
(such as Kitco’s) website.
■ Futures prices can be obtained from an exchange’s (such as the CME
Group’s) website www.cmegroup.com/market-data/index.html
■ Daily Treasury bill rates data can be found from the US Treasury’s
website www.treasury.gov/resource-center/data-chart-center/interest-
rates/Pages/TextView.aspx?data=billrates
■ Use settlement prices wherever possible. If ask and bid prices are quoted
for the spot or futures, take the average of the two.
■ Exclude futures contracts maturing in the current month as they can
have liquidity problems.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
12
The Extended
Cost-of-Carry Model
12.1 Introduction 12.4 Extended Cost-of-Carry
Models
12.2 A Family of Forward Pricing
Models 12.5 Backwardation, Contango,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Normal Backwardation, and
12.3 Forwards on Dividend-Paying
Normal Contango
Stocks
The Cost-of-Carry Model with 12.6 Market Imperfections
Dollar Dividends 12.7 Summary
The Cost-of-Carry Model with a
Dividend Yield 12.8 Cases
A Synthetic Index 12.9 Questions and Problems
The Foreign Currency Forward
Price
230 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
12.1 Introduction
Cornell University is situated in rural upstate New York. The university campus is
built on a hill overlooking beautiful Lake Cayuga and the city of Ithaca. The landscape
is scenic, dotted with hills, valleys, and even a few gorges formed during the ice age.
The winters are cold, snowy, and icy, which makes driving the roads treacherous.
When choosing a car to own in Ithaca, a resident has many options to consider.
Key among these options is the choice of rear-wheel, front-wheel, or four-wheel
drive. To safely navigate the wintry roads, four-wheel drive is the preferred selection.
Consequently, four-wheel drive cars are in great demand.
Analogous to an Ithacan deciding which car to buy in hilly Ithaca, when con-
structing a forward pricing model, one needs to consider the market setting carefully.
Commodities differ in whether they have cash flows, storage costs, or convenience
yields. Choosing the correct model for proper pricing and hedging requires matching
the model’s assumptions with market conditions. To do so otherwise leads to
“treacherous driving” when using the model in practice.
Chapter 11 presented the arbitrage-free forward pricing model under the
assumption that the underlying commodity has no cash flows, storage costs, or
convenience yields over the forward contract’s life. However, stocks and stock
indexes pay dividends, bonds make coupon payments, foreign exchange prices are
affected by both domestic and foreign interest rates, and physical commodities like
corn or gold incur storage costs but may receive a convenience yield. The model
in Chapter 11 also assumed frictionless markets and equal borrowing and lending
rates, but trading involves brokerage commissions as well as the bid/ask spread, and
borrowing and lending rates differ. How does one modify the forward pricing model
when these assumptions are relaxed to get a better approximation? The purpose of
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
this chapter is to answer this question. We do this by relaxing assumptions A1 and A4
of Chapter 11 and thereby incorporate dividends, storage costs, convenience yields,
and market imperfections into our framework.
Although this additional complexity appears daunting, it isn’t. The basic logic of
the cost-of-carry model still applies, but with bells and whistles attached. Most of
these extensions can be easily handled with only minor modifications to the basic
model. By the end of this chapter, one should be able to choose the safest “vehicle”
for forward pricing in realistic market “landscapes.”
12.2 A Family of Forward Pricing Models
The last chapter’s cost-of-carry model includes interest as the only cost-of-carry and
invokes a cash-and-carry argument to express the forward price as the future value of
the spot price (Result 11.1). As in Chapter 11, unless specified otherwise, we use the
terms asset, commodity, stock, and spot interchangeably. Spot trading and a cash market
transaction also have the same meaning. The cost-of-carry model relied on several
critical assumptions that we repeat here for convenience:
A1. No market frictions
A2. No credit risk
A FAMILY OF FORWARD PRICING MODELS 231
A3. Competitive and well-functioning markets
A4. No intermediate cash flows
A5. No arbitrage opportunities
This chapter studies the determination of forward prices in a variety of settings.
Figure 12.1 provides a road map for the family of forward pricing models considered.
We first generalize the model by relaxing the fourth assumption with the introduction
of dividends. This gives the first three extended models of this chapter:
■ The simplest dividend model assumes that the underlying stock pays fixed dollar
dividends on known dates (Result 12.1a). This model is useful for finding forward
prices for contracts written on equities and coupon-bearing Treasury bonds that
mature after the delivery date.
■ The second dividend model assumes that the underlying stock pays continuous
dividends (Result 12.1b). It’s helpful for finding forward prices when the under-
lying is a stock index.
■ An important application of the second model is to determine forward prices for
foreign currency forward contracts (Result 12.2).
FIGURE 12.1: A Family of Forward Pricing Models
Cost-of-carry model (the basic model)
Link between
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
forward prices of
different maturities
(an application)
Valuing a forward at
an intermediate date
With dividends With storage costs and
convenience yields
Currency forward Generalized
(an application) cost-of-carry
Backwardation Contango
Forward Price < Spot Forward Price > Spot
Price Price
232 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
The dividend model paves the way for richer models. One such extension is
storage costs for holding a commodity mainly used for consumption or production
like copper, corn, or crude oil. Storing the physical commodity may also provide
a convenience yield like the ability to keep a production process running in times of
temporary shortages. We include these costs and benefits to build further extensions
of the model (Results 12.3 and 12.4). Last, we introduce market frictions such as bro-
kerage costs, differential borrowing and lending rates, and limited access to proceeds
from short sales. Here buys and sells are no longer symmetric. Well-defined prices
give way to ranges within which arbitrage-free forward prices must lie (Result 12.5).
12.3 Forwards on Dividend-Paying Stocks
Most mature US companies pay their shareholders regular dividends (see Chapter 3
for a discussion of dividends). Companies dislike tinkering with the payment date or
the dividend amount to minimize the possibility that markets will ascribe negative
reasons for these actions and penalize the stock price. Firms like to wear the face of
an English monarch from which you can read nothing. Consequently, fixed payment
dates and fixed dollar dividends are reasonable assumptions for mature companies.
Many assets, such as Treasury notes and bonds, have highly predictable cash flows
and may be similarly modeled.
The Cost-of-Carry Model with Dollar Dividends
Consider a company that pays a fixed dollar dividend on a future date decided in
advance. A dividend payment actually involves several dates including the decla-
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
ration date and the report date. For our purposes, we only need to focus on the
ex-dividend date, which is the cut-off for buying the stock with the dividend. Recall
our discussion from Chapter 3.
■ Buying the stock before it goes ex-dividend gives the share plus the dividend;
purchasing after this deadline gives the share without the dividend.
■ On the ex-dividend date, the cum-dividend stock price equals the ex-dividend
stock price plus the dividend (see Result 3.1).
Dividends create an extra cash flow from holding the stock. If one adjusts the stock
price for this cash flow, the standard argument follows. Example 12.1 shows how to
do this. The resulting pricing model is identical to our cost-of-carry model, except
that the stock price net the present value of the dividend replaces the stock price in
the basic model.
FORWARDS ON DIVIDEND-PAYING STOCKS 233
EXAMPLE 12.1: The Forward Price when the Underlying Pays a Fixed Dollar
Dividend
The Data
■ Consider the data from Example 11.1. Your Beloved Machine Inc.’s (YBM) stock price S is $100
today (January 1, time 0), and the simple interest rate is 6 percent per year. Consider a newly written
forward contract on YBM that matures in one year (December 31, time T).
■ Now, assume that the stock pays a dividend div of $1 in three months (April 1, time t1 ). Figure 12.2
gives the timeline for these cash flows. Table 12.1 accommodates this dividend by adding an extra
column to our arbitrage table.
■ To determine the forward price, we use the nothing comes from nothing arbitrage principle. We
create a portfolio that has zero net payoffs on all future dates. To handle the dividend, we need to
borrow cash to generate a future liability just equal to the dividend payoff. Here are the details:
- Buy one share of YBM. This gives a dividend div = $1 after three months and a stock worth S(T)
after one year.
- Zap the dangling dividend by borrowing its present value. As B1 = 1/(1 + 0.06 × 0.25) = $0.9852
is the price of a zero-coupon bond maturing in three months, record the cash flows as B1 div =
$0.9852 today and –$1 on the ex-dividend date.
- Sell YBM forward to eliminate it from the portfolio on the maturity date.
- Finally, get rid of the forward price F by borrowing its present value. Jot down the cash flows as BF
= (1/1.06)F today and –F after one year.
■ To prevent arbitrage, today’s net cash flow must also be zero:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
−100 + 0.9852 × 1 + 0.9434F = 0
(12.1)
or, F = $104.96
The Model and Arbitrage Profits
■ Replace numbers with symbols in (12.1) and rearrange terms to get
S − B1 div = BF (12.2)
■ Suppose a dealer forgets to adjust for the dividend and quotes a forward price of $106. To obtain
arbitrage profits, create the portfolio given in Table 12.1. This has a short position in the overpriced
dealer-quoted forward and a long position in the synthetic forward. Then
−S + B1 div + BF
= −100 + 0.9852 + (106/1.06)
= $0.9852
The present value of the dividend is the immediate arbitrage profit.
■ If the quoted forward price is less than the arbitrage-free price, reverse the trades to capture the
arbitrage profit.
234 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
Formalizing this example gives our first result.
RESULT 12.1A
Cost-of-Carry Model with Dividends
Consider a forward contract written today on a stock that pays a fixed dollar
dividend div on a known future date. Then
S − B1 div = BF (12.2)
where today (time 0) ≤ ex-dividend date (time t1 ) ≤ maturity date (time T),
S is today’s spot price, F is today’s forward price for time T, and B1 ≡ B(0,
t1 ) and B ≡ B(0, T) are today’s prices of zero-coupon bonds maturing on the
ex-dividend and the delivery dates, respectively.
The zero-coupon bond price B is 1/(1 + iT) in the case of simple interest
and e−rT for continuously compounded interest, where i and r are the annual
interest rates. B1 is computed by replacing T with t1 in these expressions.
FIGURE 12.2: Timeline of Cash Flows for a Dividend
Paying Stock
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
January 1 April 1 December 31
Time 0 (Today) Time t1 (Three months later) Time T (One year later)
B1 = 1/1.015 1
B = 1/1.06 1
Forward starts. YBM pays dividend div of $1 to the Forward matures.
stockholder. Long’s payoff is
S(T) – F.
FORWARDS ON DIVIDEND-PAYING STOCKS 235
TABLE 12.1: Arbitrage Table for Finding the Forward Price of a Dividend-
Paying Stock
Portfolio Today, January 1 Dividend Date, April 1 Delivery Date, December 31
(Time 0) Cash Flow (Time t1 ) Cash Flow (Time T) Cash Flow
Buy stock, get –100 1 S(T)
dividend div after
three months
Borrow present (1/1.015) × 1 –1
value of the
dividend B1 div
today
Sell forward 0 –[S(T) – F]
Borrow present (1/1.06)F –F
value of forward
price BF
Net cash flow –100 + 0.9852 × 1 + 0 0
0.9434F
We can easily extend this to accommodate more dividends. Suppose that the stock
pays another dividend div2 at time t2 before the forward matures. Then, the result
gets modified to
S − (B1 div1 + B2 div2 ) = BF (12.3)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where div1 has a subscript to link it with the payment date and B2 is today’s price of
a zero maturing on the second ex-dividend date. In its most general form,
S − (Present value of all dividends over the forward’s life) = BF (12.4)
Why adjust for dividends? We do these adjustments because dividends differently
impact the stock and the forward: the long stock earns dividends but the long forward
doesn’t. Once you take out the dividends accruing to the stock, it puts both positions
on the same footing, and our familiar cost-of-carry argument applies. Next, we intro-
duce forwards on indexes. For these contracts, the dividend adjustment is more
complex.
The Cost-of-Carry Model with a Dividend Yield
Modern portfolio theory (MPT; say it quickly and it comes out as “empty!”)
pioneered the idea of investing in diversified portfolios that replicate the returns on
stock market indexes. An index mutual fund (or a stock index fund) pursues
this objective by holding a basket of securities in the same proportion as the index.
The fund’s share price is usually determined at day’s end. An exchange traded fund
(ETF) strives to do the same, but it trades continuously on an exchange.1 Indexes
and index funds underlie a large number of derivative securities, including forward
contracts.
236 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
What is the arbitrage-free forward price of such an indexed portfolio? Unfor-
tunately, the dividend adjustment method outlined in the previous section, which
assumes that the dividends and their payment dates are nonrandom, works poorly
for an index because there are too many dividends. This section, therefore, develops
an alternative approach that accommodates random dividends and random payment
dates. This modification uses dividend yields and the continuously compounded
version of the cost-of-carry model (see Chapters 3 and 11, respectively). To apply
this modification, we first need to understand how to adjust an index’s value for
continuously paid dividends. And this, in turn, requires a discussion of synthetic
index construction.
A Synthetic Index
An index is a portfolio of stocks (or assets) formed according to some rules. A
synthetic index is a portfolio created for the purpose of replicating the returns on a
traded index. Indexes come in two varieties: those that adjust for dividends and those
that do not. Total return indexes assume that all disbursements from the portfolio’s
stocks, including regular dividends, get reinvested in the index portfolio. For total
return indexes, the reinvestment assumption effectively removes the dividends from
consideration because there are no cash flows (see Figure 12.3). In this case, forwards
on total return indexes can be priced using the no cash flow assumption, and the
results of Chapter 11 apply.
However, most indexes, including the Dow Jones Company’s and Standard and
Poor’s, make no adjustment for regular cash dividends. In this case, dividends need to
be explicitly considered. The trick is to compute the value of the index today (time 0),
excluding the present value of the dividends paid over the life of the forward contract.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
To compute this adjustment, it is helpful first to consider how such an indexe’s value
changes if all dividends are reinvested, as in a total return index.
Let the value of the index today without reinvestment be S(0) and its value at time
T be denoted S(T). Assume that it pays a continuous dividend yield of 𝛿 percent per
year. For comparison purposes, consider an otherwise identical index with the same
time 0 value I(0) = S(0). For this index, however, assume that all of the continuously
paid dividends are reinvested back into this index. Let us call this the synthetic total
return index. Because the dividends are reinvested into I(T), its value at time T will
exceed that for the index where dividends are not reinvested, that is, I(T) > S(T). In
fact, using Result 3.2, we know that
I (T) = S (T) eᄕT
Note that S(0) represents the present value of the index at time T plus the dividends.
To get the present value of the index, less the continuously paid dividends over the
life of the forward contract, we just remove the dividend yield term; that is,
present value of S (T) = S (0) e−ᄕT
FORWARDS ON DIVIDEND-PAYING STOCKS 237
This expression gives us the modified “stock price” to use in the forward pricing
model. Notice that this is a quantity adjustment in contrast to the price adjustment used
previously for dollar dividends.
This modification assumes that the dividends on the index are paid over the year at
a uniform rate (see Figure 12.3). Although some dividends are smaller, while others
are larger, this is a reasonable assumption if there are sufficiently many stocks in the
index and dividends are paid randomly over time. Newspapers and other information
vendors regularly report these dividend yields.
Example 12.2 uses this stock price modification to determine the forward price
on an index. In practice, an index derivative usually has a multiplier, for example, the
Chicago Mercantile Exchange’s S&P 500 futures is $250 times the S&P 500 index
level, and if the previous day’s futures price is 1,000 and today’s price is 1,000.10,
then the long would have 250(1,000.10 – 1,000) = $25 credited to her brokerage
account. As we have done before, we ignore the multiplier and do calculations on a
per unit basis.
EXAMPLE 12.2: The Forward Price for an Index with a Dividend Yield
The Data
■ Consider a fictitious “INDY index” obtained by averaging stock prices and a synthetic index “INDY
spot” that replicates its performance. INDY’s current level S is 1,000, and the index does not reinvest
dividends. A newly written forward contract on the index matures in a year (time T). Let the
continuously compounded interest rate r be 6 percent per year.
■ Stocks constituting INDY index paid $20 of dividends last year and are expected to pay the same this
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
year. Then, the dividend yield 𝛿 = 20/1,000 = 0.02 or 2 percent per year.
■ If you buy e−𝛿T = e−0.02×1 = 0.9802 units of INDY spot and reinvest the dividends on a continuous basis,
then you will have one unit of the index worth S(T) at year’s end.
■ Table 12.2 records the following trades in an arbitrage table:
- Sell the forward contract on the INDY index. This gives a zero value at time 0 and –[S(T) – F] on
the delivery date T.
- Buy 0.9802 units of INDY spot, and record the cash flows as – 0.9802 × 1,000 = –980.20 today.
Reinvest all dividends until delivery back into the index to get S(T) on the maturity date.
- Borrow the present value of F. A zero-coupon bond maturing one year later is worth
B = e−rT = e−0.06 = $0.9418 today. Record the cash flows as BF = 0.9418F on January 1
and –F on December 31.
■ The resulting portfolio has a zero payoff on the delivery date. To prevent arbitrage, the net payoff is
zero today. Consequently,
−0.9802 × 1, 000 + 0.9418F = 0
(12.5)
or F = $1, 040.81
238 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
The Model and Arbitrage Profits
■ Replace the numbers with symbols in Equation 12.5 to get
−e−𝛿T S + e−rT F = 0
or F = Se(r − 𝛿)T
■ Suppose an errant dealer quotes $1,020 as the forward price for INDY index. As this is less than the
arbitrage-free price, buy the relatively underpriced traded forward and sell short the synthetic forward
as in Table 12.2. This gives an immediate arbitrage profit of
−e−rT F + e−𝛿T S
= −0.9418 × 1, 020 + 0.9802 × 1, 000
= $19.60
FIGURE 12.3: Creating a Synthetic Total Return
Index
Today (Time 0) Maturity (Time T)
(a) Total return index
I(0) I(T)
… …
All dividends reinvested in the index.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
(b) Most indexes
S(0) S(T)
… …
Underlying stocks pay numerous dividends,
but the index is not adjusted.
Now (Time 0) End Date (Time T)
(c) Synthetic total return index
S(0) … … S(T)eδT
Reinvest all dividends as paid
S(T)
e−δT <1 units of S(0)
FORWARDS ON DIVIDEND-PAYING STOCKS 239
TABLE 12.2: Cost-of-Carry Arbitrage with an Index That
Pays Dividends at a Continuous Rate
Portfolio Today (Time 0) Delivery/Maturity Date One
Cash Flow Year Later (Time T) Cash Flow
Sell one forward 0 –[S(T) – F]
contract on INDY index
Buy 0.9802 units of –0.9802 × 1,000 S(T)
INDY stock worth S
Borrow the present e−0.06 F –F
value of F
Net cash flow –980.20 + 0.9418F 0
This example gives our next result.
RESULT 12.1B
The Cost-of-Carry Model with a Known Dividend Yield
Consider a forward contract today (time 0) with delivery at time T. The
underlying stock index pays dividends at a continuous rate 𝛿 that are not
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
reinvested. Then,
F = Se(r − 𝛿)T (12.6)
where S is today’s price of the stock index, F is today’s forward price for
time T, r is the continuously compounded interest rate per year, and 𝛿 is the
dividend yield, which is the sum of all dividends paid over one year expressed
as a fraction of the stock index’s price.
Why bother about such small adjustments for the dividend yield? Although tiny,
these adjustments are needed to determine the forward price for a stock index. If you
trade without such adjustments, you are vulnerable to being arbitraged. Second, they
can be used to understand foreign currency forward prices, which are discussed next.
The Foreign Currency Forward Price
As currencies have negligible storage costs, we can analyze them in the cash-and-
carry framework. Currencies earn risk-free returns in their home countries through
investments in their respective government securities. Consequently, they are like
financial securities that earn dividends, and as our next example shows, Result 12.1B
can be applied.
240 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
EXAMPLE 12.3: The Cost-of-Carry Model for Foreign Currency Forward
Contracts
The Data
■ Top management of Americana Auto have asked the treasurer to repatriate to the United States £100
million from its operations in the United Kingdom and to make this money available in a year’s time.
The treasurer considers two choices: (1) convert now (with the help of the spot exchange rate and then
invest in domestic bonds) or (2) invest in foreign bonds and convert later (with the help of a forward
exchange rate or a forward contract).
■ Today’s spot exchange rate SA is $2 per pound in American terms, and its inverse SE = 1/2 = £0.50
per dollar in European terms. The continuously compounded annual risk-free interest rates are
r = 6 percent in the United States (domestic) and rE = 4 percent in the United Kingdom (foreign).
Using a Forward Exchange Rate (Investments/International Finance Course
Approach)
■ (see Figure 12.4) To convert now, multiply £100 million by the spot exchange rate SA . Invest this
dollar amount in one-year US Treasury bills. As the dollar return (1 + R) is erT = e0.06 × 1 = $1.0618,
the value after one year is
Initial sterling amount × Spot exchange rate in American terms × Dollar return
= £100 × $2 per pound × 1.0618 (12.7)
= $212.3673 million
■ For the second choice, invest in the United Kingdom’s risk-free Treasury gilt securities for one year.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
One pound gives 1 + RE = erE T = e0.04 × 1 = 1.0408. Consequently, £100 million would be worth
after one year:
Initial sterling amount × Value of one pound invested in zeros for one year
= 100 × 1.0408 (12.8a)
=£104.0811 million
■ Unless you find a magician like those of the “rare oul’ times” who can gaze at a crystal ball and fore-
tell the future, the spot exchange rate that prevails in the future is unknown. However, the treasurer
can fix the price for converting pounds sterling into dollars in one year’s time. She can do this in the
forward currency market, where a dealer stands ready to commit to converting in one year at the fair
price FA (In), where “In” indicates a link to an investment or an international finance course. Multiply
£104.0811 million by this quantity to get its dollar value:
(Initial sterling amount × Value of one pound invested in zeros) × FA (In)
(12.8b)
= $104.0811 × FA (In) million
FORWARDS ON DIVIDEND-PAYING STOCKS 241
■ To prevent arbitrage, they should be equal:
$212.3673 = $104.0811 × FA (In)
(12.9)
or, FA (In) = $2.0404 per pound
The Model
■ Replacing numbers with symbols in Equation 12.9 and remembering that the initial sterling amount
cancels, we get
FA (In) = SA e(r − rE )T (12.10)
This is similar to our previous result (Result 12.1B).
Finding the Forward Price Using a Forward Contract
■ Alternatively, she can convert pounds sterling to dollars by using a currency forward contract.
Table 12.3 shows the trades that generate the arbitrage-free forward price:
- Buy one foreign zero-coupon bond whose price is BE = 1/(1 + RE ) = 1/1.0408 = £0.9608 in the
United Kingdom. Obtain the dollar cost by multiplying it by the spot exchange rate in American
terms, –BE × SA = –0.9608 × 2 = –$1.9216. The bond is worth one pound at maturity, whose
dollar value is obtained by multiplying it by the spot rate S(T)A prevailing after one year.
- Remove the spot exchange rate from the portfolio on the delivery date by selling a forward contract
on pounds sterling with the forward price FA . This gives an immediate payoff of 0 and a future payoff
of –[S(T)A – FA ] in American terms.
- Borrow the present value of this forward price by multiplying it by the one year Treasury bill price
B = 1/(1 + R) = 1/1.0618 = $0.9418. This gives B × FA = 0.9418FA today and –FA in the future.
■ As the portfolio’s payoff is zero on the delivery date, to prevent arbitrage, it must also have a zero value
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
today. Consequently,
−0.9608 × 2 + 0.9418FA = 0
(12.11)
or FA = $2.0404 per pound
The Model
■ Replacing numbers with symbols in expression (12.11), we get
FA = SA BE /B = SA e(r − rE )T (12.12)
■ Compare and you will discover that the two forward prices are the same. The forward price in an
investments or international finance course is the same as the arbitrage reinforced price that we obtain
with the help of a forward contract! This result is known as covered interest rate parity.
242 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
FIGURE 12.4: Determining the Forward Exchange Rate
Time 0 (Today) Two Choices
Convert to dollars today Americana’s UK
Dollar amount today operations
(Multiply by spot ER, SA)
Invest in US zeros Invest in UK zeros
Earns US interest Earns UK interest
Time T (One Year Later)
Amount obtained Convert to dollars after one year Sterling amount
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
by Americana by [Multiply by forward ER, FA(In)] after one year
two paths must
be the same
Arbitrage Profits
■ Suppose a dealer forgets to adjust for the foreign interest rate and quotes a forward price of $2.1237
per pound. As this is greater than our arbitrage-free price of $2.0404 per pound, the treasurer can
create the portfolio given in Table 12.3 and immediately capture an arbitrage profit of
−BE × SA + B × FA
= −1.9216 + 0.9418 × 2.1237
= $0.0784
EXTENDED COST-OF-CARRY MODELS 243
TABLE 12.3: The Arbitrage Table for the Currency Forward
Price
Portfolio Now (Time 0) Cash Delivery/Maturity Date
Flow One Year Later (Time T)
Cash Flow
Buy one UK zero-coupon –1.9216 S(T)A
bond whose price is BE SA
today
Sell pound sterling forward 0 –[S(T)A – FA ]
Borrow the present value 0.9418FA –FA
of FA
Net cash flow –0.9608SA + 0.9418FA 0
We restate the pricing model (12.12) as our next result.
RESULT 12.2
The Foreign Currency Forward Price
Consider a forward contract today (time 0) for trading a foreign currency at
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
time T. Then the forward exchange rate is given by
FA = SA (BE /B) = SA e(r − rE )T (12.12)
where FA and SA are the forward and today’s spot exchange rates in American
terms, B ≡ B(0, T) is today’s price of an American (domestic) zero-coupon
bond that pays a dollar at time T, BE ≡ BE (0, T) is today’s price of a “European”
(foreign) zero-coupon bond that pays one unit of foreign currency at time T,
and r and rE are the continuously compounded domestic and foreign risk-free
interest rates, respectively.
12.4 Extended Cost-of-Carry Models
Consider a commodity that is primarily used for consumption or production, like
copper, corn, or crude oil. If one holds a long position in these commodities, she
incurs insurance costs, shipping costs, and storage costs. We call all of these storage
costs. Alternatively, the availability of the physical stock may enable someone to keep
244 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
a production process running smoothly or earn extra cash by lending them in times
of temporary shortages. These benefits are commonly known as the convenience
yield (or convenience value) of the commodity.2 They aren’t directly observed but
may be implicitly computed from market data. By contrast, forward traders do not
acquire these benefits or incur these costs. However, they exist, and Example 12.4
shows how to incorporate them in our framework.
EXAMPLE 12.4: The Cost-of-Carry Model for Commodities with Storage
Costs and Convenience Yields
The Data
■ Consider a fictitious commodity Alloyum that is costly to store but has a convenience yield. Let the
simple interest rate be 6 percent per year. Consider a newly written forward contract that matures in
a year. We create the portfolio shown in Table 12.4:
- Buy Alloyum for S = $100 today. It costs $0.05 per month to store an ounce of this commodity. This
storage cost G is 12 months × 0.05/month = $0.60 for the year, which is paid up front. This expense
does not show up on the delivery date, but you have to borrow this amount today in addition to
the commodity’s price. At maturity, repay (S + G)(1 + R) = 100.60 × 1.06 = $106.6360.
- As the holder of Alloyum, collect the convenience yield Y of $0.50 at time T.
- Borrow to finance the purchase and sell forward to eliminate the spot.
The Model and Arbitrage Profits
■ The portfolio has zero value today and a known (or a constant) value at maturity. To prevent arbitrage,
it must have a zero value on the delivery date. Replace numbers with symbols to get the pricing model:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
F − 106.6360 + 0.50 = F − (S + G) (1 + R) + Y = 0
(12.13)
or F = (S + G) (1 + R) − Y = $106.1360
■ If an errant dealer quotes F as $107, then the portfolio in Table 12.4 gives an arbitrage profit of
107 − 106.6360 + 0.50
= $0.8640
on the delivery date.
2
Using tools of modern option princing theory, Jarrow (2010) shows that convenience yields are best
understood as a label given to certain cash flows generated from storing a commodity.
EXTENDED COST-OF-CARRY MODELS 245
TABLE 12.4: The Arbitrage Table for Forward Contract Commodities
with Storage Costs and Convenience Yields
Portfolio Today (Time 0) Cash Flow Delivery Date One Year
Later (Time T) Cash Flow
Buy commodity (pay asset price S −100 − 0.60 S(T)
and storage cost G)
Get convenience value Y at maturity 0.50
Borrow to finance purchase 1000.60 –106.6360
Sell forward 0 –[S(T) – F]
Net cash flow 0 F – 106.6360 + 0.50
We write the result.
RESULT 12.3
The Cost-of-Carry Model with Storage Costs and
Convenience Yields
Consider a forward contract written today (time 0) on a commodity that
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
incurs storage costs (which are collected up front) but also gives a convenience
yield (which is received at the maturity date T). Then,
F = (S + G) (1 + R) − Y (12.13)
where 1 + R is the dollar return, S is today’s spot price, F is today’s forward
price for time T, G is the storage cost, and Y is the convenience yield.
This model introduces storage costs and convenience yield on arbitrary dates.
Why up front and at the back end, respectively? If benefits and costs happen often
enough, you can avoid this problem by using a continuously compounded version
of the model. Assume that all holding costs (like the storage cost and the interest
rate) as well as the benefits (like the dividend yield and the convenience yield) are
accrued or disbursed at a continuous rate over the life of the forward. Using a quantity
adjustment, start your trade with e−(𝛿 + 𝛾 − g)T units of the commodity. Following
a path trodden before, you can develop a generalized version of the cost-of-carry
model.
246 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
RESULT 12.4
An Extended Cost-of-Carry Model
Consider a forward contract today (time 0) that matures at time T. The
underlying asset pays cash flows at a continuous rate 𝛿. Then,
F = Se[(r + g) − (𝛿 + y)]T (12.14)
where r is the risk-free rate per year, g is the rate of continuously paid storage
costs per year, and y is the rate of continuously paid convenience yields
per year.
The basic tenets of the extended cost-of-carry model are as follows:
■ There is some deliverable amount of the spot commodity that is carried across time
and traded at a later date.
■ The spot price equals the forward price in the absence of costs and benefits to the
physical holder of the asset (see Figure 12.5).
■ Various “costs” increase while benefits decrease the forward price.
12.5 Backwardation, Contango, Normal
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Backwardation, and Normal Contango
“I can call spirits from the vasty deep,” said Glendower, in Shakespeare’s King Henry
VI. To this, Hotspur replied, “Why, so can I, or so can any man; But will they come
when you do call for them?” We can keep adding features and build fancier cost-
of-carry models that give deeper results. But are they useful? If you are interested in
determining fair value, then the answer is yes, but if you are interested in determining
risk premium or whether forward prices are unbiased predictors of future spot
prices, then the answer is no. To answer these later questions, we need to discuss
backwardation, contango, normal backwardation, and normal contango.
When the forward price is higher than the spot price, we say that the market is in
contango. This is the usual situation because for most commodities, using the cost-
of-carry model (consider the extended cost-of-carry model of Result 12.4), the costs
exceed the benefits. When the reverse happens, the market is in backwardation (see
Figure 12.5). The cost-of-carry model using convenience yields is how a financial
engineer understands pricing.
Economists, on the other hand, seek explanations via the demand and supply
mechanism.3 Their aim is to understand risk premium, the compensation for risk,
3
“Give me a one-handed economist! All my economists say, ‘On the one hand . . . on the other,’”
exclaimed Harry S. Truman (1884–1972), who was the thirty-third president of the United States
(1945–53).
BACKWARDATION, CONTANGO, NORMAL BACKWARDATION, AND NORMAL CONTANGO 247
and they are interested in the relation between forward prices and expected future
spot prices. For forwards, economists often consider a market where hedgers and
speculators interact.
Consider a market where all traders expect that in three months from today, the
spot price for corn will be $5 per bushel. Suppose that the market is dominated by
corn processors who want to hedge. They are interested in fixing a price for procuring
corn in the future to make corn oil, breakfast cereal, and other corn products. They
are willing to pay a higher price than $5 per bushel to reduce their risk for buying
an input into their production process. The demand curve for the producers (whom
we label as net hedgers) is given in Figure 12.6. Notice that this is a relatively steep
demand curve, indicating insensitivity of these hedgers to the forward price because
they need forwards to set up their hedge.
Now there are speculators who take the other side of the transaction and expect a
profit. They are willing to short forward if they receive a forward price that is higher
than the expected spot price (and go long if the opposite happens). Consequently,
they have a relatively flat demand curve that passes through the expected spot price.
Notice that we have extended the horizontal axis to the left side of the vertical axis
to accommodate short positions. Suppose that at $5.05 per bushel, the net hedging
demand from the long equals the number of contracts the speculators are willing
to go short, which is 500 in this example. This creates a situation in which the forward
price is greater than the expected future spot price. The difference is a risk premium, the
compensation to the speculators for the risk they bear. To avoid confusion with the
definition of contango, we call this normal contango.
FIGURE 12.5: Link between the Spot and the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Forward Price
Today Forward Price
Contango
(Forward price > Spot)
Interest, Storage Cost
(increase cost of carry)
Spot Forward Price
Dividend, Convenience Yield
(decrease cost of carry) Backwardation
(Forward price < Spot)
248 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
FIGURE 12.6: Normal Contango and
Normal Backwardation
Normal Contango
Forward Price
Speculators (Net)
5.05 Hedgers (Net)
5
Short Contracts –500 0 500 Long Contracts
Expected spot price is $5
Forward price is $5.05 in equilibrium
(net demand = net supply)
Normal Backwardation
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Farmer
Forward Price
Hedgers (Net) 5
Speculators (Net)
4.95
Short Contracts –450 0 450 Long Contracts
Expected spot price is $5
Forward price is $4.95 in equilibrium
MARKET IMPERFECTIONS 249
Consider another scenario where corn farmers, who want to hedge by selling
forward and fix a selling price for their corn, dominate the forward market (see
Figure 12.6). These hedgers would like to go short. Their demands are met by
speculators who are willing to go long only if the expected spot price exceeds the
current forward price. This is because for the speculators to be in the market, their
expected return must be positive. By selling at a price below the expected future spot
price, hedgers entice the speculator to go long, and this transfers the unwanted risk
to the speculator. This creates a normal backwardation because the forward price is less
than the expected future spot price. And, the risk premium is reversed.
The concepts of normal contango and normal backwardation are, thus, exactly the
notions needed to understand risk premium and whether the forward price exceeds
or is less than the expected futures spot price of the commodity. Our cost-of-carry
models cannot answer these more subtle questions. The trade-off is that the cost-of-
carry model gives a price, but the economic approach doesn’t.
12.6 Market Imperfections
How does the cost-of-carry model get modified if we start considering market
“imperfections” such as transaction costs, different borrowing and lending rates, and
restrictions on short selling? Here the cash-and-carry arbitrage and the reverse cash-
and-carry are no longer symmetric. Consequently, we get price ranges within which
the forward price lies. These bounds are likely close because it’s the trader with the
lowest transaction costs (and not retail traders like us) who take advantage of arbitrage
opportunities and determine the price bounds. Example 12.5 introduces brokerage
commissions into our basic cost-of-carry model.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
EXAMPLE 12.5: Forward Price Bounds in the Presence of Market
Imperfections
■ Consider the data from Example 11.1. YBM’s stock price S is $100 today (January 1, time 0), and the
dollar return (1 + R) is $1.06. Consider a newly written forward on YBM that matures in one year
(time T).
■ Assume that when trading YBM, a brokerage commission is charged that equals TC = 1 percent of
the stock price. For simplicity, assume that there are no transaction costs on the maturity date. Now
we need two tables, Tables 12.5a and 12.5b, for an arbitrage portfolio.
■ Table 12.5a shows, to prevent arbitrage,
−101 + 0.9434F ≤ 0
(12.15a)
or F ≤ 107.06
This gives an upper bound for the forward price.
250 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
■ Table 12.5b shows the no-arbitrage portfolio for a reverse cash-and-carry. Here
−0.9434F + 99 ≤ 0
(12.15b)
or 104.9396 ≤ F
This gives a lower bound for the forward price.
■ Combining expressions (12.15a) and (12.15b),
104.9396 ≤ F ≤ 107.0596 (12.15c)
The forward price lies between $104.9396 and $107.06.
■ Replacing numbers with symbols gives the formula
S (1 + R) (1 − TC) ≤ F ≤ S (1 + R) (1 + TC) (12.16a)
TABLE 12.5A: Arbitrage Table for Cash-and-Carry
Portfolio Today (Time 0) Cash Flow Maturity Date (Time 1) Cash Flow
Buy spot (pay 1 percent –100 × (1 + 0.01) S(T)
transaction cost today)
Sell forward 0 –[S(T) – F]
Borrow present value of F 0.9434F –F
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Net cash flow –101 + 0.9434F 0
TABLE 12.5B: Arbitrage Table for Reverse Cash-and-Carry
Portfolio Today (Time 0) Cash Flow Delivery Date (Time 1) Cash Flow
Sell spot (deduct 100 × (1 – 0.01) –S(T)
transaction costs from
short sale proceeds)
Buy forward 0 S(T) – F
Lend present value of F –0.9434F F
Net cash flow –0.9434F + 99 0
MARKET IMPERFECTIONS 251
Besides brokerage commissions, there may be a bid/ask spread.4 As transaction
costs increase, the price bounds become farther apart. However, different traders
may have different transactions costs. This opens up quasi-arbitrage (near-arbitrage)
opportunities for the lowest cost trader.
One can introduce more market frictions in a similar fashion. For example, with
unequal borrowing and lending rates, the relation becomes
S (1 + RLEND ) ≤ F ≤ S (1 + RBORROW ) (12.16b)
where 1 + RLEND is the dollar return from lending and 1 + RBORROW is that for
borrowing. Several of these relations may be combined.
RESULT 12.5
The Cost-of-Carry Model in the Presence of Brokerage
Commissions and Unequal Borrowing and Lending Rates
When stock trades are charged a brokerage commission of TC on today’s trade,
and the borrowing rates are higher than the lending rates, then the forward
price F satisfies:
S (1 + RLEND ) (1 − TC) ≤ F ≤ S (1 + RBORROW ) (1 + TC) (12.16a)
where S is today’s stock price, F is today’s forward price for time T, and
1 + RLEND and 1 + RBORROW are the dollar returns for lending and
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
borrowing, respectively.
Like playing with children’s blocks, you can combine Results 12.4 and 12.5
to build even more general models. To apply the cost-of-carry model in practice
requires a good understanding of the market and matching the market’s structure
to the model’s assumptions. For financial commodities like stock indexes or Treasury
securities, convenience yields and storage costs are not relevant, so the simpler models
apply. In contrast, for agricultural commodities like corn or wheat, convenience
yields and storage costs are extremely important and need to be included. Given
the previous discussion, all that is left for a correct application of these models is
experience and market savvy. Both of these will come in good time, and the old
saying is especially relevant in this regard: “practice makes perfect.”
4
And when you sell short, the brokerage firms may not allow you the full proceeds—you may only
earn interest on part of the funds kept with the broker. When there are restrictions on short selling, you
may have an inequality like S(1 + q × Interest) ≤ F ≤ S(1 + R), where q is the fraction of usable funds
derived from the short sale. Notice that this arbitrage only works in one direction because long and short
positions have asymmetric payoffs. This relationship is less important because rules of short selling tend
to vary from country to country.
252 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
12.7 Summary
1. We extend the basic cost-of-carry model to incorporate market frictions (relax
assumption A1 of Chapter 11) and dividends (ease A4) into the framework.
2. Assuming that interest is the only relevant carrying charge, we get three forward
pricing models for assets that generate income:
a. The simplest dividend model assumes that the underlying stock pays fixed dollar
dividends on known dates. This gives Result 12.1A, which states that the stock
price net of the present value of dividend is the discounted forward price or S
– B1 div = BF. This is useful for finding forward prices for contracts written on
dividend-paying stocks and coupon-bearing Treasury bonds that mature after
the delivery date.
b. When the underlying stock pays dividends at a continuous rate 𝛿, we get
Result 12.1B: the forward price is the compounded value of the stock price,
where compounding is done at a net rate equal to the difference between the
interest rate and the dividend yield or F = Se(r − 𝛿)T . This is useful for finding
forward prices when the underlying is an index portfolio.
c. A modification of the previous model gives covered interest rate parity, which
expresses the forward exchange rate as the spot exchange rate compounded at
a net rate equal to the difference between the domestic and the foreign interest
rates. Result 12.2 states that FA = SA (BE /B) = SA e(r − rE )T .
3. For commodities mainly used for consumption or production, storage costs can
be important. However, holders of the physical commodity may also get a
convenience yield. We include these costs and benefits to build a generalized cost-
of-carry model (Result 12.4): the forward price is the compounded value of the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
spot price, where compounding is done at a net rate equal to the costs (interest
rate r plus the storage cost rate g) minus the benefits (dividend yield 𝛿 plus the
convenience yield y) or F = Se((r + g)−(𝛿 + 𝛾))T .
4. Contango is a market where the forward price is higher than the spot price. This
happens when the net cost-of-carry is positive in the generalized cost-of-carry
model. The market is in backwardation when the reverse happens.
5. Normal contango happens when a majority of the hedgers want to set up a long
hedge. Speculators assume the other side of the trade, but they require a risk
premium for supplying liquidity: the forward price is greater than the expected
future spot price. When the net hedgers are short, the risk premium is reversed,
and a normal backwardation market is created where the forward price is less than
the expected future spot price.
6. Buys and sells are no longer symmetric when market imperfections like broker-
age costs, differential borrowing and lending rates, and short selling restrictions
are introduced. Well-defined prices give way to ranges within which forward
prices lie.
QUESTIONS AND PROBLEMS 253
12.8 Cases
Diva Shoes Inc. (Darden School of Business Case UV0265-PDF-ENG, Harvard
Business Publishing). The case studies a US-based manufacturer’s currency risk
exposure and considers whether hedging via a forward contract or a currency
option is advisable.
Leveraged Buyout (LBO) of BCE: Hedging Security Risk (Richard Ivey School
of Business Foundation Case 908N23-PDF-ENG, Harvard Business Publishing).
The case considers an equity partnership’s currency risk exposure and evaluates
various derivative instruments for hedging those risks.
Risk Management at Apache (Harvard Business School Case 201113-PDF-ENG).
The case evaluates a company’s hedging strategy and the derivatives used for this
purpose.
12.9 Questions and Problems
12.1. a. Explain the cost-of-carry model with dollar dividends in your own words.
b. Justify why the spot price considered in the model is net of the present value
of all future dividends paid over the life of the contract.
c. Why don’t we adjust for dividends that are paid after the forward’s maturity
date?
12.2. Boring Unreliable Gadget Inc.’s stock price S is $50 today. It pays a dollar
dividend after two months. If the continuously compounded interest rate is 4
percent per year, what is the forward price of a six-month forward contract on
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
BUG?
12.3. Boring Unreliable Gadget Inc.’s stock price S is $50 today. It pays dividends
of $1 after two months and $1.05 after five months. If the continuously
compounded interest rate is 4 percent per year, what is the forward price of a
six-month forward contract on BUG?
12.4. Boring Unreliable Gadget Inc.’s stock price S is $50 today. The company,
however, reliably pays quarterly dividends to shareholders. For example, BUG
paid $0.95 dividend one month back; it will pay $1 dividend two months from
today, $1.05 after five months, $1.10 after eight months, and $1.15 after eleven
months. If the continuously compounded interest rate is 4 percent per year,
what is the forward price of a six-month forward contract on BUG?
12.5. Boring Unreliable Gadget Inc.’s stock price S is $50 today. It pays dividends
of $1 after two months and $1.05 after five months. The continuously
compounded interest rate is 4 percent per year. If the six-month forward price
is $51, demonstrate how you can make arbitrage profits or explain why you
cannot.
254 CHAPTER 12: THE EXTENDED COST-OF-CARRY MODEL
12.6. Boring Unreliable Gadget Inc.’s stock price S is $50 today. It pays dividends
of $1 after two months and $1.05 after five months. The continuously
compounded interest rate is 4 percent per year. Transactions costs are $0.10
per stock traded, a $0.25 one-time fee for trading forward contracts, and no
charges for trading bonds. If the six-month forward price is $51, demonstrate
how to make arbitrage profits or explain why you cannot.
12.7. Explain how the cost-of-carry model with dollar dividends differs from the
cost-of-carry model with dividend yields.
12.8. What is a convenience yield for a commodity and why is it important to
include in the cost-of-carry model? Give an example of a commodity that
has a convenience yield.
12.9. a. What is a stock index? What happens to stock indexes when dividends are
paid?
b. What is a synthetic index? Why is it useful?
12.10. Consider SINDY Index obtained by averaging stock prices and a synthetic
index SINDY spot that replicates its performance. (1) SINDY’s current level
I is 10,000, and the synthetic index’s price S is $10,000. (2) A newly
written forward contract on the index matures after T = 0.5 years. (3) The
continuously compounded interest rate r is 5 percent per year. (4) Stocks
constituting SINDY spot paid $190 of dividends last year and are expected to
pay the same this year.
a. Compute the dividend yield 𝛿 on SINDY spot.
b. Compute the forward price F.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
12.11. Consider the data given in problem 12.11 for futures contracts on SINDY
Index. If the forward price is F = $10,231, show how you can make arbitrage
profits or explain why you cannot.
12.12. Consider the data given in problem 12.11 for futures contracts on SINDY
Index. Suppose you have to pay transaction costs: (1) When you go long or
sell short a portfolio of stocks, transaction costs (brokerage fees plus the price
impact of the trade) equal 10 basis points of the synthetic index’s price. (2)
There is a one-time fee of $15 for trading forward contracts but no charges
for trading bonds. If the forward price is F = $10,231, show how you can
make arbitrage profits or explain why you cannot.
12.13. Today’s spot exchange rate SA is $1.30 per euro in American terms. The
continuously compounded annual risk-free interest rates are r = 4 percent in
the United States (domestic) and rE = 3 percent in the Eurozone. What is the
four-month forward rate in American terms if the cost-of-carry model holds?
QUESTIONS AND PROBLEMS 255
12.14. The current dollar/Swiss franc spot exchange rate is 0.5685. If you invest one
dollar for ninety days in the US domestic riskless asset, you earn $1.0101,
and if you invest one franc for ninety days in the Swiss riskless asset, you
earn 1.0113 francs (assume continuous compounding). A broker offers you
a ninety-day forward contract to buy or sell 1 million francs at the exchange
rate of 0.55 dollars/franc. Are there arbitrage profits to be made here? If so,
compute them.
12.15. Alloyum costs $0.10 per month to store (which is paid upfront) but gives a
convenience yield of $0.12 per month (which is received on the maturity
date). If its spot price S is $200 per ounce and the continuously compounded
interest rate r is 5 percent per year, what is the five-month forward price for
Alloyum?
12.16. In the previous example, suppose a trader quotes a five-month forward price
of $206 per ounce. Demonstrate how you can make arbitrage profits from
these prices; if you cannot, explain your answer.
12.17. COMIND Index is computed by averaging commodity prices. Compute
the five-month forward price for this index if the spot price is 1,000 and
the continuously compounded annual rates for various costs and benefits are
5 percent for the interest rate, 3 percent for the dividend yield, 4 percent for
the storage cost, and 3 percent for the convenience yield.
12.18. Find the price bounds for the five-month forward price for Boring Unreliable
Gadgets when
a. BUG’s stock price S is $50 today
b. a trader can borrow money at 5 percent and lend money at 4 percent,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where the interest rates are annual simple interest rates
c. a brokerage commission of 0.5 percent of the stock price is charged today
but your broker waives transaction costs on the maturity date
12.19. Forward prices for April and June forward contracts for platinum are $400 per
ounce and $410 per ounce, respectively. The interest rate is 1 percent for the
April–June period. There is a 1 percent transaction cost whenever you trade
(buy or short sell) the spot in June, but the broker is waiving any transaction
costs for trading forward contracts and brokerage costs for trading the spot in
April. Can you generate arbitrage profits? Explain your answer. (Assume that
the spot price of platinum in June lies between $400 and $500 per ounce.)
13
Futures Hedging
13.1 Introduction 13.5 Futures versus Forward
Hedging
13.2 To Hedge or Not to Hedge
Should the Firm or the Individual 13.6 Spreadsheet Applications:
Hedge? Computing h
The Costs and Benefits of 13.7 Summary
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Corporate Hedging
13.8 Appendix
EXTENSION 13.1 Airlines and
Fuel Price Risk Deriving the Minimum-Variance
Hedge Ratio (h)
EXTENSION 13.2 A Hedged
Firm Capturing a Tax Loss Computing the
Minimum-Variance Hedge
13.3 Hedging with Futures Ratio (h)
Perfect and Imperfect Hedges Statistical Approach
Basis Risk Econometric Approach: A Linear
Guidelines for Futures Hedging Regression Model
13.4 Risk-Minimization Hedging 13.9 Cases
The Mean-Variance Approach 13.10 Questions and Problems
Limitations of Risk-Minimization
Hedging
TO HEDGE OR NOT TO HEDGE 257
13.1 Introduction
The novelist Fyodor Dostoyevsky perceptively observed in The Gambler: “as for profits
and winnings, people, not only at roulette, but everywhere, do nothing but try to
gain or squeeze something out of one another.” This aptly defines a zero-sum game of
which, excluding transaction costs, futures trading is an excellent example. There is a
pool of talented futures traders who earn a living from this business thereby depressing
the chances of winning for ordinary traders who lose money on average. Then, why
trade futures?
Cynics give a facile explanation: people are “into commodities” because it’s as
thrilling as going to the casinos. True to circus-owner P. T. Barnum’s immortal quote
that “there’s a sucker born every minute,” these gamblers gladly lose for the joy of
the ride. Perhaps others believe that they really can win the futures game. We know
no surefire strategy. Even if you stumble on such a strategy, repeated use will destroy
its efficacy. So, why trade futures?
In reality, most futures contracts are traded to hedge risks that preexist in some
line of business, and hedgers are willing to give up expected returns as payment for
this “insurance.” We discussed this issue in the context of normal backwardation in
Chapter 12. For example, Figure 13.1 shows a man selling goods on some tourist
spot, perhaps a beach. If he chooses his wares wisely, say, by carrying an assortment of
umbrellas and sunglasses, then he has hedged well. Come rain or come shine, he has
something to sell. If this is not possible, he could use futures to manage these risks.
For example, he could sell only umbrellas and buy futures on a sunglass company’s
stock, instead of selling sunglasses.
Who hedges? A majority of the Fortune 500 companies and a growing number
of smaller firms hedge risk. Farmers often hedge the selling price of their produce.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Many producers hedge input as well as output price risk.
This chapter explores the reasons for hedging, discusses its cost and benefits, and
introduces futures hedging strategies. We discuss perfect and cross hedges, long and
short hedges, and risk-minimization hedging using standard statistical techniques.
13.2 To Hedge or Not to Hedge
Unfortunately, there is no general answer to what may be phrased as (with apologies
to the immortal William Shakespeare) to hedge or not to hedge—that is the question. We
can only discuss hedging’s benefits and costs.
Should the Firm or the Individual Hedge?
The classic Modigliani and Miller (M&M) papers argued that debt policies do not
affect firm value. This irrelevance follows because a shareholder can replicate such
policies herself by trading stocks in otherwise identical firms. Therefore the no-
arbitrage principle ensures that these companies, differing only in their debt policies,
must have equal value. The same argument can be applied to a firm hedging its
balance sheet risks giving intellectual support to the irrelevance of hedging. However,
the M&M results rely on several key assumptions, including no market frictions
258 CHAPTER 13: FUTURES HEDGING
FIGURE 13.1: Wisely Hedged: Come Rain, Come Shine,
He Has Something to Sell*
Umbrellas $10
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Umbrellas $10
Sunglasses $15
Inspired by a 1980s newspaper advertisement.
(including bankruptcy costs), no taxes, and no information asymmetry. In the real
world, these assumptions fail to hold. Consequently, if reducing the firm’s risk is in
the shareholders’ best interest, then a firm can usually do a better job of hedging
than individual investors. Indeed, a company (1) is likely to have lower transaction
costs; (2) can trade larger contracts than a shareholder; (3) can dedicate competent
personnel to hedging; (4) can hedge by issuing a whole range of securities, which
individuals cannot; (5) may have private information about the company’s risks; and
(6) can hedge for strategic reasons that lie beyond an ordinary investor’s knowledge.
In all these cases, a company creates more value than an individual investor hedging
on her own. But hedging is costly to implement, both in developing and retaining
the relevant expertise within the firm, and in executing the transactions.
TO HEDGE OR NOT TO HEDGE 259
The Costs and Benefits of Corporate Hedging
A hedged business has fewer risks to worry about, but exactly why is this an advantage?
Aren’t businesses supposed to take risks? We answer these questions next, and although
we focus on the use of futures, many of these same benefits can be obtained with other
derivatives such as forwards, options, and swaps:
1. Hedging locks in a future price. Recall Chapter 8, in which we discussed the
establishment of the Chicago Board of Trade. The CBOT built orderly markets
for grain trading to avoid price crashes at harvest times and booms afterward.
Hedging with futures helps smooth such demand–supply imbalances. It allows
traders to lock in stable prices and plan production and marketing activities with
greater certainty.
2. Hedging permits forward pricing of products. For example, many airlines routinely
hedge aviation fuel prices, which is a major input cost (see Extension 13.1 for
a discussion). If fuel costs can be fixed (and other costs, such as staff salary, airplane
depreciation, airport gate rental charges, and travel agents’ commissions, can be
estimated in advance), then the airline can better set a profit margin and determine
future seat prices. Customers like knowing travel costs far in advance.
3. Hedging reduces the risk of default and financial distress. Default means a failure to pay
a contractual payment, such as debt, when promised. Default triggers unfavorable
outcomes such as lawyer fees, potential bankruptcy costs and liquidation fees,
losing control over the company’s assets, and increased costs of doing business
because suppliers fear that they may not be paid and customers may worry about
quality and service. These are called financial distress costs. A company can reduce
the likelihood of incurring these costs by hedging some of the risks it faces.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
4. Hedging facilitates raising capital. Because of the decreased risk of default, bankers
allow hedgers to borrow a larger percentage of a commodity’s value at a lower
interest rate than nonhedgers. This same logic applies to firms.
5. Hedging enables value-enhancing investments. Froot et al. (1994) argue that if external
sources of funds (like stock and bond issuances) are costlier to corporations than
internally generated funds, then hedging can help stabilize internal cash flows and
make them available for attractive investment opportunities.
6. Hedging reduces taxes. A hedge may be used to capture the benefits of a tax loss
or take advantage of a tax credit. The basic intuition is simple. Because one can
use a tax loss or a credit only when the company has positive profits, it may be
worthwhile to smooth out profits by hedging rather than letting them fluctuate.
See Extension 13.2 for an example that illustrates this notion.
However, hedging also incurs costs. When trading with a speculator, a hedger
often pays an implicit fee by trading at a price inferior to the expected future
payoff. Of course, brokers must also be paid. The presence of these trading costs
constitutes a basic argument against futures hedging. Moreover, businesses must
allocate valuable personnel to devise hedging strategies, set up adequate checks and
260 CHAPTER 13: FUTURES HEDGING
balances to prevent rogue employees from ruinously speculating with the firm’s
money, ensure compliance with the laws of the land, and meet the government’s
accounting requirements in this regard.1
Hedging is analogous to purchasing an insurance policy on some commodity’s spot
price. It pays off when spot prices move in an adverse direction. However, as with
all insurance policies, there is a cost—the premium. If the spot price does not move
adversely, one pays for insurance not used. Risk-averse individuals buy insurance
despite this cost, and analogously, firms often hedge.
To hedge or not to hedge remains an unsettled question that must be resolved on a
case-by-case basis. One often hears any number of catchy maxims: “no risk, no gain”
alongside “risk not thy whole wad.” Which to choose requires expertise, finesse, and
knowledge, which are acquired through study and experience.
EXTENSION 13.1: Airlines and Fuel Price Risk
Fuel costs can hit the airlines hard. Fuel is a major cost of the airline business; it can be between 10 percent (in
good times) and more than 35 percent (in bad times) of their average expenses.2
“When the going gets tough, the tough get going” goes a familiar saying. Crude oil prices had a huge run-up
in 2008 and peaked at over $145 per barrel during July of that year. As the oil price went up, the airlines adopted a
wide range of measures, some prudent and others drastic, to reduce fuel consumption. Most of these approaches
aimed at lowering the aircraft’s weight because a lighter plane is cheaper to fly. Airlines started charging for
checked-in luggage, which “killed two birds with one stone”: raised money and reduced weight. Other steps
included flying more fuel-efficient planes and reducing the time that plane engines stay on. Some started trading
commodities. Yet, crude oil prices started falling and jet fuel prices also declined. In November 2008, crude prices
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
closed below $60 per barrel. A report in the New York Times entitled “Airlines find it difficult, and perhaps
unwise, to hedge fuel prices” dated November 13, 2008 reported that the German airline Lufthansa’s “hedging
for the current year had been cut to 72 percent from 85 percent of its total fuel bill ... Its hedging for 2009 is
down to 57 percent, at an average of $91 a barrel.” Stunned by unprecedented oil price volatility, several other
airlines were left holding costly hedges.
Many commercial airlines as well as companies like FedEx (which uses its extensive fleet to quickly deliver
packages to numerous locations around the globe) use derivatives on oil to hedge. The hedged amount varies.
Sometimes they hedge little or none of their exposure; at other times they hedge much more (see Table 13.1 for
a description of the fuel hedging at some of the world’s largest airlines based on thier annual reports). Southwest
Airlines has a history of running a successful fuel “hedging program.” It has maintained its long track record of
profitability into the new millennium by extensive hedging even at a time when many other airlines were reeling
from losses and operating under bankruptcy protection. Consider the 2016 annual report of Southwest Airlines,
which states “Because the Company uses a variety of different derivative instruments at different price points,
the Company is subject to the risk that the fuel derivatives it uses will not provide adequate protection against
1
Barings Bank, one of England’s oldest and most prestigious merchant banks, founded in 1762 by
Sir Francis Baring, collapsed in 1995 after a rogue employee, Nick Leeson, lost $1.4 billion of company
money speculating in futures contracts. Nowadays, most companies have stronger oversight of the firm’s
overall risk situation through the office of a chief risk management officer, who often reports directly to
the chief executive officer.
TO HEDGE OR NOT TO HEDGE 261
significant increases in fuel prices and could in fact result in hedging losses, and the Company effectively paying
higher than market prices for fuel, thus creating additional volatility in the Company’s earnings. … Excluding
the impact of hedging, Fuel and oil expense would have decreased by $535 million, or 15.9 percent, compared
with 2015, due to lower market jet fuel prices.”
TABLE 13.1: Fuel “Hedging Program‘ at some of the World’s Largest Airline
Companies by Revenue (in US dollars)
Airlines Hedging Policy (as of December 2016 or for the year ending December 31, 2016)
American Airlines Group (USA) No fuel hedges as of December 31, 2016; fully exposed to fluctuations in fuel
prices.
Delta Air Lines (USA) Manages fuel price risk through a “hedging program” with options, swaps and
futures on commodities including crude oil, diesel fuel and jet fuel. Recorded fuel
hedge losses of $366 million in 2016 and $741 million in 2015.
United Continental Holdings The Company historically hedges a portion of its planned fuel requirements. No
(USA) outstanding fuel hedges as of December 31, 2016. Although the Company’s current
strategy is to not enter into transactions to hedge its fuel consumption, it regularly
reviews its strategy based on market conditions and other factors.
Deutsche Lufthansa (Germany) Fuel costs included a loss due to hedging equal to EUR 905m (One Euro was
$1.05 on December 31, 2016). Fuel expenses were about EUR 4.9 billion in 2016.
The Lufthansa Group hedges fuel prices with a time horizon of up to
24 months.
Air France-KLM (France) Based on the forward curve at December 31, 2015, a hedging strategy, approved by
the Board of Directors, sets the hedge horizon at two years
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
(a rolling 24 months) and the target hedge ratio at 60%. The hedging uses
simple futures or option-based instruments.
International Airlines (United Fuel price risk is partially hedged through the purchase of oil derivatives in forward
Kingdom) markets. The current strategy, as approved by the IAG Management Committee, is
to hedge a proportion of the anticipated fuel consumption for the next eight
quarters. The company primarily trades forward crude, gas oil and jet kerosene
derivative contracts for hedging purposes.
Southwest Airlines (USA) Discussed in text.
Southwest’s “hedging” experience demonstrates the difficulties in distinguishing hedging from speculation and
illustrates that even well-considered actions do not always deliver favorable outcomes.
2
As it is extremely difficult to predict oil prices, it is equally difficult to devise an airline’s fuel hedging strategy, which many airlines refer
to as a “hedging program.”
262 CHAPTER 13: FUTURES HEDGING
EXTENSION 13.2: A Hedged Firm Capturing a Tax Loss
Hedging with derivatives expands a company’s choices and may allow it to take advantage of tax situations. The
next example shows how futures contracts can be used to stabilize earnings so that a company can utilize past
losses to reduce current taxes.
EXT. 13.2 EX. 1: Hedging to Capture a Tax Loss
■ Goldmines Inc. (fictitious name) mines, refines, and sells gold in the world market. The company’s profits
move in tandem with gold price movements. Assume that the pretax profit is $100 million when gold prices
go up (which happens with a 50 percent chance) and 0 if gold prices go down (which also happens with
a 50 percent chance). Alternatively, Goldmines can hedge with gold futures and have a known profit of
$48 million.
■ Ext. 13.2 Fig. 1 shows these payoffs. We label them as event 1 (profit $100) and event 2 (profit $0). These
events are shown in a binomial tree, where the payoff $100 is placed on the upper branch of the tree, while
$0 is placed on the lower branch.
Expected Profit
■ The expected profit
= [(Probability of high gold price) × (Payoff $100)] + [(Probability of low gold price) × (Payoff $0)]
= 0.50 × $100 + 0.50 × 0
= $50
for the unhedged company.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ We can show the superiority of the hedging strategy by computing the expected after-tax profits for the
unhedged and hedged firm under different scenarios.
After-Tax Expected Profits (Simple Case)
■ Assume that the tax rate is 30 percent. When the gold price is high, $100 million pretax profit gives an
after-tax profit of 100 × (1 – tax rate) = $70 million. When the gold price is low, the pretax profit is 0, and
so is the after-tax profit. After-tax profits are shown in Ext. 13.2 Fig. 1.
Expected after-tax profit for unhedged firm
= 0.50 × 70 + 0.50 × 0
= $35 million
■ The hedged firm makes a profit of $48 million irrespective of the gold price:
After − tax profit for hedged firm
= 48 × 0.7
= $33.6 million
TO HEDGE OR NOT TO HEDGE 263
■ The company can choose either. The actual choice depends on the risk preferences of the company’s
management and shareholders.
After-Tax Expected Profits (When $25 Million Tax-Deductible Loss Is Carried
Forward)
■ Now suppose that the company has accumulated losses totaling $25 million. It can deduct this loss from this
year’s profit and thus lower the tax burden. Such strategies of tax reduction are known as tax shields.
■ Assume that if unutilized, this one-shot opportunity disappears. We will show that the hedged firm can always
utilize the losses to lower its taxes, but the unhedged firm can only do this half the time.
■ For the unhedged firm, total tax when the gold price is high is
(Pretax profit − Loss deducted) × (Tax rate)
= (100 − 25) × 0.3
= $22.5 million
EXT. 13.2 FIG. 1: Profits of a Hedged and an
Unhedged Firm
Pre-Tax Profit
$100.00
0.5
Unhedged Firm Hedged Firm $48
0.5
0
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Expected profit of the unhedged firm = 0.5 × 100 + 0.5 × 0 = $50 million
Expected After-Tax Profit (with No Loss Carried Forward)
$70.00
0.5
Unhedged Firm Hedged Firm $33.60
0.5
0
Expected profit = $35 million
Expected After-Tax Profit (with $25 Million Tax Deductible Loss)
$77.50
0.5
Unhedged Firm Hedged Firm $41.10
0.5
0
Expected profit = $38.75 million
264 CHAPTER 13: FUTURES HEDGING
The total tax when the gold price is low is 0. Consequently, the after-tax profit is (100 – 22.5) = $77.5 million
when the gold price is high and zero when it is low (see Ext. 13.2 Fig. 1):
Expected after-tax profit for the unhedged firm
= 0.50 × 77.5 + 0.50 × 0
= $38.75 million
■ By contrast, the total tax for the hedged firm is (48 – 25) × 0.3 = $6.90:
After-tax profit for the hedged firm
= (48 – 6.90)
= $41.10 million
■ Clearly hedging is a superior strategy: not only does the hedged firm generate greater after-tax profit but it
also removes swings and stabilizes this amount.
13.3 Hedging with Futures
Chapter 4 gave us an inkling of how a company can use forward contracts for hedging
input and output price risks. Now that you are more familiar with the workings of
a futures contract (which are easier to use than forward contracts), let us explore
hedging with futures.
Perfect and Imperfect Hedges
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Ours is an imperfect world, in which it is hard to find perfect hedges. A perfect
hedge completely eliminates spot price risk for some commodity. This happens when
(1) the futures is written on the commodity being hedged, (2) the contract matures
when you are planning to lift the hedge, and (3) the contract size and the other
characteristics of a futures unerringly fit the hedger’s need. Imperfect or cross
hedges occur when these three conditions are not satisfied. For example, a bank
may reduce the price risk of its loan portfolio, which is vulnerable to interest rate
changes, by trading Treasury bond futures contracts. This is a cross hedge because
the futures is written on US Treasury bonds, whereas the loans being hedged consist
of, say, house mortgages, car loans, and certificates of deposit. Even when futures
on the same commodity are available, the issue of a timing mismatch may occur.
Consequently, basis risk emerges as a paramount concept in analyzing, setting, and
managing a futures hedge.
Basis Risk
Basis risk is a focal point in understanding futures hedges. The basis is defined as
the difference between the spot and the futures price. It is written as Basis = Cash
price – Futures price. Example 13.1 shows how crucial basis risk is when hedging a
HEDGING WITH FUTURES 265
commodity’s spot price risk. We illustrate this in the context of two companies using
buying and selling hedges to offset input and output risks, respectively.
EXAMPLE 13.1: A Gold Futures Basis Risk
■ Today is January 1, which is time 0. Consider the gold futures contract trading on the COMEX
Division of the CME Group described in Chapter 8. The prices are reported on a per ounce. Figure
13.2 gives the timeline and the various prices. For simplicity, we assume that no interest is earned on
margin account balances.
■ On today’s date, the spot price of gold S = $990, and the June contract’s futures price F(0) = $1,000.
The basis is
b (0) = Spot price S (0) − Futures price F (0)
= 990 − 1,000
= −$10
FIGURE 13.2: Timeline and Prices for Gold Futures
Hedging Example
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Start date Closing date Delivery period
(January 1) (May 15) (June)
Time 0 Time T
Spot price, S(0) = $990 S(T) = $950
Futures price, F (0) = $1,000 F (T) = $952
Basis, b(0) = S(0) – F (0) = –$10 b(T) = S(T) – f (T) = –$2
■ The delivery period for the contract is in June. For the subsequent discussion, suppose that on May 15,
time T, the spot price S(T) is $950 and the June futures price F(T) is $952. The new basis is
b (T) = S (T) − F (T)
= 950 − 952
= −$2
266 CHAPTER 13: FUTURES HEDGING
Goldmines Inc. Sets Up a Short Hedge
■ Suppose that the mining company Goldmines Inc. goes short one June gold futures contract on
January 1 to hedge its output price risk. The company sells gold and lifts the hedge by closing out the
futures position on May 15.
■ Goldmines sells gold for $950, but it makes – [F(T) – F(0)] = – (952 – 1,000) = $48 on the futures
position. So, the effective selling price on May 15 is (see Equation 9.1b of Chapter 9)
Spot cash flow + Futures cash flow
= 950 + 48
= $998
■ One can rewrite this as
Spot − (Change in futures price)
= S (T) − [F (T) − F (0)]
(13.1)
= [S (T) − F (T)] + F (0)
= New basis + Old futures price
No matter when the company closes its position, the old futures price is fixed, and only the new basis affects
profits.
Jewelrygold Inc. Sets Up a Long Hedge
■ Now look at the same problem from the perspective of a jewelry maker, Jewelrygold Inc. As discussed
in chapter 4, the company can reduce input price risk by setting up a long hedge (a buying hedge)
and going long gold futures. When it is time to buy gold in the spot market, the company sells these
futures and removes the hedge.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ On January 1, Jewelrygold goes long a June gold futures contract to hedge input price risk. On May
15, the company buys gold for $950 and simultaneously sells futures for $952 to end the hedge.
Although Jewelrygold pays less for gold, it has to surrender nearly all of its gains through the hedge.
The cash flow from the futures position is [F(T) – F(0)] = 952 – 1,000 = –$48.
■ Jewelrygold’s effective price on May 15 is $998. By looking at the cash flows, the buying price is
= −Spot + (Change in futures price)
= −S (T) − [F (T) − F (0)]
(13.2)
= − [S (T) − F (T) + F (0)]
= − (New basis + Old futures price)
This is the same value as in the previous example, except for the minus sign.
This example shows that a futures hedged spot commodity’s portfolio value
can always be viewed as the sum of the old futures price and the new basis.
Consequently, hedgers are interested in how the basis evolves randomly through
time. This randomness gives rise to basis risk in hedging, which is often measured by
computing the basis’s variance (or standard deviation). Basis risk is, thus, fundamental
HEDGING WITH FUTURES 267
to futures hedging, and hedgers often talk about the widening or narrowing of the
basis in passionate terms. Many of them have extensive charts depicting the historic
behavior of the basis, looking for a crystal ball to foretell the future.
Guidelines for Futures Hedging
Real life is far more complex than textbook examples. Exchanges offer only a handful
of futures, whereas there are thousands of commodities. The specter of an imperfect
hedge visits us again. When setting up a hedge, it’s natural to ask, which contract
and for what maturity? Example 13.2 discusses the issues involved in answering this
question.
EXAMPLE 13.2: Hedging by Selecting a Futures from Several Competing
Contracts
Suppose that a new alloy Alloyum (a fictitious name) is developed that can replace precious metals in
industrial and ornamental use. Let Alloyum trade in the spot market. Alloyum futures may or may not
trade, and we discuss both possibilities. If you produce thirty thousand ounces of Alloyum, how would
you hedge output price risk?
Alloyum Futures Trade, and You Know When to Lift the Hedge
This case happens when the Alloyum futures contract’s maturity date exactly matches the delivery date
of the spot commodity commitment. Then, sell [(Spot position)/(Alloyum futures contract size)] =
30,000/50 = 600 Alloyum futures, assuming a contract size of fifty ounces. Basis risk eventually disappears
because the spot converges to the futures price at maturity. This can be shown by setting S(T) = F(T) in
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Equation 13.1. With zero basis risk, you fix the selling price at maturity, which is none other than today’s
futures price. Congratulations, you have set up a perfect hedge. (This perfect hedge ignores the interest
rate risk from marking-to-market. Such reinvestment risks will be considered later in the chapter.)
Alloyum Futures Trade, and You Do Not Know When to Lift the Hedge
In this case, one way to proceed is to compute the variance of the basis for different maturity futures
contracts on Alloyum and select the one that scores the lowest. Lower basis risk means that the futures
price deviates less from the spot price. You are likely to find that the futures contract maturing in the
same month as the spot sale is the best candidate. In this case, hedge with the smallest basis risk futures
contract. The number of contracts to be shorted may be found by the risk-minimization hedging strategy
discussed later in this chapter.
Alloyum Futures Do Not Trade, and You Know When to Lift the Hedge
In this case, you may first like to select a similar commodity on which a futures contract is written and
decide on the maturity month. Collect price data for Alloyum and for contender commodity futures
contracts maturing in the same month as the spot sale, and compute correlations of price changes. Select
the commodity futures that has the highest correlation to Alloyum in the overall period.
268 CHAPTER 13: FUTURES HEDGING
Alloyum Futures Do Not Trade, and You Do Not Know When to Lift the Hedge
This is the worst case for hedging. First, select a commodity futures contract as in the previous case.
Then, select a maturity month, and find the futures contract with the largest correlation with Alloyum
spot price changes and that minimizes the variance of the basis. There is, however, a famous saying in
economics that “there are no solutions, only trade-offs” (Sowell 1995, p 113). The other futures contracts
may have lower transaction costs. Liquidity is another concern because it often increases as the contract
approaches maturity, but dries up in the delivery month.
13.4 Risk-Minimization Hedging
Suppose that you want to hedge a long spot commodity position by shorting futures
contracts. Using a statistical or econometric model, assuming that past price patterns
repeat themselves enables you to find the optimal number of contracts to minimize
the variance of the hedging error. This section discusses this statistical approach to
hedging.
The Mean-Variance Approach
The mean-variance approach to risk-minimization hedging (or optimal hedging)
determines the optimal number of contracts needed to minimize the variance of the
portfolio’s price changes. The intuition of this approach can be understood by first
considering a perfect hedge:
■ A perfect hedge. Suppose you are holding a long position in the spot commodity. You
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
need to short futures contracts to hedge the spot commodity’s price risk. In the
pristine world of a perfect hedge, any change in the spot position will be exactly
offset by an equal and opposite change in the futures position. Consequently,
Change in portfolio value
(13.3)
= Change in spot position – Change in futures position = 0
The minus sign before the change in the futures position is because we are short
the futures contract and the spot and futures prices move in the same direction.
With minor modifications, we can recast the problem to focus on how changes
in the spot and futures prices per unit are related. Using data from Example 13.2,
we can hedge a long position of thirty thousand ounces of Alloyum production
by selling [(Cash position)/(Alloyum futures contract size)] = 30,000/50 = 600
Alloyum futures. We can write this as follows:
Change in portfolio value = (30, 000 × Change in spot price per ounce)
− (600 × 50 × Change in futures price per ounce)
=0
(13.4)
RISK-MINIMIZATION HEDGING 269
There is no change in the portfolio value because the spot and the futures price
changes exactly match.
■ An imperfect hedge. In reality, the hedge will be imperfect because the spot and the
futures price changes do not exactly match. Here you can find the risk-minimizing
number of futures contracts. Consider hedging our spot exposure of n ounces by
selling q futures contracts of size f ounces per contract. Then we can rewrite the
left side of Equation 13.4 as
Change in portfolio value = (n × Change in spot price per ounce)
− (q × f × Change in futures prices per ounce)
= n [S (t) − S] − qf [F (t) − F]
= nΔS − qfΔF (13.5)
where S(t) and S are the spot prices per ounce on some future date (time t) and
today (time 0), F(t) and F are futures prices per unit on those same respective dates,
and Δ compactly denotes the price change. Note that this change does not equal
zero because it is an imperfect hedge. To find the number of contracts to sell to
risk-minimize hedge the spot portfolio, compute the variance of the portfolio and
select n to minimize this variance. You can do this by taking the partial derivative
of the portfolio variance with respect to q, setting it equal to zero, and solving
for q (see the appendix to this chapter). This leads to the following result (see
Result 13.1).
RESULT 13.1
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The Risk-Minimizing Number of Futures Contracts for
Hedging a Spot Commodity Position
To minimize the risk of a long portfolio of a spot commodity, sell short q
contracts where
q = (n/f) h (13.6)
and n is the size of the spot position, f is the number of units of the under-
lying commodity in one futures contract, and h is the optimal hedge ratio (or
minimum-variance hedge ratio or risk-minimized hedge ratio).
The optimal hedge ratio is given by
covS,F sdS
h= = corr (13.7)
varF sdF S,F
where covS,F is the covariance between changes in the spot price (ΔS) and
the futures price (ΔF), varF is the variance of the change in the futures price,
sdS is the standard deviation (or volatility) of the change in the spot price
(standard deviation is the positive square root of the variance), sdF is the
270 CHAPTER 13: FUTURES HEDGING
standard deviation of change in the futures price, and corrS,F is the correlation
coefficient between ΔS and ΔF.c
c
The second part of the formula follows from the result: “covariance equals the product of the
standard deviations and the correlation coefficient” or covS,F = sdS × sdF × corrS,F .
Considering various cases reveals the intuition behind this result. Suppose that the
price changes for the spot commodity and the futures contract are perfectly correlated
(i.e., corrS,F equals 1) and their standard deviations match (i.e., sdS = sdF ). Then the
optimal hedge ratio h equals 1. This is the holy grail of a perfect hedge, which is
nearly impossible to attain in practice but is useful as an illustration.
Next assume that there is a high correlation between the price changes (corrS,F is
close to one) and the spot price change has a lower volatility than the futures price
change (i.e., sdS < sdF ). Then a unit of spot is hedged with less than a unit of futures.
Conversely, if the spot price change fluctuates more than the futures price change
(i.e., sdS > sdF ), then one unit of spot needs more than one unit of the futures to
hedge.
The optimal hedge ratio gives us an easy formula for setting up a futures hedge.
Example 13.3 shows how to implement this by utilizing spot and futures price data,
which can be easily collected from business newspapers or Internet data sources.
EXAMPLE 13.3: Setting Up a Risk-Minimization Hedge with Price Data
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Suppose that you are planning to sell thirty thousand ounces of Alloyum at some future date. To start,
you collect Alloyum spot price and platinum futures price data for sixteen consecutive trading days
(see Table 13.2).
■ Begin by computing the price differences for each of these two series of price observations. Next use
a spreadsheet to compute the parameter estimates needed for the optimal hedge ratio. (The appendix
to this chapter discusses related issues and demonstrates how to find h with the help of a simple
calculator; Section 13.6 shows how to do these calculations using the spreadsheet program Microsoft
Excel.) We get
the standard deviation of changes in the futures price per unit, sdF = 14.3368
the standard deviation of changes in the spot price sdS = 15.5002
the correlation coefficient between the spot and futures price changes, corrS,F = 0.7413
■ Use these estimates in the second part of Equation 13.7 of Result 13.1 to compute the minimum-
variance hedge ratio:
sdS 15.5002
h= corrS,F = 0.7413 = 0.80154
sdF 14.3368
RISK-MINIMIZATION HEDGING 271
■ To hedge n = 30,000 ounces of Alloyum with a platinum futures contract (contract size f = 50), you
need to sell
q = (n/f) h
= (30, 000/50) 0.8015
= 480.9 or 481 contracts (with integer rounding)
4 Or, you can use the first part of Result 13.1 to determine the optimal hedge ratio, h = cov /var = 18.3048/22.8381 =
S,F F
0.8015.
TABLE 13.2: Alloyum Spot and Platinum
Futures Price Data
Day t Alloyum Spot S(t) Platinum Futures F(t)
0 1,215 1,233
1 1,209 1,236
2 1,239 1,245
3 1,245 1,254
4 1,254 1,272
5 1,227 1,254
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
6 1,230 1,248
7 1,224 1,242
8 1,239 1,260
9 1,251 1,263
10 1,227 1,254
11 1,215 1,227
12 1,209 1,221
13 1,203 1,230
14 1,185 1,203
15 1,194 1,209
Alternatively, you can compute the hedge ratio by suitably framing the problem
and estimating a linear regression (see the appendix to this chapter).
272 CHAPTER 13: FUTURES HEDGING
Limitations of Risk-Minimization Hedging
Comedian Rodney Dangerfield often chuckled, “I don’t get no respect.” Doesn’t
this comment apply to risk-minimization hedging? It is purely computational, and
except for some ad hoc adjustments, it is a single-shot exercise with no prescription
for modifying the hedge over time. It works well if a closely related commodity can
be found and you know when to lift the hedge. It works less well otherwise and if
the hedge needs to be rolled over or rebalanced.
The idea of rebalancing is related to the concept of (exact) dynamic hedging,
which involves regular adjustments to a perfect hedge over time. Dynamic hedging
requires sophisticated analytical tools, and it applies in a complete market. It is
impossible to implement in an incomplete market (see Duffie [1989] and Jarrow and
Turnbull [2000] for examples and issues related to dynamic hedging). In complete
markets (remember Chapter 8), all kinds of securities trade that generate future
payoffs contingent on all possible future events. In reality, the markets are incomplete,
and it may be difficult to develop good futures pricing models to which dynamic
hedging applies. Consequently, risk-minimization hedging, despite all its limitations,
starts looking respectable again. In incomplete markets, it continues to attract serious
research interest.
13.5 Futures versus Forward Hedging
It’s common practice in many circles to treat forwards and futures as equivalent and
interchangeable contracts. However, this is incorrect for many reasons. Although a
forward is easy to price, ordinary investors rarely trade in the “organized” forward
market. Big banks, large corporations, and other institutions with excellent credit
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
ratings dominate this market, and to guarantee execution of the contracts, these
financial institutions usually need to post collateral. These trading restrictions are
imposed to reduce “counterparty credit risk”—the nonexecution of the contract
terms. Counterparty credit risk is a big concern in the trading of over-the-counter
derivatives, as recently evidenced by the related regulatory reforms following the 2007
credit crisis in the Dodd–Frank Wall Street Reform and Consumer Protection Act.
Second, given that forward contracts are bilateral negotiated agreements, forward
markets are illiquid and subject to significant liquidity risk. For example, if a
counterparty closes a forward contract early, significant closing costs are incurred.
By contrast, a futures (1) is an exchange-traded, standardized contract; (2) has
margins and daily settlement, which makes it safer than a forward due to the absence
of credit risk; (3) allows small traders, weaker credits, and complete strangers to
participate; and (4) usually trades in a liquid market, where traders can enter or
unwind their positions with ease.
But there is another important difference between futures and forward contracts
due to the marking-to-market of a futures contract. Marking-to-market a futures
contract introduces risks involved with reinvesting the cash flows before the contract
matures. These same reinvestment risks are not faced by a forward contract. Tracking
SPREADSHEET APPLICATIONS: COMPUTING h 273
the cash flows to these contracts, the futures trader, unlike the forward trader, earns
random and uncertain interest on margin balances.
As discussed in Chapter 9, the interest earned on a futures margin account is
dependent on the futures price changes. In addition, interest rate changes also affect
the interest earnings. Indeed, if interest rate changes are positively correlated with
futures price changes, then the futures position benefits from interest rate movements
because as cash flows are received, more interest is earned. In this case, the risk of a
futures position is reduced slightly. Conversely, if interest rate changes are negatively
correlated with futures price changes, then the futures position suffers from interest
rate movements. In this latter case, the risk is increased. This interest rate risk affects
hedging performance because of the random cash flows received. This same risk is
not present in a forward contract.
13.6 Spreadsheet Applications:
Computing h
You can easily compute the optimal hedge ratio (h) by using a standard spreadsheet
program such as Microsoft Excel. Next, we demonstrate this for Example 13.4.
EXAMPLE 13.4: Computing the Optimal Hedge Ratio (Solved Using
Microsoft Excel)
Generating the Inputs
Consider the same data as in Example 13.3. Type in the numbers and the terms exactly as follows in an
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Excel spreadsheet (see Table 13.4):
■ Type Day t in cell A1, 0 in A2, 1 in A3, 2 in A4, and so on up to 15 in cell A17. You may speed up
data entry by using the Auto Fill feature: type 0 in A2 and 1 in A3, highlight the cells A2 and A3,
and then drag down the bottom right hand corner of the cursor so that the desired values get filled
in cells A4 to A17.
■ Type S(t) in B1 and then fill out the Alloyum spot prices as given in Table 13.2: 1,215 in B2, 1,209
in B3, and so on, up to 1,194 in B17.
■ Type F(t) in C1 and then fill out the platinum futures prices F(t) as given in Table 13.2 for 16
consecutive trading days.
■ Type ΔS(t) in D1 and then type “= B3–B2” in D3, “= B4–B3” in D4, and so on. You may speed
this up by using the Auto Fill feature: type “= B3–B2” in D3 and then drag down the bottom right
hand corner of the cursor so that the desired values fill in cells D4 to D17.
■ Type ΔF(t) in E1 and then type “= C3–C2” in E3. Use Auto Fill as we have just mentioned and fill
out the desired values in cells E4 to E17.
274 CHAPTER 13: FUTURES HEDGING
TABLE 13.3: Computing h Using Microsoft Excel
A B C D E F G H
Day t S(t) F(t) ΔS(t) ΔF(t) ΔS(t) × ΔS(t) ΔF(t) ×ΔF(t) ΔS(t) ×ΔF(t)
0 1,215 1,233
1 1,209 1,236 –6 3 36 9 –18
2 1,239 1,245 30 9 900 81 270
3 1,245 1,254 6 9 36 81 54
4 1,254 1,272 9 18 81 324 162
5 1,227 1,254 –27 –18 729 324 486
6 1,230 1,248 3 –6 9 36 –18
7 1,224 1,242 –6 –6 36 36 36
8 1,239 1,260 15 18 225 324 270
9 1,251 1,263 12 3 144 9 36
10 1,227 1,254 –24 –9 576 81 216
11 1,215 1,227 –12 –27 144 729 324
12 1,209 1,221 –6 –6 36 36 36
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
13 1,203 1,230 –6 9 36 81 –54
14 1,185 1,203 –18 –27 324 729 486
15 1,194 1,209 9 6 81 36 54
Sums –21 –24 3,393 2,916 2,340
VARA 240.2571 205.5429
STDEV 15.5002 14.3368
COVAR 164.7429
CORREL 0.74138
h 0.8015 0.8015
SUMMARY 275
Computing Estimates from the Sample and the Minimum-Variance Hedge
Ratio
The entries in columns D and E are the two data series with which we work. You can do the computations
as follows (and verify that these are the same as the sample variances that you computed in Example 13.3):
■ Type in a name for the sample variance such as vara in cell A20, type “= vara(d3:d17)” in D20 and
hit return to get 240.2571, and type “= vara(e3:e17)” in E20 and hit return to get 205.5429.
■ Type stdev in A21, type “= stdev(D3:D17)” in D21 and hit return to get 15.50023, and type “=
stdev(e3:e17)” in E21 and hit return to get 14.33677.
■ As Excel’s covariance formula is given for the population, you need to multiply the estima-
tor by n/(n – 1) to get the sample estimate. Accordingly, type covar in A22, and type
“= COVAR(D3:D17,E3:E17)*(15/14)” in D22 and hit return to get 164.7429.
■ Type correl in A23 and “= correl(d3:d17,e3:e17)” in D23 and hit return to get 0.74134.
■ Type h in A24 and “= d22/e20” in D24 and hit return to get 0.801501. This is the Equation 13.6 given
in Result 13.1. You may also get h via the second result in Equation 13.6 by typing “= d23*d21/e21”
in E24 and hitting return—it’s the same answer.
13.7 Summary
1. Most futures contracts are traded to hedge spot commodity price risk. Producers
may set up a long hedge (or a buying hedge) to fix the buying price for an input
and a short hedge (or a selling hedge) to fix the selling price of an output.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
2. As with any other derivatives, the costs and the benefits must be carefully weighed
before hedging a spot commodity with futures. They involve direct and indirect
costs: the brokers must be paid, and a trade may fetch a bad price. But hedging also
has many potential benefits: it can stretch the marketing period, protect inventory
value, permit forward pricing of products, reduce the risk of default and financial
distress costs, and perhaps facilitate taking advantage of a tax loss or tax credit.
3. A perfect hedge completely eliminates price risk. It happens when (1) the futures
is written on the commodity being hedged and (2) the contract matures when you
are planning to lift the hedge. Generally, there is basis risk. The basis is defined as
the difference between the spot price and the futures price.
4. For a seller of a futures contract who hedges output price risk, the effective selling
price is (new basis + old futures price). For a buyer of a futures contract who
hedges input price risk, the effective buying price is the negative of (new basis +
old futures price).
5. We can use a statistical model, assuming past price patterns repeat themselves,
to set up a risk-minimization hedge. To minimize the risk of a long portfolio
of the spot commodity, sell short q = (n/f )h contracts, where n is the size of
the spot position, f is the contract size, and h is the optimal hedge ratio (or
276 CHAPTER 13: FUTURES HEDGING
minimum-variance hedge ratio) defined as covS,F /varF , where covS,F is the sample
covariance between the change in the spot price (ΔS) and the change in the futures
price (ΔF) and varF is the variance of the change in the futures price.
13.8 Appendix
Deriving the Minimum-Variance Hedge Ratio (h)
Suppose you are planning to sell n units of a spot commodity at some future date and
decide to hedge this exposure by selling short q futures contracts today.
■ Rewrite the change in the portfolio value between today (time 0) and some future
date (time t) as
ΔPortfolio
= (n × Change in spot price) − (q × f × Change in futures price)
(13.8)
= n [S (t) − S (0)] − q f [F (t) − F (0)]
= nΔS − q fΔF
where f is the number of units of the futures contract, S(t) is the spot price at
date t, F(t) is the futures price at date t, and Δ denotes a price change.
■ Select q so as to minimize the change in the portfolio value. First, use statistics to
compute the variance of the portfolio’s price change:
variance (aX − bY) ≡ a2 variance (X) +
(13.9)
b2 variance (Y) − 2ab [covariance (X, Y)]
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where a and b are constants and X and Y are random variables. As n, q, and f are
constant parameters, set a = n, b = q f, X = ΔS, and Y = ΔF in Equation 13.9 to get
2
var (ΔPortfolio) = n2 vars + (q f) varF − 2nq f × covS,F (13.10)
where varS is the variance of change in the spot price (ΔS), varF is the variance of
change in the futures price (ΔF ), and covS,F is the sample covariance between the
change in the spot and futures prices.
■ Using calculus, we minimize Equation 13.10 by taking the partial derivative with
respect to q and setting it equal to zero:
𝜕
[var (ΔPortfolio)] = 2q f 2 varF − 2nf covS,F = 0
𝜕q
The second partial derivative is a positive number, which indicates that this
minimizes the equation. Rearrange terms to get the number of futures contracts
to sell short (or go long in case you are setting up a buying hedge):
q = (n/f) h (13.11)
where the optimal hedge ratio h equals (covS,F /varF ).
APPENDIX 277
Computing the Minimum-Variance Hedge Ratio (h)
Example 13.4 shows how to input the values into a business calculator or a computer
to determine the minimum variance hedge ratio (h). Here we show how to directly
calculate h.
Statistical Approach
Start by deciding how frequently and over what time interval you collect your data.
Standard practice is to fix a time interval (daily, weekly, or monthly) and use the end
of the interval’s settlement prices.
DATA SERIES
■ To cross-hedge Alloyum spot with platinum futures, we need their prices as basic
inputs. These are reported in Table 13.5, whose first three columns are as follows:
- The first column is labeled Day t, and it keeps track of the days. There are
(T + 1) = 16 observations corresponding to sixteen consecutive trading days,
where t “runs” from day 0 to day t = 15.
- Columns 2 and 3 record the Alloyum spot price S(t) and the platinum futures
price F(t), respectively, for these sixteen consecutive trading days. All prices are
reported for one unit of the respective commodity.
PRICE DIFFERENCES
■ Compute price differences for each price series.
- Column 4 reports changes in the value for Alloyum spot. Denote the spot price
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
change from date (t – 1) to date t as ΔS(t) ≡ S(t) – S(t – 1). Label column 4 as
x(t) for convenience. For example, when t = 6,
x (6) ≡ ΔS (6) − S (6) – S (5) = 1, 230 − 1, 227 = 3
- Column 5 does the same for platinum futures. Label this as y(t). For example,
y (6) ≡ ΔF (6) = ΔF (t) ≡ F (t) − F (t − 1)
= F (6) − F (5) = 1, 248 − 1, 254 = −6
■ Notice that when computing the “difference,” you lose the first observation. We
now have T = 15 observations. We denote them by t, where t = 1, 2, . . . , 15.
■ We can now forget our second and third columns. We use the fourth and fifth
columns to compute the minimum-variance hedge ratio h.
278 CHAPTER 13: FUTURES HEDGING
TABLE 13.4: Price Data and Calculations for the Minimum-Variance Hedge
Ratio (h)
Day t Alloyum Platinum ΔS(t) ≡ S(t) – ΔF(t) ≡ F(t) – [ΔS(t)]2 ≡ [ΔF(t)]2 ≡ ΔS(t) × ΔF(t) ≡
Spot S(t) Futures S(t – 1) ≡ x(t) F(t – 1) ≡ y(t) x(t)2 y(t)2 x(t)y(t)
F(t)
0 1,215 1,233
1 1,209 1,236 –6 3 36 9 –18
2 1,239 1,245 30 9 900 81 270
3 1,245 1,254 6 9 36 81 54
4 1,254 1,272 9 18 81 324 162
5 1,227 1,254 –27 –18 729 324 486
6 1,230 1,248 3 –6 9 36 –18
7 1,224 1,242 –6 –6 36 36 36
8 1,239 1,260 15 18 225 324 270
9 1,251 1,263 12 3 144 9 36
10 1,227 1,254 –24 –9 576 81 216
11 1,215 1,227 –12 –27 144 729 324
12 1,209 1,221 –6 –6 36 36 36
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
13 1,203 1,230 –6 9 36 81 –54
14 1,185 1,203 –18 –27 324 729 486
15 1,194 1,209 9 6 81 36 54
Sums Σx = – 21 Σy = – 24 Σx2 = 3,393 Σy2 = 2,916 Σxy = 2,340
Some values useful for the computations follow:
T
x (t) = −21
∑
t=1
15
y (t) = −24
∑
t=1
15
x(t)2 = 3, 393
∑
t=1
15
y2 = 2, 916
∑
t=1
15
x (t) y (t) = 2, 340
∑
t=1
APPENDIX 279
COMPUTATION-SIMPLIFYING TECHNIQUES
- Three more columns are introduced to help simplify our calculations.
- Columns 6 and 7 record the square of the price changes reported in Columns 4
and 5, respectively. Finally, Column 7 reports the product of the values in Columns
5 and 6. For example, when t = 6, we get
[x(6)]2 = 32 = 9y (6)
[ y(6) ] 2 = (−6)2 = 36
x (6) y (6) = −18
STATISTICAL ESTIMATES
■ To compute h from historical data, modify our formulas for covariance and variance
by replacing the T with (T - 1) in the denominator to get an unbiased estimate.
■ We get the following parameter estimates to help us compute h (see Table 13.6):
sample covariance, covS,F = 164.7429
sample variance of ∆S, varS = 240.2571
sample variance of ∆F, varF = 205.5429
sample standard deviations, sdS = 15.5002 and sdF = 14.3367
correlation coefficient between ΔS and ΔF, corrS,F = 0.7413
■ These estimates are used to compute the minimum-variance hedge ratio h. The
first part of Equation 13.7 of Result 13.1 gives
covS,F 164.7429
h= = = 0.8015
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
(13.12)
varF 205.5429
while the second part verifies the same result:
sdS 15.5002
h= corr = 0.7413 = 0.8015 (13.13)
sdF S,F 14.3367
■ If you are estimating these values over a longer period, then you run the risk that
a futures contract may mature and no longer trade. Suppose you are hedging with
the “nearest maturity” futures contract, whose price changes are likely to have the
best correlation with the spot price changes. Suppose this contract stops trading
on Day 12. Select the futures contract that is going to mature next, get its futures
prices for Days 12 and 13, and write the price difference ΔF(13) ≡ F(13) – F(12)
in the first table in the cell corresponding to Day 13’s price difference.
280 CHAPTER 13: FUTURES HEDGING
TABLE 13.6: Statistical Estimates and Computation of the Optimal Hedge
Ratio h
Statistical Formula Computations Estimated Value
Estimate
T
1
a. Sample x(t) –21/15 –1.4000
mean, x T∑t =1
15
1
b. Sample y(t) –24/15 –1.6000
mean, y 15 ∑
t =1
T
1 1 −21 −24
c. Sample [x(t) − x] × [y(t) − y] 2, 340 − 15 164.7429
covariance of T−1 ∑
t =1
14 [ ( 15 ) ( 15 )]
15
ΔS and ΔF, 1
covS,F = x(t)y(t) − 15xy
14 {[∑
t =1 ] }
T 2
1 1 −21
d. Sample [x(t) − x]2 3, 393 − 15 240.2571
variance of T−1 ∑
t =1
14 [ ( 15 ) ]
15
ΔS, varS 1 2
= x(t)2 − 15x
14 [∑
t =1 ]
T 2
1 1 −24
e. Sample [y(t) − y]2 2, 916 − 15 205.5429
variance of T−1 ∑
t =1
14 [ ( 15 ) ]
ΔF, varF
f. Sample +√varS +√240.2571 15.5002
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
standard
deviation of
ΔS, sdS
g. Sample +√varF +√205.5429 14.3367
standard
deviation of
Δ F, sdF
covS,F 164.7429
h. Correlation 0.7413
coefficient sdS × sdF 15.5002 × 14.3367
between ΔS
and ΔF, corrS,F
covS,F 164.7429
i. Minimum- 0.8015
variance hedge varF 205.5429
ratio, h
sdS 15.5002
j. h (alternate corrS,F 0.7413 0.8015
sdF ( 14.3367 )
computation)
CASES 281
Econometric Approach: A Linear Regression Model
Econometrics is the branch of economics that studies relationships between
economic variables using mathematics and statistics. A staple of econometrics is the
linear regression model,6 in which a dependent variable may be determined from
the values of one or more independent variables, under the assumption that these
variables are linearly related. In this case, the model takes the form
ΔS (t) = 𝛼 + 𝛽ΔF (t) + u (t) (13.14)
where 𝛼 and 𝛽 are unknown parameters and u(t) are independent and identically
distributed error terms with an expected value of 0 and a constant variance.
Next you have to estimate these parameters. The most popular approach is the
least squares method, in which 𝛼 and 𝛽 are chosen so that the sum of squared
errors Σt [ΔS(t) – 𝛼 – 𝛽ΔF(t)]2 is minimized. Interestingly, the value of 𝛽 obtained
is identical to the hedge ratio h in Result 13.1. This makes life easy. Estimate the
regression Equation 13.4 by hand or by using a standard statistical package like
MATHEMATICA, MINITAB, SAS, or TSP, and the estimated value of 𝛽 will
give you the hedge ratio. Spreadsheet programs like Microsoft Excel can also run
regressions.
13.9 Cases
Lufthansa: To Hedge or Not to Hedge . . . (Richard Ivey School of Business
Foundation Case 900N22-PDF-ENG-PDF-ENG, Harvard Business Publishing).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
A short case that explores the costs and benefits of hedging and the derivatives that
may be used for hedging purposes.
Enron Gas Services (Harvard Business School Case 294076-PDF-ENG). The case
considers the risks and opportunities of selling a variety of natural gas derivatives
by a financial services subsidiary of the largest US integrated natural gas firm.
Aspen Technology Inc.: Currency Hedging Review (Harvard Business School
Case 296027-PDF-ENG). The case examines how a small, young firm’s business
strategy creates currency exposure and how one can manage such risks.
6
See standard econometrics textbooks like Amemiya (1994) or Johnston and DiNardo (1997) for a
discussion of the linear regression model.
282 CHAPTER 13: FUTURES HEDGING
13.10 Questions and Problems
13.1. Define a long hedge and a short hedge and give examples of each kind of
hedge.
13.2. What are the benefits of a corporation’s hedging? In your answer, explain
why the corporation and not the corporation’s equity holders must do the
hedging. Are there any costs to a corporation’s hedging?
13.3. What is the difference between a perfect hedge and a cross hedge? Give
examples to clarify your answer.
13.4. If one cannot create a perfect hedge, what are the alternatives? Give an
example to explain your answer.
13.5. What is the basis for a futures contract? What happens to the basis on the
delivery date?
13.6. There is no futures contract on aviation fuel. Combo Air Inc. has to buy
3 million gallons of aviation fuel in three months. Suppose you are in charge
of Combo Air’s hedging activities. You gather the following data:
TABLE 13.7: Jet Fuel Cash Prices vs. Near
Month Energy Futures Prices:
Correlations of Price Changes
1986–88 1986 1987 1988
Heating oil futures 0.54 0.76 0.89 0.32
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Gasoline futures 0.41 0.74 0.73 0.19
Crude oil futures 0.45 0.70 0.72 0.25
a. Which energy futures contract will you choose for hedging jet fuel
purchase?
b. Is this a long hedge or a short hedge?
13.7. The variance of monthly changes in the spot price of live cattle is (in cents per
pound) 1.5. The variance of monthly changes in the futures price of live cattle
for the April contract is 2. The correlation between these two price changes
is 0.8. Today is March 11. The beef producer is committed to purchasing four
hundred thousand pounds of live cattle on April 15. The producer wants to
use the April cattle futures contract to hedge its risk. What strategy should
the beef producer follow? (The contract size is forty thousand pounds.)
QUESTIONS AND PROBLEMS 283
13.8. Are hedging with forwards and futures contracts the same, or are there
different risks to be considered when using these two contracts? Explain your
answer.
13.9. When you hedge a commodity’s price risk using a futures contract, give an
example where the counterparty is also hedging. Give an example where the
counterparty is speculating.
13.10. Kellogg will buy 2 million bushels of oats in two months. Kellogg finds that
the ratio of the standard deviation of change in spot and futures prices over a
two-month period for oats is 0.83 and the coefficient of correlation between
the two-month change in price of oats and the two-month change in its
futures price is 0.7.
a Find the optimal hedge ratio for Kellogg.
b How many contracts do they need to hedge their position? (The size of
each oats contract is five thousand bushels; oats trades in the CME Group.)
13.11. Explain why hedging is like buying an insurance policy. To buy an insurance
policy, you need to pay a premium; what is the corresponding premium in
hedging? Give an example to clarify your answer.
13.12. Suppose that after you graduate, you plan to be a stock analyst for a major
financial institution. You know that if the stock market increases in value,
you will get a job with a good salary. If the stock market declines, you will
get a job, but the salary will be lower. How can you hedge your salary risk
using futures contracts? Is this a perfect or imperfect hedge?
13.13. Your company will buy tungsten for making electric light filaments in the
next three to six months. Suppose there are no futures on tungsten. How
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
would you hedge this risk? (Discuss the type of hedge, general hedging
approach, and guidelines that you would like to follow.)
13.14. You are the owner of a car rental business. If gasoline prices increase, your
car rental revenues will decline. How can you hedge your car rental revenue
risk using futures contracts? Is this a perfect or imperfect hedge?
13.15. What commodity price risk does Southwest Airlines hedge, and why? Has it
always been successful in its hedging program?
13.16. What is risk-minimizing hedging? Briefly outline how you would set up a
risk-minimizing hedge. Is a risk-minimizing hedge a perfect or imperfect
hedge? Explain your answer.
284 CHAPTER 13: FUTURES HEDGING
13.17. Canadian American Gold Inc. (CAG) has half its gold production from
mines located in Canada, while the other half is from those in the United
States. CAG uses a quarter of its production for making gold jewelry sold at
a fixed price through stores in the two nations, and the rest is sold on the
world market, where the gold price is determined in US dollars. Canadian
profits are repatriated to the United States, where CAG’s headquarters are
located. CAG’s chief executive officer wants to use futures contracts to hedge
the entire production of ten thousand ounces of gold and as many other
transactions as possible. He communicates his desire to you but seeks your
opinion one last time before the orders go out. Devise a sensible hedging
strategy that would still be in line with the CEO’s wishes (assume x is the
quantity used for making gold jewelry in the United States).
13.18. The spot price of gold today is $1,505 per ounce, and the futures price for a
contract maturing in seven months is $1,548 per ounce. Suppose CAG puts
on a futures hedge today and lifts the hedge after five months. What is the
futures price five months from now? Assume a zero basis in your answer.
13.19. Suppose that Jewelry Company is planning to sell twenty thousand ounces of
platinum at some future date. The standard deviation of changes in the futures
price per ounce sdF is 12.86, that for changes in the spot price per ounce
sdS is 14.38, and the correlation coefficient between the spot and futures price
changes corrS,F is 0.80.
a. Compute the optimal hedge ratio for Jewelry Co.
b. How many contracts do they need to hedge their position? (The contract
size is fifty ounces.)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
c. Will this be a buying or a selling hedge?
QUESTIONS AND PROBLEMS 285
13.20. (Microsoft Excel) Given the following data, compute the hedge ratio for a
risk minimizing hedge.
Day t Alloyum Spot S(t) Platinum Futures F(t)
0 1,233 1,245
1 1,219 1,256
2 1,118 1,130
3 1,246 1,264
4 1,250 1,280
5 1,219 1,223
6 1,230 1,248
7 1,227 1,280
8 1,249 1,260
9 1,225 1,289
10 1,227 1,254
11 1,223 1,255
12 1,211 1,223
13 1,203 1,267
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
14 1,189 1,213
15 1,199 1,219
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
III
Options
CHAPTER 14
Options Markets
and Trading
CHAPTER 15
Option Trading
Strategies
CHAPTER 17
CHAPTER 16
Single-Period
Option Relations
Binomial Model
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
CHAPTER 19
CHAPTER 18
The Black–Scholes– Multiperiod
Merton Model Binomial Model
CHAPTER 20
Using the Black–
Scholes–Merton Model
14
Options Markets and
Trading
14.1 Introduction Dividends and Stock Splits
14.2 Exchange-Traded Options 14.5 Options Trading, Exercising,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and the Expiration Process
14.3 A History of Options
The Trading Process
Early Trading of Options
Exercising the Option
The Year 1973: The Watershed
Year and After The Expiration Process
EXTENSION 14.1 Options on 14.6 Options Price Quotes
Futures
14.7 Regulation and Manipulation
14.4 Option Contract Features in Options Markets
Maturity Dates 14.8 Summary
Strike Prices
14.9 Cases
Margin Requirements and
Position Limits 14.10 Questions and Problems
EXCHANGE-TRADED OPTIONS 289
14.1 Introduction
Believe it or not, governments trade derivatives. Distinguished economist Lawrence
Summers recounted an interesting experience when he was deputy secretary of the
US Treasury (see Summers 1999, p. 3):
Any doubt I might have had about the globalization of economic thinking was
shattered when I met with Chinese Premier Zhu Rhongji in early 1997 in the
same pavilion where Chairman Mao had received foreign visitors. After being
offered a Diet Coke, I was asked a variety of searching questions about the
possible use of put options [emphasis added] in defending a currency, and how
they might best be structured.
A communist leader dabbling in derivatives!
A more recent example of derivative securities being used to support a govern-
ment’s fiscal and monetary policy occurred in the United States during the 2007 credit
crisis. To reduce the cost of the capital infusion by the US Treasury into financial
institutions under the Troubled Asset Relief Program (TARP), warrants—a type
of call option—were used. These warrants, issued to the US Treasury by the banks
receiving the TARP funds, provided additional upside potential to the US Treasury.
After the banks recovered, the US Treasury made significant profits on its warrant
positions, reducing the cost of the TARP program. Exactly as hoped!
This chapter continues our discussion of options and their markets started in
Chapter 5. First, we review the history of options trading, which is divided into
two phases around the watershed year of 1973. Next we describe some features
of options contracts. We end the chapter with regulatory issues, including market
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
manipulation. Although our discussions primarily revolve around equity options,
many of the features discussed here are similar to other options encountered later
in this book and elsewhere in life.
14.2 Exchange-Traded Options
Why trade options? Unlike futures, options have asymmetric payoffs, a special feature
that allows users to carve out richer payoff structures than what’s otherwise obtainable.
This makes options valuable tools for hedging, speculation, and risk management.
Because the holder is rewarded when the underlying moves favorably but has no
loss when it moves adversely, options are similar to insurance contracts and require
a premium payment at their onset. But regular use of options can get expensive,
and one must carefully weigh the costs and benefits before trading an option. We
skip discussing the applications and uses of options because of its similarity with our
earlier discussion of the applications and uses of derivatives (see section 1.3), swaps
(see section 7.5), and forwards and futures (see section 8.2).
Although options have traded over the counter (OTC) for hundreds of years,
the market really took off when exchange-traded options were introduced on the
290 CHAPTER 14: OPTIONS MARKETS AND TRADING
Chicago Board Options Exchange (CBOE) in 1973. The reasons are twofold. First,
an exchange creates a central marketplace: an orderly, efficient, and liquid market with
low transaction costs where traders can enter or exit positions with ease. Second, a
clearinghouse for the exchange guarantees contract performance and makes trades
virtually free from credit risk. The resulting popularity of plain vanilla exchange-
traded options paved the way for the development of the now gigantic OTC market
for trading customized, complex options. The causation did not happen in reverse.
Although OTC options have traded for centuries, their market remained small and
inefficient, fraught with credit and legal risks, never catching on with the investing
public until after exchange traded options became popular. Why did it happen this
way? To find out, read on.
14.3 A History of Options
“History has many cunning passages, contrived corridors,” wrote T. S. Eliot in the
poem “Gerontion.” This describes the history of options. Option-like contracts arise
naturally in economic life. One enters a call-like agreement when you pay a deposit
in advance of buying a house with the provision that your failure to complete the
purchase forfeits this amount; you enter into a put-like contract when you buy
insurance by paying a premium. The term premium also denotes the price of an option,
which reinforces the close link between options and insurance.
Despite these similarities, the options and insurance markets went their separate
ways. Insurance contracts have enjoyed popular support for centuries, but options’
acceptance has waxed and waned over the years. Options and futures traded side by
side in Amsterdam during the Dutch golden age of the seventeenth century. They
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
have traded for over two hundred years in small OTC markets in London, New
York, and several cities of Continental Europe. Options traded as bilateral contracts
that were loaded with credit, legal, and liquidity risks. They were viewed as tools for
speculation that served no worthwhile economic purpose. In fact, the US Securities
and Exchange Commission (SEC; created after the great stock market crash of 1929)
repeatedly tried to outlaw options. The market ebbed and flowed depending on the
participants. It expanded when big financiers wrote options because their reputation
and deep pockets assuaged the buyers about credit risk.
The opening of the CBOE in 1973, support from academic economists, and
the Black–Scholes–Merton pricing model helped options reach the mainstream and
remove their stigma of gambling. The information technology revolution of the
1990s ushered in the era of electronic exchanges and web-based trading. Insurance
and options markets have converged again. Nowadays options are viewed as risk
management tools that “complete the markets” by allowing traders to carve out
different portfolio payoffs. They trade on many exchanges around the globe, and
traders can transact through a single electronic platform.
A HISTORY OF OPTIONS 291
Despite this convergence, the history of options trading is sufficiently different
from that of futures trading to merit further discussion. We divide the history into
two phases:
1. Early trading of options. Option-like contracts have traded since antiquity, with
options trading in organized OTC markets in Amsterdam, London, and New
York City in past centuries.
2. The year 1973: the watershed year and after. The Nobel Prize–winning Black–
Scholes–Merton model gave traders a vital tool for pricing and hedging options.
The opening of the CBOE ended the old, inefficient OTC markets and brought
numerous innovations. Post-1973 saw the development of a variety of contracts,
the opening of new exchanges, the automation of trading, and the consolidation
and linkages among exchanges. This story is akin to the history of futures trading
discussed in Chapter 8. That’s hardly surprising because economic, political,
regulatory, and technological changes similarly affected the options markets.
Early Trading of Options
Options and futures traded in Western Europe over four centuries ago. Before
exploring that history, however, we mention that the first recorded option usage
occurs in a story telling how Thales the Milesian (624–545 BC) purchased an option
to buy the use of olive presses in the next harvest year, which he was sure would
yield a bountiful harvest. His prediction was correct, and he profited handsomely
from exercising his option.1
Let us move forward two millennia, and arrive in the Netherlands. Gelderblom and
Jonker (2005) tell the story of how sophisticated financial markets developed in West-
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
ern Europe, which, nonetheless, were vexed by problems that continue to afflict us to
this day (see Table 14.1 for a timeline of options trading). Records from Amsterdam
document how grain dealers in the city used both futures and options in the 1550s and
the following decades. Fearing that speculators could use derivatives to manipulate
staple food prices and cause social unrest, the authorities repeatedly banned forward
trading. However, the market did not disappear; it simply went underground and was
confined to the insiders. Subsequently, the market again became public, expanding
to include forwards, futures, and call and put options on equities. The 1630s saw an
extraordinary speculative frenzy over rare tulip bulbs, with futures and options helping
hedgers and the speculating public. We discuss this episode later in this chapter.
A small options market for sophisticated traders continued to operate in Continen-
tal Europe, London, and New York for over two centuries (see Malkiel and Quandt
1969). For example, during the 1690s, a well-organized options market existed in
London, and it continued to exist despite the ban on options and futures promulgated
under Barnard’s Act of 1733. In 1821, the London Stock Exchange Committee
proposed a rule that would forbid members from dealing in options. Loud protest
came from a large number of members; they even raised money to build a rival
stock exchange building. The ban was never implemented, option trading continued
1
This story is from Jowett’s translation of Chapter 11 of Book 1 of Aristotle’s Politics (etext.library.
adelaide.edu.au/a/aristotle/a8po/).
292 CHAPTER 14: OPTIONS MARKETS AND TRADING
TABLE 14.1: Early History of Options Trading
Year Development
600 bc onward Thales of Miletus traded optionlike contracts
1550–1650 Birth of options and futures trading in Amsterdam
1630s Futures and options traded on tulips in the Netherlands
1860–1900 Russell Sage extensively traded calls, puts, spreads, and straddles
1910 The Put and Call Brokers and Dealers Association was founded
(restricted but) uninterrupted, and Barnard’s Act itself was repealed in 1860. The
European options market remained small because many prominent firms refused to
participate.
A New York City options market became active during the last four decades
of the nineteenth century, when Russell Sage operated as a financier, securities
trader, options writer, and broker on a vast scale. Russell Sage’s financing and trading
activities relate to several important features of option markets:
1. The importance of collateral. An article titled “Russell Sage” (New York Times, August
5, 1903), attributed Sage’s success to his being “an excellent judge of collateral.” It
noted that “crises and panics have passed him unharmed, and when others were
eager to borrow he was as eager to lend—if the collateral was satisfactory.”
2. The invention of straddles and spreads. Sarnoff (1965) gave Sage credit for inventing
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the spread and the straddle and suggested that his extensive dealing in calls and
puts as well as these two creations led to the development of the “modern system
of selling stock-option contracts.” These activities earned him the nickname “Old
Straddle.”
3. Dividend adjustments. A dividend lowers the value of an asset; consequently, a
dividend lowers the value of a call and raises the price of a put. A client once took
the aged Sage to court in a dispute about dividend adjustments and an option’s
value (see “Dividends and Puts,” New York Times, April 6, 1887). Today options
exchanges have well-defined policies for these adjustments (see section 14.4 and
Example 14.4 for further discussions).
4. Put–call parity. Sage devised conversions, where he created synthetic loans by using
the put–call parity principle and effectively charged interest rates higher than what
the usury laws would allow.
The OTC option market in the nineteenth and the early twentieth centuries wasn’t
too different from that in historic Amsterdam. Options contracts were written in
printed forms where traders filled in the details and had them notarized. The dealers
sold options by word of mouth or by advertising in the newspapers. The Put and
Call Brokers and Dealers Association was established in 1910 to organize options
trading in the United States.
A HISTORY OF OPTIONS 293
The government continued to view options trading with suspicion for good
reasons. Particularly in the early twentieth century, speculators would often use
options to manipulate stock prices. After the stock market crash of 1929, the US
Congress sought to ban options trading, but did not do so. Options trading volume (in
terms of shares), which was about half of 1 percent of the New York Stock Exchange
(NYSE) stock trading volume in 1937, rose to around 1 percent in 1955 and hovered
around that number during the 1960s; however, trading remained anemic, and the
stigma of gambling prevented the options market from enjoying vigorous growth.
The Year 1973: The Watershed Year and After
This landscape began to change in 1973 with the opening of the CBOE and the
publication of the Black–Scholes–Merton options pricing model.
THE OPENING OF THE CHICAGO BOARD OPTIONS EXCHANGE
The CBOE opened its doors on April 26, 1973. Just 911 call options contracts traded
on sixteen stocks on the inaugural day. The trading floor was set up in the previous
Chicago Board of Trade members’ smoking lounge. The daily trading volume at the
end of the first month exceeded the average daily volume in the old, OTC options
market.
Despite its ups and downs, a remarkable expansion of options markets followed
the opening of the CBOE—see Table 14.2 for some milestones in this extraordinary
march. The mid-1970s saw the introduction of options trading on several US
exchanges and in other nations. In 1975, the clearing facility of the CBOE was
reorganized as the Options Clearing Corporation (OCC), which functions as the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
common clearinghouse for most exchange-traded listed options in the United States
and has virtually eliminated credit risk. The same year, computerized price reporting
was also introduced in the CBOE.
During the three-year period 1977–80, the SEC put a moratorium on additional
listings pending a review of the options industry’s growth. The stock market crash in
October 1987 led to a temporary waning of interest in options. Still the innovations
continued, and a variety of new contracts were introduced. The IT revolution of
the 1990s led to several key developments: automation lowered transaction costs,
many exchanges introduced electronic trading, greater links were established among
exchanges, and a wave of mergers and consolidations transformed the markets.
A VARIETY OF OPTIONS Most options varieties were introduced during the
1970s and 1980s. In 1977, put options started trading at the CBOE. The year 1982
saw the beginning of currency options trading on the Philadelphia Stock Exchange
and Treasury bond futures options on the Chicago Board of Trade, which brought
futures exchanges into the options business (see Extension 14.1). The following year,
cash-settled options on broad-based stock indexes (including options on the Standard
and Poor’s [S&P] 500 index) were launched at the CBOE. As with futures, cash
settlement allowed the creation of new kinds of options with rich payoff structures.
294 CHAPTER 14: OPTIONS MARKETS AND TRADING
TABLE 14.2: Some Milestones in Options Markets, 1973 and After
Year Development
1973 First options exchange, the Chicago Board Options Exchange (CBOE), was founded,
with call options trading on sixteen underlying stocks
1975 Call options began trading in the American Stock Exchange, the Philadelphia Stock
Exchange, and the Montreal Stock Exchange in Canada; the Options Clearing
Corporation (OCC) was established for exchange-traded options in the United States
1976 Calls began trading on the Pacific Exchange and the Australian Options Market
1977 Put options were introduced at the CBOE
1977–80 The Securities and Exchange Commission put a moratorium on the options market
1978 The European Options Exchange and London Traded Options Market were established
1982 Currency options started trading at the Philadelphia Stock Exchange
1982 Options on US Treasury bond futures started trading at the CBOT
1983 Options on broad-based stock indexes started trading at the CBOE
1987 Stock Market crash in October 1987 set back the options market, from which it took
years to recover
1993 Introduction of the CBOE Volatility Index (VIX), a key measure of market expectations
of near-term volatility conveyed by Standard and Poor’s index options prices
1998 Deutsche Borse AG and Swiss Exchange started Eurex, a joint platform for trading and
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
clearing derivatives trade
2000 Euronext was founded through the merger and consolidation of several European
exchanges
2000 First all-electronic exchange in the United States, the International Securities Exchange,
was founded
By mid-1983, exchange-traded options were available for hedging and speculating
on equity, commodity, currency, interest rates, and stock indexes. Many options
were subsequently introduced in these categories. For example, the CBOE intro-
duced long-lived equity options, LEAPS (Long-Term Equity AnticiPation
Securities), with up to three years’ life in 1990, options on sector indexes in 1992,
FLEX options (which allow traders to modify some key specifications) in 1993, and
options on country indexes in 1994. On the basis of academic research by Professor
Robert Whaley, the CBOE unveiled in 1993 the VIX (CBOE Volatility Index),
which measures the market’s estimate of near-term volatility as implied by the S&P
500 stock index option prices. The VIX index was modified in 2003 from measuring
an implied volatility to that of an expected volatility. Colloquially known as the “Fear
Gauge,” VIX is a widely used barometer of investor sentiment and market volatility.
A HISTORY OF OPTIONS 295
The CBOE introduced futures contracts on VIX in 2004 and option contracts on
VIX in 2006—we discuss the VIX in chapter 20.
NEW EXCHANGES AND THE AUTOMATION OF TRADING Many
existing exchanges embraced options trading, and new options exchanges were
formed. In 1975, call options commenced trading at the American Stock
Exchange, the Philadelphia Stock Exchange, and the Montreal Stock
Exchange in Canada. In 1976, they were introduced on the Pacific Exchange
in San Francisco and in the Australian Options Market, which became the first
options market outside North America. In 1978, two of the world’s oldest exchanges
introduced options trading in Europe: the European Options Exchange was
founded by the members of the Amsterdam Stock Exchange, and the London
Traded Options Market was established as a part of the London Stock Exchange.
By the end of the decade, BOVESPA became the first exchange to introduce options
trading in Brazil.
During the 1980s, the exchanges worked on expanding the market and jockeyed
for greater market share. The global options market suffered a severe setback during
the stock market crash of October 1987, and it took over a decade for the volume
to return to pre-crash levels. As the 1990s progressed, several discernible patterns
emerged. Competition put relentless pressure to cut costs, and IT advances made
that possible. Exchanges dismantled floor trading or began as completely electronic
exchanges. With their commitment to floor trading, US exchanges kept falling
behind. Wake-up calls came when Eurex and Euronext were founded in 1998 and
2000, respectively. Established in 2000 as an all-electronic options exchange, the
International Securities Exchange (now part of Nasdaq Inc.) became the world’s
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
largest equity options exchange. Stiff competition came from the OTC derivatives
markets, which grew faster than the exchanges. Fed chairman Alan Greenspan (recall
Chapter 10) used this fact to justify less regulation and a hands-off policy toward the
OTC markets. The US exchanges regrouped and rethought their options (no pun
intended!) and strategies. Seeing that electronic trading was the wave of the future
for trading plain vanilla securities, they acquired or merged with organizations that
focused on electronic trading or developed their own automated trading systems.
EXTENSION 14.1: Options on Futures
Why trade options on futures? Recall the observation from Chapter 1 that trading gravitates to the most liquid markets
with minimum transaction costs. As the stock market is very liquid with minimal transaction costs, it’s no surprise
that we have bustling markets for equity options. But storage costs and convenience yields for many commodities
in the spot market make it easier for traders to hedge and speculate on commodities in the more liquid futures
markets. Consequently, it’s natural for options on futures to trade on such commodities.
For options on spot (also called spot options), the spot commodity is obtained when the option is exercised.
By contrast, exercise of a futures option gives the holder a position in the underlying futures. In the United
296 CHAPTER 14: OPTIONS MARKETS AND TRADING
States, active futures as well as futures options markets exist for diverse commodities such as cattle, cotton,
corn, crude oil, gold, and silver. These options trade on the same exchange where the futures contracts trade.
Simultaneous trading of futures and futures options helps the price discovery process.
The example below shows how a futures option works. Notice that both spot options and futures options
behave similarly in response to a movement in the underlying.
EXT. 14.1 EX. 1: Call and Put Options on Futures
■ Today is January 1. Suppose the April gold futures price is $1,515 per ounce. Futures options with a contract
size of one hundred ounces and for $1,490, $1,495, $1,500, $1,510, $1,520, $1,530, and several other strike
prices (per ounce) trade in the New York Mercantile Exchange of the CME Group.
■ Consider a call option on a commodity futures. If you exercise this option, you will receive a long
position in the underlying futures contract and a cash payment equal to the excess of the futures price over
the strike price. For example, if you exercise an April 1,500 call, you end up long an April futures and receive a
payment of (1,515 − 1,500) = $15 per ounce, $1,500 in all. If the payment is negative, do not exercise—enter
the futures at zero cost in the market instead.
■ The exercise of a put option on a commodity futures puts the holder into a short position in the
underlying futures and brings in a cash amount equal to the strike price minus the futures settlement price.
For example, if you exercise an April 1,520 put, you end up being short April futures and receive a payment
of (1,520 − 1,515) = $5 per ounce.
■ Obviously, the long has to pay premiums for these options. A modified version of the Black–Scholes–Merton
model can be used to price futures options. More advanced models like the Heath–Jarrow–Morton model
(developed in 1987, published in 1992) can also be used for this purpose (see Amin and Jarrow 1991).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Notice that if you replace the futures price F(T) with the stock price S(T), then futures options payoffs are
similar to equity options payoffs. For example, consider a call futures option and a regular equity call option on
a stock with the same strike price K. If you exercise the equity call, your payoff is [S(T) − K]. If you exercise
the call futures option, you receive [F (T) − K] and a long futures position on the stock. As the futures contract
has a zero value, it can be liquidated at no cost. Exercise of a put futures option leads to a short position in the
underlying futures plus the strike minus the futures price, [K − F (T)], and a similar argument holds.
14.4 Option Contract Features
Let us examine the contract specifications for CBOE equity options.2 The CBOE
regular trading session for options is from 8:30 am to 3:00 pm Central Time (Chicago
time). CBOE equity options are American, and they can be exercised any time from
purchase until the option’s last trading day, which is the third Friday of the expiration
month. If the third Friday is a legal holiday, then the option expires the day before.
Each option is for one hundred shares of the underlying stock. If exercised, the stock
share is delivered three business days later.
2
See www.cboe.com.
OPTION CONTRACT FEATURES 297
The minimal tick size is $0.05 for options trading below $3 and $0.10 for options
trading above. However, options series are quoted in pennies ($0.01) for options
prices below $3 and in nickels ($0.05) for higher options prices.
Maturity Dates
For a particular stock, options belong to an assigned quarterly cycle, either the
January, February, or March cycle. With respect to a particular cycle, options maturities
are listed for “two near-term months plus two additional months from the January, February
or March quarterly cycles (www.cboe.com).” An example explains this procedure.
EXAMPLE 14.1: Determining the Expiration Months
■ The January cycle has the months January, April, July, and October. The February cycle has February,
May, and so on. Each stock approved for options trading belongs to one of the cycles. For example,
IBM has a January cycle, Hewlett Packard has a February cycle, and Walmart Stores has a March cycle.
■ Consider the expiration months for traded options on IBM starting January 1.
- On January 1, IBM has options expiring in January, February (the two near-term months), April,
and July (the next two months from the quarterly cycle).
- After the third Friday of January, one can trade IBM options maturing in February, March, April,
and July.
- After the third Friday of February, one can trade options maturing in March, April, July, and
October.
■ Different kinds of options have different expiration dates. For example, LEAPS are issued with three-
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
year maturities and expire on the third Friday of the expiration year. Quarterly options (or the
“quarterlies”), which can be written on some indexes and exchange-traded funds, expire on the
last business day of each calendar quarter. Weeklys, which can be traded on many stocks, ETFs, and
indexes, typically expire on Friday of each week.
Strike Prices
When an option with a particular maturity starts trading, strike prices for in-, at-,
and out-of-the-money are listed. If the stock price is over $200, the strikes are issued
$10 apart. Between $200 and $25, the strikes are issued $5 apart, and below $25,
only $2.50 apart. The CBOE has also introduced $1 strike price intervals and $2.50
strike price intervals, under certain terms and conditions, for many actively traded
stocks. Strike prices are adjusted for stock dividends and stock splits but not for cash
dividends. If the stock price moves significantly before the option expires, then new
strikes are listed to maintain the balance with respect to in-, at-, and out-of-the-
money. The next example illustrates this process.
298 CHAPTER 14: OPTIONS MARKETS AND TRADING
EXAMPLE 14.2: Determining the Strike Prices
■ Suppose the stock opens at $21 on Monday following the third Friday of the month, when new
long-maturity options are listed. This is the first trading day after the shortest maturity option expires.
Newly listed options (both calls and puts) have strike prices of $20 and $22.50, immediately above and
below the current stock price. Additional in- and out-of-the-money strikes are also listed at this time.
■ Next, suppose that only the $20 and $22.50 options are listed, and the stock rallies. When it reaches
$22.51, options with a strike price of $25 are introduced. If it goes beyond $25, options with the
next higher strike of $30 start trading, and so forth. Of course, one can still trade options with strike
prices $20 and $22.50, which were listed earlier, until these contracts expire.
■ Liquidity is usually the greatest for options with strike prices near the current stock price. Trading
volume and the number of contracts outstanding tend to decline as the stock goes deeper in- or
out-of-the-money.
Margin Requirements and Position Limits
The margin (borrowing) restrictions for long positions in options can be summarized
as follows: (1) short-term options (nine months or less) must post the full purchase
price with zero borrowing and (2) long-term options (more than nine months) can
borrow up to 25 percent of the purchase price. Options sellers are usually required
to keep the full purchase price plus some extra cash as a cushion against risk. An
example illustrates these restrictions.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
EXAMPLE 14.3: Margin Account Adjustments
■ Long call or long put (maturing in less than nine months). For example, from Figure 14.1, if Ms. Long is
buying five IBM March 2009 (IBMCR) calls worth $2.30 per share, she will need to keep (Ask price
× Number of contracts × Contract size) = 2.30 × 5 × 100 = $1,150 in her margin account.
■ Long call or long put (maturing in over nine months). The initial (maintenance) margin requirement is
75 percent of the cost (market value). For example, if Ms. Long wants to buy ten IBM January 2010
(WIBAR) contracts worth $12.50, she will have to keep at least 75 percent of 12.50 × 10 × 100 =
0.75 × 12,500 = $9,375 of her own money as margin and borrow up to $3,125 from her broker.
■ Options writer. Suppose Mr. Short is selling ten IBM October 95 (IBMVS) put options (see
Figure 14.1).
Then margin would be computed as follows:
- 100 percent of contract proceeds (13.30 × 10 × 100) $13,300
- Plus 20 percent of aggregate contract value (0.2 × 90.42 × 1,000) 18,084
- TOTAL amount (13,300 + 18,084) $31,384
Also compute
- Put option proceeds (13.30 × 10 × 100) $13,300
OPTION CONTRACT FEATURES 299
- Plus 10 percent of aggregate exercise price (0.1 × 95 × 1,000) 9,500
- TOTAL amount (13,300 + 9,500) $22,800
The margin Short has to keep is $31,384, which is the larger of the two.
If the options holder-writer has positions in other options or the underlying, the
CBOE adjusts these margin restrictions, taking into account the portfolio’s overall
risk. For example, the margin required for a long stock plus a long put position is
significantly lower than a long put position alone. The adjusted margin takes into
account the reduced risk from the long stock hedging the long put position.
Options position limits are imposed to reduce the potential for market manipula-
tion. For actively traded stocks with large capitalizations (e. g., IBM or Microsoft),
the position limit is 250,000 contracts. Stocks with smaller capitalizations have smaller
position limits. Exemptions may be granted for qualified hedging strategies.
Dividends and Stock Splits
Options are not adjusted for cash dividends or stock distributions that do not exceed
10 percent of the stock’s value. When a stock splits or pays large stock dividends,
the option’s strike price is adjusted on the ex-dividend day or the ex-split day. The
adjustment depends on whether the split or stock dividend is an integer multiple such
as 3 for 1 or a fractional such as 3 for 2. The OCC has developed these rules and
procedures in an attempt to guarantee that options traders do not suffer losses from
these events.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
EXAMPLE 14.4: Options Adjustment for Stock Splits and Stock Dividends
IBM’s stock price is $90.42 (see Figure 14.1). Consider the April 90 call contract.
An Integer Multiple split
■ Suppose IBM has a 3 for 1 split. An investor who held one share would now have three shares worth
$30.14 per share. On the ex-distribution day, the number of options will increase by a factor of 3, and
the strike price will be reduced by a factor of 3.
■ The aggregate exercise price before the split was 100 × $90 = $9,000. After the split, it equals 3 ×
100 × $30 = $9,000. They are the same—the adjustment makes the option neutral with respect to
a stock split.
A Fractional split
■ Suppose IBM has a 3 for 2 split. An investor holding two shares will now own three shares. To prevent
arbitrage, the new share’s price will be 90.42/1.5 = $60.28, where 3/2 = 1.5 is the split ratio.
300 CHAPTER 14: OPTIONS MARKETS AND TRADING
■ In this case, instead of adjusting the number of contracts, the number of shares in each contract is
multiplied by the split ratio: 100 × 1.5 = 150. The new strike price is obtained by dividing the old
strike price by the split ratio. Thus the new exercise price after the split is 90/1.5 = $60.
■ Notice that the aggregate exercise price before the split was $9,000. After the split, the aggregate
exercise price is 150 × $60 = $9,000. Again, this adjustment makes the option neutral with respect to
the stock split.
Stock Dividends
■ Large stock dividends (greater than 10 percent of the stock’s price) are called stock splits and are treated
accordingly. For example, a 3:1 split is really a 200 percent stock dividend, and a 3:2 split is a 50 percent
stock dividend.
Reverse Split
■ If IBM has a 1 for 2 reverse split, the stock price would become $180.84. The contract is adjusted so
that each original option is now an IBM April 180 call covering fifty shares.
14.5 Options Trading, Exercising, and the
Expiration Process
The Trading Process
Options trading involves the three-step process of execution, clearing, and settlement
inherent to exchange-traded securities. As with futures, only a clearing member can
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
clear trades. A clearing member must be a member of an options exchange, satisfy
OCC’s minimum capital, and meet other requirements. He must maintain efficient
operations on the trading floor and in the back office. Clearing members pay both the
exchange and the OCC a fee for each traded contract. They recoup these expenses
from the fees charged to their customers.
Exercising the Option
The process begins when the holder conveys her intention to exercise the option
to her broker. To ensure exercise on a particular day, the notification must be given
before the broker’s daily cutoff time. Next, the broker submits an exercise notice to
the OCC. The OCC sends the notice to one or more clearing members holding
short positions on the same contract. The clearing member, in turn, assigns it (either
randomly or on a first-in, first-out basis) to one or more customers who hold short
positions in the contract. The clearing member representing the short is obligated to
sell (in the case of a call) or buy (in the case of a put) the underlying shares at the
specified strike price. The OCC then arranges for the delivery of the shares of stock
and the exchange of strike price funds.
OPTIONS PRICE QUOTES 301
The Expiration Process
The expiration date is the last day an option trades. For listed stock options, this is the
third Friday of the expiration month. Brokerage firms must submit exercise notices
to the OCC.
The OCC has developed a procedure known as exercise by exception (also
called ex-by-ex) to expedite the handling of the exercise of expiring options. Every
contract in-the-money will be automatically exercised unless the clearing member
advises OCC otherwise. Currently equity and index options that are in-the-money
by at least 1 cent are automatically exercised. Nonetheless, most brokers still require
their customers to notify their intention to exercise even if an option is in-the-money.
When an equity option is exercised or assigned and it is not covered by stock
holdings in the account, sufficient purchasing power must be in the account to make
or take delivery of the shares. Otherwise, the broker may close your expiring options
positions, even without notification, on the option’s last trading day. If the expiring
position is not closed and there is insufficient purchasing power, the broker may
submit “do not exercise” instructions to the clearing member without notification.
14.6 Options Price Quotes
Wire news services, websites, and many major daily newspapers carry options price
quotes. Because each stock has calls and puts that differ in terms of exercise price and
expiration months, an exchange typically has tens of thousands of options trading.
All options contracts of the same type written on the same underlying stock belong
to the same option class. For example, different IBM calls belong to one option
class, whereas different IBM puts belong to another. Within a particular class, all
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
options with the same strike price and expiration date belong to an option series;
for example, IBM March 100 calls belong to one option series, IBM April 90 calls
belong to another option series, and IBM April 100 puts belong to yet another option
series. Example 14.5 illustrates IBM’s options price quotes.
EXAMPLE 14.5: IBM Options Price Quotes
■ IBM’s stock price is $90.42, which is up $1.80 (2.03 percent) from the previous day’s close. Figure 14.1
gives the IBM options “chain table” data from March 12, 2009, compiled from the CBOE website
(the figure has been condensed and all quotes are twenty minutes delayed). Calls are written on the left
side and puts are written on the right side of the figure, with the strike price in the middle. The first
set of numbers reports call and put prices for March 2009, with the shaded region reporting options
that are in-the-money.
■ Consider the tenth entry under “Calls,” for which $90 is the strike price. The row contains the
following data, moving from left to right:
- This option is also identified by IBMCR—C stands for the third month, March, and R stands for
the strike $90.
302 CHAPTER 14: OPTIONS MARKETS AND TRADING
- The price $2.27 is recorded under “Last Trade,” which stands for an option’s last reported trade price.
- The $0.56 denotes the price change from the previous settlement price.
- The bid price $2.25 and the ask price $2.30 are reported per option on one IBM share.
- The 6,516 under “Volume” is the number of contracts that have traded so far during the trading day.
- “Interest” 14,198 is the open interest, which is the total number of outstanding contracts.
FIGURE 14.1: IBM Options Quote from CBOE Website
(IBM) International Bus Mach Corp Com
March 2009
Calls Puts
Symbol Last Change Bid Ask Volume Interest Strike Symbol Last Change Bid Ask Volume Interest
Trade Price Trade
IBMCX 43.40 2.40 45.20 45.60 n.a. n.a. 45.00 IBMOX 0.10 0.00 0.05 0.05 35 123
IBMCU 36.00 −6.60 40.30 40.60 n.a. n.a. 50.00 IBMOU 0.05 0.00 0.05 0.05 148 2,350
IBMCV 33.40 4.40 35.20 35.60 n.a. 5 55.00 IBMOV 0.05 0.00 0.05 0.05 7 325
IBMCL 25.20 1.20 30.20 30.60 7 92 60.00 IBMOL 0.05 0.00 0.05 0.05 7 827
IBMCM 20.90 −3.00 25.30 25.60 4 117 65.00 IBMOM 0.05 0.00 0.05 0.05 34 1,309
IBMCN 19.60 0.60 20.30 20.60 14 218 70.00 IBMON 0.05 0.00 0.05 0.10 20 3,153
IBMCO 12.80 −1.70 15.40 15.60 30 578 75.00 IBMOO 0.09 −0.01 0.05 0.10 307 7,409
IBMCP 10.45 1.35 10.50 10.70 116 2,874 80.00 IBMOP 0.20 −0.05 0.15 0.20 828 10,548
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
IBMCQ 5.90 1.10 5.80 6.00 1,200 10,054 85.00 IBMOQ 0.50 −0.40 0.50 0.55 5,373 14,901
IBMCR 2.27 0.56 2.25 2.30 6,516 14,198 90.00 IBMOR 1.85 −1.05 1.80 1.90 2,741 12,079
IBMCS 0.50 0.15 0.45 0.50 3,078 14,336 95.00 IBMOS 5.00 −1.00 5.00 5.10 426 4,213
IBMCT 0.05 −0.01 0.05 0.10 148 10,760 100.00 IBMOT 10.10 −1.49 9.50 9.70 62 1,913
IBMCA 0.05 0.00 0.05 0.15 2 4,150 105.00 IBMOA 15.30 −4.10 14.50 14.80 6 75
IBMCB 0.03 −0.02 0.05 0.05 16 3,473 110.00 IBMOB 18.30 −7.00 19.40 19.80 26 10
IBMCC 0.05 0.03 0.05 0.05 n.a. 410 115.00 IBMOC n.a. n.a. 24.30 24.70 n.a. n.a.
IBMCD 0.03 −0.04 n.a. 0.05 n.a. 81 120.00 IBMOD n.a. n.a. 29.50 29.80 n.a. n.a.
IBMCE n.a. n.a. n.a. 0.05 n.a. n.a. 125.00 IBMOE 34.00 0.00 34.40 34.70 1 n.a.
OPTIONS PRICE QUOTES 303
Apr 2009
Calls Puts
Symbol Last Change Bid Ask Volume Interest Strike Symbol Last Change Bid Ask Volume Interest
Trade Price Trade
IBMDP 11.8 0.90 11.60 11.80 458 3,221 80.00 IBMPP 1.30 −0.25 1.25 1.35 1,527 12,260
IBMDQ 7.80 0.99 7.70 7.90 2,381 9,649 85.00 IBMPQ 2.40 −0.23 2.40 2.45 1,060 13,098
IBMDR 4.60 0.73 4.50 4.70 1,672 11,644 90.00 IBMPR 4.20 −0.75 4.10 4.30 2,145 10,644
IBMDS 2.40 0.60 2.30 2.40 3,405 21,552 95.00 IBMPS 6.90 −1.20 6.80 7.00 254 14,501
IBMDT 1.05 0.30 0.95 1.05 992 15,196 100.00 IBMPT 12.01 −3.19 10.5010.70 1 2,553
Jul 2009
Calls Puts
Symbol Last Change Bid Ask Volume Interest Strike Symbol Last Change Bid Ask Volume Interest
Trade Price Trade
IBMGP 14.23 0.83 14.50 14.70 88 1,605 80.00 IBMSP 4.70 −0.26 4.60 4.70 58 3,457
IBMGQ 11.40 1.00 11.20 11.40 317 1,883 85.00 IBMSQ 6.30 −0.60 6.30 6.40 264 4,300
IBMGR 8.50 0.80 8.40 8.50 600 5,110 90.00 IBMSR 8.67 −0.53 8.40 8.60 93 5,381
IBMGS 6.10 0.80 6.00 6.10 168 8,500 95.00 IBMSS 11.00 −0.70 11.0011.20 27 1,283
IBMGT 4.22 0.62 4.10 4.20 274 6,417 100.00 IBMST 15.31 −0.19 14.10 14.30 n.a. 848
Oct 2009
Calls Puts
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Symbol Last Change Bid Ask Volume Interest Strike Symbol Last Change Bid Ask Volume Interest
Trade Price Trade
IBMJP 16.23 2.33 16.30 16.60 20 397 80.00 IBMVP 7.00 −1.86 6.60 6.80 2 162
IBMJQ 13.40 2.70 13.20 13.50 3 202 85.00 IBMVQ 9.30 −1.70 8.50 8.70 12 345
IBMJR 10.59 2.29 10.50 10.70 48 119 90.00 IBMVR 11.32 −0.18 10.70 10.90 n.a. 387
IBMJS 7.40 1.10 8.10 8.30 44 519 95.00 IBMVS 14.40 −0.60 13.3013.60 44 257
IBMJT 6.00 0.40 6.10 6.30 13 273 100.00 IBMVT 17.90 −2.30 16.3016.60 43 98
Jan 2010
Calls Puts
Symbol Last Change Bid Ask Volume Interest Strike Symbol Last Change Bid Ask Volume Interest
Trade Price Trade
WIBAP 17.70 2.30 17.60 18.30 15 1,744 80.00 WIBMP 8.60 −1.80 8.40 8.70 n.a. 5.576
WIBAQ 14.10 2.30 14.60 15.20 1 1,151 85.00 WIBMQ 11.70 −0.90 10.40 10.70 8 2,616
WIBAR 11.63 1.23 12.20 12.50 n.a. 7,696 90.00 WIBMR 13.20 −1.60 12.7013.00 2 6,936
WIBAS 9.50 1.70 9.90 10.10 3 3,383 95.00 WIBMS 15.70 −0.90 15.3015.70 47 2,008
WIBAT 7.74 0.27 7.80 8.10 6 10,373 100.00 WIBMT 19.60 −3.70 18.3018.60 22 5,714
(continued)
304 CHAPTER 14: OPTIONS MARKETS AND TRADING
Jan 2011
Calls Puts
Symbol Last Change Bid Ask Volume Interest Strike Symbol Last Change Bid Ask Volume Interest
Trade Price Trade
VIBAU n.a. n.a. 69.30 70.80 n.a. n.a. 20.00 VIBMU 0.45 0.00 0.30 0.45 50 358
VIBAX n.a. n.a. 67.00 68.20 n.a. n.a. 22.50 VIBMX 0.55 −0.25 0.45 0.60 100 204
VIBAE n.a. n.a. 64.50 65.80 n.a. n.a. 25.00 VIBME 0.80 −0.12 0.65 0.75 2 105
VIBAV n.a. n.a. 59.30 60.80 n.a. n.a. 30.00 VIBMV 1.20 −0.25 1.05 1.25 n.a. 883
VIBAG 49.50 1.00 54.70 55.90 n.a. 25 35.00 VIBMG 2.10 −0.08 1.65 1.90 20 163
VIBAW 47.70 1.90 50.10 51.30 1 41 40.00 VIBMW 2.95 0.05 2.40 2.60 1 236
VIBAI 42.90 −6.20 45.80 47.00 2 7 45.00 VIBMI 3.40 −0.10 3.20 3.60 n.a. 147
VIBAZ 39.20 −6.60 41.60 43.00 10 207 50.00 VIBMZ 4.70 −0.80 4.20 4.60 10 407
VIBAK 35.40 0.00 37.80 39.00 2 2 55.00 VIBMK 6.30 0.24 5.40 5.90 10 135
VIBAL 31.50 −0.10 34.30 35.40 3 92 60.00 VIBML 7.00 −0.80 6.80 7.20 1 323
VIBAM 28.20 −5.30 31.40 31.90 2 21 65.00 VIBMM 9.60 −0.70 8.30 8.70 44 293
VIBAN 24.80 −0.70 28.00 28.80 n.a. 254 70.00 VIBMN 11.60 0.00 10.00 10.60 25 331
VIBAP 20.20 −2.84 22.00 22.90 6 293 80.00 VIBMP 16.70 1.50 14.10 14.70 40 439
VIBAQ 18.50 0.96 19.90 20.30 50 479 85.00 VIBMQ 18.80 −0.96 16.30 16.90 117 748
VIBAR 15.80 0.40 17.20 17.90 10 1,079 90.00 VIBMR 22.40 1.40 18.80 19.40 n.a. 1,107
VIBAS 14.30 1.20 14.90 15.70 45 521 95.00 VIBMS 22.60 −1.90 21.50 22.20 11 636
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
VIBAT 13.00 0.40 13.00 13.70 1 1,424 100.00 VIBMT 27.00 0.10 24.40 25.20 n.a. 676
VIBAB 9.80 0.30 9.50 10.10 4 812 110.00 VIBMB 35.50 0.10 30.70 31.60 24 512
VIBAD 6.60 −0.20 6.90 7.50 34 638 120.00 VIBMD 42.90 2.10 37.70 38.70 11 196
VIBAF 4.80 0.40 4.90 5.30 16 407 130.00 VIBMF 47.00 −4.00 45.40 46.20 3 193
VIBAH 3.60 0.60 3.20 3.70 1 599 140.00 VIBMH 54.80 −0.10 53.60 54.70 47 257
VIBAJ 2.25 0.05 2.25 2.50 67 672 150.00 VIBMJ 66.50 −2.00 62.30 62.90 42 227
Source: http://delayedquotes.cboe.com/options/options_chain.html?ASSET_CLASS=STO&ID_OSI=85502&ID_NOTATION=1551887
REGULATION AND MANIPULATION IN OPTIONS MARKETS 305
14.7 Regulation and Manipulation in
Options Markets
In the movie Mission: Impossible II (2000), a fictional character, Sean Ambrose, plans to
release a virus and then profit by selling the antidote. “I want stock!” Sean thunders.
“Stock options, to be precise.” And in the James Bond movie Casino Royale (2006),
a fictional banker, Le Chiffre, provides a guerrilla group a safe haven for its funds.
But Le Chiffre’s investments actually involve considerable risk. He buys puts on a
successful company and then engineers a terrorist attack to sink the stock values.
Stock options and market manipulation have entered pop culture!
Options have long been associated with market manipulation. For example, during
the 1920s, options gained notoriety when they were used in conjunction with stock
price manipulation schemes. One abusive practice was for speculators with large
holdings of the underlying stock to grant call options to stockbrokers (see Overby
2007). In return, brokers would recommend their clients purchase the stock.
Another market manipulation scheme is buying out-of-the-money options and
simultaneously buying the underlying shares to artificially increase the stock price so
that the options end up in-the-money. Traders on the other side, who were getting
ready to go out and celebrate happy hour on a late Friday afternoon, may suddenly
find that their short out-of-the-money call options, which were seemingly worthless,
suddenly end up in-the-money!
We introduced the notion of a price bubble in Chapter 11, which happens when
the stock price substantially deviates from its intrinsic or fundamental value. Bubbles
typically involve a dramatic price rise followed by a drastic decline. Bubbles are asso-
ciated with the dotcom boom of the 1990s and the recent real estate rise and collapse,
and options are often connected with such episodes. A classic case was the tulip bulb
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
bubble in seventeenth-century Holland. Dutch tulips are famous to this day, but their
early history was fraught with speculation, greed, and fraud (see Malkiel 2003).
The tulip bulb bubble story begins in 1593, when a botany professor brought
tulip bulbs from Turkey to Holland, hoping to vend them at high prices. This market
flourished, and tulips traded actively in the Netherlands in the years that followed.
Some time later, many bulbs caught a benign virus that patterned their petals with
bright stripes of contrasting colors. The public adored these “bizarres” and bid their
prices up. Tulip bulb prices increased to a peak during the mid-1630s, when people
from all walks of life, from noblemen to chimney sweeps, joined the fray. Some even
bartered or pawned personal belongings to buy tulip bulbs with the hope of selling
them at ever-increasing prices.
Derivatives sprang up to feed the frenzy, and speculators used these derivatives
(including options) for leverage—but they also helped the hedgers. Tulip growers and
retailers bought calls and futures for protection against price increases. Such hedging
strategies are practiced by many businesses even today. After months of outrageous
prices and tumultuous trading, the market suddenly crashed in February 1637. Bad
regulation and poor quality control have been blamed as the reasons for the crash.
It’s no surprise that the regulators have historically been hostile to options.
MacKenzie and Millo (2003) note that the proposal for the establishment of the
306 CHAPTER 14: OPTIONS MARKETS AND TRADING
CBOE was met with “instinctual hostility, based in part upon corporate memory
of the role options had played in the malpractices of the 1920s.” CBOE’s first
president, Sullivan, was told by one leading SEC official that he had “never seen a
[market] manipulation” in which options were not involved. A former SEC chairman
even compared options to “marijuana and thalidomide.” In a twist of fate, today’s
regulators support options exchanges and consider them the lesser evil compared to
the unregulated OTC derivatives markets.
14.8 Summary
1. Options trade because they provide payoffs different from forwards, futures, and
the underlying stock.
2. Historically, options have been looked at with suspicion because of a cumbersome
transaction process, small trading volume, high fees, counterparty risk, and the bad
repute of market participants. The result was that investors surreptitiously traded a
small number of contracts or shunned them altogether. The government’s response
ranged from a grudging acceptance to an outright ban.
3. Options markets got a boost in 1973 with the publication of the Black– Scholes–
Merton model and the opening of the CBOE. Options trading has seen tremen-
dous growth in the years that followed. This includes trading in options on interest
rates, sector funds, exchange-traded funds, and indexes.
4. Equity options traded on the CBOE have the following features:
a. Nature: American physical delivery options
b. Contract size: one hundred shares
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
c. Last exercise date: the third Friday of the expiration month
d. Quotes: usually quoted in pennies ($0.01) for options prices below $3 and in
nickels ($0.05) for higher options prices
e. Strike prices intervals: $2.50 when the strike price is between $5 and $25, $5
between $25 and $200, and $10 for points over $200; strikes are adjusted
for splits and recapitalizations but not for cash dividends (the CBOE has also
introduced $1 strike price intervals and $2.50 strike price intervals, under certain
terms and conditions, for many actively traded stocks)
f. Strike prices: in-, at-, and out-of-the-money strike prices are initially listed; new
series are generally added when the underlying trades through the highest or
lowest strike price available
g. Expiration months: two near-term months plus two additional months from the
January, February, or March quarterly cycles
h. Margin: short-term options buyers post the full price; long-term options buyers
(maturing in more than nine months) can borrow up to 25 percent of the
purchase price; options sellers are required to keep the full purchase price plus
extra funds as cushion against risk
QUESTIONS AND PROBLEMS 307
14.9 Cases
Chicago Board Options Exchange (CBOE) (Harvard Business School Case
205073-PDF-ENG). The case discusses institutional details behind how options
trade at the CBOE.
International Securities Exchange: New Ground in Options Markets (Harvard
Business School Case 203063-PDF-ENG). This case examines the equity options
market, the major parties involved, and the options trading process.
Milk and Money (Kellogg School of Management Case, Case KEL343-PDF-ENG,
Harvard Business Publishing). This case considers how a family dairy firm can use
regression analysis to choose the best hedges for its dairy products.
14.10 Questions and Problems
14.1. Briefly describe options trading that took place in 1500–1700 Amsterdam and
in nineteenth-century London and New York City.
14.2. a. Who was Russell Sage?
b. Why was Sage called “The Old Straddle”?
14.3. Why did options trading fall into disrepute in the United States during the
early decades of the twentieth century?
14.4. During the nineteenth century, futures trading in the United States steadily
gained acceptance while options trading was associated with suspicion.
a. Can you explain why this happened?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
b. How and when did this view of options trading get changed?
14.5. Briefly describe the events and developments that led to the founding of the
CBOE.
14.6. Describe two developments that took place in 1973 that made it a watershed
year in the history of options.
14.7. What is an option on a futures? Explain the workings of this derivative
contract.
The next two questions are based on the following data from the CME Group’s website
(prices as of 6:59:36 pm Central Standard Time on August 26, 2011).
Spot price of gold is $1,830 per ounce. Contract size is one hundred ounces.
Prices are per ounce.
14.8. a. Identify which calls are in-the-money, at-the-money, and out-of-the-
money.
b. If you exercise a call with a strike price of $1,820, what is your payoff, and
what are your holdings of the futures contracts?
c. For this call option on gold futures, what is the intrinsic value, and what
is the time value?
308 CHAPTER 14: OPTIONS MARKETS AND TRADING
Strike Call Put
1,820 85.20 107.80
1,825 83.20 110.90
1,830 81.30 113.90
1,835 79.40 217.10
1,840 77.50 120.20
14.9. a. There is an obvious mistake in the put price data—correct that first.
b. Identify which puts are in-the-money, at-the-money, and out-of-the-
money.
c. If you exercise a put with a strike price of 1,835, what is your payoff, and
what are your holdings of the futures contract?
d. For this put option on gold futures, what is the intrinsic value, and what
is the time value?
14.10. Suppose Your Beloved Machines Inc. has a February cycle for options
trading.State the months for which regular equity options on YBM (which
expire on the third Friday of the month) trade on the following dates:
a. January 1
b. January 27
c. March 1
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
14.11. State the dollar amount of margin you are required to keep with a broker
when trading one contract (on one hundred shares) of the following options
on Your Beloved Machines Inc. YBM’s current stock price is $103.
a. A long put worth $6 for an option maturing in six months
b. A long call worth $8 for an option maturing in ten months
14.12. A long is generally associated with buying and a short with selling. Is it
counterintuitive that the put holder gets the right to sell? Explain your answer.
14.13. Your Beloved Machine’s current stock price is $90. YBM December 100 calls
trade for $6.
a. Adjust the options prices and terms of the contract for a 4:1 split.
b. Adjust the options prices and terms of the contract for a 3:2 split.
14.14. Are options on the CBOE adjusted for cash dividends or stock dividends or
both?
14.15. Use Figure 14.1.
a. For the IBM April 2009 calls, is the call value increasing or decreasing in
the strike price?
QUESTIONS AND PROBLEMS 309
b. For the IBM April 2009 puts, is the put value increasing or decreasing in
the strike price?
c. For the IBM March 2009 calls, which strikes are the most actively traded?
d. For the IBM March 2009 puts, which strikes are the most actively traded?
14.16. What is exercise by exception with respect to CBOE options?
14.17. What is an options class? What is an options series? Give examples to illustrate
your answer.
14.18. What is open interest for a traded options contract? What does this tell you
about the trading interest in a particular options contract?
14.19. Give two examples of market manipulation in the options market.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
15
Option Trading Strategies
15.1 Introduction 15.6 Spread Strategies
15.2 Traders in Options Markets EXTENSION 15.1 Collars, Ratio
Spreads, Butterflies, and Condors
15.3 Profit Diagrams EXTENSION 15.2 Reinsurance
15.4 Options Strategies Contracts and Call Spreads
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Common Options Price Data 15.7 Combination Strategies
Naked Trades: The Basic Building
15.8 Summary
Blocks
15.9 Cases
15.5 Hedged Strategies
15.10 Questions and Problems
TRADERS IN OPTIONS MARKETS 311
15.1 Introduction
While teaching options and futures for the first time, one of us (Chatterjea) was
sharing some frustrations with a colleague, Mike Goldstein, about the difficulties
in motivating students. Mike advised, “That’s not the way to teach derivatives.
You should cover the interesting stuff, and show how to implement option trading
strategies. Suppose that a takeover is going on, and explain how to set up a straddle
to bet on the outcome.” He had hit the nail on the head! Focusing on equations like
C(K1 ) – C(K2 ) ≤ K1 – K2 for K1 ≥ K2 is not the best way to attract interest. Moreover,
this gives a false sense of mathematical sophistication while achieving little substance.
It’s far more exciting to take options where the action is and discuss their role for
speculating on takeover battles, placing leveraged bets on the markets, and protecting
portfolios from market crashes. This is the approach we take here.
This chapter studies some popular options trading strategies. We present profit
diagrams for the basic options trades, hedged and spread strategies, including
“butterflies” and “condors,” followed by strategies involving combinations of options.
Along the way, we discuss how options relate to insurance contracts and present a
bird’s-eye view of taxation and reporting requirements. After this chapter, you should
be able to understand option related stories in The Wall Street Journal. The strategies
we discuss are relevant to all players in the options market: the individual trader, the
financial institution, the nonfinancial corporation, and even the government.
15.2 Traders in Options Markets
Bill Lipschutz, who had taken one of the world’s first options courses from one of us
(Jarrow) at Cornell University, joined Salomon Brothers Inc. in 1981, a prestigious
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Wall Street investment banking firm at the time.1 Though Salomon prided itself as
one of the world’s most powerful trading companies, Lipschutz was disappointed to
find options trading to be nonquantitative. He found that no one seemed to know
the Black–Scholes–Merton model. Trading strategies were based on ad hoc criteria,
for example, the head of the proprietary options trading desk said, “‘I went to buy a
car this weekend and the Chevrolet showroom was packed. Let’s buy GM calls.’ That
type of stuff.” Then, one day, a trader pulled him aside and gave away the great secret:
“Look, I don’t know what Sidney is teaching you, but let me tell you everything you
need to know about options. You like ’em, buy calls. You don’t like ’em, buy puts.”
In the early 1980s, many Wall Street firms placed “leveraged outright bets”
and prayed to the “Bitch-Goddess Success.”2 As the decade unfolded, however,
most investment banks started assembling a team of derivative experts and running
1
This story is taken from the interview “Bill Lipschutz: The Sultan of Currencies” in Jack Schwager’s
(1992) book The New Market Wizards: Conversations with America’s Top Traders. Lipschutz developed an
early expertise in trading currency options, which began trading on the Philadelphia Stock Exchange in
1982. During his eight-year stint at Salomon, he routinely traded billions of dollars’ worth of currencies
and became the “largest and most successful currency trader” in the firm.
2
Professor William James of Harvard University used the expression “Bitch-Goddess Success” in a letter
to the British writer H. G. Wells.
312 CHAPTER 15: OPTION TRADING STRATEGIES
proprietary trading desks to make profits. Goldman Sachs and Company, a key player
in this transformation, lured Professor Fischer Black from the Massachusetts Institute
of Technology and made him a partner. Black helped usher in the migration of
many finance professors and numerous talented PhDs in the natural and engineering
sciences into Wall Street firms. The most mathematical of these recruits came to be
variously viewed as “quants” and “rocket scientists” whose job was to apply their math
mastery to make profits. Salomon did not lag behind. It pioneered the mortgage-
backed securities market during the 1970s, arranged the world’s first swap contract in
1981 (see Chapter 22), and built an active derivatives shop. Its consultants included
Nobel laureate Robert Engle, who was hired to develop volatility forecasts and devise
trading strategies based on cutting-edge econometrics.
Much of our discussion of options strategies in this chapter will be from the
perspective of an individual investor. Financial institutions also trade options to bet on
the markets and capture arbitrage profits. Hedge funds and proprietary trading desks
at investment banks also implement the strategies we describe. Warren Buffett noted
in his 2008 chairman’s letter in the annual report of Berkshire Hathaway Inc. that his
company had entered into 251 derivatives contracts, including writing fifteen-and
twenty-year maturity European put options on four major indexes (the Standard and
Poor’s [S&P] 500, London’s FTSE 100, Europe’s Euro Stoxx 50, and Tokyo’s Nikkei
225). He believed that each of these contracts was “mispriced at inception [emphasis
added], sometimes dramatically so.” Doesn’t this smack of arbitrage? Affirming his
belief that “the CEO of any large financial organization must be the Chief Risk
Officer as well,” Buffett declared that he had initiated these positions, would continue
to monitor them, and would take full responsibility if the company lost profits on
these derivatives trades.
We discuss later in the chapter how traders (including insurance companies)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
use options to hedge and speculate on catastrophic risks. Many foundations and
endowments, mutual funds, and pension funds actively trade options. Options often
underlie financial engineering techniques used by nonfinancial companies, which buy
inputs to produce outputs. Sometimes governments use options. Recall the example
at the beginning of Chapter 14, in which Lawrence Summers and Zhu Rhongji
discussed the feasibility of using put options to defend the value of the Chinese
currency and how the US government used warrants to design a highly effective
financial rescue package for troubled banks. Options trading strategies adopted by
different kinds of traders get regularly mentioned in the financial press and addressed
in case studies.
15.3 Profit Diagrams
As pictures say a thousand words, we use profit diagrams as our primary tool for
discussing options trading strategies. Recall that profit diagrams place the stock price
on the expiration date S(T) on the horizontal (x) axis and the net profits from trades
on the vertical (y) axis. The profit or loss is computed by vertically adding up the
profits and losses for each security within the trade. Wherever convenient, we show
the strike price K, the maximum profit (or loss), and the stock price corresponding
OPTIONS STRATEGIES 313
to the breakeven point (BEP; or the zero-profit point), where the trader recoups
the option premium. Absent market frictions, the zero-sum nature of options trading
ensures that the BEP is the same for both the buyer and the writer.
Profit diagrams have some limitations. First, you can only draw them for European
options. American options have the risk that one or more legs of the strategy may
get exercised early and disappear from the diagram. Second, they do not apply before
expiration because the time value of the options (see Chapter 5) moves their profits
away from those on the expiration date. Third, we must have just one stock as the
underlying. You cannot draw profit diagrams for a portfolio of options on different
stocks.
Still these diagrams are quite useful. If American options remain unexercised until
expiration, they have the same payoffs as European options. Although most exchange-
traded options in the US are American options, early exercise isn’t very common.
The next chapter will demonstrate that options are usually worth more alive (i.e.,
unexercised) than dead (exercised). Profit diagrams provide a useful snapshot of the
possibilities on the option’s maturity date.
We focus on equity options. With minor modifications, you can similarly analyze
other options, including those on commodities, indexes, and even futures and
forwards. For now, we examine simple options trading strategies.
15.4 Options Strategies
A story titled “Options Trading Grew Up in 2009” (Wall Street Journal, January 4,
2010) reported that many new options traders entered the market in 2009. What
attracted these options traders? The article identified several factors. Many investors
traded options to better manage their portfolio’s risks after suffering losses. Some
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
adopted complex strategies that “helped them to reduce trading costs and limit the
impact of ‘decay,’ which describes the pace at which options lose value as they
approach expiration.” To prevent losses, they bought put options or “put spreads,”
and to generate income, they tended to sell “covered calls.” This and the next two
sections discuss these options trading strategies.
Common Options Price Data
Most of our options trading strategy examples will use the following common options
pricing data (COP data) for a fictitiously named company, OPSY, which is the Dutch
name for options:
■ The current stock price S is $22.50.
■ The time to maturity T is six months.
■ The strike prices K are $17.50, $20, $22.50, and $25.
314 CHAPTER 15: OPTION TRADING STRATEGIES
TABLE 15.1: European Option Price Data for OPSY
Strike Price Call Price Put Price
K0 = $17.50 $5.50 $0.10
K1 = 20.00 3.50 0.50
K2 = 22.50 2.00 1.50
K3 = 25.00 1.00 3.00
Stock price S = $22.50; time to maturity T = 6 months; risk-free interest rate r = 5 percent per year.
■ The continuously compounded, risk-free interest rate r is 5 percent per year.
■ The European options prices are given in Table 15.1. These assume a volatility
equal to 0.26.
We ignore market imperfections such as transaction costs and also interest earned
on the option’s premium. Usually these are small. Of course, you can easily add them
to the profit or loss amounts.
Naked Trades: The Basic Building Blocks
The basic option trading strategy building blocks are stand-alone positions containing
a single option. Traders refer to these as naked trading strategies (who said options
are boring?). Long naked trades are outright purchases because you are not holding
an offsetting position in the underlying stock. Short naked trades are also known as
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
uncovered strategies because you are not covering (by reducing the risk of) the
trade by simultaneously taking a position in the underlying stock.
EXAMPLE 15.1: Going Long and Selling Short: Stocks and Options
Before illustrating the four uncovered positions for European call and put options (with strike prices of
$22.50), we draw profit diagrams for long and short stock trades (see Figure 15.1). You can evaluate
option trades vis-á-vis these key references. Naked option trading strategies are popular among traders.
Long and Short Stock
■ As stocks are limited liability assets, the maximum loss is the price paid. Long OPSY’s profit diagram
is a straight line that emanates from a loss of $22.50 along the vertical axis. It rises at a 45 degree angle,
cuts the horizontal axis at the breakeven point of $22.50, and continues upward with infinite profit
potential.
■ The short stock’s profit diagram is the mirror image of the long stock across the x axis. An oft-quoted
Wall Street saying attributed to the notorious financier Daniel Drew explains the riskiness of short
selling: “He who sells what isn’t his’n, buys it back or goes to prison.”
OPTIONS STRATEGIES 315
Long Call
■ Buying OPSY 22.50 European calls for $2 gives the right to buy the stock for $22.50 at time T.
Recall from Chapter 5 that the profit diagram is a horizontal line starting from (a maximum loss of)
$2 on the vertical axis. Then, at the strike price of K = $22.50, it shifts upward at a 45 degree angle,
breaks even where the stock price S(T) equals the strike price plus call premium (BEP = $24.50), and
increases with an unbounded profit potential.
FIGURE 15.1: Profit Diagrams for the Basic Security
Blocks (Naked Strategies)
Profit Profit
Long stock
1 22.5
1
22.5 45° 22.5
0 0
S(T) S(T)
Stock price 1
Short
at expiration 1
–22.5 stock
Call and put strike price K = $22.50
Profit Profit
Long call
21 Long put
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
24.5 22.5
2 1.5
0 45°
–2 45°
S(T) –1.5 S(T)
22.5 21
Short call –21 Short put
Breakeven point for call = $24.50 Breakeven point for put = $21
■ Applications and uses:
- Buy calls to benefit if the stock goes up while limiting losses to the premium if the stock goes down.
You can buy OPSY stocks for $22.50 per share. Alternatively, you can buy OPSY 22.50 six-month
calls for $2 per share. Each trade has its pros and cons. Action on the call happens during its short life
span. By contrast, you can hold on to the stock forever in the hope of a price increase. If OPSY goes
down to $15, you lose $7.50 per share but just $2 on the call. If OPSY goes up, you immediately
start benefiting on the long stock, but you have to wait until the stock price crosses the breakeven
point of $24.50 before profiting on the call.
316 CHAPTER 15: OPTION TRADING STRATEGIES
- Let’s look at percentage returns. A $7.50 decline is a 33.33 percent loss on the stock investment, but
a $2 loss on the premium is a 100 percent loss on the option. If the stock goes up to $30, then there
is a 33.33 percent profit on the stock but a (30 – 24.50)/2 = 275 percent profit on the call. The
dollar gain or loss is higher for the stock, but the percentage gain or loss is higher for the option. This happens
because options are leveraged investments.
- Buy calls to speculate on events, corporate developments, and sector and economic performance.
Short Call
■ Selling a call option is also known as shorting or writing a call. A short call’s profit diagram is the
mirror image across the x axis of that for a long call.
■ Naked call writing is one of the riskiest strategies, and only seasoned option traders with sufficient
financial resources are allowed to employ this strategy.
Long Put
■ A put buyer loses the entire premium of $1.50 if the stock price stays above the strike but starts
recouping his investment when S(T) dips below this mark. The breakeven point is at $21, where
the $1.50 gain exactly offsets the premium paid. Profit is maximized when S(T) hits zero. Here a
worthless stock is sold for $22.50, and deducting the premium arrives at a profit of $21.
- Put purchasers bet on the downside but limit their losses in case the bet goes wrong. As with calls,
a long put has similar pluses and minuses vis-à-vis a short stock trade.
Short Put
The uncovered put writer has the obligation to buy OPSY for $22.50 on assignment of an exercise
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
notice. A put writer is uncovered if he does not have a short stock position or has not deposited cash
equal to the strike to form a cash-secured put.
Naked put sellers take losses during market crashes. One of the most notorious examples happened
during the Black Monday crash of October 19, 1987. Buoyed by the large stock market gains during the
mid-1980s, naked put sellers viewed put premiums as free cash. But when the Dow fell 22.61 percent
on this fateful day, many of these put sellers were wiped out.
■ Naked put writers often have two objectives: to receive premium income and to acquire stock at a
cost below its market price at some future date.
■ Short put strategies are bullish.
To summarize, naked options trades generate profits that are similar to those
created by trading stocks without their downsides. But they are created for a limited
time and require regular payment of premiums to maintain the position over long
time periods.
HEDGED STRATEGIES 317
15.5 Hedged Strategies
Hedged strategies combine options and stocks in ways that reduce the overall risk
of naked option trades. They are usually covered strategies because they hedge
option transactions with the stocks that protect (cover) them.
EXAMPLE 15.2: Hedged and Covered Strategies
Covered Call Writing (Long Stock plus Short Call)
■ Covered call writing combines a short call with the underlying stock. Figure 15.2 illustrates this 1:1
hedge strategy.
- At the origin, the long OPSY trade incurs the maximum loss ($22.50), but the short call retains the
entire call premium of $2. The combined profit is – 22.50 + 2 = – $20.50.
- As the stock price increases, the short call’s profit graph (a dotted line parallel to the horizontal axis)
remains constant at $2, but the long stock’s profit graph (an upward sloping dashed line) rises by a
dollar for each dollar increase in S(T). The resulting profit diagram (shown by a solid line) rises at
a 45 degree angle, until the point at $22.50.
- At the strike price of $22.50, the profit is $2 (from short call) + 0 (from long stock) = $2. For higher
values of S(T), the short call’s profit decreases at a negative 45 degree angle, which neutralizes the
long stock position’s increase. Consequently, the covered call’s profit diagram is a flat line for all
higher values of OPSY.
■ Doesn’t the profit diagram for this concoction look like the diagram for a short put position? The
next chapter will explain why this happens.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ The timing of the trades gives this strategy different names. It’s called a buy–write if one simultaneously
buys the stock and sells the call, but it’s an overwrite if one sells the call after purchasing the share.
■ Buy–write’s popularity prompted the CBOE to develop an innovative benchmark for measuring
the performance of such strategies. Released in April 2002, the award-winning CBOE S&P 500
BuyWrite Index (BXM) is based on the total returns from hypothetically holding the S&P 500
stock index portfolio and writing a slightly out-of-the-money, one-month maturity call option on
the S&P 500 index.
■ Traders who write covered calls generally have two objectives: collecting a premium or hedging a
stock trade. By accepting the premium, the writer surrenders the opportunity to benefit from a stock
price rise above the exercise price. The call premium gives the writer some cushion against a stock
price fall, but he remains vulnerable to losses from a deep decline in the stock. Covered call writing
works well in neutral markets, particularly when the volatility is high, which translates into higher call
prices.
■ Covered call writing is a more conservative strategy than an outright stock purchase. A director of
investments at a pension fund went as far as to comment that covered call writing is one of the most
conservative strategies possible.
318 CHAPTER 15: OPTION TRADING STRATEGIES
FIGURE 15.2: 1:1 Hedges: Covered Call Writing
(Buywrite or Overwrite) and
Covered Put Buying (Married Put
or Protective Put)
Profit Long stock
20.5
Covered call writing
2
0
S(T)
22.5
Short call
1
–20.5
–22.5
Profit
21
Long put
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Covered put buying
22.5
0
–1.5 S(T)
24
Long stock
–22.5
■ Even the government views covered call writing as a less risky strategy. You can write covered calls
in some tax-advantaged retirement savings accounts. In June 2009, the Securities and Exchange
Commission approved allowing employees to sell calls against unexercised holdings of employee
HEDGED STRATEGIES 319
stock options (which are call options on the stock that are granted by the company as a form of
noncash compensation; see “Stock Options Opened for ‘Call Writing,’” Wall Street Journal, June 26,
2009).
Covered Put Buying (Long Stock plus Long Put)
■ Another highly popular strategy, covered put buying has long been used to justify the trading of
options: to buy insurance by hedging a long stock with an option. Figure 15.2 illustrates this strategy.
- At the origin, the long OPSY trade loses $22.50 but the long put position has a gain of $21, giving
a profit of – $1.50, which is the maximum loss.
- For each dollar increase in S(T), the long put decreases the covered put strategy’s profit by a dollar,
giving a profit parallel to the horizontal axis.
- When S(T) reaches K = $22.50, the long put’s profit changes and becomes parallel to the x axis, but
the long stock position keeps increasing. As a result, the profit rises at a 45 degree angle, signifying
boundless profit potential.
■ As with the previous strategy, the timing of the trades gives the strategy different names. It is called
a married put (from an old Internal Revenue Service ruling, states the Options Industry Council
website) if one simultaneously buys the put and the stock, but a protective put if one buys the put to
protect the downside risk of a stock previously purchased.
■ Again, the profit diagram looks like another option trade, a long call position. We will explain this in
the context of put–call parity in the next chapter.
■ Traders employ covered put strategies for several reasons:
- Establishing a minimum selling price for the stock. No matter how far the stock declines, your loss is
limited to the put premium plus the difference, if any, between the stock’s purchase price and the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
put’s exercise price. In this case, the minimum selling price is (Strike price – Put premium) = 22.50
– 1.50 = $21.
- Buy puts to protect unrealized profit in a long stock position. Suppose you bought OPSY when it was
trading for $15 and it has now reached $22.50. You can set up a protective put position to hedge
your profit against short-term price declines. No matter how far the stock falls, you can always
exercise the put and sell the stock at the strike price $22.50. Your profit never falls below [(Strike
price – Stock purchase price) – Put premium] = (22.50 – 15) – 1.50 = $6.
Covered Call Buying (Short Stock plus Long Call)
Suppose you are bearish and sell OPSY short. This is a high-risk strategy. What’s your protection if the
stock rallies instead? One can hedge the risk by buying a call (perhaps an out-of-the-money call like
OPSY 25, which has a price of $1).
Covered Put Writing (Short Stock plus Short Put)
A less popular strategy, covered put writing tries to generate some income for the short sellers.
320 CHAPTER 15: OPTION TRADING STRATEGIES
The 1:2 Hedge
The previous four strategies considered 1:1 hedges (see Figure 15.2), in which one option was covered
by one stock. You can also create hedged strategies by combining unequal numbers of options. A 1:2
hedge combines a long stock with two short calls (see Figure 15.3) or two long puts. The profit diagram
looks similar to the diagram of a straddle discussed later in this chapter. These strategies have long been
used by dealers who write options; for example, when options are at-the-money, a call writer may buy
half a share for each call sold, whereas a put writer may sell half a share for each put sold. The precise
quantity of stocks to hedge options trades to remove all of the portfolio’s price risk is given by the stock’s
delta, which is discussed in Chapters 17, 18, and 19 and in considerable detail in Chapter 20.
FIGURE 15.3: Long Stock and Two Short
Calls (1:2 Hedge)
Profit
Long stock
4
18.5 26.5
0
S(T)
22.5
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1:2 Hedge
–18.5
Two short calls
–22.5
15.6 Spread Strategies
Oxford University professor John L. Austin asserted during a lecture that although
there are many languages in which a double negative makes a positive, no language
uses two positives to express a negative statement. Seated in the audience was
another eminent philosopher, Columbia University professor Sidney Morgenbesser.
He dismissively dissented, “Yeah, yeah,” and the tone of his voice disproved the claim.
In finance, we can use a variation of this venerable philosopher’s assertions to describe
a spread strategy in which two risky trades combine to form a less risky portfolio.
This happens because a spread strategy subtracts “similar” risks, making the end
product safer than either of the two originals. Because spreads are two-sided hedges
SPREAD STRATEGIES 321
that combine two options of the same type but on opposite sides of the market,
the exchanges allow spreads to have lower margins than naked trades. A spread gives
a smaller profit if the underlying moves in one direction but a tiny loss otherwise.
Choice of which option to buy and which to sell determines whether it is a bullish
or a bearish spread.
Spreads come in three basic types: (1) vertical spreads (also called money,
perpendicular, or price spreads) involving options that have different strike prices
but expire on the same date; (2) horizontal spreads (or time or calendar spreads)
involving options that have the same strike but different maturity dates; and (3)
diagonal spreads involving options that differ both in terms of strike price and
maturity date. Different maturity dates prevent representing diagonal and horizontal
spreads in standard profit diagrams. Instead, traders compute options sensitivities to
understand the behavior of such portfolios. Chapter 20 explains these tools.
EXAMPLE 15.3: Bull and Bear Spreads
Bull Spread
■ You can create a bullish vertical spread by buying an option and simultaneously selling an option of
the same type but with a higher strike price, where both options have the same underlying security and
the same date of maturity.4 Traders set up bull spreads when they are optimistic about the underlying
stock. We create a bull call spread by trading OPSY calls from our COP data, as follows:
- Buy the OPSY call with a strike price K1 = $20 for $3.50 and sell a call with K2 = $22.50 for $2
and draw their profit graphs (see Figure 15.4).
- If the stock price at expiration is zero, the loss of $3.50 on the first call is partially offset by a $2
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
gain on the second call, giving a loss of $1.50.
- The loss stays constant until S(T) reaches $20. Beyond this, the long call’s profit graph increases at
a 45 degree angle, while the short call’s profit graph remains flat. Consequently, the profit graph
for the bull spread, which was $1.50 below the horizontal axis, moves up dollar for dollar for each
dollar increase in S(T), cuts the x axis at the BEP = (20 + 1.50) = $21.50, and reaches $1 when
S(T) reaches $22.50.
- Beyond $22.50, the short call position decreases at a negative 45 degree angle. This decline
neutralizes the upward rise from the long call, and the net result is a flat line that is $1 above
the x axis.
■ This is a debit spread because it costs money to create the spread, as opposed to in a credit spread,
where you receive money in the creation process. You create a credit spread when you enter into
a bullish put spread by buying a put and simultaneously selling another put with a higher strike price.
For example, by buying the OPSY 20 put and selling the OPSY 22.50 put, you create a bull spread
that has a maximum loss of $1.50, a BEP at $21.50, and a maximum profit of $1—a profit graph that
happens to be identical to the graph of our previous bull call spread (see Figure 15.4).
322 CHAPTER 15: OPTION TRADING STRATEGIES
FIGURE 15.4: Vertical Bull Spread with Calls
Profit
Long call (K1 = 20)
21.5
20
2
1 Bullish call spread
0
–1.5 S(T)
–3.5
22.5
Short call (K2 = 22.5)
FIGURE 15.5: Vertical Bear
Spread with Puts
Profit
21
Long put (K2 = 22.5)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
21.5
1.5 Bearish put spread
0.5
0
–1 S(T)
–1.5
22.5
20
Short put (K1 = 20)
–19.5
Bear Spread
■ A bearish vertical spread is created by buying an option and simultaneously selling an option of the same
type, but with a lower strike price, where both options have the same underlying security and maturity
date. We create a bear put spread by trading OPSY puts as follows:
- Sell an OPSY put with strike price K1 = $20 for $0.50 and buy a put with K2 = $22.50 for $1.50
and draw their profit diagram (see Figure 15.5).
SPREAD STRATEGIES 323
- At the origin, the long OPSY put trade gains $21, but the short put position loses $19.50, giving
a payoff of $1.50, which is the maximum profit. For higher values of S(T), the profit is a flat line
because the long put declines but the short put rises, and the two effects cancel each other.
- The profit stays constant until S(T) reaches $20. Beyond this point, the long put’s profit continues
to decline at a 45 degree angle, whereas that for the short put remains flat. The profit graph for the
bear spread, which was $1.50 above the horizontal axis, decreases dollar for dollar for each dollar
increase in S(T), cuts the x axis at the BEP (20 + 1.50) = $21.50, continues declining, and reaches
–$1 when S(T) touches $22.50. Beyond this, the long put position also becomes a flat line, and the
payoff becomes a line parallel to the x axis.
- Notice that the bear put spread of Figure 15.5 is a mirror image of the bull put spread of Figure 15.4
(which is the same as a bullish call spread) across the horizontal axis. This is no surprise because the
two spreads involve identical options that are traded in opposite directions.
- You can similarly create a bear call spread by reversing the trades that led to the creation of the
bull call spread.
You can create profit diagrams that look like vertical spreads by trading the underlying asset and
options. This is demonstrated in Extension 15.1.
4 For a bullish horizontal spread, buy an option and sell another option of the same type and strike but with shorter life. For a bullish
diagonal call spread, buy an option with a lower strike price and longer time to maturity than the one that is sold. Do the opposite
for bearish spreads.
A spread can be used as a starting point for understanding more complex strategies.
For example, you can begin with a vertical spread and tinker with it by adding more
options of the same type. This leads to ratio spreads, butterfly spreads, condors, and
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
other options trading strategies, some of which are popular and others obscure.
Extension 15.1 shows how to do this. Moreover, one can find options spreads
embedded in other contexts. For example, a reinsurance contract is equivalent to
a call spread. This connection is explored in Extension 15.2.
EXTENSION 15.1: Collars, Ratio Spreads, Butterflies, and Condors
Consider a simple example. Suppose the portfolio your boss manages tries to mimic the S&P 500 index, but it
allows some tinkering. You advise buying some puts on the index with strike K1 with cost $p per option and
financing this by selling calls with the same price $c but a higher strike price K2 . Ext. 15.1 Fig. 1 shows that the
profit graph of this zero-cost collar (long stock, long put[s], and short call[s], where the total cost of the calls
equals that of the puts) is a spread.
The strategy has its trade-off. You are pleased if the underlying declines and the put protection works but
terribly glum if the underlying rises and the gains are surrendered to the call buyer. Collars are similar to zero-
cost collars but do not require the cash flows from the calls and puts to negate each other. You can establish a
collar on a long stock position by adding a covered call and a protective put.
324 CHAPTER 15: OPTION TRADING STRATEGIES
You can create butterfly spreads and condors by modifying a vertical spread. We demonstrate this by first
drawing a vertical bull spread with our COP data. Then we create a whole range of spread strategies by adding
options of the same type (calls with calls or puts with puts), but with different strike prices.
EXT. 15.1 FIG. 1: A Zero-Cost Collar
Profit
Long stock
Long put
K1
Zero-cost collar
c
0
–p
S(T)
K2
Short call
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Here call price c = put price p
EXT. 15.1 EX. 1: Examples of a Call Spread, a Ratio Spread, Butterflies, and
Condors
Bull Call Spread
One can create a bull call spread similar to the one in Example 15.3 by buying one OPSY 17.50 call for $5.50
and selling one OPSY 20 call for $3.50. The profit diagram is initially a flat line with a loss of $2, rises at a
45 degree angle as the stock price goes from $17.50 to $20, breaks even at $19.50, and again becomes a flat line
at a profit of $0.50 when S(T) exceeds $20 (see Ext. 15.1 Fig. 2).
Ratio Spread
■ Now sell another OPSY call with strike price $20. By selling two calls, you collect a premium of $3.50 × 2
= $7, and the profit diagram declines by $2 for each dollar increase in S(T) beyond $20. This creates a ratio
spread, which is defined as an options trading strategy in which the number of calls (or puts) purchased is
SPREAD STRATEGIES 325
different from the number of calls (or puts) sold. The profit graph for this one-by-two call spread is a flat
line emanating from (– 5.50 + 7) = $1.50 on the vertical axis, which then rises at 45 degrees when S(T)
exceeds $17.50 and attains the peak value of (1.50 + 2.50) = $4 when S(T) is $20 (see Ext. 15.1 Fig. 2).5
Beyond this, the long call moves the graph up, but the two short calls drag it down twice as fast, resulting in
a line that declines at a negative 45 degree angle.
Butterfly Spread
■ A butterfly spread is created by trading four options of the same type (all calls or all puts) with three different
strike prices: two options with extreme strike prices are bought (written) and two options are written (bought)
with the middle strike price (see Ext. 15.1 Fig. 2).
■ A butterfly spread is a two-sided hedge that bets on the volatility of the underlying security: a long butterfly
spread makes a small profit if the volatility is low and a small loss otherwise. The profit graph for this long
butterfly call spread starts at a maximum loss of (–5.50 + 7 – 2) = – $0.50 along the vertical axis. Initially a
line parallel to the horizontal axis, the profit graph starts increasing when S(T) exceeds $17.50, breaks even
at $18, attains a maximum value of $2 when S(T) is $20, declines and cuts the x axis again at $22, falls to
– $0.50– $0.50 at $22.50, and again becomes a line parallel to the x axis for all higher values of S(T).
■ A butterfly spread is so named because the profit diagram looks like a butterfly’s wings. If you reverse the
preceding trades, you will create a short butterfly spread. You can also create butterfly spreads with puts.
Condor
■ Thinking that OPSY might fluctuate more than anticipated, you decide to expand the region over which
you want to profit. You modify the butterfly spread by (1) keeping a long call with K0 = $17.50 as before,
(2) (instead of two short calls at the same strike) selling one call at the next higher strike price K1 = $20 and
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
another at K2 = $22.50, and finally, (3) buying a call with the uppermost strike K3 = $25 (see Ext. 15.1 Fig. 2).
This creates a condor spread, which has four options with four strike prices: long two options with extreme
strikes and short two options with strike prices in the middle, and vice versa. The wings of the condor, the
largest flying land bird in the western hemisphere, inspire the name.
■ Our condor’s profit diagram begins on the y axis at (–5.50 + 3.50 + 2 – 1) = – $1, the amount obtained by
adding up the premiums. Initially, it’s a line parallel to the x axis because the profits from the calls are flat at this
point. When S(T) goes beyond 17.50, the call with the lowest strike becomes active and increases the payoff.
The profit graph reaches a maximum value of $1.50 when S(T) is 20. It becomes a flat line for S(T) lying
between $20 and $22.50 because the call with a strike of K0 = $17.5 pulls it up but the call with K1 = $20
pushes it down. Beyond $22.50, the third call (with K2 = $22.50) kicks in, and the profit graph falls. Finally,
when S(T) exceeds $25, the fourth option (with K3 = $25) also becomes active and increases the payoff: with
two calls pulling it up and two pushing it down, it again becomes a flat line. You can also create a condor with
puts.
326 CHAPTER 15: OPTION TRADING STRATEGIES
EXT. 15.1 FIG. 2: Bull Call Spread, Ratio
Spread, Butterfly, and
Condor
Profit
17.5 Long call (K0 = 17.5)
3.5
0.5 Bullish call spread
0
–2 S(T)
–5.5 Short call (K1 = 20)
20
Profit
7
Long call (K0 = 17.5)
4 17.5
1.5
0
1 S(T)
–5.5
1 Call ratio spread
20
2
1 Two short calls (K1 = 20)
Profit
Long call (K0 = 17.5)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
7 Long call (K2 = 22.5)
17.5 22.5
2
0
–0.5
–2 S(T)
–5.5 Butterfly call spread
20
Two short calls (K1 = 20)
Profit
17.5 25
1.5
0
–1 S(T)
Condor spread
20 22.5
SPREAD STRATEGIES 327
5
A call ratio spread typically involves buying calls and selling a greater number of calls at a higher strike price. If you sell calls and buy a
greater number of calls at a higher strike price, then you create a call backspread.
EXTENSION 15.2: Reinsurance Contracts and Call Spreads
Reinsurance contracts on a loss causing event (e. g. hurricanes) are equivalent to a call spread in which the call
options are written on the losses of the insured event. To understand this equivalence, consider the following
example.
Risks Facing the Insurance Industry
Consider some of the risks in the insurance industry.
■ Dearth of natural hedges. Insurance is a high risk business that cannot be hedged using traded financial
securities. Insurance companies employ applied mathematics and statisticians called actuaries, who calculate
the likelihood of loss events and fair premiums for insuring these risks.
■ Diversification and the law of averages. Diversification is the industry’s key mantra. Insurers try to manage risks
by diversifying their losses across a large number of policyholders. The business survives on the law of large
numbers, which says that as more and more independent and identically distributed observations are collected,
the average gets closer and closer to the mean value for the population. Although no one can foretell whether
any particular insured entity will suffer a loss, actuaries can estimate with reasonable accuracy the expected
losses in a large population. They compute fair premiums based on these expected losses. Insurance companies
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
lose money when the realized losses exceed the expected losses.
■ Reinsurance. For catastrophic events like major earthquakes and hurricanes, the law of large numbers does not
apply because these events are not independent and identically distributed across the population. Insurance
companies hedge their losses on such catastrophies by buying their own insurance. This is called reinsurance.
We discuss reinsurance after the following example.
EXT. 15.2 EX. 1: Hedging Hurricane Risk
■ Catastrophic Insurance Corporation (CatIns Corp., a fictitious name) sells homeowner’s insurance contracts
to one hundred thousand homes along the shores of the Gulf of Mexico in the states of Florida, Alabama,
Mississippi, Louisiana, and Texas. The contract provides protection against hurricane risk, which is the risk
of losses coming from powerful storms, and it has the following features:
- Annual premium $15,000.
- Fixed-amount deductible $10,000.
- Maximum coverage amount $300,000 (which is less than the value of each home).
- Contract pays for losses from one hurricane in a given year. For simplicity, assume that if a hurricane hits,
it does the same dollar damage to all homes covered by the insurance policy. These losses manifest the
328 CHAPTER 15: OPTION TRADING STRATEGIES
inability of the insurance company to diversify these losses in a large pool of the insured. Also assume that
underwriting costs and investment income net out to zero.
■ We draw the profits for CatIns in a modified profit diagram in which “profit per contract” for the insurance
company is plotted (as before) along the vertical axis and the underlying “loss per home” (instead of underlying
asset price) is depicted along the horizontal axis. Of course, you can draw the profits in a usual profit diagram.
However, this representation is more natural in the context of insurance, where losses are key. We determine
the profits as follows (see Ext. 15.2 Fig. 1; both variables are graphed in thousands of dollars).
EXT. 15.2 FIG. 1: Call Spread Diagram
Profit per contract
(thousands of dollars)
15
12 25
0
100 300 Loss per home
10 (thousands of dollars)
22
–78
CatIns’s profits (with reinsurance)
CatIns’s profits
(no reinsurance)
–275
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
No Hurricane
CatIns keeps the entire premium of $15,000, which is shown as 15 in the diagram. For one hundred thousand
homes, this turns out to be $1.5 billion.
If Hurricane Damage Happens
For the first $10,000 in losses, the company pays nothing. The profits (shown by a shaded line) remains parallel
to the horizontal axis at $15,000. For higher losses, the company pays a dollar for each dollar increase in loss.
The profit line declines at a 45 degree angle and cuts the x axis when the total loss reaches $25,000, shown as
25 in the diagram. Beyond this, the profit declines until the total loss on each home reaches the maximum loss
of $300,000, where CatIns pays $275,000. This is a profit of $27.5 billion, a staggering sum considering that the
company earned only $1.5 billion in premiums.
Insurer Buys Reinsurance
CatIns can buy reinsurance that pays for losses to a home over $100,000 for a premium of $3,000 payable to a
reinsurance company. Then, the new payoff line emanates from the y axis at (15,000 – 3,000) = $12,000, stays
COMBINATION STRATEGIES 329
parallel to the x axis until “loss per home” reaches $10,000, goes down thereafter at a 45 degree angle, cuts the hor-
izontal axis at $22,000, reaches a maximum loss of (100,000 – 22,000) = $78,000 when “loss per home” reaches
$100,000, and again becomes parallel to the x axis because the reinsurance company pays for the higher losses.
Equivalence to a Call Spread
This loss diagram shows that the profits to CatIns with reinsurance are equivalent to a call spread on the hurricane
losses. The company makes a limited profit if the hurricane’s losses are mild, and the company incurs losses that
are capped at $7.8 billion when large hurricanes hit homeowners.
Reinsurance Markets
Reinsurance has a long history. One of the earliest known reinsurance contract agreements was written in Latin
and signed by Italian merchants on July 12, 1370. Reinsurance helps insurance companies manage tail risk, which
refers to the risk of occurrence of infrequent events that cause large losses when they occur. Tail risk is so named
because these large loss events lie on the extreme end (“tail end”) of a probability distribution of such loss events.
As illustrated by the example, an insurance company’s profit diagram under a typical reinsurance contract
is equivalent to a call option spread. The call options are written with a one-year maturity on the aggregate
losses realized by the insurance company, where the strikes correspond to the initial deductible and the payment
cap. Given the call spread analogy, reinsurance contracts can be priced using the option pricing methodologies
presented in Chapters 17–20.
15.7 Combination Strategies
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
“A double privilege pays a profit, no matter which way the market goes, and costs
$212.50,” declared an advertisement in 1875 in the book Secrets of Success in Wall Street
by Tumbridge and Co., Bankers and Brokers, which had its office on 2 Wall Street,
New York City at that time. An ancestor to today’s booklets that aim at informing and
attracting traders to options, this forty-eight-page volume describes the workings of
the New York Stock Exchange and the over-the-counter options market. The traded
contracts went by the name stock privileges and were essentially American options:
calls, puts, spreads, and straddles that were customarily written on one hundred shares
of stock and matured in thirty days. Tumbridge charged $100 as the premium for a call
or a put and a $6.25 broker’s commissions for each leg of the trade. Double privilege
referred to spreads and straddles, which were created by trading two options, and
hence charged two commissions.
Straddles and strangles are examples of combination strategies that combine
options of different types on the same underlying stock and expiring on the same
date, where the options are either both purchased or both written. For example,
buying a call and a put with the same strike price creates a straddle (called a put-
to-call strategy in London); if their strike prices are different, you get a strangle.
330 CHAPTER 15: OPTION TRADING STRATEGIES
Straddles and strangles are volatility plays. Options traders use them to bet on risky
future events affecting a company that can impact the stock price in either a positive or
negative direction. Traders sometimes write them when they expect neutral markets.
EXAMPLE 15.4: Straddles and Strangles
Straddle
■ You buy a straddle (a long straddle or a bottom straddle) by buying a call and a put as follows:
- Buy OPSY 22.50 options, the call for $2 and the put for $1.50, and draw the profit diagram (see
Figure 15.6).
FIGURE 15.6: Straddles and
Strangles—Examples of
Combination Strategies
Profit
21
19 Long put
Long call
22.5 Straddle
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
19 26
0
–1.5 S(T)
–2
–3.5
Profit
19.5
17.5 Long put
Long call
Strangle
20 22.5
17.5 25
0
–0.5 S(T)
–2
–2.5
SUMMARY 331
- If the stock price at expiration is zero, the long put makes $21, but the long call loses $2, giving a
gain of $19. As S(T) increases, the long call’s payoff remains parallel to the horizontal axis, but the
long put drags the straddle’s payoff down at 45 degrees. The payoff cuts the x axis at the breakeven
point $19, keeps declining, and reaches the nadir—a loss of $3.50.
- For values of S(T) higher than $22.50, the long put’s payoff becomes parallel to the x axis, but
the long call pulls the straddle’s payoff up at 45 degrees, which cuts the x axis again at the BEP =
(22.50 + 2 + 1.50) = $26. For higher values of S(T), the payoff increases at 45 degrees, reflecting
infinite profit potential.
■ If you reverse the preceding trades and sell a call and put, then you sell a straddle (establish a top
straddle or a short straddle), a bet on low future volatility whose payoff is the mirror image of the
bottom straddle across the horizontal axis.
Strangle
■ A long strangle is created by buying options with different strike prices, where the strike price of
the put is lower than the strike price of the call. Though it’s a volatility bet like a straddle, a strangle
does have some differences:
- Buy the OPSY 22.50 call for $2, and buy the OPSY 20 put for $0.50. Plot these two strategies in
a profit diagram (see Figure 15.7).
- At the origin, the profits for the strangle is 19.50 – 2 = $17.50. For higher values of S(T), it declines
at a 45 degree angle until it reaches –$2.50 when S(T) is $20. For the stock price lying between
$20 and $22.50, the profit is flat because both the call’s and put’s profits are parallel to the horizontal
axis. Beyond $22.50, the long put’s profit stays flat, but the long call pulls up the strangle’s profit,
which cuts the x axis at $25 and keeps increasing at a 45 degree angle, indicating an unlimited profit
potential.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
15.8 Summary
1. Options trading strategies’ performances are drawn in profit diagrams for European
options with a single stock as the underlying. The basic building blocks for options
trading are simple naked (or uncovered) strategies that are stand-alone positions
buying or selling a single call or put option.
2. Hedged strategies combine options and stocks in a way that dampens the overall
risk as compared to a naked option trade. They are usually covered strategies
because they tend to back the option transaction with a stock that protects (or
covers or collateralizes) it. Hedged strategies include covered call writing (long
stock plus short call) and covered put buying (long stock plus long put).
3. Spread strategies combine several options of the same type (either both calls or
both puts) on the same underlying but on different sides of the market. These
two-sided hedges give a small profit if the underlying moves in one direction but
suffer a small loss otherwise.
332 CHAPTER 15: OPTION TRADING STRATEGIES
4. Spreads come in three basic types: (1) vertical spreads involve options that have
different strike prices but expire on the same date, (2) horizontal spreads involve
options that have the same strike but different maturity dates, and (3) diagonal
spreads involve options that differ both in terms of strike price and maturity date.
5. Several other categories of spread strategies can be created by adding options of
the same type, with the same or different strike prices, to a spread strategy:
- A ratio spread has the number of calls (or puts) bought different from the number
of calls (or puts) sold.
- A butterfly spread has four options of the same type (all calls or all puts) with
three different strike prices: buy (sell) two options with extreme strike prices
and write (purchase) two options with a middle strike price.
- A condor spread has four options with four strike prices: purchase (write) the
two extreme options and sell (buy) two options with strike prices in the middle.
6. Combination strategies combine options of different types on the same underlying
stock and expiring on the same date, where both options are either purchased or
written. A straddle combines a call and a put with the same strike price. A strangle
combines a call and a put with different strike prices.
15.9 Cases
Cephalon Inc. (Harvard Business School Case 298116-PDF-ENG). The case intro-
duces students to the use of equity derivatives as part of a risk-management
strategy, examines the application of cash flow hedging in a corporate context,
and explains the pricing of a derivative security with large jump risk.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
National Insurance Corp. (Harvard Business School Case 296036-PDF-ENG).
The case considers whether a major reinsurer should use exchange-traded insur-
ance derivatives for managing catastrophic risk.
Compensation at Level 3 Communications (Harvard Business School Case
202084-PDF-ENG). The case analyzes a compensation plan that rewards managers
for the firm’s performance only if the firm’s stock price movement exceeds a
benchmark.
15.10 Questions and Problems
The next eleven questions are based on the following data for a fictitiously named company,
OOPS.
■ Current stock price S is $22.
■ Time to maturity T is six months.
■ Continuously compounded, risk-free interest rate r is 5 percent per year.
■ European options prices are given in the following table:
QUESTIONS AND PROBLEMS 333
Strike Price Call Price Put Price
K1 = $17.50 $5 $0.05
K2 = 20 3 0.75
K3 = 22.50 1.75 1.75
K4 = 25 0.75 3.50
15.1. Use options with strike K2 = $20.00:
a. Draw a long call profit diagram.
b. Draw a short call profit diagram.
c. Draw a long put profit diagram.
d. Draw a short put profit diagram.
15.2. Use options with strike K2 = $20.00:
a. Draw a covered call (long stock plus short call) profit diagram.
b. Give a reason why a trader might want to hold a covered call position.
c. Explain the difference between a buy–write and an over-write strategy.
15.3. Use options with strike K2 = $20.00:
a. Draw a covered put (long stock plus long put) profit diagram.
b. Give a reason why a trader might want to hold a covered put position.
c. Explain the difference between a married put and a protective put.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
15.4. Use options with strike K3 = $22.50:
a. Draw the profit diagram for a long call and a short put position for options
with the same strike and maturity date.
b. Draw the profit diagram for a long stock and two short calls.
15.5 Using options with strike price K3 = $22.50, draw the profit diagram for a
long call, a short stock, and a short put position for options with the same
strike and maturity date. What other investment has this profit diagram?
15.6. a. What is a bullish vertical spread?
b. Draw a bullish vertical spread by trading put options with strike prices K2 =
$20 and K4 = $25.
15.7. a. What is a bearish vertical spread?
b. Draw a bearish vertical spread by trading call options with strike prices
K1 = $17.50 and K3 = $22.50.
15.8 Draw a butterfly spread by going long calls with strike prices K1 = $17.50 and
K3 = $22.50 and selling short two calls with a strike price in the middle.
334 CHAPTER 15: OPTION TRADING STRATEGIES
15.9 Draw a condor spread by going long calls with strike prices K1 = $17.50 and
K4 = $25 and selling short two calls with each strike price in the middle.
15.10. a. What is the aim of a long (or bottom) straddle strategy?
b. Create a long straddle by buying a call and a put with strike price K3 =
$22.50.
15.11. a. What is the aim of a short (or top) strangle strategy?
b. Create a short strangle by writing a call with strike price K3 = $22.50 and
a put with strike price K1 = $20.
The next five questions are based on the following options price data for Tel Tales
Corporations (fictitious name), where the options expire on the same date in May.
Draw profit diagrams in each case, clearly showing the stock price corresponding to zero
profit, the maximum profit and loss, and so on.
Stock Strike (K) Call Price Put Price
$29 $25 $5 $1
30 2 3
35 1 6
15.12 A 3:1 reverse hedge (buy three May 30 calls and short the stock).
15.13 A bullish spread (long call with strike price of $25 and short call with strike
price of $30).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
15.14 A butterfly spread (long put with strike prices $25 and $35 and short two puts
in the middle).
15.15 A strangle (buy call with K = $30 and buy put with K = $25).
15.16 A straddle (buy a call and a put with K = $30).
15.17 Goldminers Inc. mines and refines ore and sells pure gold in the global market.
To raise funds, it sells a derivative security whose payoff is as follows:
■ Part of the security is a zero-coupon bond (which is sold at a discount and
makes no interest payments) that pays a principal of $1,000 at maturity T.
■ Goldminers also pays an additional amount that is indexed to gold’s price
(per ounce) at maturity S(T):
0 if S (T) ≤ $1,350
$30 [S (T) − 1,350] if $1,350 < S (T) ≤ $1,400
$1,500 if $1,400 < S (T)
Analyze this derivative as a combination of bond and put options.
QUESTIONS AND PROBLEMS 335
15.18 a. What is a collar in the options market?
b. How would you create a zero-cost collar?
c. Why might a copper manufacturer find it useful to employ this strategy?
15.19 An insurance company has insured oil fields in the Middle East. Next, it
purchases reinsurance to manage its “tail risk.” How can the reinsurance
company hedge some of its risks by trading derivatives?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
16
Option Relations
16.1 Introduction 16.5 Options Price Restrictions
16.2 A Graphical Approach to EXTENSION 16.3 The Superglue
Put–Call Parity Argument
16.3 Put–Call Parity for European 16.6 Early Exercise of American
Options Options
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
No Dividends
16.4 Market Imperfections
Dividends and Early Exercise
EXTENSION 16.1 Put–Call
Parity in Imperfect Markets 16.7 Summary
EXTENSION 16.2 Dividends and 16.8 Cases
American Options
16.9 Questions and Problems
A GRAPHICAL APPROACH TO PUT–CA LL PARITY 337
16.1 Introduction
Eager to find how the options on Palm Inc.’s stock were performing on their inaugural
trading day, he switched on his computer, entered the Internet, and logged on to
his online brokerage account. Like Alice in Lewis Carroll’s children’s classic Alice’s
Adventures in Wonderland, he grew “curiouser and curiouser” and could hardly believe
his eyes. Market-traded put options were trading $4 higher than the arbitrage-free
price predicted by put–call parity. Filled with enthusiasm, he established a game plan
based on the textbook prescription to exploit the mispricing, but he stumbled at
the outset. He could not find Palm shares to borrow and sell short, a critical step in
constructing the arbitrage. He (one of us) did not make millions that morning but
acquired knowledge and, most important, a story to share with you! Later in this
chapter, we will discuss the Palm arbitrage in greater detail.
The last chapter analyzed various options trading strategies using their profit
diagrams on the maturity date. Determining an option’s price prior to expiration
is a far more difficult task. Although it’s easy to find the forward price using the
no-arbitrage argument because of the forward’s linear payoff, the nonlinear option
payoffs complicate the argument. In fact, to get an exact value, one needs to assume
an evolution for the underlying stock price (which indicates how the stock price
evolves through time), the most popular being a lognormal distribution. This is
studied in subsequent chapters. But what can you learn about options properties
without assuming such an evolution? This is the topic studied in this chapter.
Here we study three different categories of options price relations. First, we
establish put–call parity for European options. This fundamental relation stitches
together the stock, a bond, and call and put options. Put–call parity is not new:
financier Russell Sage used it to circumvent New York State usury laws in the late
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1800s. Next, we examine some restrictions that options prices must satisfy under
the no-arbitrage assumption. These provide the key steps to our final topic: the early
exercise of American options. As early exercise features are embedded in various other
financial securities, it is important to understand this topic well.
Throughout this chapter, we maintain our standard assumptions: no market frictions,
no credit risk, competitive and well-functioning markets, no intermediate cash flows (such as cash
dividends), and no arbitrage. However, we will occasionally need to relax the assumption
of “no intermediate cash flows” below.
16.2 A Graphical Approach to Put–Call
Parity
In The Put-and-Call, published in London in 1902, Leonard R. Higgins describes
the workings of London’s over-the-counter options market. In this book, he makes
the following intriguing observation: that “a put can be turned into a call by buying
all the stock” and that “a call can be turned into a put by selling all the stock.” As
observed by Higgins, a clever arrangement of stocks and European options establishes
put–call parity (PCP). Initially, we illustrate this result with payoff diagrams in the next
example, and later by other methods.
338 CHAPTER 16: OPTION RELATIONS
EXAMPLE 16.1: Establishing Put–Call–Parity by Payoff Diagrams (Replicating
a Call)
■ We can replicate a European call option by trading the underlying stock, a European put option,
and zero-coupon bonds. The options on the stock need to have a common strike price and a
common expiration date. For simplicity we use the common options pricing data (COP data) from
Chapter 15.
■ First, buy an OPSY 20 put option whose time T payoff diagram is drawn in Figure 16.1A. Next,
buy one OPSY stock and place its payoff diagram below that of the put. Now create a third payoff
diagram that sums the first two payoffs for each value of S(T). This payoff graph for the long put plus
the stock starts at a value of K = $20 on the vertical axis. When S(T) exceeds $20, the graph rises at
a 45 degree angle to the x axis.
■ For convenience, we redraw this portfolio payoff in Figure 16.1B. Draw another diagram underneath
this depicting the payoff from shorting a bond with a future liability of $20. The present value of this
cash flow today is $20B, where B is today’s price of a zero-coupon bond that pays a dollar after six
months. The short bond’s payoff is a line parallel to the x axis at – $20.
■ Now sum the payoffs of the previous positions (Long put + Long stock) and short bonds. The
portfolio (Long put + Long stock + Short bonds) replicates the payoff to a long call option with
strike price K.
■ As the time T payoffs are the same, the law of one price equates the value of the two portfolios today:
Long call = Long put + Long stock + Short bond (16.1)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Expression (16.1) is put-call parity (PCP) for European options. We can view the left side, the call, as
a market-traded asset and the portfolio on the right side as a synthetic call (a portfolio that synthetically
constructs the traded call’s payoffs).
16.3 Put–Call Parity for European Options
We follow the data–model–arbitrage approach to establish put-call parity (PCP) for
European options. As in the previous section, we consider a market-traded call and
construct a synthetic call with identical future payoffs. By the law of one price, the
traded and the synthetic call must have the same value today. This yields the PCP
relation. In this example, instead of the law of one price, we use the equivalent
nothing comes from nothing principle instead.
PUT–CALL PARITY FOR EUROPEAN OPTIONS 339
FIGURE 16.1A: Step One (PCP FIGURE 16.1B: Step Two (PCP
by Graphs): by Graphs):
Long Put plus (Long Put plus
Long Stock Long Stock)
plus Short
Payoff Bonds gives a
Long Call
20 Long put
Payoff
Long put + Long stock
0 20
K = 20 S(T)
Payoff 0
20 S(T)
Long stock
Payoff
0
0 S(T)
S(T)
–20 Short bonds
Payoff
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Long put + Long stock Payoff
20
Long call
0 0
20 S(T) K = 20 S(T)
EXAMPLE 16.2: Put–Call Parity for European Options
The Data
■ Consider the COP data of section 15.4 of Chapter 15.
- OPSY’s stock price S is $22.50 at time 0.
- The European options on OPSY have a common strike price K = $20 and mature in T = 6 months.
- Today’s put price p is $0.50. The call price c is to be determined.
340 CHAPTER 16: OPTION RELATIONS
- The continuously compounded risk-free interest rate r is 5 percent per year. Today’s price of a
zero-coupon bond paying $1 at time T is B = e−rT = $0.9753.
- No dividends are paid on OPSY’s stock over the option’s life.
■ All trades are recorded in an arbitrage table. The first column of Table 16.1 gives the trade description
and the second column records today’s cash flows. The expiration date payoffs are presented in the
last two columns. The third column reports the payoffs when S(T) is less than or equal to $20, and
the last column gives the payoffs when it’s greater than $20.
■ We first implement a stock purchase plan portfolio.
- Buy the stock in the spot market for $22.50. The cash flow is recorded as −22.50 in the second
column. The stock is worth S(T) after six months. Jot down S(T) in both the third and fourth
columns.
- Tap the bond market to borrow the present value of the strike price K = $20 by shorting zero coupon
bonds. This cash flow reduces today’s payment. Record 0.9753 × 20 today and the liability -20 in
the last two columns.
■ Next, using the options market, we create a synthetic short stock portfolio.
- Sell one OPSY 20 call. Record this trade as +c in the second column. The call expires worthless if
S(T) is less than or equal to $20. For higher values of S(T), the future payoff is a liability. Surrender
the stock worth S(T) and receive $20 in return, for a cash flow of [20 − S(T)].
TABLE 16.1: Arbitrage Table for Put–Call Parity
Expiration Date (Time T) Cash Flow
Portfolio Today (Time 0) Cash Flow
S (T) ≤ 20 20 < S (T)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Stock purchase plan
Long stock −22.50 S(T) S(T)
Short bonds to borrow 0.9753 × 20 −20 −20
the present value of strike
(K = $20, B = $0.9753)
Synthetic short stock
Short call c 0 20 − S(T)
Long put −0.50 20 − S(T) 0
Cash flows −22.50 + 0.9753 × 20 + c 0 0
−0.50
- Buy an OPSY 20 put. The premium p = $0.50 is written as a negative cash flow today. If S(T) is less
than or equal to $20, then the put holder receives $20 but has to surrender the stock worth S(T).
Record the put payoff as [20 − S(T)] in this case. The put expires worthless when S(T) exceeds
$20.
PUT–CALL PARITY FOR EUROPEAN OPTIONS 341
■ The first two trades (Long stock + Short bond) have the payoff [S(T) − 20] at expiration. The last
two trades (Short call + Long put) create an equal but opposite payoff on the expiration date.
■ The final row of Table 16.1 considers net cash flows. The portfolio’s payoff is always zero on the
expiration date.
■ The portfolio of stocks, bonds, and options created has a zero value for sure in the future. To
prevent arbitrage, using the nothing comes from nothing principle, it must have zero value today.
Consequently,
−22.50 + 0.9753 × 20 + c − 0.50 = 0
or c = $3.49
This is the European call option’s arbitrage-free price.
The Model
■ Use symbols to generalize. Replace $22.50 with S, $0.9753 × 20 with B × K, and $0.50 with p to
get the PCP for European options:
− S + BK + c − p = 0 (16.2)
Arbitrage Profits
■ A violation of PCP leads to arbitrage. Suppose you find an errant trader who quotes a call price of $4.
As this is higher than the arbitrage-free price from PCP, create the portfolio in Table 16.1: buy the
stock, short the bonds, buy the put, and sell the relatively overpriced call. This gives the cash flows
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
−S + BK + c − p
= −22.50 + 0.9753 × 20 + 4 − 0.50
= $0.51
as your immediate arbitrage profit.
■ If the trader quotes $3 instead, reverse the trades and make $0.49 as an arbitrage profit.
We use expression (16.2) to formally state our result.
RESULT 16.1
Put–Call Parity for European Options
c = p + S − BK (16.3)
where c and p are today’s prices of European calls and puts, respectively, that
have strike price K and expiration date T, B is today’s price of a zero-coupon
bond that pays a dollar at time T, and S is today’s price of the underlying stock.
342 CHAPTER 16: OPTION RELATIONS
PCP is a useful tool for understanding how to use options in forming trading
strategies. We start with the put because it is the easiest option to understand. Indeed,
a put option is like an insurance policy because it insures the value of the underlying
stock at its strike price. With this insight, we can rewrite PCP as
Call = Insurance + Stock + Borrowing (16.4)
We see that a call option is equivalent to buying the stock on margin (by borrowing)
and protecting the stock purchase with an insurance policy.
This intuition can help us to understand how to trade on information using
options. For example, suppose that you are the only person in the market who
knows that the stock is going to rise above the strike price. (Note: this is a
thought experiment—we are not considering issues like whether trading on inside
information will attract jail time!) What is the best way to trade on this information:
buying a call? No. This is the wrong answer because by buying the call, you are also
paying for the put, which is insurance on the stock that you do not need. Recall
that you are certain that the stock price will rise. In this case, the optimal strategy is
to buy the call and sell the put. Using PCP again, we also see that this is equivalent
to buying the stock and financing the purchase with borrowed cash, which creates
leverage (called gearing in Europe).
Now suppose that you expect that the stock is likely to increase but you are unsure.
In this case, is buying the call the best strategy? Maybe. The answer really depends
on your risk tolerance. Pay for the put if you like the downside insurance protection
but not otherwise.
16.4 Market Imperfections
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Market imperfections introduce difficulties, but they may also create opportunities.
Merton Miller’s observation that financial securities can help overcome rules and reg-
ulations becomes pertinent. Options and PCP can provide tools to work around these
imperfections. In the article “The Ancient Roots of Modern Financial Innovation:
The Early History of Regulatory Arbitrage,” Michael Knoll (2008) cited ancient
examples of merchants using PCP to get around restrictions on interest payments.
For example, two millennia back, Israelite financiers used PCP in this way. English
financiers did the same five hundred years ago, and interestingly, the transaction they
devised led to the development of the modern mortgage. These examples illustrate
tax or regulatory arbitrage, which involves unpackaging and rebundling cash flows
with the aim of profiting from otherwise prohibited transactions. Extension 16.1
discusses an example of PCP in action.
One can tinker with PCP to generate two extensions: (1) to accommo-
date dividends and (2) for American options. These extensions are shown in
Extension 16.1 and 16.2.
MARKET IMPERFECTIONS 343
EXTENSION 16.1: Put–Call Parity in Imperfect Markets
PCP Violations for Palm Shares
Sometimes PCP fails because of market imperfections. This happened with Palm Inc. options in an incident
discussed at the chapter’s start.
On March 2, 2000, 3Com did an equity carve-out of its subsidiary Palm, which had made the world’s first
successful handheld computer, the Palm Pilot, in 1992. Accordingly, 3Com sold 5 percent of Palm shares to the
general public in an initial public offering and declared that later in the year, it would distribute the remaining
shares in a spin-off to its existing shareholders (who would get 1.525 shares of Palm for every share of 3Com they
held). The mispricing that resulted in the equity and options markets got extensive press coverage and academic
study (see Lamont and Thaler 2003).
A hot company during the days when tech stocks led the dotcom stock market craze, Palm had an unbelievable
initial public offering (IPO). Its stock closed at the end of the first day’s trading at $95.06 per share. This implied
that 3Com’s stock price, which would equal 1.5 Palm shares plus the value of 3Com’s other assets and businesses
was $63 per share, or – $22 billion. The mishmash confused options traders. They adopted strategies like writing
call options and short selling 3Com stock to buy Palm stock, whose market capitalization then exceeded the
parent company’s valuation! “There is absolutely no rationale behind it,” interjected a trader. Predictably, Palm
started declining after the IPO.1
When options began trading on March 16, 2000, Palm stock traded around $55. For options maturing in one
month, the at-the-money (strike price K = $55) call price c was $5 and the put price p was $9. These numbers
are illustrative. The bid/ask spread was large, and the market was in a state of confusion, making it hard to find
accurate prices. A zero-coupon bond price paying $1 after one month was worth $0.995. A quick check of
put–call parity suggests
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Synthetic put price = Call price + Present value of strike –Stock price
= 5 + 0.995 × 55 − 55
= $4.725
And yet the price of the market-traded put was $9. Isn’t this an arbitrageur’s dream come true? Wouldn’t this
mean that if you sell the market-traded put and buy the synthetic put (see expression [16.2]), you can immediately
capture an arbitrage profit of 9 − 4.725 = $4.275 in one trade? Repeat a million times, and you become a
millionaire four times over!
Unfortunately, a market imperfection disrupted this trade. To create the arbitrage portfolio, an “arbitrageur”
needs to sell the relatively overpriced market-traded put and “buy” the relatively underpriced synthetic put via
buying the call, shorting the stock, and lending the present value of the strike. In this circumstance, however, the
broker could not supply the Palm shares for short selling (by borrowing them from another investor). The scarcity
of shares for short selling prevented the realization of the arbitrage profits and led to the divergence between theory
and practice. The situation did not last long. The price anomaly disappeared within days; however, a lesson was
learned. Paper profits can differ from actual realizable profits, and market imperfections do matter.
1
Palm IPO Soars, Then Retreats a Bit, Pushing Traders to Unwind Options in Parent 3Com,” W allStreetJournal, March 3, 2000.
344 CHAPTER 16: OPTION RELATIONS
EXTENSION 16.2: Dividends and American Options
Put–Call Parity Adjusted for Dividends
The parallels are eerie. One can easily modify the basic PCP by using the tools developed in Chapter 12 for
adjusting the cost-of-carry model for dividends. The intuition remains the same as before. A dividend lowers the
cost of a stock purchase, and replacing the stock price with the “stock price net of all dividends over the life of
the derivative” modifies the formula.
Assuming that a fixed-dollar dividend div is paid on a known future date t1 and B1 is today’s price of a zero
coupon bond maturing on the ex-dividend date (recall Result 12.1a), put–call parity for European options
(with a known dollar dividend) can be written as
c = p + (S − B1 div) − BK (1)
As before, this can be easily generalized to accommodate multiple dividends.
For options on indexes, we can use the insights from Result 12.1b. If the underlying pays dividends at a con-
tinuous rate 𝛿, then put–call parity for European options (with a known dividend yield) can be written as
c = p + e−𝛿T S − e−rT K (2)
A Put–Call Parity Inequality for American Options
The basic PCP does not work for American options because the option may get exercised before the expiration
date. However, a PCP inequality for American options does hold. For these inequalities, we add the assumption
that interest rates are nonnegative, or equivalently, that a zero-coupon’s price is always less than or equal to 1.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
EXT. 16.2 EX. 1: A Relation between American Calls and Puts
The Data
■ The stock price is S today. Assume that the stock pays no dividends over the life of the options.
■ American calls and puts worth cA and pA , respectively, trade on the stock. They have the same strike price K
and expire at time T but may get exercised early at time t.
■ Interest is compounded continuously at the risk-free rate r percent per year. You can invest in zero-coupon
bonds or a money market account (mma). B is today’s price and B(t) is the price at time t of zeros that mature
at time T. An investment of $1 in a mma today grows to [1 + R(t)] on the intermediate (exercise) date t and
to (1 + R) on the expiration date.
When the Portfolio Involves a Long Put (and Short Call)
■ Create a portfolio similar to the basic PCP: buy one put, write one call, buy one stock, and short bonds to
borrow the present value of the strike. Today’s portfolio value is
pA − cA + S − BK (3)
MARKET IMPERFECTIONS 345
You decide when to exercise the put. Your counterparty decides when to exercise the American call. Thus
two things can happen in the future: the holder of the American call exercises early, or she does not. Your
response depends on Long’s decision on the call.
■ If the American call is not exercised early, then the American call becomes equivalent to the European call
worth c. You can write cA = c. Now, if you don’t exercise your American put early, then it behaves like
a European put, which means pA = p. With cA = c and pA = p, expression (3) becomes equal to 0 by PCP
for European options. But you can exercise your American put before it expires. This flexibility makes the
American put worth at least as much as the European put. Thus (3) becomes
pA − cA + S − BK ≥ p − c + S − BK = 0 (4a)
■ If your counterparty exercises the American call early, then portfolio (3) has a payoff on the exercise date t (when
you buy back the bonds by paying B[t] for each zero):
pA (t) − [S (t) − K] + S (t) − B (t) K = pA (t) + [1 − B (t)] K ≥ 0 (4b)
This is nonnegative because a put cannot have a negative value and B(t) is worth less than or equal to a dollar.
■ Thus the portfolio (pA − cA + S − BK) has a value greater than or equal to zero. Combining (4a) and (4b) and
moving cA to the right side, we get our first inequality
pA + S − BK ≥ cA (5)
When the Portfolio Involves a Short Put (and Long Call)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Create the portfolio: sell the put, buy the call, short the stock, and invest K dollars in a mma. Notice that
this differs from the basic PCP portfolio because more dollars are kept in case the put is exercised early. This
portfolio’s value today is
− pA + cA − S + K (6)
Here the buyer determines whether to exercise the American put. As before, we will act in response to Long’s
decision.
■ If there is no early exercise, then the American put is the same as the European put, pA = p. If you don’t exercise
the American call early, then it behaves like the European call, which means cA = c. With cA = c and pA = p,
(6) becomes equal to KR on the maturity date (you can verify this in a payoff table). Because the last quantity
is greater than zero, (6) cannot be less than zero today:
− pA + cA − S + K ≥ 0 (7a)
Because you can exercise your American call early, this makes it even more valuable, which reinforces the
relationship in (7a).
346 CHAPTER 16: OPTION RELATIONS
■ If the buyer exercises early the American put, then portfolio (6) has a payoff on the exercise date t (when K has
grown to K[1 + R(t)]):
− [K − S (t)] + cA − S (t) + K [1 + R (t)]
(7b)
= KR (t) + cA ≥ 0
because both the call and R(t) are nonnegative.
■ Thus the portfolio (− pA + cA − S + K) cannot have a negative value today. Combining (7a) and (7b) and
rearranging terms, we get our second inequality:
cA ≥ pA + S − K (8)
■ For the PCP inequality for American options, combining (5) and (8), we get the relation between the stock,
the bond, and the American options
pA + S − K ≤ cA ≤ pA + S − BK (9)
16.5 Options Price Restrictions
This section explores various options price restrictions for several reasons.
■ They hone intuition and enhance understanding.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ They generate arbitrage opportunities when the bounds are violated.
■ Just as you stretch before you sprint, a discussion of options price relations forms
the stepping-stone to understanding early exercise of options, which is the final
topic of this chapter.
To facilitate understanding, we establish the options price restrictions in several
different ways: (1) a diagrammatic approach with profit diagrams; (2) an arbitrage table
approach, where cash flows are recorded in a table and no arbitrage yields the result;
and (3) a superglue argument, which is our name for comparing the original security
with a “restricted” but otherwise identical security and then proving the result by
arguing that “less cannot be worth more.” Why “superglue?” Read extension 16.3.
We begin with the simple and then move to the complex. We use our standard
notation and assumptions, except one: we allow the underlying stock to pay dividends.
This is an important extension because most stocks pay dividends.
OPTIONS PRICE RESTRICTIONS 347
RESULT 16.2
American Options Are Worth More Than European
Options
2a. cA ≥ c ≥ 0 (16.5a)
2b. pA ≥ p ≥ 0 (16.5b)
This result holds because European and American options are identical, except
for their exercise features. European options can only be exercised at expiration,
whereas American options can be exercised anytime. As “more cannot be worth less,”
an American option can never be priced less than an otherwise identical European
option. This simple yet robust principle underlies the superglue argument, which is
used later to prove some results (see Extension 16.2). Of course, an option’s price
must be nonnegative because the holder can always discard the option without
exercising it.
The next two results restate the option price bounds established earlier in
Chapter 5.
RESULT 16.3
Call Price Boundaries
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
3a. An American call’s price is less than or equal to the stock price, that is,
cA ≤ S (16.6)
3b. When the stock price is zero, the call price is also zero.
Result 16.3 also holds for European calls because they are less valuable than
American calls. Notice that an American call’s price cannot fall below the boundary
condition stated earlier in Chapter 5. These are depicted in the first diagram in
Figure 16.2, which reproduces Figure 5.3. Result 16.3b follows by setting S = 0
in Result 16.3a and recognizing that the call’s value must always be nonnegative.
348 CHAPTER 16: OPTION RELATIONS
RESULT 16.4
Put Price Boundaries
An American put’s price is less than or equal to the strike price K, and a
European put’s price is less than or equal to the present value of the strike
price:
4a. pA ≤ K (16.7a)
4b. p ≤ BK (16.7b)
where B is today’s price of a zero-coupon bond that pays $1 at the option’s
maturity.
A put option entitles the holder to receive the strike price, but he has to surrender
the stock in return. Thus the put price cannot exceed the strike price, which forms
an upper bound. Because a European put’s maximum payoff is the strike and only on
the expiration date, today’s put price must be less than or equal to the present value
of the strike price.
RESULT 16.5
Call Price Bound before Expiration
Before expiration, a European call’s price is greater than or equal to the larger
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
of the stock price minus the present value of the strike or zero:
c ≥ max (S − BK, 0) (16.8)
We will prove this result with payoff diagrams, by an arbitrage table, and using the
superglue argument (see Extension 16.3). The result also holds for American options,
which are more valuable than otherwise identical European options.
Figure 16.3 shows two payoff diagrams on the expiration date. The first diagram
is for a stock purchase plan that consists of buying the stock and shorting K zero-
coupon bonds. The second diagram, which is placed beneath the first, is for a long
call. Visual inspection shows that the stock purchase plan pays off less than the long
call. A consideration of today’s value of these two payoffs establishes that c ≥S − BK.
Because the call’s value must also be nonnegative, we get expression (16.8).
Our next example uses a variant of our data–model–arbitrage approach to establish
the relation.
OPTIONS PRICE RESTRICTIONS 349
FIGURE 16.2: Price Bounds for
American Calls
and Puts
Prices Upper Bound
Lower Bound
Call prices
in here
45°
0
K S, stock price
Prices
Upper Bound
K
Put prices
in here
Lower Bound
0
K S, stock price
EXAMPLE 16.3: Establishing Result 16.5 with an Arbitrage Table
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The Data, Arbitrage, and Diagram
■ Consider the data from Example 16.2 with a slight modification. OPSY’s stock price S is $22.50, the
six-month $20 strike European call price c is $2 and the put price p is $0.50, and the zero-coupon
bond price B is $0.975310.
■ Does Result 16.5 hold? We have S − BK = 22.50 − 0.9753 × 20 = $2.9938. This is greater than the
call price of $2, and the inequality in (16.8) is violated. Transferring the right side terms to the left
side, we get the expression c − S + BK = 2 − 2.9938 = −$0.9938.
■ Buy low, sell high. Create an arbitrage portfolio by buying the underpriced call and selling the over-
priced portfolio (S − BK) by shorting the stock and buying the bonds. Record these cash flows in
Table 16.2. You immediately get $0.9938. At expiration, the portfolio has a 0 payoff when the call is
in-the-money and a nonnegative payoff of [20 − S(T)] otherwise.
■ You can also see this arbitrage possibility by drawing the profit diagram for the portfolio in Table
16.2. Take $0.9938 and invest it in bonds to get $0.0190 after six months. The portfolio payoff after
six months is [20 − S(T)] when the call is out-of-the-money, and 0 otherwise. Adding $1.02 (after
rounding) to this shifts the whole curve up (see the diagram in Figure 16.4). This confirms the arbitrage
opportunity.
350 CHAPTER 16: OPTION RELATIONS
To establish the result, one simply replaces the numbers with algebraic symbols.
Writing S = $22.50, K = $20, B = $0.975310, and c = $2, the last row in the second
column of Table 16.2 shows that − c + S − BK = 0.9938. You can generalize the last
row, as c < S − BK leads to a cash inflow today and a nonnegative payoff [K − S(T)]
when S(T) ≤ K, and 0 otherwise. This is an unstable situation. To prevent arbitrage,
we must have
c ≥ S − BK (16.9)
This is Result 16.5 when combined with the fact that c ≥0.
Figure 16.4 also demonstrates an easy way to identify an arbitrage portfolio by
drawing a profit diagram and checking to see whether you can make arbitrage profits.
FIGURE 16.3: The Call Price Must Exceed
the Stock Price Minus the
Present Value of the Strike
Price (Result 16.5)
Payoff
Long stock
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
0
K = 20 S(T)
–20
Payoff
Long call
0
K S(T)
OPTIONS PRICE RESTRICTIONS 351
TABLE 16.2: Arbitrage Table Showing How a Violation of c ≥ S – BK
(Result 16.5) Creates an Arbitrage Opportunity
Expiration Date (Time T) Cash Flow
Portfolio Today (Time 0) Cash Flow
S(T) ≤ 20 20 < S(T)
Buy call −2 0 S(T) − 20
Short sell stock 22.5 −S(T) −S(T)
Buy bonds to lend the
present value of strike −0.975310 × 20 20 20
(K = $20, B = $0.975310)
Cash flows 0.9938 20 − S(T) 0
■ If the profit graph for a portfolio lies entirely above (or with some portions of it
lying along) the horizontal axis, then you have created an arbitrage portfolio.
■ If the profit graph lies entirely below (or with some portions of it lying along) the
horizontal axis, just reverse the trades to capture arbitrage profits.
■ A profit graph that crosses the axis leads to both trading profits and losses, which
can never represent an arbitrage opportunity.
FIGURE 16.4: Profit Diagram Showing an Arbitrage
Opportunity When the Call Price Is Less Than
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the Stock Price Minus the Present Value of
the Strike Price (Violation of Result 16.5)
Profit
21.02
1.02
0
K = 20 S(T)
352 CHAPTER 16: OPTION RELATIONS
Result 16.5 has its counterpart for put options.
RESULT 16.6
Put Price Bound before Expiration
6a. Before expiration, a European put’s value must be greater than or equal
to the larger of the present value of the strike price minus the stock price or
zero:
p ≥ max (BK − S, 0) (16.10a)
6b. Before expiration, an American put’s value must be greater than or equal
to its intrinsic value:
pA ≥ max (K − S, 0) (16.10b)
Result 16.6a follows easily using the superglue argument (see Extension 16.3).
Result 16.6b follows from the boundary condition because you can always exercise
an American put early and collect the strike price by surrendering the stock.
RESULT 16.7
Relation between Options with Different Strike Prices
7a. The lower the strike price, the more valuable the European call:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
c (K2 ) ≤ c (K1 ) for K1 < K2 (16.11)
7b. The higher the strike price, the more valuable the European put:
p (K1 ) ≤ p (K2 ) for K1 < K2 (16.12)
Both these results hold for American options as well.
Extension 16.3 proves several of these results using the superglue argument.
EXTENSION 16.3: The Superglue Argument
Remember superglue, the unusually strong adhesive that advertisements claim can hang a car and lift two thousand
pounds per square inch? Our use of superglue is different—we take the idea, but not the product! Consider
two identical securities, which we name Sec. We attach a “restriction” to one of these securities, something
OPTIONS PRICE RESTRICTIONS 353
unfavorable or adverse that removes one of its provisions, and assume that it gets fastened with superglue (see
Ext. 16.3 Fig. 1). As such, the restriction is permanent, and the security’s provisions are changed forever.
EXT. 16.3 FIG. 1: The Superglue
Argument for Options
Sec SecR
GLU
E
Restriction
Security Security + Restriction
SecR is identical to Sec except for a restriction that is attached by
superglue. Then, price of SecR < _ price of Sec.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The essence of the superglue argument is this: consider a security Sec and an otherwise identical security
with the restriction attached, which we call SecR (short for SecurityRESTRICTED ). A long position in SecR cannot
have a greater value than Sec because less cannot be worth more, that is, Price of SecR ≤ Price of Sec.
This is a no-arbitrage condition. If it happens otherwise, then buy the security, add the restriction (with
superglue at no cost), and sell it for more to make arbitrage profits. We can use this principle to derive all of our
options price bounds and relations. We illustrate just a few of these below.
Consider the proof of result 16.2 (directly below). Let Sec be an American option. Next, we attach a restriction
to this option with superglue: “it cannot be exercised early.” Then, by no arbitrage, we know that the price of the
restricted security—equivalent to a European call—must be less than the price of the unrestricted security—the
American call. (Figure 16.5 illustrates this approach).
Result 16.2 American Options Are Worth More Than European Options
Proof:
Sec: American option.
SecR : American option + Restriction “No early exercise allowed”
Because SecR has become an otherwise identical European option, the superglue argument gives:
European option price ≤ American option price.
354 CHAPTER 16: OPTION RELATIONS
Result 16.3b cA ≤ S
Proof:
Sec: Stock
SecR : Stock + Restriction “Pay an additional K dollars to maintain possession, otherwise throw away, and
the decision must be made before time T.” Of course, with this restriction attached, the restricted stock is an
American call.
Hence American call price ≤ Stock price.
Result 16.5 c ≥ S – BK
Proof:
Sec: European call.
SecR : European call + Restriction “Always exercise the call even if it is out-of-the money.” Today’s price of
SecR is S − BK.
The superglue argument gives the result. Result 16.2 tells us that this argument also holds for American calls.
Result 16.7a c(K1 ) ≥ c(K2 ) for K1 < K2
Sec: European call with strike price K1 .
SecR : European call with strike price K1 + Restriction “Must pay the extra amount (K2 − K1 ) in case of
exercise.” The restricted security is equivalent to a European call with strike price K2 .
The superglue argument gives the result.
A similar argument holds for American calls.
EXAMPLE 16.4: Result 16.7b by the Modified Data–Model–Arbitrage
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Approach
■ Consider the COP data from chapter 15 with minor modifications. Puts trade on OPSY that mature
in six months. The premiums are p(20) = $2 for strike price K1 = $20 and p(22.50) = $1.50 for
K2 = $22.50. These numbers violate inequality (16.12). We use the arbitrage opportunity in the data
to justify our model, which is relation (16.12).
When Both Puts Are European
■ As shown in Table 16.3, we create the arbitrage portfolio by selling the overpriced OPSY 20 put and
buying the underpriced OPSY 22.50 put. This gives a cash flow of + 2 − 1.50 = $0.50 today. The
puts’ payoffs on the expiration date depend on the stock price.
- If S(T) is less than or equal to $20, the short put has a payoff of −[20 − S(T)] and the long put has
a payoff of [22.50 − S(T)]. This gives a payoff of $2.50.
- For 20 < S(T) ≤ 22.50, the OPSY 20 put has a 0 payoff while the long OPSY 22.50 put has a
payoff of [22.50 − S(T)]. For this range for S(T), the payoff is nonnegative.
- When S(T) is greater than $22.50, both options are out-of-the-money and have 0 payoffs.
■ We showed that
p (20) = $2 > p (22.50) = $1.50
OPTIONS PRICE RESTRICTIONS 355
leads to an arbitrage opportunity. The no-arbitrage requirement dictates that p(20) ≤ p(22.50).
■ Replace numbers with symbols and you get the result
p (K1 ) ≤ p (K2 ) for K1 < K2
When Both Puts Are American
■ Suppose the mispricing holds and the prices reflect those for American puts instead. As with European
puts, create an arbitrage portfolio by selling the OPSY 20 American put and buying the OPSY 22.50
American put. This gives $0.50 today. In the future, if the short put is not exercised against you, hold
on to the long put. The payoff on the expiration date is given in Table 16.3. By contrast, if the OPSY
20 put is exercised against you, exercise the OPSY 22.50 put to eliminate the stock position, and
you are left with $2.50 on the exercise date. This trade is an arbitrage opportunity. Hence, the only
American put prices consistent with no arbitrage must satisfy result 16.7b.
TABLE 16.3: Arbitrage Table for Result 16.7b: The Higher the Strike Price, the
More Valuable Is the Put
Expiration Date (Time T) Cash Flow
Portfolio Today (Time 0) Cash Flow S (T) ≤ 20 20 <S (T) ≤ 22.50 22.50 <S(T)
Sell 20 put 2 −[20 − S(T)] 0 0
Buy 22.50 put −1.50 22.50 − S(T) 22.50 – S(T) 0
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Cash flows 0.50 2.50 22.50 – S(T) 0
RESULT 16.8
Option Prices and Time to Maturity
A shorter maturity American option cannot have a larger value
8a. cA (T1 ) ≤ cA (T2 ) for T1 < T2 (16.13)
8b. pA (T1 ) ≤ pA (T2 ) for T1 < T2 (16.14)
The results need not hold for European calls (because of dividends) or
European puts.
The next example establishes these results.
356 CHAPTER 16: OPTION RELATIONS
EXAMPLE 16.5: Establishing Result 16.8a
■ Suppose that the OPSY call with strike price $20 maturing in T1 = 6 months has a price of cA (T1 )
= cA (6 months) = $3.50 but the call maturing in T2 = 1 year is worth cA (T2 ) = cA (12 months) = $3.
Exploit this situation by selling the relatively overpriced six-month call for $3.50 and use the proceeds
to buy the relatively underpriced one-year call for $3. This gives $0.50.
■ If cA (6 months) is exercised against you before expiration, exercise cA (12 months) to satisfy the short
call provisions. And you still have your 50 cents.
■ If cA (6 months) is never exercised, you can make more money by selling cA (12 months), which has six
months left, or you can hold on to the call. Remember that all we need to show is that we can make
arbitrage profits—there is no need to demonstrate how to maximize profits.
■ Replace numbers with symbols to establish the result.
Longer-maturity European calls need not be more valuable than their shorter-
maturity counterparts. You can easily see this with a counterexample. Consider
six-month and twelve-month European calls on the OPSY stock with a common
strike price of $20. Suppose we know today that the stock will pay a liquidating
dividend and have zero value after one year but no dividends will be paid earlier.
Then, c(12 months) = 0. But c(6 months) > 0 because it’s possible for OPSY to
close higher than $20 after six months. This is true because the stock price today
is $22.50. If we knew that its price would never exceed $20 in six months’ time,
then its price must be less than $20 today to avoid immediate arbitrage. Notice that
the shareholders will not scream foul because they are not suffering any loss of value
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
owing to the liquidating dividend. It’s only the call holders who are losing out because
options are not protected for cash dividends in the United States.
Result 16.8b need not hold for European puts. We illustrate this with an example.
A money manager who had previously taught finance classes at the University
of Nebraska used the following one-hundred-year put option example in a 2008
chairman’s letter to the shareholders of his company. His name? Warren Buffett.
Consider three out-of-the-money European put prices when OPSY’s stock price
is $22.50 and the strike price is $20. One of these options will mature in a second.
The put price is obviously 0. Alternatively, if the time to maturity is six months, this
option has a premium of $0.50, as given in our COP data (or some other number
that’s greater than zero because a lot can happen to the stock in six months’ time).
Now consider a really long maturity European put. Using Result 16.4b, if the interest
rate r is 5 percent, discounting gives the present value of $20 as $0.1348 when the
time to maturity is one hundred years. The longer-maturity European put has a
smaller value than the six month option. By considering out-of-the-money very
short maturity puts and very long maturity puts (both of which are close to zero),
you can see that the European put’s premium does not necessarily increase with time
to maturity.
EARLY EXERCISE OF AMERICAN OPTIONS 357
16.6 Early Exercise of American Options
Although some traders do not bother about early exercise, professional traders
routinely exercise options early when it’s optimal to do so. Early exercise is important
because if you do not exercise when it’s optimal, you are giving away profits to the
writer. Empirical studies have found instances of both rational and irrational early
exercise decisions.
Following two simple rules yields the optimal early exercise strategy.2
RESULT 16.9
Rules of Thumb for Early Exercise
9a. Exercise an American call the first time its price is equal to the stock
price minus the strike price:
cA = S − K (16.15)
9b. Exercise an American put the first time its price is equal to the strike price
minus the stock price:
pA = K − S (16.16)
The first time early exercise is optimal, the traded option’s price will equal its
exercise value. These results are independent of the stock price evolution or dividend
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
payout policy. They depend, however, on the assumption that the markets are
arbitrage-free and well functioning, where cA and pA reflect the “true” option prices.
Note that neither option’s price can be strictly less than (16.15) or (16.16) without
generating an arbitrage opportunity (hence the equality)
It is very easy to use these rules of thumb to decide when to exercise early. Just
simultaneously monitor the stock price and the option’s price across time. If this
expression becomes satisfied, then exercise; otherwise, do not exercise. Following
these simple rules of thumb will generate maximum value for your option positions
and frustrate professional traders hoping that the options are irrationally exercised.
No Dividends
Let us start with an American option on OPSY, where the stock doesn’t pay any
dividends. Our first result is as follows.
2
The formal proof of the optimality of these rules involves solving the optimal stopping problem; see
Jarrow and Protter (2008).
358 CHAPTER 16: OPTION RELATIONS
RESULT 16.10
Early-Exercise Decisions for American Options on
No-Dividend Stocks
Assume the stock pays no dividends over the option’s life.
10a. An American call becomes identical to a European call. It should never
be exercised early.
10b. An American put should be exercised if the stock price is small enough
relative to the strike price and the time remaining to maturity.
To prove Result 16.10a, let’s consider what happens to the value of a call if it is
exercised early. If you exercise the American call before it expires, you get on the
exercise date t: [S(t) – K] = [S(t) − 20] dollars, where S(t) is OPSY’s stock price and
K = $20 is the strike. By contrast, if you sell the call, Result 16.5 ensures that you get
at least [S(t) − BK] = [S(t) − 20B]. This is a larger amount because the zero-coupon
bond price B is less than 1. So do not exercise the American call—the value of your
position in the market is larger. If you do not want to hold the option, sell it, but do
not exercise it.
Result 16.10a may be difficult to believe at first sight. To understand why it is true,
let us perform a thought experiment. Consider holding an American call option on
OPSY with a strike price of $20 when the stock is trading at $22.50. Let the option
mature in one year’s time. Now, assume that everyone in the market believes (includ-
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
ing yourself) that the stock is going to crash and be worth only $1 in a year’s time (see
Figure 16.5; the stock price crash is shown by the dashed line starting from today,
time t). Shouldn’t you exercise the call today? After all, if you hold it to expiration,
it will be worth zero dollars, and if you exercise the call today, you get $2.50.
The answer is, yes, you should exercise it today under this scenario. But doesn’t
this contradict Result 16.10a? The answer is no. The reason is that Result 16.10a
holds only under the assumption of no arbitrage. Our initial supposition creates an
immediate arbitrage opportunity, violating the assumption of no arbitrage underlying
the result. Indeed, if everyone believes the stock is worth $1 a year from now, its price
today must be less than $1 and not $22.50; otherwise, the arbitrage is to short the
stock and buy it back after the price declines.
Next, let us consider a related thought experiment that does not violate the no-
arbitrage assumption. Suppose instead that you have inside information (recall our
earlier warning: this is just a thought experiment!), and only you know the stock
is going to crash and be worth $1 in a year’s time. Here, because only you have
this information, the stock price of $22.50 is consistent with no arbitrage. Shouldn’t
you exercise this call today? The answer is no! Although you could exercise it now
to get an immediate $2.50, you can make more money by selling it in the market.
EARLY EXERCISE OF AMERICAN OPTIONS 359
FIGURE 16.5: Illustration of Result 16.10a
Prices
Stock price
S(t) = 22.50
K = 20
1
0
t T Time
Today
If everyone believes in the dashed line, the stock price will crash now
and fall below $1.
If only you believe in the dashed line, do not exercise (get $2.50),
but sell the call for more.
The market price of the American call exceeds $2.50 by Results 16.2 and 16.5:
cA ≥ c ≥ S − BK = 22.50 − 20B > 2.50
because B < 1.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Given these thought experiments, we now better understand Result 16.10a. It says
that you can make more profits by selling the OPSY call rather than by exercising it.
Thus, if no dividends are paid to the stockholder over the option’s life, then the call
option is always worth more alive than dead. Do not exercise such a call: sell it. Sell
it! SELL IT!!
Puts, though, have a different story. Unlike calls that have an unbounded profit
potential, put payoffs are capped by the strike price, the maximum payoff possible.
This leads to Result 16.10b. To understand why Result 16.10b is true, consider
the following hypothetical example. Consider a six-month original maturity put on
OPSY that began two months back. If OPSY’s stock price is S(t) = $0.0001 today,
an immediate exercise gives K − S(t) = (20 − 0.0001) = $19.9999. If the interest
rate r is 6 percent per year, investing the proceeds in a money market account over
the next day yields
[K − S (t)] erT
= [19.9999] e0.06 × (1/365)
= $20.0032
360 CHAPTER 16: OPTION RELATIONS
Waiting one day and then deciding to exercise yields at most $20, and if one waits,
the stock price may rise and even less of a payoff may occur. So early exercise is
optimal today. Of course, this is a pathological example selected to make our point,
but it does prove that early exercise is sometimes optimal if the stock price is low
enough and interest rates are high enough.
This pathological example also illustrates another observation. It is easy to believe
that as the stock price rises, there is some stock price S* where you are just indifferent
between exercising today or not. The S* just balances the interest earned on the
proceeds of early exercise with the chance of the stock price declining more (see
Figure 16.6). Such an S* does in fact exist. Unfortunately, to prove this observation
and to determine the S* precisely, we need a model for the stock price evolution. This
analysis will be pursued in chapter 18. For now, we can only prove Result 16.10b.
Dividends and Early Exercise
Dividends change the optimal exercise decision for American options. Recall that
when dividends are paid, the stock price falls. We characterized this dividend-induced
price drop as Result 3.1:3
SCDIV = SXDIV + div (3.1)
where SCDIV is stock price cum-dividend, SXDIV is stock price ex-dividend, and div
is the amount of dividend. A natural place to investigate early exercise is around the
time a stock goes ex-dividend. This leads to our next result.
FIGURE 16.6: Illustration of Result 16.10b
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Prices
K = 20
Stock price
S*
0
t T Time
Early exercise
At S*, one is indifferent between exercising or not.
Lower stock prices give higher proceeds from early exercise.
3
Exceptions to this statement can happen. We need some technical assumptions about the process
underlying stock price changes. The statement in (3.1) holds, for example, if the stock price process
follows a lognormal distribution, which underlies the Black–Scholes–Merton model. It does not hold,
for example, if the stock price process allows discrete jumps and the jump happens at the same time as
the stock goes ex-dividend; see Heath and Jarrow (1988) for a formal discussion.
EARLY EXERCISE OF AMERICAN OPTIONS 361
RESULT 16.11
Optimal Times for Exercising American Calls
The only times when it may be optimal to exercise early is just before a stock
goes ex-dividend.
The next example establishes this result with an application of Result 16.5.
EXAMPLE 16.6: Establishing Result 16.11
■ Consider a six-month OPSY 20 call. Let the next dividend div1 be paid at time t1 , which is two
months from today (time 0). Should you exercise this call at a time t, anytime over the next two
months?
■ Consider the two possible strategies: early exercise at time t or wait and exercise at time t1 . If you
exercise at time t, the value of your position is S(t) − 20, but if you exercise an instant before OPSY
goes ex-dividend at time t1 , even if the option is out-of-the-money, then you get SCDIV − 20. The
present value at time t of waiting until time t1 to exercise is
S (t) − 20B
where B is the time t price of a zero-coupon bond that pays $1 at time t1 .
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Because B is less than 1, the second strategy gives you more value than the first. Thus exercising the
American call just before the stock goes ex-dividend is better than exercising it anytime between now
and the first dividend date.
■ This result follows because exercising earlier than these ex-dividend dates leads to an early surrender
of the strike price—an action that loses interest but does not generate any additional benefit. Hence,
if the call is exercised early to obtain the dividend on the stock, it will be exercised just an instant
before the stock goes ex-dividend. This is depicted in Figure 16.7. See Jarrow and Turnbull (2000)
and other finance textbooks for a discussion of how big the dividend needs to be to trigger early
exercise.
And what about developing early exercise results for puts? Puts can get exercised
early even without dividends. As dividends would cause the stock price to drop
further, their presence is likely to delay early exercise. To get more precise insight,
we need to model the stock price evolution. Again, we address this in later
chapters.
362 CHAPTER 16: OPTION RELATIONS
FIGURE 16.7: Timeline for Exercise of American
Options
If exercise call, exercise at these points
Option starts Ex-div (div1) Ex-div (div2) Expiration
Exercise put anywhere
If exercising an American call early, exercise just before the stock goes ex-dividend.
An American put may be exercised anytime (but dividends tend to delay exercise).
16.7 Summary
1. Put–call parity for European options (see Result 16.1) states that the call price plus
the present value of strike equals the stock price plus the put price, or c + BK =
S + p, where B is today’s zero-coupon bond price paying a dollar on the options’
common maturity date and K is the common strike price.
2. PCP has a number of applications and uses:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- PCP can be used to create synthetically an option, a stock, or a bond.
- PCP can be easily modified to adjust for known dividends.
3. Option prices satisfy restrictions that can be derived using the no-arbitrage
principle. These theorems hone intuition and enhance understanding, lead to
arbitrage profits in case of violations, and form a stepping-stone to understanding
the early exercise of options.
4. Options price restrictions can be established in several different ways: (1) a
diagrammatic approach with profit diagrams, (2) an arbitrage table approach,
where cash flows are recorded in a table and no arbitrage yields the result, and
(3) a superglue approach, which is our name for comparing the original security
with a “restricted” but otherwise identical security and then proving the result by
arguing that “less cannot be worth more.”
5. Some options price properties for stocks that may pay dividends are as follows:
- Result 16.2—American options are worth more than European options: cA ≥ c
and pA ≥ p.
- Result 16.3—Call price boundaries
3a An American call’s price is less than or equal to the stock price: cA ≤ S.
3b When the stock price is zero, the call price is also zero.
SUMMARY 363
- Result 16.4—Put price boundaries
4a An American put’s price is less than or equal to the strike price K: pA ≤ K.
4b A European put’s price is less than or equal to the present value of the strike
price: p ≤ BK.
- Result 16.5—Before expiration, a European call price must exceed the stock
price minus the present value of the strike price: c ≥ S − BK.
- Result 16.6a—Before expiration, a European put must be greater than or equal
to the present value of the strike price minus the stock price: p ≥ BK − S.
- Result 16.6b—Before expiration, the value of an American put must be greater
than or equal to its intrinsic value: pA ≥ K − S.
- Result 16.7a—The lower the strike price, the more valuable is the European
call: c(K1 ) ≥ c(K2 ) for K1 < K2 .
- Result 16.7b—The higher the strike price, the more valuable is the European
put: p(K1 ) ≤ p(K2 ) for K1 < K2 .
These results hold for American options as well.
- Result 16.8—Option prices and time to maturity: the longer the time until
expiration, the greater the price of an American call or a put.
8a cA (T1 ) ≤ cA (T2 ) for T1 < T2 .
8b pA (T1 ) ≤ pA (T2 ) for T1 < T2 .
However, these results need not hold for European calls or European puts.
6. Early-exercise considerations begin with the following:
Result 16.9a—Exercise an American call the first time its price is equal to the
stock price minus the strike price: cA = S − K.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Result 16.9b—Exercise an American put the first time its price is equal to the
strike price minus the stock price: pA = K − S.
7. Dividends play a critical role in early-exercise decisions:
- Assume the stock pays no dividends over the option’s life:
Result 16.10a—The American call becomes identical to a European call. It
should never be exercised early.
Result 16.10b—The American put should be exercised if the stock price is small
enough relative to the strike price and the time to maturity.
- Absent market frictions, dividends cause the stock price to drop by the amount
of the dividend (Result 3.1). This is used to establish the following:
Result 16.11—The only times when it may be optimal to exercise an American
call early is just before a stock goes ex-dividend.
Puts can get exercised early even without dividends, but dividends are likely to
delay early exercise.
364 CHAPTER 16: OPTION RELATIONS
16.8 Cases
Boston Properties (A and B) (Harvard Business School Cases 211018 and 211041-
PDF-ENG). The case introduces options pricing, payoff diagrams, and the law of
one price and explains arbitrage as well as no-arbitrage bounds.
Smith, Barney, Harris Upham and Co. Inc. (Darden School of Business Case
UV0074-PDF-ENG, Harvard Business Publishing). The case approaches put–call
parity from the trader’s perspective and examines the practical aspects of doing
arbitrage in the options markets.
Sleepless in L. A. (Richard Ivey School of Business Foundation Case 905N11-
PDF-ENG, Harvard Business Publishing). The case discusses the Black–Scholes–
Merton model for options pricing, the concept of implied volatility, and put–call
parity. It also shows how options pricing can be used to value corporate liabilities
of a financially distressed company.
16.9 Questions and Problems
The next two questions are based on the following data for European options:
Call price = $5, risk-free continuously compounded interest rate r = 5 percent per
year, stock price S = $55, strike price K = $55, time to maturity T = 1 month.
16.1. If the put price p = $9, show how to capture arbitrage profits in this market.
16.2. Suppose short selling of stocks is not allowed in this market. Can you still
make arbitrage profits? Explain your answer.
16.3. Do the following data satisfy put–call parity for European options? If they
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
don’t, show how you can create a portfolio to generate arbitrage profits. Call
price = $6, put price = $3, stock price S = $102, strike price K = $100, time
to maturity T = 3 months, and risk-free continuously compounded interest
rate r = 5 percent per year.
16.4. How can you adjust put–call parity for known dividends on a single known
date?
16.5. Prove put–call parity for European options in the case of a single known
dividend:
c + PV (Div) + Ke−rT = p + S
where S is stock price, K is strike price, T is maturity date for the option,
r is risk-free interest rate, c is European call price, p is European put price, and
PV(Div) is the present value of dividends.
16.6. Using put–call parity, given c = $2, PV(Div) = $1, p = $1, S = K = $100,
r = 0.05 per year, and T = 0.25 years, can you make arbitrage profits? Explain.
16.7. Explain the relation between a put option (with strike price K and maturity
date T years from today) and a T-period insurance policy that insures the
stock for K dollars.
QUESTIONS AND PROBLEMS 365
16.8. Different countries, different customs, different market practices! Suppose you
go to a country where traders with inside information can easily trade. There,
you have inside information that the stock price is going to increase for sure
(no chance that it will decline), and it is legal to trade on this information.
What is the best strategy to use?
16.9. Does put–call parity always hold in financial markets? If not, give a few reasons
why it may not hold.
16.10. On the day options on Palm began trading, the share prices grossly violated
the put–call parity. Describe and explain why this happened.
16.11. Can you link Russell Sage’s actions with a distinguished economist’s view on
what drives financial innovation?
16.12. Is it true that the lower the exercise price, the more valuable the call? Explain
your answer.
16.13. Is a European put on the same stock with the same maturity worth more or
less if the strike price increases? Explain your answer.
16.14. Is it true that the more the time until expiration, the less valuable an American
put? Explain your answer.
16.15. Can you make arbitrage profits from the following European call prices?
If so, give two such examples of arbitrage, neatly showing the portfolio
construction as well as the various cash flows. The stock price is $40.
Strike Price Expiration Month
April July September
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
35 1 6 3
40 2 5 6
16.16. The following prices are given for American put options on a stock whose
current price is $100:
Strike Price Expiration Month
March June
95 1 11
100 7 5
Construct three portfolios for making arbitrage profits, showing the cash flows
from each portfolio.
366 CHAPTER 16: OPTION RELATIONS
16.17. The following prices are given for American call options on a stock whose
current price is $100:
Strike Price Expiration Month
March June
95 11 10
100 8 8
105 2 5
Construct three portfolios for making arbitrage profits, showing the cash flows
coming from each portfolio.
16.18. If an American call is written on a stock that never pays a dividend, would
you ever exercise the call option early? Explain your answer.
16.19. If an American put is written on a stock that never pays a dividend, would
you ever exercise the put option early? Explain your answer.
16.20. a. Is there a simple rule of thumb that you can use to know when to exercise
an American call early? If yes, explain the rule.
b. Is there a simple rule of thumb that you can use to know when to exercise
an American put early? If yes, explain the rule.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
17
Single-Period Binomial
Model
17.1 Introduction Arbitrage-Free Trees
17.2 Applications and Uses of the Stock Prices and Martingales
Binomial Model The Pricing Model
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The Hedge Ratio
17.3 A Brief History of Options
Pricing Models Risk-Neutral Valuation
Option Pricing Pre-1973 Actual versus Pseudo-probabilities
Options Pricing, 1973 and After 17.7 Summary
17.4 An Example 17.8 Appendix
17.5 The Assumptions Proving the No-Arbitrage
Argument
17.6 The Single-Period Model The Probabilities and Risk Premium
The No-Arbitrage Principle
17.9 Cases
Building Binomial Trees
17.10 Questions and Problems
368 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
17.1 Introduction
“Once in Hawaii I was taken to see a Buddhist temple,” wrote physicist Richard
Feynman in The Meaning of It All: Thoughts of a Citizen Scientist. “In the temple a man
said, . . . ‘To every man is given the key to the gates of heaven. The same key opens
the gates of hell.’” Derivatives may be viewed as such a key. When the financial crisis
of 2007 hit many countries, including the United States, numerous commentators,
and even the lay public, blamed it on derivatives. Indeed, credit rating agencies failed
to correctly rate complicated derivatives, and the losses on these derivatives brought
down many financial institutions. Conversely, the growth of the derivatives market
has been praised as improving economic welfare by shifting risks from those who fear
it to those who profit from it. And derivatives help solve financial problems. They
were even used by governments during the crisis to help resolve it.
For example, the US Treasury started the Troubled Asset Relief Program
during the financial crisis to commit up to “$700 billion to rescue the financial
system,” notes the July (2009) Oversight Report of the Congressional Oversight
Panel, “TARP Repayments, Including the Repurchase of Stock Warrants.” TARP
purchased stock in the stressed banks with “troubled assets,” but in addition, they
received ten-year warrants. Warrants are call options—derivatives! After the banks
recovered, the US government sold these warrants back to the banks or to others
at a “fair price,” determined with the help of the Black–Scholes–Merton (BSM)
and binomial models. In doing so, they profited substantially. Note the use of the
binomial model!
This chapter develops the single-period binomial model with an eye toward
understanding the major tenets of options pricing. We provide an illustrative example,
give the model’s intuition, and state the necessary assumptions. To value the option,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
we construct a synthetic option with identical payoffs using the stock and a money
market account. This takes us to martingale pricing and risk-neutral valuation, which
lies at the heart of modern theory and which enables us to understand the more
complex option pricing models (OPMs) that follow. Chapter 18 extends this chapter
to a multiperiod setting and makes the binomial model practical and useful. The
subsequent two chapters adopt a similar pattern: chapter 19 introduces the BSM
model, and chapter 20 discusses the model’s practical use.
17.2 Applications and Uses of the
Binomial Model
Options pricing bewildered and baffled academics until Fischer Black, Myron
Scholes, and Robert Merton developed an analytic pricing model for European
options in 1973. The mathematics underlying the model’s derivation is difficult. To
more simply derive the model, as William Sharpe noted, one can use a binomial
model and then take limits. Although the binomial model had been used as a teaching
tool at the Massachusetts Institute of Technology (MIT) and other places before its
A BRIEF HISTORY OF OPTIONS PRICING MODELS 369
publication, it was first printed in Sharpe’s classic textbook Investments. Subsequently,
Cox, Ross, Rubinstein, Rendleman, Bartter, Jarrow, and Rudd developed popular
versions of the binomial model.
The model gets its name binomial (bi means “two”) from the assumed stock price
evolution. From any point in time onward, over the next time step, the stock price
can take only one of two possible values. This raises two immediate questions: (1)
why do we need to make an assumption about stock price movements, and (2) is it
reasonable to assume that stock prices can only take two values?
The last chapter explicitly answered the first question. There we pushed as far as
we could without imposing a model for the stock price evolution. Although we got
many results, we could not obtain a pricing model. One needs to assume a model
for the stock price evolution like the binomial (which we do in this and the next
chapter) or a lognormal (chapter 19) to develop an exact OPM.
As for the second question, do stocks really only take one of two values at each
time step? No. But we need to start somewhere. We start with the simplest model to
facilitate understanding and then add the necessary layers of complexity. The simple
binomial OPM possesses some desirable features:
■ It approximates the BSM model. The binomial OPM can be repeated at each time
step in a multiperiod tree. When the parameters are properly chosen (see Extension
18.1) and the model is run over many small time intervals, the multiperiod
binomial OPM prices closely approximate BSM model prices.
■ It illustrates martingale pricing, the central idea of derivatives pricing. The binomial
approach utilizes “martingale pricing,” a powerful technique for valuation. Once
you understand this approach, you will find it easier to understand more complex
OPMs.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ It provides a useful numerical approximation technique. The binomial model is a versatile
tool that gives numerical values to approximate solutions to exact pricing models
(like the BSM model) or provides answers when no explicit analytical solutions
exist (as in the case of an American put). Later chapters show how to use the
binomial model to price interest rate derivatives. Even proprietary trading desks
of Wall Street firms sometimes use sophisticated versions of the binomial pricing
model.
17.3 A Brief History of Options
Pricing Models
The history of OPMs has two phases: before and after 1973.
Option Pricing Pre-1973
It all started more than a century ago, when a young French mathematician, Louis
Bachelier (1900), wrote a dissertation titled “The Theory of Speculation” that
370 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
developed a prototype OPM. Despite some shortcomings, it was the forerunner of
the BSM model. Bachelier was the first to model stock prices as a random walk. But
fate was unkind, and he did not get the recognition he deserved during his lifetime.
Within a decade after Bachelier’s death, his work resurfaced. It attracted the attention
of Professor Paul Samuelson, the first American to win the Nobel Prize in Economics.
During the 1960s, Samuelson and others worked on improving Bachelier’s model (see
Table 17.1 for a summary of landmark achievements in this field).
TABLE 17.1: Some Milestones in the History of Option Pricing
Models
Year Development
1900 Bachelier developed the first option pricing model (OPM).
1960s Samuelson and other scholars improved Bachelier’s model.
1973 The Black–Scholes and Merton models were published.
1970s, 1995 Merton (1974, 1977) and Jarrow and Turnbull (1995) models for pricing
derivatives with credit risk were introduced.
1978–82 The binomial model was presented in Sharpe (1978), Cox et al. (1979),
Rendleman and Bartter (1979), and Jarrow and Rudd (1982).
1979 and 1981 Harrison and Kreps (1979) and Harrison and Pliska (1981) developed the
martingale pricing methodology.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1987 and 1992 The Heath–Jarrow–Morton model (introduced in 1987, published in 1992), an
interest rate OPM, was introduced.
1997 Merton and Scholes, the surviving co-originators of the BSM model, won the
Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel.
Options Pricing, 1973 and After
During the late 1960s, two young PhDs, Fischer Black and Myron Scholes, teamed
up to work on options pricing. Around the same time, Robert Merton, a PhD
student of Paul Samuelson, started working on the same problem. Combining their
insights, 1973 saw the publication of the Black–Scholes–Merton model (Chapter 19
provides a more detailed history), which forms the foundation for today’s OPMs.
Their work stands as a towering achievement not only in finance but also in the field
of economics.
After the watershed year of 1973, a race began to build better pricing models.
Numerous OPMs resulted. These models priced different kinds of derivatives and
relaxed the BSM model’s assumptions. The race continues to this day with unabated
fury.
A BRIEF HISTORY OF OPTIONS PRICING MODELS 371
Why develop better OPMs? There are many reasons. The opening of options
exchanges and the expansion of over-the-counter options trading created a
need for better hedging. Brokers and dealers need OPMs to make effective
marketsmarkets.
The OPM extensions can be classified under five headings:
1. Improving the BSM model and using it for different underlyings. Soon after the
publication of the BSM came a slew of models that improved the original.
They included Black’s model to price commodity contracts. The stock price
evolution was generalized, and market imperfections like short selling restrictions
and transaction costs were included as well.
2. Including a term structure of random interest rates. A major drawback of the BSM model
is its constant interest rate assumption. Early attempts at pricing interest rate options
required estimating risk premia (e.g., Vasicek 1977; Brennan and Schwartz 1979)
that made the models difficult to use. The removal of the need to estimate interest
rate risk premia came in 1992 when the Heath–Jarrow–Morton (HJM) model
was published. All arbitrage-free interest rate OPMs are special cases of the HJM
model.1
3. A rigorous mathematical foundation. Harrison, Kreps, Pliska, and others generalized
the mathematics behind the BSM model. They developed martingale pricing,
which soon became the most widely used technique for pricing derivatives.
4. Pricing derivatives with credit risk. Merton (1974, 1977) pioneered the pricing of
derivatives with credit risk with the structural approach. Jarrow and Turnbull (1995)
developed an alternative method that relaxes some restrictive assumptions inherent
in Merton’s formulation. The Jarrow–Turnbull approach is called the reduced-form
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
credit risk model. See chapter 26 for additional discussion.
5. Numerical methods. With more complex pricing models came the need to compute
their values quickly and efficiently on computers. The binomial model was one
of the first successful numerical methods (see Sharpe 1978; Cox et al. 1979;
Rendleman and Bartter 1979; Jarrow and Rudd 1982). Others include numerical
methods for solving partial differential equations and Monte Carlo simulation
techniques.
To understand options pricing theory, we begin with an example that illustrates
the key ideas.
1
Baxter and Rennie (1996) state, “In the interest-rate setting, Heath–Jarrow–Morton is as seminal as
Black–Scholes. By focusing on forward rates and especially by giving a careful stochastic treatment,
they produced the most general (finite) Brownian interest-rate model possible. Other models may claim
differently, but they are just HJM with different notations.”
372 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
17.4 An Example
This section illustrates the binomial options pricing approach through an example
(see Example 17.1).
EXAMPLE 17.1: European Call Pricing
■ Your Beloved Machine Inc.’s (YBM) stock price S is $100 today (time 0). After one year (time T),
the stock price S(1) can either go up to $120.00 or down to $90.25. Let the actual probability of the
stock going up q = 3/4; then that of YBM going down is (1 – q) = 1/4.
■ Consider a European call option with strike price K = $110 and maturity date time 1. The call’s
payoff at time 1 is max[S(1) – 100, 0], that is, it is 10 = max(120 – 110, 0) if the stock goes up and 0
= max[90.25 – 110, 0] if the stock goes down.
■ Let a money market account (mma) cost $1 per unit at time 0. It earns a continuously compounded
interest rate r = 0.05 per year. A dollar invested in the mma grows to $1.0513 (see Figure 17.1).
Synthetic call construction
■ Form a portfolio V with m shares of the stock and b units of the mma. Its value today is
V (0) = m100 + b (17.1)
■ We want to choose (m, b) such that the portfolio’s payoffs at time 1, V (1), equal those of the traded
call on the expiration date. This holds if the following two equations are satisfied:
m120.00 + b1.0513 = 10 (17.2)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
m90.25 + b1.0513 = 0 (17.3)
■ Solving these equations gives m = 0.3361 and b = –28.85, where a positive sign indicates a long and
a negative sign a short. We call this portfolio the synthetic call.
■ Plugging the solutions into expression (17.1) gives V (0) ≡ m100 + b = $4.76. This is the cost of
constructing the synthetic call at time 0. To avoid arbitrage, this must be today’s price for the traded call, c.
■ It is interesting to note that the option’s value was determined without using the actual up probability
q = 3/4. This is an important observation. We will explain the reasons for this shortly.
■ Note that expression (17.1) also has a hedging interpretation. If you buy V (0) (a call option) and want
to hedge the stock price risk, then sell m shares of the stock. This creates the hedged call portfolio
V (0) – m100 = b, which is riskless because it’s equivalent to a position in the mma. We will also
explain this hedging interpretation more completely later in this chapter.
Arbitrage Profits
■ What happens when the traded call’s price differs from the price of the synthetic call? Seize the
opportunity and make arbitrage profits by buying low, selling high, or as the British would say, buying
cheap, selling dear.
AN EXAMPLE 373
■ Suppose an errant trader quotes $7 for the traded call. As the traded call is more expensive, sell it and
buy the synthetic call. Let’s do the accounting to see if it all works out.
- Sell the overpriced traded call for $7 and purchase the underpriced synthetic call by buying 0.3361
shares for $33.61 and borrow $28.85 by shorting the mma at a net cost of $4.76.
- Today’s cash flow is + 7 – 4.76 = $2.24. This is the immediate profit.
- You have no liabilities in the future. If the stock rises at maturity, you owe $10 as the writer of the
traded call. This is offset by the $10 that you get as the buyer of the synthetic call. Both calls expire
worthless if the stock falls instead.
■ If a trader quotes $3 for the call, simply reverse your strategy: buy the traded call and simultaneously
sell the more expensive synthetic call by short selling 0.3361 shares for $33.61 and lending $28.85 by
buying the mma. You immediately make $1.76, and there is no future net exposure as the assets and
liabilities perfectly match.
FIGURE 17.1: Single Period Binomial Example
and Model
Numerical Example
Stock Call
Today After 1 year Option starts Option matures
3/4 120.00 3/4 10
100 c
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1/4 90.25 1/4 0
Call: strike price K = 110, maturity time 1
One-Period Model
Stock Call
Now Maturity Time 0 Time 1
q US q cU
S c
(1 – q) DS (1 – q) cD
where U > D and 1 > q > 0
Call: strike price K, maturity time 1, payoff c(1) = max[S(1) – K, 0]
Money Market Account (mma)
1 1.0513
1 + R er t
374 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
Put Pricing
■ Consider a European put with a strike price of K = $110 and a maturity of one year. As with the
call, set up a portfolio to match the put’s time 1 value of $0 = max(110 – 120, 0) in the up state and
$19.75 = max(110 – 90.25, 0) in the down state. This yields two equations in two unknowns (m, b):
m120.00 + b1.0513 = 0
m90.25 + b1.0513 = 19.75
■ Solve the two equations to get the stock shares m = – 0.6639 and the number of mmas needed
b = 75.78.
■ As the stock and put prices move in opposite directions, it makes intuitive sense that the replicating
portfolio combines a short position in the stock and a long position in the mma.
■ The European put option’s price p is equal to the cost of construction at time 0, that is,
V (0) ≡ m100 + b = − 0.6639 × 100 + 75.78 = $9.39
Put–Call Parity
■ Alternatively, you can derive the put’s value using put–call parity for European options:
Put price = Call price + Present value of strike − Stock price
= 4.76 + 110 × (1/1.0513) − 100
= $9.39
Any other put price will open up arbitrage opportunities.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
17.5 The Assumptions
Assumptions are the key to understanding and using any model. Recall our discussion
of the cost-of-carry model and its assumptions from Chapter 11. We need those same
assumptions (A1 to A5) plus two more (A6 and A7). One is needed to describe the
behavior of interest rates, and the other is needed to characterize the binomial stock
price evolution. For emphasis as well as a better understanding of the model, we
repeat our discussion of assumptions A1–A5.
A1. No market frictions. We assume that trading involves no market frictions such
as transaction costs (brokerage commissions, bid/ask spreads), margin requirements,
short sale restrictions, and taxes (which may be levied on different securities at
different rates). Moreover, assets are perfectly divisible.
This helps us to create a benchmark model to which one can later add frictions.
In addition, it’s a reasonable approximation for many traders such as institutions.
THE ASSUMPTIONS 375
Moreover, political changes, regulatory changes, and the information technology
revolution have substantially reduced market frictions.
A2. No credit risk. We assume that there is no credit risk. Also known as default
risk or counterparty risk, this is the risk that the counterparty will fail to perform on
an obligation.
The absence of credit risk is a reasonable assumption for exchange-traded deriva-
tives. Over-the-counter derivatives, however, may possess substantial credit risks,
unless there are adequate collateral provisions.
A3. Competitive and well-functioning markets. In a competitive market,
traders’ purchases or sales have no impact on the market price; consequently, traders
act as price takers. In a well-functioning market, price bubbles are absent.
The competitive markets assumption is the workhorse of modern economics.
Spot, forward, and futures markets for many agricultural commodities and precious
metals behave like competitive markets. Pricing securities when counterparties have
market power and traders can manipulate prices is a difficult exercise. One must take
into account bargaining and strategic interactions, which are outside the scope of
this book.
A price bubble happens when an asset’s market price deviates from its intrinsic
or fundamental value. The fundamental or intrinsic value is the price paid if, after
purchase, you have to hold the asset forever. A difference between the market price
and fundamental value only occurs if one believes that selling can generate a higher
value than holding it forever. The difference is the price bubble. Recent research
shows that bubbles can invalidate the methodology we employ for pricing options.
Fortunately, however, one can show that there are no price bubbles possible in a
discrete time binomial model (see Jarrow et al. 2010).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
A4. No intermediate cash flows. Many assets reward their holders with income
or cash flows such as dividends for stocks and coupon income for bonds. For now,
we assume that the underlying asset has no cash flows over the option’s life. The next
chapter will show how to relax this assumption by introducing dividends.
A5. No arbitrage opportunities. This is self-explanatory.
These five assumptions are a repetition of the assumptions used in the cost-of-
carry model of chapter 11. Next we introduce two additional assumptions needed
for deriving the binomial OPM.
A6. No interest rate uncertainty. We assume that interest rates are constant
across time. This assumption only works for short-lived options whose underlyings
are not interest rates. It does not work well for long-dated options whose underlying
assets’ prices are correlated with interest rate changes, like long-term options on
foreign currencies. But long-dated options and interest rate derivatives are important
in today’s markets, and their pricing is a prime challenge for researchers and
practitioners. We relax this assumption in part IV of the book.
Recall that we did not require this assumption for our cash-and-carry models.
It wasn’t necessary because forwards are essentially valued in a single-period setting.
However, we will soon be introducing multiperiod and continuous time models in
376 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
which bond investments have reinvestment risks. To avoid this problem, we assume
a constant interest rate. Assumption A6 will be maintained for all OPMs in part III
of the book.
A7. Binomial process. We assume that the stock trades in discrete time and
that its price evolves according to a binomial process. For a current stock price, next
period’s stock price will either get multiplied by an up factor U > 1 and take a higher
value U × Stock price or it will be multiplied by a down factor D and take a lower
value D × Stock price. We let q > 0 denote the actual probability of going up. We
also assume U > D. Otherwise, we have mislabeled up and down.
Assumption A7 will be with us in this and the next chapter. It will eventually be
replaced by the assumption of “continuous trading” and the evolution of the stock
price according to a “lognormal process” in the context of the BSM model.
17.6 The Single-Period Model
To develop the model, we proceed through a sequence of steps. First, we explain the
no-arbitrage principle used in the example. Second, we show how to build a binomial
tree that is arbitrage-free. Next comes the martingale representation of stock prices,
which helps us to develop fancy entities called pseudo-probabilities. Then we obtain the
call’s value by building on Example 17.1. Finally, we introduce an alternative perspec-
tive on options pricing called risk-neutral valuation. The next chapter will extend this
to develop a multiperiod binomial OPM that approximates the BSM model.
The No-Arbitrage Principle
Let’s summarize the no-arbitrage principle used in the previous example to find the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
option’s value.
■ We considered a market in which both a stock and an option traded.
■ The stock took one of two possible values at the maturity date, and so did the
option because the option’s payoff is completely determined by the stock’s ending
value and the strike price.
■ We created a portfolio of the stock and mma to match the option’s payoffs in each
of these time 1 states. This replicating portfolio is called the synthetic option.
■ By the law of one price—the no-arbitrage principle—the cost of constructing the
synthetic option must equal the price of the traded option.
■ Why should they be the same? If the traded option and the synthetic option differ
in price (see Figure 17.2), then you have two distinct ways of getting the same cash
flows. Shrewd traders will make arbitrage profits by buying the cheaper option and
selling the dearer option until their prices converge.
■ This pricing methodology has an interesting by-product. A portfolio that holds
one of these options long and the other short should be entirely riskless. This kills
two birds with one stone: pricing and hedging problems are solved in a single stroke!
THE SINGLE-PERIOD MODEL 377
FIGURE 17.2: The No Arbitrage Principle
Portfolio A
(Market traded
option) Portfolio value in up state
Must be equal
without arbitrage
Portfolio value in down state
Portfolio B
(Synthetic option)
Building Binomial Trees
We need some notation to build a binomial tree. Today at time 0, the stock price is
S(0) or S. At time 1, it can move up to S(1)UP or slide down to S(1)DOWN . Let’s write
S(1)UP = SU and S(1)DOWN = SD, where the positive numbers U > D are called the
up and down factors, respectively (Figure 17.3 shows how a stock price path looks
in a single-period tree).
This multiplicative binomial model has the advantage of ruling out negative
stock prices. Moreover, in the multiperiod setting, it makes life easier by insuring
that the tree’s branches recombine. To reduce clutter, we drop the dollar sign from
the tree.
The other traded security is a mma. It earns the risk-free rate in each period.
A dollar gives (1 + R) = exp(rΔt) ≡ erΔt at the end of each period, where r is
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the continuously compounded yearly interest rate and Δt is the length of the time
FIGURE 17.3: Stock Price Evolution in a Single
Period Setting
Stock price
US Realized stock price path
S
DS Possible path
0
Today Time 1 Time
378 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
period. In Figure 17.1, a single arrow is used to show the mma’s value because it
earns a fixed return.
Consider a European call option with strike price K and maturity time 1. Let c(0)
or c denote its time 0 value. Given the call’s boundary condition, the call’s values at
date 1 are given by max[0, S(1) – K]. We label this as c(1)UP ≡ cU when the stock
price goes up and c(1)DOWN ≡ cD when it goes down (we drop some letters to reduce
clutter). Figure 17.1 shows this model under the binomial tree example. Put values
pU and pD could have been similarly introduced.
Arbitrage-Free Trees
“House built on a weak foundation, will not stand oh no,” sang “King of Calypso”
Harry Belafonte, who popularized Caribbean music. The same can happen with
a model. Because we price options with a binomial tree using the no-arbitrage
principle, the tree must be arbitrage-free for logical consistency. Otherwise, the
whole approach collapses, like a house without a strong foundation.
A simple condition Up factor > Dollar return > Down factor (U > [1 +
R] > D) is both necessary and sufficient to rule out arbitrage profits in the binomial
tree (see the appendix to this chapter for a proof). It states that neither the stock’s
nor the mma’s returns dominate the other, a necessary condition for an economic
equilibrium—and this makes intuitive sense. If one security is superior to the other,
why invest in the inferior asset? This common sense arbitrage-free condition is trivial
to impose and easy to understand. It must also hold in the multiperiod binomial tree,
and it can be generalized. For example, even if U, (1 + R), and D depend on the
node in the tree, the same condition must hold for all nodes. This will prove useful
in Part IV of the book, when we price interest rate options.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Stock Prices and Martingales
Before pricing options, let us study the no-arbitrage condition to see if we can obtain
any additional insights. Suppose that U > (1 + R) > D holds in the stock price tree
so that arbitrage is ruled out. Algebra tells us that U > (1 + R) > D holds if and only
if we can find a unique number 𝜋 strictly between 0 and 1 such that
(1 + R) = 𝜋U + (1 − 𝜋) D (17.4)
(You can see this intuitively: as U is large and D is small, you can alter 𝜋 to create any
number that lies between them.)
Dividing both sides of Equation 17.4 by (1 + R) and multiplying by S gives
𝜋US + (1 − 𝜋) DS
S= (17.5)
1+R
This expression has a useful interpretation. Look at the stock tree in Figure 17.1.
Suppose we say that the “probability” of the stock going up is 𝜋, and that of going
down is (1 – 𝜋). Then, the stock’s expected payoff computed with these probabilities
and discounted to the present by the riskless rate (1 + R) is today’s stock price! This
THE SINGLE-PERIOD MODEL 379
gives us a simple method for computing the stock’s present value that is consistent
with no arbitrage.
Given a stock price tree, it is easy to determine 𝜋. Indeed, solving Equation 17.4
gives
(1 + R) − D U − (1 + R)
𝜋= and (1 − 𝜋) = (17.6)
U−D U−D
What have we done? Our no-arbitrage condition implies the existence of some
numbers, 𝜋 and (1 – 𝜋), which we have interpreted as probabilities, and we have used
them to compute the stock’s present value. However, it is important to emphasize
that they are not the actual probabilities of the stock going up or down (which are
q = 3/4 and [1 – q] = 1/4 in this case), so we call them pseudo-probabilities (or
martingale probabilities or risk-neutral probabilities). The last two names will
make sense shortly.
Next we rewrite Equation 17.5 two more times. Why? Because after it is rewritten,
it is easier to see how the binomial model relates to the BSM OPM studied in
Chapter 19. We thus rewrite Equation 17.5 as (with T = 1)
S = E𝜋 [S (T)] e−rT (17.7a)
where E𝜋 (.) is shorthand for denoting that we are computing an expected value by
using the probabilities 𝜋 and (1 – 𝜋), and we replaced (1 + R) with erT .
This method of computing the stock’s present value as its expected payoff using the
pseudo-probabilities and discounting it backward through time with the mma’s value
(expression [17.7a]) is also known as martingale pricing. To see this interpretation,
we need to rewrite this equation using the notation for a mma’s value. Let A and
A(1) = er be the time 0 and time 1 value, respectively, of the mma. Then, Equation
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
17.7a is
S S (T)
= E𝜋 (17.7b)
A [ A (T) ]
This expression shows that S(t)/A(t) is a martingale under the pseudo-probabilities.
What is a martingale? It’s a stochastic process X(t) whose time t value equals the
expected value of X(T) at some later time T. Martingales are associated with “fair
gambles” because what you have today is what you expect to have tomorrow. In
finance, this sense of fairness gives us a pricing system with no arbitrage opportunities.
Because of Equation 17.7b the pseudo-probabilities are often called the martingale
probabilities.
The Pricing Model
Inspired by the example, we create a synthetic call using a replicating portfolio V (0)
[≡ mS + b] with m shares of the stock worth S per share and b units of the mma
priced at a dollar each.
At time 1, we want to find (m, b) to match the portfolio’s value to the traded call’s
value in each state. This yields two equations in two unknowns:
in the “up” state ∶ V(1)UP = mUS + b (1 + R) = cU (17.8a)
380 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
in the “down” state ∶ V(1)DOWN = mDS + b (1 + R) = cD (17.8b)
where
cU = max (US − K, 0) and cD = max (DS − K, 0) (17.8c)
A solution exists and solving them gives
m = (cU − cD ) / (US − DS) (17.9a)
b = (UcD − DcU ) / [(U − D) (1 + R)] (17.9b)
m is known as an option’s delta or the hedge ratio. We will soon show why it is the holy
grail of options pricing.
The cost of constructing this portfolio is
V (0) ≡ mS + b
cU − cD UcD − DcU
= S+ (17.10)
(U − D) S (U − D) (1 + R)
1 (1 + R) − D U − (1 + R)
= c + c
1 + R [( U − D ) U ( U − D ) D ]
Now to prevent arbitrage, the synthetic call’s cost of construction, V (0), must
equal the traded call’s price, c. This is the option’s arbitrage-free price! As seen in
Equation 17.10, this value depends only on (S, U, D, R, K), which we refer to
somewhat flippantly as KRUDS. These are the current stock price (S), the up and
down factors (U, D), the dollar return on the mma (1 + R), and the contract’s strike
price (K). Missing from the call’s value are q and (1 – q), the actual probabilities of
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the stock going up or down.
This omission is important and it occurs because, in the derivation of the option’s
price, we use an exact replication argument. No matter which node of the tree the
stock price moves, the synthetic call’s payoffs exactly duplicate the option’s payoffs.
Hence we don’t care what the actual probability is of moving up or down.
The omission of the actual probabilities from the option’s price is important
because it implies that if two investors agree on the KRUDS, but they disagree on q,
they will still agree on the call’s price. Of course, disagreement about the probability
q might lead to disagreement about the stock price S when (U, D, R) are fixed. But
if agreement about S occurs (which you can readily see for actively traded liquid
stocks), then the call’s price is fixed by our no-arbitrage argument.
The Hedge Ratio
Classical myths depict the Holy Grail as the cup used by Jesus Christ at the Last
Supper. Believed to possess miraculous powers, it has been lost and sought for
centuries. Indeed, the quest for the lost Holy Grail features prominently in the tales of
the legendary knights of King Arthur’s court. Options pricing theory also has its holy
grail, the hedge ratio, the number of shares of the stock to hold for each written call
to form a perfect hedge. This hedge ratio has miraculous powers because with it one
THE SINGLE-PERIOD MODEL 381
can construct a synthetic call to obtain the arbitrage-free price of a traded option.
Without it, our argument fails.
The hedge ratio for the binomial model is given by Equation 17.9a:
cU − cD
m=
(U − D) S
To prove this, consider a portfolio consisting of the short call and m shares of the
stock. The initial value of this portfolio is
− c + mS
The values of this portfolio at time 1 in the up and down states are:
cU D − cD U
− cU + mUS =
(U − D)
cU D − cD U
− cD + mDS =
(U − D)
Because these values are equal, the covered call is riskless—and we have found the holy
grail! To reemphasize, the hedge ratio works perfectly if the world behaves according
to our single-period model. Finding a hedge ratio that works for more complex and
realistic OPMs is a much harder task. Our quest for the holy grail (of options pricing)
continues in the coming chapters for these more complex OPMs.
Risk-Neutral Valuation
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Another interesting insight can be obtained using some additional algebra on
Equation 17.10. We can use the pseudo-probabilities to rewrite the call’s value in
another form, which has a nice economic interpretation. Recall that Equation 17.6
gave 𝜋 = [(1 + R) – D]/(U – D) and (1 – 𝜋) = [U – (1 + R)]/(U – D). Using them,
we can rewrite the call’s arbitrage-free value (17.10) as
c = [πcU + (1 − π) cD ] / (1 + R) (17.11)
This equation is an example of the risk-neutral valuation procedure. The
option’s value is its expected payoff using the pseudo-probabilities and discounted
using the mma’s rate. As the legendary baseball coach Yogi Berra said, “it’s like déjà
vu, all over again!” The pseudo-probabilities 𝜋 and (1 – 𝜋) are back to give us a simple
way of computing present values! This interpretation gives the pseudo-probabilities
an alternative name: risk-neutral probabilities.
Using Equation 17.8c, we can write (17.11) compactly (using T = 1) as
c = E ᄡ {max [0, S (T) − K]} e−rT (17.12)
which will be shown later to be similar to an expression used to express the BSM
model in Chapter 19.
382 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
EXAMPLE 17.2: Risk-Neutral Pricing of European Options in a Single-Period
Binomial Model
■ Consider the data from Example 17.2. The model starts at date 0 (time 0) and ends at date 1 (time T).
■ Three securities trade:
- YBM’s stock price S is $100 today. The up factor U is 1.2000 and the down factor D is 0.9025.3
- A dollar invested in the mma earns r = 5 percent per year continuously compounded and gives a
dollar return of (1 + R) = erT = e0.05 × 1 = $1.0513 in one year.
- As the strike price K is $110, the option prices at maturity are cU = max(120 – 110, 0) = 10.00
and pU = max(110 – 120, 0) = 0 when the stock goes up to $120.00 and cD = max(90.25 – 110,
0) = 0 and pD = max(110 – 90.25, 0) = 19.75 when the stock goes down to $90.25.
■ Equation 17.6 gives the pseudo-probability of an up movement 𝜋 = [(1 + R) – D]/(U – D) = 0.5001.
■ The risk-neutral pricing equation (17.9) gives today’s option prices:
c = [𝜋cU + (1 − 𝜋) cD ] / (1 + R) = $4.76
p = [𝜋pU + (1 − 𝜋) pD ] / (1 + R) = $9.39
These are the same prices that we got in Example 17.2.
3 We derive them by using 𝜎 = 0.142470 in Jarrow–Rudd specification (see Extension 18.1).
Actual versus Pseudo-probabilities
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
We noted that the pseudo-probabilities (𝜋 and [1 – 𝜋]) for up and down movement
in the stock prices, respectively, differ from the actual probabilities (q and [1 – q]).
But how are they related?
Because the two sets of probabilities denote positive fractions, we can link them
by introducing two positive terms 𝜙U and 𝜙D that denote an adjustment for risk:
𝜋 = 𝜙U q (17.13a)
(1 − 𝜋) = 𝜙D (1 − q) (17.13b)
The appendix to this chapter shows why 𝜙U and 𝜙D are an adjustment for risk. We
note that 𝜙U and 𝜙D are not independent. Knowing one, the other can be computed
since 𝜙U q + 𝜙D (1 – q) = 1.
The intuition is simple. To compute present values, it is well known that one can
take expected future values (using the actual probabilities) and discount to the present
using a risk-adjusted rate. The larger the risk, the larger is the risk-adjusted discount
rate. However, if one uses the pseudo-probabilities instead to compute a present value
as in the risk-neutral valuation Equation 17.11, then the discount rate becomes the
riskless rate. Because risk-neutral valuation does not adjust the discount rate for risk,
SUMMARY 383
to get the same present value, the adjustment for risk must occur in the use of the
pseudo-probabilities (as distinct from the actual probabilities). This is indeed the case.
Building on this insight, the appendix also discusses how these two sets of probabilities
and the risk premium can be estimated using market data.
We note that each set of probabilities has its own use: the actual probabilities
contribute (in conjunction with historical data) to the estimation of the volatility,
while the pseudo-probabilities (which we manufacture by using the risk-free interest
rate, the length of a time interval, and the volatility; see Extension 18.1) are used for
computing option prices. These actual and pseudo-probabilities will stay with us in
the chapters that follow.
17.7 Summary
1. In 1973, Fischer Black, Myron Scholes, and Robert Merton revolutionized the
field of derivatives pricing by publishing the BSM model. Since 1973, options
pricing developed along four, somewhat interlinked directions: (1) improving
the BSM model and using it for different underlyings, (2) developing models
that allow for random interest rates, (3) putting derivatives pricing on a rigorous
mathematical foundation, and (4) using computational methods, including the
binomial framework.
2. The binomial model has several advantages: (1) it is a great teaching aid that gives
a simple, intuitive introduction to options pricing, (2) it can be repeated to build a
multiperiod tree, which gives prices that closely approximate the BSM model, (3)
it utilizes martingale pricing, the paramount technique of derivatives valuation,
and (4) it is a versatile model that prices standard European and American options
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and other derivatives.
3. The binomial OPM makes some key assumptions: no market frictions, no credit
risk, competitive and well-functioning markets, no interest rate uncertainty,
no intermediate cash flows, and no arbitrage. In addition, it assumes that the
underlying trades at discrete time intervals and follows a binomial process.
4. The OPM constructs a synthetic option that replicates a traded option’s payoff.
Next, if no arbitrage opportunities exist, the two must have the same value. An
option’s arbitrage-free price is the cost of constructing the synthetic option.
5. The binomial OPM is set up as follows:
- Three securities trade: a stock, a riskless mma, and a European (call or put)
option.
- The stock price goes up and down by the factors of U or D, respectively.
- A dollar invested in the mma gives a dollar return of (1 + R) after one period.
- As the strike price is given, the option’s payoffs at maturity are determined by
the stock prices on the tree. Solve for the option price by the formula
Option price = [𝜋 × (Option payoff)UP +
(1 − 𝜋) × (Option payoff)DOWN ] / (1 + R)
384 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
where 𝜋 = [(1 + R) – D]/(U – D) is the pseudo-probability of an up movement,
and option payoffs “up” and “down” are considered at the up and down nodes.
- The method of computing the expected payoff with pseudo-probabilities and
discounting with the mma’s value is known as risk-neutral valuation.
17.8 Appendix
Proving the No-Arbitrage Argument
The no-arbitrage condition U > (1 + R) > D proves useful in many contexts.
1. Single-period binomial model: U > (1 + R) > D (“Bounded Dollar Return”): This is a
necessary and sufficient condition for no arbitrage in a single-period tree. This is
the result proven in this extension.
2. Multiperiod binomial model: This result must hold at every node in the multiperiod
binomial tree.
3. Binomial interest rate derivatives pricing model: A generalized version of this result is
required for pricing interest rate derivatives in a binomial framework: if U, (1 + R),
and D depend on the node in the tree, then the bounded dollar return condition
must hold for all nodes (see Jarrow, 2002b).
4. Pseudo-probabilities: We will soon see that this condition is intimately connected
with the existence of pseudo-probabilities (or equivalent martingale probabilities
or risk-neutral probabilities) in binomial models.
Thus the condition that we establish is linked to the core ideas at the heart of
options pricing.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
RESULT 1
No-Arbitrage Condition
The condition U > (1 + R) > D is equivalent to no arbitrage.
PROOF OF PART I: NO ARBITRAGE IMPLIES U > (1 + R) > D. Assume
the condition U > (1 + R) > D fails. We show that this gives arbitrage profits. A
violation of U > (1 + R) > D can happen in one of two ways—either (1 + R) is to
the left side of U > D (i.e., [1 + R] ≥ U > D) or to its right (i.e., U > D ≥ [1 + R]).
■ Suppose (1 + R) ≥ U > D. This makes the mma a better investment than the
stock, irrespective of the stock’s going up or down. Suppose the stock price is $100
today. Short sell the stock and invest the 100 proceeds in the mma. At the end of
the period, your arbitrage profit is 100[(1 + R) – U] ≥ 0 if the stock goes up or
100[(1 + R) – D] > 0 if the stock goes down. This is an arbitrage strategy.
APPENDIX 385
■ Suppose U > D ≥ (1 + R). This makes the mma an inferior investment to the
stock. Borrow at the risk-free rate by short selling the mma, invest the proceeds in
the stock, and collect arbitrage profits.
PROOF OF PART II: U > (1 + R ) > D IMPLIES NO ARBITRAGE. Recall
from chapter 6 that arbitrage can be expressed as a costless strategy that nonetheless
pays something in the future. With two assets, a stock, and a mma, how would you
obtain a zero-investment portfolio? Obviously, you cannot buy or sell both assets
simultaneously (how would you finance such trades?). You must buy one asset and
short the other. However, you cannot make arbitrage profits if U > (1 + R) > D
holds. Why? Because in one of the up or down nodes, the payoff to the portfolio will
be negative.
To see this, suppose that you buy the stock and short the mma obtaining a zero
initial investment portfolio. The payoff at time 1 if you go up is U – (1 + R) > 0.
The payoff if you go down at time 1 is D – (1 + R) < 0. This portfolio is not an
arbitrage opportunity. Conversely, if you short the stock and long the mma, the zero
initial investment portfolio has a negative payoff in the up node and a positive payoff
in the down node. This zero investment portfolio is also not an arbitrage opportunity.
The Probabilities and Risk Premium
RELATING ACTUAL AND PSEUDO-PROBABILITIES Let us further
explore the concepts of the actual probabilities (q, 1 – q) and the pseudo-probabilities
(𝜋, 1 – 𝜋) in the context of our single-period model and compute the risk premium
that connects them.
First, the probabilities must agree on zero-probability events. If one of the actual
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
probabilities is zero, the stock price tree will fail to branch out and the stock becomes
a riskless asset. The condition q > 0 if and only if 𝜋 > 0 prevents this collapse.
Next, recall that we linked the pseudo-and actual probabilities by Equations 17.13a
and 17.13b:
𝜋 = 𝜙U q (1a)
(1 − 𝜋) = 𝜙D (1 − q) (1b)
where 𝜙U and 𝜙D are positive numbers that adjust for risk. We note that 𝜙U and 𝜙D
are not independent. Knowing one, the other can be computed since 𝜙U q + 𝜙D (1
– q) = 1. You can write them as two values taken by the random variable 𝜙.̃ To
summarize, today’s stock price is S, and the parameters take on the following values
after one period:
US is associated with q > 0, 𝜋 > 0, and 𝜙̃ = 𝜙U
S (1) =
{DS corresponds to (1 − q) > 0, (1 − 𝜋) > 0, and 𝜙̃ = 𝜙D
It follows from this definition that the expectation of the random variable 𝜙̃ computed
using the actual probabilities is given by
E (𝜙̃ ) ≡ 𝜙U q + 𝜙D (1 − q) = 𝜋 + (1 − 𝜋) = 1 (2)
386 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
THE RISK PREMIUM Equation 17.5 allows us to write the present value of
the stock price at time 0 as the discounted expected payoffs using the pseudo-
probabilities:
𝜋US + (1 − 𝜋) DS
S= (3)
1+R
Using (1a) and (1b), we rewrite this as
q (𝜙U US) + (1 − q) (𝜙D DS)
S= (4)
1+R
Notice that we are adjusting for risk via the cash flows in the numerator. This contrasts
with the traditional way of computing a present value. The traditional method is to
adjust the denominator. You may have taken courses in capital budgeting or project
evaluation, in which you learned to compute the net present value by discounting
risky cash flows, positive or negative, with a risk-adjusted discount rate.
Interestingly, 𝜋 and q are related by a risk premium. Using algebra and statistics,
we show that the risk premium is given by cov(−𝜙,̃ S(1)/S ). To establish this, first,
note from Equation 17.17 that we have
E (𝜙S̃ (1))
S=
1+R
̃ (1))
E (𝜙S
or 1 + R = (5)
S
where E(.) denotes the expectation computed by using the actual probabilities q and
(1 – q).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
A familiar result in statistics (for two random variables x and y, cov[x, y] = E[xy]
– E[x]E[y]) gives
cov (𝜙,̃ S (1)) = E (𝜙S
̃ (1)) − E (𝜙̃ ) E (S (1)) (6)
Using (2) and transferring the last expression to the left-side, we get
cov (𝜙,̃ S (1)) + E (S (1)) = E (𝜙S
̃ (1)) (7)
We can write (5) using (7) and another result from statistics [cov(ax, y) = acov(x, y)] as
̃
E (𝜙S(1)) S (1) E (S (1))
1+R= = cov 𝜙,̃ + (8)
S ( S ) S
APPENDIX 387
Rearrangement of Equation 8 gives
E (S (1)) S (1) S (1)
= 1 + R − cov 𝜙,̃ = 1 + R + cov −𝜙,̃
S ( S ) ( S )
E (S (1)) (9)
or S =
S (1)
1 + R + cov −𝜙,̃
[ ( S )]
We rewrite Equation (9) as the following result.
RESULT 2
The Risk Premium Linking Actual and
Pseudo-probabilities
In a single-period model, the actual and pseudo-probabilities are linked by a
risk premium cov(−𝜙,̃ S(1) / S), which is given in the following expression:
q (US) + (1 − q) DS
S= (10)
S (1)
1 + R + cov − 𝜙̃ ,
( S )
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where cov( , ) is the covariance operator.
Thus the risk premium is cov(−𝜙,̃ S(1)/S). As shown, the risk premium depends
explicitly on the pseudo-probabilities, proving they give an adjustment for risk. The
risk premium can be explicitly derived in an equilibrium model. Equilibrium
models (such as the capital asset pricing model) require the equality of demand
and supply, the presence of utility functions and endowments, which determine risk
premium (defined as the expected risky return less the risk-free rate). By contrast,
arbitrage models (like our options pricing models) do not need such features and
use only the no-arbitrage condition to derive the results. Notice that the difference in
the two approaches to computing present values is that we only need to use arbitrage
pricing for expression (3), but for expression (10), we need to compute both 𝜙U and
𝜙D , which requires an equilibrium model.
ESTIMATING PROBABILITIES How would you compute the probabilities? As
noted before, the computation of the pseudo-probabilities (𝜋, 1 – 𝜋) is given in
Equation 17.6. The actual probabilities can be computed using historical data and
388 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
statistics for a single-period binomial distribution. First, using historical data, estimate
the expected return Ê (return); then set it equal to the theory’s expected return,
Ê(return) = qU + (1 – q)D, and solve for q. Remember that this is a toy model,
and although we have shown how to solve for q, the estimation issues become more
involved in the case of more complex models—but they use the same ideas as shown
here!
17.9 Cases
Leland O’Brien Rubinstein Associates Inc.: Portfolio Insurance (Harvard Busi-
ness School Case 294061-PDF-ENG). The case studies the rise and fall of Leland
O’Brien Rubinstein Associates’ portfolio insurance selling business.
Leland O’Brien Rubinstein Associates Inc.: SuperTrust (Harvard Business
School Case 294050-PDF-ENG). The case examines Leland O’Brien Rubinstein
Associates’ attempts to rebuild itself after the 1987 stock market crash by creating
new products to meet the unsatisfied needs of equity investors.
Tata Steel Limited: Convertible Alternative Reference Securities (B) (Richard
Ivey School of Business Foundation Case 910N32-PDF-ENG, Harvard Business
Publishing). The case considers valuation of a convertible bond offering by a
global top ten steel producer.
17.10 Questions and Problems
17.1. Why is the binomial model a useful technique for approximating options
prices from the Black–Scholes–Merton model? Describe some applications
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and uses of this model.
17.2. How is the no-arbitrage principle used in the binomial model to find options
prices?
17.3. In the binomial options pricing model, it is assumed that the stock price
follows a binomial process.
a. Is this a reasonable description of the actual stock price process?
b. If not, why should one study this model?
17.4. a. In the binomial options pricing model, what assumptions are made about
dividends and interest rates?
b. In the binomial stock price tree, what restrictions are needed on the up
and down factors, relative to the risk-free rate, to avoid arbitrage? Explain.
17.5. In the binomial options pricing model, what is the hedge ratio?
17.6. a. In the binomial options pricing model, what are the pseudo-probabilities?
b. In the binomial options pricing model, what does risk-neutral valuation
mean? Explain.
QUESTIONS AND PROBLEMS 389
17.7. When pricing an option using risk-neutral valuation, one is assuming that all
investors are risk neutral. Hence, if one believes that investors are risk averse,
risk-neutral valuation cannot be used. True or false? Explain your answer.
FIGURE 17.4:
Today After 1 year
120.773898
100
89.47150422
Call: strike price K = 105, option matures after 1 year.
Money Market Account (mma)
1 1.051271096
Note:
The up factor U is US/S = 120.773898/100 = 1.2077.
The down factor D is DS/S = 89.471504/100 = 0.8947
Dollar return 1 + R = 1.0513.
We report numbers to four places and the final answer to two places after
the decimal point although calculations consider ten places after the decimal
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
point. The next eight questions use the above data.
17.8. a. Given the preceding data, set up a perfect hedge and compute the call
option’s value.
b. What is the hedge ratio? What does it signify?
17.9. a. What are the pseudo-probabilities?
b. Show that the stock price is its discounted expected value using these
pseudo-probabilities.
17.10. Compute the call option’s value using risk-neutral valuation.
17.11. Suppose the market price of the call option is $6. Is there an arbitrage
opportunity? Show how you can take advantage of this price to make arbitrage
profits.
17.12. a. Set up a perfect hedge and compute the put option’s value. Let the put
have the same strike price K = $105.
b. What is the hedge ratio? What does it signify?
390 CHAPTER 17: SINGLE-PERIOD BINOMIAL MODEL
17.13. Compute the put option’s value using risk-neutral valuation.
17.14. Using the data from questions 17.8 and 17.11, verify that put–call parity holds.
17.15. For the data in question 17.13, suppose the market price of the put option is
$3. Is there an arbitrage opportunity? Explain how one would trade to exploit
this arbitrage.
The next four questions are based on the following data for a single-period binomial
model:
■ A stock’s price S is $50. After six months, it either goes up by the factor
U = 1.22095341 or it goes down by the factor D = 0.79881010.
■ Options mature after T = 0.5 year and have strike price K = $45.
■ A dollar invested in the money market account earns continuously com-
pounded risk-free interest at 2 percent per year.
17.16. Compute the call’s value.
17.17. Suppose the market price of the call option is $5. Is there an arbitrage
opportunity? Explain how one would trade to exploit this arbitrage.
17.18. Consider the following exotic option whose payoff at expiration is given
by the stock price squared less a strike price if it has a positive value, zero
otherwise:
max [S(1)2 − K, 0]
Assuming that the strike price K is $2,500, determine the value of this exotic
option under the assumption of no-arbitrage. If the market price of the call
is $600, how would you trade to exploit this arbitrage opportunity?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
17.19. Consider the following exotic option whose payoff at maturity is given by
the square root of the stock price less the strike price if it has a positive value,
zero otherwise:
max [√S (1) − K, 0]
Assuming that the strike price K is $7, determine the value of this exotic
option under the assumption of no-arbitrage. If the market price of the call
is $0.10, how would you trade to exploit this arbitrage opportunity?
17.20. (Microsoft Excel) Implied Volatility
Consider the following data for computing option prices given to you by
your professor.
QUESTIONS AND PROBLEMS 391
FIGURE 17.5:
Stock
Today After 1 year
120.77389800
100
89.47150422
Call: strike price K = 105, option matures after 1 year.
Money Market Account (mma)
1 1.051271096
You want to know how this data was generated. The professor, when asked,
apologizes and says, “I used a Jarrow-Rudd approximation but I lost the data.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
You are an Excel expert—why don’t you use ‘Goal Seek’ and determine what
the volatility is?”
18
Multiperiod Binomial
Model
18.1 Introduction Recasting the Two-Period
Example in the Multiperiod
18.2 Toward a Multiperiod Framework
Binomial Option Pricing
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The n-Period Binomial Option
Model
Pricing Model
The Stock Price Evolution
EXTENSION 18.1 Linking
Binomial Option Price Data the Binomial Model to the
The Stock Price Tree Black–Scholes–Merton Model
18.3 A Two-Period Binomial Model 18.5 Extending the Binomial Model
Backward Induction Known Dividends
Option Pricing via Synthetic Valuing American Options
Construction (Method 1)
18.6 Spreadsheet Applications
Repeat, Repeat: Risk-Neutral
Pricing (Method 2) A Two-Period Binomial Example
One-Step Valuation: Prelude to A Sixteen-Period Example
the Multiperiod Model (Method 3) 18.7 Summary
18.4 The Multiperiod Binomial 18.8 Cases
Option Pricing Model
18.9 Questions and Problems
Binomial Coefficients and
Pseudo-probabilities
TOWARD A MULTIPERIOD BINOMIAL OPTION PRICING MODEL 393
18.1 Introduction
Not too many years back, a large US consulting company valued some options for a
client. Two groups within the company were given the job. One of them used the
celebrated Black–Scholes–Merton (BSM) option pricing model (OPM), which gives
a neat analytical solution. This team (involving one of us) argued that the BSM model
is useful because a formula enables the computation of changes in the option price
in response to changes in the stock price (later, you will see how this helps traders to
hedge and manage an options portfolio). The rival group used a multiperiod binomial
OPM with thirty time periods and thirty-one branches in an Excel spreadsheet. They
argued that the binomial model is easier for clients to understand and that it gives
prices close to the BSM values. Both claims are correct, yet ease of understanding
carried the day.
This chapter extends the binomial model to a multiperiod setting and shows how
to set the parameters to get prices that closely approximate the BSM. First, we develop
a two-period binomial OPM repeating concepts and methods used earlier. Risk-
neutral valuation, when generalized, takes us to the multiperiod model. We also
show the versatility of this numerical approximation technique when it is used to
price American options and for including dividends. Finally, we discuss setting up
the binomial model in a spreadsheet program like Microsoft Excel.
18.2 Toward a Multiperiod Binomial Option
Pricing Model
To construct the multiperiod binomial OPM, we first need to understand the assumed
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
stock price process.
The Stock Price Evolution
Figure 18.1 shows how a stock price path looks under the binomial assumption. The
first figure shows the path in the single-period binomial model, which is the model
from chapter 17. It’s a line parallel to the horizontal axis. Here it corresponds to the
value US. Alternatively, it could have taken the value DS. In looking at this graph,
one can easily reject the single-period model as a good approximation over, say, a
one-year horizon when compared to the actual stock price path given in the next
figure.
The next figure also shows the stock price path in a multiperiod binomial setting.
You can still reject this evolution just by observation, but what happens if these periods
shrink and become really small? We have superimposed a realistic looking stock price
path on this graph. When the time steps become small, the multiperiod binomial
price process approximates this actual price path quite well! Here the multiperiod
binomial model may not be so easily rejected. In fact, choosing the actual probability
of going up q, the up factor U, and the down factor D appropriately, the multiperiod
binomial evolution approaches the lognormal distribution that we use in the next
chapter to obtain the BSM model.
394 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
FIGURE 18.1: The Stock Price Evolution in a Single
Period and Multiperiod Setting
Single Period
Stock price
US Realized stock price path
DS Possible path
0
Today After 1 year Time
Multiple Periods
Stock price
Actual stock path
U2S Stock price path (model)
US
S
DS
D2S
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
0
Now 1 year Time
Binomial Option Price Data
Our examples will use the binomial option pricing data (BOP data!) for Your Beloved
Machines Corp. (YBM), which are the same data used in Example 17.1 in the last
chapter.
■ YBM’s current stock price S is $100.
■ We consider a European call on YBM with strike price K = $110 and maturity
time T = 1 year.
■ The model starts today (time 0) and ends after one year at time T = 1.
■ The stock price return volatility 𝜎 is 0.142470.
■ The number of periods n varies; it can be 1, 2, 3, or more. If n is 2, we divide the
time to maturity T into two periods of length Δt = T/n = 0.5 year each. If n is
3, then Δt = 1/3 = 0.3333 year, and so on.
TOWARD A MULTIPERIOD BINOMIAL OPTION PRICING MODEL 395
■ We let the continuously compounded risk-free interest rate r be 5 percent per year.
In the one-period model, the dollar return 1 + R = e0.05 = $1.0513 after one year.
In the two-period model, the dollar return 1 + R = erΔt = e0.05×0.5 = $1.0253 after
one period and (1 + R)2 = 1.02532 = $1.0513 after two periods—at year’s end.
This is a benefit of continuous compounding. No matter how you slice and dice
the time intervals, your investment still grows by the same amount over a year.
Recall that all of our computations are performed to 16 decimal place accuracy,
although for clarity in exposition, we report rounded numbers up to either 2 or
4 decimal places.
The Stock Price Tree
The next step is to price options in a two-period binomial model. Example 18.1
presents the stock price tree and the money market account used for options pricing.
EXAMPLE 18.1: A Stock Price Tree and Option Prices at Maturity
■ As before, a stock, a money market account, and a European call option trade. We use the BOP data.
The model has three dates: today is the starting date (time 0), six months later is date 1, and the
maturity is date 2 (time T = 1 year), when the call expires. As the number of periods n is 2, there are
two time periods of length Δt = T/n = 0.5 years each.
■ The asset price evolution is shown in Figure 18.2. The dollar return is 1 + R = $1.0253 after one
period. YBM’s stock price evolves according to a multiplicative binomial model. From any node in
the binomial tree, the next period’s stock price is obtained by either multiplying the current stock
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
price by the up factor U or by the down factor D. We compute these factors using specification 1 in
Table 18.1 (see Extension 18.1): U = 1.1282 and D = 0.9224.
- Starting from a price of S = $100 at date 0, the stock can either go up to US = 1.1282 × 100
= $112.82 or down to DS = $92.24 at the end of the first period.
- If the stock price is US at date 1, then it either goes up to U(US) or down to D(US) at date 2. If it
has a value of DS at date 1, then it subsequently goes up to U(DS) or down to D(DS). Hence
U (US) = 1.1282 × 112.82 = $127.29
D (US) = 0.9224 × 112.82 = $104.07 = U (DS)
D (DS) = 0.9224 × 92.24 = $85.08
- Notice that the “up and then down” dotted path combines with the “down and then up” dashed
route because D(US) = U(DS). The branches combine because we use a multiplicative model. This
feature is very useful when there are many periods. For example, if the branches recombine, a two-
period model would have three instead of four separate nodes. However, in a 10-period model,
there would be 10 + 1 = 11 separate nodes in a recombining tree but 210 = 1,024 separate nodes
if the branches do not recombine.
- A generalization of the result (U > 1 + R > D) of chapter 17 is essential to make the tree arbitrage-
free. We assume that this condition holds at every node in the tree.
396 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
- The actual probability of the stock going up is q = 3/4 and that of YBM going down is (1 – q) =
1/4. Multiplying the probabilities together gives the probability of two up movements as q2 = 9/16,
an up and a down movement as 2q(1 – q) = 3/8, and two down movements as (1 – q)2 = 1/16.
■ You can easily compute the options prices at date 2.
- The call’s value at maturity is given by max[0, S(2) – K]. Consider the topmost node in Figure 18.2.
The stock price there is U 2 S. Thus when the stock goes “up and then up,” the call value at time
2 is c(2)UP,UP ≡ cUU = max(U2 S – K, 0). Note that “UU” appears as a subscript in the call price.
Also note that we drop “(2)” to reduce clutter. The call has a strike price K = $110. Hence cUU =
max(127.29 – 110, 0) = 17.29. The call finishes out-of-the-money at the other two nodes on date
2, so c(2)UP,DOWN ≡ cUD = 0 and c(2)DOWN,DOWN ≡ cDD = 0.
- The put’s value at date 2 is max[K – S(2), 0]. Using this formula, pUU = $0, pUD = $5.93, and
pDD = $24.92.
18.3 A Two-Period Binomial Model
We will utilize the previous data to compute European options prices using a
technique called backward induction.
Backward Induction
In Sir Arthur Conan Doyle’s first novel A Study in Scarlet (1887), Sherlock Holmes
made a classic comment about his crime detection technique: “In solving a problem
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
of this sort, the grand thing is to be able to reason backward. That is a very
useful accomplishment, and a very easy one, but people do not practise it much.”
Interestingly the approach used by Sherlock Holmes is called backward induction,
a mathematical technique where one starts at the end and then solves a problem by
working backward through time. We use it to solve the multiperiod binomial option
pricing. The steps are as follows:
■ Create the binomial stock price tree by repeatedly attaching the single-period
structure to the end of each node until the tree reaches the option’s maturity date.
■ Starting at the last date in the tree, given the strike price, you can easily compute
the option’s values on the expiration date for all possible final nodes.
■ At the end of the last period, consider the option values ordered from the top node
to the bottom node. Select the top two option values and apply the risk-neutral
pricing formula to get the topmost option price at the previous date. Next, move
down one node at the final time, and repeat this computation for the next two
option values. Continue moving down the final nodes of the tree in this fashion,
until the bottommost node is reached. When reached, this process generates all the
option prices at the beginning of the last period from the top to the bottom of the
tree.
A TWO-PERIOD BINOMIAL MODEL 397
FIGURE 18.2: Two-Period Binomial Model for Pricing a
European Option
Numerical Example
Stock Stock Stock Call Put Pseudo Prob Actual Prob
Today 6 months 1 year 1 year 1 year 1 year 1 year
127.29 17.29 0 0.25 9/16
112.82
100.00 104.07 0 5.93 0.50 6/16
92.24
85.08 0 24.92 0.25 1/16
Two-Period Model
Stock Stock Stock Call Put Pseudo Prob Actual Prob
Date 0 Date 1 Date 2 Date 2 Date 2 Date 2 Date 2
π U2S cUU pUU π2 q2
π US
q (1 – π)
S UDS cUD pUD 2π (1 – π) 2q (1 – q)
(1 – π) q
DS
(1 – q)
(1 – q) D2S cDD pDD (1 – π)2 (1 – q)2
Money Market Account (mma)
Date 0 Date 1 Date 2
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1 1.0253 1.0513
1 1.0253
Today 6 months 1 year
1 1+R (1 + R)2
■ Step back one time step (toward date 0), and compute the option values at the
beginning of the current period, using the procedure just described. Continue
doing this period by period, working backward through time. Eventually you come
to today and then the process ends. The solution is today’s option price.
Option Pricing via Synthetic Construction (Method 1)
This section uses data from Example 18.1 to demonstrate how to price options in a
two-period binomial framework. We begin by setting up a dynamic hedge using the
stock and a money market account to replicate the call’s value. Dynamic hedging is
the extension of our synthetic construction to a multiperiod setting.
398 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
We use a dynamic, self-financing trading strategy to construct a portfolio that
replicates the traded option’s payoffs. This portfolio is dynamic because it changes
over the multiple periods, and it is self-financing because there are no net cash
flows at intermediate dates. When no arbitrage is allowed, the cost of constructing
this synthetic option must equal the price of the traded option.
EXAMPLE 18.2: Pricing a Call via Synthetic Construction
■ When solving the model, like Sherlock Holmes, we have to work backward from the end. Inspired
by the example, we create at date 1 (one date before the last date) a portfolio V (1) ≡ m(1)S(1) + b(1)
with m(1) shares of the stock worth S(1) and b(1) units of the money market account priced $1 each.
This portfolio is constructed to replicate the option’s values at date 2. It is the synthetic call option.
Its cost of construction must equal the traded call value at date 1, or else arbitrage opportunities arise.
■ We have to do the replication exercise twice: once when the stock goes up to US and once more
when it goes down to DS. In each case, we have to solve two equations in two unknowns to find the
share holdings m(1), b(1).
■ We use the same data as in Example 18.1 (see Figure 18.2 for a summary of this information) to
illustrate this method. To reduce clutter, we use U and D in the subscripts instead of UP and DOWN,
for example, V (2)UP,UP is expressed as V (2)UU ⋅ At date 2, when the stock is at US (= 112.82) on
date 1,
In the “up, up” state, V (2)UU = m (1) U2 S + b (1) (1 + R) = cUU
In the “up, down” state, V (2)UD = m (1) UDS + b (1) (1 + R) = cUD
Solving these equations, we get
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
m(1)U = (cUU – cUD ) / (U2 S – UDS) = 17.29/ (127.29 – 104.07) = 0.7445
b(1)U = (UcUD – DcUU ) / [(U – D) (1 + R)]
(18.1)
= – (0.9224 × 17.29) / [(1.1282 – 0.9224) 1.0253]
= –75.57
Thus the cost of constructing the synthetic call at date 1 in the up state is V (1)U = m(1)U US +
b(1) = 8.4345 = $8.43. To rule out arbitrage, this must equal the traded call’s price cU on date 1 in the
up state. At date 2, when the stock is at DS (= 92.2363) on date 1, either solve the two equations to
find that the call is worthless at date 1, or infer that if the call is sure to be worthless at date 2, then it
must also be worthless at date 1. So when the stock goes down to 92.24, m(1)D = 0 and b(1)D = 0,
giving the synthetic option a value of cD = 0 at date 1 in the down state.
■ Finally, let us now move back to date 0. Construct a portfolio V (0) ≡ m(0)S(0) + b(0) with m(0) shares
of the stock worth S(0) and b(0) units of a money market account worth $1 each. This portfolio’s share
holdings are constructed to match the synthetic call prices cU and cD at date 1, giving two equations
in two unknowns. At date 1,
A TWO-PERIOD BINOMIAL MODEL 399
In the “up” state, V (1)U = m (0) 112.82 + b (0) 1.0253 = 8.43 = cU
In the “down” state, V (1)D = m (0) 92.24 + b (0) 1.0253 = 0 = cD
Solving these equations, we get m(0) = 0.4097 and b(0) = – 36.8536, so the cost of constructing the
synthetic call at date 0 is
V (0) = m (0) S (0) + b (0) = 4.1135 = $4.11
This must equal c(0) ≡ c, the traded call’s price today. Any other quoted price will create get-
rich opportunities for vigilant arbitrageurs. The numbers from this exercise are summarized in
Figure 18.3.
This simple example illustrates several important features of the multiperiod
binomial OPM:
■ Prices differ from the single-period model prices. Despite using the same BOP data,
the single- and two-period models give different options prices: $4.76 and $4.11,
respectively. This is because the two trees aren’t identical. Compare the figures
to notice that the stock and the option prices have different distributions at the
maturity date.
■ Self-financing. We show next that the dynamic trading strategy is self-financing.
Initiate the trading strategy at date 0 by buying 0.4097 units of the stock priced at
$100 and finance this by selling 36.8536 units of the money market account worth
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
$1 each. You still need to invest $4.11, which is the call price. Suppose the stock
goes up to $112.82 after six months. Then the portfolio’s value is $8.43. Liquidate
this portfolio and raise funds by selling 75.57 units of the money market account for
$1 each. Use the total proceeds of $84 (= $8.43 + $75.57) to buy 0.7445 units of
the stock. Obviously, this transaction is self-financing. After the up node, suppose
the stock goes up again to $127.29 at the end of the year. Then the portfolio’s value
is (0.7445 × 127.29 – 75.57 × 1.0253) = $17.29. If the stock goes down from there
to $104.07, then the portfolio value is (0.7445 × 104.07 – 75.57 × 1.0513) = 0.
These payoffs are identical to the call values at the corresponding nodes in the tree
at date 2. Suppose the stock goes down to $92.24 at date 1 instead. Then the call
becomes worthless. No trades are necessary. The call has zero value thereafter.
■ Dynamic market completion. Just two assets, a stock and a money market account,
replicate the three option values at maturity. Note that we have dynamically
completed the market, where the term complete refers to the portfolio’s ability to
match all possible option values at maturity using a dynamic trading strategy.
■ A hedging interpretation. Notice the holy grail, the hedge ratios m(0), m(1)U , and
m(1)D in the model. They tell us how to rebalance our portfolios to synthetically
construct the option values in the nodes that follow.
400 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
FIGURE 18.3: The Two-Period Binomial Model for
Pricing European Call Options
(Numerical Values)
Today (date 0) 6 months (date 1) 1 year (date 2)
cUU = 17.29
cU = 8.43 U2S = 127.29
m(1)U = 0.7445
c = 4.11 b(1)U = –75.57
US = 112.82 cUD = 0
m(0) = 0.4097 UDS = 104.07
b(0) = –36.8536 cD = 0
S = 100 m(1)D = 0
b(1)D = 0 cDD = 0
DS = 92.24 D2S = 85.08
$1 becomes 1 + R = $1.0253 after six months. Strike price K = $110.
■ Independent of actual probabilities. Note that the actual probabilities of the stock
moving up or down do not enter the valuation procedure. As in the single-period
binomial model, this is because we exactly replicate the option’s payoffs at every
node in the tree on the expiration date. Hence the probabilities of reaching these
nodes do not affect the option’s values.
Repeat, Repeat: Risk-Neutral Pricing (Method 2)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Why crank out option values by solving linear equations when we can summon the
power of algebra? Plug the algebraic values from expression (18.1) into the portfolio’s
value to get the synthetic call’s value at date 1:
cUU − cUD UcUD − DcUU
cU = V(1)U = m(1)U S(1)U + b(1)U = US +
(U’D) US (U’D) (1 + R)
Rearrange terms (keep (1 + R) in the denominator, gather cUU and cUD and rewrite)
to get
(1 + R) − D U − (1 + R)
c + c
( U − D ) UU ( U − D ) UD
cU = (18.2)
1+R
Recalling Yogi Berra, “It’s like déjà -vu, all over again!” Our pseudo-probabilities
are back. As before, we write the pseudo-probability of going up as 𝜋 = [(1 + R) –
D]/(U – D) and going down as (1 – 𝜋) = [U – (1 + R)]/(U – D). Now you can
compactly write the binomial call OPM as
cU = [𝜋cUU + (1–𝜋) cUD ] / (1 + R) (18.3)
A TWO-PERIOD BINOMIAL MODEL 401
One can develop similar formulas for call prices at the down node cD and at today’s
date c. Moreover, replacing c with p gives the formula for pricing puts! Generalize
this to get a formula for pricing European options.
RESULT 18.1
A Formula for Pricing Options by Repeated Application
of Risk-Neutral Pricing
c = [𝜋 × cU + (1–𝜋) × cD ] / (1 + R) (18.4)
where 𝜋 = [(1 + R) – D]/(U – D), U and D are the up and down factors,
respectively, and 1 + R is the dollar return.
The same formula applies for puts.
The next example uses this formula.
EXAMPLE 18.3: Pricing a Call by Repeated Application of Risk-Neutral Pricing
Using the same data as in Examples 18.1 and 18.2, we get
𝜋 = (1.0253 − 0.9224) / (1.1282 − 0.9224) = $0.50
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
cU = [𝜋cUU + (1 − 𝜋) cUD ] / (1 + R) = (17.29 × 0.50) /1.0253 = $8.43
(18.5)
cD = [𝜋cUD + (1 − 𝜋) cDD ] / (1 + R) = 0
c = [𝜋cU + (1 − 𝜋) cD ] / (1 + R) = (0.50 × 8.43) /1.0253 = $4.11
which verifies today’s call price.
One-Step Valuation: Prelude to the Multiperiod
Model (Method 3)
For pricing European options, you can simplify further and do it all in one step. In
expression (18.4), plug the expressions for cU and cD into the call value c and simplify
to get
𝜋cU + (1 − 𝜋) cD 𝜋2 cUU + 2𝜋 (1 − 𝜋) cUD + (1 − 𝜋)2 cDD
c= = (18.6)
(1 + R) (1 + R)2
This is the two-period binomial model. The numerator computes the expected
payoff using the pseudo-probabilities and the denominator discounts the cash flows.
The model’s description follows:
402 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
■ As the pseudo-probability of going from S to US is 𝜋, and from US to U 2 S is also
𝜋, the pseudo-probability of going from S to U 2 S is 𝜋2 , which is the pseudo prob-
ability of reaching cUU . Similarly, the pseudo-probability of reaching UDS (which
corresponds to the call price of cUD ) is 𝜋 (1 – 𝜋). As there are two such paths, multi-
ply by 2, giving an overall probability of 2𝜋(1 – 𝜋). The stock payoff D2 S (and call
value cDD ) has the pseudo-probability (1 – 𝜋)2 . These pseudo-probabilities are
noted in the last column of the first figure.
■ Once you have these pseudo-probabilities, compute expected payoffs by multiply-
ing each option payoff by the pseudo-probability of reaching those nodes.
■ Next, compute the overall expected payoff for the call by summing up the expected
payoffs computed in the previous step.
■ This gives the expected payoff at time T. Discount by the dollar return (1 + R)2
to get today’s expected payoff.
EXAMPLE 18.4: A Two-Period Binomial Call Pricing Model
Using the pseudo-probabilities from Example 18.3 and the option payoffs from Example 18.1 in
expression (18.6) gives today’s call price:
c = (0.25 × 17.29 + 0.50 × 0 + 0.25 × 0) /1.02532 = $4.11
One can also price European puts by any one of these three methods.1 Moreover, if you know the
value of one option, you can price the other using put–call parity.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1 For example, to replicate a put, set up a portfolio V ≡ mS + b, as before. Working backward with the second-period put prices
pUU = 0, pUD = $5.93, and pDD = $24.92, one gets pU = $2.89 and pD = $15.05 and, ultimately, today’s put price p = $8.75.
Methods 2 and 3 work similarly.
EXAMPLE 18.5: European Put Pricing by Put–Call Parity
Using the BOP data and the call price from the previous example, put–call parity for European options
(Result 16.1) gives the put price as
p = c + KB – S = 4.11 + 110 × (1/1.0513) –100 = $8.75
THE MULTIPERIOD BINOMIAL OPTION PRICING MODEL 403
18.4 The Multiperiod Binomial Option
Pricing Model
We can utilize the Binomial Theorem from mathematics to develop a generalized
version of the option pricing model. Many distinguished scholars have worked on
this theorem including Sir Isaac Newton, truly the first rocket scientist!
Binomial Coefficients and Pseudo-probabilities
The binomial coefficient (see the binomial theorem in Appendix A) provides a
general formula for computing the number of ways of reaching a final payoff in the
n
multiperiod stock price tree. The binomial coefficient (read “n choose j”)gives
(j)
the number of combinations of selecting j objects out of n objects:
n n!
= (18.7a)
( j ) j! (n − j)!
where the expression n! (read n-factorial) is defined as
n! = n × (n – 1) × (n – 2) × … × 2 × 1
(18.7b)
and 0! = 1
Because the number of up movements uniquely identifies the number of different
paths reaching a particular end-period stock payoff (which automatically fixes the
number of down movements), we can use the binomial coefficient to compute the
number of such paths. Suppose the stock price at expiration is Uj Dn−j S. As this has
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
j up movements and (n – j) down movements, the number of paths is given by
n
the binomial coefficient . To compute the total pseudo-probability of reaching
(j)
j n−j
U D S, multiply the number of ways to reach this node by the probability of each
path occurring 𝜋j (1 − 𝜋)n−j . The product is given by
n!
𝜋j (1 − 𝜋)n − j (18.8)
( j! (n − j) ! )
Next, we use this to dress up our two-period model in this binomial garb.
Recasting the Two-Period Example in the
Multiperiod Framework
In the two-period binomial tree (n = 2), the stock price reaches U2 S or UDS or D2 S
at maturity. The topmost node U2 S is reached by two ups in the stock price. Using
n = 2 and j = 2, we write the pseudo-probability of the outcome as
404 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
2 2 2!
𝜋 (1 − 𝜋)0 = 𝜋2 = 𝜋2 (18.9a)
(2) 2! (2 − 2)!
(Note that [1 – 𝜋]0 = 1.)
One reaches UDS in two ways: “up and then down” and “down and then up.”
Using n = 2 and j = 1, the total pseudo-probability of reaching UDS is
2 1 2!
𝜋 (1 − 𝜋)2 − 1 = 𝜋 (1 − 𝜋) = 2𝜋 (1 − 𝜋) (18.9b)
(1) 1! (2 − 2)!
Two downs in a two-period tree is reached by a unique path. Using n = 2 and
j = 0, the total pseudo-probability of reaching D2 S is
2 0 2!
𝜋 (1 − 𝜋)2 = (1 − 𝜋)2 = (1 − 𝜋)2 (18.9c)
(0) ( 0! (2 − 0)! )
One can compactly express expressions (18.9a) – (18.9c) as
2 j 2!
𝜋 (1 − 𝜋)2 − j = 𝜋j (1 − 𝜋)2 − j (18.9d)
(j) ( j! (2 − j)! )
where j = 2, 1, and 0, respectively
These match the pseudo-probabilities determined earlier (for convenience, we also
report these values in Figure 18.4). Recall that our two-period call valuation formula
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
(18.6) was
𝜋2 cUU + 2𝜋 (1 − 𝜋) cUD + (1 − 𝜋)2 cDD
c=
(1 + R)2
Using expression (18.9a), rewrite the first expression in the numerator as
2
𝜋2 cUU = 𝜋2 (1 − 𝜋)0 × max [U2 S − K, 0]
(2)
Notice that one can write the call payoff as max[0, U j D2 − j S − K] with j = 2 in this
particular instance.
Similarly, rewriting the other two terms in the numerator and using the summation
notation yields a compact expression for the call’s price:
2
1 2 j
c= 𝜋 (1 − 𝜋)2 − j × max [0, U j D2 − j S − K] (18.10)
(1 + R)2 ∑
j=0
( j )
THE MULTIPERIOD BINOMIAL OPTION PRICING MODEL 405
FIGURE 18.4: Two-Period Binomial Model for Pricing
European Options
Stock Stock Stock Call Pseudo Probability
t=0 t=1 t=2
2 2
U2S cUU = max (U2S – K, 0) π (1 – π)0 = π2
2
π US
2 1
S UDS cUD = max (UDS – K, 0) π (1 – π)2 – 1 = 2π(1 – π)
1
(1 – π) DS
2 0
D2S cDD = max (D2S – K, 0) π (1 – π)2 – 0 = (1 – π)2
0
Money Market Account (mma)
Date 0 Date 1 Date 2
1 1+R (1 + R)2
The n-Period Binomial Option Pricing Model
Replace 2 by n in expression (18.10) to get the general multiperiod binomial model,
which is the main result of this chapter.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
RESULT 18.2
The n-Period Binomial Option Pricing Model
n
1 n j
c= 𝜋 (1 − 𝜋)n − j × max [0, U j Dn − j S − K] (18.11)
(1 + R)n ∑
j=0
( j )
where n is the number of time periods, T is the option’s maturity date, Δt ≡
T/n is the length of each time period, 1+R ≡ erΔt is the dollar return from the
money market account for a time period, j is the numbers of up movements
n
in n periods (thus [n – j] is the number of down movements), =
(j)
n!
is the number of ways of choosing j “ups” out of n movements, 𝜋
j! (n − j)!
406 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
is the pseudo-probability for the up movement ([1 – 𝜋] is the same for the
down movement), U is the up factor for the stock price movement and D is
the down factor, S is the initial stock price, and K is the strike price.
Replacing the call payoff by max(0, K − Uj Dn−j S) gives the formula for
pricing a European put option.
Who but its creator would appreciate such a hideous-looking construct? In its
defense, expression (18.11): (1) is a generalization of the simpler formulas introduced
earlier, (2) it just uses high school algebra (of course, with some notations), and (3) it
is easily programmed into a spreadsheet program such as Microsoft Excel (see the
example at the chapter’s end).
To reiterate, expression (18.11) generalizes our simple formula for European call
pricing developed in the previous section. The first expression within parentheses
uses dollar returns to discount the cash flows. Next comes the summation sign, which
adds up the “option payoffs” multiplied by the “pseudo-probability of reaching those
payoffs” for all possible cases. The result is the option’s price!
As in the last chapter, we can rewrite Result 18.2 more abstractly to get an
expression similar to the one used in the context of the BSM model (which is
explained in the next chapter):
c = E𝜋 [c (T)] e– r T (8.12)
where E𝜋 (⋅) denotes computing the expectation under the pseudo-probability 𝜋.
Note, as emphasized before, that the option price does not depend on the actual
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
probabilities of the stock price evolution but only on the pseudo-probabilities.
Extension 18.1 shows how to choose the up and down factors to obtain a model
that approximates the BSM model when the time steps are small. In the limit, these
approximations generate a lognormal distribution that underlies the BSM model.
You can use Result 18.2 for the one-step valuation of European options in
multiperiod binomial trees. The next example demonstrates how to do this when
the model has four periods.
EXAMPLE 18.6: A Four-Period Binomial Example
■ This four-period example uses the same BOP data. Given the number of time periods n is 4, divide
the time to maturity T into four periods of length Δt = T/n = 0.25 years each.
■ Using specification 1 in Table 18.1, we get U = 1.08458539, D = 0.94056717, 𝜋 = 0.50001507,
and (1 – 𝜋) = 0.49998493. Moreover, 1 + R = erΔt = e0.05 × 0.25 = $1.0126 after one period and
(1 + R)4 = 1.01264 = $1.0513 after four periods.
THE MULTIPERIOD BINOMIAL OPTION PRICING MODEL 407
■ Using expression (18.11), we get today’s European call price:
⎧
⎪ 4 ⎫
⎪
1 4 0 4 0 4 3 1 3 1
c= ⎨ 𝜋 (1 − 𝜋) max ( 0, U D S − K ) + 𝜋 (1 − 𝜋) max ( 0, U D S − K ) ⎬
(1 + R)4 ⎪ (4) (3) ⎪
⎩ ⎭
1 4 4
=( 0.5000 × 28.3742 + 4 × 0.5000 × 10)
1.0513 ) (
= 4.0653 or $4.07
What happens to the option values if we keep on increasing the number of
periods in our numerical example? This is explored in Table 18.1, where we generate
options prices by applying specification 1 of Extension 18.1 using finer and finer time
intervals. Here Table 18.1a contains some of the Excel workings (for the binomial
model for n = 1, 2, 4, 8, 16, and 32; the second panel is a continuation of the
computations in the first panel), and Table 18.1b reports call and put prices
EXTENSION 18.1: Linking the Binomial Model to the Black–Scholes–Merton
Model
The binomial model is easy to understand and simple to implement, but will it give prices that one can
comfortably quote in real markets? The answer is yes! The binomial OPM can be used to approximate the BSM
model. The approximation works through approximating the probability distribution for the underlying asset’s
price, which the BSM model assumes is a lognormal distribution. The distribution for the binomial process
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
quickly approaches this when the parameter values are properly specified and the model is run over many small
intervals. Consequently, options values from the multiperiod binomial models converge to the BSM model.
There are several ways of achieving this convergence. “The finance literature has revealed no fewer than 11
alternative versions of the binomial option pricing model for options on lognormally distributed assets,” wrote
Chance (2008, p. 38), who noted that the three most widely used versions are Cox et al. (1979), Rendleman and
Bartter (1979), and Jarrow and Rudd (1982).
The last specification, known as the Jarrow–Rudd binomial model, was developed by Jarrow and Rudd
(1982). It builds on Rendleman and Bartter’s approach and was further modified in Jarrow and Turnbull’s
(1996) book. In their specification, as the time interval diminishes, the pseudo-probability approaches 0.50 (see
Ext. 18.1 Tab. 1).
Another specification is the Cox–Ross–Rubinstein binomial model. Though widely used in practice, it
gives erroneous results for large time intervals or small volatilities (see Ext. 18.1 Tab. 1). For both specifications,
the greater the number of intervals, the better will be the approximation. One must use at least fifty periods to
generate good options prices in the binomial models.
Which specification is superior? The answer is, they are all equivalent, because they all approach the same
BSM model prices as n goes to infinity. Chance (2008, p. 54) concludes that the (binomial) “model is not a
single model but a family of interpretations of a discrete-time process that converges to the continuous Brownian
motion process [which underlies the BSM models] in the limit and accurately prices options.”
408 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
Ext. 18.1 Tab. 18.1: Setting Up the Factors and the Pseudo-Probabilities to
Link the Binomial and BSM Models
Specification 1 (Jarrow–Rudd) Specification 2 (Cox–Ross–Rubinstein)
𝜎2
rΔt − Δt + 𝜎√Δt
Up factor (U) U=e 2 e𝜎√Δt
2
𝜎
rΔt − Δt − 𝜎√Δt
Down factor (D) D=e 2 1/U
(1 + R) − D (1 + R) − D
Pseudo-probability of stock
going up (𝜋) U−D U−D
U − (1 + R) U − (1 + R)
Pseudo-probability of stock
going down (1 – 𝜋) U−D U−D
rΔt
where Δt is the length of the time interval, r is the continuously compounded annual risk-free interest rate, 1+R = exp(rΔt) ≡ e , and 𝜎2 is the variance
of the continuously compounded return.
TABLE 18.1: Convergence of Binomial Option Prices
to the Black–Scholes–Merton Model Val-
ues
a. Excel Workings (Binomial Model for n = 1, 2, 4, 8, 16, and 32)
S K T n ∆t = T/n r 1+R Σ
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
$100 $110 1 1 1 0.05 $1.0512711 0.14247
100 110 1 2 0.5 0.05 1.02531512 0.14247
100 110 1 4 0.25 0.05 1.01257845 0.14247
100 110 1 8 0.125 0.05 1.00626957 0.14247
100 110 1 16 0.0625 0.05 1.00312989 0.14247
100 110 1 32 0.03125 0.05 1.00156372 0.14247
U D 𝜋 (1 – 𝜋) Call Put
1.199999511 0.90247088 0.500120656 0.499879344 $4.75727159 $9.39250829
1.128249231 0.92236346 0.500042629 0.499957371 4.113492 8.7487287
1.084585393 0.94056717 0.500015067 0.499984933 4.065304709 8.7005414
1.056912736 0.95562533 0.500005326 0.499994674 3.944816311 8.58005301
1.038843626 0.96741588 0.500001883 0.499998117 3.813430739 8.44866743
1.026783114 0.97634426 0.500000666 0.499999334 3.767810681 8.40304738
EXTENDING THE BINOMIAL MODEL 409
b. Options Prices
Number of Intervals Call Option Price Put Option Price
1 $4.757272 $9.392508
2 4.113492 8.748729
4 4.065305 8.700541
8 3.944816 8.580053
16 3.813431 8.448667
32 3.767811 8.403047
64 3.803293 8.438529
BSM model 3.783772 8.419009
for these periods (plus that for n = 64). The resulting values converge to the BSM
model prices; however, you will have to wait until the next chapter to learn how the
BSM model values are obtained.
18.5 Extending the Binomial Model
The multiperiod binomial model is an all-purpose tool useful for pricing derivatives
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
more complex than the options studied to this point. As such, it is useful to think of
the multiperiod model as a computational procedure for valuing arbitrary derivative
securities. We illustrate two extensions: (1) including dividends and (2) valuing
American options.
Known Dividends
US options exchanges do not alter the strike price or make adjustments for cash
dividends. Such dividends lower the value of a call but raise the value of a put. Pricing
options in the presence of dividends is tricky business. We consider two relatively
straightforward ways of adjusting the binomial model for known dividends.
As in chapters 3 and 11, the ex-dividend date is the cutoff date for a dividend
payment, and we assume that arbitrage ensures that the ex-dividend stock price falls
by the dividend (see Result 3.1). Fixed-percentage dividends are easier to model in
the binomial framework because the tree still recombines. Unfortunately, fixed-dollar
(not fixed-percentage) dividends are the norm among US corporations. Corporations
try to hold steady both the dividend date and the dollar amount. Example 18.7 shows
how to adjust the model for both fixed-dollar and fixed-percentage dividends.
410 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
EXAMPLE 18.7: Adjusting for Known Dividends (Based on Example 18.2)
Known Dividend Date and a Percentage Dividend Rate
■ In Example 18.2, assume that the stock pays a dividend of 𝛿 = 5 percent just before six months.
Then the stock price at time 1 needs to be adjusted to reflect this dividend. Accordingly, the revised
stockprices at time 1 are
US (1–dividend rate) = 112.824923 (1– 0.05) = 107.18
DS (1–dividend rate) = 92.2363 (1– 0.05) = 87.62
■ Proceed as before. Use these revised stock prices to create the binomial tree. Six-month as well as
subsequent stock prices are multiplied by (1 – dividend rate). The tree recombines as before, and our
option valuation technique works. Verify that the call value is $2.60 and the put value is $12.23.
Known Dividend Date and a Dollar Dividend
■ Suppose that in Example 18.2, the stock pays the dividend div = $5 just before six months. In
that case, compute the stock price on date 1 as US or DS as before, and then subtract $5 from
it. Now create the binomial tree with these revised prices and solve for the options values as before
(see Figure 18.5). Note that the tree does not recombine. There are four nodes after one year in a
two-period tree (had this been a 10-period tree, you would have 1,024 final nodes!). Solve for the
options values by working backward through time, and verify that the call value is $2.77 and the put
value is $12.28.
Valuing American Options
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
American options can be easily valued in a binomial model using the following steps:
■ Create a binomial stock price tree as for the European option and find the option
values after n periods when the option matures.
■ Using the risk-neutral valuation formula, compute (1) the European option’s value
at the “last but one” period (i. e., [n – 1] periods) and (2) the option’s payoff if it
is immediately exercised. Retain the larger of these two as the American option’s
value and use it in future computations.
■ Work backward period by period. At each node, determine whether the option is
worth more alive (computed value) or dead (exercised value), and retain the larger
of these two values. The procedure ends when you reach today’s date. The option’s
value at time 0 gives the solution. By following this algorithm, you have created
the American option price tree.
■ Note that a by-product of the valuation is a description of whether to exercise
or not at every node in the tree. Hence the valuation technique also provides the
optimal exercise strategy.
EXTENDING THE BINOMIAL MODEL 411
FIGURE 18.5: Two-Period Tree for Pricing
Options When the Underlying
Pays a Known Dividend
Stock Stock Stock Call Put
Today 6 months 1 year 1 year 1 year
Dividend paid
121.65 11.65 0
(112.82 – div)
107.82
99.45 0 10.55
100
98.42 0 11.58
(92.24 – 5)
87.24
80.46 0 29.54
■ Result 16.10 of Chapter 16 tells us that an American call on a stock that pays
no dividends over the option’s life is never exercised early: you get more by
selling the call rather than exercising it. Such an American call is equal to a
European call, and they have identical binomial trees. In contrast, American put
options may be exercised early, and they differ from European puts. Next, we use
Examples 18.1–18.3 to create an American put option tree in Example 18.8.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
EXAMPLE 18.8: American Option Pricing (Based on Examples 18.1–18.3)
■ Remember that a call option on a stock that pays no dividends over the option’s life is never exercised
early. Consequently, the European and American call option trees are the same as in Example 18.2.
You can verify this by performing the calculations based on the procedure stated earlier.
■ The story is different for American put options. For convenience, the first tree in Figure 18.6 gives
the European put. The second tree shows the American put.
■ Six months before maturity, the European put’s value in the up state is 2.8942. If the put is exercised,
the payoff is
110 – 112.8249 = –2.8937
The put is worth more alive than dead—don’t exercise. Retain the value 2.8937 in the American put
option tree.
■ Six months before maturity, the put option’s value in the down state is 15.0477. If the put is exercised,
the payoff is
max (110 – 92.2363, 0) = 17.7637
Here the put is worth more dead than alive. Discard the existing value 15.0494 and replace it with
17.7637 in the tree.
412 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
■ Move back to the present. For the American put, if the put is exercised, it gives
max (110 – 100, 0) = 10
Using the American put values from time 1, we find at time 0
𝜋PU + (1 − 𝜋) PD 0.50 × 2.8937 + 0.50 × 17.7637
P= = = $10.0730
1+R 1.0253
If we exercise put today, we get 110 − 100 = $10. As $10.0730 is the highest, it is recorded as today’s
American put value, and the put is not exercised today.
FIGURE 18.6: Two-period Binomial
Model for Pricing
American Options
Put Put Put
Today 6 months 1 year
European put
0
2.8942
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
8.7508 5.9345
15.0494
24.9246
American put
0
2.8942
10.0733 5.9345
15.0494
17.7637 24.9246
SPREADSHEET APPLICATIONS 413
18.6 Spreadsheet Applications
The binomial model can be implemented using spreadsheet programs like Microsoft
Excel. This section demonstrates how to do this in two steps: First, we show how to
solve a two-period binomial model, and second, we show how to solve a sixteen-
period example.
A Two-Period Binomial Example
You can program the multiperiod binomial OPM in a Microsoft Excel workbook.
EXAMPLE 18.9: Solving Example 18.3 by Excel
Generating Inputs
■ In an MS Excel worksheet, type in the following. This example uses the BOP data provided earlier
in the chapter and used in Examples 18.2 and 18.3. See Figure 18.7 for the worksheet:
- S in cell A1 and 100 in cell A2
- K in B1 and 110 in B2
- T in C1 and 1 in C2
- n in D1 and 2 in D2
- “Δt = T/n” in E1 and 0.5 in E2 (you can create Δ and other Greek symbols).
- r in F1 and 0.05 in cell F2
1 + R in G1 and “=EXP(F2*E2)” in G2 and then hit return to get 1.025315.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
-
- 𝜎 in A4 and 0.142470 in A5
■ Choose the up and down factors according to one of the specifications connecting the binomial model
with the BSM model (see Extension 18.1). We use the Jarrow–Rudd specification:
- Type U in B4. For the up factor U = exp[(r – 𝜎2 /2) Δt + 𝜎 √Δt], type “=EXP(((F2–
(A5*A5/2))*E2)+(A5*SQRT(E2)))” in B5 and hit return to get 1.128249.
- Type D in C4. For the down factor D = exp[(r – 𝜎2 /2) Δt – 𝜎√Δt], type “=EXP(((F2–
(A5*A5/2))*E2)–(A5*SQRT(E2)))” in C5 and hit return to get 0.922363.
■ Type 𝜋 in D4 and then “=(G2–C5)/(B5–C5)” in D5 and hit return to get 0.500043. Type (1 – 𝜋) in
E4 and then “=1–D5” in E5 and hit return to get 0.499957.
414 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
FIGURE 18.7: Two-Period Binomial Example with Excel
(Example 18.3)
Column A B C D E F G
Row S K T n Δt = T/n r 1+R
1
2 100 110 1 2 0.5 0.05 1.025315
3
4 σ U D π (1 – π)
5 0.142470 1.128249 0.922363 0.500043 0.499957
6
7
8 Stock (t = 0) Stock (t = 1) Stock (t = 2) Call (t = 2) Put (t = 2) Pseudo Probability
9 127.294633 17.294633 0.000000 0.250043
10 112.824923
11 100 104.065586 0.000000 5.934414 0.500000
12 92.236346
13 85.075435 0.000000 24.924565 0.249957
14
15
16 EUROPEAN OPTION TREES (METHOD 2)
17
18 Call tree
19 Call (t = 0) Call (t = 1) Call (t = 2)
20 17.294633
21 8.434532
22 4.113492 0.000000
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
23 0.000000
24 0.000000
25
26 Put tree
27 Put (t = 0) Put (t = 1) Put (t = 2)
28 0
29 2.893700
30 8.748729 5.934414
31 15.047745
32 24.924565
33
34
35 EUROPEAN OPTION PRICES (METHOD 3)
36 call = 4.113492
37 put = 8.748729
SPREADSHEET APPLICATIONS 415
Creating Binomial Stock Tree and Computing Options Prices and Pseudo-
probabilities at Maturity
■ To create the stock tree, type the headings “Stock (t = 0)” in A8, “Stock (t = 1)” in B8, and “Stock
(t = 2)” in C8 and then type the following expressions after “=” in each cell; follow up by hitting the
return key at the end:
- type “=A2” in cell A11 and hit return to get 100
- A11*B5 in B10 to get 112.824923
- A11*C5 in B12 to get 92.236346
- B10*B5 in C9 to get 127.294633
- B10*C5 in C11 to get 104.065586
- B12*C5 in C13 to get 85.075435
■ For call prices at maturity, type the heading “Call (t = 2)” in D8 and then type the following after
“=” in each cell and hit return at the end:
- MAX(C9–B2, 0) in D9 to get 17.294633
- MAX(C11–B2, 0) in D11 to get 0
- MAX(C13–B2, 0) in D13 to get 0
■ For put prices at maturity, type in the label “Put (t = 2)” in E8 and then type the following after “=”
in each cell and hit return at the end:
- MAX(0, B2–C9) in E9 to get 0
- MAX(0, B2–C11) in E11 to get 5.934414
- MAX(0, B2–C13) in E13 to get 24.924565
■ To create the pseudo-probabilities of reaching these option values at maturity, type in the label
“Pseudo-probability” in F8 and then type and hit return:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- D5*D5 in F9 and hit return to get 0.2500043
- 2*D5*E5 in F11 to get 0.500000
- E5*E5 in F13 to get 0.249957
Notice that these pseudo-probabilities sum to 1.
European Call and Put Valuation (Method 2)
■ Type “EUROPEAN OPTION TREES (METHOD 2)” in A16, “Call Tree” in A18
■ You can create the call tree through a repeated application of expression (18.4) of Result 18.1:
- Type in the headings “Call (t = 0)” in A19, “Call (t = 1)” in B19, and “Call (t = 2)” in C19.
- Copy call values at date 2—type D9 in C20, D11 in C22, and D13 in C24 to get cUU = 17.294633,
cUD = 0, and cDD = 0, respectively.
- To solve for call values at date 1, type “=(D5* C20 + E5* C22)/G2” in B21 and
“=(D5* C22+E5* C24)/ G2” in B23.
- Solve and plug in “=(D5* B21+E5* B23)/G2” in cell A22 to get today’s call value 4.113492 or
$4.11.
■ This completes the “call tree” that has all possible call values in the binomial tree. The two-period
“put tree” is similarly created below the call tree and gives a put value of 8.748729 or $8.75.
416 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
European Call and Put Valuation (Method 3)
■ Type “EUROPEAN OPTION PRICES (METHOD 3)” in A35, “call =” in A36, and “put =”
in A37. Using (18.6), we type “= (F9*D9+F11*D11+F13*D13)/(G2ˆ2)” in B36 to get today’s call
value 4.113492 or $4.11.
■ Using a variant of (18.6), where calls are substituted with put values, we type
“=(F9*E9+F11*E11+F13*E13)/(G2ˆ2)” in cell B37 to get today’s put value 8.748729 or
$8.75.
A Sixteen-Period Example
Next we extend the preceding example to build a sixteen-period model using the
same BOP data.
EXAMPLE 18.10: A Sixteen-Period Example (Excel Application)
■ We use the BOP data: S is $100, K is $110, T is 1 year, r is 5 percent per year, and 𝜎 is 0.142470.
This time, we slice up T into sixteen periods of length Δt each so that Δt = T/n = 1/16 = 0.0625
years. Table 18.3a reports the values obtained by using specification 1 ( Jarrow–Rudd) of Table 18.1:
U = 1.038843626
D = 0.96741588
𝜋 = 0.500001883
(1 – 𝜋) = 0.499998117
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ The detailed workings are reported in Table 18.3b.
- Column A has j, the number of up movements.
- Column B has the call symbols (e. g., in the row corresponding to j = 14, we indicate the call symbol
by U14D2S-K; this has stock price for fourteen ups and two downs, from which is subtracted the
strike price).
- Column C has the call values at expiration.
16
- Column D has the combination, the number of ways of getting j up movements .
(j)
- Column E has the pseudo-probability of the up movements 𝜋j .
- Column F has the pseudo-probability of the down movements (1 − 𝜋)16−j .
- Column G has the pseudo-probability of the particular path D × E × F.
- Column H has the “expected payoff ” C × G.
SPREADSHEET APPLICATIONS 417
■ Summing all the values in column H, discounted by (1 + R)16 , gives the call option value, c = $3.81.
As we have done the hard work for the call, we can use put–call parity to obtain the put premium:
p = c + KB − S
= 3.81 + (110/1.0513) − 100
= $8.45
TABLE 18.2: A Sixteen-Period Binomial Example (Solved Using
Microsoft Excel)
a. Basic Data, Computations, and European Call and Put Values
S K T n ∆t = T/n r 1+R 𝜎 U D
100 110 1 16 0.0625 0.05 1.00313 0.14247 1.038844 0.967416
𝜋 (1 – 𝜋) Call Put
0.500002 0.499998 $3.813432 $8.448667
b. Excel Workings
A B C D E F G H
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
j Call Symbol Call Combination Pseudo- Pseudo- Pseudo- Expected
Values at Probability Probability Probability Payoff C*G
Expiration of Up of Down of
Jumps 𝜋j Jumps (1 – Particular
𝜋)(16–j) Path
D*E*F
16 U16S–K 73.993666 1 0.000015 1.000000 0.000015 0.001129
15 U15DS–K 61.342818 16 0.000031 0.499998 0.000244 0.014977
14 U14D2S–K 49.561804 120 0.000061 0.249998 0.001831 0.090754
13 U13D3S–K 38.590817 560 0.000122 0.124998 0.008545 0.329768
12 U12D4S–K 28.374162 1,820 0.000244 0.062499 0.027772 0.788002
11 U11D5S–K 18.859973 4,368 0.000488 0.031249 0.066652 1.257053
10 U10D6S–K 9.999951 8,008 0.000977 0.015625 0.122194 1.221936
9 U9D7S–K 1.749117 1,1440 0.001953 0.007812 0.174562 0.305329
Sum/1.051271 = 3.813431
418 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
18.7 Summary
1. The basic binomial OPM is described by the following:
a. Three securities trade: a stock, a riskless money market account, and a European
(call or put) option.
b. The stock price goes up by factors of U or D from each node. The stock price
tree repeats this single-period structure at the end of each node.
c. A dollar invested in the money market account gives a dollar return of 1 + R ≡
erΔt after one period.
d. Given the strike price, you can compute the option’s payoffs at maturity from
the stock tree. Then solve the previous period’s option price by repeatedly
applying the formula
Option price = [𝜋 × (Option price)UP + (1 – 𝜋) ×
(Option price)DOWN] / (1 + R)
where 𝜋 = [(1 + R) – D]/(U – D) is the pseudo-probability of an up movement
and the options prices “up” and “down” are considered at the next up and down
nodes.
2. The n-period multiplicative binomial OPM is
n
1 n j
c= 𝜋 (1 − 𝜋)n − j × [max (0, Uj Dn−j S − K)]
∑
(1 + R) j=o
n {[( j) ] }
n
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where ≡ n! / [j!(n – j)!] is the number of ways of choosing j “ups” out of
(j)
n movements and n! = n × (n – 1) × (n – 2) × … × 2 × 1 is the n-factorial.
3. An American option may be valued in the binomial framework as follows:
a. Create a binomial stock price tree as in the case of the European option and
find the option values at time n, which is the maturity date.
b. Using the risk-neutral valuation formula, compute (1) the European option’s
value at time (n – 1) and (2) the option’s payoff if it is immediately exercised.
Retain the larger of these two as the American option value and use it in future
computations.
c. Work backward period by period. At each node, determine whether the option
is worth more alive (computed value) or dead (exercised value), and retain the
higher of these two values. Stop when you reach today’s date. By following this
process, you have created the American option price tree.
QUESTIONS AND PROBLEMS 419
18.8 Cases
Leland O’Brien Rubinstein Associates Inc.: Portfolio Insurance (Harvard Busi-
ness School Case 294061-PDF-ENG). The case studies the rise and fall of Leland
O’Brien Rubinstein (LOR) Associates’ portfolio insurance selling business.
Leland O’Brien Rubinstein Associates Inc.: SuperTrust (Harvard Business
School Case 294050-PDF-ENG). The case examines LOR Associates’ attempts
to rebuild itself after the 1987 stock market crash by creating new products to meet
the unsatisfied needs of equity investors.
The Value of Flexibility at Global Airlines: Real Options for EDW and CRM
(Kellogg School of Management Case, Case KEL266-PDF-ENG, Harvard
Business Publishing). The case evaluates real options for technology projects
using a binomial model.
18.9 Questions and Problems
18.1. Why is a multiperiod binomial model a better approximation to the actual
stock price process than the single period binomial model?
The next seven questions are based on the following data for a two-period binomial model
- A stock’s price S is $100. After three months, it either goes up and gets
multiplied by the up factor U = 1.163287, or it goes down and gets
multiplied by the down factor D = 0.861785.
- Options mature after T = 0.5 years and have a strike price K = $110.
- A dollar invested in the money market account grows to 1 + R = $1.012578
after three months.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The stock price tree is:
Stock Stock Stock
Today 3 months 6 months
135.3238
116.3287
100.00 100.2503
86.1785
7 4.2673
18.2. Compute the call value in the above tree using synthetic construction.
18.3. Demonstrate how you can make arbitrage profits when a trader quotes a call
price of $2.
420 CHAPTER 18: MULTIPERIOD BINOMIAL MODEL
18.4. Compute the value of the call option using risk neutral valuation.
18.5. Compute. the put value in the above tree using synthetic construction.
18.6. Suppose the market price of the put option in question 18.5 is $10, how
would you take advantage of this mispricing? Explain your answer.
18.7. Given the data in question 18.5, compute the value of the put option using
risk neutral valuation.
18.8. Consider the call and put options in questions 18.2 and 18.5. Show that they
satisfy put–call parity.
The next five questions are based on the following data for a four period binomial model
- Options mature after T = 2 years and have a strike price K = $75.
- The price of the underlying stock price is $50. The stock price evolves
according to the Jarrow–Rudd specification in this 4-period model with
volatility 𝜎 = 0.3.
- The continuously compounded risk-free interest rate r is 5 percent per year.
18.9. a Compute the up and down factors for the stock price movements and the
dollar return (1 + R) for each period.
b Use these up and down factors to create a 4-period tree for the stock price
movements.
18.10. Find the value of all call options in the tree by repeated application of risk
neutral valuation.
18.11. Explain the European call option formula at time 0 based on an n-period
binomial option pricing model.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
18.12. Find the value of all put options in the tree by repeated application of
riskneutral valuation.
18.13. Explain the European put option formula at time 0 based on an n-period
binomial option pricing model.
The next eight questions are based on the following data for a two-period binomial model
- Options mature after T = 0.5 year and have strike price K = $45.
- The price of the underlying stock price is $50. The stock price evolves
according to the Jarrow–Rudd specification in this 2-period model with
volatility 𝜎 = 0.2.
- The continuously compounded risk-free interest rate r is 2 percent per year.
18.14. Compute the up and down factors, the dollar return (1 + R), and the stock
price lattice.
18.15. Compute today’s call option price in this 2-period tree.
18.16. Compute today’s put option price in this 2-period tree.
18.17. Consider an otherwise identical American call option with the preceding data,
maturing at time 2. Compute its value. How do the American and European
call prices compare?
QUESTIONS AND PROBLEMS 421
18.18. Consider an otherwise identical American put option with the preceding data,
maturing at time 2. Compute its value. How do the American and European
put prices compare?
18.19. Compute today’s American put option price in this 2-period tree if the strike
price is $70. Does early exercise occur?
18.20. Consider the following exotic option whose payoff at maturity is given by the
stock price squared less a strike price if it has a positive value, zero otherwise:
max [S(2)2 – K, 0] .
Assuming that the strike price is $2,000, determine the value of this exotic
option under the assumption of no-arbitrage.
18.21. Consider the following exotic option whose payoff at maturity is given by
the square root of the stock price less the strike price if it has a positive value,
zero otherwise:
max [√S (2) – K, 0] .
Using the preceding data except for assuming a new strike price is $5,
determine the value of this exotic option under the assumption of no-
arbitrage.
(Microsoft Excel) Consider the following
- Options mature after T = 0.5 year and have strike price K = $65.
- The price of the underlying stock S is $63. The stock price evolves
according to the Jarrow–Rudd specification in this 3-period model with
volatility 𝜎 = 0.3.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- The continuously compounded risk-free interest rate r is 0.045 per year.
18.22. a. Using Excel, compute today’s call option value with the preceding data.
b. Using Excel, compute today’s put option value with the preceding data.
19
The Black–Scholes–Merton
Model
19.1 Introduction 19.9 Extending the
Black–Scholes–Merton Model
19.2 Nobel Prize–Winning Works
Adjusting for Dividends
(1973)
Foreign Currency Options
19.3 The Assumptions Valuing American Options
EXTENSION 19.1 Bubbles and EXTENSION 19.3 Exotic Options
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Option Pricing
19.10 Summary
19.4 The Pricing and Hedging
Argument 19.11 Appendix
Modeling the Stock Price Evolution
19.5 The Black–Scholes–Merton
Formula Continuously Compounded
(Logarithmic) Return
19.6 Understanding the
Introducing Uncertainty
Black–Scholes–Merton Model
The Central Limit Theorem
Stock Prices and Martingales
First Derivation of the
Risk-Neutral Valuation
Black–Scholes–Merton Formula
Actual versus Pseudo-probabilities (as a Limit of the Binomial Model)
19.7 The Greeks Second Derivation of the
Interpreting the Greeks Black–Scholes–Merton Formula
(Using Risk-Neutral Valuation)
Some Road Bumps Ahead
The Probabilities and the Risk
EXTENSION 19.2 Market Premium
Manipulation and Option Pricing
Proof of (24)
19.8 The Inputs
19.12 Cases
Observable Inputs
Volatility: The Elusive Input 19.13 Questions and Problems
NOBEL PRIZE–WINNING WORKS (1973) 423
19.1 Introduction
The year 1973 saw the publication of the Black–Scholes–Merton (BSM) option
pricing model (OPM) and the birth of the first modern options exchange, the
Chicago Board Options Exchange. Professor Mark Rubinstein declared in his 1994
American Finance Association presidential address, “This model [BSM model] is
widely viewed as one of the most successful in the social sciences and has perhaps been
(including its binomial extension) the most widely used formula, with embedded
probabilities, in human history.” We share the professor’s enthusiasm, but we are a
little less sure about the usage data.1
This chapter presents the BSM model. First, we consider some applications and
uses. Next, we provide a brief history of the events that led to its development
and publication. We describe the model’s assumptions, present the Black–Scholes
formula for valuing European options, and discuss the theory underlying the model’s
structures. We discuss how to gather the inputs, including the most elusive, the
“volatility.” Finally, we show how to extend the BSM model to adjust for dividends,
to price foreign currency options, and we discuss the pricing of American options.
An insert presents a bird’s-eye view of exotic options.
19.2 Nobel Prize–Winning Works (1973)
During the late 1960s, a Harvard-trained applied mathematician, Fischer Black, was
working as a consultant in Boston. Black applied the capital asset pricing model to
value options and derived a partial differential equation (pde) that implicitly solved
the option’s value. He teamed up with Massachusetts Institute of Technology (MIT)
assistant professor Myron Scholes to solve the equation and derive the formula that
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
bears their names. At the same time, a doctoral student of Paul Samuelson, Robert
Merton, was pioneering the use of stochastic calculus to tackle a variety of problems,
including portfolio theory and option valuation. Black and Scholes solved the pde
using the Capital Asset Pricing model. Later, Merton showed Black and Scholes how
to derive their pde using a perfectly hedged portfolio consisting of the stock and the
option.
The Black–Scholes paper had difficulty getting published, Professors Merton
Miller and Eugene Fama used their personal influence to get the paper published in
the Journal of Political Economy as “The Pricing of Options and Corporate Liabilities”
in 1973. The same year also saw the publication of Merton’s (1973a) “Theory of
Rational Option Pricing” in the Bell Journal of Economics and Management Science.
The rest is history. The crowning academic glory came in 1997 when Merton and
Scholes won the Bank of Sweden Prize in Economic Sciences in Memory of Alfred
1
In George Bernard Shaw’s play “O’Flaherty V. C.,” an English general told an Irish war hero, “Well, in
recruiting a man gets carried away. I stretch it a bit occasionally myself. After all, it’s for king and country.
But if you won’t mind my saying it, O’Flaherty, I think that story about your fighting the Kaiser and the
twelve giants of the Prussian guard singlehanded would be the better for a little toning down. I don’t ask
you to drop it, you know; for it’s popular, undoubtedly; but still, the truth is the truth. Don’t you think
it would fetch in almost as many recruits if you reduced the number of guardsmen to six?”
424 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
Nobel. Black, whose contributions were duly acknowledged in the Nobel citation,
unfortunately died from cancer two years earlier and missed the honor.
19.3 The Assumptions
The BSM model, like its binomial approximation, requires several important assump-
tions. We repeat here assumptions A1–A6 of Chapter 17, which are common to both
models and restate much of the discussion. The last assumption (binomial trading
process) is replaced by two assumptions (A7 and A8) that characterize the stock price
evolution under continuous trading.
A1. No market frictions. We assume that trading involves no market frictions
such as transaction costs (brokerage commissions and bid/ask spreads), margin
requirements, short sale restrictions, and taxes (which may be levied on different
securities at different rates). Moreover, assets are perfectly divisible. This helps create
a benchmark model to which one can later add frictions, and it’s a reasonable
approximation for institutional traders.
A2. No credit risk. We assume that there is no credit risk. Also known as default
risk or counterparty risk, this is the risk that the counterparty will fail to perform on
an obligation. The absence of credit risk is a reasonable assumption for exchange
traded derivatives. Over-the-counter derivatives, however, may possess substantial
credit risks, unless there are adequate collateral provisions.
A3. Competitive and well-functioning markets. In a competitive market,
traders’ purchases and sales have no impact on the market price; consequently, traders
act as price takers. In a well-functioning market, price bubbles are absent.
The competitive markets assumption is the workhorse of modern economics. The
relaxation of this assumption is discussed in Extension 19.2, after we present the BSM
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
model. A price bubble happens when an asset’s price substantially deviates from its
intrinsic or fundamental value. Recent research shows that bubbles can invalidate the
methodology we employ for pricing options (see Extension 19.1).
A4. No intermediate cash flows. Many assets reward their holders with cash
flows such as dividends for stocks and coupons for bonds. The BSM model assumes
that the underlying asset has no cash flows over the option’s life. Later in the chapter,
we will show how to relax this assumption.
A5. No arbitrage opportunities. This is self-explanatory.
EXTENSION 19.1: Bubbles and Option Pricing
As discussed in Chapter 17, a price bubble happens when an asset’s price substantially deviates from its intrinsic or
fundamental value. The fundamental or intrinsic value can be defined in two equivalent ways: the first is the present
value of its future cash flows; the second is the price paid if, after purchase, you have to hold the asset forever. A
difference only occurs if one believes that selling can generate a higher value. This difference is the bubble.
THE ASSUMPTIONS 425
OPMs, like any other models, are valid under a given set of assumptions. When an assumption fails, the model
may no longer be valid. If assumption A3 (competitive and well-functioning assumption) fails and a bubble exists,
then many results of option pricing theory no longer hold (see Jarrow, Protter, and Shimbo 2010b).
Fortunately, with respect to the BSM model, assumption A8 is consistent with assumption A3. This is because
the assumption that the stock price follows a lognormal distribution excludes bubbles. Thus, the lognormal
distribution is a very strong assumption. If one believes that there are stock price bubbles, then the lognormal
distribution assumption needs to be modified, and the BSM model no longer applies.
It can be shown that in a market with no arbitrage opportunities or dominated assets, asset price bubbles only
exist when the market is incomplete. Because most option pricing theories assume a complete market (where, recall
from Chapter 8, securities trade that help investors attain any desired future payoffs), so that synthetic construction
is possible, we see that the standard approach to option pricing excludes bubbles. Again, this is a good result, as
long as one believes that there are no stock price bubbles.
But how do you price options when there are asset price bubbles in an incomplete market? In such
markets, alternative pricing methodologies need to be invoked. The two common approaches are using
indifference/utility-based pricing and identifying the risk-neutral probabilities from traded derivatives. Because
exact hedging is no longer possible, we cannot price by synthetic construction; however, risk-neutral valuation
may still apply. When there are bubbles in the underlying asset, pricing is subtle. Recent research suggests that
some results stay the same while others change:
■ Put–call parity still holds for European options in the presence of bubbles.
■ Risk-neutral valuation works for the put’s price because it has a bounded payout, but it does not work for
the call; it turns out that if the underling asset has a price bubble, so does the call.
■ Forwards and futures both inherit the underlying’s price bubble, and even more interesting, futures prices can
have their own price bubbles.
Bubbles are a thorn in the side of the policy makers, a source of problems for asset holders, an opportunity for
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
hedge fund managers, and a hot research area for professors. The three witches in William Shakespeare’s Macbeth
sang together, “Double, double, toil and trouble/Fire burn and cauldron bubble.” Perhaps we should rephrase:
“Bubble, bubble, toil and trouble!”
A6. No interest rate uncertainty. We assume that interest rates are constant
across time. This assumption works for short-lived options whose underlying asset
isn’t sensitive to interest rate movements. This assumption is relaxed in part IV of the
book.
Next, we introduce two assumptions that are crucial for developing the BSM
model
A7. Trading takes place continuously in time. This assumption gives a
realistic description of transacting actively traded stocks. Stocks do not trade only
on a fixed time grid but rather at any time instant during the trading day. This is
particularly true in today’s markets, where a significant amount of trading volume
is attributed to high-frequency computerized trading strategies pursuing fleeting profit
opportunities in the blink of an eye.
426 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
A8. The stock price follows a lognormal probability distribution. This
assumption means that the continuously compounded returns on the stock follow
a normal distribution. We will explain this assumption in two ways: first, by some
diagrams, and second, by adding more structure to the continuously compounded
rate of return concept familiar from Chapter 2.
Does the lognormal assumption give a realistic model of stock price behavior?
As pictures say a thousand words, let’s go back to our diagrams. In Figure 19.1, the
first two diagrams repeat possible stock price evolutions from Figure 18.1, for the
single period and multiperiod binomial models, respectively. As “seeing is believing,”
you can easily surmise, “These don’t look like stock price paths!” The multiperiod
binomial evolution diagram leaves open the intriguing possibility that it might start
looking like a stock price path if the time step shrinks to zero. This brings us to the
third figure, which depicts a stock price path for the lognormal distribution. Here
you can no longer rule out the evolution just by visual inspection. A rejection of this
evolution now requires the use of formal statistical procedures. Hence it is the first
“real” or “realistic” stock price evolution that we have considered.
Next, let’s use Result 2.1c (with slightly enhanced notation): an investment of
L(0) ≡ L dollars in a risk-free account today grows to L(T) ≡ L(0)erT dollars after T
years, where r is the continuously compounded risk-free rate of return per year. Is
there a similar result for stocks? Yes. We can write
S (T) = S (0) eyT (19.1)
where S(T) and S(0) are the stock prices at times T and 0, respectively, and y is the
continuously compounded rate of return on the stock per year.
However, this return is not risk-free. To characterize its properties, we need a
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
model of returns that (1) generates prices that look like the stock price patterns
observed in financial markets, (2) never allows stock prices to be negative (because they are
of limited liability), (3) but allows stock price returns to be positive, zero, or negative (as you
find in the markets), and (4) provides a mathematically tractable model. Samuelson,
Merton, and other researchers achieved these objectives in one stroke by replacing
Bachelier’s arithmetic Brownian motion with a geometric Brownian motion. The
modification gives us a structure similar to (19.1).
The geometric Brownian motion assumption is that y is normally distributed with
mean (𝜇 − 𝜎2 /2)T and standard deviation 𝜎√T. This is expressed as
𝜎2 T
𝜇T − + 𝜎√Tz
S (T) = S (0) e 2 (19.2)
where 𝜇 is the expected return on the stock per year, 𝜎 is the standard deviation
or volatility of the stock’s return per year, and z is a standard normal random
THE ASSUMPTIONS 427
variable with mean 0 and variance 1; that is, z has the probability density function
(a plot of which gives the familiar bell curve)
2
z
1 −
q (z) = e 2 (19.3)
√2𝜋
FIGURE 19.1: Stock Price Evolution under
Discrete and Continuous Trading
Single Period
Stock price
US Realized stock price path
S (single period model)
DS Possible path
0
Today After 1 year Time
Multiple Periods
Stock price
Actual stock path
... Stock price path
US (multiperiod model)
S
DS
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
...
0
Now 1 year Time
Continuous Time Path
Stock price
Actual stock path
Stock path
S (lognormal model)
0
Now 1 year Time
The probability q(z) is the actual probability of the random variable z, which generates
the randomness in the stock price S(T).
428 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
Note that Equation 19.2 satisfies properties 1, 2, and 3 trivially. Property (4)
follows because the normal distribution is so well studied. The appendix to this
chapter provides an axiomatic justification for modeling the stock price with a
lognormal distribution based on the efficient markets hypothesis.
19.4 The Pricing and Hedging Argument
The original proof of the BSM model utilizes the idea of hedging an option position
with an opposing stock trade, a notion used by options writers in London more than
a century ago (Nelson 1904). An understanding of this hedging argument is essential
to understanding the BSM model. The intuition is easily explained in three steps (see
Figure 19.2).
Step 1
■ Call prices and the underlying stock price move together. This happens because as
the stock price goes up, the call has a higher chance of ending up in-the-money,
which makes the call more valuable today.
■ Consequently, the call value is a function of the underlying stock price, and because
the option eventually matures, its price must also depend on time.
Step 2
■ If a long call and a long stock position move together, then a long call and a short
stock position must move in opposite directions. Recall from Chapter 15 that this
creates a partial hedge because the increase in one is offset by a decrease in the
other.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ But they don’t perfectly balance, so it is not a perfect hedge.
Step 3
■ We can modify the partially hedged position in the long call and short stock to
make it a perfect hedge. One can do this by short selling a fraction of a stock for
each long call. Then, for any movement in the stock price, the change in the call
value exactly offsets the change in the short stock position.
■ The perfectly hedged portfolio uniquely identifies the option’s arbitrage-free price.
This happens because the hedged portfolio requires a known initial investment, and
it is riskless over the next time period. To avoid arbitrage, it must earn the risk-free
rate. The equation implied by this restriction, a partial differential equation, when
solved yields the option’s price. The mathematics for this argument is contained in
the appendix to Chapter 20.
The replication argument as just described utilizes the hedge ratio, which we called
the holy grail of option pricing. Jarrow (1999, p. 233) notes that “the idea of
constructing a perfectly hedged portfolio is the key insight of the Black–Merton–
Scholes model, more important than the valuation formula itself. Indeed, if one
considers the meaning of a perfectly hedged portfolio, it becomes apparent that
THE BLACK–SCHOLES–MERTON FORMULA 429
FIGURE 19.2: The Pricing and Hedging
Argument
Stock and call prices move together
Call and short stock positions move in opposite directions
Call and a fraction of a short stock creates a perfectly hedged portfolio
A riskless portfolio of the
stock and option changing
continuosly in time earns the
risk-free rate
Black–Scholes–Merton model
This gives a pde whose
solution is the BSM model
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
it implies that a position in the stock and the riskless asset can be created that exactly duplicates
the changes in value to the call option” [emphasis added]. The creation of synthetic options
is the basis for most OPMs and will be used throughout this book.
19.5 The Black–Scholes–Merton Formula
Under assumptions A1–A8, one can derive the Black–Scholes–Merton formula for
pricing European options. There are three ways of proving this formula: (1) taking the
limit of the multiperiod binomial model of Chapter 18 as the time steps go to zero,
(2) directly applying risk-neutral valuation, and (3) using the hedging argument of
the previous section to obtain a partial differential equation (pde) whose solution yields
the formula. The appendix to this chapter provides the proof using the multiperiod
binomial model, the simplest derivation, and the proof using risk-neutral valuation,
the paramount technique of modern derivatives pricing. The appendix to Chapter 20
provides the pde proof, the derivation originally used by Black and Scholes (1973)
but the least extendable methodology of the three.
430 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
The formula is given by the following result (see Result 19.1).
RESULT 19.1
The Black–Scholes–Merton Formulas for Pricing
European Options
The European call price is
c = SN (d1 ) − Ke−rT N (d2 ) (19.4a)
where S ≡ S(0) is today’s spot price, N( . ) is the cumulative standard normal
distribution function which has mean 0 and standard deviation 1,
S 𝜎2
log ( ) + r + T
K ( 2)
d1 =
𝜎√T
d2 = d1 − 𝜎√T
K is the strike price, r is the continuously compounded risk-free interest rate
per year, T is the time to maturity measured in years (time 0 is today), and 𝜎
is the stock return’s volatility per year, which is the square root of the stock
return’s instantaneous variance.
The European put price is given by
p = Ke−rT N (−d2 ) − SN (−d1 ) (19.4b)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
One can show that the put price follows from the call’s price using put–call parity
for European options along with the result that N(d) + N(–d) = 1.
The first observation to make about the BSM formula is that the call’s value
depends on the stock price (S), the strike price (K), the time to maturity (T), the
riskless interest rate (r), and the stock return’s volatility (𝜎). It does not depend,
surprisingly, on the stock’s expected return (𝜇). This omission is significant and has
two important implications.
First, the stock’s expected return is the most difficult input to estimate. The
estimation of the other inputs is easier and discussed in a subsequent section. In
fact, had the option’s value depended on 𝜇, the formula would not have been used.
Estimating 𝜇 is equivalent to estimating the stock’s risk premium, and no two financial
economists agree on how to do this. This is a well-known problem in portfolio
theory. Consequently, avoiding the estimation of 𝜇 is a blessing in disguise.
The second implication is that because the option value does not depend on 𝜇,
two traders who agree on (S, K, T, r, 𝜎) but disagree on the stock’s expected return,
𝜇, will still agree on the call’s value. Of course, if they disagree on the stock’s expected
return, 𝜇, the traders will probably disagree on the stock price S. However, if they
agree on S, regardless of their views of 𝜇, they agree on the call’s value.
THE BLACK–SCHOLES–MERTON FORMULA 431
Example 19.1 demonstrates how to use Result 19.1, assuming that you are given
all of the inputs.
EXAMPLE 19.1: Using the Black–Scholes–Merton Formula
Consider the binomial option pricing (BOP) data of Chapter 18:
■ YBM’s current stock price S is $100.
■ The strike price K is $110.
■ Time to maturity T is 1 year.
■ The volatility 𝜎 of the stock’s return is 0.142470 per year.
■ The continuously compounded risk-free interest rate r is 5 percent per year.
The European Call Price
■ Compute d1 and d2 :
S 𝜎2 100 (0.14247)2
log ( ) + r + T log ( + 0.05 + (1)
K ( 2) 110 ) [ 2 ]
d1 = = = − 0.2468
𝜎√T 0.14247√1
d2 = d1 − 𝜎√T = − 0.2468 − 0.14247 = −0.3893
■ One can get the cumulative normal distribution function from (1) a spreadsheet program such as
Excel, (2) a calculator, or (3) a table of values for the standard normal distribution (Appendix A at the
end of the book shows how to read this table)2 . Using Excel, we get
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
N (d1 ) = N (−0.2468) = NORMSDIST (−0.2468) = 0.4025
N (d2 ) = N (−0.3893) = NORMSDIST (−0.3893) = 0.3485
■ The zero-coupon bond price is B = erT = e−0.05 = 0.9513.
■ The European call price using the BSM formula is
c = SN (d1 ) − Ke−rT N (d2 )
= 100 × 0.4025 − 110 × 0.9513 × 0.3485
= $3.78
The European Put Price
■ For European puts, we need different values from the cumulative normal distribution function:
N (−d1 ) = N [− (−0.2468)] = N (0.2468) = NORMSDIST (0.2468) = 0.5975
N (−d2 ) = NORMSDIST (0.3893) = 0.6515
432 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
■ The European put price is
p = Ke–rT N (−d2 ) − SN (−d1 )
= 110 × 0.9513 × 0.6515 − 100 × 0.5975
= $8.42
The European Put Price (by Put–Call Parity)
■ We can double-check our computations using put–call parity:
p = c + Ke−rT − S = 3.78 + 110 × 0.9513 − 100 = $8.42
2 In Excel, the command “=NORMSDIST(z)” gives the standard normal cumulative distribution function (in Lotus Notes, use
the command @NORMAL[z, 0, 1, 0]). The distribution has a mean of 0 and a standard deviation of 1.
19.6 Understanding the
Black–Scholes–Merton Model
To better understand the BSM model, we apply the same argument used to obtain
the binomial model in Chapter 17. This argument starts by introducing an analogous
notion to that of the pseudo-probability.
Stock Prices and Martingales
In an arbitrage-free market, it can be shown (as in the binomial model) under the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
lognormal assumption that there exist pseudo-probabilities such that the stock price
today is its discounted expected value at time T, that is,
S (t) = E𝜋t [S (T)] e−r(T−t) (19.5)
where 𝜋(z) are the pseudo-probabilities. This is analogous to Equation 17.7.
Equation 19.5 is a present value formula. In this formula, the adjustment for risk
occurs in the use of the pseudo-probabilities when taking the expectation and not in
the discount rate, which is riskless. Rearranging terms in Equation 19.5, we get
S (t) S (T)
= E𝜋t (19.6)
e rt [ erT ]
where the t corresponds to the date the expectation is taken. In this form, anal-
ogous to Chapter 17, we see that the stock price normalized by the price of a
money market account is a martingale under the pseudo-probabilities. As noted in
Chapter 17, martingales in probability theory are associated with fair games, that
is, games in which the gambler has an equal chance of losing or winning. In our
context, this is an implication of the no arbitrage assumption. In fact, although not
UNDERSTANDING THE BLACK–SCHOLES–MERTON MODEL 433
proven here, the implication also goes in the reverse direction. These insights are the
basis for many generalizations of the BSM formula. Next, we apply this approach to
value a European call option.
Risk-Neutral Valuation
Under the lognormal distribution assumption, the market is complete, and we can
construct a portfolio in the stock and money market account that exactly replicates
the payoffs to the traded call option. As before, the cost of constructing this synthetic
option yields the arbitrage-free price of the traded call, and exactly as in Chapter 17,
this implies that the call price can be computed as its discounted expected payoff using
the pseudo-probabilities and the risk-free rate; that is, risk-neutral valuation applies:
c = E𝜋 {max [0, S (T) − K]} e−rT (19.7a)
It is shown in the appendix to this chapter that one can rewrite this equation in the
following way:
c = E𝜋 {S (T) |S (T) ≥ K} Prob𝜋 {S (T) ≥ K} e−rT
(19.7b)
−KProb𝜋 {S (T) ≥ K} e−rT
Making the identification of the terms in Equation 19.7b with those in the BSM
formula (19.4), we see that
S (0) N (d1 ) = E𝜋 {S (T) |S (T) ≥ K} Prob𝜋 {S (T) ≥ K} e−rT
(19.8)
KN (d2 ) e−rT = K Prob𝜋 {S (T) ≥ K} e−rT
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The identifications help explain the meaning of the BSM formula.
We now see that the first term in the BSM formula, S(0)N(d1 ), can be inter-
preted as the discounted expected stock price at maturity, given that it ends up in-
the-money multiplied by the probability of being in-the-money. This interpretation
makes intuitive sense. Of course, as just noted, the expectation is computed using the
pseudo-probabilities to adjust for the risk of the cash flows. We note that because the
current stock price S(0) is not the expected stock price at maturity, given that it ends
up in-the-money, N(d1 ) is not the probability of the stock being in-the-money.
The second term in the BSM formula, KN(d2 )e−rT , can be interpreted as the
discounted strike price (Ke−rT ) times the pseudo-probability of ending up in the
money, that is, N(d2 ) = Prob𝜋 {S(T) ≥ K}. Again, this interpretation makes intuitive
sense. This explains why the hedge ratio N(d2 ) is always between zero and 1, because
the probability of being in-the-money is always between 0 and 1. It is 0 only when
the stock price will never be in-the-money, and it is 1 only when the stock price will
be in-the-money for sure.
Actual versus Pseudo-probabilities
Are the actual (q[z]) and pseudo (𝜋[z]) probabilities linked? Interestingly, and
not surprisingly, they are linked by a risk premium. The argument is a simple
434 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
generalization of our discussion in Chapter 17. The details are given in the appendix
to this chapter.
First, note that
𝜋 (z) > 0 if and only if q (z) > 0 (19.9a)
that is, the actual and pseudo-probabilities must agree on zero-probability events.
This follows from the no arbitrage assumption. Consistent with this, one can show
that the pseudo-probabilities equal
𝜋 (z) = 𝜙 (z) q (z) (19.9b)
2
where 𝜙(z) = e−(𝜃 /2)+𝜃z > 0 and 𝜃 = −[(𝜇 − r)/𝜎]√T.
The quantity 𝜙(z) in Equation 19.9b represents an adjustment for risk, and
𝜃 is, in fact, the stock’s risk premium. The appendix to this chapter provides the
justification for these statements. Notice that the result looks identical to that from
Chapter 17.
19.7 The Greeks
We can gain a better understanding of the BSM model by understanding how its
prices change when we change the inputs. This is easily accomplished by taking
partial derivatives (given in Result 19.1) with respect to the stock price (S), time
to maturity (T), the riskless interest rate (r), and the stock return’s volatility (𝜎). In
addition, we also compute the second partial derivative with respect to the stock
price. These partial derivatives are called the Greeks because they are often written
using the Greek alphabet—although one of them has a fictitious Greek name! These
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Greeks are given in Table 19.1. They are also useful for hedging and portfolio risk
management, which are discussed in Chapter 20.
Interpreting the Greeks
We first consider the delta, which measures the change in the option’s value when
there is a small change in the underlying stock price. For the call, we have
𝜕c Change in the value of the call
Delta ≡ = = N (d1 ) (19.10a)
𝜕S Change in the value of the stock
The term on the right side of this equation is defined in Result 19.1. This shows
that the delta, the hedge ratio, is a number strictly between zero and one. We can
similarly compute the delta for a put.
The sensitivity of the delta to changes in the underlying stock price is called
the gamma. This is the second partial derivative of the BSM call price formula with
THE GREEKS 435
respect to the stock price, that is
𝜕 (Delta) Change in the delta 𝜕2 c
gamma ≡ = =
𝜕S Change in the value of the stock 𝜕S2
(19.10b)
1
= N′ (d1 ) > 0
S𝜎√T
where N′( . ) is the standard normal density function and all of the remaining symbols
are defined in Result 19.1. The gamma shows that as the stock price increases,
the delta increases. The gamma for the put is similarly computed by taking the
partial derivative with respect to the underlying stock price of the delta for the
put. The gamma for the put and the call are identical. Example 19.2 illustrates these
computations.
EXAMPLE 19.2: Deltas and Gammas for European Options
■ Consider the BOP data (see Example 19.1): S = $100, K = $110, T = 1 year, 𝜎 = 0.142470 per
year, and r = 0.05 per year. Using the formulas from Table 19.1 and computations from the previous
example, we get d1 = −0.2468. Then,
For the call, deltaC = N(d1 ) = 0.4025
For the put, deltaP = −N(−d1 ) = −0.5975
■ We use Microsoft’s Excel spreadsheet program for computing deltas and plotting their values. The
deltas are computed using the NORMSDIST function for stock prices S at $5 intervals ranging from
$50 to $150 (see Table 19.2) and are drawn in Figure 19.3a, where the stock price is depicted on the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
horizontal axis and the delta values are along the vertical axis. Notice that both the deltas are upward
sloping curves.
■ To understand these figures, consider the behavior of the option’s price for extreme stock price
values.
- For stock prices near $150, the call is deeply in-the-money and increases by almost a dollar for
a dollar increase in the stock price, which gives a delta close to 1. In contrast, the put is deeply
out-of-the-money and has a near-zero response to any stock price movement, signifying that the
put’s delta is close to 0.
- For very low stock prices, the call’s delta is almost 0, while the put option’s price decreases by almost
a dollar for each dollar increase in the stock price, yielding a delta close to −1.
■ The gamma is the same for calls and puts. Using the NORMDIST function in Microsoft Excel’s
spreadsheet to compute N′(d1 ), we get
gamma = N′(d1 ) / (S𝜎√T) = 0.0272
We compute the gammas for stock prices S = $50, 55, … ,145, 150; report them in Table 19.2; and
plot them in Figure 19.3b.
436 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
TABLE 19.1: Partial Derivatives of the Black–Scholes–Merton
Model
Derivative Call Option Put Option
𝜕BS
Delta ≡ 0 < N(d1 ) < 1 −1< −N(−d1 )< 0
𝜕S
𝜕2 BS 1
Gamma ≡ N′(d1 ) > 0 Same as for European call
𝜕S2 S𝜎√T
𝜕BS SN′(d1 )𝜎 SN′(−d1 )𝜎 >
Theta ≡ − − rKe−rT N(d2 ) < 0 − + rKe−rT N(−d2 ) 0
𝜕t 2√T 2√T <
𝜕BS
Vega ≡ S√TN′(d1 ) > 0 Same as for European call
𝜕𝝈
𝜕BS
Rho = TKe−rT N(d2 ) ≥ 0 −TKe−rT N(−d2 ) ≤ 0
𝜕r
BS denotes European option prices c and p as obtained from Result 19.1; N( · ) is the cumulative normal distribution
function with mean 0 and variance 1; and N ′ ( · ) is the normal density function.
The sensitivity of the option value to changes in the time to expiration is called
theta:
𝜕c Change in the value of the option
theta ≡ =
𝜕t Decrease in the time to expiration
SN′ (d1 ) 𝜎
=− − rKe−rT N (d2 ) < 0
2√T
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The theta of the call is negative. This makes intuitive sense because as time passes
(which means ∆t is positive) and the time to maturity decreases, the option becomes
less valuable. This follows for two reasons: first, the variance of the stock’s return
(given by 𝜎2 T) decreases, and second, the present value of the strike (which is the
cost of exercising) increases, and both changes work in the same direction to make the
call less valuable. The put’s theta has an ambiguous sign. Here the first effect is similar
to that of a call. However, as time passes, the present value of the strike price the put
holder pays increases, which increases the put’s value. In a given context, it’s hard to
tell which of these two opposing effects dominate.
The sensitivity of the option value to changes in the volatility is called vega (which
is not in the Greek alphabet but has found popular usage nonetheless!):
𝜕c Change in the value of the option
vega ≡ =
𝜕𝜎 Change in the volatility (19.10d)
= S√TN′ (d1 ) > 0
As the volatility increases, the value of the call increases. The same is true for the
put option. Increased volatility is a good thing for the option’s value! The larger the
volatility, the more spread out the stock price distribution is and the larger is the tail
THE GREEKS 437
probability that the option ends up in-the-money. It’s no wonder that most traders
view options as “volatility bets”!
TABLE 19.2: Deltas and Gammas for
Calls and Puts
Stock Price, S Call deltaC Put deltaP Gamma
$50 0.0000 −1.0000 0.0000
$55 0.0000 −1.0000 0.0000
$60 0.0001 −0.9999 0.0000
$65 0.0005 −0.9995 0.0002
$70 0.0030 −0.9970 0.0009
$75 0.0117 −0.9883 0.0029
$80 0.0349 −0.9651 0.0068
$85 0.0826 −0.9174 0.0126
$90 0.1620 −0.8380 0.0191
$95 0.2720 −0.7280 0.0245
$100 0.4025 −0.5975 0.0272
$105 0.5381 −0.4619 0.0265
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
$110 0.6636 −0.3364 0.0233
$115 0.7686 −0.2314 0.0186
$120 0.8492 −0.1508 0.0137
$125 0.9065 −0.0935 0.0094
$130 0.9446 −0.0554 0.006
$135 0.9685 −0.0315 0.0037
$140 0.9828 −0.0172 0.0021
$145 0.9909 −0.0091 0.0012
$150 0.9953 −0.0047 0.0006
438 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
FIGURE 19.3A: Deltas for Calls and Puts
Deltas
1.00
Call delta
0.50
0.00
–0.50
–1.00
Put delta
50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150
Stock Prices
FIGURE 19.3B: Gamma for a Call (Same as That for
the Put)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Gamma 0.03
0.025
0.02
0.015
0.01
0.005
0
50
55
60
65
70
75
80
85
90
95
100
105
110
115
120
125
130
135
140
145
150
Stock Prices
THE GREEKS 439
Last, the sensitivity of the option value to changes in the interest rate is called rho:
𝜕c Change in the value of the option
rho ≡ =
𝜕r Change in the value of the interest rate (19.10e)
= TKe−rT N (d2 ) ≥ 0
As the interest rate increases, the call’s value increases. This is due to the decrease in the
present value of the strike price paid at time T if the option ends up in-the-money.
But a European put’s rho is negative. This happens because a rise in the interest
rate lowers the present value of the exercise price that the holder might receive at
expiration, and this reduces the put’s value.
Some Road Bumps Ahead
The Greeks are useful for understanding the sensitivity of option values to changes
in the inputs. Sometimes the Greeks are also used for hedging option price risks
caused by changes in these parameters. A warning needs to be raised at this time!
We will show in Chapter 20 that delta and gamma hedging are valid procedures
when the BSM assumptions are satisfied (Extension 19.2 shows what happens when
the competitive markets assumption is violated and market manipulation is possible).
However, even if the BSM assumptions are valid, rho hedging (which has been largely
abandoned by the practitioners) and vega hedging (which should also disappear but
exists today in many institutions) have serious conceptual problems. Let us drop a
hint. We held volatility and interest rates constant throughout our derivation of the
BSM formula, so the risk of their changing was not included in the price. This is
in contrast to the stock price, whose risk was explicitly considered. Now, all of a
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
sudden, we cannot start tinkering with something that we kept fixed. We will return
to these issues in Chapter 20.
EXTENSION 19.2: Market Manipulation and Option Pricing
In the derivation of the BSM model, assumption A3, the competitive and well-functioning markets assumption is
crucial. Extension 19.1 already discussed how option pricing changes when stock price bubbles exist that violate
the well-functioning assumption. This extension discusses the difficulties in pricing options when the competitive
markets assumption is relaxed.
A competitive market is one in which traders’ trades do not affect the stock price. When this assumption
is relaxed and traders’ trades change the price, stock market manipulation is possible. In a market that can be
manipulated, the BSM model is no longer valid. In fact, the BSM value and delta can be completely wrong. We
illustrate such a dramatic failure of the BSM model with a simple example.
Let us consider a European call option with strike price K and maturity T written on a stock with time t
price S(t), and volatility 𝜎.
Suppose that the stock price has evolved across time, as graphed in Ext. 19.2 Fig. 1, and that currently we are
standing at time T−𝜀 for a small 𝜀 > 0, where the stock price is out-of-the-money, that is,
S (T − 𝜀) < K
440 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
Using the BSM formula, it can easily be shown that the value of the call option and its delta are approximately
zero, that is,
c ∼ 0 and delta ∼ 0 (1)
This is because the stock is trading out-of-the-money, and with so little time remaining before the option expires,
it is unlikely to move much above the strike price.
Now suppose that trades can change the stock price. This is the violation of the competitive market assumption.
Furthermore, suppose that a manipulator purchases a large quantity of the stock’s shares purposely to cause the
stock price to jump to S* > K. In light of this manipulation, the true price of the call option and its delta are
easily seen to be
c = S∗ − K and delta = 1 (2)
As indicated, these values are dramatically different from those predicted by the BSM model. If a risk manager had
used the BSM model to hedge an option position using a delta value near zero (see Chapter 20 for an elaboration
on delta hedging), he would have lost significant wealth because of the market manipulator’s actions.
EXT. 19.2 FIG. 1: Manipulated Stock
Price
S(t)
S*
K
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Stock price
0
T–ε T Time
this example illustrates, the BSM model no longer applies when the competitive markets assumption is relaxed.
Pricing options in a world in which the stock price can be manipulated is a fruitful area for research (see Bank
and Baum 2004; Jarrow 1994).
19.8 The Inputs
Most of the inputs for the BSM formula are easily obtained. The most elusive input
to measure is the volatility. Because of its importance, volatility estimation is discussed
in this section and also in the next chapter. This chapter concentrates on historical
volatility, while Chapter 20 focuses on calibration and implied volatilities.
THE INPUTS 441
Observable Inputs
The strike price (K) and time to maturity (T) are specified in the option contract.
The stock price (S) is easy to obtain. For actively traded stocks, use the last transaction
price, and for less liquid stocks, use the average of the bid and ask prices.
The continuously compounded risk-free interest rate (r) is computable from
Treasury security prices. Select the Treasury bill that matures closest to the option’s
expiration date, take the average of the bid and ask yields, and compute the Treasury
bill’s price. Last, solve for the continuously compounded annual interest rate from
this zero-coupon bond price (see Chapter 2). This is the interest rate to use in the
BSM formula.
Volatility: The Elusive Input
Volatility is the hardest input to estimate. There are two standard methods for
estimating the volatility: an explicit method using past stock prices, called the historical
volatility, and an implicit method using option market prices, called the implied
volatility.
HISTORICAL VOLATILITY The volatility when estimated using statistical
methods and past price data is called a historical volatility. An estimate of the
volatility is the stock return’s sample standard deviation, although numerous other
estimators can also be used. Historical volatility estimation, using actual stock price
realizations, is based on nature’s “draws” from the actual probability distribution
Equation 19.1. This implies, of course, that the estimates of both the stock’s expected
return 𝜇 and volatility 𝜎 are obtained under the actual probability q. Example 19.3
illustrates how to estimate the stock’s volatility using the return’s sample standard
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
deviation.
EXAMPLE 19.3: Computing Historical Volatility
■ Stock price data are available from newspapers or the Internet.
■ The data. Table 19.3 reports closing stock prices for Your Beloved Machines Corp. (YBM). Column
1 is labeled Week t. There are (T + 1) = 17 observations corresponding to 17 consecutive weeks,
where t runs from week 0 to week 16. Column 2 records weekly closing prices for YBM. Seventeen
observations are chosen to illustrate the computation. In practice, a larger number of observations are
recommended. This issue is discussed after the example.
■ The price relative. The price relatives of the stock prices are reported in column 3. This is given by
S(t)/S(t−1), where S(t) is week t’s closing price. For example, when t = 6, the price relative is
S (6) /S (5) = 103/100 = 1.03
Notice that when computing the “difference,” the first observation is lost. We now have n =16 data
points, which we denote by t = 1, 2, … ,16.
442 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
TABLE 19.3: Computing 𝜎 from Weekly Price Data
16
Week (t) Stock Price S(t) Price Log of Price Relative [x(t) − 𝜇]̂ 2
S(t) ∑
Relative S (t) t =1
S(t − 1) log ≡ x(t)
[ S (t − 1) ]
0 100
1 102 1.02 0.019802627 0.000344658
2 99 0.970588235 −0.029852963 0.000966627
3 100 1.01010101 0.010050336 7.76632E-05
4 101 1.01 0.009950331 7.59106E-05
5 100 0.99009901 −0.009950331 0.000125171
6 103 1.03 0.029558802 0.000802087
7 104 1.009708738 0.009661911 7.09679E-05
8 101 0.971153846 −0.029270382 0.000930741
9 101 1 0 1.53181E-06
10 100 0.99009901 −0.009950331 0.000125171
11 101 1.01 0.009950331 7.59106E-05
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
12 100 0.99009901 −0.009950331 0.000125171
13 102 1.02 0.019802627 0.000344658
14 102 1 0 1.53181E-06
15 99 0.970588235 −0.029852963 0.000966627
16 102 1.03030303 0.029852963 0.000818835
Sums Σx = 0.019802627 0.005853263
Mean Σx/16 = 0.001237664
■ The logarithm of the price relative. The natural logarithm of the price relative, log[S(t)/S(t−1)], is
computed and recorded in the fourth column. For example, we have
log [S (6) /S (5)] = log (1.03) = 0.029558802
For simplicity, write this as log[S(t)/S(t−1)] = x(t).
THE INPUTS 443
■ Computing the sample variance. One can proceed in several ways:
- If the data are input into a spreadsheet or a computer program, the sample variance is computed
in one step. For example, running Excel command VARA over the sixteen observations in cells
D4 = 0.019802627 to D19 = 0.029852963 gives VARA(D4:D19) = 0.000390218.
- Alternatively, one can plod through the calculations by hand, first computing the sample mean,
T
1
𝜇̂ = x (t) = 0.019802627/16 = 0.001237664
T t∑
=1
and then reporting in column 5 the deviations of the x(t)’s values from the mean, squared, and then
computing the average of these squared deviations using (T − 1) = 15 in the denominator to get
the sample variance:
T
1
𝜎̂ 2 = [x(t) − 𝜇]̂ 2 = 0.005853263/15 = 0.00390218
T − 1 t∑
=1
■ Computing the annualized volatility. To convert the computed sample variance to the annualized variance,
because the variance is proportional to the time period (see Result 1 in the appendix to this chapter),
we need only multiply by the number of weeks in a year, that is,
𝜎̂ 2Annual = 𝜎̂ 2Week (Number of weeks) = 0.00390218 × 52 = 0.02029131
The positive square root is the annualized volatility 𝜎̂ = 0.14245.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Converting to annualized figures. Suppose you compute the volatility using daily data instead. How
would you annualize this? Although people initially used 365 days, empirical studies suggest that a
better volatility estimate is obtained if one uses the number of trading days instead (see French 1980;
French and Roll 1986).3 Assuming there are 250 trading days in a year, one gets
𝜎̂ 2Annual = 𝜎̂ 2Daily (250)
3 If you compute sigma using monthly data (one stock price data per month), then use the square root of 12 as the adjustment
factor to compute annualized volatility.
In computing the sample variance, a number of choices need to be made. First,
how many observations should you use? Theoretically speaking, more observations
will give a better estimate, everything else held constant. But everything else is
almost never constant! A long time series of prices may not be stationary because
the underlying structure of the company may change, and consequently, so, too, the
stock’s volatility. A good rule of thumb is to get a minimum of 50 observations (where
the central limit theorem from statistics kicks in, which enables the use of the normal
distribution for hypothesis testing). Second, what should the time interval be between
444 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
stock price observations? Shorter intervals imply more price observations, but at short
intervals, market frictions and market microstructure issues can introduce significant
noise into the estimation. Daily or weekly intervals have less of these microstructure
problems. A good rule of thumb is to use daily or weekly data.
IMPLIED VOLATILITY The volatility estimated using an OPM in conjunction
with contemporaneous option market prices, instead of past stock prices, is called an
implied volatility. This section explains how to obtain such an estimate using the
BSM formula.
For the purposes of this estimate, we can rewrite the BSM call price as BSM(𝜎),
where 𝜎 is the volatility. The implied volatility 𝜎Imp is defined as the volatility that solves
the following equation:
cMarket = BSM (𝜎Imp ) (19.11)
In words, the implied volatility is that volatility that equates the market price for an
option equal to its BSM value.
Thus, for the implied volatility, we use (S, K, r, T) and cMarket as inputs and then
compute the volatility 𝜎 by solving this equation. Technically, since the BSM formula
is an increasing function of the volatility, a solution always exists. Unfortunately,
no closed form solution is available. However, using numerical approximation
techniques, the option’s implied volatility 𝜎Imp can be easily determined.
This estimation method is an example of calibration, in which one estimates
a parameter from a model by using market prices as inputs rather than outputs.
Calibration and implied volatilities are often misused, so care should be exercised
when employed. Chapter 20 discusses the use and misuse of calibration and implied
volatilities in detail.4
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
19.9 Extending the Black–Scholes–Merton
Model
This section extends the BSM model in three key directions: one is to dividends, two
is to foreign currency options, and three is to American options. All these extensions
are important applications of the BSM model.
Adjusting for Dividends
The assumption that the underlying stock pays no dividends underlies the BSM
model. But most stocks pay dividends, which lowers the price of calls and raises the
price of puts. Fortunately, one can adjust the basic model for dividends. We present
two such approaches. The first adjusts the BSM formula for a known dollar dividend,
whereas the second allows the stock to pay random dividends at a continuous rate.
Here the discussion of dividend adjustments presented in the cost-of-carry model
(Chapter 12) and the multiperiod binomial model (Chapter 18) are relevant.
4
Note that because implied volatility estimation uses market prices and the BSM formula, this estimation
is done under the pseudo-probabilities 𝜋 and does not depend on the actual probabilities.
EXTENDING THE BLACK–SCHOLES–MERTON MODEL 445
When the dividend date and dollar dividend are known, replace the actual stock price
with the current stock price minus the present value of all the dividends paid over
the option’s life. Then continue with business as usual: employ the BSM model with
this modified stock price. Thus Equation 19.4a gets modified to
c = [S − PV (Dividends)] N (d1 ) − Ke−rT N (d2 ) (19.12)
where
S − PV (Dividends) 𝜎2
log + r+ T
[ K ] ( 2)
d1 =
𝜎√T
d2 = d1 − 𝜎√T
When the dividend rate is known, again the stock price is replaced by the current
stock price less the present value of the dividends paid over the option’s life, but here
there is a fractional reduction (e−𝛿T ), that is,
c = Se−𝛿T N (d1 ) − Ke−rT N (d2 ) (19.13a)
p = Ke−rT N (−d2 ) − Se−𝛿T N (−d1 ) (19.13b)
where 𝛿 is the dividend yield:
S 𝜎2
log ( ) + r − 𝛿 + T
K ( 2)
d1 =
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
𝜎√T
d2 = d1 − 𝜎√T
The dollar dividend is random in this case because the random stock price is
multiplied by a constant dividend rate, yielding a random dollar dividend payment.
This model, originally derived by Robert Merton (1973a), is particularly useful for
valuing index and currency options for the reasons we had discussed before (see
Chapter 12). In the special case when 𝛿 = 0, this model reduces to the BSM model.
Example 19.4 illustrates their use.
EXAMPLE 19.4: Pricing a European Option on an Index
The Data
■ Consider the data from Chapter 12:
- INDY index’s current level I = 1,000.
- The continuously compounded interest rate r is 6 percent per year.
446 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
- The dividend yield 𝛿 = 20/1,000 = 0.02 or 2 percent per year because the stocks constituting INDY
paid $20 in dividends last year.
The European Call Price
■ Consider a European call on INDY with maturity T = 1 year and a strike price of K = 1,100. Let
the index’s return volatility 𝜎 be 0.142470 per year. The payoff to the call on the maturity date is
100[S(T) −1,100] dollars if it ends in-the-money, and is zero otherwise, where 100 is the contract
multiplier. This call is valued using Equation 19.13a, as follows:
- Compute d1 and d2 :
S 𝜎2
log ( ) + r − 𝛿 + T
K ( 2)
d1 =
𝜎√T
1, 000 (0.14247)2
log + 0.06 − 0.02 + (1)
( 1, 100 ) [ 2 ]
=
0.14247√1
= −0.3170
d2 = d1 − 𝜎√T = − 0.3170 − 0.14247 = −0.4595
- Using the NORMSDIST function of Excel, we get the cumulative standard normal distribution
function values
N (d1 ) = N (−0.3170) = 0.3756
N (d2 ) = N (−0.4595) = 0.3230
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- The zero-coupon bond price is B = e−rT = e−0.06 = 0.9418.
- The present value of the dividend yield e−𝛿T = e−0.02 = 0.9802.
- The European call price for the index option is therefore
c = e−𝛿T SN (d1 ) − Ke−rT N (d2 )
= 0.9802 × 1, 000 × 0.3756 − 1, 100 × 0.9418 × 0.3223
= $33.6286
- Using the contract multiplier 100, the European call’s price is $3,362.86.
Foreign Currency Options
Pricing foreign currency options using the BSM formula is just a clever use of the
BSM model with a dividend yield. We illustrate this application with an example (see
Example 19.5).
EXTENDING THE BLACK–SCHOLES–MERTON MODEL 447
EXAMPLE 19.5: Pricing a European Option on a Currency
The Data
■ Consider the data from Chapter 12:
- Let today’s spot exchange rate SA be $2 per pound (in American terms).
- Let the continuously compounded annual risk-free interest rates be r = 6 percent in the United
States (domestic) and rE = 4 percent in the United Kingdom (foreign). The UK rate is interpreted
as the “dividend yield” 𝛿 in this context.
The European Call Price
■ Consider a European call on pound sterling with maturity T = 1 year and a strike price of
K = $2.10. Let the pound’s volatility 𝜎 be 0.10 per year. The payoff to the call on the maturity date
is 100[S(T)−2.10] dollars if it ends in-the-money, and zero if otherwise, where 100 is the contract
multiplier. This call is valued using Equation 19.13a:
- Compute d1 and d2 :
2 (0.1)2
log ( + 0.06 − 0.04 + (1)
2.1 ) [ 2 ]
d1 = = − 0.2379
0.1√1
d2 = d1 − 𝜎√T = − 0.2379 − 0.1 = − 0.3379
- Moreover, we get
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
N (d1 ) = N (−0.2379) = 0.4060
N (d2 ) = N (−0.3379) = 0.3677
The zero-coupon bond price (US) is B = e−rT = e−0.06 = 0.9418. The zero-coupon bond price
(foreign) BE = e −rE T = e−0.04 = 0.9608.
- The European call price for the currency option is therefore
c = e −rE T SN (d1 ) − Ke−rT N (d2 )
= 0.9608 × 2 × 0.4060 − 2.1 × 0.9418 × 0.3677
= $0.0529
Using the contract multiplier 100, the European currency call’s price is $5.29.
Valuing American Options
The BSM formula and its dividend modifications only hold for European options.
Yet, American options trade. We know from Chapter 16 that when there are
dividends, American options are more valuable than European options because of an
early exercise premium. Unfortunately, we can only obtain formulas for American call
448 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
options in some very special cases, and there are no closed-form solutions for pricing
American put options. This implies that to price American options in practice, we
need to turn to numerical procedures. Numerical procedures compute values using
the computer instead of plugging into an analytic formula.
You have already been introduced to numerical procedures in Chapter 18 when
pricing American calls and puts using the binomial model. Because the multiperiod
binomial model can be used to approximate a lognormally distributed stock price, the
multiperiod binomial model provides an important tool for approximating American
call and put values in the BSM model as well. This approach works well but can be
computationally time consuming.
Other numerical procedures for pricing American call and put options are available
as well. Numerical approximation techniques can be divided into two categories:
(1) methods that numerically solve the partial differential equation (pde) implied
by the option’s price (see Chapter 20 for a discussion of the BSM pde) and (2)
those that approximate the stock price evolution to compute the option’s risk-
neutral value (a discounted expectation). To solve pdes, numerical techniques like
finite difference methods and numerical integration can be employed. To compute
discounted expectations, the binomial model and Monte Carlo simulation techniques
are useful. Unfortunately, both methods are beyond the scope of this textbook and
are left for more advanced courses (see Glasserman [2003] for Monte Carlo methods,
Wilmott [1998] for pdes, and Duffie [2001] for a discussion of both).
EXTENSION 19.3: Exotic Options
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Exotic options are derivative securities that differ from the simple, plain vanilla forwards, futures, and European
and American call and put options that we have encountered so far. There are many exotic options with names like
“digital options,” “gap options,” “options on the maximum or minimum,” “down-and-out options,” “down-and-
in options,” and “average options.” The defining characteristic of any such exotic option is its payoff at maturity
or at an intermediate time, conditional on certain events related to the path of the underlying stock price being
satisfied. The easiest method of pricing such options is to use risk-neutral valuation and compute the discounted
expected payoffs to these options under the pseudo-probabilities. We illustrate this general approach with two
examples under the BSM assumptions.
Digital options
A digital put option has a maturity date T and a strike price K. Its payoff at maturity is
1 if S (T) < K
digital (T) =
{0 if S (T) ≥ K
This payoff is similar to an ordinary European put, except in its payoff, which equals just $1 if the option ends
up in-the-money.
SUMMARY 449
Using risk-neutral valuation, the time 0 value of a digital put option is
digital (0) = E𝜋 {digital (T)} e−rT = e−rT N (d − 𝜎√T)
where
S
log ( ) − rT
K 𝜎
d= + √T
𝜎√T 2
Barrier options
Barrier options are similar to European options, except the payoffs depend on whether the stock price hits some
barrier over the option’s life. We illustrate one such option.
A down-and-out call option has a maturity date T and a strike price K. Its payoff at maturity is that of an
ordinary European call option, but only if a lower barrier (b) is not hit over the option’s life. If the lower barrier
is hit, the option expires worthless, regardless of its value at maturity. The payoff can be written as follows:
down&out (T) = 1{𝜏 > T} max [S (T) − K, 0]
where
1 if min S (t) > b for all t ∈ [0, T]
1{𝜏>T} =
{0 otherwise
Using risk-neutral valuation, the time 0 value of a down-and-out call option is
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
down&out (0) = E𝜋 {down&out (T)} e − rT
A complex analytic formula for this equation exists; see Jarrow and Turnbull (2000).
19.10 Summary
1. The year 1973 saw the publication of the Black–Scholes–Merton option pricing
model. The BSM model forms the foundation of more advanced derivatives
pricing models.
2. As with the binomial model, the BSM model assumes no market frictions, no
credit risk, competitive and well-functioning markets, no interest rate uncertainty,
no intermediate cash flows, and no arbitrage. Here, however, we introduce a
different assumption for the stock price evolution. The BSM model assumes
continuous trading and that the stock price returns follow a lognormal distribution.
450 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
3. The BSM formula for pricing a European call option is
c = SN (d1 ) − Ke−rT N (d2 )
where S ≡ S(0) is today’s spot price, N( . ) is the cumulative standard normal
distribution function which has mean 0 and standard deviation 1,
S 𝜎2
log ( ) + r + T
K ( 2)
d1 =
𝜎√T
d2 = d1 − 𝜎√T
K is the strike price, r is the continuously compounded risk-free interest rate per
year, T is the time to maturity measured in years, and 𝜎 is stock return’s volatility
per year.
4. As in Chapter 17, it can be shown that the stock price, normalized by the price of
a money market account, is a martingale under the pseudo-probabilities. This is
an implication of our no arbitrage assumption. Under the lognormal distribution
assumption, the market is complete, and risk-neutral valuation holds using the
pseudo-probabilities and gives the call price as
c = E𝜋 {max [0, S (T) − K]} e−rT
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
5. To understand the BSM formula, we compute the option sensitivities (popularly
known as the Greeks) by taking partial derivatives with respect to the stock price
(S) to get delta, the time to maturity (T) to get theta, the riskless interest rate (r)
to get rho, and the stock return’s volatility (𝜎) to get vega. In addition, the second
partial derivative with respect to the stock price is the gamma. The Greeks are
useful for options hedging and portfolio risk management, a topic we address in
Chapter 20.
6. Volatility is the hardest input to estimate. There are two standard methods for
estimating the volatility: an explicit method using past stock prices, called the
historical volatility, and an implicit method using option market prices, called the
implied volatility. The implied volatility is discussed in Chapter 20.
7. Three common extensions of the BSM model are to adjust for dividends, to price
foreign currency options, and to price American options. American options don’t
have analytical solutions. They can be solved by numerical methods belonging to
one of three types: (1) a multiperiod binomial option pricing model, (2) partial
differential equations, and (3) Monte Carlo simulation techniques.
APPENDIX 451
19.11 Appendix
Modeling the Stock Price Evolution
This appendix justifies the widely used lognormal assumption for stock prices using
fundamental axioms for stock price changes. We build this argument in several
steps: we (1) divide the time horizon into shorter periods over which stock prices
follow a continuously compounded (logarithmic) return, (2) introduce uncertainty
via additional assumptions on these returns, and (3) apply the central limit theorem
to show that stock prices are lognormally distributed.
Continuously Compounded (Logarithmic) Return
We introduce a discrete time multiperiod structure like the one introduced in
Chapter 18. We do this by chopping up the time horizon [0, T] into n shorter
time periods of length T/n = ∆t each (see Figure 1). Let us assume that over any time
interval [t −∆t, t], the stock price has a continuously compounded return zt that is
inspired by Equation 19.1:
S (t) = S (t − Δt) ezt (1)
The stock price grows by continuous compounding. Setting t = ∆t, 2∆t,.. given
relations like
S (Δt) = S (0) ez1 over [0, Δt] ,
S (2Δt) = S (Δt) ez2 over [Δt, 2Δt] , ...
S (T) = S (T − Δt) ezT over [T − Δt, T]
We can use the preceding relations to write the stock price at the end of the time
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
horizon as
S (T) S (T − Δt) S (Δt)
S (T) = … S (0)
[ S (T − Δt) S (T − 2Δt) S (0) ]
(2)
= ezT ezT−1 … ez1 S (0)
= S (0) ez1 +…+zT−1 +zT
Now, if we define the sum of these continuously compounded stock price returns
over the short intervals as Z(T), that is,
z1 + ... + zT−1 + zT ≡ Z (T) (3a)
then we can simply rewrite (2) as
Z (T) = log [S (T) /S (0)] (3b)
Thus Z(T) corresponds to the continuously compounded (or logarithmic) return
over the time horizon [0, T]. Notice that (3a) is a linear relationship that has a number
of attractive features and makes the model amenable to mathematical analysis.
452 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
FIGURE 1: Modeling the Stock Price Evolution
Stock price
0
∆t 2∆t T – ∆t T Time
This graph divides the time horizon [0, T] into n intervals of length ∆t each.
Introducing Uncertainty
Next, we introduce uncertainty in our model. We do this by adding two assumptions,
which are often-used statistical properties for a collection of random variables.
A1. The returns {zt } ≡ {z1 , z2 , … , zT } are independent. Early empirical
studies suggested that stock price returns over successive intervals are independent of
each other.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
A2. The returns {zt } are identically distributed. Early empirical studies also
suggested that stock price returns over successive intervals tend to follow the same
distribution.
Notice that we have not specified any probability distribution for the continuously
compounded stock price returns. These two assumptions yield the property that is
often referred to by the abbreviation iid (for independent and identically distributed).
Any process satisfying these assumptions is called a random walk. Stock prices
following a random walk are linked with the efficient markets hypothesis introduced
in Chapter 6 because random walks occur when information comes in randomly and
gets immediately reflected in stock prices.
Next, we introduce two more assumptions to ensure that stock prices behave as the
time interval shrinks and approaches zero. We do this by assuming that the expected
return and the variance of the return are proportional to the length of the time
interval.
A3. The expected continuously compounded return can be written as
E (zt ) = 𝜇Δt (4)
where 𝜇 is a constant that corresponds to the expected continuously compounded
return per year.
APPENDIX 453
A4. The variance of the continuously compounded return can be written as
var (zt ) = 𝜎2 Δt (5)
where 𝜎2 is a constant that corresponds to the variance of the continuously
compounded return per year.
Assumptions A1−A4 allows us to write
E [Z (T)] = E (z1 + … + zT−1 + zT )
= 𝜇 (Δt + Δt + ... + Δt)
(6a)
[because, by A3, the returns have identical mean 𝜇]
= 𝜇T
and
var [Z (T)] = var (z1 + ... + zT–1 + zT )
= var (z1 ) + var (z2 ) + ... + var (zT ) [by assumption A1]
n (6b)
= 𝜎2 Δ [by assumption A4]
∑
j=1
= 𝜎2 T
We used expression (6b) earlier in the chapter when estimating the stock’s historical
volatility.
The Central Limit Theorem
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Next, we apply the central limit theorem (CLT). It states that the average of
iid random variables (provided it has a well-defined mean and variance), when
“standardized” (by subtracting from the average the common mean and dividing the
expression by the standard deviation of the average), starts behaving like a standard
normal distribution (see Mood et al. 1974; Cox and Miller 1990). An application of
this theorem yields the following result (see Result 1).
RESULT 1
Lognormal Distribution of Stock Price Returns
Stock prices are lognormally distributed with mean 𝜇T and variance 𝜎2 T over
the time horizon [0, T].
454 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
Hence we can write the continuously compounded return as
S (T) 𝜎2
Z (T) = log = 𝜇− T + 𝜎√Tz (7)
( S (0) ) ( 2)
where z is a standard normal random variable with mean 0 and variance 1. Because z
is normally distributed, Z(T) is normally distributed, and hence S(T) is said to have
a lognormal distribution. This result is reported as Equation 19.2 in the text, and it
is used in Example 19.3.
First Derivation of the Black–Scholes–Merton
Formula (as a Limit of the Binomial Model)
This appendix provides a formal derivation of the BSM formula based on the intuition
obtained from Figure 19.1, the second graph of the multiperiod binomial model. In
this figure, you can see that as the time interval becomes smaller, the multiperiod
model appears more and more realistic. In fact, it can be shown that it approaches
the lognormal distribution! Not surprisingly, this gives us a way to derive the BSM
formula.
MULTIPERIOD BINOMIAL PARAMETERS FOR THE LOGNORMAL
APPROXIMATION
■ To formalize this intuition, we start with assumption A7 of Chapter 17 as applied
to the multiperiod binomial model. We divide our time interval [0, T] into
subintervals of length ∆t, yielding T/∆t = n intervals.
■ Next, we use the Jarrow–Rudd specification of the parameters
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
𝜎2 𝜎2
rΔt − Δt + 𝜎√Δt rΔt − Δt − 𝜎√Δt
U=e 2 and D = e 2 (8a)
given in Extension 18.1. Because the option value does not depend on the actual
probability of moving up, we leave the actual probability unspecified, and work
under the pseudo-probabilities. We choose the pseudo-probability 𝜋 = 1/2. This
choice guarantees that the binomial process for the log return on the stock over
[0, T] has an expectation and volatility equal to rT − (𝜎2 /2)T and 𝜎2 T, respectively.
Using expression (8a), the stock price process is
S (T) = S (0) U j D(n−j) (8b)
■ Taking logarithms of expression (8b) yields
log S (T) = log S (0) + j log U + (n − j) log D (9a)
APPENDIX 455
Next, substitution of U and D into this expression gives
𝜎2
log S (T) = log S (0) + j rΔt − Δt + 𝜎√Δt
( 2 )
2 (9b)
𝜎
+ (n − j) rΔt − Δt − 𝜎√Δt
( 2 )
Substituting ∆t = T/n and simplifying this is transformed to
𝜎√T
log S (T) = log S (0) + 2j
( √n ) (9c)
𝜎2
+ rT − T − √n𝜎√T
( 2 )
Last, rearranging terms gives
S (T) 𝜎2 n
log − rT − T
s (0) ( 2 ) (j − 2 )
= (9d)
𝜎√T √n
2
TAKING LIMITS
■ The famous de Moivre–Laplace theorem states that a binomial random variable
approaches a normally distributed random variable as the number of realizations
(time intervals) approaches infinity, that is,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
⎛ n ⎞
⎜j − ⎟
Prob ⎜⎜ 2 ≤ x⎟ → N (x) as n → ∞ (10a)
⎟
⎜ √n ⎟
⎝ ⎠
2
Hence, by substitution of expression (9d), we see that the logarithm of the assumed
stock price process approaches a normally distributed random variable, that is,
⎛ S (T) 𝜎2 ⎞
⎜ log S (0) − (rT − 2 T) ⎟
Prob ⎜⎜ ≤ x⎟⎟ → N (x) as n → ∞ (10b)
⎜ 𝜎√T ⎟
⎝ ⎠
In simpler notation, this implies that as n → ∞,
𝜎2
r− T+𝜎√Tz
( 2 )
S (T) = S (0) e (11)
where z is a standard normal random variable with mean zero and unit variance
under the pseudo-probability 𝜋(z). This is our lognormally distributed stock price!
456 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
MARTINGALE PRICING
■ The normal distribution has the property that
𝜎2
𝜋 𝜎√Tz T
E [e ]=e2 (12a)
Taking the expectation of expression (11) yields
𝜎2
r− T
( 2 )
E𝜋 [S (T)] = S (0) e E𝜋 [e
𝜎√Tz
] (12b)
The use of expression (12a) in (12b) yields, after rearranging terms, the martingale
pricing relation expression (19.6) given earlier in the text, that is,
S (0) = E𝜋 [S (T)] e−rT (13)
where E𝜋 ( . ) is expectation using the pseudo-probabilities 𝜋(z) associated with
the lognormal distribution; that is, the current stock price S(0) is the expected
discounted value of the stock price at time T. This is analogous to expression
(17.5).
RISK-NEUTRAL VALUATION
■ Next, using the multiperiod binomial call option formula from Chapter 18,
Equation 18.11, with the Jarrow–Rudd parameters, we have
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
n
n j
cn = e−rT 𝜋 (1 − 𝜋)n−j × [max (0, Uj Dn−j S (0) − K)] (14)
∑ {[( j ) ] }
j=0
■ Using expression (8b) again, this can also be written as (see [18.12])
cn = E𝜋 {max [0, S (T) − K]} e−rT (15)
This is called risk-neutral valuation.
TAKING LIMITS AGAIN
■ Taking the limit as n → ∞, one gets
cn → E 𝜋 {max [0, S (T) − K]} e−rT (16)
where 𝜋(z) are the pseudo-probabilities associated with the random variable z as in
expression (11). This follows from interchanging the limit and expectation operations
(see Billingsley 1995, p. 291).
The next section shows that this yields the BSM formula.
APPENDIX 457
Second Derivation of the Black–Scholes–Merton
Formula (Using Risk-Neutral Valuation)
Using expression (16) above, we have
c = E𝜋 {max [0, S (T) − K ]} e−rT (17)
EVALUATING THE EXPECTATION Taking conditional expectations, one can
write the call value as the sum of two terms:5
c = E𝜋 {max [S (T) − K, 0] |S (T) ≥ K} Prob𝜋 {S (T) ≥ K} e−rT
(18)
+E𝜋 {max [S (T) − K, 0] |S (T) < K} Prob𝜋 {S (T) < K} e−rT
Recognizing that if S(T) ≥ K, max[0, S(T)−K] = S(T) − K, and if S(T) < K,
max[0, S(T) − K] = 0, we can rewrite the last expression as
c = E𝜋 {S (T) |S (T) ≥ K} e−rT Prob𝜋 {S (T) ≥ K} − KProb𝜋 {S (T) ≥ K} e−rT (19)
One can show using the properties of normal distributions (see Jarrow and Rudd
1982) that
E𝜋 {S (T) |S (T) ≥ K} Prob𝜋 {S (T) ≥ K} e−rT = S (0) N (d1 )
(20)
KProb𝜋 {S (T) ≥ K} e−rT = KN (d2 ) e−rT
Combined, these give the BSM formula. This completes the derivation.
The Probabilities and the Risk Premium
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
In our derivation of the BSM formula, we showed that one could write the stock
price today as its discounted value using the pseudo-probabilities, that is,
S (0) = E𝜋 [S (T)] e−rT (21)
where E𝜋 represents expectation using 𝜋(z).
Recall that the actual and pseudo-probabilities are related by the expression
𝜋 (z) = 𝜙 (z) q (z) (22)
2
where 𝜙(z) = e−(𝜃 /2)+𝜃z
> 0 and 𝜃 = − [(𝜇 − r) /𝜎] √T. Using 𝜙(z), we can rewrite
S (0) = E [𝜙 (z) S (T)] e−rT
∞ (23)
= ∫ 𝜙 (z) S (T) q (z) dz e−rT
[ −∞ ]
5
This follows from the fact that E (X) = ∑Y E (X|Y) Prob (Y), where X and Y are random variables.
458 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
This expression transforms the expectation taken in terms of the pseudo-probabilities
to an expectation taken in terms of the actual probabilities.
There are two ways to compute a present value. One is as given in expression
(21). Here the present value is computed by taking an expectation using the pseudo-
probabilities and discounting at the risk-free rate. The adjustment for risk occurs
in the use of the pseudo-probabilities when taking the expectation. The second and
traditional way of computing a present value is to take the expectation using the actual
probabilities and adjust the discount rate for risk. Expression (23) can be manipulated
to generate this second method of computing a present value. Indeed, as proved
immediately, this expression can be equivalently written as
∞ S(T)
−(rT+cov[log( ,− log 𝜙(z)]
S (0) = ∫ S (T) q (z) dz e S(0) )
[ −∞ ]
(24)
S(T)
−(rT+cov[log( ,− log 𝜙(z)])
= E [s (T)] e S(0) )
where the discount rate is adjusted instead by the covariance term cov(log
[S(T)/S(0)], −log [𝜙(z)]), a risk premium. We see that the risk-adjusted discount
rate depends on 𝜙(z), which provides an alternative confirmation that 𝜙(z) is an
adjustment for risk.
Proof of (24)
This proof just uses simple calculus and knowledge of normal distributions. We start
with ∞
S (0) = ∫ 𝜙 (z) S (T) q (z) dz e−rT (25)
[ −∞ ]
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Plugging in 𝜙(z) and S(T) gives
z2
∞ 𝜃2 𝜎2 −
− + 𝜃z 𝜇T− T + 𝜎√Tz e 2
S (0) = S (0) ∫ e 2 e 2 dz e−rT
−∞ √2𝜋
∞ 1
𝜇T−rT + 𝜎√T𝜃 −(𝜎√T + 𝜃−z) 2 dz
= S (0) e ∫ e 2
−∞ √2𝜋
which we get by completing the square. However,
∞ 1
−(𝜎√T + 𝜃−z) 2 dz
∫ e 2 =1
−∞ √2𝜋
since this is normal with mean (𝜎√T + 𝜃) and variance 1. Hence S (0) =
S (0) e𝜇T−rT+𝜎√T𝜃 , which implies
𝜇T − rT + 𝜎√T𝜃 = 0
or 𝜃 = − [(𝜇 − r) /𝜎] √T
QUESTIONS AND PROBLEMS 459
Last,
S (T)
cov log , − log [𝜙 (z)]
( [ S (0) ] )
𝜎2 𝜃2
= cov 𝜇T − T + 𝜎√Tz, − 𝜃z
[ 2 2 ]
= −𝜎√T𝜃cov (z, z)
= −𝜎√T𝜃
because cov(z, z) = var(z) = 1. Thus
S (T)
𝜇T = rT + cov log , − log [𝜙 (z)]
( [ S (0) ] )
Finally, note that under q(z), from expression (13) above
S (0) = E [S (T)] e−𝜇T
and substitution completes the proof.
19.12 Cases
MW Petroleum Corp. (A and B) (Harvard Business School Cases 295029 and
294050-PDF-ENG). The cases focus on evaluation and execution of a creative
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
financing structure that allows reallocation of oil price risk in the context of an
acquisition.
MoGen Inc. (Darden School of Business Case UV1054-PDF-ENG, Harvard Busi-
ness Publishing). The case shows how to price a convertible bond offering as a
straight bond plus the conversion option.
Sara’s Options (Harvard Business School Case 201005-PDF-ENG). The case stud-
ies the valuation of stock options in pay packages offered to a graduating MBA
student.
19.13 Questions and Problems
19.1. Explain why pricing and hedging go hand in hand in the Black–Scholes–
Merton option pricing framework.
19.2. What does the Black–Scholes–Merton model assume about the evolution of
the stock price, about dividends paid on the stock, and about interest rates?
19.3. Do you think the Black–Scholes–Merton model assumptions of no dividends
paid on the stock and constant interest rates are reasonable assumptions? If
not, why do you think they are imposed?
460 CHAPTER 19: THE BLACK–SCHOLES–MERTON MODEL
19.4. In the assumption of the Black–Scholes–Merton model,
a. what is the stock’s return volatility?
b. what is assumed about the stock’s return volatility? Is this assumption
reasonable or not? If not, why do you think it is used?
The next two questions use the following information for pricing European options.
Special Motors Corporation’s stock price S is $59, the strike price K is $60,
the maturity T is forty-four days, the implied volatility 𝜎 is 30 percent per
year, and the risk-free interest rate r is 3.3 percent per year.
19.5. Using the Black–Scholes–Merton formula, find the price of a call option.
19.6. Using the Black–Scholes–Merton formula, determine the price of a put
option.
19.7. Verify whether the options prices obtained in the previous two questions
satisfy put–call parity.
19.8. Consider two investors who agree on the stock’s price and volatility but who
do not agree on the stock’s expected return. One believes that the stock price
will earn 15 percent over the next year, while the second believes that it will
have a negative 5 percent return. Will they agree or disagree on the value of
a one-year call option on the stock when using the Black–Scholes–Merton
model? Explain your answer.
19.9. Explain how to compute the Black–Scholes–Merton model using the pseudo-
probabilities. What is this principle called?
19.10. What is the difference between the actual probabilities and the pseudo-
probabilities, and on what does it depend? Explain your answer.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
19.11. What are the Greeks in the Black–Scholes–Merton formula? Which two
Greeks correspond to changes in the stock price?
19.12. Comment on the following questions, carefully explaining your answers.
a. As the stock price input increases, everything else constant, does a call
option’s price increase or decrease?
b. As the volatility input increases, everything else constant, does a call
option’s price increase or decrease?
19.13. Given the data for Special Motors Corp. given earlier (i.e., S is $59, K is $60,
T is forty-four days, 𝜎 is 30 percent per year, and r is 3.3 percent per year),
compute the call option’s delta and gamma.
19.14. a. Which inputs to the Black–Scholes–Merton model are observable and
which need to be estimated?
b. Describe some of the difficulties in estimating the unobservable input.
19.15. What is the implied volatility when using the Black–Scholes–Merton model?
Does this estimate depend on the stock’s time series of past stock prices?
Explain your answer.
QUESTIONS AND PROBLEMS 461
19.16. FunToy’s stock has the following weekly closing prices: $40, 41, 43, 42, 42,
46, 43, 44, 47. Assuming fifty-two trading weeks in a year, compute the
annualized, sample standard deviation (volatility) 𝜎.
19.17. Fallstock Inc.’s stock price S is $45. European options on Fallstock have a
strike price K = $50, a maturity T = 180 days, the risk-free interest rate r =
5 percent per year, and the let the volatility 𝜎 = 0.25 per year. The company
is going to pay a dividend of 50 cents after 125 days.
a. Compute the price of a call option using the Black–Scholes–Merton
model.
b. Compute the price of a put option.
19.18. SINDY index is currently at I = 11,057. European options on SINDY have
a strike price K = 11,000, a maturity T = 45 days, a dividend yield 𝛿 = 1.5
percent per year, the risk-free interest rate r = 5 percent per year, and the
estimated implied volatility 𝜎 = 16 percent per year.
a. Compute the price of a call option using the Merton formula.
b. Compute the price of a put option using the appropriate version of put–
call parity (see Chapter 16).
19.19. A European call on euro matures after T = 6 months. The call pays on the
maturity 100[S(T) − 30] dollars if it ends in-the-money, and zero is otherwise,
where 100 is the contract multiplier and $1.30 is the strike price K. Euro’s
volatility 𝜎 is 12 percent per year. Today’s spot exchange rate SA is $1.4 per
euro (in American terms). The continuously compounded annual risk-free
interest rates are r = 5 percent in the United States (domestic) and rE =
4.5 percent in the Eurozone. Compute the price of the call option using the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Merton formula.
19.20. How would you value an American call option in the Black–Scholes– Merton
model?
20
Using the
Black–Scholes–Merton
Model
20.1 Introduction 20.6 The Black–Scholes–Merton
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Model: A Postscript
20.2 Hedging the Greeks
Delta Hedging 20.7 Summary
Gamma Hedging 20.8 Appendix
20.3 Hedging a portfolio of options The Mathematics of Delta,
Gamma, and Vega Hedging
20.4 Vega Hedging
A Third Derivation of the
EXTENSION 20.1 Stochastic Black–Scholes–Merton Formula
Volatility Option Models
20.9 Cases
20.5 Calibration
Theoretical and Econometric
20.10 Questions and Problems
Models Overview
Using Calibration Data (from the Wall Street
Journal, March 3, 2000)
Implied Volatilities
The Project
EXTENSION 20.2 VIX: The Fear
Index
HEDGING THE GREEKS 463
20.1 Introduction
More than a century ago, financiers who sold options for strategic and manipulative
purposes dominated the New York City options market. In contrast, London dealers
acted more like insurance agents employing a “model” to estimate the option’s
premium as the stock’s average price less the strike price plus various costs and profits.
These sellers would also often “delta hedge” their positions. New Yorker Samuel
A. Nelson (1904, p. 14) remarked that the London options market, “where the
business is more of a science” prospered more than its Wall Street counterpart. Isn’t it
interesting that a model helped organize a trader’s thinking, placed trading on a more
scientific footing, and helped the market grow more than a century before our time!
Today, models are actively used in financial markets. In fact, model misuse and
abuse significantly contributed to the 2007 financial crisis. Many financial institutions
misused models and, consequently, underestimated the risk of their positions. These
errors led to insufficient equity capital in the economy, resulting in the failures of
numerous financial institutions. If the financial institutions had better understood the
model subtleties, perhaps this credit crisis could have been avoided. Understanding
the content of this chapter will set you on the path to avoiding such errors.
This chapter explains how to use the Black–Scholes–Merton (BSM) model
to facilitate risk management. We show that both delta and gamma hedging are
consistent with options pricing theory. We also explain why two other once popular
techniques, the now-discarded rho hedging and the still-practiced vega hedging, are
not. Next we revisit the knotty problem of volatility estimation via the computation
of implied volatilities. This is an example of calibration, a popular technique that also
has significant scope for abuse. We end this chapter with a discussion of the BSM
model’s extensions and generalizations.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
20.2 Hedging the Greeks
Hedging the Greeks is a risk management technique used to eliminate market risks from
a stock or options portfolio. Borrowing terms from the title of the classic western
movie The Good, the Bad and the Ugly, we label delta and gamma hedging as “the
good” and faulty practices like vega and rho hedging as “the bad and the ugly.” Both
delta and gamma hedging are “the good” because they are consistent with the theory
(and assumptions) underlying the BSM model. In contrast, vega and rho hedging
are “the bad and the ugly” because they are inconsistent with the theory. Hence, as
argued later, they do not and, in fact, cannot fulfill their desired objectives.
Delta Hedging
The delta (or hedge ratio) is the number of shares of the stock to trade for each
option held to remove all price risk from an option position. For this reason, in
earlier chapters, we characterized the hedge ratio as the holy grail of options pricing.
464 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
To construct a delta-hedged position, we consider a portfolio V (t) at time t
consisting of a long call with price c(t) and n shares of the underlying stock worth
S(t) each. Suppressing the time symbol t for simplicity, the initial value of the
portfolio is
V = c + nS (20.1a)
The change in this portfolio’s value over a small time interval [t, t + ∆t] can be
written as
ΔV = Δc + nΔS (20.1b)
where ∆V ≡ V (t + ∆t) – V (t), ∆c ≡ c(t + ∆t) – c(t), and ∆S ≡ S(t + ∆t) – S(t).
The goal of delta hedging is to choose n such that the variance of the change in the
portfolio’s value is equal to zero, that is, var(∆V ) = 0, where var( . ) denotes variance.
Then, the portfolio will be riskless.
Recognizing that the call’s price is a function of time and the underlying stock
price, using a Taylor series expansion from first-year college calculus, one can solve
for this n (see the appendix to this chapter). The solution is the option’s delta. We
state this observation as a result.
RESULT 20.1
Delta Hedging in the Black–Scholes–Merton Model
To remove price risk from a long call position, short
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
𝜕c
deltaC ≡ = N (d1 ) shares of stock (20.2a)
𝜕S
where
S 𝜎2
log ( ) + r + T
K ( 2)
d1 = (20.2b)
σ√T
N( . ) is the cumulative standard normal distribution function (which has mean
0 and standard deviation 1), S is the stock price, K is the strike price, r is the
continuously compounded risk-free interest rate, T is the time to maturity,
and 𝜎 is the stock return’s volatility (see Result 19.1).
To delta hedge a long put position, buy (–deltaP ) ≡ N(–d1 ) stocks.
We illustrate the effectiveness of this result with a numerical example.
HEDGING THE GREEKS 465
EXAMPLE 20.1: Delta Hedging
■ Suppose we buy a European call option on Your Beloved Machine’s (YBM) stock at time t by paying
a price c. We assume that the stock price S is $100, the strike price Kis $110, the time to maturity T
is 1 year, the volatility 𝜎 of the stock’s return is 0.142470 per year, and the continuously compounded
risk-free interest rate r is 5 percent per year.
■ To hedge one long call position, short deltaC shares of YBM. This gives a delta-neutral portfolio
V = c – (deltaC )S. We compute the values of ∆V = ∆c – (deltaC )∆S for various stock prices in
Table 20.1 to see how the delta hedge works. For simplicity, we assume that the stock price can
magically change without changing any of the other inputs to the model!1
TABLE 20.1: Delta Hedging Errors for Calls and Puts
S d1 DeltaC = DeltaP = Call Put ∆V (Call) ∆V (Put)
N(d1 ) −N(–d1 )
$90 –0.9863 0.1620 0.8380 $1.0294 $15.6646 $1.2709 $1.2709
95 –0.6068 0.2720 0.7280 2.1030 11.7382 0.3319 0.3319
99 –0.3173 0.3755 0.6245 3.3948 9.0300 0.0135 0.0136
100 –0.2468 0.4025 0.5975 3.7838 8.4190 0 0
101 –0.1770 0.4298 0.5702 4.1999 7.8351 0.0136 0.0136
105 0.0957 0.5381 0.4619 6.1367 5.7719 0.3403 0.3403
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
110 0.4222 0.6636 0.3364 9.1477 3.7829 1.3386 1.3386
- Row 4 reports S = $100 in column 1, d1 = −0.2468 in column 2, N(d1 ) = deltaC = 0.4025 in
column 3, and the call price $3.7838 in column 5 (see Example 19.2).
- Column 5 reports call values corresponding to the stock prices recorded in column 1.
- Column 7 reports changes in the hedged portfolio’s value for different stock prices:
ΔV = Δc − (deltaC ) ΔS
= [c (t + Δt) − c (t)] − (deltaC ) [S (t + Δt) − S (t)]
= (New call price − 3.7838) − 0.4025 (New stock price − 100)
For example, for S = $95, we get ∆V = (2.1030 – 3.7838) – 0.4025(95 – 100) = $0.3319.
466 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
■ As seen in Table 20.2, delta hedging works well for small movements in the stock price. Indeed, ∆V is smaller
when the new stock price is close to the original stock price (e. g., when S = 99, ∆V = 0.0135), and
it has larger values when it deviates further (e. g., when S = 110, ∆V = 1.3386).
■ Following the same approach, one can make a long put trade delta-neutral by buying
n = –deltaP = –[–N(–d1 )] = 0.5975 shares. This is reported in row 4 of Table 20.2.
- We get a delta-neutral portfolio V = p + (deltaP )S at time t.
- Using values from column 6 of Table 20.2, the change in value of this portfolio over [t, t + ∆t] is
ΔV = Δp + (deltaP ) ΔS
= (New put price − 8.4190) + 0.5975 (New stock price − 100)
For S = $95, we get ∆V = (11.7382 – 8.4190) + 0.5975(95 – 100) = $0.3319. As before, the hedge
works well for small fluctuations in the stock price.
1This gives a rough-and-ready estimation of hedging performance. For better approximations, you can track ∆V over a longer
period (such as a day) and make allowance for the interest cost V r∆t.
Aren’t these delta-neutral trades akin to the simple 1:2 hedges (which combine a
long stock with two short calls or two long puts) that we discussed in chapter 15?
In terms of actual implementation of the position, delta hedges are minor variants
of such trades. However, the difference lies in the trader’s objective as well as in her
performance. Delta hedging, in theory, creates portfolios that are riskless for small
fluctuations in the underlying stock price, as the previous example illustrates, whereas
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the 1:2 hedges are not riskless.
Gamma Hedging
“In theory, there is no difference between theory and practice. But, in practice, there
is,” mused baseball player Yogi Berra. This quote applies well to delta hedging. If the
world behaved as per the BSM assumptions, then hedging using delta is sufficient
to make an option position riskless. One can continuously rebalance the hedge to
perfectly replicate the option’s payoff. The theory assumes, as noted, that the time
between rebalancing, ∆t, is infinitesimal, and therefore the corresponding changes in
the stock price are small. But continuous trading is impossible. For weekly or even
daily hedging, changes in the stock price are not small, and the hedging error can be
significant. In addition, there are the familiar transaction costs that accompany any
trade, which also makes continuous trading prohibitively expensive.
To address this problem, we need to hedge both small and large changes in the stock
price over the hedging interval. We do this by also hedging the risk from changes in
the “stock price squared.” The second hedge uses the option’s gamma, obtained by
taking the second derivative of the BSM call price with respect to the stock price.
The objective of the second hedge is to anticipate changes in the delta itself (the delta
HEDGING THE GREEKS 467
changes with changes in the underlying stock price) and to compensate for these
changes.
To hedge these two risks, we need two securities so that we can establish a gamma-
hedge in addition to a delta hedge. We can hedge a short call by trading the underlying
stock and another option on the same stock. Consider forming a portfolio V (t)
consisting of one long call with price c1 (t), n1 shares of the stock with price S(t),
and n2 units of the second option with price c2 (t). The value of the portfolio is
V = c1 + n1 S + n2 c2 (20.3a)
The change to this portfolio’s value over the time interval [t, t + ∆t] is
ΔV = Δc1 + n1 ΔS + n2 Δc2 (20.3b)
The objective of delta and gamma hedging is to choose n1 and n2 such that the
variance of the position is zero, that is, var(∆V ) = 0. Using a Taylor series expansion,
the appendix to this chapter shows that the solution is as follows.
RESULT 20.2
Delta and Gamma Hedging a Long Call Position
To remove delta and gamma price risk from a long call with price c1 , hold
n1 shares of stock with price S and n2 shares of another option on the same
stock with price c2 , where
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
n1 = (gammaC1 /gammaC2 ) deltaC2 − deltaC1 (20.4a)
and n2 = −gammaC /gammaC (20.4b)
1 2
The Greeks are defined by the general formulas
S 𝜎2
log ( ) + r + T
𝜕c K ( 2)
deltaC ≡ = N (d1 ) where d1 = (20.4c)
𝜕S σ√T
𝜕2 c 1
and gammaC ≡ = N′ (d1 )
𝜕S 2
Sσ√T
where N ′( . ) is the cumulative standard normal distribution function and
N ′ ( . ) is the standard normal density function.
The next example illustrates the effectiveness of both gamma and delta hedging.
468 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
EXAMPLE 20.2: A Gamma- and Delta-Neutral Portfolio
■ Suppose we buy a European call on YBM at time t for c1 dollars. From the previous example, S =
$100, K = $110, T = 1 year, 𝜎 = 0.142470 per year, and r = 0.05 per year.
■ Let us delta and gamma hedge this position by trading n1 shares of YBM and n2 shares of a second call
also on YBM. Let the second European call’s price be c2 , with a strike price of $105 and a maturity
of nine months.
■ The change in this portfolio’s value over the time interval ∆t is given by ∆V = ∆c1 + n1 ∆S + n2 ∆c2 .We
show the performance of this hedge in Table 20.2.
■ In row 4, columns 2–5 use S = $100, K = $110, T = 1 year, 𝜎 = 0.142470, and r = 0.05 to compute
the relevant values for call 1: (d1 )C1 = – 0.2468 in column 2, deltaC1 = N ′ (d1 )C1 = 0.4025 in column
3, Call1 = $3.7838 in column 4, and gammaC1 = N ′ (d1 )/(S𝜎√T) = 0.0272 in column 5.
■ In row 4, columns 6–9 do the same for call 2 with the strike price 105 and time to maturity 0.75
years: (d1 )C2 = –0.0298 in column 6, deltaC2 = N(d1 )C2 = 0.4881 in column 7, Call2 = $4.4000 in
column 8, and gammaC2 = N ′ (d1 )/(S𝜎√T) = 0.0323 in column 9.
TABLE 20.2: Delta and Gamma Hedging Errors for Calls
S (d1 )C1 DeltaC1 Call1 GammaC1 (d1 )C2 DeltaC2 Call2 GammaC2 ∆V
= N(d1 ) = N(d1 )
$90 –0.9863 0.1620 1.0294 0.0191 –0.8837 0.1884 $1.0857 0.0243 –$0.0092
95 –0.6068 0.2720 2.1030 0.0245 –0.4455 0.3280 2.3630 0.0308 –0.0048
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
99 –0.3173 0.3755 3.3948 0.0269 –0.1113 0.4557 3.9281 0.0325 –0.0001
100 –0.2468 0.4025 3.7838 0.0272 –0.0298 0.4881 4.4000 0.0323 0
101 –0.1770 0.4298 4.1999 0.0273 0.0508 0.5203 4.9043 0.0320 0.0000
105 0.0957 0.5381 6.1367 0.0265 0.3656 0.6427 7.2344 0.0288 0.0117
110 0.4222 0.6636 9.1477 0.0233 0.7427 0.7712 10.7826 0.0223 0.1154
■ Using the formulas in Result 20.2, we compute
n1 = (gammaC /gammaC ) deltaC2 − deltaC1 = 0.007684
1 2
n2 = −gammaC /gammaC = −0.840422
1 2
HEDGING A PORTFOLIO OF OPTIONS 469
■ Column 10 reports changes in the delta- and gamma-hedged portfolio’s value for different stock prices
over a short time interval:
ΔV = Δc1 + n1 ΔS + n2 Δc2
= [c1 (t + Δt) − c1 (t)] + n1 [S (t + Δt) − S (t)] + n2 [c2 (t + Δt) − c2 (t)]
= (New call price1 − 3.7838) + 0.007684 (New stock price − 100)
−0.840422 (New call price2 − 4.4000)
For S = $95, we get ∆V = (2.1030 – 3.7838) + 0.007684(95 – 100) – 0.840422(2.3630 – 4.4000) =
– $0.0048.
■ This is less than the hedging error 0.3310 that we found for a corresponding stock price change in
Example 20.1 where we only delta hedged.
■ A comparison of the last columns in Tables 20.2 and 20.3 indicates the superiority of delta and gamma
hedging over delta hedging alone.
20.3 Hedging a portfolio of options
The magic of finance is that what works for a single security can usually be extended
to a portfolio of securities. Using this magic, we extend the idea of delta and gamma
hedging to a portfolio of options on one or more underlying stocks. Let’s begin with
a simple example.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
EXAMPLE 20.3: Computing the Delta of a Portfolio Consisting of Two
Options on a Single Stock
Suppose that you create a diagonal spread (see Chapter 15) by buying a hundred one-year YBM 110
calls and writing the same number of nine-month YBM 105 calls. When the stock price is $100, we
compute their deltas, as reported in the previous example, where deltaC1 = 0.4025 and deltaC2 = 0.4881,
respectively. The portfolio delta is the number of options times the deltas, then summed, that is,
deltaPort = 100 × 0.4025 − 100 × 0.4881 = −8.5575
The portfolio’s delta implies that the portfolio’s value will decline by approximately $8.56 for a dollar
increase in stock price.
We can generalize this example to develop rules for portfolio risk management.
■ Computing the delta ( and gamma) of an option portfolio on a single stock. Suppose that
you are holding ni options (ni is positive for long holdings, negative for short)
where i = 1, . . . , N denotes European calls or puts on YBM with different
strikes and expiration dates. The delta of this portfolio is just the number of shares
470 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
of each option times the deltas of each of the options summed across all the options:
N
deltaPort = ni × deltai (20.5)
∑
i =1
You can similarly compute the gamma of an option portfolio on a single stock
as the weighted sum of the individual gammas, where the weights are the share
holdings.
■ Delta (and gamma) hedging an option portfolio on multiple stocks. In this case, you need
to delta hedge the option portfolio with respect to each of the different underlying
stocks.
- First, partition the portfolio based on the underlying stocks into subportfolios.
- Next, compute the delta of each subportfolio by using expression (20.5).
- Finally, delta hedge each subportfolio with respect to its underlying stock.
- Analogously, you can extend this line of argument to gamma hedge a portfolio
of options.
20.4 Vega Hedging
When constructing the BSM model, by assumption, there is only one risk: the risk
of stock price changes. Consequently, the OPM reflects the pricing of this risk. No
other risks are present in the model’s formulation.
Delta and gamma hedging using the BSM model are consistent with the theory
underlying the BSM model because these quantities reflect the sensitivity of the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
option’s value to stock price risk. The inability to trade continuously is what makes
gamma hedging useful. All the other inputs to the BSM model, the interest rate (r)
and the volatility (𝜎), are assumed to be constant. The risks of changing interest rates
and volatility are therefore not reflected in the BSM formula. It is nonsensical to
hedge changes in these constant parameters using the Greeks derived from the BSM
formula. A simple example illustrates why this statement is true. After studying the
example, we will apply this insight to vega hedging.
EXAMPLE 20.4: Errors in Hedging a Constant Parameter
■ Consider a European call with strike price K and expiration date T that trades on Your Beloved
Machines (YBM). We will trade YBM stocks to delta hedge a long position in this call option.
A Stock Growing at a Constant Rate
■ Let us perform a thought experiment and introduce an unrealistic assumption. Let’s assume that today’s
stock price S(t) is nonrandom and grows at the risk-free rate r. We agree that this assumption is silly,
but is it any more silly than assuming that the stock’s return has a constant volatility?
VEGA HEDGING 471
■ Continuing, given this assumption, at time t, we know whether the option is going to end in-the-
money. Discounting the call’s known time T payoffs at the risk-free rate gives the arbitrage-free call
price:
c = max [S − Ke−r(T−t) , 0] (20.6)
■ There is no need to hedge our option position because there is no risk.
Delta Hedging Using Expression (20.6)
■ We notice, however, that when applied to market data, the stock price is stochastic and our nonhedge
doesn’t work. The hedging error is the entire change in the option’s price ∆c.
■ As an ad hoc fix, we attempt to hedge the Greek. In this case, the stock price is the relevant parameter,
so we want to delta hedge using expression (20.6).
■ As justification, we employ the Taylor series expansion of the call formula (20.6) around the stock
price over [t, t + ∆t], as discussed previously. We delta hedge by choosing n shares of the stock to short
for each long call, where
𝜕c
n= (20.7)
𝜕S
For this model, expression (20.6), we can compute the option’s delta:
𝜕c 1 if S − Ke−r(T−t) > 0
n= = (20.8)
𝜕S {0 otherwise
■ Here the delta hedge is to hold n shares of the stock as in expression (20.8). This implies that if the
stock is currently in-the-money, n = 1, and if it is out-of-the-money, n = 0.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- If the call is out-of-the-money, then delta hedging using expression (20.8) is equivalent to not
hedging! In this case, hedging the Greek based on expression (20.8) is useless.
- If the call is in-the-money, then delta hedging using expression (20.8) is equivalent to shorting
the entire stock. In this case, let us investigate the improvement, if any, in the hedging based on
expression (20.8) versus doing nothing at all.
Hedging Based on Expression (20.8) Does Not Work
■ To prove that delta hedging based on the formula in expression (20.8) doesn’t work when the call is
in-the-money, we need to know the true random evolution of the stock price.
■ Because we are now familiar with the BSM model, let’s assume that the stock price actually satisfies
the BSM lognormal distribution assumption from Chapter 19. Then, the “true” price of the option
is the BSM value, and we can approximate the change in the true option value using the BSM delta
and a Taylor series expansion, that is,
𝜕c 𝜕c 𝜕c
Δc = Δt + ΔS = Δt + N (d1 ) ΔS
𝜕t 𝜕S 𝜕t
472 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
■ Suppose that we delta hedge using expression (20.8) with one share of the stock. By direct substitution,
we see that change in the value of the delta-hedged portfolio is
𝜕c 𝜕c 𝜕c
ΔV = Δc − ΔS = Δt + ΔS − ΔS = Δt + [N (d1 ) − 1] ΔS (20.9)
𝜕t 𝜕S 𝜕t
If the delta hedge works, then the change in this portfolio’s value should be just (𝜕c/𝜕t)∆t. Why?
Because this term is deterministic, and it would imply that var(∆V ) = 0. In contradiction, expression
(20.9) shows that the hedging error of the hedged portfolio V is [N(d1 ) – 1] ∆S.
■ Now if we didn’t hedge at all, the change in the value of the portfolio would equal the change in the
value of the option, that is,
𝜕c 𝜕c 𝜕c
ΔV = Δc = Δt + ΔS = Δt + N (d1 ) ΔS (20.10)
𝜕t 𝜕S 𝜕t
■ We can now see that delta hedging has really not improved our situation very much!
- The hedging errors are [1 – N(d1 )] ∆S in expression (20.9) versus N(d1 )∆S in expression (20.10).
- For just in-the-money options, N(d1 ) ≈ 0.6. So the delta-hedged portfolio change (expression
[20.9]) and the naked call portfolio (expression [20.10]) are quite close because [1 – N(d1 )] ≈ 0.4
and N(d1 ) ≈ 0.6.
Summary
■ Why does delta hedging work for the BSM model and not for this example? Because in the BSM
model, the stock price is stochastic and its risk is priced within the model. Hedging its Greek is
appropriate. In our example, however, the stock price is nonrandom, so its risk is not priced within
the model. Hedging its Greek, the delta, using this call model doesn’t work.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Now let us turn to vega hedging in the BSM model. Just as in our previous
example, because the BSM model from which the vega is obtained has no volatility
risk included, volatility risk is not priced within the model. The Taylor series
expansion is irrelevant because the volatility is constant (∆𝜎 = 0) in the derivation of
the BSM model. In the BSM model, vega measures how the BSM’s price changes
if the volatility is changed once and forever constant thereafter. This is not the case with
stochastic volatility. Hence vega hedging is meaningless in this context and will not
work. The magnitude of the error in vega hedging, like our example, is understood by
looking at an OPM that relaxes the constant volatility assumption. These are called
stochastic volatility option models. Extension 20.2 illustrates the bias in vega
hedging and provides a brief look into stochastic option volatility models.
VEGA HEDGING 473
EXTENSION 20.1: Stochastic Volatility Option Models
This extension shows the magnitude of the errors that occur in vega hedging when using the BSM formula. To
see the bias, we consider a more general stochastic volatility OPM in which the stock’s volatility is stochastic. In
this more general model, in which volatility risk is priced, we can compute the correct vega and compare it to the
vega from the BSM model. The difference between the two vegas characterizes the error in using the BSM vega.
For a stochastic volatility option pricing model, we need to change the constant volatility assumption A8 of
Chapter 19. This is done by retaining the same structure as expression (19.2) of Chapter 19 but by modifying
the volatility term from a constant 𝜎 to a new expression 𝜎(t), which is random. The stock price evolution is
given by
𝜎(t)2
log S (t + Δt) = log S (t) + μ − Δt + 𝜎 (t) √Δtz (1)
( 2 )
where S(t + ∆t) is the stock price at time t + ∆t, μ is the expected return on the stock per year, z is a standard
normal random variable with mean 0 and variance 1, and the evolution of the volatility is given by
𝜎 (t + Δt) = 𝜎 (t) + 𝜂 (t) Δt + 𝛽1 (t) √Δtz + 𝛽2 (t) √Δtw (2)
where 𝜂(t) is the expected change in the volatility per year, 𝛽1 (t) and 𝛽2 (t) are volatilities of the volatilities, and w
is another standard normal random variable with mean 0 and variance 1 that is independent of z.
In this model, both the stock price and the stock’s volatility follow stochastic processes. It can be shown (see
Eisenberg and Jarrow 1994) that the price of a European call option with strike price K and maturity T is
∞
c (0) = ∫ BSM (𝜃0 ) 𝜋 (𝜃0 |𝜎0 ) d𝜃0 (3)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where BSM(𝜃0 ) is the BSM formula,
T
𝜃0 = ∫ 𝜎s ds
0
and the pseudo-probability of 𝜃0 conditional on 𝜎0 is
𝜋 (.|𝜎0 )
We see that the call option’s true price is a weighted average of BSM values. The correct hedge ratio is therefore
∞
𝜕c (0) 𝜕BSM (θ0 ) 𝜕θ0 𝜕𝜋 (θ0 |𝜎0 )
=∫ 𝜋 (θ0 |𝜎0 ) + BSM (𝜃0 ) dθ0 (4)
𝜕𝜎0 0 [ 𝜕θ0 𝜕𝜎0 𝜕𝜎0 ]
Clearly this is different from the BSM vega, which is 𝜕BSM(𝜎0 )/ 𝜕𝜎0 in the notation of this extension. As such, the
difference between these two vegas documents the error in using the BSM vega when the volatility is stochastic.
The result is that hedging using the BSM vega when volatility is stochastic doesn’t work.
474 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
20.5 Calibration
In the world of applied sciences and engineering, calibration is the process of setting
a device’s measurements in tune with the measurements of another device with a
known magnitude. For example, the unit of distance meter was measured from 1889
to 1960 in terms of a prototype “meter bar” kept in the International Bureau of
Weights and Measures in France. To calibrate a second rod equal to one meter, one
needs to compare it with the meter bar and mark off an identical distance.
In derivatives pricing, calibration is a technique that estimates a derivatives
pricing model’s parameters by equating the model’s price to the market price of a
derivative. Before understanding the use and misuse of calibration, especially with
respect to the BSM model, we need to refine our understanding of economic models.
There are two types of models: theoretical and statistical.2
Theoretical and Econometric Models
■ A theoretical model relies on assumptions regarding the structure of the economy
and economic theory to derive a set of outputs (Figure 20.1 shows the basic
structure of such models). The purpose is to understand cause and effect. For
example, the BSM model is a theoretical model. In this case, we have discussed
the assumptions underlying the model. Its parameters are estimated using historical
stock price data, as described in Chapter 19, and the outputs are the option values
and the Greeks (delta and gamma). The model is designed to understand the impact
of the inputs (cause) on the outputs (effect).
■ A statistical model or, equivalently, an econometric model (because our
model arises in an economic context) identifies patterns in historical data—
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
correlations—that are expected to continue into the future. Econometric mod-
els use statistical techniques to model the patterns in historical data. Recall
that more than a century ago, London dealers used a very simple statistical
model to price options. A similar statistical model for pricing options can
be generated as follows. Using historical option prices, compute the average
call price conditioned on the strike price, the time to maturity, the stock’s
volatility, and the level of the stock price. Use this conditional average call
price as an estimate of the traded option’s value. Here the conditional averages
are capturing the correlations between the inputs (strike, maturity, volatility,
stock price) and the output (option value). A more sophisticated statistical
model for pricing call options is discussed later, which uses implied volatilities.
Both types of models have their advantages and disadvantages. Models are
designed to understand the structural shifts in a dynamic and changing economy.
Theoretical models hope to understand the casual relationships in these structural
changes. If modeled correctly, a theoretical model is useful both before and after
the structural shifts. The disadvantage of a theoretical model is that it is hard to
2
See pp. 2–3 of Applied Time Series Analysis f or Managerial Forecasting by Charles R. Nelson (1973) for
a discussion of theoretical models (which have mathematical functions to represent “causal relationships”
in some economic environment) vs. statistical (econometric) models.
CALIBRATION 475
FIGURE 20.1: Theoretical Vs. Econometric Models
Theoretical Model (Cause and Effect):
One validates a theoretical model by testing the consistency of the outputs and
assumptions against real-world data.
Assumptions
Real
Model Outputs
world
Validation
Econometric/Statistical Model (Correlations):
One validates a statistical model by testing the outputs against real-world data.
Real
Model Outputs
world
Validation
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
get it right. In contrast, statistical models are often easier to construct, but they
frequently do not capture structural shifts. Consequently, if a structural shift occurs,
the statistical model will be in error until it is reestimated.
Statistical models are indispensable when no theoretical models can be built
(we discussed an example of this in the context of risk minimization hedging in
Chapter 13). Sometimes, as we illustrate later in this chapter, a theoretical model
can degenerate into a statistical model.
Using Calibration
Calibration has two uses with respect to derivatives pricing models. These two uses
depend on whether the model is accepted or rejected when it is tested with respect
to historical data.
DERIVATIVES PRICING MODEL ACCEPTED If the derivatives pricing
model is accepted (consistent with historical data), then calibration can be used
as a quick method for estimating the model’s parameters. Note that in this
case, the estimated model’s parameters will be the same regardless of whether
they are estimated using historical data or via calibration. However, this is
476 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
not the usual circumstance under which calibration is used. The more usual
circumstance is when the derivatives pricing model is rejected by historical data.
DERIVATIVES PRICING MODEL REJECTED If the derivatives pricing
model is rejected (inconsistent with historical data), then calibration is used to make
the model’s price consistent with the market price. Note that in this case, calibration is
used to fix or overcome a pricing error embedded in the derivatives pricing model.
The process of calibration thus transforms a theoretical model into an econometric model.
This econometric model provides a nonlinear estimate of the derivative’s price. The
nonlinear estimator results from the use of the inverse of the derivatives pricing
formula.
Given that calibration generates a statistical model formulated to match market
prices, the derivatives model should be used for no other purpose.3 In this case, the
calibrated model should only be used to price derivatives. The theoretical model,
being rejected, can no longer be used to perform delta and gamma hedging. We will
return to this comment later, when we discuss implied volatilities, but before this, we
provide a simple example to illustrate these insights.
EXAMPLE 20.5: Calibration and Hedging Errors
To illustrate the incorrectness of using a calibrated derivatives pricing model to hedge when the
underlying model is rejected by the data, we return to Example 20.4, in which the stock price S(t) is
non-random and grows at the risk-free rate r. This model is easily rejected using historical stock price data.
■ Recall that under this assumption, the value of a European call option with strike price K and
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
maturity T is
c (t) = max [S (t) − Ke−r(T−t) , 0] (20.11)
and the option’s details is
𝜕c 1 if S − Ke−r(T−t) > 0
= (20.12)
𝜕S {0 otherwise
If we were to use this model to price options, we would discover that the model price does not
match the market price. This is not surprising.
■ Nonetheless, we can still use the model as a statistical model for pricing options if we use calibration.
In this case, the parameter to calibrate is the stock price itself. We use calibration to find an implicit
stock price, that is, that stock price S* that equates the model price to the market price:
cMkt (t) = S∗ − Ke−r(T−t)
3
Thomas J. Sargent of New York University discusses similar problems in macroeconomic models; see
“An Interview with Thomas J. Sargent by George W. Evans and Seppo Honkapohja” in Macroeconomic
Dynamics 9 (2005).
CALIBRATION 477
Such an implied stock price S* always exists, and of course, the implied stock price will not equal
the market price of the stock. However, this is not a concern because the purpose for using the
model is to match options prices (not stock prices). Note the analogy of the implied stock price to
an implied volatility in the context of the BSM model, in which volatility is held constant.
■ Interestingly, we can then use this statistical model with the implied stock price to value different
strike and different maturity options, and it will fit them reasonably well. This is a valid use of the
calibrated statistical model.
■ An improper use would be to use the statistical model, with the implied stock price, to hedge an option.
The option’s delta is either 1 or 0, which as noted before provides a very poor hedge.
Implied Volatilities
Computing an implied volatility with respect to the BSM model is the prime example
of calibrating a derivatives pricing model. Hence we can apply our insights from the
previous discussion of calibration to this situation. Before that, however, we recall the
definition of the BSM implied volatility.
Consider a European call option on a stock with strike price K, maturity T,
continuously compounded risk-free rate r, and stock return volatility 𝜎. Denote the
BSM model of chapter 19 by BSM(t, S, 𝜎, r, K, T). The implied volatility is that
𝜎IV such that the model call price matches the market price; that is, 𝜎IV is chosen to
make the following equation hold:
BSM (t, S, 𝜎IV , r, K, T) = cMkt (20.13)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Thus, for implied volatility, we use (t, S, r, K, T) and cMKt as inputs and then compute
𝜎IV by solving this equation. Technically, since the BSM formula is an increasing
function of the volatility, a solution always exists. Because the equation has no explicit
analytical solution, numerical approximation techniques are used to compute the
implied volatility 𝜎IV . We can rewrite the implied volatility in a more abstract form as
𝜎IV = BSM−1 (cMkt ∶ t, S, r, K, T) (20.14)
where BSM−1 ( . ) denotes the inverse function. As indicated, this expression gives a
nonlinear estimator for the option’s volatility as a function of the inputs (cMkt , t, S, r,
K, T).
Next, we illustrate the computation of an implied volatility.
478 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
EXAMPLE 20.6: Computing Implied Volatilities
■ Table 20.3 reports closing option prices for International Business Machine Corp. (IBM) on February
3, 2011, along with relevant input data:
- Column 1 reports six strike prices $155 to $175 at $5 intervals.
- Column 2 reports the closing stock price on IBM.
- Column 3 reports the continuously compounded risk-free interest rate r = 0.10 per year
(chapter 2 shows how to compute r from Treasury bill prices).
- Column 4 reports the time to maturity. Because options expire on February 19, there are sixteen
days remaining before maturity. Thus, T = 16/365 = 0.0438 years.
- Column 5 reports the call prices from the market. We assume that these options are European.
- Column 6 reports the implied volatility. The implied volatility in column 6 are obtained in Excel
by first guessing a value for the implied volatility, using this and the other inputs to determine cBSM ,
and then using Excel’s “Goal Seek” command to solve for the implied volatility that would set the
difference (cMkt – cBSM ) to zero.4
TABLE 20.3: Implied Volatilities
K S r T cMkt Implied Volatility
150 163.53 0.001 0.0438 13.575 0.1938
155 163.53 0.001 0.0438 8.6 0.1405
160 163.53 0.001 0.0438 4.0 0.1234
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
165 163.53 0.001 0.0438 1.08 0.1249
170 163.53 0.001 0.0438 0.17 0.1324
175 163.53 0.001 0.0438 0.055 0.1648
■ Figure 20.2 plots the strike prices (column 1) along the horizontal axis and implied volatilities
corresponding to each strike along the vertical axis. This gives a pattern that is referred to as a volatility
smile.
■ When computing implied volatilities, if the BSM model were consistent with the data, then for all
strikes, one would get the same value. The graph would be a flat line. The implied volatilities are not
equal, hence the BSM model is rejected by this example’s data.
CALIBRATION 479
FIGURE 20.2: Implied Volatility Smiles
Implied
volatility 0.25
0.20
0.15
0.10
0.05
0.00
150 155 160 165 170 175
Strike Prices
4 You can solve for implied volatility by (1) trial and error, (2) by a spreadsheet program as outlined above (see the problems at
the end of this chapter for more details), or (3) by using our Priced! software.
To understand the proper use of the calibrated BSM with an implied volatility,
we first need to determine if the BSM model is accepted or rejected by historical
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
data. Unfortunately, it is well known that the BSM model is rejected by historical data.
The most direct verification of this rejection has been studies computing implied
volatilities (as in the previous example). When computing implied volatilities, if the
BSM model were correct, then for all the other inputs (t, S, K, r, T), one would get
the same value for the volatility, and it would be equal to the historical volatility.
Unfortunately, when estimating implied volatilities from equity options, the
evidence shows that (1) implied volatilities differ across strikes and maturities (the
pattern is a smile or sneer; see Figure 20.2 for a typical pattern); (2) implied volatilities
differ across time, stock prices, and interest rates; and (3) implied volatilities do not
equal historical volatilities (see Bakshi et al. 1997; Pan 2002). This evidence strongly
rejects the BSM model’s validity.
Given that the BSM model is rejected, one can still use a calibrated BSM as a
nonlinear statistical model for estimating options prices. This is typical industry practice.
A common procedure is to estimate the implied volatility using a subset of the
available options and then to use the resulting implied volatility estimate to price
the remaining set of options or options at a later date. This is a perfectly valid use of the
BSM model.
480 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
However, because the BSM model is rejected by historical data, it is incorrect to use a
calibrated BSM to delta or gamma hedge. This is a common misuse of the BSM model.
The hedging implications of the theoretical model are no longer valid. To hedge
in this circumstance, one needs to employ either an alternative and more complex
theoretical model or a purely statistical model (as was done for futures prices in
Chapter 13). Vega hedging will not work, contrary to common belief, for two reasons:
first, it is a wrong Greek to hedge because it was held constant in the original BSM
model (volatility risk is not reflected in the BSM price), and second, the BSM model
is rejected by historical data, hence its partial derivatives—the Greeks—are irrelevant.
For a study showing the poor hedging that results from using the BSM model with
delta, gamma, and vega hedging, see Engle and Rosenberg (2000).
EXTENSION 20.2: VIX: The Fear Index
If you think implied volatility is scary, it has been used to develop something even scarier—the “fear index,”
called the VIX. The Chicago Board Options Exchange introduced the VIX in 1993 “to measure the market’s
expectation of the 30-day volatility implied by at-the-money S&P 100 Index (OEX) option prices.”5 In 2003,
the VIX index was modified to represent an expected volatility instead of an implied volatility. Today, the VIX
has developed into a widely watched barometer of stock market volatility, reflecting investor anxiety. The VIX
index increases during financial crises. Today, various derivatives trade on VIX, yielding investors and financial
institutions a method to hedge or speculate on stock market volatilities
5
See “The CBOE Volatility Index – VIX” (www.cboe.com/micro/vix/vixwhite.pdf), page 2.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
20.6 The Black–Scholes–Merton Model: A
Postscript
As shown, because the BSM model is rejected by historical data, although a calibrated
BSM model is still useful for pricing options, it is not useful for hedging. A calibrated
BSM model is a statistical model in which the implied volatility estimate captures
any model errors into its magnitude. These model errors, however, invalidate its use
for hedging. So what can one do to overcome the hedging limitation of the BSM
model?
There are two approaches. First, one can build a statistical model relating changes
in the option’s price to changes in the stock’s price, analogous to that used in Chapter
13 for futures prices. Such a statistical model will inherit all the advantages and
disadvantages of such models in predicting the future. Second, one can build a better
mousetrap! The better mousetrap would generalize the assumptions so that the option
model is no longer rejected by historical data.
A preliminary guide to these extensions is as follows. Extension 20.2 discusses
how to extend the BSM model to include volatility risk. Chapters 21–25 of this
SUMMARY 481
book discuss how to include interest rate risk in OPMs. Asset price bubbles and their
impact on option pricing were discussed in Chapter 19. Finally, credit and liquidity
risk extensions of OPMs are discussed in Chapter 26. Transaction costs and trading
restrictions are also worthy of study, but we leave their discussion to more advanced
textbooks and the academic literature.
20.7 Summary
1. Delta hedging is removing the price risk from a long call position by selling short
𝜕c
deltaC ≡ = N (d1 ) shares of stock
𝜕S
where
S 𝜎2
log ( ) + r + T
K ( 2)
d1 =
σ√T
A delta hedge works well when the time interval over which the hedge is executed
is small.
2. Delta and gamma hedging removes the price risk from a long call with price c1 by
buying n1 shares of stock with price S and selling n2 shares of another option on
the same stock with price c2 , where
n1 = (gammaC1 /gammaC2 ) deltaC2 − deltaC1
and n2 = −gammaC1 /gammaC2
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The Greeks are defined by the general formulas
𝜕c 𝜕2 c 1
deltaC ≡ = N (d1 ) and gammaC ≡ 2 = N′(d1 )
𝜕S 𝜕S Sσ√T
where N( . ) is the cumulative standard normal distribution function and N ′ ( . ) is
the standard normal density function. Delta and gamma hedging are designed to
overcome the errors introduced when the hedging interval is not small.
3. Vega hedging is a nonsensical procedure because it is based on an option valuation
formula, the BSM model, that assumes volatility is a constant. Hence volatility risk
is not embedded in the BSM formula, so the formula’s sensitivity to changes in
the stock’s volatility, the vega, does not capture volatility risk in an option’s price.
4. Calibration is an often-misused estimation procedure. In the context of the BSM
formula, implied volatility estimation is an example of calibration. Because the
BSM is rejected by historical data, using a calibrated BSM model transforms the
theoretical BSM model into a statistical model. This statistical model can be used
to estimate options prices, but it is invalid to use a calibrated BSM model to delta
(and/or gamma) hedge an option position.
482 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
20.8 Appendix
The Mathematics of Delta, Gamma, and Vega
Hedging
Hedging using the Greeks is based on a Taylor series expansion. A Taylor series
expresses a well-behaved function as a sum of terms involving its derivatives computed
at a single point. For a function f of two variables x and y, the Taylor series about the
point (a, b) can be written as (where [x – a] is ∆x, [y – b] is ∆y, f [x, y] – f [a, b] is ∆f ;
see Appendix A)
𝜕f 𝜕f 1 𝜕2 f 2 𝜕2 f 2 𝜕2 f
Δf = Δx + Δy + Δx + Δy + 2 ΔxΔy + error
𝜕x 𝜕y 2 [ 𝜕x2 𝜕y2 𝜕x𝜕y ]
where the higher-order terms of the expansion are considered insignificant “error
terms” when ∆x and ∆y are small.
DELTA HEDGING Suppose one can trade quickly enough so that the infinitesimal
time change dt is approximately equal to the small time change ∆t that we consider
in practice, that is, dt ≈ ∆t. Then the BSM theory implies that delta hedging removes
all of the stock price risk from an option position. We now explain how this happens.
■ The option and the stock portfolio. Consider a portfolio consisting of one European
call option with time t price c(t). We want to trade n shares of a stock worth S(t)
to remove the stock price risk from this option position. Denote the time t value
of the portfolio as V (t). In symbols, we can write the evolution of the change in
the portfolio’s value over a small time interval [t, t + ∆t] as
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
ΔV = Δc + nΔS (1)
where ∆V ≡ V (t + ∆t) – V (t) and the other terms are similarly defined.
■ The Taylor series expansion for a delta hedge. Because the BSM call price is a function
of time t and the stock price S(t) (other terms like the volatility, the interest rate,
and the strike price are constant parameters and not variables), we can use a Taylor
series expansion to write
𝜕c 𝜕c
Δc = Δt + ΔS + error (2)
𝜕t 𝜕S
where (error/∆t) goes to zero as the time interval shrinks to zero (∆t → 0). This last
observation implies that for small changes in time, ∆t, the error term is small and
can be ignored. Ignoring these terms, substitution of expression (2) into expression
(1) gives the change in the value of our portfolio as
𝜕c 𝜕c
ΔV = Δt + ΔS + nΔS (3)
𝜕t 𝜕S
APPENDIX 483
■ Determining the hedge ratio and constructing the delta-neutral portfolio. To make this
portfolio riskless for small changes in the stock price, choose n so that the term
involving ∆S disappears. This is the only random term in expression (3). This is
obtained by setting
𝜕c
ΔS + nΔS = 0 (4)
𝜕S
Solving for n yields
𝜕c
n=− = −N (d1 ) ≡ −deltaC (5)
𝜕S
The number of shares held has the opposite sign to the option position. Because we
began with one long call, we need to short deltaC ≡ N(d1 ) shares of the underlying
stock to hedge the exposure.
■ The delta is the holy grail of options pricing. The hedge ratio n represents the number
of shares of the stock to trade for each call to construct a riskless portfolio. Being
riskless, this portfolio is equivalent to a position in a money market account. To
see this, substituting n into expression (3) gives
𝜕c
ΔV = Δt (6)
𝜕t
Thus the change in the portfolio’s value equals 𝜕c/𝜕t (which equals the call’s theta;
see Table 19.1) times the time interval ∆t. This term is nonrandom, and hence it
has zero variance. To avoid arbitrage, this portfolio must earn the riskless rate. This
observation (combined with an extension of this approach, as described in the next
section in the context of gamma hedging) leads to an alternative derivation of the
BSM model. The portfolio V that we have constructed is said to be delta neutral.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Notice that delta hedging involves the first-order derivatives. Such approximations
work well over small time periods, during which the stock price does not move
much.
GAMMA HEDGING How does one remove the price risk of an option position
over a time interval that is longer than what we had considered earlier: a day, or
perhaps even a week (in other words, the hedging interval ∆t is significantly larger
than the infinitesimal time interval dt considered in the model)? We can achieve this
by adding a gamma hedge to the delta hedge.
■ The Taylor series expansion for delta and gamma hedging. Suppose that we are long
one European call option and we want to remove the stock price risk from our
portfolio over the time interval [t, t + ∆t]. As before, we start with a Taylor series
expansion of the call’s value, but this time, we include the second-order term:
𝜕c 𝜕c 1 𝜕2 c
Δc = Δt + ΔS + (ΔS)2 + error (7)
𝜕t 𝜕S 2 𝜕S2
where (error/∆t) goes to zero as ∆t→ 0. In the calculation, we ignore the error
term because it is small.
484 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
■ Adding a stock and another option to the option portfolio. Notice that there are “two”
risks in the call’s price movements—one comes from the change in the stock price
and the other from the change in the stock price squared. To hedge two risks, we
need two securities. We can use the underlying stock as one security and another
option on the same stock as the second. At time t, we create a portfolio V (t)
consisting of one long call whose price is c1 , n1 shares of the stock whose price is
S(t), and n2 number of the second option whose price is c2 :
V (t) = c1 + n1 S + n2 c2 (8a)
■ Constructing the delta- and gamma-neutral portfolio. To determine the portfolio’s
composition, we write its value change over [t, t + ∆t] as
ΔV (t) = Δc1 + n1 ΔS + n2 Δc2 (8b)
- Substitution of expression (7) into (8b) for the change in the two calls’ values
gives
𝜕c1 𝜕c 1 𝜕2 c1
ΔV = Δt + 1 ΔS + (Δs)2 + n1 ΔS
𝜕t 𝜕S 2 𝜕S2
(9)
𝜕c 𝜕c2 1 𝜕2 c2
+n2 2 Δt + n2 ΔS + n2 (ΔS)2
𝜕t 𝜕S 2 𝜕S2
- To make this portfolio have zero risk, we want to choose n1 and n2 such that
the coefficients of the terms involving ∆S and (∆S)2 disappear. This gives two
equations in two unknowns:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
𝜕c1 𝜕c
+ n1 + n2 2 = 0 (10a)
𝜕S 𝜕S
𝜕2 c1 𝜕2 c2
+ n2 =0 (10b)
𝜕S2 𝜕S2
Notice that expression (10a) involves the Greek deltas and that the second
expression (10b) involve the Greek gammas (see Table 19.1).
- Rearrange terms in (10b) to get
𝜕2 c1 𝜕2 c2
n2 = − / (11a)
( 𝜕S2 ) ( 𝜕S2 )
The negative sign indicates that we need to sell short n2 options. Alternatively,
you can write n2 as –gammaC1 /gammaC2 .
APPENDIX 485
- Plug this value into (10a) and rearrange terms to get
𝜕2 c1 𝜕2 c2 𝜕c2 𝜕c1
n1 = / − (11b)
[( 𝜕S2 ) ( 𝜕S2 )] 𝜕S 𝜕S
This can also be written as (gammaC1 /gammaC2 )deltaC2 – deltaC1 . The hedge
involves going long n1 shares of the stock if the preceding expression is positive
and selling them short if the expression is negative.
- Substituting these shares into the portfolio’s value shows that the resulting
portfolio is riskless:
𝜕c1 𝜕2 c1 𝜕2 c2 𝜕c2
ΔV = Δt − / Δt (12)
𝜕t [( 𝜕S 2 ) ( 𝜕S 2 )] 𝜕t
Such a portfolio, which consists of nonrandom terms, is said to be delta- and
gamma-neutral.
VEGA HEDGING As we discuss in the text, vega hedging is “the bad and
the ugly” and should not be used. To see why this is an invalid procedure, let
us begin by following the logic that we employed for both delta and gamma
hedging.
- The Taylor series expansion for a delta and vega hedge. We start with the BSM
formula, which depends on the stock price, time, and volatility, written c[S(t), t,
𝜎]. To hedge both the delta and vega over [t, t + ∆t], we proceed to use a Taylor
series expansion, as before:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
𝜕c 𝜕c 𝜕c
Δc = Δt + ΔS + Δσ + error (13)
𝜕t 𝜕S 𝜕𝜎
where (error/∆t) goes to zero as ∆t → 0. Because the error term is small, we
ignore it in the subsequent calculations.
- Adding a stock and another option to the portfolio. There are two risks in the call’s
price movements: one is the change in the stock price and the other is the change
in the stock’s volatility. As before, we hedge them using the underlying stock and
a second option on the same stock. This gives the portfolio’s value at time t:
V (t) = c1 + n1 S + n2 c2 (14)
- Constructing the delta-and vega-neutral portfolio. We write the change in the
portfolio’s value over [t, t + ∆t] as
ΔV (t) = Δc1 + n1 ΔS + n2 Δc2 (15)
486 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
Substitution of the change in the call’s value gives an expression analogous to
expression (9):
𝜕c1 𝜕c 𝜕c
ΔV = Δt + 1 ΔS + 1 Δ𝜎 + n1 ΔS
𝜕t 𝜕S 𝜕𝜎
(16)
𝜕c2 𝜕c2 𝜕c
+n2 Δt + n2 ΔS + n2 2 Δ𝜎
𝜕t 𝜕S 𝜕𝜎
- To make this portfolio have zero risk, we choose n1 and n2 such that the terms
involving ∆S and ∆𝜎 disappear. This is accomplished by requiring
𝜕c1 𝜕c
+ n1 + n2 2 = 0 (17a)
𝜕S 𝜕S
𝜕c1 𝜕c
+ n2 2 = 0 (17b)
𝜕𝜎 𝜕𝜎
- Rearrange terms in (17b) to get
𝜕c1 𝜕c2
n2 = − / (18a)
( 𝜕𝜎 ) ( 𝜕𝜎 )
A negative sign indicates that we need to sell short vegaC1 /vegaC2 options.
- Plugging this value into expression (17a) and rearranging terms yields
𝜕c1 𝜕c2 𝜕c2 𝜕c1
n1 = / − (18b)
[( 𝜕𝜎 ) ( 𝜕𝜎 )] 𝜕S 𝜕S
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ However, there is a conceptual problem in this application. We have kept the
parameter 𝜎 fixed throughout our derivation of the BSM model, and now we
want to change it! This is nonsensical because the risk of a changing volatility is
not reflected in the option’s value. Consequently, vega hedging is an invalid procedure
that should never be used.
A Third Derivation of the Black–Scholes–Merton
Formula
The original derivation of the BSM formula in 1973 for valuing a European option
was obtained as the solution to a partial differential equation. It follows from the
Taylor series expansion used for delta and gamma hedging, with one small twist. Of
course, this approach needs the underlying assumptions given in chapter 19.
Assume that the European call option’s value depends only on the price of the
stock and time, that is, c(t) = c[t, S(t)]. Then, you can create a “hedged position” by
combining a long position in the stock with a short position in the option. Letting
the short position in the stock be (𝜕c/𝜕S) shares, the portfolio’s value is
𝜕c
V=c− S (19)
𝜕S
CASES 487
■ Suppose that the stock price changes by ∆S ≡ S(t + ∆t) – S(t) over a small interval
[t, t + ∆t] and the call price changes by ∆c. Then, the change in the value of our
portfolio is
𝜕c
ΔV = Δc = ΔS (20)
𝜕S
■ Use of Ito’s lemma and stochastic calculus. We can use stochastic calculus to develop an
alternate expression for the change in the portfolio’s value:
- If the option position changes continuously, Ito’s lemma yields
𝜕c 𝜕c 1 𝜕2 c
Δc = Δt + ΔS + (ΔS)2 (21)
𝜕t 𝜕S 2 𝜕S2
This is similar to equation (7).
- Expression (18) can be further simplified by using another result from stochastic
calculus, which allows us to write the squared change in the stock price as (∆S)2
= 𝜎2 S2 ∆t. This is the “one small twist” mentioned earlier.
- Substituting into (21) and applying it to (20) yields
𝜕c 𝜕c 1 𝜕2 c 2 2 𝜕c
ΔV = Δt + ΔS + 𝜎 S Δt − ΔS
𝜕t 𝜕S 2 𝜕S 2 𝜕S
(22)
2
𝜕c 1𝜕c 2 2
= Δt + 𝜎 S Δt
𝜕t 2 𝜕S2
■ A riskless hedge. Because all terms on the right side of (22) are deterministic, no
arbitrage necessitates that the hedged portfolio V must earn the risk-free rate over
the time period ∆t, that is, rV ∆t. Substitution and canceling ∆t from both sides
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
gives
𝜕c 1 𝜕2 c 2 2
rV = + 𝜎S (23)
𝜕t 2 𝜕S2
■ Black–Scholes–Merton partial differential equation. Substituting (19) into (23) and
rearranging terms gives the famous BSM partial differential equation (pde)
𝜕c 𝜕c 1 𝜕2 c
= rc − rS − 𝜎2 S2 2 (24)
𝜕t 𝜕S 2 𝜕S
■ Derivation of the Black–Scholes–Merton formula. Using the call option’s boundary
condition c(T) = max[S(T) – K, 0], the solution to this pde is the BSM formula.
20.9 Cases
J&L Railroad (Darden School of Business Case UV0251-PDF-ENG, Harvard Busi-
ness Publishing) and J&L Railroad: The Board Meeting (Darden School of
Business Case UV2566-PDF-ENG, Harvard Business Publishing). The first case
studies a railroad company’s decision to hedge next year’s expected fuel demand.
488 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
The second case, which may be done in conjunction with the first case, considers
these issues in greater detail.
Pine Street Capital (Harvard Business School Case 201071-PDF-ENG). The case
examines issues faced by a technology hedge fund in trying to decide whether
and/or how to hedge equity market risk by short selling and trading options.
Sleepless in LA (Richard Ivey School of Business Foundation Case 905N11-PDF-
ENG, Harvard Business Publishing). The case discusses the Black–Scholes–
Merton model for options pricing, the concept of implied volatility, and put–call
parity. It also shows how options pricing can be used to value the corporate
liabilities of a financially distressed company.
20.10 Questions and Problems
The first five questions use the following information for pricing European options.
Special Motors Corporation’s stock price S is $59, the strike price K is $60, the
maturity T is forty-four days, the implied volatility 𝜎 is 30 percent per year, and the
risk-free interest rate r is 3.3 percent per year.
20.1 a. Given the preceding data, compute the call option’s delta using the Black–
Scholes–Merton model.
b. Given the preceding data, compute the put option’s delta using the Black–
Scholes–Merton model.
20.2 a. Describe how to set up the delta hedge for a long call option position based
on the computations in problem 20.3.
b. If the stock price changes to $60 over the next trading day, what is the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
delta hedging error? Assume the Black–Scholes–Merton model accurately
describes movements in the call’s value.
20.3 a. Describe how you will set up the delta hedge for a short put option position
based on computations in problem 20.3.
b. If the stock price changes to $60 over the next trading day, what is the
delta hedging error? Assume the Black–Scholes–Merton model accurately
describes movements in the call’s value.
20.4 a. Given the data in question 20.3, compute the call option’s gamma on the
basis of the Black–Scholes–Merton model.
b. Given the data in question 20.3, compute the put option’s gamma on the
basis of the Black–Scholes–Merton model.
20.5 a. Consider a portfolio consisting of 50 shares of the stock, long 10 calls and
short 30 puts. What is the portfolio’s delta? (Use delta values from question
20.3.)
QUESTIONS AND PROBLEMS 489
20.6 a. What is vega hedging?
b. Describe some conceptual problems with this once-popular practice.
20.7 If you want to hedge volatility risk in a call option position, how would you
go about doing so?
20.8 What is calibration with respect to the BSM model? Explain how to use
calibration correctly. When is calibration used incorrectly?
Implied Volatilities
K cMkt Implied Volatility
150 13.00
155 8.00
160 3.75
165 1.00
170 0.20
175 0.05
20.9 Briefly describe the two types of models that we introduced, highlighting their
strengths and weaknesses.
20.10 When constructing a theoretical model for pricing a derivative, if the theo-
retical model is rejected by historical data, can one still use the model to price
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
derivatives? Explain your answer using the BSM as an example.
20.11 Complete the following table using the following information:
Comment on what this implies about the validity of the Black–Scholes–
Merton model?
- Column 1 reports six strike prices $155 to $175 at $5 intervals.
- Column 2 is for the closing stock price on YBM, $160.
- Column 3 is for the risk-free interest rate r = 10 percent per year.
- Column 4 is for the time to maturity, which is twenty days (20/265 =
0.054795 years).
- Column 5 reports European call option prices from the market.
- Column 6 reports the implied volatility that you need to calculate.
20.12 a. When using historical prices, is the Black–Scholes–Merton model accepted
or rejected by the data?
b. If the model is accepted, explain why. If the model is rejected, give one
probable cause of rejection.
490 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
20.13 If historical data rejects the Black–Scholes–Merton model,
a. can the model still be used for pricing options? Explain your answer.
b. can the model be used for hedging option positions? Explain your answer.
20.14 What is a volatility smile with respect to the Black–Scholes–Merton model?
20.15 a. What is delta hedging? If the Black–Scholes–Merton model is correct, do
you need delta hedging?
b. What is gamma hedging? Under what conditions do you need gamma
hedging on top of delta hedging?
20.16 In hedging using the Black–Scholes–Merton model, we used the phrase “the
good, the bad, and the ugly.” Identify “the good,” and “the bad and the ugly”
and explain why these descriptions are accurate or may be inappropriate.
20.17 Given the evidence supports that the stock’s volatility is stochastic and the
BSM model is rejected by the data, how can one hedge options?
20.18 (Microsoft Excel Project) We thank Professor Craig W. Holden of Indiana
University, Bloomington, for providing a similar problem (copyright-free) on
his website (www. kelley. iu. edu/cholden). His website www. excelmodeling.
com has several books teaching hands-on practical finance with the help of
Microsoft Excel.
Overview
The purpose of this assignment is to introduce you to the science of pricing options.
The project has two parts:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ You will calculate and graph the value of European-style calls and puts using the
built-in spreadsheet functions, such as the cumulative normal.
■ Using actual options data from the Wall Street Journal, you will also calculate
the implied volatilities using the solver module and graph the “smile” pattern of
implied volatility.
Data (from the Wall Street Journal, March 3, 2000)
The closing price for the Standard and Poor’s 500 (S&P 500) index (symbol SPX)
was 1,381.76. As options expire on the third Friday of the expiration month (17th
for March contracts, 21st for April contracts), choose the (annualized) ask yield rate
for the Treasury bill that matures nearest to the expiration date. Express T as fraction
of a year: 14/365 for March options and 49/365 for April options.
QUESTIONS AND PROBLEMS 491
S&P 500 Index Put Option Prices
Strike Last (Price) Strike Last (Price) Strike Last (Price)
Mar 1325c 67 Mar 1375c 27 Mar 1425c 6 1/4
Mar 1325p 5 1/2 Mar 1375p 17 Mar 1425p 47 1/2
Apr 1325p 18 3/8 Apr 1375c 51 Apr 1425c 24
Mar 1350c 44 Apr 1375p 32 Apr 1425p 54
Mar 1350p 10 Mar 1400c 13 7/8 Mar 1450c 2 1/4
Apr 1350c 62 1/2 Mar 1400p 29 Mar 1450p 67
Apr 1350p 24 Apr 1400c 35 1/4 Apr 1450c 14 1/2
Apr 1400p 44 Apr 1450p 72 1/2
A “c” denotes call price and a “p” denotes put price.
T-Bill Prices
Maturity Days to Maturity Ask Yield
Mar 16 13 5.28
Mar 23 20 5.13
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Mar 30 27 5.30
Apr 06 34 5.31
Apr 13 41 5.48
Apr 20 48 5.61
Apr 27 55 5.76
The Project
Do the project in Excel worksheets. Print out and turn in all of your results.
Deliverables are noted below.
a. Cumulative Normal and Normal Density Graphs
Examine a standard normal distribution. Find the cumulative normal and the
normal density for values of d from –3.0 to 3.0 in increments of 0.1. Record
this in a table in your Excel worksheet. Then graph both the cumulative normal
and the normal density on the same diagram with d on the x axis.
492 CHAPTER 20: USING THE BLACK–SCHOLES–MERTON MODEL
Deliverable: (1) The table with cumulative normal distribution function and the
standard normal density function’s values.
(2) The graph generated.
b. Option Pricing Using Black–Scholes–Merton Formula and Intrinsic Value
Computation
Use the Black–Scholes–Merton formula to estimate the call price and the put
price for options with the following data:
Strike price K= 10
Risk-free interest rate r = 3 percent or 0.03 per year
Volatility 𝜎 = 60 percent or 0.6 per year
Time to maturity T = 1 year
S(0) ranging from 0 (actually 0.0000001, a very small value) to 20,
in increments of 2
Record these numbers and computations in a table.
Deliverable: (3) Table with various inputs and corresponding call and put prices
estimated using the BSM model.
c. Intrinsic Value Computation and Graph
Now calculate call intrinsic values and put intrinsic values for the same options
(K = 10 and S(0) ranging from 0 to 20 in increments of 2).
Graph both the call price and the call intrinsic value on the same diagram with
S(0) on the x axis.
Similarly, graph both the put price and the put intrinsic value on the same
diagram with S(0) on the x axis.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Deliverable: (4) Table with the call price and the call intrinsic value, and the
put price and the put intrinsic value. (You may combine (3) and (4) and present
it in a single table.)
(5) The call graph.
(6) The put graph.
d. Implied Volatility Computation and Graph
Given data from the March 3, 2000, issue of the Wall Street Journal, calculate
the implied volatility from S&P 500 index put options (not index call options)
for K (strike) values of 1325, 1350, 1375, 1400, and 1425 for the March
contract.
Then calculate the implied volatility from S&P 500 index put options (not
index call options) for K values of 1325, 1350, 1375, 1400, and 1425 for the
April contract.
Graph the implied volatility diagram with the exercise price X on the x axis
and the March and April implied volatilities on the y axis.
Deliverable: (7) The table with the implied volatilities.
QUESTIONS AND PROBLEMS 493
(8) The graph with the implied volatilities.
( Suggestions and useful information, relevant excel commands, and implied volatility
computation)
Goal Excel Command
1. To calculate the natural log: =LN(A1)
2. To calculate the normal density: =NORMDIST(A1,0,1,FALSE)
3. To calculate the cumulative normal: =NORMDIST(A1,0,1,TRUE)
4. To calculate the exponential: =EXP(A1)
5. To solve implied volatility: Goal Seek
To estimate implied volatilities using Goal Seek command in Excel 2007 or later
versions:
■ In an Excel sheet, record different strike prices for March in one column and the
various model inputs in other columns.
■ In the column with the heading “Implied volatility,” start with an arbitrary volatility
number (say 0.2) and use it to compute March put price.
■ Record market-traded March put prices corresponding to the different strike
prices. In another column, note the “Put Price Difference” to record the difference
between “market traded put price” and the “model generated put price” solved
from the model.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Next, select a cell under “Put Price Difference.” On the Data tab, in the Data Tools
group, click What-If Analysis, and then click Goal Seek.
- In the “Set cell” box, record the current cell box.
- In the “To value” box, type in 0.
- In the “By changing cell” box, type in the location of the cell corresponding to
the implied volatility from the same row.
- Click OK to get the implied volatility in its own column. (In this case, the zero-
price difference is obtained by changing the standard deviation.)
■ Hold and drag the mouse down to similarly solve implied volatilities corresponding
to March and April put options corresponding to different strike prices.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
IV
Interest Rate Derivatives
CHAPTER 21
Yields and
Forward Rates
CHAPTER 22
Interest Rate Swaps
CHAPTER 23 CHAPTER 24
Single-Period Multiperiod
Binomial HJM Model Binomial HJM Model
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
CHAPTER 25
The HJM
Libor Model
CHAPTER 26
Risk
Risk Management
Management
Models
21
Yields and Forward Rates
21.1 Introduction Using Forward Rates
EXTENSION 21.1 Home EXTENSION 21.3 Computing
Ownership and Derivatives Forward Rates from Coupon
Bond Prices
21.2 Yields
Revisiting Zeros 21.5 The Basic Interest Rate
Derivatives Contracts
Yields and the Yield Curve
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
A Brief History
EXTENSION 21.2 The
Expectations Hypothesis Forward Rate Agreements
Interest Rate Futures
21.3 The Traditional Approach
EXTENSION 21.4 Treasury
Duration
Futures
Modified Duration Hedging
The Equivalence between
Applications and Limitations Forward and FRA Rates
21.4 Forward Rates 21.6 Summary
The Definition
21.7 Cases
Understanding Forward Rates
21.8 Questions and Problems
INTRODUCTION 497
21.1 Introduction
In 1909, the founder of Ford Motor Company, Henry Ford, announced that
the company would build only one car, the Model T, and famously declared,
“Any customer can have a car painted any colour that he wants so long as it is
black.”1 Similarly, mortgage loans used to be simple as well. In the 1950s and
1960s, home mortgages were typically made by savings and loans (S&Ls), whose
standard offering was a no-frill 30-year, fixed-rate loan. S&Ls were known as
3-6-3s because they paid depositors 3 percent, charged borrowers 6 percent, and
the managers went to play golf at 3:00 pm! But times have changed.
Today home mortgages come with a variety of embedded options and complicated
provisions (see Extension 21.1). There is a bewildering assortment of fixed-and
adjustable- rate mortgages. Some have teaser rates with ceilings and floors on interest
rates charged. Some don’t pay principal in the early years, whereas others are
extendible. Most can be repaid without penalty. To understand these provisions—
the embedded options—one needs to understand interest rate derivatives (IRDs).
This is the purpose of this and the remaining chapters.
Who uses IRDs? All players in the economy: individuals, corporations, govern-
ments, and financial institutions:
1. Individuals. A family’s assets and liabilities are all sensitive to interest rate move-
ments. On the liability side, interest rates affect payments on home mortgages,
car loans, personal loans, and credit card balances. On the asset side, they affect
earnings on savings accounts, certificates of deposit, and money market funds.
2. Corporations and governments. Corporations routinely use IRDs to structure their
interest rate–sensitive liabilities and assets. Mutual funds, hedge funds, pension
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
funds, insurance companies, and college and university endowments use IRDs
to manage their investments. Last, they are essential for managing the complex
treasury operations of local, state, and federal government.
3. Financial institutions. Financial institutions use advanced knowledge of IRDs to
develop products for their clients and to manage their own interest rate risks.
Our discussion of IRDs parallels our earlier presentation of spots, forwards, futures,
and options on commodities and equities; however, there’s a twist. For IRDs, an array
of different maturity default-free zero-coupon bonds replace the single commodity or
stock price as the underlying asset. This small change raises some major hurdles for
IRD pricing models. Still, despite its reputation for complexity, this subject is easy to
understand if it is introduced gradually and with forethought. Magically, and perhaps
without your awareness, the early chapters have already built this foundation for you!
For simplicity, we focus on Eurodollar IRDs. They are simpler than derivatives on
Treasury securities. Moreover, Eurodollar IRDs attract the greatest trading volume,
1
See Chapter IV of My Life and Work (1922) by Henry Ford in collaboration with Samuel Crowther
from Project Gutenberg website www.gutenberg.org/dirs/etext05/hnfrd10.txt, February 23, 2011.
498 CHAPTER 21: YIELDS AND FORWARD RATES
and you can have your cake and eat it too. Once you understand Eurodollar IRDs,
it is easy to comprehend derivatives based on Treasuries.
EXTENSION 21.1: Home Ownership and Derivatives
Suppose you find your dream house and submit an offer to buy the house for $500,000. A flurry of activity now
begins. If the seller accepts the offer, the two of you first create an option contract. The seller’s broker requires
you to place a deposit, say, $5,000, with them. Should you decide to back out, the $5,000 will be forfeited;
otherwise, it goes toward the purchase price. This deposit creates a call option to buy the house.
You now go to a bank or a mortgage broker to arrange a loan. They obtain your credit score indicating how
you have handled debt in the past: credit card as well as other loan payments. Your credit score helps determine
whether you get the loan plus the magnitudes of the down payment required and the interest rate charged.
On the basis of this evaluation, suppose that the lender agrees to loan you $450,000 for the house. This loan
is called a mortgage. You have to pay this principal back in fifteen or thirty years with interest paid in monthly
installments, and you have to come up with a $50,000 down payment. This 10 percent down payment provides
a cushion against a decline in the house’s value if you decide to leave town without a forwarding address! Unless
you make a 20 percent down payment, the lender sometimes requires you to pay for mortgage insurance, which
is a put option on the risk of your not repaying the principal (a credit derivative!).
Mortgages come in two varieties. A fixed-rate mortgage (FRM) charges a fixed interest rate over the loan’s
life. For a FRM, you have to pay back the loan in equal monthly installments over fifteen or thirty years. An
adjustable-rate mortgage (ARM) usually charges a lower initial interest rate for one to five years, with the
remaining payments adjusted depending on future interest rates. ARM mortgages come with caps and floors on
the interest paid per period to protect both the lender and the borrower from adverse interest rate movements.
These caps and floors are call and put options on interest rates—more IRDs!
Your expertise in modern finance should come in handy. If you are planning to stay in your home for a short
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
period, or if you are betting that interest rates will remain low, or if you are anticipating a significant increase in
your personal wealth in the near future, then an ARM may make sense. In addition, in the United States, most
loans come without a prepayment penalty, so you can choose to pay back the loan before it matures. If interest
rates fall, you also have the option of refinancing your home and locking in a cheaper interest rate.
In years past, the major players in the mortgage market were banks and savings and loan associations (S&Ls).
The money for mortgage loans was raised and lent in local markets. Thanks to modern finance, a mortgage broker
can now lend you the money for a house mortgage and immediately sell that loan in a national market. Typically,
a loan is split into two parts. The right to service the loan is held by entities like the Bank of America or JP
Morgan Chase, which regularly collects your mortgage payments and coaxes you to buy other products. The
loan itself may be combined with other mortgage loans and sold to Wall Street wizards, who then combine these
mortgage loans into a pool and sell bonds against the pool to cover the cost of their purchases. These bonds are
called asset-backed securities or, in this case, mortgage-backed securities.
Did you really think the mortgage business was staid and stodgy?
On our journey to understanding IRDs, this chapter introduces yields and forward
rates, which are fundamental concepts. We start by revisiting zero-coupon bonds,
from which we extract yields, and the yield curve, concepts that are often introduced
in basic finance and investment courses. We also discuss the once-popular but now
YIELDS 499
fading techniques of duration and convexity based interest rate risk management.
Then, we introduce forward rates, which are at the heart of modern interest rate risk
management. We wrap up this chapter with a discussion of forward rate agreements
(FRAs) and interest rate futures, which are analogous to forwards and futures on
commodities and equities.
21.2 Yields
To understand IRDS, we need to start at the beginning and understand yields. Yields
are obtained from zero-coupon bond prices.
Revisiting Zeros
Our basic building blocks are risk-free zero-coupon bonds of different maturities.
US Treasury bills and STRIPS and Eurodollar deposits fit our description of zeros.
We introduced Treasuries and Eurodollars in Chapter 2. Although derivatives on the
Treasuries ruled the markets during the early days (1970s and 1980s), derivatives on
Eurodollars took over thereafter, and became the dominant market with the greater
trading volume. Moreover, IRDs on Eurodollars are often simpler than IRDs on
Treasuries.
Recall that Eurodollars are dollar deposits in European banks (or European
subsidiaries of US banks) that earn interest in dollars. They were concocted during
the 1950s to evade the jurisdiction (and thus regulatory controls) of the US
government. Moreover, unlike Treasury securities, for which large auctions affect the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
available supply, the larger supply of dollars changes less frequently and is less “lumpy.”
Less regulation and a less managed supply enable Eurodollar rates to adjust more
freely to market forces than do Treasuries. This makes Eurodollar markets attractive
to hedgers and speculators. These factors have contributed to the tremendous growth
of Eurodollar markets, which are measured in trillions of dollars.
Recall from Chapter 2 that there are three major interest rates relevant to
Eurodollars. First is the Eurodollar rate, which is the interest rate paid on Eurodollar
deposits. Second is the generic London Interbank Offer Rate, which is popularly known
as libor. It is the interest rate at which major London banks offer to lend funds (in
different currencies) to each other. Libor quotes on Eurodollars are among the most
popular rates in the interest rate markets. Third is an international interest rate index
based on average libor rates. The International Continental Exchange (ICE) gets
quotes from numerous banks as to their offer rates on Eurodollar deposits of all
maturities ranging from one day to one year. For each maturity deposit, some of
the highest and lowest quotes are dropped, and the rest are averaged to compute
a libor rate index. Such indexes are computed for many currencies, with multiple
maturities for each.
This libor rate index for Eurodollars is the most widely used benchmark or
reference rate for short-term interest rates worldwide. As it has some credit risk,
500 CHAPTER 21: YIELDS AND FORWARD RATES
this libor index rate exceeds the corresponding rate on a similar maturity Treasury
security, the difference being known as the Treasury–Eurodollar (TED) spread.
Because an index is harder for any single financial institution to manipulate, this libor
rate index is used as the reference rate for over-the-counter interest rate derivatives
contracts.2 Like Treasury securities, Eurodollar interest rate computations assume
360 days in a year.
Yields and the Yield Curve
In Chapter 2, we introduced zero-coupon bonds for the primary purpose of moving
cash across time. Now we employ zeros for extracting the necessary interest rates. We
only consider zero-coupon bonds that are default-free. Extensions of these insights
to risky bonds are discussed briefly in Chapter 26.
For simplicity, we denote time as progressing in unit increments t = 0, 1, 2,
3, . . . Of course, in practice, one needs to replace the unit increment with the actual
days between the relevant cash flows.
From now on, we also need to attach time notations to our familiar symbols. For
example, the price at time t of a zero-coupon bond that pays a dollar at time T is
denoted B(t, T). A zero-coupon bond’s price at time 0 is denoted B(0, T). This
notation is illustrated in Figure 21.1.
The yield (or yield to maturity or internal rate of return) on a zero-coupon
bond is that interest rate that equates the discounted bond’s cash flows to its current
price. It is the interest rate earned each period if the zero-coupon bond is held until
maturity. The yield of a default-free zero is sometimes called the spot rate, although
this terminology is fading in usage.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
FIGURE 21.1: A Zero-Coupon Bond’s Cash Flow and
Prices
Time 0 1 2 … t … T
B(0, T) B(t, T) $1
B(t, T) is time t price of a zero-coupon bond that pays a dollar at time T.
2
However, the libor index was nonetheless manipulated during the financial crisis of 2007.
YIELDS 501
In symbols, the yield R(0, T) at time 0 (today) is defined by the equation
1
B (0, T) = (21.1)
[1 + R(0, T)]T
Notice that yields generalize the concept of a dollar return used before. Example 21.1
shows how to compute yields and create a yield curve that plots the zero-coupon
bonds’ yields (on the vertical axis) as a function of maturity (on the horizontal axis).
This example also introduces the Common Zero-Coupon Bond Price Data utilized
in this and subsequent chapters.
EXAMPLE 21.1: Computing the Yield Curve
■ Today is time 0. Consider zero-coupon bond prices with different maturity dates time T denoted
by B(0, T), where T = 1, 2, 3, 4. The common zero-coupon bond price data are shown in
Table 21.1.
- The first column reports the various times to maturity T = 1, 2, 3, 4. The longest-term zero-
coupon bond matures at time 4.
- The second and third columns give today’s zero-coupon bond prices: B(0, 1) = $0.97,
B(0, 2) = $0.93, and so on.
■ These numbers are used to compute the yields in the fourth and fifth columns. A rearrangement of
Equation 21.1 gives the yield as
1
R (0, T) = − 1.
[B(0, T)]1/T
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- For example, the yield on the two-period zero is R(0, 2) = 1/[B(0,2)]1/2 – 1 = 0.0370.
- A plot of these yields against maturities gives the yield curve. In this case, it is also called the
zero-coupon yield curve (or zero curve) (see Figure 21.2).
TABLE 21.1: The Common Zero-Coupon Bond Price Data and
Yields
Times to Maturity T Today’s Zero-Coupon Bond Prices, B(0,T) Yields, R(0,T)
1 B(0, 1) 0.97 R(0, 1) 0.0309
2 B(0, 2) 0.93 R(0, 2) 0.0370
3 B(0, 3) 0.88 R(0, 3) 0.0435
4 B(0, 4) 0.82 R(0, 4) 0.0509
502 CHAPTER 21: YIELDS AND FORWARD RATES
FIGURE 21.2: Zero-Coupon Yield Curve
Yields
0.06
0.05
0.04
0.03
0.02
0.01
0.00
1 2 3 4
Time to maturity
Several concepts are related to yields.
■ Standing at time t, the yield on the shortest-maturity zero-coupon bond maturing
at time (t + 1) has a special name: the spot rate of interest (or the spot rate).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Using our notation for the yield, this is R(t, t + 1), but for simplicity, we write it
as R(t). We emphasize that whenever the second argument of the yield notation is
missing, it will always denote the spot rate of interest.
■ For future reference, we note that a riskless money market account earns the spot rate
of interest. Standing at time t, a money market account’s return is known over the
next time step [t, t + 1]. Of course, we do not know at time t what the time
(t + 1)’s or any future time period’s spot rate will be. Future spot rates of interest
are random.
■ The graph of the yield curve is called the term structure of interest rates.
Figure 21.3 graphs the yield curve at five different dates for Treasury secu-
rities. Note that the yield curve can be flat or upward or downward
sloping. The shape of the yield curve reflects various factors such as
(1) expectations of the market participants regarding the future movement of
interest rates, (2) risk aversion of traders, and (3) the supply and demand
considerations of the underlying zero-coupon bonds. The hypothesis that
the shape of the yield curves only reflects expectations regarding the future
movements of interest rates is known as the expectations hypothesis.
Extension 21.2 provides a brief discussion of the expectations hypothesis as well
as several other competing theories explaining the term structure of interest
rates.
YIELDS 503
FIGURE 21.3: Yield Curves for Treasury Securities
at Selected Dates
Yields
18%
16%
14%
12% 1981-May
10% 1981-Aug
1987-Nov
8% 1995-Nov
6% 1998-Sept
4%
2%
0%
0.08333
0.25
0.5
1
2
3
4
5
10
15
20
25
30
Time to maturity
Consider a default-free coupon bond that trades in the market at time 0. The yield
(or yield to maturity, also called the internal rate of return) on the coupon bond
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
is the interest rate that equates the discounted coupon bond’s cash flows to the bond’s
current price PT . In symbols, the yield y satisfies
C C C C+L
PT = + +…+ +
(1 + y) (1 + y)
2 (1 + y) T−1 (1 + y)T
T (21.2)
C L
= +
∑ (1 + y)t (1 + y)T
t =1
where C denotes the yearly coupon, L denotes the principal (or par or face value),
and T denotes the time to maturity in years.3
One can interpret the yield as the interest rate earned per period if (1) one invests
today in a coupon bond that matures at time T, (2) one holds the bond until maturity,
and (3) all the coupons received get reinvested (in the future) at today’s yield y. Of
course, of these three conditions, the third is the most unreasonable because future
interest rates are random, and they will be unlikely to equal today’s yield.
3
We assume annual coupon payments for simplicity. When coupons are paid every six months, the
formula is modified with C/2 replacing C, the equation in the denominator becoming (1 + y/2), and
the annual yield is written as (1 + y/2)2 – 1.
504 CHAPTER 21: YIELDS AND FORWARD RATES
EXTENSION 21.2: The Expectations Hypothesis
The classical (macro) economics term structure of interest rate literature characterizes the risk premium embedded
in the term structure of interest rates according to hypotheses about investor behavior: the market segmentation
hypothesis, the liquidity preference hypothesis, and the expectations hypothesis. Recall that risk premia are the
compensation (in terms of an increase in expected returns above the spot rate) paid for the risks embedded in
different securities. Given an understanding of risk premia, one can both price zero-coupon bonds and forecast
future spot rates of interest.
The market segmentation hypothesis, which is normally associated with Professor John M. Culbertson
(1957), states that the different maturity equilibrium zero-coupon bond prices are determined in isolation of each
other by distinct market clienteles demanding payoffs at different horizons. Thus there is market segmentation
across different future time horizons. This implies that the risk premia paid for bonds of different maturities are
determined in isolation of the risk premia paid for the other maturity bonds, and that there might be arbitrage
across market segments.
In contrast, Nobel laureate economist Sir John Richard Hicks’s (1946) liquidity preference hypothesis
is that the equilibrium zero-coupon bond prices of different maturities are jointly determined by supply and
demand considerations, in particular, liquidity needs across time, and the interest rate risk premia reflect these
considerations. Professors Franco Modigliani and Richard Sutch’s (1966) preferred habitat theory argues
that borrowers and lenders have their own preferred maturity habitats (as suggested by market segmentations
hypothesis) and would only move to another habitat if they were given an increased premium to compensate for
the risk and cost of moving. Modern equilibrium models of the term structure of interest rates can be viewed as
the representation of the liquidity preference (or preferred habitat) hypothesis in a continuous time framework
(see Cox et al. 1985; Dunn and Singleton 1986; Sundaresan 1984).
The expectations hypothesis is associated with economists Irving Fisher (1930) and Friedrich A. Lutz (1940).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
It has as its basic premise that zero-coupon bonds of different maturities are perfect substitutes. In the literature,
this perfect substitutes hypothesis has been formalized (in terms of the mathematics) in three distinct ways. Each
of these formalizations can be characterized by a formula for the zero-coupon bond’s price. Jarrow (1981) and
Cox et al. (1981) independently discovered this critique of the various expectations hypotheses. The following
discussion is presented for the general case in terms of today (time 0), an intermediate date (time t), and the
maturity date for a long-term bond (time T).
The local expectations (LE) hypothesis is that
B (t + 1, T) − B (t, T)
Et = R (t) (1a)
[ B (t, T) ]
where Et [.] denotes expectation, B(t, T) is the time t price of a zero-coupon bond that pays a dollar at maturity
date T, and R(t) is the spot rate at time t (which is the one-period rate for immediate borrowing or lending; see
Equation 21.1).
This formalization interprets “perfect substitutes” as meaning that the different maturity zero-coupon bonds
earn no risk premia, that is, the expected return on all zero-coupon bonds is equal to the spot rate of interest.
YIELDS 505
Equivalently, in terms of bond prices, the LE hypothesis can be written as:
1
B (0, T) = E (1b)
{ [1 + R (0)] [1 + R (1)] … [1 + R (T − 1)] }
This implies, as in Equation 1a, that the T-maturity bond’s price is its expected discounted payoff using the spot
rate of interest with no adjustment for risk.
The return-to-maturity expectations (RME) hypothesis is that
1
− 1 = E {[1 + R (0)] [1 + R (1)] … [1 + R (T − 1)]} − 1. (2a)
B (0, T)
This formalization of the “perfect substitutes” condition is that the T-period holding period return from a zero-
coupon bond (the left side of Equation 2a) is equivalent to the expected return from holding the money market
account over the same horizon (the right side of Equation 2a). To understand this second assertion, looking at
the right side of Equation 2a, the first term corresponds to the expected interest earned by investing a dollar in
a money market account at time 0 and letting it roll over until time T. Interest is earned on interest. Subtracting
the original dollar gives us the expected percentage return on the money market account over [0, T].
Equivalently, in terms of bond prices, the RME hypothesis can be written as
1
B (0, T) = (2b)
E {[1 + R (0)] [1 + R (1)] … [1 + R (T − 1)]}
It can be shown that the LE and RME hypotheses are mutually exclusive (unless interest rates are deterministic)
by Jensen’s inequality.
The unbiased expectations (UE) hypothesis is that
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
f (0, t) = E {R (t)} . (3a)
where f (0, t) is a one-period borrowing or lending rate that you can contract at time 0 for the time interval
[t, t + 1]; see Equation 21.9 later in the text. Here the perfect substitutes condition is formalized to mean that
forward rates provide an unbiased estimate of the expected future spot rate.
Equivalently, in terms of bond prices, the UE hypothesis can be written as
1
B (0, T) = . (3b)
[1 + R (0)] [1 + E {R (1)}] … [E {R (T − 1)}]
It can also be shown that the LE and UE hypotheses are mutually exclusive.
If the term structure of interest rates has risk premia, then all three forms of the expectations hypotheses
are not true, and the pricing of zero-coupon bonds becomes more complex. Unfortunately, empirical evidence
strongly supports the existence of interest rate risk premium and the rejection of all three forms of the expectations
hypothesis. But they are interesting hypotheses none-the-less because they help us to better understand the pricing
of zero-coupon bonds. Knowing what is not true sometimes helps us understand what is, and why!
506 CHAPTER 21: YIELDS AND FORWARD RATES
EXAMPLE 21.2: Computing a Coupon Bond’s Yield
■ Consider a two-year coupon bond with price P2 = $101.86, principal L = $100, and a yearly coupon
of C = $6. Using Equation 21.2, the coupon bond’s yield y is the solution to the equation
6 106
101.86 = + ,
(1 + y) (1 + y)
2
or, 101.86y2 + 197.72y − 10.14 = 0.
■ Solving this quadratic equation (analytically, by trial and error, or by numerical methods) gives4
y = 0.05 or 5 percent.
■ Microsoft Excel enables you to compute the yield with one formula, which has the syntax:
YIELD(settlement,maturity,rate,pr,redemption,frequency,basis).
■ The inputs are as follows:
- Settlement is the bond’s settlement date—the day from which the bond earns interest. Let us assume
it is January 15, 2012, which is input into Excel as DATE (2012,1,15).
- Maturity is the date the bond expires, or January 15, 2014.
- Rate is the annual coupon rate, or 6 percent.
- Pr is the bond’s price per $100 face value, which is $101.86.
- Redemption is the bond’s redemption value per $100 face value, which is $100.
- Frequency is the number of coupon payments per year, 1 for annual payments, 2 for semiannual
payments, 4 for quarterly payments, and so on.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- Basis is the type of day count basis to use. Here we use 0 to denote that we are assuming 30 days in
a month and 360 days in a year. Other examples of basis are actual/actual, actual/360, and so on.
The yield is given by
YIELD(DATE(2012,1,15), DATE(2014,1,15),6%,101.86,100,1,0) = 0.049999687
4 A quadratic equation ax2 + bx + c = 0 has the solutions x = [−b ± √(b2 − 4ac)]/2a. Select the positive value for y.
A special case when computing a yield should be noted. If the price of the bond
in Example 21.2 were 100 instead (equal to par), then the yield would equal the
coupon rate of 0.06. When this happens, the coupon rate is the same as the yield and
is called the par bond yield.
Yields are often used to select bonds for inclusion into an investment portfolio.
However, this approach to bond portfolio selection is problematic. Just like warnings
on cigarette cartons that alert one to the dangers of smoking, a warning needs to be
posted at this point with respect to using yields for bond portfolio management. As
noted earlier, the yield on a coupon bond is equivalent to the internal rate of return
(IRR). The IRR is a well-studied object in capital budgeting, which is the science of
choosing among investment alternatives (in our case, bonds). The recommendation
THE TRADITIONAL APPROACH 507
in early texts was to select projects with the highest IRR (read as yields). However,
in their classic Principles of Corporate Finance, Brealey and Myers (2003) debunk IRR’s
desirability and point out several pitfalls. The most critical pitfall from our point of
view is that when using IRR in making investment decisions, one is assuming that
the reinvestment rate for future cash flows is non-random and equal to the IRR. Of
course, in actual markets, interest rates are random, and this assumption is violated.
Brealey and Myers therefore recommend not using the IRR when making investment
decisions, and they provide suitable alternatives.
As a consequence, astute readers like you are also warned against using a bond’s
yield (the IRR) as the sole determinant for choosing among different bonds to hold
in an investment portfolio. We will provide suitable alternative strategies for bond
selection in Chapters 23–25, in the context of the Heath–Jarrow–Morton (HJM)
framework.
21.3 The Traditional Approach
In 1938, Frederick R. Macaulay proposed the concept of duration (also known as
Macaulay’s duration), which led to the development of the traditional approach to
interest rate risk management. We next discuss how duration and its related concept,
convexity, are used for fixed-income securities portfolio management.
Duration
First, using Equation 21.2, we write the bond’s price as
T
C L
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
P= + (21.3a)
∑ 1+y t T
t =1 ( ) (1 + y)
where y is the bond’s yield.
Now assume that there is a small change in the bond’s yield, Δy, which equals
the new yield minus the original yield. Next, use a Taylor series expansion to get an
equation for the approximate change in value of the bond’s price:5
1 C L
ΔP ≈ − ∑T + Δy
1 + y [ t =1 (1 + y)t (1 + y)T ]
(21.3b)
Δy
= −PDur
1+y
5
A Taylor series expansion allows us to write the change in the bond price as
P (y + Δy) − P (y) = (dP/dy) (Δy) + (1/2) (d2 P/dy2 ) (Δy)2 + … (21.1)
This relation is an approximation because only the first term on the right side is retained. We get a better
approximation if the second derivative is used. This takes us to the notion of “convexity” (which we
introduce later), but it also increases computational complexity.
508 CHAPTER 21: YIELDS AND FORWARD RATES
where ≈ denotes approximation and Dur is duration, where
T
1 C L
Dur = ×t+ ×T (21.4)
∑ [
P { t =1 (1 + y)t ] [ (1 + y)T ] }
A bond’s duration (Equation 21.4) has a useful economic interpretation. The
terms in the braces in Equation 21.4 are the discounted cash flows. They become
“weights” when expressed as a fraction of the bond price. These weights add up to 1,
giving Equation 21.4 the interpretation of a weighted average. In the summation, the
weights are averaging the times to the cash payments t = 1, 2, … T. Thus duration
can be interpreted as the bond’s average life. Using this interpretation, it is easy to see
that the duration of a zero-coupon bond is the same as the time to maturity because there
are no early cash flows and the equation within the braces is just the zero-coupon
bond’s price times the time to maturity. Equation 21.3b is rewritten as the following
result (see Result 21.1).
RESULT 21.1
A Modified Duration Formula for a Bond’s Price Change
For a small change in the bond’s yield to maturity y, the change in the bond’s
price is
ΔP ≈ −PDurM Δy (21.5)
where P is the original bond price,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
ΔP is the change in the bond price (Δ denotes New value – Original
value),
DurM ≡ Dur/(1 + y) is the modified duration, and
Δy is the change in the yield to maturity.
Equation 21.4 shows that when yields change, the change in the bond’s price can be
approximated as minus the change in the bond’s yield times the product of the bond’s
price and its modified duration. The negative sign in Equation 21.5 reflects the fact
that prices and yields move in opposite directions. The formula also shows that all else
constant, a bond with a higher modified duration will have a greater price change
for a given change in the yield. As such, modified duration can be interpreted as a
“risk measure” for the bond.
The next example shows how to compute duration, modified duration, and to
approximate changes in the bond’s price when yields change only slightly.
THE TRADITIONAL APPROACH 509
EXAMPLE 21.3: Using Modified Duration to Approximate a Bond’s Price
Change
■ Let a newly issued two-year coupon bond have a par value of $100, a coupon rate of 6 percent ($6),
and a yield y = 0.05 or 5 percent per year. We present the following computations in Table 21.2:
- Columns 1 and 2 record the timing and magnitude of the cash flows.
- Column 3 reports the present value of the cash flows. They are computed by dividing $6 by
(1 + y) = 1.05 and $106 by (1 + y)2 , respectively. Sum these numbers to get the bond’s price,
$101.8594. This is the same as the bond price P2 in Example 21.2.
- Column 4 is presented for computational convenience. The entries in each cell are derived by
multiplying the present value of the cash flow by its timing (column 3 times column 1). Their sum
equals $198.0045. Dividing this number by the bond price gives the duration:
198.0045
Dur = = 1.9439 years
101.8594
■ Suppose that the yield increases by 25 basis points. Discount the bond’s cash flow by the new rate 5.25
percent to get the new bond price of $101.3896. The bond price change equals
ΔP = New price − Old price
= −$0.4698
■ Using Equation 21.5 instead, we compute the bond’s price change as
ΔP ≈ −P DurM Δy
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1.9439
= −101.8594 × × 0.0025
1.05
= −$0.4714.
Comparing the two changes, the bond price actually declines to $0.4698 but the duration-based mea-
sure approximates a decline of $0.4714. The error is a little over 0.15 cents on a principal amount of $100.
TABLE 21.2: Bond Price and Duration Computation
Time t (years) Cash Flows Present Value of Cash Flows Present Value of Cash Flow × t
1 6 5.7143 5.7143
2 106 96.1451 192.2902
Sum $101.8594 (= bond price) 198.0045
Duration, Dur = 198.0045/101.8594 = 1.9439 years.
510 CHAPTER 21: YIELDS AND FORWARD RATES
Notice an analogy. Just as we had used delta to get a linear approximation to
changes in options prices for small changes in stock prices, we use modified duration
to get a linear approximation to changes in bond prices for small changes in the
yield. Both of these approaches are based on a first-order Taylor series expansion, and
just as we improved the approximation in the case of options by incorporating the
second derivative gamma, we can do the same and get a better approximation to the
price-yield curve by introducing the second derivative with respect to the bond’s
yield. This gives a price change formula that uses both duration and convexity.
Because the convexity approximation is identical to that obtained when using an
option’s gamma, we omit the details and refer the reader to the relevant discussion in
Chapter 20.6
Modified Duration Hedging
Consider a bond portfolio consisting of ni units of the coupon bond with coupon
payment Ci , maturity Ti , principal Li , and price Pi for bonds i = 1, . . . N. The time
0 value of this portfolio is
N
V= ni Pi . (21.6a)
∑
i =1
The bond portfolio’s modified duration is7
N
DurMV = wi DurMi (21.6b)
∑
i =1
ni Pi
where wi = and DurMi is the modified duration of bond i. Note that the weights
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
V
N
sum to 1, that is wi = 1.
∑
i =1
The bond portfolio’s modified duration is just the weighted average of the
individual bond’s modified durations, where the weight for bond i corresponds to
the percentage of the portfolio’s value represented by the position in that bond.
Given an existing bond portfolio, modified duration hedging is the process of
modifying the composition of the bond portfolio through buying, selling, and/or
shorting bonds such that the new portfolio has a zero-modified duration. Under
certain conditions (to be shown below), duration hedging will remove the interest
rate risk from a bond portfolio.
6
The new price change formula is ΔP ≈ −PDurM Δy + (P/2) Convexity(Δy)2 , where we define
T
1 t(t + 1)C T(T + 1)L
Convexity = + .
P(1 + y)2 {∑
t =1
[ (1 + y)t ] (1 + y)T }
7
One can derive Equation 21.6b as follows. Define y to be the yield on portfolio V, viewing V as a
“single bond” whose coupons differ across time. Take the derivative of both sides of Equation 21.6a
with respect to y. This uses Equation 21.3, which gives P as a function of an arbitrary yield y. Next,
use Equation 21.5 for each bond, including V when viewed as a “bond” by itself, and algebra to obtain
Equation 21.6b. In Equation 21.6b, each bond’s modified duration is evaluated at the portfolio’s yield y,
and not their own yields.
THE TRADITIONAL APPROACH 511
During the last few decades of the 20th century, duration hedging was a widely
used technique for managing the interest rate risk of a bond portfolio. However, it is
currently being supplanted by the more sophisticated hedging techniques based on
the HJM model.
To understand the conditions under which traditional duration hedging removes
the interest rate risk from a bond portfolio, let’s consider the simplest bond portfolio
consisting of: (1) a single coupon-bond with coupon payment C, maturity T,
principal L, and price P and (2) n zero-coupon bonds with maturity T and price
B. We note that the coupon bond has a modified duration of DurM and that the
T-period zero-coupon bond has a modified duration of T/(1 + yB ), where yB is the
yield on the zero-coupon bond.
The portfolio’s value is
V = P + nB. (21.7a)
To duration hedge we choose n such that the right side of Equation 21.7b is zero,
which implies that the modified duration of the portfolio is approximately zero,8 that
is
P nB T
DurMV ≈ ( ) DurM + ( ) = 0. (21.7b)
V V 1 + yB
Algebra gives the number of zero-coupon bonds to trade as:
P × DurM × (1 + yB )
n=− . (21.7c)
B×T
This is called the modified-duration hedge ratio.
Now let’s study the conditions under which this zero-modified duration portfolio’s
value does not change when interest rates change. First, we can write the change in
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the portfolio’s value as
ΔV = ΔP + nΔB. (21.8a)
Second, using the modified-duration approximation given in Result 21.1 for both
the coupon and zero-coupon bonds, we can rewrite this as
T
ΔV ≈ −PDurM ΔyP − nB Δy (21.8b)
1 + yB B
where ∆yP is the change in the yield on the coupon-bond P and ∆yB is the change
in the yield on the zero-coupon bond B. Last, substitution of the modified-duration
hedge ratio n from Equation 21.7c yields:
ΔV ≈ −PDurM (ΔyP − ΔyB ) . (21.8c)
8
The duration of the portfolio V is only approximately equal to zero because we replace the portfolio’s
yield y on the right side of Equation 21.7b with both the bond’s and zero-coupon bond’s yields,
respectively, when computing their durations. The hedge ratio in Equation 21.7c is the appropriate
one to remove interest rate risk, as shown below.
512 CHAPTER 21: YIELDS AND FORWARD RATES
This is the result for which we were looking! Examining this equation, we see
that the zero-modified-duration bond portfolio’s value will be neutral to changes in
interest rates if and only if the coupon- and zero-coupon bonds’ yields are equal when
interest rates change, that is ∆yP = ∆yB . This is a very strong condition. It implies that
modified-duration hedging is effective if and only if the yield curve shifts in a parallel fashion.
At this point, you can rightly sound an alarm. If you recall Figure 21.3, you see
that the yield curve does not evolve through time via parallel shifts! Consequently,
modified-duration hedging doesn’t really work, especially when the long and short
ends of the term structure of interest rates behave differently.
But modified-duration hedging has another problem as well. Recall that modified-
duration hedging only uses the first-order approximation from the Taylor series
expansion of the bond’s price as a function of the yield. The first order approximation
only works well when yield changes are small. If yield changes are large, then the
second-order approximation in the Taylor series expansion needs to be included. This
is called convexity hedging. Because convexity hedging with bonds is analogous to
gamma option with options, we omit the remaining details and refer the reader to the
relevant discussion in Chapter 20. Unfortunately, convexity hedging does not solve
the non-parallel shifting yield curve problem. For this reason, we do not recommend
its use.
Applications and Limitations
Duration, its modifications and extensions, has been widely used in the past to manage
interest rate risk.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1. Active bond portfolio management. One can modify a portfolio’s duration to bet on
the direction of future interest rate movements. For example, if one believes that
interest rates are going to fall (which would increase the value of a bond portfolio
because bond prices will rise), just increase the portfolio’s duration by eliminating
low duration bonds and buying high duration bonds. High duration bonds are
more sensitive to interest rate changes, so that an increase in the portfolio’s duration
will increase profits as interest rates fall. Do the opposite to soften the blow of an
interest rate hike.
If desired, one can construct the portfolio to be duration neutral, that is, to
have zero duration. As noted previously, making a bond portfolio duration neutral
reduces, but does not eliminate, interest rate risk from the bond portfolio.
2. Matching assets and liabilities. A typical bank has liabilities (like checking and savings
accounts) with very low duration (because customers can remove them at a
moment’s notice) and assets (like loans to consumers and businesses) with high
durations (because they are typically for longer periods). In its natural state, a bank
holds a speculative position in terms of the shape of the yield curve. A bank can
FORWARD RATES 513
reduce this vulnerability by managing the gap between the assets’ and liabilities’
durations. For example, it can increase the duration of the liabilities by encouraging
customers to invest in long-term deposits, and it can lower the duration of the
assets by issuing floating-rate loans such as adjustable-rate mortgages and lines of
credit that charge variable rates tied to short-term interest rates.
Duration-based hedging has many shortcomings that limit its efficacy. Duration
hedging strategies do not completely eliminate interest rate risk from a bond portfolio
because of the inherent weaknesses of this approach. First, it only works for small
changes in interest rates, and in reality, interest rates sometimes change by larger
magnitudes before one can rebalance the portfolio. Second, it only works for parallel
shifts of the yield curve, which assumes that all of the bonds’ yields change by the same
amount. However, this rarely happens in practice because short-term interest rates are
much more volatile than long-term interest rates (see Figure 21.3). Combined, these
two weaknesses greatly reduce the effectiveness of duration hedging.
Although some traders may still use the concept of duration to manage a bond
portfolio’s interest rate risk, duration-based hedging is on its way out. Term structure
modeling, which we develop in this and the subsequent chapters, is a far superior
method for interest rate risk management because it works for all possible evolutions
of the yield curve.
21.4 Forward Rates
“The old order changeth, yielding place to new,” reflected a dying King Arthur
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
[in Alfred Lord Tennyson’s poem Le Morte D’Arthur (1845)]. So is it with respect
to duration and traditional interest rate risk management. Newer techniques have
been developed that overcome all of duration hedging’s limitations. These techniques
consider the impact of the changing shapes of the term structure of interest rates
on bond prices. The approach is based on the HJM model discussed in Chapters
23–25.
This section introduces the concept of a forward rate, which forms the basis of
modern interest rate risk management. Forward rates are handy for several reasons.
1. Forward rates are unit-free. Rates provide a better measure to compare across bonds
than do bond prices. To see this, consider a dollar increase in two identical stocks
where one had a 2:1 split the day before. Does a dollar stock price increase convey
the same meaning after the split as it did before? The answer is no because the
benefit of a $1 increase in the stock’s price depends on its initial price. Yet a
1 percent change for both stocks either before or after the split is comparable
because it is unit-free. The same is true for forward rates when considering changes
in bond prices.
514 CHAPTER 21: YIELDS AND FORWARD RATES
2. Forward rates are more stable than bond prices. Again, that rates are unit-free becomes
relevant. Consider a stock that earns 10 percent return per year, a stable rate. Its
prices are $100 today, $110 after one year, $121 after two years, $133.10 after three
years, and so on. The price changes differ across time and are not stable; however,
the 10 percent return is constant. Similarly, zero-coupon bond prices change
across time, yet forward rates can be stable. These considerations become important
in estimating and modeling the term structure of interest rates in Chapters 23–25.
3. Versatility of forward rates. The building blocks of modern interest rate management
theory are interlinked:
Zero − coupon bond prices ↔ Yields ↔ Forward rates
where we use ↔ to indicate that given one of these quantities, the others can be
computed. Since all of these quantities are interchangeable, we select the simplest for
analysis. This takes us to forward rates.
The Definition
To understand the definition, let us first study an example. Consider two zero-coupon
bonds trading today (time 0) with maturities at times 3 and 4, and with prices
B(0, 3) = $0.88 and B(0, 4) = $0.82, respectively. Their price ratio adjusted by
subtracting 1 is B(0, 3)/B(0, 4) – 1 = 0.0732, or 7.32 percent. This adjusted
ratio captures the implicit interest earned on the four period bond between times
3 and 4. Indeed, the extra interest earned over the time period [3, 4] is the only
difference between the three and four period zero-coupon bonds.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Extending this example, we can formally define the time 0 forward rate for the
future time period [T – 1, T] as
B (0, T − 1)
f (0, T − 1) = −1 (21.9)
B (0, T)
where B(0, T– 1) and B(0, T) are today’s zero-coupon bond prices for bonds maturing
at times (T –1) and T, respectively.9
As in the example, one can understand the forward rate by looking at
Figure 21.4. This figure shows that the forward rate f (0, T – 1) isolates the implicit
interest earned on the longer maturity bond over the last time period [T – 1, T].
Example 21.4 illustrates this computation.
9
We could have written the symbol as f (t,T – 1,T) to denote the forward rate, but we do not need the
last symbol T because this is a one-period rate.
FORWARD RATES 515
FIGURE 21.4: The Forward Rate
f(0, T – 1)
Time 0 1 2 … t … (T – 1) T
B(0, T – 1) $1
B(0, T) $1
Forward rate f(0, T – 1) is the one-period rate that you can contract
at time 0 for investing a dollar over the time horizon [T – 1, T]
EXAMPLE 21.4: Computing Forward Rates
Table 21.3 uses the data from Example 21.1.
■ The first column reports the various zero-coupon bond maturities denoted by T = 0, 1, 2, 3, 4. The
second and third columns give today’s zero-coupon bond prices. B(0,0) is trivially 1.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ These numbers are used to compute the forward rates, reported in the fourth and fifth columns. By
Equation 21.9, f (0,0) = B(0,0)/B(0,1) – 1 = 0.0309. Similarly, one can compute the other rates such
as f (0,2) = [B(0,2)/B(0,3)] – 1 = 0.93/0.88 – 1 = 0.0568, and so on.
TABLE 21.3: Zeros, Forward Rates, and Yields
Maturity Today’s Zero-Coup on Bond Prices Today’s Forward Rates Yields, R(0, T)
0 B(0, 0) $1 f (0, 0) 0.0309 R(0, 1) 0.0309
1 B(0, 1) 0.97 f (0, 1) 0.0430 R(0, 2) 0.0370
2 B(0, 2) 0.93 f (0, 2) 0.0568 R(0, 3) 0.0435
3 B(0, 3) 0.88 f (0, 3) 0.0732 R(0, 4) 0.0509
4 B(0, 4) 0.82
516 CHAPTER 21: YIELDS AND FORWARD RATES
■ Figure 21.5 plots the forward rate curve that depicts the forward rates versus time to maturity. For
comparison, also shown is the zero-coupon bond yield curve (from Example 21.1).
■ We will often use the zero-coupon bond prices, forward rates, and yields to maturity from this table
in subsequent numerical examples.
FIGURE 21.5: Forward and Zero-Coupon Yield Curves
Forward Rates and Yields
0.08
0.07
0.06
0.05
Forward Rate
0.04
Yield
0.03
0.02
0.01
0
1 2 3 4
Time period
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Understanding Forward Rates
We can understand forward rates in three related ways:
1. The first interpretation follows from the definition, which is the conventional
approach in the academic literature and finance industry. Taking the ratio of two
consecutive zero-coupon bond prices gives the implicit interest rate earned over
the relevant future time period.
2. A second interpretation is as the rate one can contract today on riskless borrowing
or lending over the future time period [T – 1, T]. One can see this interpretation
using a payoff table (see Table 21.4), where the numbers from Example 21.4 are
employed to illustrate how to replicate the borrowing rate over a future time period
by trading in zero-coupon bonds today. The procedure is as follows:
a. Today is time 0 and the longest bond’s maturity date is T = 4.
b. Today short one unit of the three-period zero-coupon bond worth B(0,3) =
$0.88 and buy [B(0,3)/B(0,4)] = 0.88/0.82 = 1.0732 units of the four-period
zero-coupon bond worth B(0,4) = $0.82.
c. This portfolio has zero value today and hence zero cash flow.
d. The first cash flow occurs at time 3 and is a negative dollar. This is the same as
borrowing $1 at time 3.
FORWARD RATES 517
e. The next cash flow is at time 4, when the four-period zero-coupon bond
matures. This is a positive cash flow of [B(0,3)/B(0,4)] = $1.0732, which equals
[1 + f (0,3)] by the definition of a forward. This is equivalent to paying back
the borrowing of $1 plus interest.
To obtain the general result, replace time 4 by time T. This substitution shows
that the cash flows obtained from the zero-coupon bond portfolio in the payoff
table exactly replicate the cash flows from contracting at time 0 to risklessly
borrow a dollar over [T – 1, T].
3. The third interpretation is presented in the next section, where we show that the
forward rate corresponds to the rate quoted on a forward rate agreement contract.
TABLE 21.4: Payoff Table for a Portfolio of Zero-Coupon Bonds That
Generates the Forward Rate
Portfolio Time 0 Time T – 1 = 3 Time T= 4
Short one unit of a three-period B(0, 3) –1
zero-coupon bond worth B(0,3)
today
B(0, 3) 0.88 B(0, 3) B(0, 3)
Buy = = − B(0, 4)
B(0, 4) 0.82 B(0, 4) B(0, 4)
1.0732 units of a four-period
zero-coupon bond
worth B(0,4) today
B(0, 3)
Sum 0 –1 1.0732 = = f(0, 3) + 1
B(0, 4)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Using Forward Rates
This section shows how you can express other rates and prices in terms of forward
rates. The next chapter will show you how to do the same with swaps, and
Chapters 23–25 will reveal the critical role forward rates play in modern interest
rate derivatives pricing models. We start with an example (see Example 21.5).
EXAMPLE 21.5: Computing a Zero-Coupon Bond Price Using Forward Rates
■ Continuing with our previous example, using definition (21.9), we can write the first zero-coupon
bond’s price in terms of forward rates:
1
f (0, 0) = − 1,
B (0, 1)
(21.10a)
1
or B (0, 1) =
1 + f (0, 0)
518 CHAPTER 21: YIELDS AND FORWARD RATES
where f (0, 0) is also the spot rate of interest.
■ Considering a forward rate that starts one period later, at time 1, we get
B (0, 1)
f (0, 1) = − 1,
B (0, 2)
1 1 1 (21.10b)
or B (0, 2) = B (0, 1) =
1 + f (0, 1) [1 + f (0, 0)] [1 + f (0, 1)]
where the last term on the right uses the value of B(0,1) from Equation 21.10a.
■ We can generalize by continuing these successive substitutions to derive an equation for a longer-
maturity zero-coupon bond.
The generalization of this example is our next result (see Result 21.2).
RESULT 21.2
A Zero-Coupon Bond’s Price and Yield in Terms of
Forward Rates
1
B (0, T) = (21.11a)
[1 + f (0, 0)] [1 + f (0, 1) … [1 + f (0, T − 1)]
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
R (0, T) = ([1 + f (0, 0)] [1 + f (0, 1)]
1 (21.11b)
… [1 + f (0, T − 1)]) T − 1
where B(0,T) is the time 0 price of a zero-coupon bond maturing at
time T,
R(0,T) is the zero-coupon bond’s yield to maturity, and
f (0, t) represents the forward rate contracted at time 0 for [t, t + 1].
Equation 21.11a expresses a zero-coupon bond’s price B(0, T) as the discounted
value of a dollar received at time T, where the discounting is done by the relevant
forward rates. Equation 21.11b states that a zero coupon bond’s yield is the geometric
average of the corresponding forward rates minus 1. This result follows readily from
the definition of a zero coupon bond’s yield as given in Equation 21.1 at the start
of this chapter. Equation 21.11b also provides an additional insight: the spot rate of
interest (denoted by R[0]) is the forward rate for immediate borrowing and lending
for one period, that is,
R (0) = R (0, 1) = f (0, 0) . (21.12)
FORWARD RATES 519
Although zero-coupon bonds are easier to work with, they don’t often trade. In
this case one can determine forward rates from coupon bond prices. This method is
demonstrated in Extension 21.3.
The concepts of forward rates and yields are useful for understanding the basic
interest rate derivatives, forward rate agreements, and interest rate futures, which are
discussed in the next section.
EXTENSION 21.3: Computing Forward Rates from Coupon Bond Prices
Computing the forward rates from coupon bond prices is a simple exercise in solving a system of linear equations
for a collection of unknowns. First, we need the following no-arbitrage relation for a coupon bond.
Consider a coupon bond with price PT , principal L, maturity date T, and a coupon payment C paid every
period. To avoid arbitrage, in a frictionless and competitive market, it must be true that
PT = CB (0, 1) + CB (0, 2) + … + CB (0, T − 1) + (C + L) B (0, T) . (1)
This follows by noting that the cash flows from a portfolio of zeros consisting of C units of B(0,1), C units of
B(0,2), . . . , C units of B(0, T– 1), and (C + L) units of B(0, T) are the same as from the coupon bond. Hence
the cost of constructing this portfolio must equal the price of the traded coupon bond, or else there is an arbitrage
opportunity. The right side of Equation 1 is the cost of constructing the portfolio.
Next, suppose we are given a collection of T distinct coupon bond prices. Then, each coupon bond has an
equation similar to Equation 1. This gives us T equations in the T unknowns {B(0,1), B(0,2), . . . , B(0, T)}.
One can solve this system of linear equations using linear algebra to obtain the zero-coupon bond prices. If the
system does not have a solution, one can find the zero-coupon bond prices that minimize the sum of errors
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
squared between the right side of Equation 1 and the coupon bond prices on the left side. Finally, one can input
the zero-coupon bond prices obtained into Equation 21.9 to get the forward rates.
EXT. 21.3 EX. 1: Computing Forward Rates from Coupon Bond Prices
■ Today is time 0. Consider three US Treasury securities trading with par value L = $100.
- The one-year Treasury bill has a price of P1 = $97.
- The two-year Treasury note has a price P2 = $100.60 and pays a coupon C2 = $4 annually.
- The three-year Treasury note has a price P3 = $101.90 and pays a coupon C3 = $5 annually.
Computing Zero-Coupon Bond Prices
■ Fortunately, the first T-bill gives the zero’s price. Equation 1 gives
P1 = LB (0, 1)
97 = 100B (0, 1) (2a)
or B (0, 1) = $0.97
520 CHAPTER 21: YIELDS AND FORWARD RATES
■ Using the two-year T-note, Equation 1 gives
P2 = C2 B (0, 1) + (C2 + L) B (0, 2)
100.60 = 4 × 0.97 + (4 + 100) B (0, 2) (2b)
or B (0, 2) = (100.60 − 3.88) /104 = $0.93
■ Finally, for the three-year T-note, Equation 1 gives
P3 = C3 B (0, 1) + C3 B (0, 2) + (C3 + L) B (0, 3)
101.9 = 5 × 0.97 + 5 × 0.93 + (5 + 100) B (0, 3) (2c)
or B (0, 3) = (101.9 − 4.85 − 4.65) /105 = $0.88
Computing Forward Rates
As these zero-coupon bond prices are the same as the corresponding values in Example 21.4, we have already
computed the forward rates and reported them in Table 21.3.
Once you have determined the zero-coupon bond prices, you can easily compute their yields and plot them
against time to maturity to obtain the zero-coupon bond price curve.
21.5 The Basic Interest Rate Derivatives
Contracts
A Brief History
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Interest rate derivatives (IRDs) were created during the 1970s when the placid world
of sluggish interest rates gave way to an era of unprecedented interest rate volatility.
Recall from our discussions in Chapter 8 that this happened because oil shocks and
other supply-side disturbances hiked up inflation, making interest rates more volatile
than before. Businesses started seeking tools for managing interest rate risks, which
generated innovations in exchange-traded IRDs:
■ In 1975, the Chicago Board of Trade (CBOT) launched the first IRD, the GNMA
(Government National Mortgage Association) futures, which could be used to
hedge long-term interest rate risk.
■ In 1976, thirteen-week Treasury bill futures contracts started trading. Currently
these are the oldest exchange-traded interest rate futures contracts.
■ During the late 1980s, GNMA futures contracts became obsolete because US
Treasury bond futures, which the CBOT introduced in 1977, provided a better
hedge for the same risk.
■ In 1981, the Chicago Mercantile Exchange (CME) introduced the world’s first
cash-settled derivative, Eurodollar futures. It provided a hedge for short-term
interest rate risk, and thirteen years later, it replaced the T-bond futures as the
world’s most actively traded futures contract.
THE BASIC INTEREST RATE DERIVATIVES CONTRACTS 521
■ In 1982, the CBOT introduced options on T-bond futures, whose pricing and
hedging are discussed in Chapters 23–25.
■ In 1992, the CME launched a post-market global electronic transaction system
Globex. It has helped the growth of the IRD market, which has long been a
twenty-four-hour market owing to the influence of world events on interest rates.
Diverse IRDs trade in today’s markets. Recall that the basic equity (and commod-
ity) derivatives on an underlying stock (or commodity) are forward contracts, futures
contracts, European calls, and European puts. Analogous contracts trade for IRDs.
In this case, the underlying “asset” is the spot interest rate, which is an index and not
a traded security. The basic derivatives are the same but this time they have different
names: forward rate agreements (FRAs), interest rate futures, caplets, and floorlets.
Not all of these basic securities trade, however. Traded are over-the-counter FRAs
and exchange-traded futures. Caplets and floorlets trade only as parts of portfolios
called caps and floors, respectively. Moreover, portfolios of FRAs and zero-coupon
bonds also trade. These are interest rate swaps, which were introduced in Chapter 7.
This fact is demonstrated in the next chapter.
The next section introduces FRAs and interest rate futures contracts. We first
present FRAs, which are easy to understand, and related to forward rates. Next follow
Eurodollar futures, which are immensely popular and similar to T-bill futures (which
are discussed in Extension 21.4).
Forward Rate Agreements
A forward rate agreement is analogous to a forward contract and is sometimes simply
known as a “forward.” Just as a forward fixes a price for buying a commodity at a later
date, a forward rate agreement fixes an interest rate for borrowing (or lending) over
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
some future time period. The largest market for this over-the-counter contract is in
London, where various dealers quote FRA rates for loans denominated in eurodollars,
euribor (Euro Interbank Offered Rate, a libor-like interest rate index computed on euro
interbank term deposits), pounds sterling, Swiss francs, and Japanese yen, among
others.The next example illustrates the workings of an FRA.
EXAMPLE 21.6: A Forward Rate Agreement
■ Today is January 1 (time 0). International Treats and Restaurants Inc.’s (INTRest, a fictitious name)
treasurer plans to borrow $100 million five months from today to build a factory. She likes the current
interest rate and decides to lock it in by entering into a long position in a forward rate agreement.
■ London based Dealerbank Corp. (fictitious name) quotes on a 5 × 8, 100 million Eurodollar FRA
a rate of 5.95 percent to 6 percent. Figure 21.6 shows the workings of this FRA (where we have
simplified some market conventions related to the settlement date).
- The 5 × 8 means that the FRA begins five months from today (June 1) and terminates on the eighth
month.
522 CHAPTER 21: YIELDS AND FORWARD RATES
- The notional principal LN is $100 million. This amount never changes hands but is used for
computing interest payments. This is a Eurodollar FRA because the interest rates are computed
in terms of the libor index rates for eurodollars.
- The treasurer decides to buy the FRA from the dealer at the fFRA rate of 6 percent. This becomes
the fixed borrowing rate that the company locks in. Alternatively, if INTRest wants to loan funds,
it can sell the FRA for 5.95 percent, which becomes the effective lending rate.
■ Five months later, on June 1, the 3 month Libor rate index is 8 percent per year. The buyer of the
FRA collects a payment because this floating three-month libor rate index (known as the reference
rate) is higher than the fixed FRA rate contracted in advance.
■ The treasurer gets paid on the payment date September 1 (which is the termination date)
Notional principal × (Reference rate − fFRA ) × Time period
= 100, 000, 000 × (0.08 − 0.06) × (92/360) (21.13a)
= $511, 111.11
■ Some FRAs pay on the settlement date, which is the day the reference rate is announced (five months from
today in this example). In that case, to get the payment, discount the cash flow in Equation 21.13a
FIGURE 21.6: Timeline and Cash Flows of an FRA
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Now (start date) Settlement date Payment date
(Time 0 = January 1) (June 1) (September 1)
The company buys an FRA on LN = On June 1, the three- On September 1,
$100 million with an FRA rate fFRA = month libor rate index the company gets
6 percent per year. is 8 percent per year. $511,111.11.
using 8 percent, which is the market interest rate for this 92-day period. Then Long’s company collects
on June 1:
511, 111.11
= $500, 871.08. (21.13b)
92
(1 + 0.08 × 360 )
■ INTRest will have to pay 8 percent if it borrows money in June. The effective borrowing cost is only
6 percent because of the payment received on the FRA.
THE BASIC INTEREST RATE DERIVATIVES CONTRACTS 523
■ The protection lasts only for three months. To hedge cash flows for a longer time horizon, INTRest
can enter into a series of forward rate agreements of different maturities. The company would pay 6
percent and receive the floating reference rate for the relevant periods. Fortunately, this contract trades
in the market as a package. As you may guess, this is the interest rate swap contract that we introduced
in Chapter 7! The next chapter will formalize this intuition and show why an interest rate swap is a
collection of FRAs.
Notice the following features of an FRA:
1. The FRA uses the actual/360-day count convention. Interest is computed over the actual number of
days in the period, assuming there are 360 days in the year. FRAs on Eurodollars and Euribors follow
this convention, while FRAs on pounds sterling follow the actual/365-day count convention.
2. The payment is discounted with the realized libor rate index by using simple interest rates.
3. We have assumed that there is no risk of default on the part of the writer of the FRA, implying
that the payment contracted is the payment received. As an FRA is an over-the-counter instrument,
collateral is needed to support the transaction, which the counter-parties negotiate.
4. As with a forward contract, the payoff from an FRA may be positive, zero, or negative. Thus, an FRA
is a “bet” on the future movements of the libor rate index.
The FRA’s payoff in this example can be formalized. Consider a forward rate
agreement with notional principal LN , settlement date tSet , and payment date T. The
forward rate agreement’s payoff on the payment date (time T) is
LN × (Ref − fFRA ) (T − tset ) /360 (21.14a)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where Ref is the reference rate (the libor rate index or Euribor in most cases)
determined at time tSet that applies over the time period (T – tSet ) in days and where
fFRA is the FRA rate.
If the cash flow takes place on the day the reference rate is determined, then the
FRA payoff on the settlement date (time tSet ) is
LN × (Ref − fFRA ) (T − tset ) /360
(21.14b)
[1 + Ref × (T − tset ) /360]
where the payoff in Equation 21.14a is discounted to the settlement date by the
reference rate.
In our discrete time model, the FRA payoff at time T simplifies to
LN × [R (T − 1) − fFRA ] (21.14c)
because the reference rate is R(T – 1), the spot rate of interest at time (T – 1), and
the time period has unit length (not 360 days).
524 CHAPTER 21: YIELDS AND FORWARD RATES
Interest Rate Futures
As you may guess, an interest rate futures contract is the exchange-traded “fraternal
twin” of an FRA. This is because it’s also a contract that fixes a borrowing (or a
lending) rate at some future date. It differs from an FRA in the same way that a
futures contract differs from a forward. A host of features distinguish them such as
standardization, a clearinghouse, margins, and daily settlements. Some interest rate
futures (like those on Treasuries) end in physical delivery, whereas others (like those
on Eurodollars) end in cash settlement.
Futures prices get quoted the same way as the underlying security’s prices are
quoted. Consider some highly popular interest rate futures contracts that trade at
the CME Group’s exchanges. Futures prices for Treasury bonds, notes, and STRIPS
are quoted in 32nds, with $100 as the par value. By going long a futures contract
on one of these securities, you are promising to buy the underlying bond at a fixed
price at a future date. However, Eurodollar futures, like Treasury bill futures, follow
a different quoting convention. We discuss this in the context of Eurodollar futures,
which are free from the complications related to the embedded options present in the
physical delivery feature of T-bill futures contracts (Treasury futures are discussed in
Extension 21.4).
Eurodollar futures contracts’ payoffs are based on the three-month libor rate index
at the future delivery date of the contract, but instead of directly quoting futures
interest rates, Eurodollar futures contracts are quoted in terms of futures prices,
which are defined as
Quoted futures price = 100 − Futures interest rate. (21.15)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
For example, a quoted futures price of 94 implies a futures interest rate of
(100 – 94) = 6 percent.
Notice the following features of Eurodollar futures.
1. Obviously, dividing the futures interest rate by 100 gives the interest rate as
a decimal, 0.06 in this case. The decimal point format is used in subsequent
computations.
2. Ninety-four is the quoted futures price for a Eurodollar deposit with $100 par
value paid three months after the contract matures. The quoted Eurodollar futures
price, a number between 0 and 100, sets the market value of a newly created
Eurodollar futures contract to zero.
3. This method of quoting futures prices has yet another benefit: it restores the
familiar relation between ask and bid prices seen in other markets. Recall that
in most markets the ask price (which is the buying price for the customer, abc)
is higher than the bid price (which is the selling price of the customer). The
same relationship holds in the interest rate market in term of prices: the ask price
paid to buy a bond is higher than the bid price at which it is sold. However,
bonds usually get quoted in terms of an interest rate: the yield. Due to an inverse
relationship between a bond’s price and its yield, the ask yield rate is lower than
the bid yield rate. Similarly, the quoted futures price, as Example 21.7 illustrates,
places the quoted ask futures price higher than the quoted bid futures price (see
Figure 21.7), and the ask futures interest rate is lower than the bid rate.
THE BASIC INTEREST RATE DERIVATIVES CONTRACTS 525
EXAMPLE 21.7: A Eurodollar Interest Rate Futures Contract
Eurodollar Futures Contract
■ Suppose today is January 1, and the June Eurodollar futures has a quoted ask price of 94.00 and a
quoted bid price of 93.99. This means the ask futures interest rate is 100 – 94 = 6 percent per annum,
while the bid futures interest rate is 100 – 93.99 = 6.01 percent (see Figure 21.7). The underlying asset
is a 90-day 1 million Eurodollar three-month deposit. One million dollars is the notional amount,
and the contract is cash-settled. The futures interest rate is always computed over a 90-day period (see
Table 21.5 for the CME Group–traded Eurodollar futures contract’s specifications).
■ It is important to note that the quoted futures price (94 for ask) is the convention that the market uses
to determine the futures interest rate. Here, the ask futures interest rate is 6 percent or ifut = 0.06 in
decimal terms.
■ The futures interest rate is used to determine the payments to the futures contract. The time t quantity
that determines the payments to the futures contract is given by the formula
90
1, 000, 000 × (1 − ifut × (21.16a)
360 )
where ifut is the futures interest rate expressed as a decimal. For convenience, we call this quantity the
futures-contract-price.
■ The difference between the quoted futures price and the futures-contract-price is an adjustment for
the notional amount (from 100 to 1 million dollars) and an adjustment for the three-month time
period over which the simple interest rate applies (note in Equation 21.15 that the quoted futures
price is implicitly for a one-year time period).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ The ask futures-contract-price is 1,000,000 × [1 – 0.0600 × (90/360)] = $985,000, and the bid
futures-contract-price is 1,000,000 × [1 – 0.0601 × (90/360)] = $984,975 (see Figure 21.7). Notice
that if the quoted futures price goes down by one basis point, or 0.0001, then the Eurodollar futures-
contract-price goes down by $25.
■ The futures-contract-price on the delivery date is
90
1, 000, 000 × (1 − i × (21.16b)
360 )
where i is the Libor spot interest rate index at the delivery date expressed as a simple interest rate.
Note that on the delivery date, we get convergence of the futures interest rate to the spot rate.
■ The futures-contract-price on the delivery date represents the value on that date of a Eurodollar
deposit paying a million dollars in three months time. This computation uses a simple interest rate
based on the 360 day year convention. For example, if i = 0.03 on the delivery date, then the futures
contract price is 1,000,000 × [1 – 0.03 × (90/360)] = $992,500. This is the purchase price on the
delivery date of a Eurodollar deposit that pays one million dollars after three months.
526 CHAPTER 21: YIELDS AND FORWARD RATES
FIGURE 21.7: Futures Price, Futures Rate, and Futures-
Contract-Price for Eurodollar Futures
Ask futures-contract-price = $985,000
Ask futures price = $94
Bid futures rate = 6.01 percent
Ask futures rate = 6 percent
Bid futures price = $93.99
Bid futures-contract-price = $984,975
Hedging Example
■ Let us tailor this contract to meet INTRest’s hedging needs as enumerated in the previous example:
- INTRest needs one hundred contracts to match the $100 million.
- INTRest must sell Eurodollar interest rate futures contracts to fix the borrowing cost: if the interest
rate increases, the futures price declines, the futures generates a negative payment, and being short
INTRest receives a positive cash payment.
- For example, if the futures price for the ask declines to 91.99 (which corresponds to a futures interest
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
rate of 100 – 91.99 = 8.01 percent), then the new futures-contract-price is 1,000,000 × [1 – 0.0801
× (90/360)] = $979,975.
■ Being short one hundred contracts, INTRest has a cash flow of
−100 × [New contract value (for ask) − Old contract value (for bid)]
= −100 × (979, 975 − 984, 975)
= $500, 000
The contract’s performance is similar to that of the FRA contract (see payoff in Equation 21.13b).
Margins
■ We have considered the bid/ask spread but neglected the other transactions costs (such as brokerage
commissions) as well interest earned or paid on the margin account. Margin requirements for
Eurodollar futures are complex. They depend on the trading strategy as well as other factors, and
speculators have higher margin requirements than do hedgers or exchange members.
■ January 1
- Ms. Long buys one June Eurodollar contract at a futures price of 94. As noted earlier, this
corresponds to a futures-contract-price of $985,000.
THE BASIC INTEREST RATE DERIVATIVES CONTRACTS 527
- No money changes hands today. Long keeps $1,000 initial margin with her broker. Her maintenance
margin is $750 per contract.
■ January 2
- Settlement for the futures price is 94.05. The futures interest rate is 100 – 94.05 = 5.95 percent or
0.0595 as a decimal. The futures-contract-price is
1,000,000 × [1 – 0.0595 × (90/360)] = $985,125.
- The margin account adjustment
= Futures-contract-price (Jan 2) – Futures-contract-price (Jan 1)
= 985,125 – 985,000 = $125. Her margin account balance is 1,000 + 125 = $1,125.
- Notice that with the futures price increasing, the long position has gained value.
■ January 3
- Settlement for the futures price is 93.75. The futures-contract-price is
1,000,000 × [1 – 0.0625 × (90/360)] = $984,375.
- The margin account adjustment = 984,375 – 985,125 = – $750.
- As her margin account’s value 1,125 + (–750) = $375 has fallen below the maintenance margin
level, she gets a margin call from her broker to deposit (1,000 – 375) = $625 or more in cash.
TABLE 21.5: Eurodollar Futures Contract Specifications
Specification Description
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Underlying instrument Eurodollar time deposit having a principal value of $1
million with a three-month maturity
Tick size (minimum fluctuation) One-quarter of one basis point (0.0025 = $6.25 per
contract) in the nearest expiring contract month;
one-half of one basis point (0.005 = $12.50 per contract) in
all other contract months
Contract months March, June, September, December, extending out ten
years (total of forty contracts) plus the four nearest serial
expirations (months that are not in the March quarterly
cycle). The new contract month terminating ten years in
the future is listed on the Tuesday following expiration of
the front quarterly contract month
Last trading day The second London bank business day prior to the third
Wednesday of the contract expiry month; trading in the
expiring contract closes at 11:00 am London time on the
last trading day
(Continued )
528 CHAPTER 21: YIELDS AND FORWARD RATES
TABLE 21.5: (Continued )
Specification Description
Final settlement Expiring contracts are cash settled to 100 minus the
three-month US dollar libor rate index on the last trading
day; final settlement price will be rounded to four decimal
places, equal to 1/10,000 of a percent or $0.25 per contract
Position limits None
Trading hours (Central Time) Open outcry, Monday–Friday: 7:20 am – 2:00 pm;
CME GLOBEX, Sunday–Friday: 5:00 pm – 4:00 pm
Source: www. cmegroup. com/trading/interest-rates/stir/eurodollar_contract_specifications.html
The Eurodollar’s futures-contract-price in the preceding example can be formal-
ized. Consider a Eurodollar futures contract at time t with delivery at time T and
notional value LN = $1 million on a three-month Eurodollar deposit. The futures-
contract-price of the Eurodollar futures at time t is given by the formula
90
F (t) = LN × [1 − ifut (t) × (21.16c)
360 ]
where F(t) is the time t futures-contract-price and ifut (t) is the time t futures interest
rate expressed as a decimal. Note that the quoted futures price does not explicitly
appear in this equation. The futures price is just the convention used by the market
to quote the futures interest rate.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Marking to market over the time period [t, t + 1] is given by the change in the
futures-contract-prices, that is
90
F (t + 1) − F (t) = −LN × [ifut (t + 1) − ifut (t)] × . (21.16d)
360
As the futures interest rate increases, the futures-contract-price declines, as indicated
by the minus sign in this equation. This is intuitive because futures-contract-prices
and futures interest rates move inversely to each other.
The settlement futures-contract-price for the futures contract on the delivery
date is
90
F (T) = LN × [1 − i (T) × (21.16e)
360 ]
where F(T) is the time T futures-contract-price and i(T) is the 90-day libor rate
index spot rate at time T expressed as a simple interest rate.
As we approach the delivery date, note that we get convergence of the futures
interest rate ifut (t) to the spot rate of interest; that is, as t goes to T, ifut (t) approaches
i(T).
THE BASIC INTEREST RATE DERIVATIVES CONTRACTS 529
In our discrete time model, setting the notional value of the Eurodollar futures
contract to a dollar, the futures-contract-prices at time t and the delivery date time
T simplify to
F (t) = [1 − ifut (t)] (21.16f)
and
F (T) = [1 − R (T)] (21.16g)
because the time period has unit length, and the spot rate of interest (see Equation
21.12) is also the simple interest in this case.
We note that in this discrete time case, the quoted futures price is the futures-
contract-price expressed as a percentage (multiplied by 100). For our subsequent
modeling, therefore, the distinction between the quoted futures price and the futures-
contract-price is no longer needed.
EXTENSION 21.4: Treasury Futures
Treasury Bill Futures Price
Introduced in 1976, the 13-week Treasury bill futures contract is the oldest exchange-traded interest rate
derivative. A highly popular contract for hedging and speculating on short-term interest rate risk, it shares many
common features with the Eurodollar futures contract: both contracts are written on a notional principal of
$1 million, both assume that interest applies over 90 days, both use 360 days in a year, and both are structured so
that a one–basis-point movement in the index translates into a $25 change in the futures-contract-price. The key
difference is that a T-bill futures contract has a T-bill as the underlying asset, which must be delivered physically
if the short delivers, while a Eurodollar futures contract is cash-settled with no future delivery requirement.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
To see how a 13-week T-bill futures works, let us start with an ask futures price of $94.00. Similar to the
Eurodollar futures (see Example 21.7), this translates into an interest rate of 100 – 94 = 6 percent per year. This is
the banker’s discount yield that you saw in Equation 2.7b. Using this interest rate as the ask, Equation 2.7b gives
an ask futures-contract-price for the bill with a face value of $1 million and 90 days to maturity:
T
PAsk = 1, 000, 000B = 1, 000, 000 1 − (banker’s discount yield)
[ ( 360 )]
90
= 1, 000, 000 (1 − 0.06 ×
360 )
= 985, 000
If the quoted futures price is $94.01, then the banker’s discount yield is 5.99 percent, and the bond’s ask futures-
contract-price is $985,025. Similar to the Eurodollar futures, a rise in the futures price by one basis point is
associated with a $25 increase in the 13-week Treasury bill futures-contract-price, a gain that will accrue to the
buyer of the T-bill futures.
The physical delivery requirement makes the contract more cumbersome than Eurodollar futures. At the
maturity of the T-bill futures, a newly issued T-bill becomes available with 13 weeks to maturity. The CME
Group allows the substitution of 91-day and 92-day T-bills with appropriate price adjustments to enhance the
liquidity of the deliverable security.
530 CHAPTER 21: YIELDS AND FORWARD RATES
Treasury Note and Bond Futures
Longer-maturity Treasury futures work similar to T-bill futures, but they are quoted differently. Treasury bonds,
notes, and STRIPS are quoted in 32nds with $100 as the par value. By going long a futures contract on one of
these securities, you are promising to buy the underlying bond at a fixed price at a future date.
As with T-bill futures, the short initiates the delivery and decides on which bond to deliver. The contracts
allow substitution among several possible bonds with appropriate price adjustment via conversion factors. This
gives the short a valuable embedded option in the form of providing the cheapest-to-deliver bond. Moreover,
the short has another “option,” known as the wild card option, because the T-bond futures stop trading at
2:00 pm. (Central Time), while bonds trade until 4:00 pm and the short has until 8:00 pm to issue the notice
of intention to deliver. These embedded options significantly complicate the valuation of Treasury futures
contracts.
The Equivalence between Forward and FRA Rates
Both forward rates and forward rate agreements lock in an interest rate for borrowing
or lending starting at some future date. Is there a link between the two? The answer
is yes, they are equal. Let us illustrate this with a simple example (see Example 21.8).
EXAMPLE 21.8: Forward and FRA Rates
■ Using the assumption of no arbitrage, we will determine the FRA rate and demonstrate that it is
equivalent to the forward rate.
- Consider a forward rate agreement written at time 0 that starts at time 3 and matures at time 4.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- Denote the FRA rate by fFRA (0,3) ≡ fFRA . The contract is on a notional principal of LN = $100
million.
- The FRA contract is written on the spot interest rate’s value at time 3, R(3). This spot rate becomes
known at time 3 and is effective over the time interval [3,4] (see Equation 21.14c).
At the FRA’s Maturity Date, Time 4
■ The cash payoff at the FRA’s maturity date is
[R (3) − fFRA ] 100
(21.17a)
= [1 + R (3) − (1 + fFRA )] 100
where we rewrite the payoff by adding and subtracting 1.
At Time 3
■ Moving back one period and computing the time 3 present value gives
[1 − (1 + fFRA ) B (3, 4)] 100. (21.17b)
THE BASIC INTEREST RATE DERIVATIVES CONTRACTS 531
The first term is 1 because we get [1 + R(3)] at time 4 by investing a dollar in a money market account
at time 3, and the second term is obtained by multiplying (1 + fFRA ) at time 4 by the zero-coupon
bond price B(3,4) to get its present value at time 3.
At Time 0, When the FRA is Negotiated
■ To compute the time 0 present value, note two facts (see Figure 21.1): the time 0 present value of a
dollar at time 3 is B(0,3); similarly, the time 0 present value of B(3,4) is B(0,4). The present value of
the preceding cash flow is therefore
[B (0, 3) − (1 + fFRA ) B (0, 4)] 100 (21.17c)
■ As with a conventional forward contract, the FRA rate sets the value of the FRA to zero on the day
it is written (time 0). Equating Equation 21.17c to zero and solving for the FRA rate gives:
[B (0, 3) − (1 + fFRA ) B (0, 4)] 100 = 0 (21.17d)
or fFRA (0, 3) ≡ fFRA = [(B (0, 3) /B (0, 4)] − 1 ≡ f (0, 3)
where the last term follows from the definition of the forward rate as given in Equation 21.9.
The argument in the previous example works more generally. Replacing time 4
by T in Equation 21.17d yields the following result (see Result 21.3).
RESULT 21.3
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The FRA Rate and Forward Rates
Consider an FRA with notional LN , maturity date T, and settlement date
T – 1 in the discrete time model. Then,
the FRA rate, fFRA (0, T − 1) = f (0, T − 1) , the forward rate (21.18)
This result is just as it should be. Recall from the previous section that the forward
rate with maturity (T – 1) is that rate one can contract at time 0 for riskless borrowing
and lending over [T – 1, T]. In contracting, the borrower is betting that the spot rate
of interest will be greater than the forward rate—but this is the same bet made when
going long an FRA. When going long an FRA, one is betting that the reference rate,
here the spot rate of interest, is greater than the FRA rate. Therefore, the two rates
must be the same. This is a powerful insight: the abstract forward rate that we had
defined earlier is the same as the interest rate that emanates from the market-traded
forward rate agreement contract. The power of forward rates is further illustrated in
the context of other derivatives in the next three chapters.
532 CHAPTER 21: YIELDS AND FORWARD RATES
21.6 Summary
1. The discussion of interest rate derivatives (IRDs) parallels the earlier presentation
of spots, forwards, futures, and options on commodities and equities. For IRDs,
an array of different maturity default-free zero-coupon bonds replace the single
commodity or stock price as the underlying asset. Yields and forward rates
extracted from the zeros are foundational concepts. We focus on Eurodollar
IRDs, which are simpler than derivatives on Treasury securities and attract the
greatest trading volume.
2. Eurodollars are dollar deposits in European banks (or a European subsidiary of
a US bank) that earn interest in dollars. There are three major interest rates for
Eurodollars: the Eurodollar interest rate paid on Eurodollar deposits, the generic
libor at which London banks offer to lend dollars to each other, and a libor rate
index (which is an average of libor quotes collected from major London-based
banks.
3. Assuming a discrete framework in which time progresses in unit increments
t = 0, 1, 2, 3, . . . , the yield (or yield to maturity or internal rate of return) on
a zero-coupon bond is the interest rate that equates the discounted bond’s cash
flows to its current price:
1
B (0, T) =
[1 + R(0, T)]T
where B(0, T) is the time t price of a zero-coupon bond that pays a dollar at time
T, and R(0, T) is the yield at time t on a zero-coupon bond with maturity T.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
4. Standing at time t, the yield on the shortest-maturity zero-coupon bond that
matures at time (t + 1) has a special name—the spot rate of interest R(t). A
riskless money market account earns the spot rate of interest each period.
5. The yield curve plots the zero-coupon bonds’ yields as a function of each bond’s
time to maturity. This graph is called the term structure of interest rates. A yield
curve can take various shapes influenced by factors such as (1) the expectations
of the market participants regarding the future movement of interest rates, (2)
risk aversion, and (3) supply and demand considerations for the underlying zero-
coupon bonds.
6. A default-free coupon bond’s yield (or yield-to-maturity, also called the internal
rate of return) y is defined by
C C C C+L
PT = + +…+ +
(1 + y) (1 + y)2 (1 + y)T−1 (1 + y)T
T
C L
= +
∑ [ (1 + y)t ] (1 + y)T
t =1
where PT denotes the coupon bond price, C denotes the yearly coupon, L
denotes principal (or par or face value), and T denotes time to maturity in years.
SUMMARY 533
If the yield equals the coupon rate, the bond’s price is identical to the par
value, and the yield is called the par bond yield.
7. Traditional interest rate management involves computing Macaulay’s duration
and using it to approximate a bond’s price change for a given change in the
yield. Result 21.1 states that for a small change in the bond’s yield to maturity y,
the change in the bond’s price is approximately
ΔP ≈ −P DurM Δy
where P is the original bond price, ∆P is the change in bond price, DurM ≡ Dur/
(1 + y) is the modified duration, and ∆y is the change in the yield-to-maturity.
The duration is computed by
T
1 C L
Dur = t ×t+[ ×T
P {∑
t =1 [ (1 + y) ]
(1 + y)T ] }
where C is yearly coupon, L is the bond’s par (or principal) value, T is the bond’s
maturity, and y is the original yield to maturity. The price approximation can be
improved by a convexity adjustment, which uses the second-order term in a
Taylor series expansion for the price change formula.
8. Duration is a weighted average of the time periods to the receipt of the cash
flows. A bond with higher duration undergoes a greater price change given a
change in the interest rate. The duration of a zero-coupon bond is the same as
the time to maturity. A bond portfolio’s duration is a weighted average of the
durations of the individual bonds.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
9. Duration-based hedging has many shortcomings that limit its efficacy. It only
works for small changes in interest rates and for parallel shifts of the yield curve,
which means that all yields change by the same amount. The yield curve rarely
changes by parallel shifts.
10. The time 0 forward rate for the future period [T – 1, T] is defined as
B (0, T − 1)
f (0, T − 1) = −1
B (0, T)
where B(0, T– 1) and B(0, T) are today’s zero-coupon bond prices for bonds
maturing at times (T – 1) and T, respectively.
11. The zero-coupon bond price and yield can be expressed in terms of forward rates
as
1
B (0, T) =
[1 + f (0, 0)] [1 + f (0, 1)] … [1 + f (0, T − 1)]
1
R (0, T) = ([1 + f(0, 0)][1 + f(0, 1)] … [1 + f(0, T − 1)]) T − 1.
534 CHAPTER 21: YIELDS AND FORWARD RATES
where B(0, T) is the time 0 price of a zero-coupon bond maturing at time T,
R(0, T) is the zero-coupon bond’s yield to maturity, and f (0, t) represents the
forward rate contracted at time 0 for time t.
12. The spot rate of interest (denoted by R[t]) is also the forward rate for immediate
borrowing and lending for one period, that is, R(t) = f (t,t).
13. A forward rate agreement (FRA) is an over-the-counter contract that fixes the
interest rate for borrowing (or lending) over some future time period. It has
several features: (1) an FRA uses the actual/360-day count convention, (2) the
payment is discounted with the realized libor rate index using simple interest,
and (3) like a forward, the payoff from an FRA can be positive, zero, or negative.
The FRA rate is the same as the forward rate.
14. Consider an FRA with notional principal LN , settlement date tSet , and payment
date T. The payoff on the payment date is
LN × (Ref − fFRA ) (T − tSet ) /360
where Ref is the reference rate (the libor rate index or euribor in most cases)
determined at time tDec that applies over the time period (T – tSet ) in days, and
where fFRA is the FRA rate.
15. An interest rate futures is an exchange-traded contract that fixes a borrowing (or
a lending) rate at some future date. It differs from an FRA in the same way as a
futures contract differs from a forward. A host of features distinguish them such
as standardization, a clearinghouse, margins, and daily settlements.
16. Eurodollar futures prices, like Treasury bill futures, get quoted in terms of an
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
index, which is defined as
Futures price = 100 − Futures interest rate
17. The asset underlying a Eurodollars futures contract is a 90-day 1 million
Eurodollar three-month deposit. The futures interest rate is always computed
over a 90-day period, and the contract is cash-settled. The Eurodollar futures-
contract-price at time t, which determines the payments to the futures contract,
is given by the formula
90
1, 000, 000 × [1 − ifut ×
360 ]
where ifut is the futures interest rate obtained from the preceding formula for the
futures price expressed as a decimal.
18. Consider an FRA with notional LN , maturity date T, and settlement date
T – 1 in our discrete time model. Then,
the FRA rate, fFRA (0, T – 1) = f (0, T – 1), the forward rate.
QUESTIONS AND PROBLEMS 535
21.7 Cases
Pension Plan of Bethlehem Steel—2001 (Harvard Business School Case
#202088-PDF-ENG). The case analyzes the company’s pension fund assets and
liabilities by using the traditional tools of interest rate risk management.
Union Carbide Corp.: Interest Rate Risk Management (Harvard Business School
Case #294057-PDF-ENG). The case studies how the firm can manage its
exposure to interest rates by matching the duration of its liabilities to the duration
of its assets.
Deutsche Bank: Finding Relative Value Trades (Harvard Business School Case
#205059-PDF-ENG). The case considers how the bank can find yield curve trades
that are of interest to clients as well as to its proprietary trading desk.
21.8 Questions and Problems
21.1. What is the difference between a Eurodollar, a Eurodollar rate, and the US
T-bill rate?
21.2. What is libor? Explain how it is similar or not similar to a libor rate index.
21.3. Compute the yield curve for the following zero-coupon bond prices.
Times to Maturity T Today’s Zero-Coupon Bond Prices B(0, T)
1 B(0, 1) 0.98
2 B(0, 2) 0.96
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
3 B(0, 3) 0.94
4 B(0, 4) 0.90
21.4. How does the bond’s yield relate to an internal rate of return? Is the bond’s
yield a good measure for its expected return over the next year?
The next three questions use the following information.
Bondy Bond (fictitious name) is a coupon bond that matures in 10 years. The
coupon rate is 5 percent per year. The bond’s principal is $10,000 and the
market price is $8,500.
21.5. (Microsoft Excel) Compute the yield for Bondy Bond, showing how you use
the Excel program for this purpose.
21.6. Compute Bondy Bond’s duration using the yield computed in the previous
problem.
21.7. Suppose the yield increases by 0.0005 over a week.
a. Compute the new price of Bondy Bond.
536 CHAPTER 21: YIELDS AND FORWARD RATES
b. Compute the actual change in bond prices.
c. Use a duration based formula to predict the change in Bondy Bond’s price.
How does this compare with the actual change?
21.8. A coupon bond has a life of 5 years, makes semiannual coupon payments at
the rate of 6 percent per year, has a principal of $100, and a market price of
$95.
a. Using Excel, compute this bond’s yield to maturity.
b. Compute the duration of the bond.
21.9. At time 0, what is the duration of a zero-coupon bond with maturity T?
Prove your answer using the definition of duration.
21.10. What are the two problems with using duration as a risk measure for changes
in bond prices?
21.11. Compute the forward rates for the following zero-coupon bond prices.
Times to Maturity T Today’s Zero-Coupon Bond Prices B(0, T)
1 B(0, 1) 0.98
2 B(0, 2) 0.96
3 B(0, 3) 0.94
4 B(0, 4) 0.90
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
21.12. Compute the forward rates and yields for the following zero-coupon bond
prices.
Times to Maturity T Today’s Zero-Coupon Bond Prices B(0, T)
1 B(0, 1) 0.93
2 B(0, 2) 0.90
3 B(0, 3) 0.88
4 B(0, 4) 0.85
21.13. Using a payoff table similar to Table 21.4 in the text, explain the portfolio of
zero-coupon bonds that generates the forward rate f (0,3).
21.14. Compute the zero-coupon bond prices from the following forward rates.
QUESTIONS AND PROBLEMS 537
Times to Maturity T Today’s forward Rates f (0, T)
1 f (0, 1) 0.02
2 f (0, 2) 0.03
3 f (0, 3) 0.04
4 f (0, 4) 0.05
21.15. What is the spot rate of interest R(t)? Explain how it relates to forward rates
and yields.
21.16. What is a forward rate agreement? Give a numerical example to demonstrate
the timing and magnitude of the payoffs to an FRA.
21.17. What is an interest rate futures? Give examples of two popular interest rate
futures.
21.18. What are the differences between a forward rate agreement and an interest rate
futures contract? In your answer, consider both the institutional differences
and the economic differences.
21.19. Given the difference between forward rate agreements and interest rate futures
contracts, would you expect the FRA rate to equal the futures interest rate,
for otherwise identical contracts? Explain your answer.
21.20. What is the relation between an FRA rate for a contract that matures at time
T and a forward rate of interest for time T – 1?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
22
Interest Rate Swaps
22.1 Introduction 22.4 Valuation
22.2 A Brief History 22.5 Variations of Interest Rate
The Introduction of Swap Swaps
Contracts 22.6 Swaps and FRAs
From a Brokerage to a Dealership Synthesizing Swaps with
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Market Eurodollars and FRAs
ISDA and Standardization of
Contracts The Yield and Swap Curve
EXTENSION 22.1 Computing
22.3 Institutional Features Forward Rates from Swap Rates
Entering a Swap Contract
22.7 Summary
Documentation
Closing a Swap Position 22.8 Cases
22.9 Questions and Problems
A BRIEF HISTORY 539
22.1 Introduction
“Trust you couldn’t; gamble you shouldn’t.” This characterizes the lessons learned from
a fixed-for-floating interest rate swap deal struck in 1993 between Bankers Trust and
Procter & Gamble (P& G). Enticed by the fact that the swap would pay a low floating
rate, P& G apparently downplayed the importance of a “spread” in its payment terms,
which could potentially increase its liability. Interest rates rose, the company suffered
huge losses, and it sued, arguing it didn’t understand the complex swap. Suffering
a serious erosion of trust, the bank eventually forgave most of the money due and
settled the suit. The moral of this story is clear: don’t transact in derivative securities
that you do not understand, especially those with complicated terms like an exotic
interest rate swap! Understanding such swaps is a purpose of this chapter.
Although you have seen swaps before in Chapter 7, we discuss them here in greater
detail. The chapter begins with a brief history of swaps. Next, we discuss institutional
features of swap contracts. Swaps enjoy great flexibility because they trade in the over-
the-counter markets, but the trades are guided by standardized rules. We then develop
a valuation formula for plain vanilla interest rate swaps. Last, we demonstrate how to
construct swap curves, and we link swaps to forward rate agreements (FRAs) and
forward rates.
The swap valuation formula presented in this chapter is obtained in a world with
stochastic interest rates. But the valuation formula is derived without specifying a
particular model for the evolution of the term structure of interest rates. This is due to
the linear structure of the swap’s payoffs in both the underlying fixed and floating rate
cash flows. The swap’s value is shown to only depend on the current term structure of
zero-coupon bond prices. This is analogous to the option pricing relations obtained
in Chapter 16, prior to the study of the binomial option pricing model of Chapters
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
17 and 18 or the Black–Scholes–Merton model of Chapters 19 and 20. If you recall,
the nonlinear payoffs of call and put options required the specification of a stock
price process to obtain valuations. The same is true for the more general interest rate
derivatives discussed in the chapters that follow. The nonlinear payoff structure of
interest rate derivatives require the specification of an evolution for the term structure
of interest rates, a key element in understanding the Heath–Jarrow–Morton (HJM)
model.
22.2 A Brief History
In 1981, the Wall Street firm Salomon Brothers Inc. organized a transaction between
the World Bank (often referred to as “the Bank”) and the International Business
Machines Corporation (IBM) hailed as the first swap. The World Bank found that
interest rates in West Germany and Switzerland were lower than in the United States,
but the bank had already borrowed the maximum amount it could in these markets.
IBM had loans denominated in Deutsche marks and Swiss francs. An appreciation of
the dollar lowered the value of these loans, and IBM wanted to convert them into
dollar loans. So, the Bank borrowed dollars by selling Eurobonds and converted them
into marks and francs in the spot market. The deal was that the World Bank and IBM
540 CHAPTER 22: INTEREST RATE SWAPS
embraced each other’s loan obligations. The net result was the first swap because it
involved an exchange of cash flows at future dates.
The swap market has enjoyed phenomenal growth—Table 22.1 lists some mile-
stones. We classify the market developments into four categories: (1) the introduction
of new swap contracts, (2) the shift from a brokerage business to a dealership market,
(3) the founding of the International Swaps and Derivatives Association (ISDA) and
the standardization of contracts, and (4) the use of technology and the automation of
trading.
The Introduction of Swap Contracts
The year 1982 saw a financial arrangement usually recognized as the first interest
rate swap (see Johnson 2004; Greenbaum and Thakor 2007). The first interest rate
swap cleverly matched two parties—Sallie Mae issued an intermediate fixed-rate
loan, ITT Corporation issued a floating-rate note—and they swapped their payment
obligations. This deal opened the flood gates with respect to swap creation and
issuance. Besides developing exotic versions of interest rate and currency swaps, in
1986, Chase Manhattan Bank intermediated the first commodity swap based on
oil prices; in 1989, Bankers Trust organized the first equity swap; and in 1995,
J. P. Morgan issued the first credit default swap. Swap markets have enjoyed enormous
growth. Interest rate and currency swaps were actively traded almost 15 years before
either the credit default swap or equity derivatives markets traded significant notional
quantities, which happened in 2001 and 2002, respectively.
TABLE 22.1: Some Milestones in the History of Swaps
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Year Development
1981 The first currency swap was arranged by Salomon Brothers
between IBM and the World Bank.
1982 The first interest rate swap was arranged between Sallie Mae and
ITT Financial.
1985 The International Swap Dealers Association (ISDA; later renamed
the International Swaps and Derivatives Association) was
chartered. ISDA introduced SWAPS, a document template useful
for standardizing swap contracts.
1986 The first commodity swap was intermediated by Chase
Manhattan Bank.
1987 ISDA published the first Master Agreement, newer versions of
which were introduced in 1992 and 2002.
1989 The first equity-linked swap was organized by Bankers Trust.
1995 The first credit default swap was traded by J. P. Morgan.
A BRIEF HISTORY 541
From a Brokerage to a Dealership Market
By 1983, investment banks were arranging one-off swap transactions for their clients
on a fully matched basis. Banks earned substantial brokerage fees and had no credit
risk—but there were road bumps along the way. Every transaction had to be
separately negotiated with the counterparties and carefully documented. The front
office (traders) wanted to proceed but the back office (accountants) moved at a snail’s
pace, and the trader’s numbers would sometimes differ from the accountant’s. In a
particularly telling incident, acute back-office problems forced J. P. Morgan to scale
down its swap business for about six months in 1987.
The swap banks (also known as swap facilitators) discovered that by becoming
swap dealers, they could cut their back-office costs by standardizing the contracts
and trading them like homogeneous commodities. The swap banks also started
warehousing swaps. They would enter into different swaps on both the buy and
the sell side of the market and hedge the residual risk with interest rate derivatives.
The next example illustrates how this works.
EXAMPLE 22.1: Managing a Swap Book
■ Suppose that Swapbank Inc. (a fictitious name) enters into a plain vanilla interest rate swap on a
notional principal of $100 million, where it receives the floating-rate six-month libor index rate and
pays a fixed rate per year. Like any other good dealer, Swapbank tries to manage its books by perfectly
hedging its portfolio.
■ The bank enters into an offsetting swap with matching terms, but it receives four basis points higher
interest on a notional principal of $80 million. The bank has hedged its interest rate risk on $80
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
million and earns four additional basis points, which equals
Principal amount × Spread × Time period over which interest is earned
= 80 million × 0.0004 × 0.5
= $16, 000 every six months
■ Swapbank has counterparty credit risk. However, given the collateral agreements in the swap market,
this risk is small.
■ There is still risk from the remaining $20 million notional principal. Swapbank can enter into another
offsetting swap to remove this risk. Swapbank can also remove this risk by trading other interest rate
derivatives. We explain how to do this at the end of this chapter.
ISDA and Standardization of Contracts
As mentioned previously, to trade swaps quickly and to minimize transaction costs,
standardized contracts were needed. Efforts at developing such documents began in
1984, when eighteen swap dealers joined together to form the International Swap
Dealers Association (ISDA, pronounced “iz-da”; later renamed the International Swaps
542 CHAPTER 22: INTEREST RATE SWAPS
and Derivatives Association to reflect its broader scope). Chartered in 1985, ISDA
published the Code of SWAPS (Code of Standard Wording, Assumptions, and
Provisions for Swaps; the “Code”), which provides standardized documents known
as the Master Agreements for over-the-counter (OTC) derivatives contracts.
The ISDA’s Master Agreement establishes contract standards for privately
negotiated derivatives so that legal uncertainty is reduced and credit risk is lowered
through collateral provisions, netting arrangements, and procedures to follow in the
case of a default. The Master Agreement itself has two parts: (1) a preprinted form
that cannot be amended and (2) a schedule that allows the two counterparties to
note changes and amend provisions on the printed form (see Allen and Overy 2002).
The Master Agreement establishes noneconomic terms (such as “representations
and warranties, events of default, and termination events”) that govern future
OTC derivatives transactions between the two counterparties. Each time these two
counterparties execute a future trade, all they need do is negotiate the economic terms
and record them in a Confirmation. The Master Agreement, the confirmations, and
other supporting documents make up the ISDA Agreement Structure (as of 2006).
22.3 Institutional Features
It is instructive to follow a swap transaction from beginning to end and discuss the
relevant institutional features met along the way (see Figure 22.1 for the timeline).
Entering a Swap Contract
Suppose Your Beloved Machines (YBM) decides to become a swap issuer (or a
swap initiator). Then YBM contacts one or more swap providers who stand ready to
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
be their counterparty. Of course, in selecting the swap bank, YBM needs to consider
their credit rating, their prior experience with similar trading, the documentation
they provide, and the price they quote.
There are three basic ways to initiate a swap contract:
1. The traditional way is by phone, during which the basic economic terms of the
transaction (price or rate, principal amount, term, and frequency of payment)
are negotiated and any credit support requirements (margin, or collateral, or a
guarantee) are discussed. The counterparties are committed to transact as per this
“oral agreement.” Then, one counterparty (or broker, if involved) sends a letter
containing the economic terms to the other and asks for a confirmation.
2. A second way is via a web-based trading platform, which displays a range of
potential swaps. Once YBM has prearranged trading agreements with a bank,
the company can check quotes and transact a plain vanilla swap in a matter of
minutes. The bank has already made YBM sign the necessary paperwork and set
up a system that allows it to trade quickly but safely.
3. A third way is by running an auction. This is more common in the case
of municipal swaps that have a state, local, or municipal government as a
counterparty.
INSTITUTIONAL FEATURES 543
FIGURE 22.1: Timeline for a Plain Vanilla Swap
Periodic net payments (dark boxes)
… … …
Starting date Intermediate dates Maturity date
Swap initiator and bank Counterparties make periodic Swap ends with a
execute trade by verbal payments (semi-annually for final payment.
agreement, an electronic most swaps).
trade, or an auction.
Necessary documentation
follows.
Unless it is a forex or a Payments are netted (if they are
currency swap, notional in the same currency); usually
principal is not not netted if they are in
exchanged. different currencies.
No cash changes hands Traders can close out positions
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
(a par swap). by (1) a buyout, (2) an
assignment, or (3) an offset.
Counterparties may No secondary market. OTC
have to keep collateral. contract—credit risk is present.
Documentation
“Partnerships often finish in quarrels” wrote Benjamin Franklin in his 1791 autobi-
ography, “but I was happy in this, that mine were all carried on and ended amicably,
owing, I think, a good deal to the precaution of having very explicitly settled, in our
articles, every thing to be done by or expected from each partner [emphasis added], so that
there was nothing to dispute, which precaution I would therefore recommend to all
who enter into partnerships.” This practical advice on partnerships, and by extension
swaps, is the reason for documentation.
The swap contract’s documentation must describe (1) the currencies of the cash
flows, (2) the interest rates, (3) the principal amount, (4) whether the principal is
544 CHAPTER 22: INTEREST RATE SWAPS
exchanged, and (5) whether the counterparties make future cash payments in their
entirety or through a net payment. Next comes a timetable: (6) the tenor (or maturity)
of the swap and (7) the frequency of cash payments (every six months for a majority
of simple swaps). Finally, the contract must specify other issues such as (8) how to
terminate a swap early and (9) what to do if one side defaults. Of course, web-
based trading platforms have most of this material prepackaged so that trades can be
executed quickly.
Closing a Swap Position
If all contract terms are fulfilled, a swap terminates at its maturity date. But there are
three ways of ending a swap before it matures. Suppose that YBM decides to close
out an existing swap today and calculations show that the swap has a positive value of
$1 million. Then, the process of closing is as follows:
1. Buyout. A buyout involves YBM’s counterparty paying YBM $1 million and then
ending the swap by mutual consent.
2. Assignment. With the counterparty’s approval, YBM can assign the swap to a
third party. The third party would pay YBM $1 million, replace YBM in the
transaction, embrace all the benefits, and accept all the obligations for the swap’s
remaining life.
3. Offset. YBM can enter into an identical swap with equal but opposite terms as
the original swap. YBM’s counterparty in the new swap will have to pay YBM
$1 million because that is the value of the swap. In the combined swap position,
YBM will have zero net cash flows in the future. Notice that unlike the first two
transactions, here YBM does not completely leave the market and will continue
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
to have credit risk until both swaps mature.
22.4 Valuation
This section values a plain vanilla interest rate swap when interest rates are stochastic.
The valuation formula is derived without specifying a particular evolution of the term
structure of interest rates. This is due to the linear structure of the swap’s payoffs in
both the underlying fixed and floating cash flows. This is analogous to the pricing of
forward contracts in Chapters 11 and 12.
Swap valuation involves a clever trick: (1) decompose the swap into a fixed rate and
a floating rate loan corresponding to each leg of the swap, (2) compute the present
value of each of these loans, and (3) take their difference. Of course, we know how
to compute the present value of fixed cash flows, but how does one value the floating
cash flows? We solve this problem with a simple observation.
VALUATION 545
The simple observation is that the value of a floating rate loan is always equal to its
par value on the payment and reset dates.1 This happens because the loan pays out the
interest owed on these dates. For example, a floating rate loan of $100 million for
three years pays out the relevant interests perhaps quarterly (by whatever method of
interest computation devised) and at each quarterly payment date, the value of the
loan outstanding is $100 million.
The clever trick in valuing a swap is to include the principal payments while
valuing the fixed-rate and floating-rate cash flows. This inclusion does not matter
because the principal payments cancel when taking the difference between the two
loans’ cash flows. And, it simplifies our task because it is easy to value both the fixed
and floating rate loans. To see how this works, consider both a fixed rate loan paying
interest for three years and then returning a principal of $100 million, and a three-year
floating rate loan with a principal of $100 million. Next, value both of these loans.
Now, if you take the difference between these values, the present value of the $100
million notional cancels, leaving just the present value of the fixed-rate payments less
the floating rate payments. This is precisely the value of a swap.
Armed with these insights, let us value an interest swap.
EXAMPLE 22.2: Valuing a Plain Vanilla Interest Rate Swap (Fixed versus
Floating)
A Numerical Example
■ Consider the data from Example 7.4.
- Fixed Towers Inc. borrowed $100 million for three years at a floating rate of one-year libor but now
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
wants a fixed-rate loan.
- Floating Cruisers Co. raised the same amount, for the same time period, at a fixed interest rate of
i = 0.06 or 6 percent per year and wants to switch to a floating-rate loan.
- They do an interest rate swap by embracing each other’s liabilities. Fixed Towers buys the swap by
agreeing to pay the fixed rate of 6 percent to Floating Cruisers, who, in turn, sells the swap by
agreeing to pay the one-year libor rate to Fixed Towers (see Figure 22.2, which repeats Figure 7.2).
A swap buyer pays a fixed rate and receives the floating rate, whereas a seller does the opposite.
- The notional principal of LN = $100 million is used to compute interests that are paid at year’s end.
- We discount the fixed rate payments using the zero-coupon bond prices in Table 22.2.
1
It is a simple matter to compute the present value of the next floating interest rate payment between
the payment dates. The present value of the next floating interest rate payment plus the par value is the value of
the floating rate loan to the issuer between the payment dates. The next floating interest rate payment (say, X) is
known at the last repayment date (say, time a) and before it is due (say, time b). The next floating interest
rate payment is equal to the current floating interest rate times the notional. Then, the present value at
time a of the next floating interest rate payment at time b is B(a, b)X. At time b, when the interest is paid,
the floating rate loan resets to par value.
546 CHAPTER 22: INTEREST RATE SWAPS
FIGURE 22.2: Plain Vanilla Interest Rate Swap
Transforming a Fixed Rate Loan Into a
Floating Rate Loan (and Vice Versa)
After Entering into Swap for 3 years
libor
Fixed Towers Floating Cruisers
Net cost 6 percent per year Net cost libor per year
6 percent
libor 6 percent
Present Value of the Fixed Rate Loan
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ First, let us look at the fixed rate loan underlying the swap. The fixed rate loan has a coupon payment
of C dollars per period plus a principal of LN at the maturity date. The periodic payments are obtained
by multiplying the notional principal by the fixed interest rate i. Therefore
The fixed payment at the end of each period,
C = LN × i
= 100 × 0.06 (22.1)
= $6 million
The present value of the fixed rate loan is
PVFix = B (0, 1) × C + B (0, 2) × C + B (0, 3) × C + B (0, 3) × C + B (0, 3) × LN
= (0.97 + 0.93 + 0.88) × 6 + 0.88 × 100 (22.2)
= $104.68 million
VALUATION 547
TABLE 22.2: Common Zero-Coupon Bond Price
Data
Times to Maturity T Today’s Zero-Coupon Bond Prices B (0, T)
1 B(0, 1) 0.97
2 B(0, 2) 0.93
3 B(0, 3) 0.88
4 B(0, 4) 0.82
Present Value of the Floating Rate Loan
■ Recall our observation: the floating rate loan is always equal to its par value at the payment dates:
PVFloat = LN = $100 million (22.3)
Swap Valuation
■ Consequently, the value of the swap to Fixed Towers, who pays fixed and receives floating, is computed
by subtracting (22.2) from (22.3):
Present value of the floating rate loan − Present value of the fixed rate loan
= PVFloat − PVFix
(22.4)
= 100 − 104.68 million
= −$4.68 million
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ To enter the swap, Fixed Towers needs to be paid $4.68 million today.
■ Alternatively, one can adjust the interest rate to make this a par swap with zero value. To do this,
one needs to find the fixed rate that equates the present value of the two legs, that is,
LN = B (0, 1) C + B (0, 2) C + B (0, 3) C + B (0, 3) LN
Remembering from (22.1) that C = LN × i, rearranging terms, and canceling LN , we get
1 − B (0, 3)
Swap rate iS = (22.5)
B (0, 1) + B (0, 2) + B (0, 3)
The fixed interest rate for the par swap, iS , is commonly referred to as the swap rate.
■ Plugging in the values of the zero-coupon bonds, we get
iS = (1 − 0.88) / (0.97 + 0.93 + 0.88) = 0.043165 or 4.32 percent (approximately)
This example motivates the following general result.
548 CHAPTER 22: INTEREST RATE SWAPS
RESULT 22.1
Valuing a Plain Vanilla Interest Rate Swap
Consider an interest rate swap between two counterparties starting today (time
0) and ending at time T. At the end of each period (at time t, where t = 1,
2,.., T), the swap buyer pays a fixed amount LN × i (where LN is the notional
principal and i is the fixed interest rate) and receives a floating payment known
only at the beginning of that period.
The value of the swap to the buyer equals the present value of a floating-rate
loan minus the present value of a fixed rate loan, that is,
VSwap = PVFloat − PVFix (22.6)
The present value of the floating rate loan at the payment dates always equals
par, that is,
PVFloat = LN (22.7)
and the value of the fixed rate loan is the present value of the fixed cash flows,
that is,
PVFix = [B (0, 1) + B (0, 2) + ... + B (0, T)] LN × i + B (0, T) × LN
T (22.8)
= B (0, t) LN × i + B (0, T) × LN
[∑
t =1 ]
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where B(0, t) is the time 0 price of a zero-coupon bond that pays $1 at time t.
In the case of a par swap, the fixed interest rate iS (commonly known as
the swap rate) is computed by setting PVFloat = PVFix .
For simplicity, our previous example assumed annual interest payments.
Example 22.3 shows how cash flows are computed in the more realistic case of
semiannual payments.
EXAMPLE 22.3: Plain Vanilla Interest Rate Swap with Semiannual Interest
Payments
■ Let us continue with the original swap in Example 22.2 but assume that the payments are made every
six months. The fixed rate i = 6 percent is quoted on a semiannual bond-equivalent basis. It
assumes a 365-day year and computes interest over the actual number of days in this period. If there
VARIATIONS OF INTEREST RATE SWAPS 549
are 182 days during the next six-month period, Fixed Towers pays after six months:
Notional prinicpal × Fixed interest rate × (Days in period/365)
= 100, 000, 000 × 0.06 × (182/365)
= $2, 991, 780.82
■ The floating side is quoted on a money market yield basis, which assumes 360 days in a year
instead of 365. If the current six-month libor rate is 5 percent per year, Floating Cruisers pays after
six months
Notional prinicpal × Six − month libor rate × (Days in period/360)
= 100, 000, 000 × 0.05 × (182/360)
= $2, 527, 777.78
■ Fixed Towers pays Floating Cruisers a net amount equal to
2, 991, 780.82 − 2, 527, 777.78
= $464, 003.04
■ Every six months, Fixed Towers pays $2,991,780.82 or a similar amount computed using the actual
number of days over the previous six-month period. After six months are over, the counterparties
check the market and finds the new six-month bbalibor rate that applies over the next six months.
With this rate and the actual number of days in the period, the next floating payment is calculated.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
22.5 Variations of Interest Rate Swaps
This section explores the different kinds of interest rate swaps that can be created by
adding features to plain vanilla interest rate swaps. Interest swaps have become quite
sophisticated these days. Here are some examples created by adding various features
to plain vanilla interest rate swaps:
■ Setting a maximum value (or a cap) on the floating rate in a plain vanilla swap
creates a rate-capped swap. Later you will see that this simple change embeds a
call option into the swap. If a minimum value (or a floor) on the floating rate is
also specified, then this security embeds an interest rate collar.
■ A yield swap, which has both payments floating, may involve the exchange of a
bond with a lower yield for one with a higher yield. Usually the higher-yield bond
has a longer maturity and a lower credit rating, which implies a higher interest rate
and credit risk.
■ The notional principal in an amortized swap gets reduced over the swap’s life
according to a predetermined schedule. This is useful for hedging a portfolio of
mortgages, which typically earns a fixed rate on the dollars invested and which
amortizes over time owing to principal repayments.
550 CHAPTER 22: INTEREST RATE SWAPS
■ There are other swaps that involve changing notional principals. For example,
accreting swaps have a notional principal increasing over time. These swaps are
useful for hedging the risk associated with a construction project, where more
money is invested over time. Seasonal swaps are useful in the retailing business,
where the committed capital depends on the time of the year. The most general
form of such a swap would be a roller-coaster swap, whose notional principal is
structured by the dealer to suit the financing needs of the client’s business.
■ A cancellable swap gives one counterparty the right to cancel the swap without
additional payments or penalties. They come in two types. A callable swap gives
the fixed-rate payer the right to end the swap and is likely to be called when floating
rates decline below a certain level. In contrast, a putable swap gives the fixed-rate
receiver the right to put or cancel the swap and is likely to be put when floating
rates rise above a threshold level.
■ A forward swap (or a deferred swap) has its terms and conditions set at
origination but begins only at a later date. This delaying tactic may be used for
accounting or tax purposes.
■ A constant maturity swap (or CMS) involves the exchange of fixed-rate
payments for floating rate payments where the floating rate (which is the constant
maturity rate) is reset at each period according to a fixed maturity market rate with
a duration extending beyond that of the swap’s reset period (such as the one-year
constant maturity Treasury note rate).
To value and hedge these many variations of plain vanilla interest rate swaps, one
needs the Heath–Jarrow–Morton model to price the embedded options, which is
discussed in the next three chapters.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
22.6 Swaps and FRAs
The previous chapter noted that forward rates are the key rates underlying the
valuation of interest rate derivatives. As the forward rate is identical to the FRA
rate, it is not surprising that FRAs can be used to synthetically construct a swap. Let’s
see how.
Synthesizing Swaps with Eurodollars and FRAs
This section shows how to construct a synthetic swap as a portfolio of FRAs and
zeros (Eurodollar deposits). Synthetic construction has several benefits: (1) it provides
an alternative way for pricing a swap as the cost of this synthetic construction, (2) it
shows how to hedge a swap’s cash flows using FRAs and zeros, and (3) it provides
a tool for identifying arbitrage opportunities across the FRA and swap markets. The
next example demonstrates this construction.
SWAPS AND FRAS 551
EXAMPLE 22.4: Synthetic Construction of an Interest Rate Swap with FRAs
and Eurodollar Deposits
The Construction
■ A careful examination of a fixed-for-floating interest rate swap’s cash flows reveals how the synthetic
construction works. Consider the swap from Example 22.2.
- The swap has a notional principal of LN = $100 million. It starts today (time 0) and matures after
three years.
- The fixed interest rate i is 0.06 per year.
- The floating rate is the one-year libor index rate for Eurodollars, R(t). These rates are known at the
beginning of each period, that is, R(0) is determined at time 0, R(1) at time 1, and R(2) at time 2.
■ The swap buyer Fixed Towers pays fixed and receives floating. For example, its cash flow is
100 [R (1) − 0.06] (22.9a)
after two years. Figure 22.3 depicts the buyer’s cash flows.
■ Now, consider an FRA on a notional principal of $100 million that matures after two years. The FRA
buyer’s cash flow is given by expression (21.11c) of Chapter 21:
100 [R (1) − fFRA (0, 1)] (22.9b)
where fFRA (0,1) is the forward rate negotiated today that starts at the end of the first year and applies
over the second year.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
FIGURE 22.3: A Swap Buyer’s Cash Flows (Pays
Fixed, Receives a Floating Rate)
Time 0 1 2 3
Spot R(0) R(1) R(2)
rates
Cash 100[R(0) – 0.06] 100[R(1) – 0.06] 100[R(2) – 0.06]
flows to
the Buyer
■ Except for the last quantity in each expression, the cash flows in expressions (22.9a) and (22.9b) are
the same. To adjust for this difference, add to the FRA trade a long position in 100[fFRA (0,1) – 0.06]
zero-coupon bonds maturing after two years. This could be a positive or negative quantity, depending
552 CHAPTER 22: INTEREST RATE SWAPS
on whether the spot rate is greater than or less than the FRA rate. The long FRA and the long zeros’
portfolio’s payoff at time 2 will be
100 [R (1) − fFRA (0, 1)] + 100 [fFRA (0, 1) − 0.06]
(22.9d)
= 100 [R (1) − 0.06]
This is the same as the swap’s time 2 cash flow expression (22.9a).
■ As this argument is not specific to time 2, you can synthetically construct each of the swap’s cash flows
at the other payment dates in an identical manner.
■ Summing the FRAs and zeros across all the cash flow payment dates generates a portfolio of FRAs
and zero-coupon bonds that yield the same cash flows as the swap.
We gather these ideas together as our next result.
RESULT 22.2
Synthetically Constructing a Swap Using FRAs and Zeros
Consider an interest rate swap starting today (time 0) and ending at time T.
At the end of each period t = 1, 2,.., T, the swap buyer pays a fixed amount
LN × i (where LN is the notional principal and i is the fixed interest rate) and
receives a floating amount, the spot rate R(t − 1).
The swap buyer’s cash flows at time t are
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
LN × [R (t − 1) − i] (22.10)
This swap’s payoffs are identical to a long portfolio of FRAs with the same
notional principal maturing at times t = 1, 2,.., T and LN [ fFRA (0, t − 1) − i]
zero-coupon bonds also maturing at times t = 1, 2,.., T.
Given that swaps are portfolios of FRAs and zeros, it makes sense that one
can use swap rates to extract forward rates. This extraction is demonstrated in
Extension 22.1.
The Yield and Swap Curve
This section relates the Treasury (par bond) yield curve to the swap curve. Recall
that the yield curve for US Treasury securities is obtained from T-bills for maturities
of one year or less and from T-notes and T-bonds for maturities of more than one
year. For the coupon-bearing bonds, the par bond yields are graphed. Recall that the
par bond yields correspond to the yields on Treasury securities that are priced at par.
Such a hypothetical Treasury yield curve is graphed in Figure 22.4.
SWAPS AND FRAS 553
EXTENSION 22.1: Computing Forward Rates from Swap Rates
Computing forward rates from swap rates is analogous to computing forward rates from coupon bond prices (see
Extension 21.3). This is done by solving a system of linear equations for a collection of unknowns. The first step
is to compute the zero-coupon bond prices.
Consider a par swap with value VSwap = 0, notional (LN ), maturity date (T), and a swap rate (iS ) paid every
period. Setting expression (22.7) equal to (22.8) in Result 22.1 and rearranging terms gives us an expression
involving the zero-coupon bond prices, which are the unknowns, that is,
T
1 − B (0, T) = B (0, t) × is (1)
[∑
i =1 ]
Next, suppose we are given a collection of T distinct maturity swap rates. Then, each swap has an equation
similar to expression (1). This gives us T equations in the T unknowns {B(0, 1), B(0, 2),..., B(0, T)}. Solve them
to obtain the zero-coupon bond prices. If the system does not have a solution, one can find those unknowns that
minimize the sum of errors squared between the right side of expression (1) and the bond prices across all bonds.
EXT. 22.1 EX. 1: Computing Forward Rates from Swap Rates
■ Today is time 0. Consider three plain vanilla fixed-for-floating swaps with swap rates is = 0.0309 for a swap
maturing in one year, 0.0368 for a swap maturing in two years, and 0.0432 for a swap maturing in three years.
Computing Zero-Coupon Bond Prices
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ For the one-year swap, expression (1) gives
1 − B (0, 1) = B (0, 1) iS
(2a)
or B (0, 1) = 1/ (1.0309) = 0.97
■ For the two-year swap, expression (1) gives
1 − B (0, 2) = [B (0, 1) + B (0, 2)] iS
(2b)
or B (0, 2) = (1 − 0.97 × 0.0368) / (1 + 0.0368) = 0.9300
■ Finally, for the three-year swap, expression (1) gives
1 − B (0, 3) = [B (0, 1) + B (0, 2) + B (0, 3)] iS
(2c)
or B (0, 3) = (1 − 0.97 × 0.0432 − 0.9300 × 0.0432) / (1 + 0.0432) = 0.88
These zero-coupon bond prices are the same as the corresponding values in Example 21.3, and we have already
computed the forward rates in Example 21.5, completing this example.
554 CHAPTER 22: INTEREST RATE SWAPS
Now let’s construct a similar curve for the Eurodollar market. Recall that
Eurodollar deposit rates are quoted for up to one year. These rates correspond to
the T-bill yields and can be graphed directly. Par swap rates for swaps with maturities
of more than one year complete the curve. Recall that a par swap rate is the rate
that equates the present value of the cash flows from the swap’s underlying coupon
bond (expression [22.8]) to the face value of the bond (the present value of the swap’s
underlying floating-rate bond). Hence the par swap rate is equivalent to a par bond
yield when viewed from this perspective. Such a hypothetical Eurodollar swap curve
is also graphed in Figure 22.4.
Thus, when constructed in this manner, the swap curve is almost identical to the
Treasury yield curve. The difference is not in the formulas for the rates graphed—they
are the same and provide an apples-to-apples comparison. Rather, the difference is
due to the differing credit risks between the Treasury and Eurodollar rates as measured
by the Treasury-Eurodollar (TED) spread. If it were plotted on the same graph, the
Eurodollar rate would exceed the Treasury rate by the TED spread. In fact, one
reason for graphing the two curves together is to visually see their differences (the
TED spread)! When markets are calm, the TED spread tends to be small in magnitude
and relatively constant across maturities. When markets are in turmoil, however, the
TED spread becomes large in magnitude, with the biggest differences on the short
end of the curve. It is a useful statistic to gauge the market’s pricing of credit risk for
the largest European banks in times of calm and turmoil.
FIGURE 22.4: Yield Curve and Swap Curve
Interest rate
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Yield curve (Treasuries)
0
1 year Time
Short-maturity bonds Long-maturity bonds
(yields) (par bond yields)
Interest rate
Swap curve (Eurodollars)
0
1 year Time
Short-maturity deposits Long-maturity swaps
(bbalibor rates) (par swap rates)
QUESTIONS AND PROBLEMS 555
22.7 Summary
1. The first swap was a currency swap arranged by Salomon Brothers between IBM
and the World Bank. We divide the market developments into three categories: (1)
the introduction of new swap contracts, (2) a shift from a brokerage business to a
dealership market, (3) the founding of ISDA and the standardization of documents,
and (4) the use of technology and the automation of trading.
2. There are three ways of initiating a swap contract: through phone calls, via a
web-based platform, and by auctions. Auctions are more common in the case
of municipal swaps. A swap position is closed through a buyout, through an
assignment, or by entering into an offsetting swap.
3. A swap contract’s documents must describe the cash flows, a timetable for the cash
payments, and other relevant considerations like how to terminate a swap early
and what to do if one side defaults.
4. The value of a plain vanilla interest rate swap to the fixed-rate payer equals the
present value of the floating-rate note minus the present value of the fixed-rate
loan implicit in the swap. A par swap sets the fixed interest rate so that the swap
has a zero value.
5. A swap can be synthetically constructed as a portfolio of forward rate agreements
and zero-coupon bonds.
6. The Treasury par bond yield curve and the Eurodollar swap curve graph identical
rates, with the exception of credit risk. The difference between these two curves
is called the Treasury over Eurodollar deposit rate (TED) spread.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
22.8 Cases
Procter and Gamble versus Bankers Trust: Caveat emptor (Thunderbird Case
A06-05-0001, European Case Clearing House). The case analyzes a complex
interest rate swap between Procter & Gamble and Bankers Trust.
Wal-Mart’s Use of Interest Rate Swaps (Harvard Business School Case 108038-
PDF-ENG). The case studies Walmart’s use of interest rate swaps to hedge the
fair value of its fixed-rate debt against changing interest rates.
Banc One Corp.: Asset and Liability Management (Harvard Business School
Case 294079-PDF-ENG). The case examines a bank’s use of interest rate swaps
to manage the sensitivity of its earnings to changes in interest rates and as an
attractive investment alternative to conventional securities.
22.9 Questions and Problems
22.1. If you enter a plain vanilla interest rate swap, what risks do you face?
22.2. a. To reduce counterparty risk in a swap, what conditions are written into
the swap contract?
556 CHAPTER 22: INTEREST RATE SWAPS
b. Consider both the structure of the regular payments and the notional. Are
there any other restrictions in the ISDA swap documentation that reduce
counterparty risk?
22.3. a. What are the three ways one can initiate a swap transaction?
b. What are the four ways one can use to close a swap position?
22.4. Explain how a plain vanilla interest rate swap can change a fixed-rate loan into
a floating-rate loan, and include a diagram to support your explanations.
22.5. What is the value of a floating-rate loan? Explain your answer.
22.6. a. What is the warehousing of swaps?
b. Have swap dealers been successful in managing their books of plain vanilla
interest swaps? Explain your answer.
22.7. Compute the value of a four-year fixed-rate loan with a coupon of 5 percent
paid yearly on a principal of $100. Use the zero-coupon bond prices in the
following table:
Times to Maturity T Today’s Zero-Coupon Bond Prices B(0, T)
1 B(0, 1) 0.97
2 B(0, 2) 0.93
3 B(0, 3) 0.88
4 B(0, 4) 0.82
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
22.8. Explain why a cancellable swap might be of interest to a corporate treasurer.
22.9. Explain how to construct a swap using a floating-rate bond and a fixed-rate
bond.
22.10. Explain why a swap has zero value when it is initiated. Why might this be a
useful convention?
22.11. Based on the examples in the text, come up with an argument as to why plain
vanilla interest rate swaps do not exchange principal, while a plain vanilla
currency swap does.
22.12. Consider the swap in Example 22.2. Now, assume that the payments are made
every six months. Zero-coupon bond prices (maturing every six months) are
given in the following table:
QUESTIONS AND PROBLEMS 557
TABLE 22.3: Zero-Coupon Bond Prices
Time to Maturity (in Years) Zero-Coupon Bond Prices B (0, T)
0.5 B(0, 0.5) 0.99
1.0 B(0, 1.0) 0.97
1.5 B(0, 1.5) 0.95
2.0 B(0, 2.0) 0.93
2.5 B(0, 2.5) 0.91
3.0 B(0, 3.0) 0.88
Compute the value of this swap.
22.13. How can one synthetically construct a swap using a portfolio of forward rate
agreements and zero-coupon bonds?
22.14. Ignoring credit risk, is the swap rate on a T-year swap the same as the par-
bond yield on a T-year Treasury bond? Explain your answer.
22.15. Describe the two different ways of synthetically constructing a swap. Which
method is likely to be the easier to implement?
22.16. In the Bankers Trust versus Procter & Gamble situation, P&G argued that
it did not fully understand the complex interest rate swap it had entered.
Do you think it is prudent to enter a derivative that you do not completely
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
understand? Explain your answer.
22.17. a In chapter 1, we listed five different kinds of risks that were considered
important by the Basel Committee on Banking Supervision and the
International Organization of Securities Commissions. Pick three of these
risks and explain why they are important to traders in the swap market.
b Discuss the role played by the International Swaps and Derivatives Associ-
ation in the swap and over-the-counter derivatives markets.
c What is a Master Agreement?
22.18. Can you explain why an amortized interest rate swap might be a useful
instrument for managing the risk of residential home mortgages?
22.19. After it is written and time has passed, will an interest rate swap always be
equal to zero value? Explain.
22.20. Why are the swaps considered excellent examples of financial engineering?
23
Single-Period Binomial
Heath–Jarrow–Morton
Model
23.1 Introduction The Hedge Ratio (the Holy Grail)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
23.2 The Basic Interest Rate Risk-Neutral Valuation
Derivatives Valuing a Floorlet
Uses of Caps, Floors, and Collars Valuing Various Interest Rate
Derivatives
23.3 A Brief History of Interest
Multiple Factors
Rate Derivatives Models
23.6 Summary
23.4 The Assumptions
23.7 Appendix
23.5 The Single-Period Model
The Equal Pseudo-probability
Arbitrage-Free Evolutions
Condition
Zero-Coupon Bond Prices and
Proof of Sufficiency
Martingales
Proof of Caplet and Floorlet Parity
Understanding the Equal
Pseudo-probability Condition 23.8 Cases
Actual versus Pseudo-probabilities 23.9 Questions and Problems
Caplet Pricing
INTRODUCTION 559
23.1 Introduction
This chapter introduces the Heath–Jarrow–Morton (HJM) model for valuing interest
rate derivatives in the context of a simple binomial model. This chapter is similar to
yet different from the binomial option pricing model in Chapter 17. It is similar in that
the no-arbitrage and hedging arguments are invoked to value interest rate derivatives.
It is different in that a key assumption is changed. No longer is there only one stock
with random returns and one interest rate that is constant. Instead, we introduce a
term structure of interest rates that fluctuates randomly across time. Just like Alice in
her adventures in Wonderland,1 we have fallen into a new world, the world of interest
rate derivatives. It is similar to but different from our more familiar world of equity
derivatives. In this new world, like the Cheshire Cat and Mad Hatter, there are strange
and unusual characters—forward rates and zero-coupon bond prices—that will help
us in searching for our white rabbit, which is the interest rate derivative’s hedge ratio.
Although we considered interest rate derivatives (forward rate agreements [FRAs]
and swaps) in Chapters 21 and 22, they had linear payoffs. Consequently, we could
value the derivatives without assuming a particular evolution for the term structure
of interest rates. This and the next few chapters extend this analysis by considering
the basic interest rate derivatives with nonlinear payoffs, including futures contracts
as well as interest rate call and put options. In this new world, calls and puts are called
caplets and floorlets. Portfolios of these options trade, called caps and floors. This chapter
discusses caps and floors and their uses.
A brief history of interest rate derivatives models follows. Then, we study the
single-period binomial HJM model. Because the purpose of the HJM model is
to price interest rate derivatives, just as in the binomial option pricing model of
Chapter 17, we need to assume a stochastic evolution for the term structure of
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
interest rates. The presentation of the material in this chapter purposely parallels
the development of the binomial option pricing model (Chapter 17). As such, this
chapter piggybacks on that material to make our presentation simpler and easier to
follow. First we state the necessary assumptions. Second, to value a traded caplet, we
construct a synthetic caplet with identical payoffs using a zero-coupon bond and a
money market account (mma). The cost of construction yields the caplet’s arbitrage-
free price. This takes us to determining the hedge ratio, martingale pricing, and
risk-neutral valuation. To value a floorlet, we introduce caplet and floorlet parity,
the analogue of put–call parity. A discussion of the valuation of alternative interest
rate derivatives concludes the chapter. Chapter 24 picks up where this chapter ends
by extending the single period to a multiperiod setting, thereby generating an HJM
model that is useful in practice.
One last comment is necessary before reading Chapters 23 and 24. All com-
putations in the next two chapters are performed using 16-digit accuracy. When
reporting results in various formulas, however, we often round to four decimal places.
The resulting equations, if evaluated using the reported (and rounded) numbers, may
give different answers from those reported based on sixteen-plus digit accuracy. This
1
Alice’s Adventures in Wonderland is a classic book published in 1865 by Charles L. Dodgson under the
pseudonym Lewis Carroll.
560 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
is a typical problem when reporting numerical computations in this fashion (see
Extension 2.1). Being forewarned, you should have no problems understanding the
computations that follow.
23.2 The Basic Interest Rate Derivatives
As seen earlier in this book, four basic derivatives are written on any underlying
security or index: forwards, futures, call options, and put options. The same is true
in the world of interest rate derivatives. In this context, the underlying “index” is
the default-free spot rate of interest, and call and put options go by different names:
caplets and floorlets, respectively. Yet in this strange new world, caplets and floorlets
do not trade by themselves but only in portfolios called caps and floors, respectively.
The reason these securities trade as portfolios is to economize on transaction costs.
These securities, as will be shown, are useful for hedging fixed- and floating-rate
coupon bonds. As coupon bonds have multiple cash flows, it makes sense that the basic
hedging instruments need multiple cash flows as well. Interestingly, in this interest
rate world, forwards trade separately as FRAs and also in portfolios, called interest
rate swaps. We summarize these four basic derivatives in Table 23.1 and describe
them subsequently.
■ Forwards. A forward rate agreement (FRA) is an over-the-counter (OTC) contract
that fixes an interest rate for borrowing (or lending) over some future time period.
The FRA rate is equivalent to the forward rate, which is the fundamental building
block for our theory. An interest rate swap is a portfolio of zero-coupon bonds and
FRAs (see Chapter 22).
■ Futures. An interest rate futures is an exchange-traded contract that fixes a borrowing
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
(or a lending) rate at some future date. It differs from an FRA the same way
that a futures contract differs from a forward contract. A host of features such
as standardization, a clearinghouse, margins, and daily settlements make futures
safe and easy to trade. Some interest rate futures (like those on Treasuries) end in
physical delivery, while others (like those on Eurodollars) end in cash settlement.
■ Call options. A caplet is a European call option on an interest rate that is effective
for a single period. The buyer gets paid the spot rate minus the cap rate when
it’s positive, and nothing otherwise. A cap is a portfolio of caplets. Caps are cash
settled contracts usually based on a libor rate index.
■ Put options. A floorlet is a European put option on an interest rate that is effective
for a single period. The buyer is paid that period’s floor rate minus the spot rate
only when it’s positive. A floor is a portfolio of floorlets. Floors are cash settled
contracts usually based on a libor rate index.
Of course, other, more exotic interest rate derivatives trade. One of these is a
swaption, which is an option with an interest rate swap as the underlying. They
come in two basic types: a payer swaption is an option to enter into a swap where
the holder pays the fixed rate and receives the floating rate, while a receiver swaption
is an option to enter into a swap where the holder pays the floating rate but receives
the fixed rate.
THE BASIC INTEREST RATE DERIVATIVES 561
TABLE 23.1: The Basic Interest Rate Derivatives
Today Intermediate Date Maturity Date
1. Forwards: forward rate agreements (FRAs) No cash flows. Short pays buyer the spot rate
Buyer negotiates FRA rate (which is equivalent minus the FRA rate, applied
to the forward rate) and contract terms with the on a notional amount and
seller. No cash changes hands today. computed over a fixed time
interval.
2. Futures: Eurodollar futures contracts Marking-to-market cash Cash settlement based on a
Regulated, standardized contracts betting on a flows. notional quantity computed
simple interest rate (libor) to be realized at a over a fixed time period.
future date. No cash changes hands today.
3. Call options: caplets Caplet buyer receives the spot Same as in intermediate date.
A cap buyer pays a premium and buys a portfolio rate less the cap rate if
of European calls of different maturities on the positive, or zero otherwise.
same notional principal. The underlying is a
simple interest rate (libor). A caplet is one of
those European calls.
4. Put options: floorlets Floorlet buyer receives the Same as in intermediate date.
A floor buyer pays a premium and buys a floor rate minus the spot rate
portfolio of European put options of different if positive, or zero otherwise.
maturities on the same notional principal. The
underlying is a simple interest rate (libor). A
floorlet is one of those European puts.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Although the pricing of caps and floors is the primary goal in this chapter and the
next two, we also discuss how to price swaptions. Before discussing pricing, let us
consider how financial institutions use caps and floors.
Uses of Caps, Floors, and Collars
Example 23.1 illustrates how an interest rate cap can be used to speculate or hedge
interest rate risk.
EXAMPLE 23.1: An Interest Rate Cap
Speculation
■ International Treats and Restaurants Inc.’s (INTRest, a fictitious name) treasurer buys a cap today (time
0) with a cap rate k = 5 percent per year on a notional of LN = $100 million. This OTC contract
matures in two years and has the six-month libor rate index as the underlying floating rate of interest.
This cap is a collection of four caplets, each of which makes payments based on the realized spot rates
at six-month intervals.
562 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
■ Figure 23.1 shows the cap’s payoffs. The dotted line shows the six-month libor rate index. Spot rates
are determined at the beginning of each time period, and payments based on these rates are made at
the end of each time period.
- Today’s spot rate R(0) is 0.06. The first caplet pays INTRest after six months:
LN × [R (0) − k] × Time period
= 100 × (0.06 − 0.05) × 0.5 year (23.1a)
= $0.5 million
- Six months later (time 1), the six-month libor rate index is R(1) = 0.05. The second caplet is
at-the-money and no payment is made.
- As R(2) = 0.04 is less than the cap rate, the third caplet also expires worthless at time 3.
- The spot rate R(3) is 0.07 after a year and a half. The fourth caplet gives INTRest a payoff of 100
× (0.07 – 0.05) × 0.5 year = $1 million at time 4.
FIGURE 23.1: A Cap Buyer’s Payoffs
[R(0) – k] = 0.01 determined at time 0 but paid at time 1
Cap payoff
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Spot rate
0.07
0.06
k = 0.05 Cap Rate
0.04 [R(3) – k] = 0.02
determined at
time 3 but paid
at time 4
0 1 2 3 4 Time t
Today 6 months 2 years
Hedging
■ A cap protects the holder against interest rate increases by putting a ceiling on the floating borrowing
rate. To see this, suppose INTRest buys this cap and simultaneously issues floating rate debt paying
the libor rate index plus 50 basis points with L = $100 million par value and with a similar payment
date structure as the cap.
THE BASIC INTEREST RATE DERIVATIVES 563
- After six months, INTRest’s cash flow is
Cash flow from bond + Cash flow from first caplet
= −L × [R (0) + 0.005] × Time period + LN × [R (0) − k] × Time period
(23.1b)
= −100 × (0.005 + k) × 0.5
= 2.75 million
- The second and third cash out flows are just $2.75 million and $2.25 million, respectively, because
INTRest benefits from the low floating rates at times 1 and 2, with R(1) = 0.05 and R(2) = 0.04.
- At time 4, the total payment is again capped at $2.75 million.
■ The cap places a ceiling on INTRest’s borrowing cost at the cap rate + 50 basis points.
The cap’s payoff given in Equation 23.1a can be formalized as follows. Consider a
cap dates t = 1, 2,..., T. If the cap has a notional of LN and a cap rate k per year, then
a caplet maturing at time t has a payoff equal to
ct = max [0, R (t − 1) − k] × LN × Time period (23.2a)
where R(t – 1) is the spot rate (libor or euribor in most cases) determined at time (t –
1) that applies to the time period [t – 1, t]. This time period is determined according
to market conventions (such as actual/360 for Eurodollars and Euribor or actual/365
for pound sterling deposits). The cap’s payoff is the sum of the individual caplet’s
payoffs.
In our discrete time model, the tth caplet’s payoff at time t simplifies to
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
ct = max [0, R (t − 1) − k] (23.2b)
where we have assumed that the notional principal is $1 and the time period has unit
length.
A floor allows the holder to speculate or hedge on an interest rate decrease. A floor
is a collection of floorlets, which are European put options on interest rates. A floor
can ensure a minimum interest rate on an existing bond holding (which is analogous to
a protective put strategy from Chapter 15) or if one simultaneously buys the bond and
a floor (which is analogous to a married put). Many endowment funds, hedge funds,
mutual funds, pension funds, and financial institutions buy OTC traded interest rate
floors to achieve a minimum return on their bond holdings while retaining the upside
potential.
We can similarly describe a floorlet’s payoff. Consider a floor consisting of T
floorlets with maturity dates t = 1, 2, . . . , T. If the floorlets have a notional of
LN and a floor rate k per year, then a floorlet maturing at time t has a payoff equal to
pt = max [0, k − R (t − 1)] × LN × Time period (23.3a)
564 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
where R(t – 1) is the spot rate determined at time (t – 1) that applies to the time period
[t – 1, t]. Similarly, a floor’s payoff is the sum of the individual floorlet’s payoffs.
Analogously, a floorlet’s payoff at time t in the case of our discrete model with a
notional principal of $1 and a time period of unit length is
pt = max [0, k − R (t − 1)] (23.3b)
A commonly used combination of interest rate derivatives is an interest rate collar
that combines a cap and a floor with identical terms and conditions but with different
strike rates. Let’s consider Example 23.2.
EXAMPLE 23.2: An Interest Rate Collar
■ Suppose that INTRest Inc.’s treasurer buys the interest rate cap of the previous example. It consists
of four caplets each with a notional LN of $100 million. The cap rate k1 is 5 percent per year, the
maturity date is 2 years, the six-month libor rate index is the underlying floating rate of interest, and
the payments are based on the realized spot rates at six-month intervals.
■ Next, the treasurer thinks that she can save the company money by selling a floor with identical terms,
except a floor rate of k2 = 2 percent per year. The premium earned on the floor offsets the cap’s cost
and lowers the effective cost of the overall position.
■ Now if the realized spot rate exceeds 5 percent, INTRest will receive a payment from the cap so
that the company’s effective rate is 5 percent. Conversely, if the spot rate falls below 2 percent, say, to
0.5 percent, INTRest will have to pay on the floor – (k2 – R) = – 0.015 or an outflow of 1.50 percent.
The company has confined the interest rate paid to within the band of k2 to k1 or 2 to 5 percent. The
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
short floor traded surrenders the opportunity to benefit from very low interest rates.
■ This combination of caps and floors is similar to collars in the equity market introduced in Chapter 15,
and if the prices of the caps and floors match, then it is a zero-cost collar.
Before pricing interest rate derivatives like caps and floors, let us begin by providing
a bird’s-eye view of the historical development of interest rate derivatives pricing
models.
23.3 A Brief History of Interest Rate
Derivatives Models
In their book Financial Calculus—An Introduction to Derivatives Pricing, Baxter and
Rennie (1996, p. 203) wrote,
In the interest-rate setting, Heath–Jarrow–Morton is as seminal as Black–Scholes. By
focusing on forward rates and especially by giving a careful stochastic treatment, they
produced the most general (finite) Brownian interest-rate model possible. Other models
may claim differently, but they are just HJM with different notations.
THE ASSUMPTIONS 565
To understand the relevance of these comments, we step back and look at the history
of interest rate derivatives models.2 Interest rate derivatives pricing models began with
a collection of papers focusing on the evolution of the spot rate of interest, called “spot
rate” models, and it eventually ended with the HJM model.
■ Spot rate models. This class of models focused on the evolution of the spot rate,
R(t). These models include those by Vasicek (1977), Brennan and Schwartz (1979),
and Cox et al. (1985). Spot rate models have two major limitations. First, in
these models, the values of interest rate derivatives depend on interest rate risk
premia or, equivalently, on the expected return of default-free bonds. Recall that
estimating a risk premium was the nagging problem facing option pricing in the
pre–Black–Scholes–Merton (BSM) era, which the BSM model overcame. Second,
these models can not easily match the initial yield curve. This yield curve matching
is essential for accurate pricing and hedging of interest rate derivatives because any
discrepancies in yield curve matching may generate “false” arbitrage opportunities
in the priced derivatives.
■ HJM models. To address the limitations of the spot rate model, the HJM model was
developed. Motivated by consulting, Professor David Heath of Cornell University’s
School of Operations Research and Industrial Engineering and a PhD student,
Andrew Morton, were working on this problem. Robert Jarrow, a professor
of finance at Cornell’s Johnson Graduate School of Management, joined them.
Together they developed the HJM model, which was first presented as a working
paper in 1987 and published as a series of papers in 1990 and 1992. The HJM
model is a very general continuous-time and multifactor model. In fact, all of the
previous spot rate models are special cases. A popular subcase of the HJM model
known as the libor model (or the market model or the BGM model) is widely used by
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Wall Street for the pricing and hedging of caps, floors, and swaptions. Sandmann et
al. (1995), Miltersen et al. (1997), and Brace et al. (1997) independently developed
the libor model. This model generates a formula for a caplet that is analogous to the
BSM European call option formula. As such, it is familiar territory for derivatives
traders, which explains its widespread popularity (see Chapter 25 for a discussion
of this model).
23.4 The Assumptions
The assumptions underlying the single binomial HJM model are similar to those used
to generate the binomial option pricing model of Chapter 17, but with an important
modification. The old assumption A6 that there is no interest rate uncertainty is now
relaxed, and the old assumption A7 is replaced with a new assumption characterizing
the random evolution of the term structure of interest rates. This new assumption is
similar to, but more complex than, the old assumption A7 describing the evolution
of stock prices.
2
See Hughston (1996) and Ho and Lee (2004), Chapter 5, for additional histories of interest rate
derivatives models.
566 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
For completeness, we first restate assumptions A1–A5.
A1. No market frictions. We assume that trading involves no market frictions
such as transaction costs (brokerage commissions and bid/ask spreads), margin
requirements, short sale restrictions, and taxes (which may be levied on different
securities at different rates). Moreover, assets are perfectly divisible. This helps us to
create a benchmark model to which one can later add frictions, and it’s a reasonable
approximation for institutional traders.
A2. No credit risk. We assume that there is no credit risk. Also known as default
risk or counterparty risk, this is the risk that the counterparty will fail to perform on an
obligation. The absence of credit risk is a reasonable assumption for both exchange-
traded derivatives and many OTC derivatives because of collateral provisions.
A3. Competitive and well-functioning markets. In a competitive market,
traders act as price takers and prices cannot be manipulated. In a well-functioning
market, price bubbles (which create a wedge between an asset’s market price and its
intrinsic or fundamental value) are assumed to be absent.
A4. No intermediate cash flows. Many assets reward their holders with cash
flows such as dividends for stocks and coupon income for bonds. For now we assume
that assets provide no such cash flows to their holders. As you will see later, such cash
flows are easily incorporated into the model.
A5. No arbitrage opportunities. This is self-explanatory.
We now relax the old assumption A6 (of Chapter 17) that there is no interest rate
uncertainty. Instead, we allow the entire term structure of interest rates to fluctuate
randomly through time. The relevant term structure is that of forward rates. Recall
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
that Equation 21.9 in Chapter 21 gives the definition of the time t forward rate for the
future period [T – 1, T] as
B (t, T − 1)
f (t, T − 1) = −1 (23.4)
B (t, T)
where B(t, T) is the time t price of a zero-coupon bond maturing at time T. When
the time t forward rate corresponds to the immediate future period [t, t + 1], the
forward rate gets a special name, the spot rate, and is denoted by f (t, t) = R(t).
Given the term structure of forward rates, one can easily compute the term
structure of zero-coupon bond prices as shown in Chapter 21 (see Result 21.2).
Conversely, given the term structure of zero-coupon bond prices, one can easily
obtain the term structure of forward rates using Equation 23.4. This equivalence
implies that the term structure of forward rates and the term structure of zero-coupon
bond prices contain the same information.
Thus one can describe the evolution of the term structure of interest rates by
imposing an assumption on either the forward rates or the zero-coupon bond prices.
In discrete time models, for pedagogical reasons—to see the economic reasoning
more clearly—it is easier to impose the term structure evolution assumption on the
zero-coupon bond prices. For parameter estimation, however, one must transform
THE ASSUMPTIONS 567
the zero-coupon bond price evolution to that of the forward rates because the
forward rates’ evolution is more time stationary, a property compatible with statistical
estimation procedures (see Chapter 21).
In continuous-time models, however, the situation reverses. It is easier to impose
the term structure evolution assumption directly on forward rates. This is the
approach taken in Chapter 25 when discussing the HJM libor model. Of course,
as we now know, these are equivalent ways of imposing the same assumption.
To model the term structure evolution, two changes are made to the binomial tree
of Chapters 17 and 18. First, the spot rate of interest is now random across time; call
it R(t). Second, stock prices are irrelevant and dropped from the model. Instead, each
node now has a vector of zero-coupon bond prices B(t, T) with maturities T = 1, 2,
3, . . . for times t = 0, 1, 2, . . . , T
A6. Single-period binomial process. The zero-coupon bond prices B(t, T)
for t = 0, 1 and T = 1, 2, 3 trade in discrete time, and their prices evolve according
to a binomial process. After time 0, there are two states called “up” and “down.”
First, if the zero-coupon bond matures at time 1, then the current bond price goes
to $1 regardless of the state. Second, the T > 1 maturity zero-coupon bond price is
multiplied by the factor u(0, T) in the up state and takes the price u(0, T) × B(0,T)
with the actual probability q(0) > 0, or it is multiplied by the factor d(0, T) in the
down state and takes the price d(0, T) × B(0, T) with probability 1 – q(0). We assume
that u(0, T) > d(0, T).
The single-period evolution for the term structure of zero-coupon bond prices
is given in Figure 23.2. We first explain this evolution via an example (see
Example 23.3).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
EXAMPLE 23.3: The Single-Period Term Structure Evolution
Let today be time 0. There are four securities in the model: three zero-coupon bonds and a money-
market-account (mma). We describe their evolution over a one-year period.
The Zero-Coupon Bond Prices
■ The time 0 prices of zero-coupon bonds maturing in 1 year (time 1), 2 years (time 2), and 3 years
(time 3), respectively, are given in the following matrix:
⎡B (0, 3) = $0.88⎤
⎢B (0, 2) = $0.93⎥ (23.5)
⎢ ⎥
⎣B (0, 1) = $0.97⎦
These prices are taken from the common zero-coupon bond price data used in Example 21.1. To
reduce clutter, we drop the dollar signs from the tree.
568 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
■ After one period, bond prices can move up or down. We assume that this happens with the actual
probabilities q(0) = 0.3 and [1 – q(0)] = 0.7, respectively. The bond prices in the up state (labeled by
the subscript u) are
⎡B (1, 3)u = 0.9167⎤
⎢B (1, 2) = 0.9620⎥ (23.6a)
⎢ u ⎥
⎣B (1, 1)u = 1 ⎦
The bond prices in the down state (labeled by the subscript d) are
⎡B(1, 3)d = 0.8977⎤
⎢B(1, 2) = 0.9555⎥ (23.6b)
⎢ d ⎥
⎣B(1, 1)d = 1 ⎦
Note that at time 1, the bond maturing at time 1 pays $1 in both the up and down states and then
disappears from the market.
Risky Versus Riskless Securities
■ In this evolution, there are two riskless securities, the bond that matures at time 1 and the mma. Both
the two- and three-period maturity bonds are risky. This can be seen by computing their returns in
both the up and down states and noticing that they are different:
⎡[B(1, 3)u − B (0, 3)] /B (0, 3) = (0.9167 − 0.88) /0.88 = 0.0418⎤
⎢ ⎥
⎢[B(1, 2)u − B (0, 2)] /B (0, 2) = (0.9620 − 0.93) /0.93 = 0.0344⎥ (23.7a)
⎣[B(1, 1)u − B (0, 1)] /B (0, 1) = (1 − 0.97) /0.97 = 0.0309 ⎦
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
⎡[B(1, 3)d − B (0, 3)] /B (0, 3) = (0.8977 − 0.88) /0.88 = 0.0201⎤
⎢ ⎥
⎢[B(1, 2)d − B (0, 2)] /B (0, 2) = (0.9555 − 0.93) /0.93 = 0.0274⎥ (23.7b)
⎣[B(1, 1)d − B (0, 1)] /B (0, 1) = (1 − 0.97) /0.97 = 0.0309 ⎦
This observation will prove useful later in this section.
The lower part of Figure 23.2 gives the formalization of this example using the
symbols introduced in assumption A6.
THE ASSUMPTIONS 569
FIGURE 23.2: One Period Evolution of Bond Prices in A
Binomial Framework
Time 0 Time 1
Zero-Coupon Bond Prices (Example)
B(1, 3)u = 0.9167
B(1, 2)u = 0.9620 (“Up” state)
q(0) = 0.3 B(1, 1)u = 1
B(0, 3) = 0.88
B(0, 2) = 0.93 R(0) = 0.0309
B(0, 1) = 0.97
B(1, 3)d = 0.8977
B(1, 2)d = 0.9555 (“Down” state)
B(1, 1)d = 1
Zero-Coupon Bond Prices (Symbols)
u(0, 3)B(0, 3) = B(1, 3)u
u(0, 2)B(0, 2) = B(1, 2)u
q(0) 1 = B(1, 1)u
B(0, 3)
B(0, 2) R(0)
B(0, 1)
d(0, 3)B(0, 3) = B(1, 3)d
d(0, 2)B(0, 2) = B(1, 2)d
1 = B(1, 1)d
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Money Market Account (mma)
A(0) = 1 A(1) = [1 + R(0)] = 1.0309
THE ZERO-COUPON BOND PRICE EVOLUTION
■ Today is denoted by time 0. The zero-coupon bond prices are given by B(0, 1),
B(0, 2), and B(0, 3). They pay a dollar on their maturity dates: times 1, 2, and 3,
respectively. Only time 1 is shown in this evolution.
■ The actual probability of moving “up” is q(0) > 0.
■ At time 1, the prices of the zeros move to the up state (in which case they get
multiplied by the up factor) or the down state (where they get multiplied by the
down factor).
- If zero-coupon bond prices go up, the three-period bond has a new price of
B(1, 3)u , which is obtained by multiplying B(0, 3) by the up factor u(0, 3).
- Similarly, the two-period bond’s price is u(0, 2)B(0, 2) = B(1, 2)u .
570 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
- The one-period bond’s price is B(1,1)u = 1 because it matures, pays $1, and
disappears from the market.
- Note that the subscript u is attached to the prices of the zeros at time 1 to
distinguish them from the down state prices.
- If zero-coupon bond prices go down, the three-period bond attains a new price
of d(0, 3)B(0, 3) = B(1, 3)d . Similarly, d(0, 2)B(0, 2) = B(1, 2)d and B(1, 1)d = 1.
Here we use the subscript d for the down state.
THE SPOT RATE
■ Irrespective of the state that occurs, the shortest-maturity bond pays $1 at time
1. This bond is thus riskless and earns the spot rate:
R (0) = [1 − B (0, 1)] /B (0, 1)
= (1 − 0.97) /0.97 (23.8a)
= 0.0309 or 3.09 percent
THE MONEY MARKET ACCOUNT By definition, the mma grows at each
period’s spot rate of interest. If one invests $1 in the mma at time 0, it pays a dollar
return of [1 + R(0)] at time 1. Hence the mma’s values at times 0 and 1 are
A (0) = $1 and
(23.8b)
A (1) = 1 [1 + R (0)] = $1.0309
23.5 The Single-Period Model
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
We now present the single-period binomial HJM model. We apply this model to price
caplets, although as discussed later, the model is easily used to price other interest rate
derivatives as well. The first step in this methodology is to guarantee that the assumed
evolution is arbitrage-free.
Arbitrage-Free Evolutions
“The rain came down, the streams rose, and the winds blew and beat against that
house; yet it did not fall, because it had its foundation on the rock,” but the house
built on sand “fell with a great crash” (see Matthew 7:25 and 7:27, King James New
International Version). Similarly, we want to build a robust binomial tree that will not
crumble when used to price interest rate derivatives. The tree (the method) crumbles
if the tree’s evolution admits arbitrage. Why? Because how can one determine the
arbitrage-free price for a derivative when there exist arbitrage opportunities in the
prices of the zero-coupon bonds themselves? Obviously, one cannot. Hence we need
to guarantee that the assumed evolution is “built on rock” by being arbitrage-free.
The task is far more complex than it was in the binomial model for pricing equity
derivatives because here we have to ensure that a trader cannot make arbitrage profits
by trading any combination of these zero-coupon bonds.
NECESSARY CONDITIONS We can use our insights obtained from the
binomial model in Chapter 17 to determine the arbitrage-free conditions. In
THE SINGLE-PERIOD MODEL 571
Chapter 17, there was only one risky asset’s price on the binomial tree. Recall the
no-arbitrage condition using the notation of this chapter:
Up factor > 1 + R (0) > Down factor
Here, however, we have two risky bonds at each node. Hence, if there is no arbitrage,
then this condition must be true for both of the risky zero-coupon bonds B(0,2) and
B(0,3). Otherwise, just as argued in Chapter 17, one can form a portfolio of a risky
zero-coupon bond and the mma to form an arbitrage opportunity.
This logic implies then that a necessary condition for no arbitrage is
u (0, T) > 1 + R (0) > d (0, T) for T = 2, 3 (23.9a)
where u(0, T) and d(0, T) are the T-period zero-coupon bond’s up and down factors
from time 0 to time 1.
But this condition is not sufficient to guarantee that the tree is arbitrage-free. The
reason is because it only relates each risky bond to the mma. It does not compare
the two risky bonds’ returns to each other. Such a comparison will generate another
necessary condition, which, in conjunction with Equation 23.9a, will turn out to
be sufficient. Before determining that remaining condition, however, it is useful to
rewrite Equation 23.9a in an alternative form.
ALTERNATIVE NECESSARY CONDITIONS Simple logic shows that
Equation 23.9a is true if and only if there exists a real number 𝜋(0,T) with
0 < 𝜋(0, T) < 1 such that
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1 + R (0) = 𝜋 (0, T) u (0, T) + [1 − 𝜋 (0, T)] d (0, T) (23.9b)
for T = 2, 3. That is, the dollar return on the mma is a weighted average of
the up and down returns on the risky zero-coupon bonds. Note that the weight
𝜋(0, T) depends on the current time (which is given by the first argument 0) and the
particular risky bond under consideration (which is given by the second argument
T). Solving Equation 23.9b gives a formula for this weight:
1 + R (0) − d (0, T) u (0, T) − [1 + R (0)]
𝜋 (0, T) = and 1 − 𝜋 (0, T) =
( u (0, T) − d (0, T) ) ( u (0, T) − d (0, T) )
(23.9c)
for T = 2, 3. With this formula, the necessary condition for no arbitrage simplifies
to just checking to see if 0 < 𝜋(0, T) < 1 for each T.
Last, we can rewrite Equation 23.9b one more time. Dividing both sides of (23.9b)
by [1 + R(0)] and multiplying both sides by B(0, T) gives our final result:
𝜋 (0, T) B (0, T) u (0, T) + [1 − 𝜋 (0, T)] B (0, T) d (0, T)
B (0, T) = for T = 2, 3
1 + R (0)
(23.9d)
We will return to this condition shortly and explain its significance.
572 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
SUFFICIENT CONDITIONS The previous necessary conditions (23.9a)–(29.9d)
do not compare the dollar returns of the two risky bonds. To compare these dollar
returns, to see if there is any arbitrage embedded in their price evolutions, we can
compare the cash flows to the traded two-period zero with a synthetic two-period
zero, constructed using a portfolio of the three-period zero and the mma. If the price
of the traded two-period zero equals the cost of constructing the synthetic two-period
zero, then we know the dollar returns on the two- and three-period zeros admit no
arbitrage. It is shown in the appendix to this chapter that for this to be true,
𝜋 (0, 3) B (0, 2) u (0, 2) + [1 − 𝜋 (0, 3)] B (0, 2) d (2)
B (0, 2) = (23.10)
1 + R (0)
Equation 23.9d for the two-period zero is identical to Equation 23.10 but with 𝜋(0,2)
instead of 𝜋(0,3). This can only happen if
𝜋 (0, 3) = 𝜋 (0, 2) with 0 < 𝜋 (0, 2) < 1 (23.11)
which is our last necessary condition.
It turns out that Equation 23.11 is also a sufficient condition for the tree to be
arbitrage-free. The proof of this assertion is contained in the appendix to this chapter.
We summarize this discussion as a result (see Result 23.1).
RESULT 23.1
Equal Pseudo-probability Condition
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The single-period zero-coupon bond price evolution is arbitrage-free if and
only if there exists a real number 𝜋(0) strictly between 0 and 1 such that
𝜋 (0) = 𝜋 (0, T) for T = 2, 3 (23.12)
where 𝜋(0, T) = ([1 + R(0) – d(0, T)]/[u(0, T) – d(0, T)]), R(0) is the time
0 spot rate, and [u(0, T), d(0, T)] are the dollar returns on the T maturity
zero-coupon bonds in the up and down states, respectively.
Because 𝜋(0) is a number between zero and one, we can interpret it as a probability,
and we call it the pseudo-probability. We call Equation 23.12 the equal pseudo-
probability condition. Fortunately, it greatly simplifies the task of checking to see
if a zero-coupon bond price tree is arbitrage-free.
THE SINGLE-PERIOD MODEL 573
EXAMPLE 23.4: Verifying That a Single-Period Binomial Tree Is
Arbitrage-Free
Consider the term structure evolution given in Example 23.2 and Figure 23.2. To check whether this
evolution is arbitrage-free, we compute the up and down factors for the two zero-coupon bonds as they
evolve from time 0 to time 1.
■ For the zero-coupon bond maturing at time 3, we have
up factor, u (0, 3) = B (1, 3)u /B (0, 3) = 0.9167/0.88 = 1.0418 and
down factor, d (0, 3) = B (1, 3)d /B (0, 3) = 0.8977/0.88 = 1.0201
■ For the zero-coupon bond maturing at time 2, we have
up factor, u (0, 2) = B (1, 2)u /B (0, 2) = 0.9620/0.93 = 1.0344 and
down factor, d (0, 2) = B (1, 2)d /B (0, 2) = 0.9555/0.93 = 1.0275
Next, we compute the pseudo-probabilities for the two- and three-period zeros:
1 + R (0) − d (0, 3) 1.0309 − 1.0201
𝜋 (0, 3) = = = 0.5
( u (0, 3) − d (0, 3) ) (1.0418 − 1.0201)
1 + R (0) − d (0, 2) 1.0309 − 1.0275
𝜋 (0, 2) = = = 0.5
( u (0, 2) − d (0, 2) ) (1.0344 − 1.0275)
■ Because both these pseudo-probabilities are equal and strictly between 0 and 1, the single-period tree
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
is arbitrage-free.
■ For this tree, 𝜋(0) = 0.5 is the unique pseudo-probability of an up movement for zero-coupon bonds
maturing at times 2 and 3.
Zero-Coupon Bond Prices and Martingales
Using the new notation, under the no-arbitrage condition (23.12), we can rewrite
Equation 23.9d as
𝜋 (0) B(1, T)u + [1 − 𝜋 (0)] B(1, T)d
B (0, T) = for T = 2, 3 (23.13a)
1 + R (0)
Suppose we view the “probability” of B(0, T) going up as 𝜋(0) and the probability
of going down as [1 – 𝜋(0)], where T = 2 and 3. Then, this equation shows that
the zero-coupon bond’s expected discounted payoff using the pseudo-probabilities
is today’s zero-coupon bond price! This gives a method for computing the present
value of a zero-coupon bond.
Notice how similar this is to Equation 17.5 in Chapter 17. There the stock’s
price evolved as a martingale so that the stock price equaled its discounted expected
574 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
payoff at time 1 using the pseudo-probabilities. Now we do the same for both zero-
coupon bonds. The final step is applying the pseudo-probabilities to price interest
rate derivatives. It is hoped that you are beginning to understand why martingale
pricing lies at the heart of modern derivatives valuation.
EXAMPLE 23.5: Bond Prices as Expected Discounted Values
■ Consider the term structure evolution given in Example 23.3 and Figure 23.2. At time 1, the three-
period zero’s price is B(1,3)u in the up state with pseudo-probability 𝜋(0) and B(1,3)d in the down
state with pseudo-probability [1 – 𝜋(0)]. As a dollar invested in the mma today gives [1 + R(0)] at
time 1, application of Equation 23.13a gives
𝜋 (0) B(1, 3)u + [1 − 𝜋 (0)] B(1, 3)d 0.5 × 0.9167 + 0.5 × 0.8977
= = 0.88
1 + R (0) 1.0309
which matches B(0,3), today’s price of the zero-coupon bond maturing at time 3.
■ Applying the same formula to the two-period zero whose prices at time 1 are B(1, 2)u and
B(1, 2)d with pseudo-probabilities 𝜋(0) and [1 – 𝜋(0)], respectively, we get
𝜋 (0) B(1, 2)u + [1 − 𝜋 (0)] B(1, 2)d 0.5 × 0.9620 + 0.5 × 0.9555
= = 0.93
1 + R (0) 1.0309
which matches B(0,2), today’s price of the zero-coupon bond maturing at time 2.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Using now familiar abstract notation, we rewrite these equations as the following
result (see Result 23.2).
RESULT 23.2
Risk-Neutral Valuation for Zero-Coupon Bond Prices
The time 0 value of a zero-coupon bond is
E𝜋 [B (1, T)]
B (0, T) = for T = 2, 3 (23.13b)
1 + R (0)
where E𝜋 [.] denotes expectation using the pseudo-probability 𝜋(0), R(0) is
the time 0 spot rate, and B(1, T) is the T maturity zero-coupon bond’s price
at time 1.
This result shows that the time 0 price of a zero-coupon bond equals its expected
discounted time 1 price, where the discount rate is the spot rate and the expectation
is computed using the pseudo-probabilities.
THE SINGLE-PERIOD MODEL 575
Finally, using the notation for the value of the mma also gives
B (0, T) B (1, T)
= E𝜋 for T = 2, 3 (23.13c)
A (0) ( A (1) )
This equation shows that the zero-coupon bond’s price, normalized by the value of
the mma, is a martingale under the pseudo-probabilities.
Alternatively stated, a necessary and sufficient condition for the tree to be
arbitrage-free is the existence of a real number 𝜋(0) strictly between 0 and 1
such that the normalized zero-coupon bond prices are martingales under this
pseudo-probability. It is this form of the no-arbitrage condition that readily generalizes to
more complex models for the evolution of the term structure of interest rates.
Understanding the Equal Pseudo-probability
Condition
The equal pseudo-probability condition has a nice economic interpretation. To
understand this interpretation, we note the obvious fact that
𝜋 (0, 3) = 𝜋 (0, 2) holds if and only if [1 − 𝜋 (0, 3)] = [1–𝜋 (0, 2)] (23.14a)
Writing this equation out in full using Result 23.1, we see that the tree is arbitrage-
free if and only if
u (0, 3) − [1 + R (0)] u (0, 2) − [1 + R (0)]
= (23.14b)
( u (0, 3) − d (0, 3) ) ( u (0, 2) − d (0, 2) )
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
First, note that the numerator in Equation 23.14b is the compensation for bearing
the risk of a zero-coupon bond versus holding the mma. This is because the possibility
of receiving the larger return on the risky zero in the up state u(0, T) is the reason
one holds a risky bond instead of the mma. Figure 23.3 illustrates this difference.
FIGURE 23.3: The Bond’s Risk Premium
u(0,T)
reward
$1 R(0)
d(0,T)
risk
576 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
Second, the denominator measures the spread, or the risk, of a zero-coupon bond.
Indeed, the larger is the distance between the up and down returns on the zero-
coupon bond, the riskier is the payoff, because both the gains and losses are magnified.
Hence the distance between the largest and worst dollar returns on a zero-coupon
bond is a measure of risk.
Combined, the ratio measures the reward per unit of risk, called the risk premium.
The no-arbitrage condition can thus alternatively be stated as follows: the tree is
arbitrage-free if and only if all zero-coupon bonds earn the same risk premium. This last
observation makes good economic sense. If the zero-coupon bonds had different risk
premiums, traders would desire to hold the zero with the largest compensation per
unit of risk. In fact, if a trader also cleverly shorts a bond with a smaller compensation
per unit of risk, all of the interest rate risk can be removed from the resulting portfolio
to generate an arbitrage opportunity. This is the logic underlying Result 23.2.
Actual versus Pseudo-probabilities
As in Chapter 17, the pseudo-probabilities 𝜋(0) and 1 – 𝜋(0) for the up and down
movements in the tree are both strictly positive if and only if the actual probabilities
q(0) and 1 – q(0) are strictly positive. This condition implies that there exist two
strictly positive numbers 𝜙u and 𝜙d such that
𝜋 (0) = 𝜙u q (0) and
1 − 𝜋 (0) = 𝜙d [1 − q (0)]
These numbers 𝜙u and 𝜙d transform the actual probabilities to the pseudo-
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
probabilities, which are useful for pricing derivatives using the technique of risk-
neutral valuation (see below). The quantities 𝜙u and 𝜙d adjust the actual probabilities
for the risk of the cash flows when computing the arbitrage-free price as a discounted
expected value. The proof of this statement is identical to the proof given for the
binomial model of Chapter 17 (see the appendix to that chapter) and is therefore
omitted.
Caplet Pricing
Having built this structure, it is now easy to price a caplet using the single-period
HJM model. A caplet is a European call option on the spot rate of interest. Thus a
caplet’s payoff with maturity date 2 and strike rate k is given by
c (2) = max [R (1) − k, 0] (23.15a)
Note that the caplet’s payoff at time 2 is based on the spot rate at time 1. This is by
market convention.
THE SINGLE-PERIOD MODEL 577
Because this payoff is known at time 1, the present value of this payoff at time 1 is
obtained by discounting (23.15a) by the spot rate R(1):
max [R (1) − k, 0]
c (1) = (23.15b)
1 + R (1)
Thus, we can alternatively view the caplet as having the maturity date 1 with this revised
payoff. This is justified because all of the payoff ’s uncertainty is resolved at date 1. As
in this example, to facilitate computation, we introduce the following convention in
this and the next chapter.
Caplet and Floorlet Maturity Convention: For valuation purposes, we consider
a caplet’s and floorlet’s maturity as being one time step before the actual payment date
with a payoff equal to the present value of the actual payment one time step later.
The caplet’s payoffs are illustrated in Figure 23.4. We next show how to price this
caplet, first with an example and then in general .
FIGURE 23.4: A Caplet Evolution
Time 0 Time 1
Example
0.0395 – 0.04
c(1)u = max 0, =0
1.0395
c(0) = 0.003
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
0.0465 – 0.04
c(1)d = max 0, = 0.0062
1.0465
Symbols
R(1)u – k
c(1)u = max 0,
1 + R(1)u
c(0)
R(1)d – k
c(1)d = max 0,
1 + R(1)d
EXAMPLE 23.6: Pricing a Caplet
■ As in the binomial model of Chapter 17, we construct a portfolio consisting of a zero-coupon bond and
mma to replicate the traded caplet’s payoffs in both the up and down states. The cost of constructing
this synthetic caplet is the arbitrage-free price of the traded caplet.
578 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
■ Using the zero-coupon bond price evolution in Figure 23.2, we can compute the values for the spot
rate at time 1 in the up and down states, R(1)u and R(1)d , respectively:
R (1)u = 1/B (1, 2)u − 1 = 1/ (0.9620) − 1 = 0.0395
R (1)d = 1/B (1, 2)d − 1 = 1/ (0.9555) − 1 = 0.0465
The Caplet’s Time 1 Payoffs
■ Consider the caplet of Figure 23.4 maturing at time 1 with strike rate k = 0.04 per year. In the up
state, the caplet ends out-of-the-money with zero value, as the following computation shows:
cu = max [(R (1)u − k) / (1 + R (1)u ) , 0]
(23.16a)
= max [(0.0395 − 0.04) / (1.0395) , 0] = 0
In the down state, the caplet ends up in-the-money with value
cd = max [(R (1)d − k) / (1 + R (1)d ) , 0]
(23.16b)
= max [(0.0465 − 0.04) / (1.0465) , 0] = $0.0062
Constructing the Synthetic Caplet
■ Form a portfolio using m shares of the two-period zero-coupon bond with price B(0, 2) and n units
of the mma with value A(0). The cost of construction is
V (0) ≡ mB (0, 2) + nA (0) (23.17a)
At time 1, we want to find (m, n) to match the portfolio’s value to the traded caplet’s value in each
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
state. This yields two equations in two unknowns:
In the up state ∶ V (1)u ≡ mB (1, 2)u + nA (1) = c (1)u
(23.17b)
or m0.9620 + n1.0309 = 0
In the down state ∶ V (1)d ≡ mB (1, 2)d + nA (1) = c (1)d
(23.17c)
or m0.9555 + n1.0309 = 0.0062
Subtracting Equation 23.17c from Equation 23.17b gives
m × (0.9620 − 0.9555) = −0.0062
or m = −0.9556
Plugging m’s value into Expression 23.17b and solving we get
−m0.9620
n= = 0.8917
1.0309
THE SINGLE-PERIOD MODEL 579
■ So by going short 0.9556 of the two-period zero-coupon bond and buying 0.8917 units of the mma,
we can construct a synthetic caplet. The cost of construction is
V (0) ≡ mB (0, 2) + n
= −0.9556 × 0.93 + 0.8917 = $0.003
This is the arbitrage-free price of the traded caplet.
■ One can repeat this construction using the mma and the three-period zero. Try it and see that since
the term structure evolution is arbitrage-free, both approaches give the same caplet price.
The Hedge Ratio (the Holy Grail)
The holdings in the risky bond, m, is called the caplet’s delta or the hedge ratio, the holy grail of option
pricing. It is the ratio of change in the caplet’s price to a change in the zero-coupon bond’s price.
Looking at Equation 23.17b and 23.17c, we see that if one holds the traded caplet and shorts m units
of the two-period zero-coupon bond, the resulting portfolio will be riskless and earn the mma’s rate of
return.
An Arbitrage Opportunity
Suppose that the market price of the caplet is $0.005. Then, an arbitrage opportunity exists. The arbitrage
opportunity is to sell the traded call for $0.005 and buy the synthetic call at a cost of $.003. At time 1,
the payoffs to the short traded call and the long synthetic call are equal in magnitude and opposite in
sign, so the net payoff is zero.
We can formalize this example to obtain some general conclusions regarding our
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
method for pricing caplets.
THE CAPLET’S TIME 1 PAYOFFS First, we note that the caplet’s time 1 payoffs
are (see Figure 23.4)
R(1)u − k
In the up state ∶ c (1)u = max 0, (23.18a)
[ 1 + R(1)u ]
R(1)d − k
In the down state ∶ c (1)d = max 0, (23.18b)
[ 1 + R(1)d ]
CONSTRUCTING THE SYNTHETIC CAPLET We can create a synthetic
caplet with cost of construction
V (0) = mB (0, 2) + nA (0)
with m shares of the two-period zero worth B(0,2) and n units of the mma priced at
A(0) = 1.
580 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
Next, at time 1, we match this portfolio’s value with that of the caplet’s:
In the up state ∶ V(1)u = mu (0, 2) B (0, 2) + n [1 + R (0)] = c (1)u (23.19a)
In the down state ∶ V(1)d = md (0, 2) B (0, 2) + n [1 + R (0)] = c (1)d (23.19b)
Solving for m and n yields
c (1)u − c (1)d
m= (23.20a)
[u (0, 2) − d (0, 2)] B (0, 2)
u (0, 2) c (1)d − d (0, 2) c (1)u
n= (23.20b)
[u (0, 2) − d (0, 2)] [1 + R (0)]
To avoid arbitrage, the cost of constructing the synthetic caplet portfolio must
equal the traded caplet’s price. Substitution of m and n into this equation and
simplification, therefore, gives the arbitrage-free price of the caplet:
c (0) = mB (0, 2) + n
c (1)u − c (1)d u (0, 2) c (1)d − d (0, 2) c (1)u
= B (0, 2) +
[u (0, 2) − d (0, 2)] B (0, 2) [u (0, 2) − d (0, 2)] [1 + R (0)]
(23.21)
Notice that this caplet’s value does not depend on the actual probability q(0); it only
depends on the up and down factors for the two-period zero [u(0, 2), d(0, 2)], the
two-period zero’s price B(0, 2), and the spot rate of interest R(0). This is similar to
what we found in the binomial option pricing model in Chapter 17.
This is an important observation. It implies that if two investors agree on these
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
inputs, but disagree on the actual probability of going up, q(0), they will still agree
on the caplet’s price. Of course, disagreement about the actual probability of going
up q(0) might lead to a disagreement about the initial zero’s price B(0, 2) given fixed
[u(0, 2), d(0, 2), R(0)]. However, if agreement about the zero’s price B(0, 2) occurs,
then the caplet’s value is fixed by our no-arbitrage argument, and the two investors
will agree on the caplet’s price.
This omission occurs because in the derivation of the caplet’s price, we use exact
replication. No matter to which node of the tree the zero’s price moves, the synthetic
caplet’s payoffs exactly duplicate the traded caplet’s payoffs. Hence we are indifferent
to what the actual probability is of moving up or down.
Finally, we note that the synthetic construction illustrated uses the two-period
bond and the mma. An analogous synthetic construction could have been done using
the three-period bond and the mma. This construction generates different holdings
in the three-period bond and mma, as distinct from those in the two-period bond
and mma, but the arbitrage-free price is identical. This is guaranteed by the equal
pseudo-probability condition given by Equation 23.12.
THE SINGLE-PERIOD MODEL 581
The Hedge Ratio (the Holy Grail)
The method for constructing the synthetic caplet also yields the method for hedging
the risk of a long caplet position. Determining this hedge ratio, of course, is the holy
grail of option pricing (see Chapter 17). If one shorts m shares of the two-period zero
for each caplet held long, the resulting portfolio is riskless. The holdings in the risky
bond, m, is called the caplet’s delta or hedge ratio. To see why the portfolio is riskless,
we just need to rearrange the terms in Equation 23.21 to obtain the time 0 hedged
caplet portfolio:
c (0) − mB (0, 2)
Its payoffs at time 1 are
c (1)u − c (1)d
In the up state ∶ c (1)u − mu (0, 2) B (0, 2) = c (1)u − u (0, 2)
[u (0, 2) − d (0, 2)]
u (0, 2) c (1)d − d (0, 2) c (1)u
= = n [1 + R (0)]
[u (0, 2) − d (0, 2)]
c (1)u − c (1)d
In the down state ∶ c (1)d − md (0, 2) B (0, 2) = c (1)d − d (0, 2)
[u (0, 2) − d (0, 2)]
u (0, 2) c (1)d − d (0, 2) c (1)u
= = n [1 + R (0)]
[u (0, 2) − d (0, 2)]
In both the up and down states, the portfolio has the same value and is therefore
riskless. As shown, it earns the spot rate of interest. The hedge ratio, m, given by
Equation 23.20a therefore removes all of the interest rate risk from the hedged caplet
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
portfolio.
Risk-Neutral Valuation
We can rearrange Equation 23.21 for the caplet’s arbitrage-free price to obtain another
insight. Some algebra yields
1 1 + R (0) − d (0, 2) u (0, 2) − [1 + R (0)]
c (0) = c (1)u + c (1)d
[1 + R (0)] [( u (0, 2) − d (0, 3) ) ( u (0, 2) − d (0, 2) ) ]
Using the definition of the pseudo-probability in Result 23.1 under the no-arbitrage
condition, we can rewrite this equation abstractly as
𝜋 (0) c (1)u + [1 − 𝜋 (0)] c (1)d
c (0) = (23.22a)
1 + R (0)
The caplet’s arbitrage-free price is seen to be its expected discounted value using the
pseudo-probabilities. This gives a present value formula. The pseudo-probabilities
adjust for the risk of the caplet’s payoff, and therefore one can use the riskless rate
582 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
to discount to the present. As in Chapter 17, if one uses this equation to value the
caplet, then it is called risk-neutral valuation.
Using familiar abstract notation, we can summarize this result (see Result 23.3).
RESULT 23.3
Risk-Neutral Valuation for a Caplet
A caplet with strike rate k and maturing at time 1 is given by the formula
E𝜋 [c (1)]
c (0) = (23.22b)
1 + R (0)
where E𝜋 [ . ] denotes expectation using the pseudo-probability 𝜋(0), R(0) is
the time 0 spot rate, and B(1, T) is the T maturity zero-coupon bond’s price
at time 1.
As noted earlier, the caplet’s value does not depend on the actual probability
of moving to the up state. It only depends on the pseudo-probability. Because
of Equation 23.22b, the pseudo-probability is often called the risk-neutral
probability.
We note in passing that this more abstract equation will prove useful in deriving
a put–call parity theorem for caplets and floorlets analogous to that for ordinary calls
and puts.
Risk-neutral valuation greatly simplifies the computation of the caplet’s arbitrage-
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
free price, as the following example shows (see Example 23.7).
EXAMPLE 23.7: Risk-Neutral Valuation
Using the binomial tree in Figure 23.2 and the caplet payoffs from Figure 23.4, we get
𝜋 (0) c (1)u + [1 − 𝜋 (0)] c (1)d 0.5 × 0 + 0.5 × 0.0062
c (0) = = = 0.003
1 + R (0) 1.0309
As illustrated, using the risk-neutral valuation method is much easier than constructing the synthetic
caplet and solving the two equations for the two unknowns—and it gives the same value, which should
not be a surprise!
Valuing a Floorlet
Risk-neutral valuation can also be used to value a floorlet. Alternatively, the price of
a floorlet can be obtained from that of a caplet by using a put–call parity relationship.
This is given by the following result (see Result 23.4).
THE SINGLE-PERIOD MODEL 583
RESULT 23.4
Caplet–Floorlet Parity
[f (0, 1) − k] B (0, 2) = c (0) − p (0) (23.23)
where time 0 is today, time 1 is the maturity date, f (0,1) is the time 0 forward
rate for date 1, B(0,2) is today’s price of a zero-coupon maturing at time 2,
and c(0) and p(0) are today’s prices of a caplet and a floorlet, respectively, with
identical terms and conditions including strike rate k and maturity date 1.
This relation is proved in the appendix to this chapter. Caplet–floorlet parity shows
that a position long a caplet and short a floorlet equals a long position in [f (0,1) – K]
zero-coupon bonds, represented by the left side of this equation. Given the value of a
caplet, the value of the floorlet follows directly from this equation. The next Example
23.8 illustrates the use of caplet–floorlet parity for valuing a floorlet.
EXAMPLE 23.8: Valuing a Floorlet
■ Consider a floorlet that has strike rate K = 0.04 per year and maturity date 2. Its time 2 payoff is
p (2) = max [k − R (1) , 0]
■ The floorlet is a European put option on the spot rate of interest. Note that the floorlet’s payoff at
time 2 is based on the spot rate at time 1. Because this payoff is known at time 1, the present value of
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
this payoff at time 1 is obtained by discounting by the spot rate:
max [k − R (1) , 0]
p (1) =
1 + R (1)
As with the caplet, since the floorlet’s payoff is known at time 1, using our convention, we consider
time 1 as the maturity of the floorlet.
■ Let the term structure evolution be as in Figure 23.2 and Example 23.3. Consider a caplet with the
same maturity and the same strike rate as the floorlet’s. We determined the time 0 price of this caplet as
c(0) = 0.003 dollars in Example 23.7. The forward rate at time 0 for date 2 is given by Equation 23.4 as
B (0, 1) 0.97
f (0, 1) = −1= − 1 = 0.0430
B (0, 2) 0.93
■ Hence, using caplet–floorlet parity, we have that
p (0) = c (0) − [f (0, 1) − k] B (0, 2)
= 0.003 − (0.0430 − 0.04) 0.93 = $0.0002
which is the floorlet’s value at time 0. You can check to see that using risk-neutral valuation gives
the same answer, as it should!
584 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
Valuing Various Interest Rate Derivatives
The method we have developed for pricing caplets via risk-neutral valuation is
surprisingly general. It can be applied to any interest rate derivative to obtain its
arbitrage-free price. The only modification needed is to replace the caplet’s time
1 payoff with that of the alternative interest rate derivative. We will show how to
apply this technique to price American options, swaptions, and futures contracts
after studying the two-period model in Chapter 24. We cannot study them here
because in a single period model, the complexity of these securities disappear: futures
contracts become forward contracts, American options become European options,
and swaptions become deterministic.
Multiple Factors
The single-period binomial HJM model is called a one-factor model because at
each node of the tree, only one of two possibilities can happen: up and down. A
two-factor model has three branches at each node in the tree: up, middle, and
down; a three-factor model has four branches at each node, and so forth.
A factor corresponds to an economic shock affecting the evolution of the term
structure of interest rates. Such shocks are due to the natural randomness in an
economy resulting from the ordinary economic activities of production, consump-
tion, technological advances, population growth, and government regulatory policy
(monetary and fiscal). One can understand these random shocks by thinking about
how to simulate an evolution through an HJM tree.
For a one-factor tree, starting at time 0 and progressing forward, flipping a single
coin at each node generates a path through the tree. Of course, the coin would need
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
to be biased with the probability of heads being the actual probability of moving up.
A coin is the simplest random generating device one can think of. Because only
one random generating device is needed to create a path through a binomial tree,
it is called a one-factor model. For a two-factor model, at each node in the tree, one
needs to flip two coins to decide which branch to take. Recall that a two-factor
model has three branches at each node. One coin is needed to decide between up
and “the remaining two branches.” A second coin is needed to choose between the
remaining two branches: middle or down. Hence you need two of these simplest
random generating devices to simulate an evolution through a two-factor HJM tree.
Each random generating device is a “factor.” The same logic extends for three or
more factors.
The valuation and hedging methods for multifactor models are straightforward
extensions of those illustrated for the one-factor model. The only difference is that
more branches need to be considered, and hedging requires more zero-coupon bonds
at each time step because there is more risk (see Jarrow 2002b).
SUMMARY 585
When modeling the evolution of the term structure of interest rates, one should
include enough factors to capture all of the relevant randomness affecting the
economy. Macroeconomics can help in understanding the sources of this randomness,
as discussed earlier. Researchers have estimated multifactor models and discovered
that at least three or four factors are needed to provide term structure evolutions
consistent with historical experience.
23.6 Summary
1. The basic equity derivatives are forwards, futures, calls, and put options. The basic
interest rate derivatives are forwards (called forward rate agreements), futures, call
options (called caplets), and put options (called floorlets). Caplets and floorlets do
not trade, but portfolios of caplets, called caps, and portfolios of floorlets, called
floors, trade over-the-counter.
2. The single-period binomial HJM model assumes that the zero-coupon bond prices
B(t, T) for T = 1, 2, 3 and t = 0, 1 trade in discrete time and that their prices
evolve according to a binomial process. First, if the zero-coupon bond matures the
next period, that is, at time 1, then the current bond price goes to $1 regardless of
the state. Second, the current T > t + 1 maturity zero-coupon bond price either
gets multiplied by an up factor u(t, T) and takes the price u(t, T) × B(t, T) with
the actual probability q(t) > 0, or it is multiplied by a down factor d(t, T) and takes
the price d(t, T) × B(t, T) with probability 1 – q(t).
3. The single-period zero-coupon bond price binomial tree is arbitrage-free if and
only if there exist pseudo-probabilities 𝜋(0) strictly between 0 and 1 such that
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
𝜋 (0) = 𝜋 (0, T) for T = 2, 3
1 + R (0) − d (0, T)
where 𝜋 (0, T) =
[ u (0, T) − d (0, T) ]
4. Consider a caplet maturing at time 2 with a strike rate of K. The present value of
its payoff at time 1 is c(1) = max [0, (R(1) – k)/(1 + R(1))]. Because this payoff is
known at time 1, we consider the caplet’s maturity as time 1, and not time 2. The
arbitrage-free price of a caplet can be determined using risk-neutral valuation
c(0) = E𝜋 [c(1)]/[1 + R(0)], where E𝜋 [·] denotes expectation using the pseudo-
probabilities. The hedge ratio m = [c(1)u – c(1)d ]/{[u(0,2) – d(0,2)]B(0,2)} can be
used to form a riskless portfolio consisting of a long position in the caplet and
shorting this number of two-period zero-coupon bonds.
5. Caplet–floorlet parity is
[f (0, 1) − k] B (0, 2) = c (0) − p (0)
where f (0, 1) is the time 0 forward rate for date 1, B(0, 2) is the time 0 zero-
coupon price with maturity at time 2, and both the caplet with time 0 price c(0)
and the floorlet with time 0 price p(0) have strike price k and maturity 1.
586 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
23.7 Appendix
The Equal Pseudo-probability Condition
As with the binomial model for pricing equity options (see Chapter 17), we can
trade two assets to synthetically construct the price of a third asset. Closely related to
this synthetic construction is the equivalence of the pseudo-probabilities, which the
following proof shows.
Proof of Necessary Condition
■ We first show how to construct the evolution of the traded two-period zero-
coupon bond’s price using a portfolio of the mma and the three-period zero-
coupon bond.
■ The payoffs to the traded two-period zero at time 1 in the up and down states are
u(0, 2)B(0, 2) and d(0, 2)B(0, 2), respectively.
■ Let us construct a portfolio consisting of n units of the mma and m shares of the
three-period bond with time 0 value V (0) ≡ mB(0, 3) + nA(0) such that in both
the up and down states, the portfolio’s value matches the price of the two-period
zero, that is,
V(1)u = mB(1, 3)u + nA (1) = u (0, 2) B (0, 2)
and
V(1)d = mB(1, 3)d + nA (1) = d (0, 2) B (0, 2)
■ Using the notation from Figure 23.2, we can rewrite this system of equations as
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
mu (0, 3) B (0, 3) + n [1 + R (0)] = u (0, 2) B (0, 2)
mu (0, 3) B (0, 3) + n [1 + R (0)] = d (0, 2) B (0, 2)
Solving this set of equations gives
[u (0, 2) − d (0, 2)] B (0, 2) [u (0, 3) d (0, 2) − u (0, 2) d (0, 3)] B (0, 2)
m= and n =
[u (0, 3) − d (0, 3)] B (0, 3) [u (0, 3) − d (0, 3)] [1 + R (0)]
Substituting these back into V (0) gives the cost of construction as
[u (0, 2) − d (0, 2)] B (0, 2) [u (0, 3) d (0, 2) − u (0, 2) d (0, 3)] B (0, 2)
V (0) = B (0, 3) + A (0)
[u (0, 3) − d (0, 3)] B (0, 3) [u (0, 3) − d (0, 3)] [1 + R (0)]
Algebra yields
1 + R (0) − d (0, 3) u (0, 3) − (1 + R (0))
B (0, 2) u (0, 2) + B (0, 2) d (0, 2)
( u (0, 3) − d (0, 3) ) ( u (0, 3) − d (0, 3) )
V (0) =
1 + R (0)
APPENDIX 587
Finally, using Equation 23.9c, we can write the cost of construction as
𝜋 (0, 3) B (0, 2) u (0, 2) + [1 − 𝜋 (0, 3)] B (0, 2) d (0, 2)
V (0) = (1)
1 + R (0)
■ No arbitrage implies that the cost of constructing the synthetic two-period zero
must equal the price of the traded two-period zero, that is,
𝜋 (0, 3) B (0, 2) u (0, 2) + [1 − 𝜋 (0, 3)] B (0, 2) d (0, 2)
B (0, 2) = (2)
1 + R (0)
Equation 2 is the third necessary condition (see Equation 23.10).
Proof of Sufficiency
■ To we note that Equation 23.12 implies that Equation 23.13b holds. We use this
condition subsequently.
■ Let condition 23.12 hold: 𝜋(0) = 𝜋(0, T) for T = 2, 3 in our interest rate binomial
tree. We suppose that an arbitrage opportunity exists and then generate a contra-
diction of Equation 23.12, thereby proving that the evolution is arbitrage-free.
■ Consider an arbitrage portfolio consisting of a units of the mma, b units of the
two-period zero, and c units of the three-period zero-coupon bond. Then, by the
definition of an arbitrage trading strategy this portfolio has a zero initial investment:
a + bB (0, 2) + cB (0, 3) = 0 (3)
■ After one period, the arbitrage portfolio’s values must be non-negative always and
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
one must be strictly positive, that is:
In the up state ∶ aA (1) + bB(1, 2)u + cB(1, 3)u ≥ 0
In the down state ∶ aA (1) + bB(1, 2)d + cB(1, 3)d ≥ 0
where at least one of the last two inequalities is strictly positive.
■ Dividing the last two equations by A(1) and taking expectations using the pseudo-
probability 𝜋(0) gives
B (1, 2) B (1, 3)
E𝜋 a + b +c >0 (4)
( A (1) A (1) )
But using Equation 23.13b on this last equation (which follows from Equation
23.12) yields a contradiction of Equation 3:
a + bB (0, 2) + cB (0, 3) > 0 (5)
Thus, when unique pseudo-probabilities exist, we cannot have an arbitrage
opportunity in the tree.
588 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
Proof of Caplet and Floorlet Parity
To prove caplet and floorlet parity in the single-period model, consider a caplet and
floorlet both with maturity 1 and strike rate k. Their time 2 payoffs are
c (2) = max [0, R (1) − k] and
p (2) = max [0, k − R (1)]
where R(1) is the spot rate.
It is easy to show that the following equation holds:
R (1) − k = max [0, R (1) − k] − max [0, k − R (1)] = c (2) − p (2) (6)
One can easily verify that this identity is true by testing values for the spot rate that
are greater than or less than the common strike rate k.
Add and subtract 1 on the left side and divide by the mma’s time 2 value, A(2).
This gives
[1 + R (1)] /A (2) − (1 + k) /A (2) = max [0, R (1) − k] /A (2) − max [0, k − R (1)] /A (2)
or [1 + R (1)] /A (2) − (1 + k) /A (2) = c (2) /A (2) − p (2) /A (2)
(7)
Take expectations of (7) with respect to the pseudo-probability to get
1 + R (1) 1 c (2) p (2)
E𝜋 − (1 + k) E𝜋 = E𝜋 − E𝜋 (8)
[ A (2) ] [ A (2) ] [ A (2) ] [ A (2) ]
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The risk-neutral valuation method implies that
1
B (0, 2) = E𝜋 (9a)
[ A (2) ]
1 + R (1) 1
E𝜋 = E𝜋 = B (0, 1) (9b)
[ A (2) ] [ A (1) ]
c (2)
c (0) = E𝜋 (9c)
[ A (2) ]
p (2)
p (0) = E𝜋 (9d)
[ A (2) ]
Substitution of Equations 9a–9d in Equation 8 yields the result
B (0, 1) − (1 + k) B (0, 2) = c (0) − p (0) (10)
but the left side of this equation equals
B (0, 1)
− 1 B (0, 2) − kB (0, 2) = [f (0, 1) − k] B (0, 2)
[ B (0, 2) ]
QUESTIONS AND PROBLEMS 589
where f (0,1) is obtained from the definition of forward rate (see Equation 23.4).
Substitution yields the final result:
[f (0, 1) − k] B (0, 2) = c (0) − p (0) (11)
23.8 Cases
Spencer Hall (A) (European Case Clearing House Case 9A95B045). The case ana-
lyzes issues surrounding the hedging of interest rate risk with instruments like
interest rate swaps, caps, and collars.
Spencer Hall (B) (European Case Clearing House Case 9A97N004). An extension
of the previous case (Spencer Hall [A]), the case considers issues surrounding the
cancellation of an interest rate cap by accepting a premium payment.
Markborough Properties Inc. (Ivey Publishing Case 9A87B002, European Case
Clearing House Case). The case examines how a company can hedge its interest
rate risk by using derivatives like interest rate swaps and caps.
23.9 Questions and Problems
23.1. Show that the following tree is not arbitrage-free.
FIGURE 1: Zero-coupon Bond Price Evolution
Time 0 Time 1
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Zero-Coupon Bond Prices
B(1, 3)u = 0.9167
B(1, 2)u = 0.9620 (“Up” state)
B(1, 1)u = 1
q(0) = 0.3
B(0, 3) = 0.83
B(0, 2) = 0.93 R(0) = 0.0309
B(0, 1) = 0.97
B(1, 3)d = 0.8977
B(1, 2)d = 0.9555 (“Down” state)
B(1, 1)d = 1
Money Market Account (mma)
A(0) = 1 A(1) = [1 + R(0)] = 1.0309
23.2. Compute the forward rate evolution implied by Figure 23.2 (this is identical
to the preceding tree with B[0,3] = 0.88).
590 CHAPTER 23: SINGLE-PERIOD BINOMIAL HEATH–JARROW–MOR TON MODEL
23.3. In the caplet example 23.6, construct the synthetic caplet using the three-
period zero and mma (the caplet has a strike rate of k = 0.04 or 4 percent).
Show that the cost of construction is 0.003, the same as that with the two-
period zero-coupon bond.
23.4. In the floorlet example 23.7, value the floorlet using risk-neutral valuation
and show you get the same answer as in the text.
23.5. In the caplet example 23.6, suppose the traded caplet has a market price of
$0.002. What is a trading strategy that will generate an arbitrage opportunity?
23.6. In the caplet example 23.6, suppose you have a long position in the caplet.
Form a hedged portfolio of the caplet and the two-period zero-coupon bond
that has no interest rate risk.
23.7. Use Figure 23.2. Consider a European call option on the three-period zero-
coupon bond with maturity time 1 and strike price of k = $0.90. Compute
the value using risk-neutral valuation.
23.8. Use Figure 23.2. Consider a European put option on the 3-period zero-
coupon bond with maturity time 1 and strike price of K = $0.90. Compute
the value using risk-neutral valuation.
FIGURE 2: Zero-coupon Bond Price
Evolution
B(1, 3)u = 0.932788
B(1, 2)u = 0.973744
B(1, 1)u = 1
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
B(0, 3) = 0.897710
B(0, 2) = 0.942596
B(0, 1) = 0.970874
B(1, 3)d = 0.916495
B(1, 2)d = 0.968004
B(1, 1)d = 1
23.9. For the single period evolution in this figure, compute the spot rates at time 0
and at time 1 in both the up and down states. Prove that this tree is arbitrage-
free (ignore round errors beyond two decimal places in the computation of
the pseudo-probabilities).
23.10. For the single-period evolution in the preceding figure, consider a caplet with
maturity time 1 with strike rate k = 0.03.
a. What is the value of this caplet at time 0?
b. What is the delta for this caplet in terms of the three-period zero-coupon
bond?
QUESTIONS AND PROBLEMS 591
23.11. For the single-period evolution in the preceding figure, what is the value of a
floorlet with maturity time 1 and strike rate 0.03? Compute the value using
risk-neutral valuation.
23.12. For the caplets and floorlets in examples 23.10 and 23.11, show that they
satisfy caplet-floorlet parity.
23.13. For the single-period evolution given in the preceding figure, consider a
European call option with maturity date 1 and strike price K = $0.92 on
the three-period zero-coupon bond. What is the arbitrage-free price of this
call option?
23.14. For the single-period evolution given in the preceding figure, consider an
American call option with maturity 1 and strike price K = $0.92 on the
three-period zero-coupon bond. What is the arbitrage-free price of this call
option? Is early exercise (at time 0) optimal?
23.15. For the single-period evolution given in the preceding figure, consider a
European put option with maturity 1 and strike price K = $0.92 on the
three-period zero-coupon bond. What is the arbitrage-free price of this put
option?
23.16. For the single-period evolution given in the preceding figure, consider a
European put option with maturity date 1 and strike price K = $0.94 on
the three-period zero-coupon bond. What is the arbitrage-free price of this
put option?
23.17. For the single-period evolution given in the preceding figure, consider an
American put option with maturity date 1 and strike price K = $0.94 on the
three-period zero-coupon bond. What is the arbitrage-free price of this put
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
option? Is early exercise (at time 0) optimal?
23.18. Why do you think caps and floors trade as portfolios of caplets and floorlets in
interest rate markets, whereas in equity markets, calls and puts directly trade?
23.19. Why is it that Eurodollar futures and forward rate agreement rates are not
equal?
23.20. When one compares forward rate agreement contracts to Eurodollar futures
contracts, which have greater counterparty risk? Do you think this might
influence the difference between FRA and futures rates? If yes, why?
24
Multiperiod Binomial
HJM Model
24.1 Introduction 24.5 Forwards, Futures, and
Swaptions
24.2 The Assumptions
Forward Rate Agreements
24.3 The Two-Period Model
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Eurodollar Futures
Arbitrage-Free Evolutions
Comparing Forward and Futures
Zero-Coupon Bond Prices and Rates
Martingales
Swaptions
Caplet Pricing
24.6 Summary
Floorlet Pricing
Valuing Various Interest Rate 24.7 Appendix
Derivatives Linking the Binomial HJM Model
Multiple Factors and the HJM Libor Model
24.4 The Multiperiod Model 24.8 Cases
24.9 Questions and Problems
THE ASSUMPTIONS 593
24.1 Introduction
The single-period Heath–Jarrow–Morton (HJM) model is useful as a teaching tool
to introduce the economic logic underlying the pricing and hedging methodology.
However, even as a teaching tool, the single-period model has its limitations. Many
interest rate derivatives need at least two periods to capture their complexity. For
example, in a single-period setting, an American call option trivializes to a European
call, a futures contract simplifies to a forward, and a swaption sheds its uncertainty
and becomes a deterministic security! Therefore, even for pedagogical reasons, one
needs to extend the HJM model to a multiperiod setting.
More important, the real limitation of a single-period model is that just as in the
binomial option pricing model of Chapter 18, the term structure evolution only
becomes realistic when the time periods become small and the number of periods
becomes large. To obtain a model useful for practice, one must extend the HJM
model to a multiperiod setting. The two-period model of this chapter starts us along
this path. The multiperiod extension just involves more complicated computations
as the number of periods increase, easily done on a computer, once the economic
logic of the two-period model is mastered (see Jarrow 2002b).
This chapter presents the multiperiod binomial HJM model. We start with the
term structure of interest rate evolution in a two-period setting. Next, the two-
period HJM valuation model is presented. The extension to multiple periods and
the approximation to the continuous time HJM libor model, the topic of Chapter
25, is discussed. Finally, the multiperiod model is applied to the pricing of caplets,
floorlets, forward rate agreements (FRAs), Eurodollar futures, and swaption contracts.
For ease of presentation, the material parallels the presentation used in the multiperiod
binomial model of Chapter 18. The next chapter presents a continuous time analogue
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
to the Black–Scholes–Merton model for pricing caplets and floorlets.
24.2 The Assumptions
We retain assumptions A1–A5, which have been with us since Chapter 17, repeated
here for convenience:
A1. No market frictions
A2. No credit risk
A3. Competitive and well-functioning markets
A4. No intermediate cash flows
A5. No arbitrage opportunities
We extend assumption A6 to two periods.
A6. Two-period binomial process. The zero-coupon bond prices B(t, T) for
t = 0, 1, 2 and T = 1, 2, 3 trade in discrete time, and their prices evolve according to
a binomial process. At each time in the tree, there are two states “up” and “down.”
First, if the zero-coupon bond matures in the next period, that is, T = t + 1, then
the current bond price goes to $1 regardless of the state. Second, standing at time t,
the current T > t + 1 maturity zero-coupon bond price either gets multiplied by
594 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
the factor u(t, T) in the up state and takes the price u(t, T) × B(t, T) with the actual
probability q(t) > 0 or it is multiplied by the factor d(t, T) in the down state and takes
the price d(t, T) × B(t, T) with probability 1− q(t). We assume that u(t, T) > d(t, T).
Note that the verbal description is almost identical to that in the single-period
model. The only difference is in the number of time steps considered. The two-
period evolution is given in Figure 24.1. It is the same as the single-period evolution
up to time 1 in both the up and down states. The difference occurs after time 1. For
this reason, when describing the evolution, we focus only on discussing what happens
after the first period.
EXAMPLE 24.1: The Two-Period Term Structure Evolution
Zero-Coupon Bond Prices
■ If we are in the up state at time 1, then zero-coupon bond prices can go up again, or they can go down.
The actual probability of going up is q(1)u = 0.2. The prices of the zeros in the up and down states
are identified by the subscripts uu and ud, respectively.
- Consider the longest-maturity zero-coupon bond. If the up state occurs, then this zero’s price is
B(2, 3)uu = 0.9555. If the down state occurs, then the price of this zero is B(2, 3)ud = 0.9504.
- Consider the two-period bond. If either state occurs, the bond pays off a dollar. Thus B(2, 2)uu =
B(2, 2)ud = 1.
■ If we are in the down state at time 1, the actual probability of going up is q(1)d = 0.6. The prices of the
zeros in the up and down states are identified by the subscripts du and dd, respectively.
- Consider the three-period zero. If the up state occurs, then this zero’s price is B(2, 3)du = 0.9440.
If the down state occurs, then the zero’s price is B(2, 3)dd = 0.9349.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- Consider the two-period bond. If either state occurs, the bond pays off a dollar. Thus B(2, 2)du =
B(2, 2)dd = 1.
■ Note that over the time period (1, 2), there is one less zero-coupon bond trading in the term structure
of zeros. This is because the one-period bond at time 0 matures at time 1 and disappears from the
market.
■ We note that the binomial tree, unlike that in the binomial model of chapter 17, does not recombine
to form a lattice. This is because this example is based on a discrete approximation to the HJM libor
model that does not recombine (see the appendix to this chapter).
The lower part of Figure 23.3 gives the formalization of this example using the
symbols introduced in assumption A6.
Zero-Coupon Bond Prices
■ If we are in the up state at time 1, and if up occurs again, then the three-period
bond’s price is u(1, 3)u B(1,3)u = B(2, 3)uu . If down occurs instead, then the three-
period bond’s price is d(1, 3)u B(1, 3)u = B(2, 3)ud . The two-period bond pays $1
regardless of the state at time 2.
THE ASSUMPTIONS 595
■ If we are in the down state at time 1, and if up occurs again, then the three- period
bond’s price is u(1, 3)d B(1, 3)d = B(2, 3)du . If down occurs instead, then the three-
period bond’s price is d(1, 3)d B(1, 3)d = B(2, 3)dd . The two-period bond pays $1
regardless of the state at time 2.
The binomial tree in its most general form does not recombine to form a lattice.
It could be designed to do so. Subcases of the HJM model where the tree recombines
to form a lattice include the Ho and Lee (1986) and Vasicek (1977) models.
The Spot Rate
■ If we are in the up state at time 1, then the spot rate is
R(1)u = [1 − B(1, 2)u ] /B(1, 2)u
(24.1a)
= (1 − 0.9620) /0.9620 = 0.0395 or 3.95 percent
■ If we are in the down state at time 1, then the spot rate is
R(1)d = [1 − B(1, 2)d ] /B(1, 2)d
(24.1b)
= (1 − 0.9555) /0.9555 = 0.0465 or 4.65 percent
The Money Market Account
■ The money market account (mma) earns the spot rate each period:
1 + R (1)u if up
A (2) = [1 + R (0)] ×
{ 1 + R (1)d if down
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1.0717 if up
= (24.2)
{ 1.0788 if down
At this point, it is easy to understand how to extend the two-period model to
three or more periods. Starting at time 0, the number of zero-coupon bonds trading
at each node should exceed the number of periods in the model by at least one.
Then, one just adds more time steps. At each time step, the shortest- maturity bond
matures, pays $1, and then disappears from the tree. Otherwise, the tree’s structure is
no different than the two-period evolution.
596 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
FIGURE 24.1: Two-Period HJM Evolution
Time 0 Time 1 Time 2
Zero-Coupon Bond Prices (Example)
B(2,3)uu = 0.9555
q(1)u = 0.2 B(2,2)uu = 1
B(1,3)u = 0.9167
B(1,2)u = 0.9620 R(1)u = 0.0395
B(1,1)u = 1
q(0) = 0.3
B(2,3)ud = 0.9504
B(2,2)ud = 1
B(0,3) = 0.88
B(0,2) = 0.93 R(0) = 0.0309
B(0,1) = 0.97 B(2,3)du = 0.9440
q(1)d = 0.6 B(2,2)du = 1
B(1,3)d = 0.8977
B(1,2)d = 0.9555 R(1)d = 0.0465
B(1,1)d = 1
B(2,3)dd = 0.9349
B(2,2)dd = 1
Zero-Coupon Bond Prices (Symbols)
Time 0 Time 1 Time 2
u(1,3)uB(1,3)u = B(2,3)uu
q(1)u B(2,2)uu = 1
u(0,3)B(0,3) = B(1,3)u
u(0,2)B(0,2) = B(1,2)u R(1)u
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
B(1,1)u = 1
q(0)
d(1,3)uB(1,3)u = B(2,3)ud
B(2,2)ud = 1
B(0,3)
B(0,2) R(0)
B(0,1) u(1,3)dB(1,3)d = B(2,3)du
q(1)d B(2,2)du = 1
d(0,3)B(0,3) = B(1,3)d
d(0,2)B(0,2) = B(1,2)d R(1)d
B(1,1)d = 1
d(1,3)dB(1,3)d = B(2,3)dd
B(2,2)dd = 1
Time 0 Time 1 Time 2
Money Market Account (mma)
Earns R(0) Earns R(1)
[1 + R(0)] × [1 + R(1)u] = A(2)u
= 1.0717
A(0) = 1 A(1) = [1 + R(0)]
= 1.0309 [1 + R(0)] × [1 + R(1)d] = A(2)d
= 1.0788
THE TWO-PERIOD MODEL 597
24.3 The Two-Period Model
The two-period binomial HJM model is analogous to the single-period model, once
one employs the backward induction argument used in the multiperiod binomial
option pricing model of chapter 18. The term structure evolution for the two-period
economy is given in Figure 24.1 and was previously described.
Let us introduce some simplifying notation. At time 2, we see there are four states
possible corresponding to the branches (up, up), (up, down), (down, up), and (down,
down). We write the prices at time 2 as B(2, T)su and B(2, T)sd , where s ∈ {u, d}
(which is a compact way of writing that s takes on either the value u or d) and
T = 2, 3. For example, if s = d and T = 3, then the first expression takes the value
B(2,3)du . This is the time 2 price of a three-period zero-coupon bond at the node
reached by going down at time 1 and then up at time 2. Figure 24.1 shows that this
zero has a price of 0.9440. Notice that the time paths are read from left to right: in
du, the letter d denotes that the first movement is down, and the next letter u denotes
that the second movement is up.
Arbitrage-Free Evolutions
Just as in the single-period model, we need to determine necessary and sufficient
conditions for the two-period binomial tree to be arbitrage-free.
NECESSARY AND SUFFICIENT CONDITIONS It is clear that if the tree
is arbitrage-free, then at each node and time, the single-period evolution at that
node must be arbitrage-free. Thus Result 23.1 must hold at every node in the tree—
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and it turns out that if Result 23.1 holds at every node in the tree, then the tree is
arbitrage-free. The proof of this last statement is identical to the sufficiency proof for
Result 23.1 in chapter 23 and is therefore omitted. Consequently, we have obtained
the following result.
RESULT 24.1
Equal Pseudo-Probability Condition
The two-period zero-coupon bond price evolution is arbitrage-free if and
only if there exists real numbers 𝜋(0), 𝜋(1)u , 𝜋(1)d strictly between 0 and 1
such that the pseudo-probabilities at any node must be identical, that is,
𝜋 (0) = 𝜋 (0, T) for T = 2, 3 (24.3a)
𝜋(1)s = 𝜋(1, 3)s for s ∈ {u, d} (24.3b)
598 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
where 𝜋(0, T) = [1 + R(0) – d(0, T)]/[u(0, T) – d(0, T)]; 𝜋(1, 3)s = [1 +
R(1)s − d(1, 3)s ]/[u(1, 3)s – d(1, 3)s ]; R(0) and R(1)s are the time 0 and time
1 spot rates, respectively; and u(0, T), d(0, T), u(1, 3)s , d(1, 3)s are the dollar
returns on the zero-coupon bond in the up and down states at times 0 and 1,
respectively.
In this characterization result, the no-arbitrage condition at time 0 is the same as
in the single-period model. Note that at time 2, the last period in the tree, there
is only one pseudo-probability. This is because there is only one risky zero-coupon
bond trading over the last time step. Hence, over the last time step, the no-arbitrage
condition is the same as in Chapter 17. The pseudo-probability just needs to be
strictly between zero and one.
When considering models with more than two time steps, both of these obser-
vations will always be true. First, for every time step except the last, there will be
multiple pseudo-probabilities, all of which must be equal and strictly between zero
and one for these nodes to be arbitrage-free. Second, over the last time step, there is
only one pseudo-probability that must be strictly between zero and one.
The next example shows that this arbitrage-free condition is satisfied by the two-
period binomial evolution in Figure 24.1.
EXAMPLE 24.2: An Arbitrage-Free Two-Period Binomial Tree
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
This example verifies that the term structure evolution in Figure 24.1 is arbitrage-free.
■ First, we have already checked the node at time 0 in Example 23.4 and verified that it is arbitrage-free.
Hence we only need to check the two nodes at time 1 in the up and down states. We need to compute
the up and down returns on the three-period zero-coupon bond for all possible states at time 1. These
are as follows:
u (1, 3)u = B (2, 3)uu /B (1, 3)u = 0.9555/0.9167 = 1.0423
d (1, 3)u = B (2, 3)ud /B (1, 3)u = 0.9504/0.9167 = 1.0367
u (1, 3) d = B (2, 3)du /B (1, 3)d = 0.9440/0.8977 = 1.0516
d (1, 3)d = B (2, 3)dd /B (1, 3)d = 0.9349/0.8977 = 1.0415
■ Then, using expression (24.3), we compute the pseudo-probabilities:
1 + R(1)u − d(1, 3)u 1.0396 − 1.0367
𝜋(1, 3)u = = = 0.5
( u(1, 3)u − d(1, 3)u ) ( 1.0423 − 1.0367 )
1 + R(1)d − d(1, 3)d 1.0465 − 1.0415
𝜋(1, 3)d = = = 0.5
( u(1, 3)d − d(1, 3)d ) ( 1.0516 − 1.0415 )
THE TWO-PERIOD MODEL 599
Since this is the last time step and each of these pseudo-probabilities is between zero and one, the tree
is arbitrage-free. We denote the value of these pseudo-probabilities as 𝜋(1)u and 𝜋(1)d in the up and
down states, respectively.
Zero-Coupon Bond Prices and Martingales
Assuming that each node in the tree is arbitrage-free, just as in the single-period
model, we can show that the zero-coupon bond prices must equal their expected
discounted $1 payoff, using the pseudo-probabilities. To do this, we use backward
induction.
Standing at time 1 in state s ∈ {u, d}, because the node is arbitrage-free, by
definition of 𝜋(1)s , algebra yields
𝜋(1)s B(2, T)su + [1 − 𝜋(1)s ] B(2, T)sd
B(1, T)s = for T = 2, 3 (24.4a)
1 + R(1)s
or, more abstractly,
E𝜋1 [B (2, T)]
B(1, T)s = for T = 2, 3 (24.4b)
1 + R(1)s
This is analogous to the single-period model, but viewed from time 1 in state s.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Next, stepping back to time 0, the logic used for the one-period model shows
that
𝜋 (0) B(1, T)u + [1 − 𝜋 (0)] B(1, T)d
B (0, T) = for T = 2, 3 (24.5a)
1 + R (0)
Substitution of expression (24.4a) into this expression and algebra yields
B(2, T)uu
B (0, T) = 𝜋 (0) 𝜋(1)u
[1 + R (0)] [1 + R(1)u ]
B(2, T)ud
+𝜋 (0) [1 − 𝜋(1)u ]
[1 + R (0)] [1 + R(1)u ]
(24.5b)
B(2, T)du
+ [1 − 𝜋 (0)] 𝜋(1)u
[1 + R (0)] [1 + R(1)d ]
B(2, T)dd
+ [1 − 𝜋 (0)] [1 − 𝜋(1)u ]
[1 + R (0)] [1 + R(1)d ]
or, more abstractly, we get our final result.
600 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
RESULT 24.2
Risk-Neutral Valuation for Zero-Coupon Bond Prices
The time 0 value of the zero-coupon bonds can be written as
1
B (0, 2) = E𝜋 (24.6a)
( [1 + R (0)] [1 + R (1)] )
and
B (2, 3)
B (0, 3) = E𝜋 (24.6b)
( [1 + R (0)] [1 + R (1)] )
where E𝜋 (⋅) denotes expectation using the pseudo-probabilities [𝜋(0),
𝜋(1, s)]; R(0) and R(1) are the time 0 and 1 spot rates, respectively; and B(2,3)
is the three-period zero-coupon bond’s price at time 2.
These two expressions show that the zero-coupon bond’s price is its expected
discounted $1 payoff using the pseudo-probabilities.
Result 24.2 proves that the local expectations hypothesis, introduced in
chapter 21 (see Extension 21.2), holds under the pseudo-probabilities and not the
actual probabilities. The local expectations hypothesis states that all zero-coupon
bonds earn the same expected return over any time period. Essentially, all zero-
coupon bonds are equally risky. We see that the local expectations hypothesis is true
under the pseudo-probabilities, and as seen later, because the pseudo-probabilities
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
differ from the actual probabilities via a risk adjustment, the local expectations
hypothesis cannot hold under the actual probabilities.
Last we note that the denominators in expressions (24.6a) and (24.6b) correspond
to the time 2 value of the mma A(2). As such, this immediately implies that the
price of a zero-coupon bond normalized by the price of the mma, B(t, T)/A(t), is a
martingale under the pseudo-probabilities, generalizing the same result proven in the
single-period model. The next example applies risk-neutral valuation to zero coupon
bonds.
EXAMPLE 24.3: Risk-Neutral Valuation for Zero-Coupon Bonds
Using the two-period evolution given in Figure 24.1, we can check to see if the risk-neutral valuation
formula holds for the zero-coupon bonds maturing at time 2. The formula is
B(2, 2)uu B(2, 2)ud
𝜋 (0) 𝜋 (1)u + 𝜋 (0) [1 − 𝜋 (1)u ]
[1 + R (0)] [1 + R(1)u ] [1 + R (0)] [1 + R(1)u ]
B(2, 2)du
+ [1 − 𝜋 (0)] 𝜋 (1)u
[1 + R (0)] [1 + R(1)d ]
THE TWO-PERIOD MODEL 601
B(2, 2)dd
+ [1 − 𝜋 (0)][1 − 𝜋 (1)u ]
[1 + R (0)] [1 + R(1)d ]
1 1
= 0.25 + 0.25
[1.0309] [1.0396] [1.0309] [1.0396]
1 1
+0.25 + 0.25
[1.0309] [1.0465] [1.0309] [1.0465]
= 0.97 = B (0, 2)
validating the result in the two-period example. As shown, the price of the three-period zero is equal
to its value in the four states at time 2 multiplied by the pseudo-probability of each state occurring and
discounted to time 0 using the spot rates of interest. This is expression (24.6b) written out in longhand.
Caplet Pricing
We are now ready to price a caplet in this two-period economy. Although they are not
shown on the two-period tree, the spot rate at time 2 in each state can be computed
as
1 − B(2, 3)us
R(2)us = (24.7a)
B(2, 3)us
and
1 − B(2, 3)ds
R(2)ds = for s ∈ {u, d} (24.7b)
B(2, 3)ds
Consider a caplet maturing at time 3 with strike rate k. The caplet’s time 3 payoff
is c(3) = max[0, R(2) −k]. This value is known at time 2, hence the time 2 present
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
value of this payoff is
R (2) − k
c (2) = max 0, (24.8)
[ 1 + R (2) ]
Using the convention introduced in the previous chapter, for valuation purposes, we
will consider the caplet’s maturity as time 2. These are the payoffs used in the tree to
compute the time 0 value of this caplet.
BACKWARD INDUCTION Let us recall the (fictional) super sleuth Sherlock
Holmes, who popularized the method of backward induction: start at the end, work
backward through time, eventually solving the problem at the beginning. To price
the caplet using this method of backward induction, we pretend we are standing at
time 1 and value the caplet at that time. Then, having the time 1 values, we move
back to time 0 and value the caplet at that date, which gives our answer.
METHOD 1: CONSTRUCTION OF THE SYNTHETIC CAPLET AT
TIME 1 Standing at time 1 at state s 𝜖 {u, d}, we want to construct the portfolio
of the three-period zero and mma account such that the payoffs to the portfolio
match the payoffs to the traded caplet. The method is identical to that used in the
602 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
one-period model, except the notation is now different. We choose n shares of the
three-period zero and m shares of the mma with cost
V(1)s = mB(1, 3)s + n A (1)
such that at time 2,
in the up state: V(2)su = mu(1, 3)s B(1, 3)s + nA(1)[1 + R(1)s ] = c(2)su (24.9a)
in the down state: V(2)sd = md(1, 3)s B(1, 3)s + nA(1)[1 + R(1)s ] = c(2)sd (24.9b)
Solving for m and n yields
c(2)su − c(2)sd
m= (24.10a)
[u(1, 3)s − d(1, 3)s ] B(1, 3)s
u(1, 3)s c(2)sd − d(1, 3)s c(2)su
n= (24.10b)
[u(1, 3)s − d(1, 3)s ] A (1) [1 + R(1)S ]
Substitution of m and n into the cost of construction and simplification gives the
arbitrage-free price of the caplet:
c (1)s = mB (1, 3)s + nA (1)
c(2)su − c(2)sd u(1, 3)s c(2)sd − d(1, 3)s c(2)su (24.11a)
= B (0, 2) +
[u(1, 3)s − d(1, 3)s ] B (0, 2) [u(1, 3)s − d(1, 3)s ] [1 + R(1)s ]
METHOD 2: RISK-NEUTRAL VALUATION AT TIME 1 Using the arbitrage-
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
free condition and the pseudo-probabilities, algebra yields
𝜋(1)s c (2)su + [1 − 𝜋(1)s ] c (2)sd
c (1)s = (24.11b)
1 + R(1)s
or, more abstractly,
E𝜋1 [c (2)]
c (1)s = (24.11c)
1 + R(1)s
We note again that this is the same method we used in the single-period model,
except the time and state notation has changed. This gives a present value formula.
The caplet’s time 1 value is its discounted expected payoff at time 2 using the pseudo-
probabilities.
PRICING THE CAPLET AT TIME 0 Having priced the caplet at time 1 in both
the up and down states, we now move back to time 0 to price the caplet. Note that
this is the same problem solved in the single-period model, except that we replace
the previous payoffs with expression (24.11c). The arbitrage-free price of the caplet
is therefore
𝜋 (0) c(1)u + [1 − 𝜋 (0)] c(1)d
c (0) = (24.12a)
1 + R (0)
THE TWO-PERIOD MODEL 603
Substitution of the caplet’s time 1 values from expression (24.11b) and algebra yields
c (2)uu
c (0) = 𝜋 (0) 𝜋(1)u
[1 + R (0)] [1 + R(1)u ]
c (2)ud
+ 𝜋 (0) [1 − 𝜋(1)u ]
[1 + R (0)] [1 + R(1)u ]
(24.12b)
c (2)du
+ [1 − 𝜋 (0)] 𝜋(1)u
[1 + R (0)] [1 + R(1)d ]
c (2)dd
+ [1 − 𝜋 (0)] [1 − 𝜋(1)u ]
[1 + R (0)] [1 + R(1)d ]
Using more abstract notation, we can alternatively write this as our next result.
RESULT 24.3
Risk-Neutral Valuation Formula for a Caplet
A caplet with strike rate k and maturity date 2 is given by
c (2)
c (0) = E𝜋 (24.13)
( [1 + R (0)] [1 + R (1)] )
where E𝜋 (⋅) denotes expectation using the pseudo-probabilities [𝜋(0), 𝜋(1,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
s)]; R(0) and R(1) are the time 0 and 1 spot rates, respectively; and c(2) is the
caplet’s value at time 2.
Expression (24.13) illustrates risk-neutral valuation for pricing a caplet, and it gives
a present value formula! Today’s price is the expected discounted value of the caplet’s
time 2 payoff, where the expectation uses the pseudo-probabilities. The pseudo-
probabilities adjust for risk, consequently discounting is with the riskless spot rate of
interest. For the purposes of computing the arbitrage-free price of a caplet, expression
(24.13) is easy to use, as illustrated in the next example.
EXAMPLE 24.4: Pricing a Caplet
This example uses the risk-neutral valuation method to price a caplet with maturity time 2, strike rate
k = 0.05, and notional of $1.
604 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
■ First, we need to compute the caplet’s payoffs at time 2 in all states. To do this, we must first compute
the spot rate of interest at time 2 in all states. Using Figure 24.1, we get
1 − B(2, 3)uu 1 − 0.9555
R(2)uu = = = 0.0465
B(2, 3)uu 0.9555
1 − B(2, 3)ud 1 − 0.9504
R(2)ud = = = 0.0522
B(2, 3)ud 0.9504
1 − B(2, 3)du 1 − 0.9440
R(2)du = = = 0.0593
B(2, 3)du 0.9440
1 − B(2, 3)dd 1 − 0.9349
R(2)dd = = = 0.0696
B(2, 3)dd 0.9349
■ The caplet’s payoffs at time 2 are
R(2)uu − k 0.0465 − 0.05
c (2)uu = max 0, = max [0, ]=0
[ 1 + R(2)uu ] 1.0465
R(2)ud − k 0.0522 − 0.05
c (2)ud = max 0, = max [0, ] = 0.0021
[ 1 + R(2)ud ] 1.0522
R(2)du − k 0.0593 − 0.05
c (2)du = max 0, = max [0, ] = 0.0088
[ 1 + R(2)du ] 1.0593
R(2)dd − k 0.0696 − 0.05
c (2)dd = max 0, = max [0, ] = 0.0183
[ 1 + R(2)dd ] 1.0696
These payoffs are illustrated in Figure 24.2.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Finally, using the risk-neutral valuation formula, the caplet’s time 0 value is
c (2)uu
c (0) = 𝜋 (0) 𝜋(1)u
[1 + R (0)] [1 + R(1)u ]
c (2)ud
+𝜋 (0) [1 − 𝜋(1)u ]
[1 + R (0)] [1 + R(1)u ]
c (2)du
+ [1 − 𝜋 (0)] 𝜋(1)u
[1 + R (0)] [1 + R(1)d ]
c (2)dd
+ [1 − 𝜋 (0)] [1 − 𝜋(1)u ]
[1 + R (0)] [1 + R(1)d ]
0 0.0021
= $0.25 + 0.25
[1.0309] [1.0396] [1.0309] [1.0396]
0.0088 0.0183
+0.25 + 0.25
[1.0309] [1.0465] [1.0309] [1.0465]
= $0.0068
THE TWO-PERIOD MODEL 605
Floorlet Pricing
The same method can be applied to price a floorlet. The arbitrage-free valuation
formula will be expression (24.13), with the payoff of the floorlet replacing that of
the caplet. Alternatively, given the price of the caplet, one can use caplet–floorlet
parity to value the caplet. The next result gives caplet–floorlet parity in the two-
period setting.
RESULT 24.4
Caplet–Floorlet Parity
[f (0, t − 1) − k] B (0, t) = c (0) − p (0) (24.14)
where f (0, t − 1) is the time 0 forward rate for date t − 1, B(0, t) is the time
0 zero-coupon bond price with maturity at time t, and both the caplet with
time 0 price c(0) and the floorlet with time 0 price p(0) have strike rate k and
maturity date t − 1 for t = 2, 3.
The proof is identical to that of the single-period model and is therefore omitted
(see the appendix to Chapter 23). The next example shows how to price a floorlet
using caplet–floorlet parity.
FIGURE 24.2: Caplet Evolution
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Time 0 Time 1 Time 2
0.0465 – 0.05
c(2)uu = max 0, =0
1.0465
c(1)u
0.0522 – 0.05
c(2)ud = max 0, = 0.0021
1.0522
c(0) = 0.0068
0.0593 – 0.05
c(2)du = max 0, = 0.0088
1.0593
c(1)d
0.0696 – 0.05
c(2)dd = max 0, = 0.0183
1.0696
606 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
EXAMPLE 24.5: Valuing a Floorlet
■ Consider a floorlet that has strike rate k = 0.05 and maturity time 2. Its time 3 payoff is
p (3) = max [k − R (2) , 0]
■ The floorlet is a put option on the spot rate of interest. Note that the floorlet’s payoff at time 3 is based
on the spot rate at time 2. Because this payoff is known at time 2, the present value of this payoff at
time 2 is obtained by discounting by the spot rate. For valuation purposes, this is the floorlet’s maturity
date.
max [k − R (2) , 0]
p (2) =
1 + R (2)
■ Let the term structure evolution be as in Figure 24.1. Consider a caplet with the same maturity and
the same strike price as the floorlet’s. We determined the time 0 price of this caplet as c(0) = $0.0068
in Example 24.5.
■ The forward rate at time 0 for date 2 is
B (0, 2) 0.93
f (0, 2) = −1= − 1 = 0.0568
B (0, 3) 0.88
Hence, using caplet–floorlet parity (see expression [24.14]), we have that
p (0) = c (0) − [f (0, 2) − k] B (0, 3)
= 0.0068 − (0.0568 − 0.05) 0.88 = 0.0008
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The floorlet has a value of $0.0008 at time 0 on a notional of $1.
Valuing Various Interest Rate Derivatives
The same approach used for pricing a caplet can be used to price any interest rate
derivative. The only modification is to replace the payoffs of the caplet with those
of the interest rate derivatives under consideration. For American-type interest rate
derivatives, just as in Chapter 18, in the backward induction argument, one needs to
make a decision regarding early exercise or not at each node of the tree. Otherwise,
the procedure is identical to that illustrated earlier.
Multiple Factors
The extension to multiple factors in the two-period model is the same as in the
single-period model. For example, a two-factor model has three branches at each
node in the tree. The pricing methodology extends easily, see Jarrow (2002b) in this
regard.
FORWARDS, FUTURES, AND SWAPTIONS 607
24.4 The Multiperiod Model
The extension of the binomial HJM model to more than two time periods is
computationally demanding but straightforward in terms of the ideas and concepts.
When building the tree, more zero-coupon bonds are needed in the term structure
as the time horizon in the model grows. There should be at least one more zero-
coupon bond in the term structure than there are periods in the model. Except for
the extended size of the bond price vectors at each node, the analysis is identical
to that already presented. As the time progresses in the tree, one zero-coupon bond
matures and leaves the term structure. When valuing interest rate derivatives, just as
in the two-period model, backward induction is used. All of the previous results hold
in a T-period model (see Jarrow [2002b] for more on this topic).
The multiperiod binomial HJM model can be used to approximate a continuous
time HJM model by making the time steps small in size. The link between the
multiperiod binomial HJM model and the HJM libor model discussed in the next
chapter is provided in the appendix to this chapter. The appendix shows how to
choose the up and down factors in the zero-coupon bond price evolution so that the
implied simple forward rates have a lognormal distribution in the limit.
24.5 Forwards, Futures, and Swaptions
Let us revisit FRAs, Eurodollar futures, and swaptions in the context of the two-
period binomial model, although the results hold more generally in a multiperiod
setting. As noted in the introduction to this chapter, the complexities of these
contracts are only evidenced in multiperiod models.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Forward Rate Agreements
We previously discussed FRAs in Chapter 21. Recall that in our discrete time model,
the FRA payoff at time T can be written as
LN × [R (T − 1) − fFRA (0, T − 1)] (24.15)
where LN is the notional value of the contract and fFRA (0, T− 1) is the time 0 FRA
rate for time T− 1. The payment on the FRA occurs one period after the uncertainty
in the contract’s payoff is determined.
Consider an FRA with maturity date 3 on a notional value of $1. Using the
previous expression, we have that the FRA’s time 3 payoff is
R (2) − fFRA (0, 2) (24.16)
A long position in the FRA receives at time 3 the difference between the spot rate
and the agreed-on FRA rate fFRA (0,2). At time 0, the market-clearing FRA rate is
determined such that the value of the FRA is zero. It was shown in Chapter 21 that
608 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
this market-clearing condition implies that the FRA rate is equal to the forward rate,
that is,
B (0, 2)
fFRA (0, 2) = f (0, 2) = −1 (24.17)
B (0, 3)
The forward rates at time 0 are easily computed from the zero-coupon bond prices
at time zero in the two-period evolution of Figure 24.1:
⎡f (0, 2) = 0.0568⎤
⎢f (0, 1) = 0.0430⎥
⎢ ⎥
⎣f (0, 0) = 0.0309⎦
Eurodollar Futures
We previously discussed Eurodollar futures contracts in Chapter 21 as well. Recall
that in our discrete time model, a futures contract with delivery date T on a notional
value of LN has a futures price equal to
F (t) = LN × [1 − ifut (t)] at time t (24.18a)
and
F (T) = LN × [1 − R (T)] at time T (24.18b)
where ifut (t) is the time t futures rate.
The cash flows over the life of the futures contract are determined by marking-
to-market. We have that the cash flows at times t = 1, 2, 3,.. are
[F (t) − F (t − 1)] = −LN × [ifut (t) − ifut (t − 1)] (24.18c)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Consider a Eurodollar futures contract with maturity 2 on a notional value of
$1. We can easily determine the arbitrage-free futures price in our two-period HJM
model.
ARBITRAGE-FREE FUTURES PRICES We use backward induction in con-
junction with the risk-neutral valuation formula to determine the arbitrage-free
futures prices.
Standing at time 1, using the risk-neutral valuation formula, the value of the futures
contract is
E𝜋1 [F (2) − F (1)]
V (1) =
1 + R (1)
where the 1 denotes taking the expectation at time 1. The time 1 futures price F( 1)
is set such that the contract has zero value. This implies that
E𝜋1 [F (2) − F (1)]
0=
1 + R (1)
Some simple algebra yields the result
F (1) = E𝜋1 [F (2)] (24.19)
FORWARDS, FUTURES, AND SWAPTIONS 609
The time 1 futures price is seen to be equal to the expected time 2 futures price,
where the expectation is computed using the pseudo-probabilities.
Last, moving back to time 0, again using the risk-neutral valuation formula, the
value of the futures contract is
E𝜋 [F (1) − F (0)]
V (0) =
1 + R (0)
As before, the time 0 futures price F(0) is set such that the contract has zero value.
This implies that
E𝜋 [F (1) − F (0)]
0=
1 + R (0)
or, equivalently,
F (0) = E𝜋 [F (1)] (24.20)
Similarly, the time 0 futures price is equal to the expected time 1 futures price under
the pseudo-probabilities. Expressions (24.19) and (24.20) prove the fact that futures
prices are a martingale under the pseudo-probabilities.
ARBITRAGE-FREE FUTURES RATES We can now determine the arbitrage-
free futures interest rates using expressions (24.18a) and (24.18b). Substitution of these
expressions into the arbitrage-free futures prices gives
1 − ifut (1) = E𝜋1 [1 − R (2)] and 1 − ifut (0) = E𝜋 [1 − ifut (1)]
(24.21)
or ifut (1) = E𝜋1 [R (2)] and ifut (0) = E𝜋 [ifut (1)]
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The futures rates are also seen to be martingales under the pseudo-probabilities.
One last computation gives us our final result. Using the notation of Figure 24.1,
we can write expression (24.21) as
ifut (1) = 𝜋(1)s R(2)su + [1 − 𝜋(1)s ] R(2)sd (24.22a)
for s ∈ {u, d}. Then, substitution of this expression into (24.21) gives
ifut (0) = 𝜋 (0) {𝜋 (1)u R (2)uu + [1 − 𝜋 (1)u ] R (2)ud }
(24.22b)
+ [1 − 𝜋 (0)] {𝜋 (1)d R (2)du + [1 − 𝜋 (1)d ] R (2)dd }
Writing this expression more abstractly yields the following result.
RESULT 24.5
Eurodollar Futures Rates
Consider a Eurodollar futures contract with maturity date 2 on a notional
value of $1. The market clearing futures rate is
ifut (0) = E𝜋 [R (2)] (24.23)
610 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
where E𝜋 [⋅] denotes expectation using the pseudo-probabilities [𝜋(0), 𝜋(1,s)]
and R(2) is the spot rate at time 2.
As evidenced by this result, the futures interest rate is the expected futures spot rate
using the pseudo-probabilities. In contrast to common belief, the futures rate is not
the expected future spot rate using the actual probabilities. Because the pseudo- and
actual probabilities differ by a risk adjustment, we see that Eurodollar futures rates
reflect both risk premia and the actual probabilities, generating a biased estimate of
the future spot rate of interest.
Comparing Forward and Futures Rates
We can now compare the forward and futures interest rates. Using the definition of
the spot rate R(2), we can rewrite expression (24.23) as
1
ifut (0) = E𝜋 −1 (24.24)
( B (2, 3) )
but
⎛ 1 ⎞ 1
E𝜋
1 ⎜ A (2) ⎟ ( A (2) ) B (0, 2)
E𝜋 = E𝜋 ⎜ ⎟> = = 1 + f (0, 2)
( B (2, 3) ) ⎜ B (2, 3) ⎟ B (2, 3) B (0, 3)
𝜋
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
⎝ A (2) ⎠ E ( (2) )
A
The first strict inequality is due to a well-known result in probability called Jensen’s
inequality (given a random variable x, E [1/x] > 1/E[x]). The second to last equality
uses the risk-neutral valuation formulas for the zero-coupon bonds. The last equality
is just the definition of the forward rate. Substitution of this insight into expression
(24.24) shows that
ifut (0) > f (0, 2) (24.25)
The futures rate is strictly greater than the forward rate. Both contracts are betting
on the future spot rate of interest R(2). However, interest rate futures contracts have
marking-to-market, whereas FRAs do not. This implies that futures contracts have
cash flow reinvestment risk that FRAs do not. Hence, to induce a speculator to
enter a futures contract at zero value, a rate higher than the FRA rate needs to be
included. This reinvestment risk difference between the two contracts is the cause
of the inequality in expression (24.25). As stated many times before, futures and
forward contracts are fraternal, not identical, twins! The next example illustrates the
determination of eurodollar futures rates.
FORWARDS, FUTURES, AND SWAPTIONS 611
EXAMPLE 24.6: Eurodollar Futures Rates
■ This example computes the time 0 futures rates for a Eurodollar futures contract maturing at time 2
using the term structure evolution in Figure 24.1. Example 24.4 already computed the spot rates at
time 2, which are: R(2)uu = 0.0465, R(2)ud = 0.0522, R(2)du = 0.0593, and R(2)dd = 0.0696. Then,
ifut (0) = E𝜋 [R (2)] = 0.25 (0.0465 + 0.0522 + 0.0593 + 0.0696)
= 0.0569
■ For comparison, the time 0 forward rate for time 2 is
B (0, 2) 0.93
fFRA (0, 2) = f (0, 2) = −1= − 1 = 0.0568
B (0, 3) 0.88
As the theory showed, the futures rate is greater than the corresponding FRA rate.
Swaptions
A swaption is an option (call or put) on the value of a swap at a future date. As an
option, there is a strike price K and maturity date T, and the underlying—the swap—
needs to be completely specified, receiving fixed and paying f loating (or conversely).
We illustrate the valuation of a swaption using the example in Figure 24.1.
EXAMPLE 24.7: Swaption Valuation
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
To value a swaption, one must first value the underlying swap.
■ Consider a swap receiving floating and paying fixed with a notional of $100 million and maturing at
time 2. Let the fixed rate on the swap be 0.04 per year. The floating rate is the spot rate R(t).
■ Using the valuation formula in Chapter 22 (Result 22.1), the swap’s value at time 1 can be written
abstractly as the difference between the value of the floating-rate loan less the value of the fixed-rate
loan:
VSwap (1) = PVFloat (1) − PVFixed (1)
= 100 − [100 × (0.04) × B (1, 2) + 100 × B (1, 2)]
Here a floating rate loan is always valued at par ($100 million), and the value of a fixed-rate loan is
the present value of its coupon and principal payment at time 2. Using the zero-coupon bond prices
from Figure 24.1, we obtain
VSwap (1)u = 100 [1 − (0.04) 0.9620 − 0.9620] = −0.0475
VSwap (1)d = 100 [1 − (0.04) 0.9555 − 0.9555] = 0.6248
612 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
■ Consider a swaption written on this swap. Formally, let’s consider a swaption that is a European call
option on this swap with a maturity date time 1 and strike price K = 0. The payoff on the swaption
at its maturity, time 1, is
Swaption (1) = max [VSwap (1) , 0]
If the swap’s value is greater than zero, the option is exercised and the long gets the underlying swap
with value VSwap (1).
■ Using the values of the swap computed earlier, the swaptions’ values at time 1 in the up and down
state are
Swaption(1)u = max [−0.0475, 0] = 0
Swaption(1)d = max [0.6248, 0] = 0.6248
■ Finally, using the risk-neutral valuation formula, the time 0 value of the swaption is
𝜋 (0) Swaption(1)u + [1 − 𝜋 (0)] Swaption(1)d
Swaption (0) =
1 + R (0)
0.5 × 0 + 0.5 × 0.6248
= = 0.3030
1.0309
From an economic perspective, a swaption is an insurance policy that enables the
long to enter a swap at a later date, at terms determined at time 0. Such a contract
would be of interest to a corporate treasurer who entered into a floating-rate loan at
time 0, and who may want to switch to a fixed-rate loan at a later date if interest rates
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
increase, but at the fixed rates determined at time 0. If exercised at a later date, his
floating-rate loan plus the swap changes the floating-rate loan into a fixed-rate loan,
with the time 0 fixed rates—and there is no cost to exercise because the strike price
is zero! This is a wonderful insurance product for the corporate treasurer, but as with
all insurance contracts, it has a cost—premium—to be paid. The previous example
demonstrated how to determine the arbitrage-free value of this premium.
24.6 Summary
1. The binomial HJM model assumes that the zero-coupon bond prices B(t, T) for
t = 0, 1, 2 and T = 1, 2, 3 trade in discrete time and that their prices evolve
according to a binomial process. First, if the zero-coupon bond matures the next
period, that is, T = t + 1, then the current bond price goes to $1 regardless
of the state. Second, standing at time t, the current T > t + 1 maturity zero-
coupon bond price either gets multiplied by an up factor u(t, T) and takes the price
u(t, T) × B(t, T) with the actual probability q(t) > 0 or it is multiplied by a down
factor d(t, T) and takes the price d(t, T) × B(t, T) with probability 1− q(t).
APPENDIX 613
2. The two-period zero-coupon bond price binomial tree is arbitrage-free if and
only if there exist pseudo-probabilities 𝜋(0), 𝜋(1)u , 𝜋(1)d strictly between 0 and 1
such that
𝜋 (0) = 𝜋 (0, T) for T = 2, 3
and
𝜋(1)s = 𝜋(1, 3)s for s ∈ {u, d}
where 𝜋(0, T) = [1 + R(0) – d(0, T) ] / [u(0, T) – d(0, T) ] and 𝜋(1, 3)s =[1 + R(1)s
– d(1, 3)s ] / [u(1, 3)s – d(1, 3)s ]
3. Consider a caplet maturing at time 2 with a strike rate of k. Its payoff at time 2 is
c(2) = max {0, [R(2) – K ] / 1 + R(2)]}. The arbitrage-free price of a caplet can be
determined using the risk-neutral valuation formula c(0) = E𝜋 (c(2) / ( [1 + R(0)]
[1 + R(1)])), where E𝜋 (⋅) denotes expectation using the pseudo-probabilities.
4. Caplet–floorlet parity is
[f (0, 1) − k] B (0, 2) = c (0) − p (0)
where f (0,1) is the time 0 forward rate for date 1, B(0,2) is the time 0 zero-coupon
bond price with maturity at time 2, and both the caplet with time 0 price c(0) and
the floorlet with time 0 price p(0) have strike rate k and maturity 1.
5. Consider a Eurodollar futures contract maturing at time 2. The arbitrage-free
futures rate at time 0 is ifut (0) = E𝜋 [R(2)]. The futures rate is always larger than
the corresponding FRA rate.
24.7 Appendix
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Linking the Binomial HJM Model and the HJM Libor
Model
The binomial interest rate model is easy to understand and to implement. But is it
really useful? The answer is yes. In fact, the binomial model can approximate the
popular HJM libor model, which is commonly used by financial institutions to price
caps, floors, and swaptions. This appendix shows how to set the parameters of the
binomial HJM model so that it approximates the HJM libor model of Chapter 25.
With interest rate models, the specification is more complex than that used in
the case of the binomial option pricing model in Chapter 18 to approximate the
Black–Scholes–Merton model. The logic of the procedures, however, is identical.
We provide the convergence parameters for a one-factor HJM model. These
parameters are selected such that the simple forward rate i(t, T) converges to a
lognormal distribution, which yields the libor model.
First, impose a discrete time step of length ∆t. Fix a partition of time such that
N∆t = 𝛿. Let T* be the maximum maturity zero-coupon bond under consideration
with t ≤ T ≤T*, where T* = t + M𝛿 = t + M ⋅ N ⋅ ∆t. A total of M ⋅ N time steps
of length ∆t are in the time horizon.
614 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
The forward rate and spot rate over the time interval [t, t + ∆t] are defined by
B (t, T) 1
f (t, T) = −1
( B (t, T + Δt − 1) ) Δt
and
1 1
R (t) = −1
( B (t, t + Δt) ) Δt
We would like to investigate the limit of the multinomial tree as ∆t → 0. We note
that the forward rate f (t, T) will converge to the continuously compounded forward
rate f (t, T) over the time period [t, t+ dt].
We are interested in the limit distribution of the simple forward rate, defined by
B (t, T) 1
i (t, T) = −1 for 0 ≤ t ≤ T ≤ T∗
( B (t, T + 𝛿) )𝛿
where [T, T+ 𝛿] corresponds to the time interval in the future over which the forward
rate applies; 𝛿 is a fixed time period, say, 0.25 years.
Choose at each time step t (see Jarrow 2002b, p. 286)
𝜋t = 1 > 2
T−Δt
𝛾(t, j)Δt3/2
[1 + R (t)] ∑
u (t, T) = e j = t+Δt
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
⎛ T−Δt ⎞
cosh ⎜⎜ 𝛾 (t, j) Δt3/2 ⎟⎟
∑
⎝j = t+Δt ⎠
T−Δt
𝛾(t, j)Δt3/2
[1 + R (t)] ∑
d (t, T) = e j = t+Δt
⎛ T−Δt ⎞
cosh ⎜⎜ 𝛾 (t, j) Δt3/2 ⎟⎟
∑
⎝j = t+Δt ⎠
where cosh(x) ≡ (e−x + ex )/2 and the volatility functions for the continuously
compounded forward rates satisfy
𝛿 ⋅ i (t, T) 𝜎 (t, T)
𝜕
( 1 + 𝛿 ⋅ i (t, T) )
𝛾 (t, T + 𝛿) = 𝛾 (t, T) + (1)
𝜕T
for the given inputs {B(t, T): T ≥ t} and {𝜎(t, T): T ≥ t}, where 𝜎(t, T) is a
deterministic function of time t (see Shreve 2004, p. 442).
CASES 615
The 𝜎(t, T) correspond to the instantaneous volatilities of the simple forward rates
i(t, T) in the libor model.
𝛿 ⋅ i (t, T) 𝜎 (t, T)
Define X (t, T) =
( 1 + 𝛿 ⋅ i (t, T) )
𝜕X (t, T)
Approximate on the partition by its backward-looking derivative, that is,
𝜕T
𝜕X (t, T) X (t, T − Δt) − X (t, T)
≈
𝜕T Δt
where X(t, t − ∆t) ≡ 0.
The solution to expression (1) is given by the following. Let T = t + j∆t + K𝛿,
where j = 0, … , N and K = 1, … , M− 1. Then, for fixed 𝛿, and for each j = 0, … ,
N− 1,
K
𝜕X (t, t + jΔt + k𝛿)
𝛾 (t, T) = 𝛾 (t, t + jΔt) + (2)
∑ 𝜕T
k =0
To specify a process, r( t, t + j∆t) for j = 0, . . . , N − 1 can be arbitrarily specified.
We suggest setting them equal to zero.
Given are B(t, T), 𝜎(t, T) for T = t + n∆ for n = 0, . . . , MN. From B(t, T),
one computes R(t) and i(t, T). From i(t, T), 𝜎(t, T), one computes 𝛾(t, T). Finally,
substitution gives u(t, T), d(t, T). Then, the distribution for simple forward rates
i(T, T)– i(t, T) converges to a lognormal distribution as ∆t → 0.
For implementation, to estimate the parameters, one often chooses
𝜎 (t, T) = 𝜎 (T − t)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
to be deterministic functions of time to maturity.
It is important to note that the binomial tree generated by this approximation is
not a lattice. The nonlattice structure results from the desire to obtain lognormally
distributed simple forward rates in the limit. To extend this approximation to a
multifactor model, see Jarrow (2002b).
24.8 Cases
Liability Management at General Motors (Harvard Business School Case
293123-PDF-ENG). The case studies issues related to managing the liability
structure of General Motors using interest rate derivatives.
Columbia River pulp Company Inc.—Interest rate hedging Strategy (Ivey
Publishing Case 9A90B037, European Case Clearing House Case). The case
examines issues related to hedging interest rate risk in a company’s floating rate
debt using interest rate derivatives like caps, collars, and swaps.
Spencer Hall (A) (European Case Clearing House Case 9A95B045). The case ana-
lyzes issues surrounding the hedging of interest rate risk with instruments like
interest rate swaps, caps, and collars.
616 CHAPTER 24: MULTIPERIOD BINOMIAL HJM MODEL
24.9 Questions and Problems
24.1. Consider Figure 24.1 where B(0,1) = $0.95 replaces $0.97. Prove that the
new tree is not arbitrage-free.
24.2. In the caplet example 24.4, suppose the strike rate k = 0.06. Value the caplet.
24.3. In the floorlet example 24.5, suppose the strike rate k = 0.06. Value the
floorlet.
24.4. Show that the caplet and floorlet values obtained in problems 24.2 and 24.3
satisfy caplet–floorlet parity.
24.5. Use Figure 24.1. Consider a European call with maturity time 2 and strike
price 0.95 on the three-period zero. What is the value of this call option?
24.6. Use Figure 24.1. Consider a European put with maturity time 2 and strike
price 0.95 on the three-period zero. What is the value of this call option?
24.7. Consider the Eurodollar futures contract in example 24.6.
a. Compute the time 1 futures rates in the up and down states s : [ifut (1)
=E𝜋 [R(2)]].
b. Compute the forward rate agreement rate at time 1 for maturity 2 in the up
and down states. Show that these two are the same because both contracts
mature in one period.
24.8. For the swap in example 24.7, let the strike price be K = $0.50. Compute
the value of the swap using risk-neutral valuation.
24.9. Why are swaptions a useful interest rate derivative? Give an example to explain
your answer.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
24.10. Suppose a borrower has a 10-year floating-rate loan but wants a fixed-rate
loan. Explain how the borrower can change the floating-rate loan into a fixed-
rate loan using caps and floors.
24.11. Suppose the term structures of Eurodollar futures rates are increasing for times
1, 2, 3, and 4. Does this mean that the spot rate of interest is expected to
increase over the next four time periods?
QUESTIONS AND PROBLEMS 617
The next questions are based on this figure.
FIGURE 24. A: Zero-coupon Bond Price Evolution
B(2, 3)uu = 0.962057
B(2,2)uu = 1
B(1, 3)u = 0.932788
B(1, 2)u = 0.973744
B(1, 1)u = 1
B(2, 3)ud = 0.953824
B(2, 2)ud = 1
B(0, 3) = 0.897710
B(0, 2) = 0.942596
B(0, 1) = 0.970874 B(2, 3)du = 0.951966
B(2, 2)du = 1
B(1, 3)d = 0.916495
B(1, 2)d = 0.968004
B(1, 1)d = 1
B(2, 3)dd = 0.941612
B(2, 2)dd = 1
24.12. Compute the spot rates of interest at every node in the tree.
24.13. Is the two-period evolution arbitrage-free? Prove your answer.
24.14. Consider a caplet with maturity date time 2 and strike rate k = 0.03. Compute
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the arbitrage-free value of this caplet.
24.15. Consider a floorlet with maturity date time 2 and strike rate k = 0.03. What
is the value of this floorlet?
24.16. Using caplet–floorlet parity, verify the answers in problems 24.14 and 24.15.
24.17. Consider a European call with maturity time 2 and strike price $0.95 on the
three-period zero. What is the value of this call option?
24.18. Consider a European put option with maturity date time 2 and strike price
$0.93 on the three-period zero-coupon bond. What is the value of this put
option?
24.19. Consider an American put option with maturity date time 2 and strike price
$0.96 on the three-period zero-coupon bond. What is the value of this put
option?
24.20. Compute the Eurodollar futures rates for a contract maturing at time 2 and
compare them to the forward rate agreement rate for a contract maturing at
time 3.
25
The Heath–Jarrow–Morton
Libor Model
25.1 Introduction Forward Rates and Martingales
25.2 Why Caplets? Risk-Neutral Valuation
Actual versus Pseudo-probabilities
25.3 A History of Caplet Pricing
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Black’s Model 25.8 Caps and Floors
The Heath–Jarrow–Morton Model Caps
The HJM Libor Model Floorlets
EXTENSION 25.1 Black’s Model Caplet–Floorlet Parity
for Pricing Commodity Options Floors
and Caplets
25.9 Using the HJM Libor Model
25.4 The Assumptions The Greeks
25.5 Pricing Caplets Delta Hedging
Caplet and Floorlet Pricing (CFP) Gamma Hedging
Data 25.10 American Options, Futures,
Caplet Pricing Model Swaptions, and Other
25.6 The Inputs Derivatives
Observable Inputs 25.11 Summary
The Average Forward Rate 25.12 Appendix
Volatility
25.13 Cases
25.7 Understanding the HJM
Libor Model 25.14 Questions and Problems
WHY CAPLETS? 619
25.1 Introduction
Once upon a time, an expert on interest rate derivatives and his wife went to a bank
to open a certificate of deposit that would earn a fixed rate of interest. The bank
gave them an option to change the interest rate once during the investment’s lifetime
and lock in the then-current rate. After explaining the various terms and conditions,
the bank officer asked, “Do you understand this option?” The professor chuckled to
himself and was tempted to reply, “Yes, and I even know how to value the option, do
you?” But he just said yes, under his wife’s glaring eyes. They finished the formalities
and went on their way. As illustrated, even standard banking products contain interest
rate derivatives requiring prudent handling!
This chapter continues our study of the Heath–Jarrow–Morton (HJM) model by
extending the discrete time model of Chapters 23 and 24 to a continuous-time model.
You may recall that we assumed no interest rate risk in the continuous-time Black–
Scholes–Merton (BSM) model. Now it is back with a vengeance! This chapter shows
how to price the most basic interest rate derivative, a caplet, in the continuous-time
HJM libor model. Of course, a caplet is just a European call option on an interest rate,
the same derivative studied in the BSM model. As such, this chapter is the analogue
of Chapters 19 and 20 on the BSM model.
The next section explains why pricing caplets is a significant milestone in our
study of derivatives. After, we provide a brief history of caplet pricing models. Next
are the assumptions underlying the HJM libor model, the caplet pricing formula,
and a discussion of the inputs. Continuing, we reintroduce martingales, riskneutral
valuation, and the actual and pseudo-probabilities. The presentation is couched in
terms of forward rates because of their desirable properties. The pricing of floorlets
follows using caplet–floorlet parity. Then, we discuss delta and gamma hedging. The
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
valuation of American options, futures contracts, swaptions, and other interest rate
derivatives completes the chapter.
25.2 Why Caplets?
Why does pricing a caplet in continuous time cap (pun intended!) our study of
derivatives?
■ Our mantra has been to focus on the key interest rate derivatives: forwards, futures,
calls, and puts. Once you understand them well, you can understand most interest
rate derivatives by applying the build and break approach (see Chapter 7).
■ We studied forwards in Chapter 21. We discussed futures, calls, and puts on interest
rates in Chapters 23 and 24 in a discrete time model. This Chapter repeats the
discussion of futures, calls, and puts on interest rates in a continuous time setting.
■ This takes us to caplets, which are call options on interest rates. Once one prices a
caplet, a floorlet follows with caplet–floorlet parity. Futures are studied after caplets
and floorlets.
620 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
So the time has come for the real stuff! First, let’s take a peek at the history of
caplet pricing models.
25.3 A History of Caplet Pricing
Black’s Model
The first widely used model for pricing caplets was based on Fischer Black’s (1976)
formula for pricing commodity options (see Extension 25.1 for a discussion of the
model, and Figure 25.1 for a diagram showing the evolution of these models).
Black’s model required the assumption of constant continuously compounded libor
spot interest rates. Obviously this is an invalid assumption in a world with random
interest rates. Nonetheless, the model was used in the beginning because little else
was available.
FIGURE 25.1: Evolution of Caplet Pricing
Models
Black–Scholes–
Merton model Well understood by traders
1973
Black–Scholes–Merton model
Black’s model modified to price forwards
1976
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and futures options
Black’s model Arbitrage within model (spot
modified to price rate constant and forward rate
caplets stochastic)
Heath–Jarrow–
Arbitrage-free term structure
Morton model
evolution
1992
HJM libor model Assumes lognormally
for pricing caplets distributed simple
1995, 1997 forward rates
A HISTORY OF CAPLET PRICING 621
The Heath–Jarrow–Morton Model
Early interest rate option pricing models (OPMs), like the early equity OPMs,
required the estimation of risk premium. The estimation of risk premium, which is
equivalent to estimation of expected returns on zero-coupon bonds, is an impossible
task. Just as the BSM model bypassed this thorny issue of risk premium estimation
in the context of equity options, the HJM model did the same for interest rate
derivatives. Consequently, it fostered the expansion of the swap and interest rate
derivatives markets.
The HJM Libor Model
Black’s model for pricing caplets admits arbitrage because it is impossible for both
continuously compounded spot rates to be constant and simple forward rates to be
random. This inconsistency was noted in the original HJM paper (Heath, Jarrow,
and Morton 1992). Following this demonstration, a search began to find a stochastic
evolution for interest rates where continuously compounded spot rates are random
yet forward rates are lognormally distributed. The lognormal distribution was needed
to maintain the essence of Black’s formula.
EXTENSION 25.1: Black’s Model for Pricing Commodity Options and Caplets
In 1976, Black published “The Pricing of Commodity Contracts” in the Journal of Financial Economics. This paper
extended the BSM model to price an option on commodity futures. Using notation similar to Result 19.1,
Black’s formula for a European commodity futures call option is
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
c = B (0, T) [FN (d1 ) − KN (d2 )] (1)
where today is time 0, B(0, T) = e−rT is today’s price of a zero-coupon bond that matures at time T, r is the
continuously compounded risk-free interest rate per year; T is the common maturity date for both the option
and forward (or futures) contracts; F is the forward (or futures) price determined today; N(.) is the cumulative
standard normal distribution function, d1 = [log(F /K) + (𝜎2 T / 2)] /𝜎√T and d2 = d1 − 𝜎√T; K is the strike
price; and 𝜎 is the volatility of the continuously compounded forward (or futures) price.
As noted in part II of the book, forward and futures prices are the same if interest rates are constant. Note that
formula (1) can be obtained by replacing the stock price S in the BSM model with FB(0,T).
Black’s model easily extends to price options on forward contracts where the option matures at date T but
the forward matures at some later date. The new call price formula is the same as in Equation 1, except that the
zero-coupon bond price and the forward price correspond to the later date.
Next, to apply this model to caplets, one considers the forward rate i(0,T) computed as simple interest (no
compounding). Under this forward rate, Black’s formula for pricing a caplet that matures at time T with strike
rate k is
c = B (0, T + 𝛿) [i (0, T) N (d1 ) − kN (d2 )] 𝛿 (2)
622 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
This is the formula in Equation 1 with i(0,T) replacing F, the volatility of the simple forward rate 𝜎B replacing
𝜎, and an additional term 𝛿 denoting the time interval over which the caplet’s interest payment applies. Note the
similarity of this formula to the HJM libor caplet pricing formula of Result 25.1.
Unfortunately, Black’s model does not satisfy our assumption A5 of no arbitrage. This happens because the
model makes two assumptions that cannot both hold at the same time: first, the spot rate of interest (the riskless
rate) is constant, and second, the forward rate is stochastic. These assumptions are mutually inconsistent. If the
spot rate is always constant, then the forward rate must be nonrandom as well. Assuming that they hold together
generates arbitrage opportunities within the model!
This inconsistency was the major motivation for the development of the HJM libor model. By giving an
arbitrage-free justification for Black’s formula (with modified inputs), the HJM libor model overcame these
difficulties and allowed the pricing of caplets and floorlets.
Klaus Sandmann, Dieter Sondermann, Kristian Miltersen (1995); Kristian Mil-
tersen, Klaus Sandmann, Dieter Sondermann (1997); and Alan Brace, Dariusz
Gatarek, Marek Musiela (1997) independently discovered the solution. They came
up with the observation that although continuously compounded forward rates cannot be
lognormally distributed in an HJM model, simple forward rates can. This special case of the
HJM model is known as the libor model. It also goes by the names libor market
model, forward libor model, or the market model.
25.4 The Assumptions
The first five assumptions underlying the HJM libor model are identical to those
in Chapter 19, and the sixth assumption is analogous to the lognormal assumption
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
underlying the BSM model. This is no surprise because of the libor model’s lineage
to the BSM model via Black’s (1976) model (see Figure 25.1).
Recall from the last two chapters that the caplet’s underlying assets are the term
structure of Eurodollar deposits (and swaps) or, equivalently, the term structure
of Eurodollar zero-coupon bonds. The relevant assumptions from Chapter 19 are
assumed to hold for these traded zero-coupon bonds. They are repeated here for
convenience (and emphasis).
A1. No market frictions. We assume that trading involves no market frictions
and that the assets are perfectly divisible. This helps us to create a benchmark
model to which one can add frictions, and it’s a reasonable approximation for
institutional traders. Furthermore, frictionless markets is even becoming a reasonable
first approximation for small traders because the accumulated advances in information
technology have drastically reduced trading costs.
A2. No credit risk. This is a reasonable assumption for exchange-traded
derivatives but may be less so for over-the-counter (OTC) derivatives. In reality, OTC
transactions often impose collateral provisions on the counterparties to lower such
credit risk. We assume, as a first approximation, that the OTC collateral provisions
remove credit risk from these OTC contracts. Interestingly, derivatives subject to
THE ASSUMPTIONS 623
credit risk are priced by extending this term structure modeling framework, which
is studied in Chapter 26.
A3. Competitive and well-functioning markets. In a competitive market,
traders act as price takers. In a well-functioning market, price bubbles are absent.
Despite this being a reasonable assumption, fines have been levied against Barclays
bank for manipulating libor index rates during the financial crisis of 2007.
A4. No intermediate cash flows. As zero-coupon bonds do not have cash flows
before they mature, this assumption is automatically satisfied.
A5. No arbitrage opportunities. This assumption is self-explanatory.
We discard old assumption A6, no interest rate uncertainty, and we now allow
interest rates to fluctuate randomly. In this regard, we need to impose an assumption
on the evolution of the term structure of interest rates.
To facilitate an understanding of this assumption, let us first revisit the concept of
a forward rate from earlier chapters. Recall that Equation 21.9 gives today’s forward
rate (for a discrete time interval) for the future period [T, T + 1] as
B (0, T)
f (0, T) = −1 (25.1)
B (0, T + 1)
where B(0, T) and B(0, T + 1) are today’s (time 0’s) prices for the zero-coupon bonds
maturing at times T and (T + 1), respectively.
This definition is reasonably general. It holds irrespective of whether interest rates
are stochastic or deterministic, but the limitation is that the time interval is discrete
and of unit length. We need to generalize this aspect of the definition. The question
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
is, how would you generalize this definition to compute a forward rate over any time
period [T, T + 𝛿], even for an infinitesimal time interval? Example 25.1 helps to
answer this question.
EXAMPLE 25.1: Forward Rate Computations
Forward Rates (for a Discrete Time Interval)
■ We compute forward rates and report them in Table 25.1.
- The first column reports time T: time 0 is today, time 1 is one year, and so on.
- The second column reports today’s prices of the zero-coupon bonds, B(0, T), which mature at time
T. Their numerical values are reported in the third column. For example, today’s price of a zero
that matures after one year is B(0, 1) = $0.97, after two years is B(0, 2) = $0.93, and so on (these
numbers are taken from Table 21.3). B(0, 0) = $1 by definition.
624 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
TABLE 25.1: Computing Different Types of Forward Rates (See Example
25.1)
Time T Price of a B(0, T) Length of Forward Rates Forward Rate Computations
Zero B(0, T) Values Interval
B(0, 0)
0 B(0, 0) 1 1 f (0, 0) ≡ R(0) − 1 = 0.0309
B(0, 1)
B(0, 1)
1 B(0, 1) 0.97 1 f (0, 1) − 1 = 0.0430
B(0, 2)
i 1 B(0, 1)
1.0027 B(0, 1.0027) 0.9699 1/365 = 0.0027 (0, 1) − 1 × 365 = 0.0421
365 [ B(0, 1.0027) ]
B(0, 1)
1.5 B(0, 1.5) 0.95 0.5 i0.5 (0, 1) − 1 × 2 = 0.0421
[ B(0, 1.5) ]
B(0, 2)
2 B(0, 2) 0.93 1 f (0, 2) − 1 = 0.0568
B(0, 3)
B(0, 3)
3 B(0, 3) 0.88 1 f (0, 3) − 1 = 0.0732
B(0, 4)
■ We compute forward rates using definition Equation 25.1.
- The forward rate for the interval [0, 1] is f (0, 0) = B(0, 0)/B(0, 1) − 1 = 0.0309 and is given in the
fourth and fifth cells of the second row. Relation (21.12) tells us that this is also the spot rate, R(0).
Several of these computations are also reported in Table 21.3.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
- The forward rate over the interval [1, 2] is
B (0, 1)
f (0, 1) = − 1 = 0.0430 (25.2a)
B (0, 2)
■ We have two dates in the forward rate’s notation: today (time 0), when the rate is quoted, and the
date (time 1) when the rate becomes effective. Because we know the ending date (time 2), we skip
writing it as a third argument. Although many of our examples consider the length of the time
period to be one year, it need not be so in a continuous-time setting.
Simple Forward Rates (in Continuous Time)
■ Suppose you want to compute a forward rate as in Equation 25.2a over a fixed but not a unit time
interval such as six months. For this, we use the concept of simple interest, and we consider a simple
forward rate. To compute this rate, we need a zero maturing at time 1.5: assume it has a price of B(0,
1.5) = $0.95. Considering the time interval [1, 1.5], we write the simple forward rate as per year.
B (0, 1) 1 0.97
i0.5 (0, 1) = −1 = − 1) 2 = 0.04210526 (25.2b)
( B (0, 1.5) ) 0.5 ( 0.95
THE ASSUMPTIONS 625
Note that in the definition of a simple interest rate, the time interval over which the interest rate
applies needs to be indicated. The first time argument in the simple interest rate notation, t = 0,
corresponds to the current date, while the second time argument corresponds to the start date of the
future time interval T = 1. The subscript 0.5 following the symbol “i” indicates the length of the
time interval over which the interest rate applies: half of a year. Finally, note that multiplication by 2
converts the six-month interest rate into an annual rate.
■ Why go back to the simple interest discussed in Chapter 2? First, the Eurodollar market is a market in
which the derivatives (such as swaps, caplets, and floorlets) use simple interest rates to compute cash
flows, and second, it will enable us to price caplets using a modified Black’s formula.
Continuously Compounded Forward Rates (in Continuous Time)
■ Suppose instead that you want to compute the simple forward rate over an infinitesimal time inter-
val one year into the future, [1, 1 + Δt], as Δt → 0. Let us first approximate this by setting Δt = 1/365
= 0.002740 in Equation 25.2b which is one day measured in years. This requires a zero-coupon bond
that matures after one year plus one day with a price of B(0, 1.002740) = $0.969888 today. Then, the
simple forward rate over the small time interval Δt = 1/365 is
B (0, 1)
i 1 (0, 1) = − 1 × 365 = 0.000115 × 365 = 0.04214920 (25.2c)
( B (0, 1.002740) )
365
per year. Notice that multiplication by 365 converts the daily interest rate into an annual rate.
■ Using this rate, if you invest a dollar starting in one year for one day, then you get $1.000115 back in
one year and one day.
■ Taking smaller and smaller time intervals enables us to define the continuously compounded forward rate
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
via Equation 25.2c as
B (0, 1) 1
f ̃ (0, 1) = lim iΔt (0, 1) = lim −1 (25.2d)
Δt→ 0 Δt→ 0 ( B (0, 1
+ Δt) ) Δt
Here, we introduce a different notation for the continuously compounded forward rate, as opposed to
a simple interest rate over an infinitesimal time interval. This new notation avoids using a time interval
subscript because the length of the time interval to which this forward rate applies is always fixed at an
instant.
The example contains two important ideas that we rewrite for emphasis. First,
we define the simple forward interest rate. Replacing 1 with T and 0.5 with 𝛿 in
Equation 25.2b gives us the simple forward rate over [T, T + 𝛿] (i.e., the forward
rate expressed as a simple interest rate). We define this at time 0 for date T as
B (0, T) 1
i (0, T) = −1 × (25.3a)
( B (0, T + 𝛿) ) 𝛿
Note that we delete the subscript because for the remainder of this chapter, the
time interval over which the simple forward rate applies is fixed and always equals to
𝛿. To reduce clutter, we started the clock at time 0, although our analysis holds even
if h l k i t
626 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
Second, we define the continuously compounded forward rate by replacing 1
with T and (1+ Δt) with (T + Δt) in Equation 25.2d, where Δt is the time
interval over which the interest rate applies. Taking limits as Δt goes to zero gives
the continuously compounded forward rate:
B (0, T ) 1
f̃ (0, T ) = lim −1 (25.3b)
Δt→0 ( B (0, T + Δt) ) Δt
Now, we add our assumption on the stochastic evolution of simple forward rates.
A6. The simple forward rate i (T, T ) over [T, T+ 𝛿] follows a lognormal
distri-bution. This assumption is that the natural logarithm of the simple forward
rate follows a normal distribution. In symbols,
1
∫0T 𝜇s ds − 𝜈2T (T) + 𝜈T √T⋅z
i (T, T) = i (0, T) e 2 with probability
1 2 (25.4a)
− z
q (z) = e 2 /√2𝜋
where time 0 is today, time T is the maturity date, 𝜇T is the mean percentage change
per year in the simple forward rate,
∫0T 𝜎2s ds
𝜈2T = (𝜈 is a Greek letter, pronounced “nu”) (25.4b)
T
𝜎s is the time s – maturity simple forward rate [i(0, s)]’s volatility per year, q(z) are
the actual probabilities, and z is a standard normal random variable with zero mean
and unit variance. Note that the relevant volatility in this evolution is 𝜈T , which
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
is the “average” of the simple s – maturity forward rates’ instantaneous volatilities
over the option’s life. The mean is assumed to be a deterministic function of time
to maturity T. The s – maturity forward rate’s volatility 𝜎s is assumed to be a
constant.
This assumption is similar to the BSM model assumption A8 for the stock price’s
evolution presented in relationship (19.2). Here, however, the simple forward rate
replaces the stock price as the variable with a lognormal distribution.
Other assumptions for the evolution of the simple forward rate will generate
different term structure evolutions and different caplet pricing models. However,
these alternative assumptions are usually imposed on the evolution of the continuously
compounded forward rates f̃ (0, T) and not the simple forward interest rates as in the
HJM libor model. Some of these alternative models go by the names of the assumed
evolution: the Gaussian model, the extended Vasicek model, and the affine model
(see Jarrow 2010). These alternative models are left for more advanced courses.
We complete our list of assumptions by retaining assumption A7 from Chapter 19
with respect to the trading of the Eurodollar zero-coupon bonds.
PRICING CAPLETS 627
A7. Trading takes place continuously in time. This assumption is self-
explantory.
These assumptions set the stage for introducing the HJM libor model.
25.5 Pricing Caplets
Caplets are European call options on the simple forward rate. For many of our
examples in this chapter, we use the following data.
Caplet and Floorlet Pricing (CFP) Data
■ The caplets and floorlets start trading today (time 0) and expire after one year
(time T).
■ The notional principal LN is $100 million.
■ The strike rate (which may be the cap rate or the floor rate) k is 4 percent per year.
■ The payoffs are based on the six-month libor rate index, the simple forward rate
i(0,T). Assuming that this six-month period has 182 days, 𝛿 = 182/365 = 0.4986
years is the investment period over which simple interest is computed.
■ The average forward rate volatility 𝜈T is 0.15, or 15 percent per year.
■ If the caplet (or the floorlet) is in-the-money, the payment is made 0.4986 years
after the expiration date (at time [T + 𝛿]).
■ The zero-coupon bond prices and simple forward rates are taken from the previous
example (see Table 25.1).
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Caplet Pricing Model
Before introducing the pricing model, let us illustrate the workings of a caplet in
continuous time (see Example 25.2).
EXAMPLE 25.2: An Interest Rate Caplet’s Payoffs
■ International Treats and Restaurants Inc.’s (INTRest, a fictitious name) treasurer buys a caplet today
(time 0) with maturity T = 1 year, a strike rate of k = 4 percent, and notional of LN = $100 million
(see Figure 25.2 for a timeline).
628 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
FIGURE 25.2: Timeline for a Caplet (In Continuous Time)
Start date Intermediate dates Maturity date Payment date
Time 0 Time T Time (T + δ)
(Today) (1 year) (1.4986 years)
Time interval, δ = 0.4986 year
Buyer and writer Long can exercise libor i(T, T) determined for the period δ. If i(T, T ) > k, long’s
negotiate caplet’s an American payoff is
terms: the notional caplet cT+δ = LN × [i(T, T ) – k] × δ
LN, the strike rate k,
If i(T, T ) ≤ k, caplet
the maturity date T
expires out-of-the- money
Long pays premium Traders can close Value of Long’s payoff on this date, Zero-sum game—
and becomes owner out their positions cT = B(T, T + δ ) × LN × max[i(T, T ) – k] × δ, Long’s gain is Short’s loss
of the caplet by reverse trades where the zero-coupon bond
pays $1 at time (T + δ )
■ Suppose the six-month libor rate index realized after one year is 5 percent. This is the simple forward
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
rate, i(T, T) = i(1, 1) = 0.05.
■ Use the caplet and floorlet pricing (CFP) data:
- The caplet’s payoff at time (T + 𝛿) = 1.4986 years is
c1,4986 = max (0.05 − 0.04, 0) × (182/365) × 100 million
(25.5a)
= $498, 630.14
The payment is received 𝛿 years (a fraction of a year) after the caplet’s maturity at time T. It equals
the yearly simple interest rate difference [i(T, T) − k] times the period 𝛿 over which the interest is
earned times the notational.
- The present value of the payment on the maturity date is obtained by multiplying it by the relevant
zero-coupon bond’s price. Suppose B(1, 1.4986) = 0.9850 is the price of a zero-coupon bond at
time 1 (at time T) which matures after six more months (at time [T + 𝛿]). Then, the time 1 value
of the caplet’s payoff is
c1 = 0.9850 × max (0.05 − 0.04, 0) × (182/365) × 100 million
(25.5b)
= $491, 150.68
PRICING CAPLETS 629
The caplet’s payoff from this example can be easily formalized. Consider a caplet
with notional principal LN , strike rate k, maturity date T, and payment date (T + 𝛿)
and where the time period over which the simple interest rate is computed is 𝛿 years.
The caplet’s payoff on the payment date is
cT+𝛿 = max [i (T, T) − k] × 𝛿 × LN (25.6a)
where the simple forward rate i(T, T) is the interest rate at time T that applies over
the time period [T, T + 𝛿].
To obtain the value of this payoff at time T, discount this cash flow by multiplying
it by the relevant zero-coupon bond’s price. The caplet’s value on the maturity
date is
cT = B (T, T + 𝛿) max [i (T, T) − k] × 𝛿 × LN (25.6b)
where B(T, T + 𝛿) is the time T price of a zero-coupon bond that pays a dollar at
time (T + 𝛿).
For computation, one can think of time T as both the maturity date and the payoff
date of the caplet, but in this case, one needs to use Equation 25.6b as the caplet’s
payoff.
Under the HJM libor model, it can be shown that the caplet’s time 0 value is given
by the following formula (see Result 25.1).
RESULT 25.1
HJM Libor Model Formula for Caplet Pricing
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
A caplet’s price is
c = B (0, T + d) [i (0, T) N (d1 ) − kN (d2 )] × 𝛿 × LN (25.7a)
where B(0, T + 𝛿) is today’s price of a zero-coupon bond that matures at
time (T + 𝛿), T is the caplet’s maturity date, 𝛿 is the time period for interest
computation, i(0, T) is the simple forward rate per year that applies to the
interval [T, T + 𝛿], N( . ) is the cumulative standard normal distribution
function (with mean 0 and standard deviation 1),
i (0, T ) 𝜈2 × T
log + T
[ ( k ) 2 ]
d1 = and d2 = d1 − 𝜈T √T (25.7b)
𝜈T √T
k is the strike rate (which is the cap rate), 𝜈T is the average forward rate
volatility over the caplet’s life, and LN is the notional principal.
630 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
We discuss the proof of this result later in this chapter when discussing risk-neutral
valuation.
The caplet formula is similar in appearance to the standard BSM formula but
applied to the simple forward rate at time T, i(0, T), and not the stock price (which
is irrelevant in the world of interest rate derivatives). The key difference from the
standard BSM formula is that the stock price is replaced by the bond price B(0, T +
𝛿) multiplied by the simple interest rate i(0, T). This occurs because B(0, T + 𝛿) ×
i(0, T) corresponds to the present value of i(T, T) dollars received at time (T + 𝛿).
In addition to the simple forward rate i(0, T), the caplet’s value depends on the
zero-coupon bond price B(0, T + 𝛿), the strike rate k, the maturity date T, and the
average forward rate volatility 𝜈T . As with the BSM formula, the mean percentage
change in the simple forward rate does not appear in this equation. Had it appeared,
the thorny problem of risk premium computation would have remained with us. This
“important omission” makes the formula usable. Analogous to the BSM model, this
fact implies that if two traders agree on the inputs [i(0, T), B(0, T + 𝛿), k, T, 𝜈T ] but
disagree on the mean percentage change in simple forward rates, they will still agree
on the caplet’s price.
Example 25.3 demonstrates Result 25.1.
EXAMPLE 25.3: Pricing a Caplet
■ Consider the CFP data.
- The caplet is traded today (time 0) and matures at T = 1 year.
- The notional principal LN = $100 million and the strike rate k = 4 percent per year.
- The contract applies to the six-month libor rate index (which is the simple forward rate). The
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
interest is computed over 182 days, or 𝛿 = 182/365 = 0.4986 year.
- The simple forward rate is i(0, T) = i(0, 1) = 0.0421 (see the cell in the fifth row and fifth column
in Table 25.1).
- The average forward rate volatility 𝜈T is 0.15, or 15 percent per year.
- A zero-coupon bond maturing after 1.4986 years is worth B(0,T + 𝛿) = B(0,1.4986) = $0.95 today.
■ Compute d1 and d2 :
i (0, T) 𝜈2 × T 0.0421 0.152 × 1
log + T log ( +
( k ) 2 0.04 ) 2
d1 = = = 0.4170 and
𝜈T √T 0.15√1
d2 = d1 − 𝜈T √T = 0.4170 − 0.15 = 0.2670
■ Using Excel, we get cumulative standard normal distribution function values:
N (d1 ) = N (0.4170) = NORMSDIST (0.4170) = 0.6616 and
N (d2 ) = N (0.2670) = NORMSDIST (0.2670) = 0.6052
THE INPUTS 631
■ The caplet’s price using the HJM libor model formula (25.7a) for a $100 million notional is
c = B (0, T + 𝛿) [i (0, T) N (d1 ) − kN (d2 )] × 𝛿 × LN
= B (0, 1.4986) [i (0, 1) N (0.4170) − 0.04N (0.2670)] × 0.4986 × 100 million
= 0.95 × (0.0421 × 0.6616 − 0.04 × 0.6052) × 0.4986 × 100 million
= $172, 842.45
25.6 The Inputs
Collecting inputs for the HJM libor caplet pricing model is analogous to gathering
the inputs for the BSM model. Most of the inputs are easily obtained, except for the
average forward rate volatility.
Observable Inputs
The easiest inputs to obtain are the strike rate k, expiration date T, and time period for
interest computation 𝛿. These are written into the contract. What about the simple
forward rate? Had the underlying been a Treasury security, one could observe zero-
coupon bond prices from the market—but we are using Eurodollars and libor rate
indexes. This difference only presents a minor complication.
As the forward rate is equivalent to the FRA rate (see Result 21.3), one can obtain
the simple forward rate i(0, T) by observing the rates from FRA contracts trading in
the OTC markets. The quoted FRA rates are based on the simple interest rates as
defined in Equation 25.3a.1
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
A Eurodollar zero-coupon bond’s price can be created in several ways from these
FRA rates. For example, you can use the FRA rates in the following formula to
compute the zero-coupon bond prices:
1
B (0, T) = (25.8)
[1 + i (0, 0) × 𝛿] [1 + i (0, 𝛿) × 𝛿] ⋯ [1 + i (0, T − 𝛿) × 𝛿]
where i(0, 0) is the spot rate that applies over [0, 𝛿], i(0, 𝛿) is the simple forward
rate computed over the next time period of length 𝛿, and so on. This is a known
relationship (Equation 21.1a of Result 21.2) that has been slightly modified to account
for the time intervals not being of unit length, but equal to 𝛿 time units.
1
We point out that when we defined the forward rate in earlier chapters, for simplicity, we always fixed
the time interval over which the rates applied to be equal to one unit of time. Here, to correspond to
practice, it is necessary to make the time interval differ from one time unit. It is set equal to 𝛿 time units.
632 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
Alternatively, you can obtain these zero-coupon bond prices directly from
Eurodollar rates (for maturities up to one year) and swap rates (for maturities beyond
one year). The procedure for obtaining these zero-coupon bond prices from swap
rates was given in Extension 22.3.
The Average Forward Rate Volatility
Like the thorny problem of volatility estimation encountered in the context of the
BSM model, the average forward rate volatility is also the hardest input to estimate
in the HJM libor model. As before, there are two standard methods for volatility
estimation: an explicit method using past simple forward rates, yielding a historical
volatility, and an implicit method—calibration—using caplet market prices, yielding
an implied volatility.
FIGURE 25.3: Integration and the
Average Forward Rate
Volatility
y
y = g(x)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
0
a b x
s-maturity simple forward
rate variance σs2
0 0.25 … 1.25 1.5 Maturity s
THE INPUTS 633
Note that the equation for the volatility differs from that used for equity options.
Here it equals the forward rate’s average volatility over the caplet’s life, that is,
∫0T 𝜎2s ds
𝜈T = √ (25.9a)
T
where 𝜎s is the time s – maturity simple forward rate i(0, s)’s volatility per year.
To compute this equation, we need to estimate each individual s – maturity
forward rate’s volatility 𝜎s . The trick here is that we need to use constant maturity
forward rates and compute their volatility just as we did the stock’s return volatility
in Chapter 19. Except for these two differences, the methods for estimating the
simple forward rate’s volatility per unit time are analogous to those for equity options
discussed in Chapters 19 and 20.
The relevant equation to compute the s – maturity simple forward rate i(t, s)’s
historical volatility is
i (t + Δt, s + Δt) 1
var log = 𝜎s (25.9b)
√ [ ( i (t, s) )] Δt
This is the standard deviation of the percentage change2 in the s – maturity
simple forward rate over [t, t + Δt]. Multiplication by (1/Δt) annualizes the volatility
estimate.
HISTORICAL VOLATILITY To compute the average forward rate volatility as
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
given in expression (25.9a), we need to use some basic ideas from integral calculus.
Consider a continuous function y = g(x) as depicted in Figure 25.3, where x is plotted
along the horizontal axis and y is plotted along the vertical axis. Calculus tells us that
b
the integral ∫ g(x)dx measures the area under this curve between x = a and x = b.
a
Recall that one can approximate this area using the set of rectangles lying below the
curve (see Figure 25.2). Of course, the area of each rectangle equals its base times its
height. The limit of the area covered by these rectangles as the bases go to zero yields
the integral of g(x) with respect to x from a to b. We use this simple approximation to
compute the average forward rate volatility. The procedure is as follows: (1) collect
time series observations of the constant s – maturity forward rates for the entire term
structure of maturities s ranging from 0 to time T, (2) compute the variance of the
percentage changes in these simple forward rates using Equation 25.9b, (3) add up
the variances to get the approximation to the numerator of relationship (25.9a), and
finally, (4) divide by the length of the entire time interval T to compute the average.
Finally, taking the square root gives the average forward rate volatility.
2
The ratio log[i(t + Δt, s + Δt)/i(t, s)] gives the percentage change in the s — maturity simple forward
rate over the time interval. Note that these forward rates correspond to the time intervals [t, t + 𝛿] and
[t + Δt, t + Δt + 𝛿], respectively, where 𝛿 is fixed. In our example, 𝛿 corresponds to six months.
634 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
As the historical volatility is estimated using actual forward rate realizations that
nature “draws,” this estimate is obtained under the actual probability distribution.
Using our earlier terminology, we refer to this as “estimation under the probability
q.” The next example illustrates this method.
EXAMPLE 25.4: Computing the Historical Average Forward Rate Volatility
■ Both the maturity of the caplet, time T, and the simple forward rate’s investment interval, 𝛿, are given
by the contract’s specifications. We choose T = 1.25 years and 𝛿 = 0.5 years, representing a typical
caplet’s parameters.
■ First, one must decide on the frequency (daily, weekly, or monthly) of the data. Then, one must
collect the FRA quotes from dealers in the OTC markets for all maturity FRAs from time 0 to
time T. Weekly observation intervals are a good choice, balancing microstructure noise versus
providing a sufficient number of observations.
■ Table 25.2 reports hypothetical constant maturity forward rates for maturities ranging from time 0
to T = 1.25 years, collected each week for twenty consecutive weeks. Each forward rate applies to a
six-month investment period, that is, 𝛿 = 182/365 = 0.4986 year.
- The first column is labeled Week n, and it keeps track of the dates when the data points were
collected.
- The next six columns give the constant s – maturity forward rates. These simple forward rates always
correspond to a constant s – years in the future. Each rate is for a six-month investment period.
For example, on week 1, the forward rate for maturity time 0 (the spot rate) is 0.1000. The spot
rate corresponds to the investment period [0,0.5]. Next, the forward rate for maturity time 0.25 is
0.1050, corresponding to the rate over the investment period [0.25, 0.75], and so forth.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ There are several ways to compute the sample variance:
- There’s the tedious way of hand computing. Although inefficient, it builds mental muscles. When
computing by hand, one needs to compute both the mean and variance for each s – maturity
forward rate time series:
N
1
Mean ∶ 𝜇̂ s = i (n, s) (25.10a)
N∑ n =1
N
1 2
Variance ∶ 𝜎̂ 2s = [i(n, s) − 𝜇̂ s ] (25.10b)
N−1∑
n =1
- Alternatively, one can use a programmable calculator, or a spreadsheet, or a statistical package on a
computer. We input the data in Table 25.2 in Microsoft Excel and compute the variance in one step.
For example, running Excel command VARA over the twenty observations in cells B2 = 0.1000
to B21 = 0.0920 gives VARA(B2:B21) = 0.00004779. We denote this by 𝜎̂ 0s and record this below
the three-month spot rate series. Similarly, we compute and record in the table 𝜎̂ 20.25 = 0.00004984,
𝜎̂ 20.5 = 0.00005241, and so on.
■ Once we have these sample variances, it’s easy to compute the average forward rate volatility.
THE INPUTS 635
TABLE 25.2: Computing the Average Forward Rate Volatility Using Weekly
Price Data
Constant Maturity Simple Forward Rate i(n, s)
Week n s=0 s = 0.25 s = 0.5 s = 0.75 s = 1.00 s = 1.25
1 0.1000 0.1050 0.1100 0.1150 0.1200 0.1150
2 0.1020 0.1060 0.1120 0.1160 0.1230 0.1160
3 0.1050 0.1070 0.1140 0.1160 0.1260 0.1190
4 0.1080 0.1090 0.1150 0.1180 0.1270 0.1230
5 0.1090 0.1110 0.1150 0.1210 0.1280 0.1250
6 0.1100 0.1130 0.1180 0.1210 0.1290 0.1280
7 0.1100 0.1150 0.1190 0.1250 0.1300 0.1310
8 0.1110 0.1150 0.1200 0.1240 0.1340 0.1310
9 0.1120 0.1160 0.1210 0.1270 0.1340 0.1350
10 0.1120 0.1170 0.1210 0.1280 0.1350 0.1360
11 0.1110 0.1180 0.1210 0.1270 0.1360 0.1310
12 0.1090 0.1170 0.1200 0.1260 0.1350 0.1280
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
13 0.1070 0.1130 0.1180 0.1230 0.1290 0.1210
14 0.1030 0.1090 0.1100 0.1200 0.1230 0.1240
15 0.0990 0.1050 0.1100 0.1160 0.1210 0.1210
16 0.0980 0.1030 0.1080 0.1180 0.1200 0.1190
17 0.0960 0.1010 0.1060 0.1090 0.1160 0.1160
18 0.0940 0.0990 0.1010 0.1050 0.1140 0.1110
19 0.0920 0.0970 0.1000 0.1030 0.1110 0.1090
20 0.0920 0.0950 0.0990 0.1010 0.1090 0.1080
Variance 0.00004779 0.00004984 0.00005241 0.00006542 0.00006747 0.00006940
Variance × 0.25 0.00001195 0.00001246 0.00001310 0.00001635 0.00001687 0.00001735
Summed variance × 0.25/ T 0.00007046
Volatility × √52 0.060530
Tis 1.25 years; 𝛿 is 0.5 years.
636 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
- Multiply each of these variance estimates by 0.25 (the base of the rectangle) and record them in
Table 25.2. Summing gives the area under the curve, that is,
T
∫ 𝜎2s ds ≈ 𝜎̂ 2s × 0.25 = (𝜎̂ 20 + 𝜎̂ 20.25 + … + 𝜎̂ 21.25 ) × 0.25 = 0.00008808 (25.10c)
∑
0 s
- Dividing (25.10c) by the length of the time interval T gives
∑s 𝜎̂ 2s × 0.25
𝜈2T = = 0.00007046 (25.10d)
T
The positive square root is the average forward rate volatility 𝜈T = 0.008394, or 0.8394 percent per
week. To get the volatility per year, multiply by the square root of 52, to get 0.060530 or 6.0530 percent.
This represents the average volatility (per year) of the constant maturity forward rates over the caplet’s
life.
In computing the historical average forward rate volatility, a number of issues
still remain. First, how many observations should you use? Theoretically speaking,
more observations will give a better estimate, but long time series data may not be
“stationary.” A good rule of thumb is to use at least 50 observations so that the
central limit theorem applies. Second, how long should the time interval be between
observation dates? If observations are collected at short intervals, then market frictions
and market microstructure issues can significantly impact the estimation. Daily or
weekly intervals have fewer of these problems. Finally, what should be the time
interval between the maturities of the forward rates? The smaller the time interval,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the better that the sum of the rectangles approximates the area under the curve. In
practice, quarterly time intervals between the different maturity forward rates, as in
Table 25.2, yield a reasonable approximation.
IMPLIED VOLATILITY Just as for the BSM formula, one can use market prices of
caps, as portfolios of caplets, to find that average forward rate volatility 𝜈T that best
matches market prices. This is calibration. The procedure yields an implied volatility.
In practice, just as in the BSM model, implied volatilities for caplets differ from
historical volatilities, and when graphed across strikes and maturities, they exhibit a
smile or sneer (see Jarrow et al. 2007; Rebonato 2002). This implies, as in the case of
the BSM formula, that the HJM libor model is formally rejected by the data.
Nonetheless, the model can still be used with implicit volatilities but, in this case,
only as a statistical model for pricing caplets. When used as a statistical model, its
purpose must be restricted to pricing caplets (and floorlets) only, and its use as a
hedging tool is no longer appropriate. This is because the theoretical model has been
rejected (see Chapter 20 for a detailed discussion of these issues). The comments and
criticisms previously mentioned with respect to the BSM formula apply to the HJM
libor model as well.
As with options in other markets, many dealers express their quotes in terms of
volatilities, which they refer to as “vols.” The other inputs being known, with the
HJM libor model you can determine the caplet price.
UNDERSTANDING THE HJM LIBOR MODEL 637
25.7 Understanding the HJM Libor Model
Recall martingales, pseudo-probabilities, and risk-neutral valuation from equity
option pricing. We introduced them in part III of the book to provide an enriched
understanding of the option pricing methodology. We do the same here. In addition,
these concepts also enable us to provide a quick sketch of the proof for the caplet
pricing formula in Result 25.1.
Forward Rates and Martingales
As discussed in the context of the binomial interest rate option pricing model of
Chapter 23, it can be shown (see the appendix to this chapter)3 that the no-arbitrage
assumption implies the existence of pseudo-probabilities 𝜋T+𝛿 (z) > 0 such that
i (0, T) = E𝜋T+𝛿 [i (T, T)] (25.11)
where i(0,T) is today’s (time 0) simple forward rate for the time interval [T,T + 𝛿], T
is the caplet’s expiration date, 𝛿 is the time period over which the interest is applied,
and E𝜋T+𝛿 [.] is expectation based on these pseudo-probabilities 𝜋T+𝛿 (z).
Here the pseudo-probabilities depend on the maturity of the forward rate.
Equation 25.11 states that the current forward rate equals its expected time T value,
when using the pseudo-probabilities. Interestingly, this shows that the unbiased
expectations hypothesis (which was introduced in Extension 21.2) holds under the
pseudo-probabilities. Because the pseudo-probabilities contain an adjustment for risk
(see below), this also shows that the unbiased expectations hypothesis will not hold
under the actual probabilities.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Risk-Neutral Valuation
To see how a caplet is priced by risk-neutral valuation, let us first note that as shown
in the appendix to this chapter the caplet’s value can be written as
c (0) = B (0, T + 𝛿) E𝜋T+𝛿 {max [i (T, T) − k, 0] ⋅ 𝛿} × LN (25.12)
where E𝜋T+𝛿 {.} is expectation based on the pseudo-probabilities 𝜋T+𝛿 (z). In this
equation, the caplet’s payment date is time (T + 𝛿).
This expression states that the caplet’s time 0 value is equal to its discounted
expected payoff, using the pseudo-probabilities. As such, it is analogous to the risk-
neutral valuation procedure used with respect to the BSM model.
Equation 25.12 provides us with a convenient method for proving Equation 25.7a
of Result 25.1. As noted earlier, under the pseudo-probabilities, the forward rate
i(T,T) has a lognormal distribution with mean zero and variance 𝜈2T . A direct
3
It can be shown that the lognormally distributed simple forward rates imply a stochastic process for the
corresponding continuously compounded forward rates. Condition (25.11) can then be used to show
that the HJM arbitrage-free drift restriction holds for the continuously compounded forward rate process,
proving that the implied term structure evolution is arbitrage-free.
638 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
evaluation of the expectation in Equation 25.12 using this lognormal distribution
gives the HJM libor formula. The proof is identical to that given for the BSM model,
with the exception that only the inputs (and their interpretations) have changed.
Actual versus Pseudo-probabilities
The difference between the actual and pseudo-probabilities represents an adjustment
for risk, just as for the BSM model. To see this, we start with assumption A5. Under
assumption A5, the simple forward rate i(0,T) at time 0 has a lognormal distribution.
We can write this as
T
i (T, T) 1
log = ∫ 𝜇s ds − 𝜈2T T + 𝜈T √T ⋅ z (25.13)
( i (0, T) ) 0
2
where z is a standard normal random variable with mean 0, variance 1, with
1 2
− z
probability q(z) = e 2 /√2𝜋. The Equation 25.13 holds at any time t between
time 0 and the expiration date T.
These q(z)s are the actual probabilities. As with the BSM model, it can be shown
(see the appendix to this chapter) that the actual probabilities and the pseudo-
probabilities must agree on zero probability events, that is,
𝜋T+𝛿 (z) > 0 if and only if q (z) > 0 (25.14)
Otherwise, traders will not agree on what constitutes an arbitrage opportunity.
Second, this implies that the actual and pseudo-probabilities are related by:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
𝜋T+𝛿 (z) = 𝜑T+𝛿 (z) q (z) (25.15a)
where
𝜃2T + 𝛿
− + 𝜃 T +𝛿 z
𝜑T+𝛿 (z) = e 2 and (25.15b)
∫0T 𝜇s ds
𝜃T+𝛿 = − (25.15c)
( 𝜈 √T )
T
The quantity 𝜑T + 𝛿 (z) is an adjustment for risk. The justification for this statement
is the same as that given for the BSM formula in Chapter 19. This risk adjustment in
the pseudo-probabilities is the reason that the risk-neutral valuation procedure works
in our world with risk-averse traders!
25.8 Caps and Floors
The HJM libor model for pricing caplets is all we need! One can use it to price
a cap (which is a collection of caplets), a floorlet (via caplet–floorlet parity), and a
floor (which is a collection of floorlets). Let’s now see how to price these interest rate
derivatives.
CAPS AND FLOORS 639
Caps
A cap is a financial security with a notional amount LN , a strike rate k, an expiration
date T, and a “payment” interval 𝛿. The cap buyer receives the payment
max [i (t − 𝛿, t − 𝛿) − k, 0] × 𝛿 × LN (25.16)
at the payment dates t = 𝛿, 2𝛿, ..., (T + 𝛿).
For example, the first payment happens at time 𝛿, which is computed using the
simple forward rate i(0, 0) that applies over the time interval [0, 𝛿]. Recall that this sim-
ple forward rate effective immediately is also the spot rate. The second payment hap-
pens at time 2𝛿 and is based on the forward rate i(𝛿, 𝛿) known at time 𝛿, and so forth.
Hence, a cap is a portfolio of caplets, with each caplet having the same strike rate
k as the cap, and each with a different maturity date corresponding to the payment
dates. We can find this cap’s value by valuing each individual caplet and adding the
resulting values.
Floorlets
A floorlet is a European put option on the spot rate of interest. As with a caplet, the
payment is received 𝛿 years (a fraction of a year) after the floorlet’s stated maturity at
time T. The payment equals the notational times the simple interest rate [k − i(T,
T)]𝛿 adjusted for the time period 𝛿 over which the interest applies.
To dress this up in more formal terms, consider a floorlet with notional principal
LN , strike rate k, and expiration date T, where the time period over which interest
is computed is 𝛿 years and the payment date is (T + 𝛿). The floorlet holder’s payoff
on the payment date is
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
pT+𝛿 = max [k − i (T, T) , 0] × 𝛿 × LN (25.17)
where the simple forward rate i(T, T) is the interest rate determined at time T that
applies over the time period [T, T + 𝛿].
To obtain the value of this payoff on the floorlet’s maturity date, multiply the
preceding payoff by B(T, T + 𝛿), which is the time T price of a zero-coupon bond
that pays a dollar at time (T + 𝛿). Once you have priced a caplet, it’s easy to value a
floorlet by using our old friend caplet–floorlet parity.
Caplet–Floorlet Parity
Just as the price of otherwise identical European calls and puts on the same under-
lying are related by put–call parity, so are the prices of caplets and floorlets. The
following argument is a generalization of that given to derive caplet floorlet parity in
the HJM binomial model. We start with a mathematical identity:
i (T, T) − k = max [i (T, T) − k, 0] − max [k − i (T, T) , 0] (25.18a)
You can easily check that this identity is true by inputting values for the simple forward
rate i(T, T) that are greater than or less than the common strike rate k.
Notice that the payoffs on the right side are those at time (T + 𝛿) for the caplet and
floorlet, respectively. The left side represents the payments to an FRA with maturity
640 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
time (T + 𝛿). Next, we apply the risk-neutral valuation procedure: take expectations
E𝜋T+𝛿 {⋅} using the pseudo-probabilities 𝜋T + 𝛿 (z) and then discount the payoffs by
multiplying by the zero-coupon bond’s price that matures at time (T + 𝛿). This gives
B (0, T + 𝛿) E𝜋T+𝛿 [i (T, T) − k] × 𝛿 × L
= B (0, T + 𝛿) E𝜋T+𝛿 {max [i (T, T) − k, 0]} × 𝛿 × LN (25.18b)
− B (0, T + 𝛿) E𝜋T+𝛿 {max [k − i (T, T) , 0]} × 𝛿 × LN
where we have also multiplied the equations on both sides by 𝛿 × LN , the time period
over which interest is computed times the notional.
Because the expectation of i(T, T) is today’s simple forward rate, the expected
value of the constant k is itself and the discounted expected value of the caplet’s and
floorlet’s payoffs are today’s prices, we get the final result (see Result 25.2).
RESULT 25.2
Caplet–Floorlet Parity
c − p = B (0, T + 𝛿) [i (0, T) − k] × 𝛿 × LN (25.19)
where c and p are today’s (time 0) prices of caplets and floorlets that have
strike rate k and expiration date T, respectively, B(0, T + 𝛿) is today’s price of
a zero-coupon bond that pays $1 at time (T + 𝛿), i(0, T) is the simple forward
rate that applies to the period [T, T + 𝛿], and LN is the notional principal.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Example 25.5 uses caplet–floorlet parity to price a floorlet.
EXAMPLE 25.5: Pricing a Floorlet Using the HJM Libor Model
■ Consider the CFP data.
- Today is time 0, the floorlet’s expiration date T is one year, interest applies over six months, a 𝛿 =
182/365 = 0.4986 year period, and the payoff date is (T + 𝛿).
- The notional principal LN = $100 million and the strike rate k = 4 percent per year.
- The simple forward rate applicable over 𝛿 is i(0,T) = i(0,1) = 0.0421.
- The average forward rate volatility 𝜈T is 0.15, or 15 percent per year.
- A zero-coupon bond maturing after 1.4986 years is worth B(0, T + 𝛿) = B(0, 1.4986) = $0.95
today.
■ Example 25.3 gave the caplet’s price as
c = B(0, T + 𝛿) [i (0, T) N(d1 ) − kN(d2 )] × 𝛿 × LN
= $172, 842.45
USING THE HJM LIBOR MODEL 641
■ Using the caplet–floorlet parity relation (25.19), the floorlet’s price is
p = c − B (0, T + 𝛿) [i (0, T) − k] × 𝛿 × LN
= 172, 842.45 − 0.95 × (0.0421 − 0.04) × 0.4986 × 100 million
= $73, 116.42
Floors
A floor is a financial security on a notional amount LN with a strike rate k, a maturity
date T, and a “payment” interval 𝛿. A floor holder receives payments
max [k − i (t − 𝛿, t − 𝛿) , 0] × 𝛿 × LN (25.20)
at times 𝛿, 2𝛿, . . . , T, (T + 𝛿). Thus a floor is a portfolio of floorlets, each with the
same strike rate k as the floor but with different maturity dates corresponding to the
payment dates. To value a floor, price the constituent floorlets and just add them up!
25.9 Using the HJM Libor Model
Using a derivative pricing model really means “hedging.” After all, a perfect hedge
is the basis for derivatives pricing, and the holy grail of option pricing is the
determination of the option’s delta. Let’s see how we can hedge in the HJM libor
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
model.
Stepping back for one moment, hedging interest rate risk in an HJM model is very
flexible because every traded security can be used to hedge anything else. However,
because of the special structure of the HJM libor model, this section confines itself to
a discussion of delta and gamma hedging for caplets and floorlets. There are several
reasons for this.
1. Caplets and floorlets have closed-form solutions in the HJM libor model. This
enables one to easily compute partial derivatives yielding exact formulas for delta
and gamma hedges. Unfortunately, closed form solutions are not available for other
interest rate derivatives.
2. Our discussion of the dangers in using vega hedging in a model where the vola-
tility is nonrandom and not priced within the model applies to the HJM libor
model. This is confounded by the fact that the HJM libor model is rejected by
historical data, exactly analogous to the BSM model (see Chapter 20).
3. The delta and gamma hedging techniques discussed here do not extend to more
complex interest rate derivatives like American options and futures contracts,
which do not have closed-form solutions. They must be hedged using numerical
methods in a more complex HJM model, examples of which were discussed in
642 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
chapter 24 in the context of a two-period binomial model. This is because the
continuously compounded forward rates in an HJM libor model, which are needed
to price interest rate derivatives that are not just simple extensions of caplets and
floorlets, do not follow a simple probability distribution.
The Greeks
Hedging the Greeks is a risk management technique that uses the HJM libor model
and a Taylor series expansion to hedge the interest rate risk of a cap or floor posi-
tion. The approach is similar to the one discussed in Chapter 20 in the context of the
BSM model. Before introducing Greek hedging, let us first compute the deltas and
gammas needed in the hedges.
■ Taking the partial derivative of the caplet pricing formula in Result 25.1 with
respect to i(0,T) ≡ i, it can be shown that the caplet’s delta is
𝜕c 𝛿
DeltaC ≡ = − c + B (0, T + 𝛿) N (d1 ) 𝛿LN (25.21a)
𝜕i [1 + i (0, T) 𝛿]
This makes sense because all else being equal, an increase in the simple forward
rate raises a caplet’s value.
■ We can rewrite the definition of the simple forward rate given in Equation 25.3a
as
B (0, T)
B (0, T + 𝛿) = (25.21b)
1+i×𝛿
Taking the partial derivative of this equation with respect to i(0,T) gives
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
𝜕B (0, T + 𝛿) B (0, T + 𝛿) 𝛿
=− <0 (25.21c)
𝜕i [1 + i (0, T) 𝛿]
This equation also makes sense because an increase in the interest rate lowers a
zero-coupon bond’s value.
■ Taking the second partial derivative of Equation 25.21a, using Equation 25.21c,
and simplifying yields the caplet’s gamma:
𝜕2 c B (0, T + 𝛿) ′ 2𝛿2
GammaC ≡ = N (d1 ) 𝛿L N + c
𝜕i2 i (0, T) vT √T [1 + i(0, T)𝛿]2
2B (0, T + 𝛿)
− N (d1 ) 𝛿2 LN
[1 + i (0, T) 𝛿]
(25.21d)
where N ′ (∙) is the standard normal density function.
USING THE HJM LIBOR MODEL 643
Delta Hedging
The idea of a hedged portfolio is simple. Consider a portfolio consisting of a single
caplet. Instead of selling the caplet, we want to take an offsetting position in another
interest rate sensitive security, to remove all the interest rate risk from our long
caplet position. For this purpose, we can use some zero-coupon bonds. Hence we
want to buy or sell some zero-coupon bonds to delta hedge the caplet so that the
resulting portfolio becomes delta-neutral, and its value remains the same despite small
fluctuations in the simple forward rate.
If we can trade quickly enough so that the time interval between trades Δt
approximates an infinitesimal time interval dt, then the theory tells us that the
delta hedge removes all interest rate risk from a caplet position. The result is a
perfectly hedged portfolio that earns the risk-free rate under our assumptions (see the
appendix of Chapter 20 for a detailed discussion of this approach). Unfortunately, it
is impossible to trade continuously as prescribed by the theory. Consequently, we
must necessarily consider delta hedging over a finite but small time interval [0, Δt].
To delta hedge a long caplet position, consider the resultant portfolio V consisting
of the long caplet worth c and n shares of the zero-coupon bond worth B(0,T + 𝛿) ≡
B. We implement this hedge at time 0, but you can do this at any time t. The initial
value of the portfolio is
V (0) = c (0) + nB (0, T + 𝛿) (25.22a)
In symbols, we can write the evolution of our portfolio’s value over the time interval
[0, Δt] as
ΔV = Δc + nΔB (25.22b)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where ΔV ≡ V (Δt) − V (0) and ΔB ≡ B(Δt) − B(0).
Writing out the first-order terms in a Taylor series expansion in terms of changes
in the simple forward rate Δi, we see that the portfolio is delta-neutral if we can find
n such that
𝜕c 𝜕B
ΔV = Δi + n Δi = 0 (25.22c)
𝜕i 𝜕i
Note that the partial derivatives with respect to the model’s other parameters are
absent because they are fixed and do not change.
The preceding equation gives
𝜕c
𝜕i
n=− (25.22d)
𝜕B (0, T + 𝛿)
𝜕i
Using Equation 25.21a in the numerator and Equation 25.21c in the denominator
yields
c
n=− + N (d1 ) [1 + i (0, T) 𝛿] LN (25.22e)
B (0, T + 𝛿)
where we used Equation 25.21b to simplify further.
644 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
This gives the holy grail for the HJM libor model! The hedge ratio n represents
the number of shares of the zero-coupon bond to hold for each caplet to construct
a riskless portfolio (equivalent to a position in a money market account). Isn’t it
puzzling to see that a long position in a caplet is hedged by a long position in a zero-
coupon bond? No. The reason is, of course, that bond prices and simple forward
rates move in opposite directions, hence, bond and caplet prices move in opposite
directions too (see Result 25.3). It all makes sense!
RESULT 25.3
Delta Hedging in the Libor Model
To remove the interest risk from a long caplet position, buy
c
n=− + N (d1 ) [1 + i (0, T) 𝛿] LN (25.23)
B (0, T + 𝛿)
shares of a zero−coupon bond worth B (0, T + 𝛿)
where d1 = (log[i (0, T) /k] + 𝜈2T T/2)/𝜈T √T, N ( . ) is the cumulative standard
normal distribution function, today is time 0, i(0, T) is the simple forward
rate, T is the expiration date, 𝛿 corresponds to the period over which interest
applies, LN is the notional principal, B(0, T + 𝛿) is today’s price of a zero-
coupon bond, k is the strike rate, and 𝜈T is the average forward rate volatility.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Example 25.6 illustrates the effectiveness of delta hedging a caplet.
EXAMPLE 25.6: Delta Hedging a Caplet
■ Consider the CFP data.
■ Today is time 0, T = 1 year, i(0, T) = i(0, 1) = 0.0421, 𝛿 = 0.4986 year, B(0, T + 𝛿) = B(0, 1.4986) =
$0.95, k = 0.04, and 𝜐T = 0.15. Today’s caplet price of $172, 842.45 is based on a notional principal
LN = $100 million (see Example 25.3).
■ As just discussed, a delta-neutral portfolio is obtained by hedging one long caplet position with a
purchase of n = N(d1 )[1 + i(0, T)𝛿] shares of the zero-coupon bond. This creates a delta-neutral
portfolio with initial value V = c + nB(0, T + 𝛿).
■ We compute the values of ΔV = Δc + nΔB(0, T + 𝛿) to see how the delta hedge works for different
simple forward rate changes. The results are reported in Table 25.3. For simplicity, we assume that the
simple forward rate can magically change without changing time.
■ The first column reports the different simple forward rates i(0,1) starting from 0.0421, given in the
middle row. The rates differ from the starting value by 100, 10, or 1 basis points.
- Column 2 lists the bond B(0, T + 𝛿) prices corresponding to the different rates obtained from
Equation 25.21b.
- Column 3 reports the caplet values corresponding to the different rates.
USING THE HJM LIBOR MODEL 645
- To see how the hedge performs, we first compute the number of shares of the zero-coupon bonds
needed to hedge one caplet. This is given by Equation 25.22e. Using N(d1 ) with our basic data, we
get
c
n=− + N (d1 ) [1 + i (0, T) 𝛿] LN
B (0, T + 𝛿)
= 0.6616 × (1 + 0.0421 × 0.4986) × 100 million (25.24a)
= 67, 553, 564.09
- The final column reports the changes in the hedged portfolio’s value for the different simple forward
rates:
ΔV = Δc + nΔB (0, T + 𝛿)
= [c (Δt) − c (0)] + n [B (Δt) − B (0)]
= (New call price − 172, 852.22) + 67, 553, 564.09 × (New bond price − 0.950054)
(25.24b)
- For example, in row 4, for i(0,1) = 0.0420, we get ΔV = (169,739 - 172,852.22) + 67,553,564.09
× (0.950100 - 0.950054) = $22.20.
TABLE 25.3: Delta Hedging Errors
for a Caplet
i(0, 1) B(0, T+ 𝛿) c ΔV
0.0321 0.954716 8,088.79 150,212.61
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
0.0411 0.950518 142,989.05 1,495.93
0.0420 0.950100 169,739.89 22.20
0.0421 0.950054 172,852.22 0
0.0422 0.950007 175,991.68 5.23
0.0431 0.949590 205,424.69 1,244.00
0.0521 0.945436 575,720.16 90,953.48
■ The table shows that for small changes in forward rates, the delta-neutral portfolio’s value changes
are small (for example, when i[0, 1] = 0.0420, ΔV = 22.20), but for larger changes in forward rates,
the hedging error increases significantly (for example, when i[0, 1] = 0.0521, ΔV = 90, 953.48).
Consistent with the underlying theory, this shows that for small changes in simple forward rates, the
hedgeperforms quite well.
646 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
Gamma Hedging
When Δt is small, delta hedging works well because the change in the simple for-
ward rates will be small. But if the hedging interval Δt is large, then the change in the
simple forward rates will be large, and one needs to gamma hedge. Gamma hedging is
hedging the changes in delta in addition to the changes in interest rates (akin to the
approach in Chapter 20).
Gamma hedging is a valid procedure because the risk of changes in interest rates
is already priced into the valuation methodology. We note, however, that gamma
hedging is less important for interest rate derivatives as compared to stocks because
the changes in forward rates i(t,T) are smaller over daily and weekly trading intervals
than price changes are for a stock.
As discussed in Chapter 20, adding a gamma hedge on top of a delta hedge requires
an additional hedging security. Two choices seem natural here: another zero-coupon
bond or another caplet. If we use another zero-coupon bond, then we need to know
the impact of changing i(0,T) on this other zero-coupon bond. This is a complex
problem that is not easily handled in the HJM libor model. Alternatively, we can use
another caplet on the same forward rate but with a different maturity. We take the
second approach because this caplet’s delta is available as an analytic equation. As with
delta hedging, we set up a delta–gamma hedge at time 0, but we could have done
this at any time t.
Let c2 be another caplet on the same maturity forward rate but with a different
strike rate. We form a portfolio consisting of the initial caplet c1, n1 number of shares
of the zero-coupon bond B(0,T + 𝛿) ≡ B, and n2 number of shares of the additional
caplet with a different strike rate. Then, the change in the value of the resultant
portfolio over the time interval [t, t + t] is
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
ΔV = Δc1 + n1 ΔB + n2 Δc2 (25.25a)
Then, to be delta and gamma neutral, we need to choose both n1 and n2 such that
𝜕c1 𝜕B 𝜕c
ΔV = Δi + n1 Δi + n2 2 Δi
( 𝜕i 𝜕i 𝜕i )
(25.25b)
1 𝜕2 c1 2 𝜕2 B 2 𝜕2 c2 2
+ Δi + n1 Δi + n2 Δi = 0
2 ( 𝜕i2 𝜕i2 𝜕i2 )
Equating the terms in the first parentheses (which has the coefficients of Δi) and the
second parentheses (which has the coefficients of Δi2 ) to zero, we get two equations
in two unknowns. Their solution is
𝜕2 c1 𝜕B 𝜕c1 𝜕2 B 𝜕c2 𝜕2 B 𝜕2 c2 𝜕B
n2 = − ÷ − 2 (25.25c)
( 𝜕i2 𝜕i 𝜕i 𝜕i2 ) ( 𝜕i 𝜕i2 𝜕i 𝜕i )
𝜕c1 𝜕c 𝜕B
n1 = − + n2 2 ÷ (25.25d)
( 𝜕i 𝜕i ) 𝜕i
SUMMARY 647
These give the prescribed holdings for setting up a delta-gamma-neutral portfolio.
Because the computation is otherwise similar to that contained in Chapter 20 we do
not provide a numerical example to illustrate this procedure.
25.10 American Options, Futures,
Swaptions, and Other Derivatives
How does one price and hedge American options, futures, and other interest rate
derivatives? Unfortunately, the HJM libor model’s formulas are of little help. To
understand why, we have to explore some technical considerations. Note that in the
specification of the HJM libor model (assumption A6), we did not need to specify the
number of factors generating the evolution of the term structure of zero-coupon bond
prices. This is because the pricing of a single caplet only requires the specification
of the relevant simple forward rate and the (T + 𝛿) maturity bond. However, for
more complex interest rate derivatives, a specification of the factor structure and the
evolution of the entire term structure of zero-coupon bond prices are required. This is
similar to what we did in Chapter 24 but extended to the continuous-time setting.
For pricing American options, just as in the BSM model, numerical procedures
need to be employed. The discrete-time HJM model as presented in Chapters 23 and
24 can be used in this regard. The parameters to input into the discrete-time HJM
model to obtain the HJM libor model as the time step goes to zero are given in the
appendix to Chapter 24.
Similarly, the pricing of futures, swaptions, and other interest rate derivatives
requires the use of numerical methods.4 This is because under the HJM libor
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
model, the implied continuously compounded forward rates do not have a lognormal
distribution. Consequently, the evolution of the term structure of continuously
compounded forward rates is quite complex. The discrete-time HJM model as
presented in Chapters 23 and 24 can provide a numerical procedure for use in this
regard. A presentation of the analogous continuous-time HJM model is left for more
advanced courses (see, e.g., Jarrow [2002b] for an introduction to such models).
25.11 Summary
1. Caplets and floorlets are European call and put options, respectively, on interest
rates (as measured by simple forward rates). They trade as bundles: a cap is a col-
lection of caplets with identical terms and conditions, except for their expiration
dates, and a floor is a similar collection of floorlets.
2. Fischer Black (1976) developed a model that was later modified to price caplets.
Black’s original model was inconsistent with the no-arbitrage assumption.
4
If the simple forward rate is assumed to be lognormal, then the swap rate, being an average of forward
rates, is not log normally distributed so that a simple Black-type model does not apply.
648 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
3. Sandmann et al. (1995), Miltersen et al. (1997), and Brace et al. (1997) indepen-
dently developed the HJM libor model, which assumes that simple forward rates
are lognormally distributed in an HJM model. The simple forward rate is defined
at time 0 and applies over the time interval [T,T + 𝛿], that is,
B (0, T) 1
i (0, T) = −1 ×
( B (0, T + 𝛿) ) 𝛿
4. The HJM libor model formula for caplet pricing is
c = B (0, T + 𝛿) [i (0, T) N (d1 ) − kN (d2 )] × 𝛿 × LN
where B(0, T + 𝛿) is today’s price of a zero-coupon bond that matures at time
(T + 𝛿), T is the caplet’s maturity date, 𝛿 is the time period that interest is earned,
i(0, T) is the simple forward rate per year effective over the interval [T, T + 𝛿],
N( . ) is the cumulative standard normal distribution function (with mean 0 and
standard deviation 1), d1 = (log [i(0, T)/k] + 𝜈2T T/2)/𝜈T √T and d2 = d1 − 𝜈T √T, k
is the strike rate, 𝜈T is the average forward rate volatility over the caplet’s life, and
LN is the notional principal.
5. The easiest inputs to obtain are the strike rate k, expiration date T, and time period
that interest is earned 𝛿 because they are written into the contract. Forward rates
are the same as forward rate agreement rates, which are quoted in the OTC market.
FRA rates or Eurodollar rates (for maturities up to one year) and swap rates (for
maturities beyond one year) can be used to compute zero-coupon bond prices.
6. The average forward rate volatility is computed by the equation
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
∫0T 𝜎2s ds
𝜈T = √
T
where 𝜎s is the time s – maturity simple forward rate i(0,s)’s volatility per year.
7. Just as for the BSM formula, one can use market prices of caps, as portfolios of
caplets, to find that average forward rate volatility 𝜈T that best matches market
prices. This is the implied volatility.
8. Caplet–floorlet parity is given by
c − p = B (0, T + 𝛿) [i (0, T) − k] × 𝛿 × LN
where c and p are today’s (time 0) prices of caplets and floorlets that have strike rate
k and expiration date T, respectively; B(0,T + 𝛿) is today’s price of a zero-coupon
bond that pays $1 at time (T + 𝛿); i(0,T) is the simple forward rate over the period
[T,T + 𝛿]; and LN is the notional principal.
9. The HJM libor model can be used for delta and gamma hedging interest rate risk.
c
The delta hedge is obtained by buying − + N(d1 )[1 + i(0, T)𝛿]LN
B(0, T + 𝛿)
APPENDIX 649
shares of a zero-coupon bond worth B(0,T + 𝛿). Gamma hedging takes a similar
approach. However, these techniques do not extend to more complex interest
rate derivatives like American options and futures contracts, which do not have
closed-form solutions.
25.12 Appendix
This appendix proves the equation in the text related to risk neutral valuation.
Proof of Equation 25.11
First, Jarrow (2010) shows that no arbitrage implies the existence of pseudo
probabilities 𝜋T+𝛿 such that at time t,
B(t, K) 𝜋 B(T + 𝛿, K) 𝜋
= Et T+𝛿 = Et T+𝛿 [B(T + 𝛿, K)] for all K ≤ T + 𝛿
B(t, T + 𝛿) [ B(T + 𝛿, T + 𝛿) ]
𝜋
and 0 ≤ t ≤ K where Et T+𝛿 [⋅] denotes the time t conditional expectation using
the pseudo probabilities 𝜋T+𝛿 . We note that the third equality uses the fact that
B(T + 𝛿, T + 𝛿) = 1.
B(t, K)
This equation states that the ratio is a martingale under the 𝜋T+𝛿
B(t, T + 𝛿)
probabilities. The difference in this ratio, as contrasted with the ratio in Chapters
23 and 24, is that the denominator is the zero-coupon bond B(t, T + 𝛿) and not the
mma A(t).
Rewritten, this no arbitrage condition implies that at time 0:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
B (0, K) 𝜋 B (t, K)
= E T+𝛿
B (0, T + 𝛿) [ B (t, T + 𝛿) ]
for any t ∈ [0, K].
Next, replace t = T and K = T in this equation. Algebra and using the definition
B(t, T)
i(t, T) = − 1 for t = 0 and t = T yields Equation 25.11:
B(t, T + 𝛿)
𝜋T+𝛿
i (0, T) = E [i (T, T)] .
Proof of Equation 25.15
From Chapter 19, we have Equation 19.2:
1
∫0T 𝜇ds− 𝜎2 (T)+𝜎√T⋅z
S (T) = S (0) e 2
2
∫0T 𝜎2 ds ∫0T 𝜇ds
where 𝜎 = and 𝜇 = .
T T
650 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
S(t)
We have that under the pseudo probabilities 𝜋, rT is a martingale, i.e., Equation
e
19.6 holds:
S (t) 𝜋 S (T)
= Et T+𝛿 .
ert ( erT )
From Chapter 19, we have that
⎛ ∫0T 𝜇ds ⎞
⎜ − r⎟
𝜃 = − ⎜⎜ T ⎟ √T
⎟
⎜ 𝜎 ⎟
⎝ ⎠
and
𝜃2
− +𝜃z
𝜙 (z) = e 2 .
For the the pseudo probabilities 𝜋T+𝛿 , identify: S(0) = i(0, T), r = 0, 𝜎 = vT , and 𝜇
= 𝜇s . We see that Equation 19.6 becomes
𝜋
i (t, T) = Et T+𝛿 (i (T, T))
which is desired Equation 25.11. So the change in probabilities are given by
⎛ ∫0T 𝜇s ds ⎞
⎜ ⎟ ∫0T 𝜇s ds
⎜ T ⎟
𝜃 = −⎜ ⎟ √T = − (
⎜ vT ⎟ vT√T )
⎝ ⎠
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
and
𝜃2
− +𝜃z
𝜙 (z) = e 2 .
Changing notation: 𝜃 = 𝜃T+𝛿 and 𝜙(z) = 𝜑T+𝛿 (z), we get the final result:
∫0T 𝜇s ds
𝜃T+𝛿 = −
( vT √T )
and 𝜃2T+𝛿
− +𝜃T+𝛿 z
𝜑T+𝛿 (z) = e 2 .
APPENDIX 651
We just need to check that 𝜋T+𝛿 is a probability as defined in Equation 25.25a.
Indeed, we have that
2
𝜃2 z
1 − +𝜃z−
𝜋T+𝛿 (z) = 𝜑T+𝛿 (z) q (z) = e 2 2 ≥0
√2𝜋
and
∞
∫ 𝜋T+𝛿 (z) dz
−∞
∞
= ∫ 𝜑T+𝛿 (z) q (z) dz
−∞
∞ 𝜃2 z2
1 −+𝜃z−
= ∫ e 2 2 dz
√2𝜋 −∞
∞ 1
1 (𝜃−z)2
−
= ∫ e 2 dz = 1
√2𝜋 −∞
since this is normally distributed with mean 𝜃 and variance 1.
Risk Neutral Valuation.
Let xT be a random variable at time T.
Risk neutral valuation using the pseudo probabilities 𝜋T+𝛿 gives
x0 xT
= E𝜋T+𝛿 .
B(0, T + 𝛿) ( B(T, T + 𝛿) )
Or,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
xT
x0 = B (0, T + 𝛿)E𝜋T+𝛿 .
( B (T, T + 𝛿) )
Alternatively, letting xT+𝛿 be a random variable at time T + 𝛿, this simplifies to
x0 xT+𝛿
= E𝜋T+𝛿 = E𝜋T+𝛿 (xT+𝛿)
B (0, T + 𝛿) ( B (T + 𝛿, T + 𝛿))
since B(T + 𝛿, T + 𝛿) = 1, i.e.
x0 = B (0, T + 𝛿) E𝜋T+𝛿 (xT+𝛿 ) .
Proof of Equation 25.12
We are given the following payoff table:
0 T T+𝛿
− − − − − −
Time resolution payment
c0 cT = B (T, T + 𝛿) max (i (T, T) − K, 0) 𝛿LN max (i (T, T) − K, 0) 𝛿LN
652 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
Applying the risk neutral valuation formula
B (T, T + 𝛿) max (i (T, T) − K, 0) 𝛿LN
c0 = B (0, T + 𝛿) E𝜋T+𝛿
( B (T, T + 𝛿) )
𝜋T+𝛿
= B (0, T + 𝛿) E (max (i (T, T) − K, 0) 𝛿LN ) .
This is Equation 25.12.
25.13 Cases
Risk at Freddie Mac (Stanford Business School Case F270, European Case Clearing
House Case). The case considers the interest rate risk management of Freddie
Mac’s mortgage portfolio.
Student Educational Loan Fund Inc. (Abridged) (Harvard Business School Case
201083-PDF-ENG). The case studies how an organization that makes student
loans can use an interest rate derivative to change the loan terms from a variable-
rate loan with semiannual payments to a fixed-rate loan with equal monthly
payments.
Remy Cointreau: Debt Management and Yield Curves (Case 111-039-1, Euro-
pean Case Clearing House Case). The case examines interest rate risk management
at a leading wines and spirits manufacturer, which keeps a large inventory and
which has a significant amount of financial debt of different maturities.
25.14 Questions and Problems
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
25.1. In the Heath–Jarrow–Morton libor model, what are the two assumptions that
differ relative to the Black–Scholes–Merton model?
25.2. a In the HJM libor model, is it a simple interest rate or a continuously
compounded interest rate that follows a lognormal distribution?
b Why is this difference important (hint: relates to Black’s formula)?
c In the HJM libor model, is the volatility of the simple forward rate of interest
a constant, or does it depend on time?
25.3. Discuss the history of evolution of interest rate option pricing models that
started from the BSM model and ended with the HJM libor model.
25.4. In pricing a caplet with maturity T and strike rate k, at what date is the payment
received, and what is the payoff at time T?
25.5. Use the caplet valuation formula, Equation 25.5, to answer the following
question. Suppose two investors disagree on the expected percentage change
in the simple forward rate over the caplet’s life. Will the two investors still agree
on the caplet’s price?
25.6. Suppose that the time 0 simple forward rate i(0,2) is 0.045 per year, the notional
principal is LN is $40 million, the strike rate k is 0.04 per year, the time period
QUESTIONS AND PROBLEMS 653
for interest computation 𝛿 is three months, the caplet matures in T = 2 years,
the average forward rate volatility is 0.15 per year, and a zero-coupon bond
maturing after 2.25 years is worth B(0,T + 𝛿) = $0.89 today. What is the value
of the caplet?
25.7. In the HJM libor caplet value, because the volatility of the forward rate changes
over the caplet’s life, what volatility is input into the formula? How does this
differ from the volatility input into the BSM model for an equity option?
25.8. There are two ways to estimate the forward rate’s average volatility over the
caplet’s life. What are they? What are their advantages and disadvantages?
25.9. Given the following data, compute the historical volatility of the constant-
maturity forward rates.
Weekly Forward Price Data
Constant Maturity Simple Forward Rate i(n, s)
Week n s=0 s = 0.25 s = 0.5 s = 0.75 s = 1.00 s = 1.25
1 0.1020 0.1060 0.1120 0.1160 0.1230 0.1160
2 0.1050 0.1070 0.1140 0.1160 0.1260 0.1190
3 0.1080 0.1090 0.1150 0.1180 0.1270 0.1230
4 0.1090 0.1110 0.1150 0.1210 0.1280 0.1250
5 0.1100 0.1130 0.1180 0.1210 0.1290 0.1280
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
6 0.1100 0.1150 0.1190 0.1250 0.1300 0.1310
7 0.1110 0.1150 0.1200 0.1240 0.1340 0.1310
8 0.1120 0.1160 0.1210 0.1270 0.1340 0.1350
9 0.1120 0.1170 0.1210 0.1280 0.1350 0.1360
10 0.1110 0.1180 0.1210 0.1270 0.1360 0.1310
11 0.1090 0.1170 0.1200 0.1260 0.1350 0.1280
12 0.1070 0.1130 0.1180 0.1230 0.1290 0.1210
T is 1.25 years; 𝛿 is 0.5 years.
25.10. Given the market price of the caplet is $209,801.727, and using the following
inputs for the caplet (notional $100 million, strike rate k = 4 percent, maturity
1 year, six-month bbalibor i(0,1) = 0.421 with 𝛿 = 0.4986, B(0,T + 𝛿) =
$0.95) compute the implied average forward rate volatility over the caplet’s
life.
654 CHAPTER 25: THE HEATH–JARROW–MORTON LIBOR MODEL
25.11. Is the forward rate an unbiased estimate of the future spot rate of interest?
Explain your answer.
25.12. Why are the actual and pseudo-probabilities different? Do the actual proba-
bilities influence the caplet’s value? Explain your answer.
25.13. Is the caplet’s value equal to the expected discounted value of its payoff at
maturity using the actual probabilities? If not, provide and explain the correct
answer.
25.14. Consider a floorlet that has the identical terms as the caplet in problem Using
caplet–floorlet parity, compute the floorlet’s value.
25.15. If you want to hedge a caplet using a zero-coupon bond of maturity (T + 𝛿),
what is the hedge ratio that you should use?
25.16. Given the caplet solution in problem 25.6, compute the caplet’s delta.
25.17. Why would one use gamma hedging with the HJM libor model? Why would
one not use vega hedging with the HJM libor model?
25.18. What is caplet–floorlet parity? Explain your answer.
25.19. If one wants to price an American option using the HJM libor model, is there
a closed-form solution? If not, how can one price such an option?
25.20. How would one price an interest rate futures in the HJM libor model?
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
26
Risk Management Models
26.1 Introduction EXTENSION 26.2 Real Options
26.2 A Framework for Financial Reduced-Form Models
Risk Management EXTENSION 26.3 Credit Default
Swaps
26.3 Computing the Loss
Liquidity Risk
Distribution
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Operational Risk
26.4 Value-at-Risk and Scenario
Analysis 26.6 The Future of Models and
Traded Derivatives
Value-at-Risk
Model Risk
Scenario Analysis
Derivatives
EXTENSION 26.1 Risk Measures
26.7 Summary
26.5 The Four Risks
Market Risk 26.8 Cases
Credit Risk 26.9 Questions and Problems
Structural Models
656 CHAPTER 26: RISK MANAGEMENT MODELS
26.1 Introduction
Even in the rarefied world of global investment banking, Goldman Sachs Group Inc.,
whose star-studded list of former employees includes a prime minister of Italy, several
US Treasury secretaries, a World Bank president, and some central bank governors,
enjoys a pristine reputation. According to Nocera, in December 2006, Goldman
Sachs discovered that various risk measures, including value-at-risk (VaR), indicated
that something was wrong in the mortgage-backed securities market.1 The company
convened a meeting of risk managers and senior executives. They scrutinized every
trading position, debated the results from their financial models, discussed the market,
and decided to “get closer to home” (which means, in trader lingo, to reduce their
mortgage-backed securities risk exposure). This move enabled Goldman Sachs to
avert billions of dollars in losses when the housing market crashed in summer 2007
and to avoid the fate of many failed Wall Street firms like Bear Stearns and Lehman
Brothers.
This story also highlights what Bernstein (1998) described as “a persistent tension
between those who assert that the best decisions are based on quantification and
numbers, determined by the patterns of the past [emphasis added], and those who base
their decisions on more subjective degrees of belief about the uncertain future [emphasis
added].”2 What good are models that fail when they are most needed? This is an
often-heard criticism coming from practitioners, prophets, and even some professors
who experienced the financial crisis of 2007.
We disagree! Our reading of the crisis is very different. In the Goldman Sachs
example, both “mind” and “model” worked exactly as they should. By identifying
trouble brewing, the models did their job. A refined understanding of models—their
benefits and limitations—saved the day. Alternatively stated, it is not the models being
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
poor but the poor use of models that helped caused the financial crisis.
This chapter concludes our study of derivatives pricing models. The first twenty-
five chapters have shown us how to price and hedge derivatives subject to market
risk, which includes the risks generated by equity prices, commodity prices, foreign
currencies, and interest rates. This chapter combines this knowledge and adds
the remaining risks—credit, liquidity, and operational—to build a unified risk
management framework, unique in its construction and presentation.
An outline for this chapter is as follows. First, we present the financial risk
management framework using the firm’s balance sheet as the focal point for analysis
(see Figure 26.1). The firm’s risk is characterized by the firm’s equity loss distribution.
Well-known risk management tools such as VaR and scenario analysis are easily
understood from this perspective. Next, we study the four risks, market risk, credit
risk (from which the concept of real options follows), liquidity risk, and operational
risk, and discuss how to model each of these in a consistent fashion. We conclude
this chapter with comments on the future of models and traded derivatives.
1
“Risk Mismanagement,” New York Times, January 4, 2009.
2
See Against the Gods: The Remarkable Story of Risk by Peter L. Bernstein (1998, p. 6).
INTRODUCTION 657
FIGURE 26.1: A Framework for Corporate Financial
Risk Management
Accounting identity: At – Lt Et
Prob{∆Et ≤ x}
Computation of the loss distribution
Value-at-risk Operational
(an application) risk
Scenario analysis
(an application)
Liquidity
Market risk
risk
The topic of this book
(BSM and HJM)
Credit risk
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Credit default Real Structural Reduced-
swaps options approach form
(an application) approach
The first four sections of this chapter on risk management are written at a
higher level than the previous twenty-five chapters were. This is done because
risk management and generalizing the assumptions underlying the Black–Scholes–
Merton (BSM) and Heath–Jarrow–Morton (HJM) models are the next steps in
derivatives pricing and hedging, and we wanted to give the reader a flavor of what’s
to come. As such, these four sections give a quick overview of the relevant topics and
the mathematics involved, not an in-depth analysis. The last section, on the future of
derivatives and models, however, is written at an introductory level and requires no
new analytical techniques.
658 CHAPTER 26: RISK MANAGEMENT MODELS
26.2 A Framework for Financial Risk
Management
From a society’s point of view, the most important use of derivatives models is in
their application to managing the risk of a firm’s equity capital (corporation, financial
institution) or an individual’s wealth. By managing this equity risk, the key players in
the economy hold assets and liabilities more aligned with their needs and preferences.
These firms and individuals hedge risks they want to avoid and speculate on risks for
which they have special knowledge. This risk management facilitates the optimal
allocation of risk across the economy, increasing societal welfare and economic
growth. The models presented in the first twenty-five chapters of this book can be
used as the foundation for financial risk management.
Financial risk management is the management of the risk of a firm’s (corporate
or financial) equity capital. To understand the issues involved in managing a firm’s
equity risk, we need to visit the typical firm’s balance sheet presented in Table 26.1.
A balance sheet gives a snapshot of a firm’s financial condition by summarizing its assets,
liabilities, and ownership equity on a specific date.
On the left side of the balance sheet are the firm’s assets, which provide economic
benefits to the firm. We write the time t asset’s value as At (see Chapter 1 for the
definitions of assets, liabilities, and other relevant terms). Depending on whether the
assets are for the short term (one year or less) or the long term, assets (and liabilities)
are further classified as current or noncurrent, respectively. Current assets include cash,
accounts receivable, and inventory. Long-term assets consist of both financial and real
assets—property and equipment.
On the right side of the balance sheet are the firm’s liabilities and equity. The
liabilities are a firm’s obligations that require payment at some future date. We denote
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
their time t value by Lt . Equity is the shareholder’s wealth, analogous to an individual’s
wealth, its time t value denoted by Et .
TABLE 26.1: A Firm’s Balance Sheet
Assets At Liabilities Lt
Current assets Accounts payable
■ Cash and cash equivalents
■ Financial liabilities
■ Accounts receivable
■ Pension fund obligations
■ Inventories
Long-term assets Equity Et
■ Financial assets ■ Ownership shares
■ Property, plant, and equipment
A FRAMEWORK FOR FINANCIAL RISK MANAGEMENT 659
By definition, the difference between the firm’s assets and liabilities is the share-
holder’s equity. Hence the familiar accounting identity holds:
At − Lt ≡ Et (26.1)
Risk management can be understood by focusing on the various components of this
basic accounting identity. The firm’s management controls the purchase and sale of
the firm’s assets (called the capital budgeting decision), the accumulation of the firm’s
liabilities (the composition, term structure, and magnitude), and the magnitude of
the equity (dividends, new issuance). These decisions are the domain of corporate
finance (see Ross et al. 2004).
For risk management purposes, the equity holders are concerned with changes in
the equity’s value over some time interval [t, t + Δt]. For many decisions, Δt might
be a day, a week, or a year. In symbols, the change in the equity value is defined as
ΔEt ≡ Et+Δt − Et (26.2)
Since this is for internal risk management, before bankruptcy and/or liquidation
proceedings, it is possible that given a large enough loss, the value of the firm’s equity
could become strictly negative, Et+Δt < 0, wiping out all of the firm’s equity and
even more.
Of course, from expressions (26.1) and (26.2), the change in equity value is
determined by the changes in the value of the assets and liabilities:
ΔEt ≡ ΔAt − ΔLt (26.3)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
where ΔAt ≡ At+Δt − At and ΔLt ≡ Lt+Δt − Lt .
The primary tool to use for analyzing changes in the firm’s equity is its probability
distribution, that is,
Prob {ΔEt ≤ x} = Prob {ΔAt − ΔLt ≤ x} (26.4a)
for all values x.
From this distribution, one can determine all risk management measures of interest
to equity holders and regulators. For example, the probability that the firm will be
insolvent over the time interval [t, t + Δt] is given by
Prob {ΔAt − ΔLt ≤ −Et } (26.4b)
The firm is insolvent if the change in the value of the equity (the left side of the
inequality) is negative and greater than or equal to the existing value of equity. In
terms of the balance sheet, this occurs if the change in the assets minus the change
in the liabilities is negative and greater than the magnitude of the equity cushion at
660 CHAPTER 26: RISK MANAGEMENT MODELS
time t. For example, if assets increase by $1 billion and liabilities increase by $5 billion,
and the value of shareholders’ equity is $3.5 billion, then
ΔEt ≡ ΔAt − ΔLt = 1 − 5 = −4 < −3.5 = −Et
and the firm is insolvent.
Note that this condition implies that since the equity shares are of limited liability,
after bankruptcy proceedings and/or liquidation, the value of the equity becomes
zero, and the liability holders eventually suffer the extra losses that the equity holders
do not.
Regulators are concerned with the ability of financial institutions and some large
corporations, deemed too big to fail, to remain solvent, that is, to have enough
equity capital to fulfill the obligations promised to its liabilities. For these capital
determination decisions, one focuses on the lower tail of this distribution, where the
change in equity value is negative (see Figure 26.2). The change in the value of equity
is negative if the firm has losses. This occurs if the change in the value of the assets is
less than the change in the value of the liabilities. Taking this regulatory perspective,
the distribution for the changes in the value of the firm’s equity is called the loss
distribution. For these risk management decisions, we need to understand how to
compute this loss distribution. This is the topic of the next section.
26.3 Computing the Loss Distribution
Computing the loss distribution for changes in the firm’s equity value is a difficult task.
To facilitate this task, one can think of the value of the firm’s assets and liabilities as
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
functions of a vector of state variables, Xt , describing the state of the economy at time
t, that is, At (Xt ) and Lt (Xt ). The state variables could be macroeconomic quantities
like inflation, gross domestic product, and unemployment, or they could be the prices
of the primary or basic securities themselves: zero-coupon bond prices, stock prices,
commodity prices, or foreign currency exchange rates. When computing these asset
and liability values, of course, one must first decompose these values into their
component parts and model each component part separately.
It is important to recognize that the functions At ( Xt ) and Lt ( Xt ) are those formulas
determined by the derivatives pricing theory studied in the previous chapters, for
example, the BSM formula for stock options or the HJM formula for caps, where
Xt corresponds to the price of a stock and the term structure of zero-coupon bond
prices, respectively. When computing these formulas, four risks need to be included:
market, credit, liquidity, and operational. These four risks were discussed briefly in
Chapter 1. Let us recall these descriptions.
Market risk is the risk from random movements in stock prices, interest rates,
foreign currencies, and commodity prices. Credit risk corresponds to the risk that a
contractual claim will not be executed as agreed, thereby leading to default. Liquidity
risk is the risk that when trading large quantities of a financial security in a short time
period, there will be a quantity impact on the price to the trader’s disadvantage. Last,
COMPUTING THE LOSS DISTRIBUTION 661
operational risk is the risk that the firm’s operations may generate losses because of
errors, mismanagement, fraud, accidents, or legal considerations.
The first twenty-five chapters of this book have concentrated on market risk. In
these chapters, by assumption, we excluded credit, liquidity, and operational risk. The
extension of our models to these other three risks is discussed later on in the chapter.
For the moment, however, let us suppose that we know how to include all four risks
in our valuation formulas.
Next, given an assumed probability distribution for ∆Xt that incorporates all four
of these risks, one can then determine the probability distribution for the firm’s
assets, liabilities, and, finally, equity using the functions determined by derivatives
pricing theory:
Prob{ΔEt (Xt ) ≤ x} = Prob{ΔAt (Xt ) − ΔLt (Xt ) ≤ x} for all x
Summarizing, a method for computing the probability distribution for losses is:
1. Choose the primary securities or state variables describing the state of the market
or the economy.
2. Assume an arbitrage-free random evolution for the primary securities in a
complete market setting or a random evolution for the state variables.
3. Value all of the firm’s assets and liabilities as derivatives on the primary securities or
state variables. This gives the functional forms for their values: At (Xt ) and Lt (Xt ).
4. Compute the distribution for the changes in the assets and liability values
based on the functional forms and the assumed evolutions. This computation
could be generated using Monte Carlo simulation methods or analytic formulas.
If necessary, to simplify this computation, one can use linear and quadratic
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
approximations for the values via the derivative’s deltas and gammas.
Note that correlations across the changes in the collection of asset and liability
values within the firm are generated by the dependence of the functional forms
on the common primary securities or state variables vector Xt . This state variable
dependence includes changes in equity value resulting from changes in the health
FIGURE 26.2: Graph of the Loss Distribution
Probability density
Probability of insolvency
–Et 0 ∆At – ∆Lt
662 CHAPTER 26: RISK MANAGEMENT MODELS
of the economy as reflected in the business cycle. In turn, the dependence on
the business cycle captures correlated defaults within the firm’s asset and liability
portfolio. In practice, the computation becomes complex and is often generated using
Monte Carlo simulation techniques rather than analytic formulas. The details of these
methods are left for future reading and can be found in more advanced books.
26.4 Value-at-Risk and Scenario Analysis
The approach we have presented for computing the loss distribution underlies two
common risk management statistics—VaR and scenario analysis—both of which are
used extensively by corporations, policy makers, and regulators.
Value-at-Risk
Value-at-Risk (VaR) was originated in 1989 by Sir Dennis Weatherstone, then the
chairman of J. P. Morgan & Co. Sir Dennis desired a one-page report that could
measure the total risk faced by his global banking firm. He wanted it placed on his
desk within fifteen minutes of the New York Stock Exchange’s close. This 4:15 report
summarized the risk of J. P. Morgan’s balance sheet using VaR.
VaR is a summary measure of the loss distribution for the change in the firm’s
equity value over a fixed period of time [t, t + Δt], where Δt could be a day, week,
or a year. Using the accounting identity relating the firm’s equity to the firm’s assets
and liabilities, VaR (VaR𝛼 ) for an 𝛼 loss probability is that number such that
Prob {ΔEt ≤ −VaR𝛼 } = Prob {ΔAt − ΔLt ≤ −VaR𝛼 } = α
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
In words, VaR𝛼 is that dollar amount that the firm’s losses will exceed with probability
𝛼 (see Figure 26.3). In practice, 𝛼 is often set to be 1 or 5 percent. A minus sign appears
before VaR𝛼 in this expression so that the computed VaR𝛼 > 0. Let us consider some
VaR computations to help us understand its application and significance.
FIGURE 26.3: Value-at-Risk
Probability density
α percent probability
–Varα 0 ∆At – ∆Lt
VALUE-AT-RISK AND SCENARIO ANALYSIS 663
To understand how to use VaR𝛼 in risk management decisions, let us suppose
that 𝛼 = 0.01 is the insolvency probability that the firm and the regulators feel is
appropriate. Given the VaR𝛼 , the firm’s management can consider if it is properly
capitalized for its insolvency risk by comparing VaR𝛼 to the firm’s existing equity
capital position Et .
EXAMPLE 26.1: Computing Value-at-Risk
■ To compute VaR, one views the firm as a portfolio of “securities,” where the securities are the firm’s
assets and liabilities. The assets have a positive portfolio weight, while the liabilities have a negative
weight. Of course, the portfolio weights must sum to one.
■ Consider a firm which has an asset and liability portfolio consisting of three companies (fictitious
names) Aztec, Inca, and Mayan corporations, with portfolio weights as given in Table 26.2A.
Furthermore, let the daily returns on these three companies be normally distributed with means,
volatilities, and correlations as given in Tables 26.2A and 26.2B. This characterizes the random
evolution of the primary securities underlying the firm’s assets and liability portfolio.
■ To compute a daily VaR based on this normal distribution, we need to first compute the expected
daily return and volatility of the firm’s asset and liability portfolio. The daily expected return of this
portfolio is 3
E (rp ) = wi E (ri )
∑
i =1
= (3/4) × (0.001) + (3/4) × (0.0005) − (1/2) × (0.001)
= 0.000625
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The daily return variance of this portfolio is
3
𝜎2p = w2i,j covariance (ri , rj )
∑
i,j =1
= (3/4)2 (0.0134164)2 + (3/4)2 (0.0089443)2 + (−1/2)2 (0.0178885)2
+2 (3/4) (3/4) (0.0134164) (0.0089443) (0.5)
+2 (3/4) (−1/2) (0.0134164) (0.0178885) (0.9)
+2 (3/4) (−1/2) (0.0089443) (0.0178885) (0.5)
= (0.0001012 + 0.000045 + 0.00008) + (0.0000675 − 0.000162 − 0.00006)
= 0.0002262 − 0.0001545
= 0.0000717
where we have used the well-known result that covariance (ri , rj ) = 𝜎i × 𝜎j × correlation (ri , rj ), where
i and j denote the companies, which are labeled 1, 2, and 3, respectively.
664 CHAPTER 26: RISK MANAGEMENT MODELS
■ The daily volatility is the square root of the variance:
𝜎p = +√0.0000717 = 0.0084676
TABLE 26.2A: Expected Return, Volatility, and Portfolio Weights
Aztec Corp. Inca Corp. Mayan Corp.
(Company 1) (Company 2) (Company 3)
Expected daily return 0.001 0.0005 0.001
Daily volatility of the returns 0.0134164 0.0089443 0.0178885
Portfolio weights 3/4 3/4 -1/2
TABLE 26.2B: Correlation of Daily
Expected Returns
Aztec Corp. Inca Corp. Mayan Corp.
Aztec Corp. 1 0.5 0.9
Inca Corp. 0.5 1 0.5
Mayan Corp. 0.9 0.5 1
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
■ Third, we can now compute the daily VaR at the 5 and 1 percent levels. Recall from statistics that
for a normal probability distribution, 95 percent of the values lie within 1.65 standard deviations of
the mean.
- The 5 percent left tail is 1.65 × 0.0084676 = 0.0139715 units away from the mean.
- Hence the 95 percent confidence level or −VaR0.05 is 0.000625 − 0.0139715 = − 0.0133465.
- This means that there is only a 5 percent chance that the firm will lose more than 1.33 percent of
its value on any given trading day.
■ As 99 percent of the values lie within 2.33 standard deviations of the mean, we get the following:
- The 1 percent left tail is 2.33(0.0084676) = 0.0197295 units away from the mean.
- The 99 percent confidence level or −VaR0.01 is 0.000625 − 0.0197295 = − 0.0191045.
- Thus there is only a 1 percent chance that the firm will lose more than 1.91 percent of its value
on any given trading day.
■ If VaR𝛼 < Et , then the firm’s insolvency probability is less than the 𝛼 probability, and
the firm is properly capitalized. Indeed, if a loss equal to VaR𝛼 occurs, then there
is enough equity capital to buffer this loss. The firm does not become insolvent.
VALUE-AT-RISK AND SCENARIO ANALYSIS 665
■ If VaR𝛼 > Et , then the firm’s insolvency probability is greater than 𝛼, because there
is not enough equity capital to cover the VaR𝛼 losses. The firm needs either to (1)
add more equity capital to its balance sheet or (2) restructure the risk of its assets
and liabilities to reduce the loss probability. For example, (1) the firm can issue new
equity or cut dividend payments and (2) it could sell assets and use these funds to
reduce its liabilities.
If it is used as the sole measure to quantify a firm’s risk, VaR𝛼 has a number of
failings. First, it does not consider the magnitude of the losses beyond the 𝛼 level. For
example, if the VaR𝛼 is $100 million, a 1 percent loss of an additional $150 million
beyond the VaR𝛼 is treated equally in the VaR𝛼 computation as an additional $500
million loss beyond VaR𝛼 . Clearly these different losses should not be treated as of
equal risk to the firm. Second, in particular situation, VaR𝛼 penalizes diversification.
Traditional portfolio theory views diversification of an asset and liability portfolio as
beneficial because it reduces risk. The fact that VaR𝛼 can increase instead of decrease
when a portfolio becomes more diversified is problematic. This possibility is illustrated
in the next example.
EXAMPLE 26.2: VaR𝛼 Penalizes Diversification
■ Strangely, VaR𝛼 can increase when a portfolio becomes more diversified! This is counter to the
intuition that you developed taking a modern investment course where diversification reduces risk.
This implies, of course, that VaR𝛼 is not always a good measure of a firm’s risk.
■ Suppose that there are two loans A and B that can be used to constitute a firm’s assets. Let A and B
have the loss distributions given in Table 26.3. Loans A and B are similar in that both have no loss with
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
probability 0.9991 and a $1 loss with probability 0.0009. We assume that the probabilities of default
are statistically independent for loans A and B.
TABLE 26.3
Loss Prob(loss A) Prob(loss B) Prob(loss [A+ B]/2)
$0 0.9991 0.9991 0.99820081
$0.5 0 0 0.00179838
$1 0.0009 0.0009 0.000000081
■ We compare two hypothetical firms. One firm invests a dollar in loan A. The second firm invests $0.50
equally in both loans A and B. The second firm, of course, has the more diversified loan portfolio
because it is spreading its $1 investment across two loans whose losses are statistically independent.
■ The second firm’s return is given in the third column. With probability (0.0009)(0.0009) =
0.000000081, both loans A and B will default, giving a $1 (= $0.5 + $0.5) loss to the second firm.
With probability 2(0.0009)(0.9991), either loan A or loan B will default, but not both, giving a $0.50
666 CHAPTER 26: RISK MANAGEMENT MODELS
(= $0.5 + $0) loss to the second firm. With probability (0.9991)(0.9991), neither loan will default,
giving a $0 loss to the second firm.
■ Let us choose 𝛼 = 0.001 for our VaR computation. The nondiversified firm consisting of only loan A
has VaR0.001 (A) = $0 because no losses occur with probability greater than 0.001. But the diversified
firm has VaR0.001 ([A + B]/2) = $0.50 because with probability 0.00179838, a loss of $0.5 happens.
The diversified firm’s VaR is larger! Hence the diversified portfolio would require more capital based
on this measure alone, which is an incorrect decision.
Because diversification reduces risk, this last failing is a serious flaw in using VaR𝛼 as
a risk measure. Used in isolation for the determination of equity capital, VaR𝛼 is
a dangerous statistic because of the two failings just discussed. However, used in
conjunction with other risk measures, it is a useful tool in the risk manager’s toolbox.
Scenario Analysis
During the financial crisis of 2007, the Federal Reserve Board required banks
receiving funds under the Troubled Asset Relief Program to compute an alternative
risk measure to decide if they could repay their TARP funds.3 This alternative measure
is based on scenario analysis, another useful technique in one’s risk management
toolbox.
Scenario analysis (which also goes by the popular name stress testing) starts with
a selection of a particular set of scenarios that the firm wants to protect against. These
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
scenarios correspond to various states of the economy represented by paths of the
state variables ΔXt over the time period [t, t + Δt]. For example, if Xt represents
a collection of macroeconomic variables (such as inflation, gross domestic product,
unemployment, and interest rates), then a set of scenarios might be that interest rates
increase by 1 percent, unemployment increases by 5 percent, inflation increases by
5 percent, and GDP falls by 2 percent. For these scenarios, the firm wants to compute
the losses realized on its assets and liabilities to see if it has sufficient equity capital to
avoid insolvency.
Using our notation, the firm picks a particular ΔXt and computes
ΔEt (Xt ) = ΔAt (Xt ) − ΔLt (Xt )
Given the formulas for the values of the assets and liabilities as a function of the
state variables, one then computes the change in the firm’s equity capital, and if it is
negative and exceeds the current equity capital position, one knows that this scenario
would cause the firm to become insolvent.
3
“Pay up! The U.S. Federal Reserve Asks Banks to Submit TARP Repayment Plans,” New York Daily
News, November 24, 2009.
VALUE-AT-RISK AND SCENARIO ANALYSIS 667
EXTENSION 26.1: Risk Measures
Value-at-risk is a useful risk measure, but as noted in the text, it has some flaws. As might be expected, there are
other risk measures that could be used either in conjunction with or instead of VaR𝛼 .
The study of risk measures for use in the determination of equity capital began with the paper by Artzner
et al. (1999). They defined coherent risk measures. To understand what these are, we need to step back and begin
with the definition of a risk measure.
A risk measure is a positive real valued function that quantifies the “risk” to changes in a firm’s equity
capital—losses—over some time period. The loss to a firm’s equity capital is, of course, a random variable. By
this definition, there are an unlimited number of risk measures. For example, the maximum loss, the expected
loss, the standard deviation of the loss, VaR𝛼 for an arbitrary 𝛼, and the present value of the loss are all risk
measures.
This leads to the question, how does one decide which risk measures are useful?
The answer is that to be of any use, a risk measure needs to possess some desirable properties or axioms that
make it a “good” risk measure. Artzner et al. defined four such axioms. To state these axioms, we need some
additional notation.
Let 𝜌( . ) be a risk measure and let ΔE be the change in a firm’s equity over a fixed time period. Let r be the
return on a riskless asset (a money market account) over this same time period.
Now we can state the four axioms:
(Translation invariance) 𝜌 (ΔE + α) = 𝜌 (ΔE) − α
(Subadditivity) 𝜌 (ΔE1 + ΔE2 ) ≤ 𝜌 (ΔE1 ) + 𝜌 (DE2 )
(Positive homogeneity) For λ ≥ 0, 𝜌 (λΔE) = λ𝜌 (ΔE)
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
(Monotonicity) For ΔE1 ≤ ΔE2 , 𝜌 (ΔE2 ) ≤ 𝜌 (ΔE1 )
The translation invariance axiom states that if one adds 𝛼 units of equity capital to the firm and invests it in
a riskless asset, the risk measure declines by exactly 𝛼 units. Subadditivity implies that a diversified portfolio
(combining two firms’ equity capital) is no more risky than the sum of each separate firm’s capital positions.
Positive homogeneity states that if you expand a firm’s assets and liabilities proportionately, the risk increases
proportionately as well. Last, monotonicity states that if changes in a firm’s equity capital always exceed another
firm’s (both plus and minus), then the firm with the larger changes is more risky. Any risk measure that satisfies
these four axioms is called a coherent risk measure.
As one might expect, there are many coherent risk measures and many risk measures are not coherent. For
example, VaR𝛼 is not a coherent risk measure because as shown in a previous example in the text, it violates
the subadditivity axiom. Perhaps surprising the premium paid for an insurance policy to insure against the firm’s
default is not a coherent risk measure either because it violates the translation axiom.
An example of a coherent risk measure is the expected value of ΔE. A more sophisticated coherent risk measure
̃
is the expected shortfall, defined as E{ΔE|ΔE ≤ −VaR𝛼 }, where E{.|ΔẼ ≤ −VaR𝛼 } denotes a conditional
expectation given that losses in equity values exceed VaR𝛼 . The proof that expected shortfall is a coherent
measure can be found in Follmer and Schied (2004).
Whether coherent risk measures are the proper risk measures depends on whether the four axioms are deemed
reasonable, and there is considerable disagreement on this point. For example, Jarrow (2002a) does not like the
668 CHAPTER 26: RISK MANAGEMENT MODELS
translation invariance axiom precisely because it excludes the premium paid for an insurance policy to insure
against the firm’s default as a coherent risk measure. To rectify the situation, Jarrow proposes an alternative set of
axioms that generates what he calls insurance risk measures. Other sets of axioms are also possible, and other risk
measures can be characterized in this manner. The study of risk measures is an exciting area of current research
(for a good summary of the different axiom selections, see Follmer and Schied 2004).
In this circumstance, if it is believed that this scenario is likely, the firm can either
(1) add more equity capital to its balance sheet or (2) restructure the risk of its assets
and liabilities to reduce the loss probability. Given that the loss is scenario specific, it
might also be possible to hedge this loss using derivatives, as discussed in the previous
chapters.
The problem with scenario analysis is that it does not assign probabilities to
the various scenarios selected. Hence the firm’s management does not know the
probability that this scenario will occur. Of course, this failing can be overcome
by selecting the scenarios to be considered in conjunction with the probability
distribution for ΔXt . Unfortunately, this procedure and related computations is left
for more advanced courses.
26.5 The Four Risks
The Basel Committee’s Risk Management Guidelines for Derivatives (July 1994) iden-
tified the following risks in connection with an institution’s derivative activities.4 In
contrast to the Basel Committee, we include legal risk in operational risk. Otherwise,
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
the two sets of definitions are identical.
■ Market risk is the risk to a firm resulting from adverse movements in market prices.
■ Credit risk is the risk that a counterparty will default on an obligation.
■ Liquidity risk is the risk that a firm may not be able to liquidate an asset or liability
position without incurring losses due to a quantity impact on the market price.
■ Operational risk is the risk that law suits, human error, operation failures, and
inadequate controls will result in losses.
This section discusses each of these risks and how to model them using the derivatives
pricing technology.
Market Risk
The pricing and hedging of market risk using derivatives has been the emphasis of
this book. The market risks considered in Chapters 1–25 included changes to stock
prices, commodity prices, foreign currency exchange rates, and interest rates. Our
models were generated under a set of assumptions that excluded credit risk (by
4
See www.bis.org.
THE FOUR RISKS 669
assumption A2), liquidity risk (by assumptions A1 and A3), and operational risk
(obviously!). To include these additional risks, we need to relax the assumptions
and consider more general models. The remainder of this section discusses these
generalizations.
Credit Risk
Credit risk is the risk that a counterparty will default on a contract. This is an important
consideration in trading activity because all counterparties may default. An example
serves to illustrate this risk. On June 26, 1974, the West German bank Bankhaus
Herstatt, which had received Deutschemark payments, was forced into liquidation
by West German regulators before it could make corresponding dollar payments in
New York (which had a six-hour time difference), causing losses for its counterparties.
This incident alerted the market about cross-currency settlement risk, also known
as Herstatt risk (which Reuters Financial Glossary defines as “the risk that when
dealing with an overseas client a bank may not be able to recover its funds if the
counter party defaults on its payment obligation”5 ). This event led to the development
of a continuously linked settlement system for foreign exchange payments.
There are two approaches to modeling credit risk: structural and reduced-
form models. Structural models are sometimes called contingent claims models,
as originated by Merton (1974). Jarrow and Turnbull (1992, 1995) originated
reduced-form models to overcome a restrictive assumption contained in structural
models. Structural models assume that all of the firm’s assets trade in frictionless and
competitive markets and, therefore, their prices are observable. This assumption is
usually violated in practice. Most firms’ assets are very illiquid and often do not trade,
for example, a firm’s home office building or manufacturing plant, including the
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
manufacturing equipment. In contrast, reduced-form models replace this assumption
with a weaker restriction. Reduced-form models only assume that a subset of the
firm’s liabilities (debt) trade. The subset of the firm’s liabilities that must trade is
the one for which pricing and hedging is being considered. This small difference,
however, leads to a big change in valuations and risk management.
From an economic perspective, structural models are best applied to internal
management decision making, in which, even though the assets do not trade in
frictionless and competitive markets, the management can “observe” their prices.
In this context, structural models aid economic intuition. In contrast, reduced-form
models are best applied to pricing, hedging, and risk management decisions that
involve the trading of credit risky securities in financial markets. As such, they are the
preferred modeling approach to risk management in our current market environment.
They are easily estimated, and empirical studies show that they fit market prices well
(for a review, see Jarrow 2009).
5
http://glossary.reuters.com/index.php/Herstatt_Risk.
670 CHAPTER 26: RISK MANAGEMENT MODELS
TABLE 26.4: A Firm’s Time t Balance
Sheet
Assets At Liabilities Lt
Zero-coupon bond Maturity T, Face value K
Equity Et
Structural Models
To understand the intuition behind the structural models, it is best to start with
Merton’s (1974) original formulation. We consider a simple firm with the liability
structure given in Table 26.4. As before, this firm has assets with time t value At against
which it has issued debt. The debt consists of a single zero-coupon bond with face
value K and maturity date T. We let Lt be the time t value of the firm’s debt and
Et be the time t value of the firm’s equity. This is the simplest liability structure a firm
can possess. This firm is too simple for this model to be used in practice, but it is just
right for us to understand the economics.
Merton’s structural model is a transformation of the BSM model of chapter 19.
The structural model uses assumptions A1–A8 from the BSM model, with two
modifications. First, assumption A2, no credit risk, is removed because the purpose of
the structural model is to understand when the firm will default on its debt. Second,
the firm’s asset value replaces the stock price in assumptions A7 and A8. In particular,
the structural model assumes that the firm’s assets trade continuously in time and that
the asset value process follows a lognormal distribution. Assumption A6, no interest
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
rate uncertainty, is retained.
We let the continuously compounded riskless spot rate of interest per year be
denoted r. Hence a default-free zero-coupon bond paying a dollar at time T has time
t value
B (t, T) = e−r(T−t)
In this firm, the time T value of the firm’s debt and equity is completely determined
by the debt’s covenants. At the debt’s maturity, the equity holders decide whether to
pay off the debt. They will do so if and only if the value of the firm’s assets exceeds
the debt’s face value. Otherwise, default occurs. Hence the time T value of the equity
can be written as
ET = max [AT − K, 0] (26.5)
We see that the equity’s value is equivalent to the payoff from a European call option
on the firm’s value, with maturity date T and strike price K.
In essence, the debt holders, and not the equity holders, own the firm. The equity
holders have retained the option to buy the firm back from the debt holders at time
T for K dollars. They will do so at time T if the value of the assets exceeds this
purchase price. Hence we can write the value of the equity at time t to be equal to
Et = c (At , K, T) (26.6)
THE FOUR RISKS 671
where c(a, b, T) represents the value of a European call option at time t on an asset
with current price a, strike price b, and maturity date T.
Using the accounting identity relating the value of the firm’s equity to its assets
and liabilities, we see that the debt’s time T value is equal to the value of the assets
less the value of the equity, that is,
LT = AT − ET = AT − max [AT − K, 0] (26.7)
Simple algebra shows that we can rewrite the debt’s time T value as
LT = min [K, AT ] = K − max [K − AT , 0] (26.8)
The debt’s value at maturity is seen to be equal to the smaller of the face value of the
debt or the firm’s asset value. Alternatively, the debt’s time T payoff is equal to the
face value of the debt less the payoff to a European put option on the firm’s value,
with maturity date T and strike price K. The time t value is therefore
Lt = Ke−r(T−t) − p (At , K, T) (26.9)
where p(a, b, T) represents the value of a European put option at time t on an
asset with current price a, strike price b, and maturity date T. This shows that the
firm’s risky debt is valued less than a similar maturity default-free zero-coupon bond
(Ke−r(T−t) ). The decrease in value is equal to the put option’s value in expression
(26.9).
Using the lognormal distribution assumption, we can give an explicit formula for
the debt’s time t value:
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Lt = At N (−d1 ) + Ke−r(T−t) N (d2 ) (26.10)
where N( . ) is the cumulative standard normal distribution function which has mean
0 and standard deviation 1,
At 𝜎2
log + r+ (T − t)
(k) ( 2)
d1 =
𝜎√T − t
d2 = d1 − 𝜎√T − t, and 𝜎 is the asset return’s volatility per year.
This expression has a nice economic interpretation. The debt’s value is equal to
the firm’s value multiplied by N(−d1 ) plus the risk-adjusted probability of no default
times the discounted face value of the debt. The two terms represent the possible
payoff scenarios for the firm’s debt, depending on if the firm defaults (the first term)
or does not (the second term).
It is easy to show that as the firm’s asset volatility increases, the value of the firm’s
debt decreases. Since the firm’s asset volatility 𝜎 is under the control of management,
the decreasing value of the debt as 𝜎 increases leads to the shareholder/debt holder conflict
672 CHAPTER 26: RISK MANAGEMENT MODELS
(see Ross et al. 2004, p. 640). Understanding the shareholder/debt holder conflict is
just one use of the structural model. Another is that it can be used to provide insights
into the firm’s investment/capital budgeting decisions. This is the theory of real options
(see Extension 26.2).
The structural model can be extended to consider more complex liabilities issued
by the firm. However, this extension becomes quite difficult to solve analytically
because the entire liability structure of the firm needs to be specified and all of the
liability values solved simultaneously. In this regard, numerical procedures need to be
employed (for some of these extensions, see Jones et al. 1984; Leland 1994).
EXTENSION 26.2: Real Options
Option pricing theory can be used to understand how a firm should make its investment decisions. In the past,
this field was called capital budgeting. Today it is called real options, the title of this extension. This is because
embedded in most of the firm’s investment decisions are options, which were often ignored in classical capital
budgeting.
Real options theory usually proceeds under the same assumptions as the structural model. The model uses
assumptions A1–A8 from the BSM model, with the same two modifications: first, assumption A2, no credit risk,
is removed; second, the firm’s asset value replaces the stock price in assumptions A7 and A8. In particular, it is
assumed that a particular subset of the firm’s assets trades continuously in time and that this asset’s value process
follows a lognormal distribution. Assumption A6, no interest rate uncertainty, is retained. We let the riskless spot
rate of interest per year be denoted r. Of course, just as in the structural model, many of these assumptions can
be relaxed, as a result increasing analytical and computational complexity.
We explain real options theory using an example. Let us consider a gold mining company that is thinking
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
about purchasing a new gold mine. The purchase price for the mine is P0 dollars. The management has studied
the cost of extracting the gold and believes it can extract one production unit (one hundred tons) for a cost of
U0 dollars. Gold trades, and it has a market price of At dollars per unit at time t. It takes T time units to extract
the gold from the mine.
Last, the management also realizes that after the initial unit of gold is extracted from the mine at time T, it can
invest an additional K dollars to extract another k units of gold, where 0< k < 1. For simplicity (although this is
not necessary for the analysis), we assume that the additional k units each with a present value of AT at time T.
This extra opportunity to invest at time T is an embedded real option that the management wants to include
in its decision. This embedded option is a European call option on gold with a strike price of K and a maturity
date time T. Hence its payoff is
CT = max [kAT − K, 0] (1)
Under the no-arbitrage and complete market assumption, we know that we can use risk-neutral valuation to
determine the present value of the gold within the gold mine. The present value is the sum of the initial unit of
gold plus the embedded option, that is,
Present value = E𝜋 (AT ) e−rT − U0 + E𝜋 (max [kAT − K, 0]) e−rT (2)
THE FOUR RISKS 673
Using the fact the price of gold is a martingale under the risk-neutral probabilities, and using some simple algebra
on the value of the embedded call option, we can rewrite this as
Present value = A0 − U0 + kc0 (A0 , K/k, T) (3)
where c0 (a, b, T) is the time 0 value of a European call option on an asset with price a, strike price b, and maturity
date T.
Using the lognormal assumption for the evolution of the gold’s price, we can determine the value of the
project to be
Present value = A0 − U0 + kA0 N (d1 ) − Ke−rT N (d2 ) (4)
where N( . ) is the cumulative standard normal distribution function, which has mean 0 and standard deviation 1,
A0 𝜎2
log + r+ T
( K/k ) ( 2)
d1 =
𝜎√T
d2 = d1 − 𝜎√T, and 𝜎 is the price of gold’s return volatility per year. Given this present value, management
should purchase the gold mine if the present value exceeds the purchase price P0 dollars.
As illustrated by this example, real options theory applies derivatives pricing theory to any and all options
embedded within a firm’s investment decisions. A classic textbook on this topic is Dixit and Pindyck (1994).
Reduced-Form Models
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
The second class of credit risk models are called reduced-form models. To understand
the intuition behind the reduced-form model, it is best to start with Jarrow and Turn-
bull’s (1992, 1995) original formulation. The reduced-form model is an extension
of the HJM model of Chapters 23–25. The reduced-form model uses all of the
assumptions A1–A7 from the HJM libor model, with two modifications. One,
assumption A2, no credit risk, is removed because the purpose of the reduced-form
model is to price and hedge risky debt. The second modification is that the simple
(default free) forward rates need not follow a lognormal distribution. Other evolutions
are possible, and for the purposes of this section, we keep the evolution completely
arbitrary.
Note that unlike the structural model discussed in the previous section, reduced-
form models consider an economy with random interest rates. This is of particular
importance in the risk management of fixed-income securities where interest rate
and credit risk are paramount.
We let rt denote the random default-free spot rate of interest (continuously
compounded), and B(t, T) the time t value of a default-free zero-coupon bond paying
a dollar at time T. In addition to assumptions A1–A7, the reduced-form model needs
two more assumptions to characterize a firm’s risky debt. For easy comparison to the
structural model, we consider a firm that has as one of its liabilities a zero-coupon
bond with face value K and maturity date T. This assumption is easily relaxed in
674 CHAPTER 26: RISK MANAGEMENT MODELS
a reduced-form model, unlike the complexity that arises when relaxing this same
assumption in a structural model.
The first additional assumption is as follows.
A8. The firm’s zero-coupon bond trades continuously in time.
Note that unlike the structural model, the reduced-form model only assumes
that this particular debt issue trades. The firm’s other liabilities need not trade. This
overcomes the key limitation of the structural model discussed previously.
The second additional assumption characterizes the payoffs to the risky debt issue
if the firm defaults. This characterization, therefore, must include two components
(1) an assumption about when default occurs (i.e., specifying a random default time)
and (2) an assumption about the loss rate on the debt issue in the case of default.
A9. The firm’s default time 𝜏 is the first jump time of a Poisson process
with intensity 𝜆 > 0. In the event of default, the loss rate is (1 − 𝛿), where
0 < 𝛿 < 1 is the debt’s recovery rate.
The default intensity 𝜆 is the probability of default per unit time, given that the
firm has not yet defaulted. The firm’s default time is denoted by 𝜏. The loss rate is
given by the fraction (1 − 𝛿), where 𝛿 is the recovery rate. All of these parameters are
easily estimated using historical data, a topic we leave to more advanced courses.
The payoff to the debt can be written abstractly as
K1{𝜏 >T} + 𝛿K1{𝜏 ≤ T} (26.11)
where
𝜏>T
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
1 if
1(𝜏>T) =
{0 if 𝜏≤T
The symbol 1{𝜏 > T} is called an indicator function. It takes the value 1 if default does not
occur by time T and the value 0 otherwise. Using this indicator function, expression
(26.11) shows that the debt’s payoff is equal to K dollars if no default occurs prior
to the debt’s maturity and 𝛿K dollars otherwise. The payoff in default is only 𝛿 < 1
percent of the original face value of the debt K.
Let Lt denote the time t value of this debt issue. Under the no arbitrage assumption,
just as in the HJM model, it can be shown that there exists risk-neutral probabilities
𝜋 such that all default-free zero-coupon bonds and the firm’s debt are martingales,
after normalization by the value of a money market account. This implies that
T
B (t, T) = E𝜋t (e− ∫t r𝜎 ds
) (26.2a)
T 𝜏
Lt = E𝜋t (k1{𝜏>T} e− ∫t rs ds
+ 𝛿K1{𝜏≤T} e− ∫t rs ds
) (26.12b)
THE FOUR RISKS 675
where the subscript “t” indicates the expectation is taken at time t. Expression
(26.12a) is just the familiar risk-neutral valuation formula. The price of a default-
free zero-coupon bond is just the expected discounted value of a dollar paid at time
T, where one discounts using the spot rate of interest. Expectations are computed
using the risk-neutral probabilities.
Expression (26.12b) also uses risk-neutral valuation. This expression shows that
the debt’s value is equal to its expected discounted payoff across the two possibilities
regarding default: (1) the debt does not default before maturity, in which case it
receives its face value K, and (2) default occurs prior to maturity, in which case
the debt receives its recovery value. One discounts, as before, using the spot rate of
interest, and expectations are computed using the risk-neutral probabilities.
Using properties of the Poisson process and expression (26.12a), one can show that
(see Jarrow 2009)
T
̃ ̃
Lt = e−𝜆(T−t) KB (t, T) + ∫ 𝜆ẽ −𝜆(s−t) 𝛿KB (t, s) ds (26.13)
t
where 𝜆̃ is the risk-neutral default probability per unit time.
We see here that the risky zero-coupon bond is worth less than the default-
free zero-coupon bond KB (t,T). The first term represents the expected discounted
̃
value of the promised payment conditioned on no default. Here e−𝜆(T−t) is the risk-
neutral probability of no default. The second term represents the expected discounted
̃
value of the recovery value conditioned on default, where 𝜆ẽ −𝜆(s−t) is the risk-neutral
probability of default at time s. One sums across all times s where default is possible,
yielding the integral in the second term.
The risk-neutral probability of default 𝜆̃ per unit time differs from the actual
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
default probability 𝜆 by a risk premium. If default risk is idiosyncratic (firm specific),
however, then the risk-neutral and actual default intensities are equal (see Jarrow
2009). After conditioning on the state of the economy, it is likely that what causes
one firm to default versus another is idiosyncratic. It is an open research question as
to whether this is true.
Unlike the structural model, it is very easy to extend the reduced-form model to
have more complex default processes, to price other traded liabilities is- sued by the
firm, and to price credit derivatives. Extension 26.3 shows how to apply reduced-
form models to price credit default swaps that were introduced in Chapter 7.
676 CHAPTER 26: RISK MANAGEMENT MODELS
EXTENSION 26.3: Credit Default Swaps
Credit default swaps are term insurance contracts written on traded bonds with a maturity date, a notional
value, and a premium. If the insured bond defaults during the life of the contract, then the CDS seller pays off
the difference between the notional and market value of the bond. Of course, the market value declines due to
default. For this coverage, the CDS buyer pays a regular premium payment.
The remainder of this extension shows how one can use the reduced-form model to price CDS. We use the
same model as in the previous section and consider a CDS written on a bond issued by our firm with maturity
date T*, coupon payments, and a notional value of $1. We denote the time t price of this coupon bond by Lt .
We let 𝜏 be the firm’s random default time, and if the firm defaults prior to the coupon bond’s maturity date,
we assume that the bond incurs a loss of (1 - 𝛿) dollars, where 𝛿 is the recovery rate.
We can now value a CDS on this coupon bond. We let the CDS have a maturity date T < T* and a notional
value equal to $1. In this CDS, the protection seller agrees to pay the protection buyer the loss on the debt if the
firm defaults before the maturity date of the CDS. In return, the protection buyer pays a constant dollar spread c
times the notional value at the fixed intermediate dates t, t + 1, … , T - 1. These payments continue until the
swap’s maturity or the default date, whichever happens first.
Using the risk-neutral valuation procedure, the time t value of the CDS to the protection seller is therefore
T−1
T T
CDSt = E𝜋t c1{𝜏>k} e− ∫t rs ds
− (1 − 𝛿) 1{𝜏≤T} e− ∫t rs ds
(1)
(∑
k =t )
This is the present value of the premiums received (the summation term) less the losses paid if default occurs. It
can be shown that this can be simplified to
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
T−1 T
̃ ̃
CDSt = c e−𝜆(k−t)
B (t, k) − (1 − 𝛿) 𝜆̃ ∫ e−𝜆s (s − t) B (t, s) ds (2)
∑
k =t t
where 𝜆̃ is the risk-neutral default probability per unit time.
The market clearing CDS rate c is that spread that gives the CDS zero value, that is,
̃
∫t T e−𝜆s (s − t) B (t, s) ds
c = (1 − 𝛿) 𝜆̃ ≈ (1 − 𝛿) 𝜆̃ (3)
T−1
̃
e−𝜆(k−t) B (t, k)
∑
k =t
We see that the CDS rate is approximately equal to the risk-neutral expected loss (per year)—the risk-neutral
default probability times the loss rate—on the firm’s debt. The CDS price takes into account risk aversion through
the use of the risk-neutral probabilities.
CDS trade on many firms’ debt issues, and the market uses CDS spreads and expression (3) to infer the market’s
view of the risk of default as reflected in the implied risk-neutral default probability (for a given recovery rate).
The reduced-form methodology is also easily employed to understand and price even more complex credit
derivatives like collateralized debt obligations (see Jarrow 2009).
THE FOUR RISKS 677
Liquidity Risk
Liquidity risk corresponds to the relaxation of the competitive markets assumption.
Recall that the competitive markets assumption implies that all traders act as price
takers, that is, their trades have no impact on the buying or selling price. Liquidity
risk occurs when there is a quantity impact of a trade on the price obtained.
Quantity impact occurs in actual markets due to market power (size) or asymmetric
information (which means the counterparties have differential information, e.g.,
a seller is likely to know more about a used car’s quality than the buyer). Market
power is easy to understand. Buying a large quantity of any security changes the price
because the buyer needs to induce the seller to sell. Asymmetric information causes a
quantity impact because if one believes that the seller is informed and selling because
the security is worth less than its current price, clearly the selling price must decline.
In conjunction, these two reasons imply that selling (buying) large quantities of a
security usually decreases (increases) the price obtained. When markets are volatile
and disrupted, the quantity impact is larger than when markets are quiet and stable.
The failure of Long-Term Capital Management (LTCM) in 1998 is perhaps the
most infamous example of the importance of liquidity risk. LTCM was a hedged
fund started by John Meriwether, ex–Salomon Brothers bond traders, and several
finance academics in 1994. Using convergence arbitrage trading strategies in bond
markets, they had phenomenal profit performance for the first four years of their life,
providing returns in excess of 20 percent. A characteristic of their bond arbitrage
trading strategies was that they required large leverage to transform small differences
in bond spreads into large returns for their clients. This, of course, implies a small
capital position.
LTCM’s performance turned bad in May and June 1998, when mortgage-backed
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
securities lost significant value. Then, Russia’s defaulting on its debt in August 1998,
and the resulting market reactions, caused significant losses in LTCM’s asset value.
This loss in asset value resulted in LTCM receiving margin calls for its futures positions
and collateral calls for its swap positions. To obtain the additional capital, LTCM had
to sell its assets. Selling assets when the market is spooked and declining is problematic.
These asset sales resulted in increased value losses for LTCM as the process of selling
large quantities of securities caused the market to drop further. A cascade of additional
margin and collateral calls and market drops resulted, eventually leading to the failure
of LTCM and the bailout of LTCM organized by the New York Federal Reserve
Bank to avoid a financial market collapse (for a detailed history, see Jorion 2000).
There are two approaches to modeling liquidity risk, which differ based on the
permanence of the quantity impact of a trade on the price. If the quantity impact on
the price is temporary, then liquidity risk is analogous to an endogenous transaction
cost. By temporary, we mean that the quantity impact on the price only lasts for the
trade under consideration, and after the trade, the price reverts back to its proper
value. Çetin et al. (2004) were the first to study this approach in the context of
derivatives pricing. Fortunately, the derivatives pricing methodology extends under
this form of liquidity risk. Hence all of the results previously studied in Chapters
1–25 apply directly, with few, if any, changes.
678 CHAPTER 26: RISK MANAGEMENT MODELS
If the quantity impact on the price is permanent, however, then liquidity risk takes
on a different characterization. In this case, traders can act strategically and manipulate
the price to their advantage. Here liquidity risk needs to include the consideration
of market manipulation. Jarrow (1992, 1994) was to first to study manipulation in
the context of derivatives pricing. Unfortunately, the standard derivatives pricing
methodology fails to hold under this form of liquidity risk. Pricing and hedging
become context specific, depending on the traders, their strategies, and the market
setting. Here all of the results previously studied in Chapters 1–25 fail to apply, and
new models are needed. The extensions in this case are the subject of current research.
Operational Risk
Operational (and legal) risk is the risk inherent in running a financial institution.
These risks usually cause only small losses, but sometimes they have resulted in large
blowouts. A stunning example of operational risk happened on December 8, 2005,
when a broker at Japan’s Mizuho Securities Company transposed the quantity and
the selling price in the company’s order processing system. He mistakenly typed an
order to sell 610,000 shares of a small company J-Com at 1 yen instead of the other
way round. The error was detected only a minute and a half later, but the Tokyo
Stock Exchange’s all-electronic trading system thwarted all cancellation efforts. The
result was a loss of $335 million.6 In February 7, 1991, traders at now-defunct Wall
Street firm Salomon Brothers submitted an unauthorized customer bid in the amount
of $1 billion as what Salomon termed a “practical joke” to tease an employee, but
they failed to stop those bids on time, and an extra $1 billion worth of bonds got
purchased. Eventually, the acquired securities were transferred to Salomon’s account
at no loss to the firm (see Chatterjea 1993).
Good management controls and procedures can minimize operational risk. Con-
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
sequently, the control and minimization of this risk really falls under the heading of
business management or organizational behavior and not finance. Nonetheless, for
pricing and hedging this risk, finance is again useful.
Operational risk can be modeled using the same mathematics as used in the
reduced-form credit risk models. Indeed, operational risk can be viewed as a negative
event, which if it occurs, generates a loss to the firm. This is exactly how the reduced-
form credit risk models view default. Consequently, the modeling techniques used
in credit risk can be applied to operational risk (see Jarrow 2008; Jarrow et al. 2010a).
The only substantive difference between these two models is that there is more
publicly available data on credit risk events (realizations and losses) then there are on
operational risk events. However, these differences are narrowing as various industry
groups are creating operational risk databases.
6
See “Mizuho Settles Trade Debacle; Japanese Minister Scolds Brokers for Taking Advantage of Error,”
Wall Street Journal, December 14, 2005.
THE FUTURE OF MODELS AND TRADED DERIVATIVES 679
26.6 The Future of Models and Traded
Derivatives
The credit crisis of 2007 and its resulting aftermath created a discomfort with
risk management models and derivatives. This discomfort raises the question as to
the future of risk management models and derivatives. This section addresses this
question, in light of what we have now studied in this book. The following discussion
is based on Jarrow (2011).
Model Risk
When using financial risk management models, one must recognize that there is
model risk. Model risk occurs because models are approximations—simplifications—
of a complex reality. Statisticians Box and Draper (1987, p. 74) said, “Remember
that all models are wrong: the practical question is how wrong do they have to be to
not be useful.”
A model’s usefulness—the quality of the approximation—needs to be judged
relative to a purpose. Very crude models can be useful for decision making in some
contexts, whereas more sophisticated models are needed in others. The usefulness of
the model should be tested and continually validated to assure continued accuracy of
the approximation.
To understand how to test a model, we first need to divide a model’s assumptions
into two types: robust and critical assumptions. A robust assumption is one in which
the implications of the model only change slightly if the assumption is modified only
slightly. In contrast, a critical assumption is one in which the implications of the
model change discretely if the assumption is only changed slightly.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
This distinction is important because since models are approximations of a complex
reality, we may not get the assumptions exactly correct. With robust assumptions, we
do not need to worry too much. For small errors in the robust assumptions, the
implications only change by a small amount. We need to be careful, however, with
the critical assumptions. If we get the critical assumptions wrong, by just a little, the
implications completely change.
Now to test a model, one needs to test all the model’s relevant implications and
all the assumptions, if possible. Sometimes, however, not all the assumptions can be
tested. In this case, one must test all the critical assumptions. If any of the implications
or critical assumptions are rejected, the model should not be used. Why? Because
using an incorrect model generates poor risk management decisions.
If a model is accepted for use, the implications and critical assumptions need to
be continually monitored and validated because markets are dynamic and constantly
changing. It is always possible that if the market’s structure shifts because of regulatory
or institutional changes, the underlying assumptions may no longer be reasonable
approximations. If this happens, then pricing and hedging will no longer work.
The best way to understand models and their use is to consider an analogy. Models
are analogous to medical prescription drugs. Prescription drugs have great medical
benefits if used properly, with educated use. If used wrongly, however, prescription
drugs can have negative consequences, even death. Because prescription drugs can
680 CHAPTER 26: RISK MANAGEMENT MODELS
cause death if used incorrectly, this does not mean that we should stop using them. It
does mean, however, that we need educated and regulated use. In fact, prescription
drugs should probably be used more because they save and prolong lives. The same
is true of models.
Financial markets have become too complex to navigate without risk management
models. Determining a price—fair value—is sometimes not an issue because in many
cases, expert judgment may provide reasonable estimates. However, the following
are true:
■
■ There is no way to hedge a portfolio, that is, determine hedge ratios, without a
model.
■ There is no way to price a derivative in an illiquid market without a model.
These issues are at the heart of risk management. Hence financial risk management
models are here to stay.
Derivatives
Derivatives have existed for thousands of years. They have continued to exist (trade)
across different centuries, across different governments, and across different economic
systems. Despite occasional uproars about their evil uses, derivatives have continued to
trade because they help complete markets and, consequently, improve societal welfare.
If history shows us nothing else, it shows us that derivatives are here to stay!
Given this insight, the best way to live with derivatives is through education and
knowledge. We have started you on this journey to understanding with the content
of this book, but this is just the beginning! There is much more to learn and master.
We have just walked with you on only the first part of your journey to understanding
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
derivatives, and it is now time for us to leave. However, we leave you knowing that
you are well prepared for the obstacles ahead.
26.7 Summary
1. A firm’s risk management decisions can be understood using the firm’s balance
sheet and computing the firm’s loss distribution defined as Prob{ΔAt − ΔLt ≤ x},
where the time t asset’s value is At , the time t liability is Lt , and ΔAt ≡ At+Δt − At
and ΔLt ≡ Lt+Δt − Lt correspond to the change in the value of the assets and
liabilities, respectively, over the interval [t, t + Δt].
2. Value-at-risk (VaR𝛼 ) for an 𝛼 loss probability is that number such that Prob{ΔAt −
ΔLt ≤ −VaR𝛼 } = 𝛼. In some cases, VaR𝛼 penalizes diversification, which is a
failing in this risk management statistic.
3. Scenario analysis (which also goes by the popular name stress testing) starts with
a selection of a particular set of scenarios (or states of the economy) that the firm
wants to protect against. For these scenarios, the firm wants to compute the losses
realized on its assets and liabilities to see if it has sufficient equity capital to avoid
QUESTIONS AND PROBLEMS 681
insolvency. Scenario analysis’s major failing is that it does not assign probabilities
to the various scenarios.
4. Credit risk is the risk that a counterparty will fail to execute on the terms of a con-
tract. There are two models for quantifying credit risk: structural and reduced-
form models. The structural model is useful for internal management decision
making and understanding the economics of the firm’s balance sheet, whereas the
reduced-form model is useful for pricing and hedging credit-risky instruments
trading in the financial markets.
5. Both liquidity and operational risk can be modeled using the techniques studied
in this book. Operational risk uses the same models as credit risk.
6. Models are needed for risk management in complex financial markets. To avoid
model misuse, better education on models is needed.
26.8 Cases
Amaranth Advisors: Burning Six Billion in Thirty Days (Richard Ivey School of
Business Foundation Case 908N03-PDF-ENG, Harvard Business Publishing).
This case studies the staggering losses encountered by a hedge fund and introduces
concepts like liquidity risk, value-at-risk, spread trades, and the use of derivatives.
Delphi Corp. and the Credit Derivatives Market (A) (Harvard Business School
Case 210002-PDF-ENG). This case considers the use of credit derivatives to hedge
credit risk and/or to speculate on the price of corporate debt and also how credit
derivatives affect the incentives of creditors to negotiate with a distressed company
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
attempting to avoid bankruptcy.
CITIC Tower II: The Real Option (University of Hong Kong Case HKU199-PDF-
ENG, Harvard Business Publishing). This case studies real options involved in
the decision to purchase newly reclaimed land in Hong Kong for the purpose of
building an office building at an appropriate time in the future.
26.9 Questions and Problems
26.1. When using value at risk at the 5 percent level to determine equity capital,
do the debt holders need to worry about the magnitude of the losses at the 1
percent level? Explain.
26.2. Consider a firm whose business is to purchase distressed debt, that is, bonds
of firms that have defaulted, hoping to profit from the bankruptcy resolution
process. Which of the four risks (market, credit, liquidity, and operational)
does this distressed debt investment firm consider when evaluating the
distressed debt? Explain.
682 CHAPTER 26: RISK MANAGEMENT MODELS
26.3. Suppose you enter into a $5 bet with a friend on the winner of a soccer game.
What risks do you face in this gamble?
26.4. In the structural credit risk model, it is claimed that the debt holders own the
firm, and not the equity holders. Explain why.
26.5. Since the equity holder’s position is equivalent to a call value on the firm’s
assets, and the equity holders manage the firm’s assets, can they increase the
value of their equity by increasing the firm’s asset volatility? Explain your
answer.
26.6. What are the three risks faced by a bond in the reduced-form credit risk
model?
26.7. Which risks caused the hedge fund Long-Term Capital Management to fail?
26.8. Explain why operational risk and credit risk can be modeled using similar
techniques.
26.9. What assumption makes the structural model not useful for actual pricing and
hedging but makes the reduced-form model appropriate for these activities?
26.10. Your grandmother gives you an old silver flower vase as a birthday gift. You
want to sell it to buy a new computer. What risks do you face with respect
to this asset?
26.11. Internet markets—like eBay—decrease which risk of holding assets?
26.12. Why are models analogous to prescription medical drugs? Are derivatives
likely to discontinue trading? Explain.
The next two questions are based on the following data for three companies.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Expected Return, Volatility, and Portfolio Weights
Ali Co. Barb Co. Cao Co.
(Company 1) (Company 2) (Company 3)
Weekly expected 0.0020 0.0025 0.0030
return
Weekly volatility 0.03 0.04 0.05
of the returns
Portfolio weights 1/3 1/3 1/3
QUESTIONS AND PROBLEMS 683
Correlation of Expected Returns
Ali Co. (Company 1) Barb Co. (Company 2) Cao Co. (Company 3)
Ali Co. 1 0.4 0.8
Barb Co. 0.4 1 0.5
Cao Co. 0.8 0.5 1
26.13. Compute the weekly 95 percent and 99 percent value-at-risk for Ali Co.
26.14. Compute the weekly 95 percent and 99 percent value-at-risk for a portfolio
consisting of the three stocks in equal proportions.
26.15. Now that you have completed the book and are on your way to mastering
derivatives, briefly discuss your views as to whether derivatives are good or
bad? Feel free to use the internet to get some quotes to support your answer.
Copyright © 2018. World Scientific Publishing Company Pte. Limited. All rights reserved.
Appendix A
Mathematics and Statistics
A.1 Exponents and Logarithms A.3 The Normal Distribution
Exponents A.4 A Taylor Series Expansion
Logarithms Single Variables
Continuous Compounding Multiple Variables
A.2 The Binomial Theorem
EXPONENTS AND LOGARITHMS 685
A.1 Exponents and Logarithms
Exponents and logarithms are useful for valuing cash flows across time and for
computing continuously compounded returns.
Exponents
Arithmetic classes taught us that 2 times 2 is 4, and algebra classes showed us that
we can write this as 22 (read “2 raised to the power of 2”). Here 2 is an exponent
denoting the number of times 2 multiplies itself.
We can generalize this example and write “x multiplied by x” as x2 . We can
generalize further to denote “n products of x” as
xn ≡ x × x … n times (1.1)
The ≡ sign denotes that “this is true by definition.”
An exponent may also be a fraction that denotes a root or a radical. For example,
x1/2 means x1/2 × x1/2 = x; we also write x1/2 as √x and read it as “square root of x ”.
When generalized, x1/n is the “nth root of x,” and it is also expressed as √n
x, which
means
x1/n × x1/n … n times ≡ x (1.2)
Negative exponents are used to create an inverse. For example, 2−1 is 1/21 = 1/2.
This is generalized to define
x−n ≡ 1/(x × x … n times) ≡ 1/xn (1.3)
Of course, the preceding expression is only defined when x is not equal to zero
(in notations, x ≠ 0); moreover, x0 = 1, when x ≠ 0. These puzzling results can be
explained by the law of exponents. You can write x raised to the power 0 as
x0 = xm−m = xm × x−m = xm /xm = 1
This holds true when x does not equal zero. When x = 0 and we raise it to the power
0, then we have zero in the denominator, and we are essentially dividing by 0, which
is forbidden by mathematicians because it gives nonsensical results.
Next we write some useful properties of exponents (see Table A.1). We state the
laws and give an example illustrating their use.
Logarithms
We only consider natural logarithms in this book (see Table A.2). The natural
logarithm, written as log(x), is defined by
log (x) = y is equivalent to ey = x (1.4)
686 APPENDIX A: MATHEMATICS AND STATISTICS
TABLE A.1: Laws of Exponents
Laws of Exponents Examples
1. xm xn = xm+n x2 x3 = x5
1 1
2. x−m = m x−2 = 2
x x
xm x5
3. xm−n = n x5−3 = 3 = x2
x x
1 x5
4. x0 = xm−m = xm = 1, when x ≠ 0 0 5−5
x =x = 5 =1
xm x
5. (xm )n = xmn (x2 )3 = (x2 ) (x2 ) (x2 ) = x6
6. xm ym = (xy)m and xm /ym = (x/y)m x5 y5 = (xy)5 and x4 /y4 = (x/y)4
TABLE A.2: Laws of Logarithms
Laws of Exponents Examples
1. log(xy) = log x + log y log 50 = log 5 + log 10
2. log(x/y) = log x – log y log(10/2) = log 10 − log 2
= log 5 + log 2 − log 2 = log 5
3. log(x)n = n log x log 23 = log 8 = log(2 × 2 × 2)
= 3 log 2 and log 21/5 = (1/5) log 2
4. log 1 = 0 log 1 = 0 because e0 = 1
where e is the Euler number, approximately 2.711828, defined as the limit
1 n
e = lim (1 + ) (1.5)
n→∞ n
when n goes to infinity or as the infinite series
x x2 x3 xn
ex = 1 + + + +⋯+ +⋯ (1.6)
1 1×2 1×2×3 n!
for any real number x. Notice that a representative term in expression (1.6) is given
by the (n + 1)th term xn /n!, where n! is read as n-factorial and is defined by
n! = n × (n − 1) × … × 3 × 2 × 1 and 0! = 1.
THE BINOMIAL THEOREM 687
Continuous Compounding
If X dollars is invested today at an interest rate r that is compounded m times per year,
and m becomes infinitely large, then it has the time T value
X[(1 + r/m)m ]T = X[(1 + r/m)m/r ]rT → XerT (1.7)
because (1 + r/m)(m/r) ≡ [1 + 1/(m/r)](m/r) = (1 + 1/n)n goes to e as n = m/r goes to
infinity.
This can be used to model dividend yields. Consider buying one unit of a stock
that pays dividends at a continuous rate of 𝛿 per year. Assume that the dividends are
reinvested back into the asset (see section 12.3).
To model this, we start by dividing the year into m periods of equal length and
assume that a dividend of 𝛿/m times the stock value is paid at the end of each interval.
At each payment date, we use the dividend to buy more of the stock. Consequently,
the value of the position in the stock grows just like compound interest. Here, 𝛿/m
is compounded m times per year, where m becomes infinitely large. Then, our one
unit in the stock grows to
[(1 + 𝛿/m)m ]T ≡ [(1 + 𝛿/m)(m/𝛿) ]𝛿T = e𝛿T units after T years (1.8)
This happens because (1 + 𝛿/m)(m/𝛿) ≡ [1 + 1/(m/𝛿)](m/𝛿) = (1 + 1/n)n goes to e as
n = m/𝛿 goes to infinity.
A.2 The Binomial Theorem
One of the first facts you learn in algebra is how to square the sum of two variables
x and y:
(x + y)2 = x2 + y2 + 2xy
The expression (x + y) is called a binomial because it is the sum of two terms. Next
comes the cubic expansion of a binomial:
(x + y)3 = x3 + 3x2 y + 3xy2 + y3
One can expand this binomial for an arbitrary integer n. The binomial theorem
provides a method for doing this:
n n 0 n n−1 1 n n−j j n
(x + y)n = xy + x y ⋯+ x y ⋯+ x1 y(n−1)
(0) (1) (j) (n − 1)
n (2.1)
n 0 n n n−j j
+ xy = x y
(n) ∑ (j)
j=0
688 APPENDIX A: MATHEMATICS AND STATISTICS
where the binomial coefficient (which is the number of ways of selecting j things
out of a possible number of n things) can be expressed as
n n!
= (2.2)
( j ) j! (n − j) !
Here the expression n! is called n-factorial and is defined as
n! = n × (n − 1) × (n − 2) × … × 2 × 1 (2.3)
For example, 4! = 4 × 3 × 2 × 1 = 24 and 1! = 1, and by definition, 0! = 1. Moreover,
n! = n(n − 1)! = n(n − 1)(n − 2)!, and so on.
A.3 The Normal Distribution
A standard normal distribution function has a mean equal to 0 and a standard deviation
equal to 1. The normal distribution function is denoted by N(d). This gives the
probability that the underlying normal random variable z will take a value less than
or equal to d, that is, N(d) = Prob(z ≤ d).
This probability can be obtained from a normal distribution table (see Table A.3)
or using computer software (e. g., spreadsheets). Usually, normal distribution tables
give probabilities for z greater than 0.
The normal density represents the mathematical derivative of the normal
distribution function. The normal density function is symmetric around mean 0
(see Figure A.1) and is shaped like a bell. Because the normal density function is
symmetric around 0, if you want to compute N(−d) for some negative value, use
N(d) + N(−d) = 1.
EXAMPLE A.1: Cumulative Standard Normal Distribution
■ From the cumulative standard normal distribution table (Table A.3),
N (0.75) = 0.7734
This is the probability that a normal random variable will have a value of 0.75 or less.
■ Likewise,
N (0.55) = 0.7088
■ Or you can use spreadsheet programs like Excel or Lotus Notes to compute cumulative normal.
Excel will use the command = NORMSDIST(z), where z is the underlying normal random variable
(in Lotus Notes, use the command @NORMAL(z,0,1,0)).
THE NORMAL DISTRIBUTION 689
FIGURE A.1: Normal Density Function
Probability density
N(d)
–∞ 0 d ∞ Value
TABLE A.3: Cumulative Normal Distribution Function
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0 0.5 0.504 0.508 0.512 0.516 0.5199 0.5239 0.5279 0.5319 0.5359
0.1 0.5398 0.5438 0.5478 0.5517 0.5557 0.5596 0.5636 0.5675 0.5714 0.5753
0.2 0.5793 0.5832 0.5871 0.591 0.5948 0.5987 0.6026 0.6064 0.6103 0.6141
0.3 0.6179 0.6217 0.6255 0.6293 0.6331 0.6368 0.6406 0.6443 0.648 0.6517
0.4 0.6554 0.6591 0.6628 0.6664 0.67 0.6736 0.6772 0.6808 0.6844 0.6879
0.5 0.6915 0.695 0.6985 0.7019 0.7054 0.7088 0.7123 0.7157 0.719 0.7224
0.6 0.7257 0.7291 0.7324 0.7357 0.7389 0.7422 0.7454 0.7486 0.7517 0.7549
0.7 0.758 0.7611 0.7642 0.7673 0.7704 0.7734 0.7764 0.7794 0.7823 0.7852
0.8 0.7881 0.791 0.7939 0.7967 0.7995 0.8023 0.8051 0.8078 0.8106 0.8133
0.9 0.8159 0.8186 0.8212 0.8238 0.8264 0.8289 0.8315 0.834 0.8365 0.8389
1 0.8413 0.8438 0.8461 0.8485 0.8508 0.8531 0.8554 0.8577 0.8599 0.8621
1.1 0.8643 0.8665 0.8686 0.8708 0.8729 0.8749 0.877 0.879 0.881 0.883
1.2 0.8849 0.8869 0.8888 0.8907 0.8925 0.8944 0.8962 0.898 0.8997 0.9015
1.3 0.9032 0.9049 0.9066 0.9082 0.9099 0.9115 0.9131 0.9147 0.9162 0.9177
1.4 0.9192 0.9207 0.9222 0.9236 0.9251 0.9265 0.9279 0.9292 0.9306 0.9319
1.5 0.9332 0.9345 0.9357 0.937 0.9382 0.9394 0.9406 0.9418 0.9429 0.9441
(Continued )
690 APPENDIX A: MATHEMATICS AND STATISTICS
TABLE A.3: (Continued )
1.6 0.9452 0.9463 0.9474 0.9484 0.9495 0.9505 0.9515 0.9525 0.9535 0.9545
1.7 0.9554 0.9564 0.9573 0.9582 0.9591 0.9599 0.9608 0.9616 0.9625 0.9633
1.8 0.9641 0.9649 0.9656 0.9664 0.9671 0.9678 0.9686 0.9693 0.9699 0.9706
1.9 0.9713 0.9719 0.9726 0.9732 0.9738 0.9744 0.975 0.9756 0.9761 0.9767
2 0.9772 0.9778 0.9783 0.9788 0.9793 0.9798 0.9803 0.9808 0.9812 0.9817
2.1 0.9821 0.9826 0.983 0.9834 0.9838 0.9842 0.9846 0.985 0.9854 0.9857
2.2 0.9861 0.9864 0.9868 0.9871 0.9875 0.9878 0.9881 0.9884 0.9887 0.989
2.3 0.9893 0.9896 0.9898 0.9901 0.9904 0.9906 0.9909 0.9911 0.9913 0.9916
2.4 0.9918 0.992 0.9922 0.9925 0.9927 0.9929 0.9931 0.9932 0.9934 0.9936
2.5 0.9938 0.994 0.9941 0.9943 0.9945 0.9946 0.9948 0.9949 0.9951 0.9952
2.6 0.9953 0.9955 0.9956 0.9957 0.9959 0.996 0.9961 0.9962 0.9963 0.9964
2.7 0.9965 0.9966 0.9967 0.9968 0.9969 0.997 0.9971 0.9972 0.9973 0.9974
2.8 0.9974 0.9975 0.9976 0.9977 0.9977 0.9978 0.9979 0.9979 0.998 0.9981
2.9 0.9981 0.9982 0.9982 0.9983 0.9984 0.9984 0.9985 0.9985 0.9986 0.9986
3 0.9987 0.9987 0.9987 0.9988 0.9988 0.9989 0.9989 0.9989 0.999 0.999
3.1 0.999 0.9991 0.9991 0.9991 0.9992 0.9992 0.9992 0.9992 0.9993 0.9993
3.2 0.9993 0.9993 0.9994 0.9994 0.9994 0.9994 0.9994 0.9995 0.9995 0.9995
3.3 0.9995 0.9995 0.9995 0.9996 0.9996 0.9996 0.9996 0.9996 0.9996 0.9997
3.4 0.9997 0.9997 0.9997 0.9997 0.9997 0.9997 0.9997 0.9997 0.9997 0.9998
3.5 0.9998 0.9998 0.9998 0.9998 0.9998 0.9998 0.9998 0.9998 0.9998 0.9998
3.6 0.9998 0.9998 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999
3.7 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999
3.8 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999 0.9999
3.9 1 1 1 1 1 1 1 1 1 1
4 1 1 1 1 1 1 1 1 1 1
A TAYLOR SERIES EXPANSION 691
A.4 A Taylor Series Expansion
Formally introduced by the English mathematician Brook Taylor in 1715, a Taylor
series allows us to express special functions as an infinite sum of terms involving
polynomials.
Single Variables
The Taylor series expansion of a single variable function f (x) that is infinitely
differentiable in a neighborhood of a number a is given by the power series
f 1 (a) f 2 (a) f (n) (a)
f (x) = f (a) + (x − a) + (x − a)2 + ⋯ (x − a)n + ⋯ (4.1)
1! 2! n!
where n! = n × (n − 1) × … × 2 × 1 denotes the factorial of n and f (n) (a) denotes
the nth derivative of f evaluated at the point a. Notice that f 0 is defined to be f itself
and (x − a)0 and 0! are both defined to be 1; f 1 and f 2 are customarily written as f ′
and f ′′, respectively.
Consider the function graphed in Figure A.2. A Taylor series approximation of
this function at the point x = a using just the first two terms is given by
f (x) ≈ f (a) + f ′ (a) (x − a) (4.2)
Algebra tells us that the right side of this expression is a straight line tangent to the
function f (x) at x = a. Expression (4.2) shows that this straight line (shown by a dashed
line in the figure) is an approximation to the function f (x) at x = a.
FIGURE A.2: Taylor Series Approximation
y = f(x) y = f(a) + f´(a)(x – a)
f´(a)
f(a)
0
a x
692 APPENDIX A: MATHEMATICS AND STATISTICS
Such approximations work well for small changes in x around x = a. A better
approximation for larger changes in the variable x is obtained by retaining the second
derivative as given by
1
f (x) ≈ f (a) + f ′ (a) (x − a) + f ′′ (a) (x − a)2 (4.3)
2
Multiple Variables
The Taylor series for a well-behaved function of multiple variables can be expressed
as
∞ ∞
(x1 − a1 )n1 ⋯ (xd − ad )nd 𝜕n1 +⋯+nd f
f (x1 , … , xd ) = ⋯ (a1 , … , ad )
∑ ∑ n1 ! … nd ! n n
( 𝜕x11 ⋯ 𝜕xdd )
n1 =0 nd =0
(4.4)
To illustrate this general result, consider a function f of two variables x and y. The
Taylor series expansion to the second order about the point (a, b), writing (x − a) as
Δx, and (y − b) as Δy, is
𝜕f (a, b) 𝜕f (a, b)
f (x, y) ≈ f (a, b) + Δx + Δy
𝜕x 𝜕y
(4.5a)
2 2
1 𝜕 f (a, b) 2 𝜕 f (a, b) 2 𝜕2 f (a, b)
+ Δx + Δy + 2ΔxΔy
2! [ 𝜕x2 𝜕y2 𝜕x𝜕y ]
Writing f (x, y) − f (a, b) as Δf, and moving the first term on the right side of this
expression to the left side gives the simpler expression
𝜕f 𝜕f 2
1 𝜕f 2 𝜕f 2
2
𝜕2 f
Δf ≈ Δx + Δy + Δx + Δy + 2ΔxΔy (4.5b)
𝜕x 𝜕y 2! 𝜕x2 𝜕y2 𝜕x𝜕y
These results are used in Chapters 20 and 25.
Glossary
Algorithmic trading or algo Trading using an algorithm Basel Committee on Banking Supervision Established
(which is a set of rules or sequence of steps) to identify in 1974 by the Group of Ten countries’ central-bank
patterns in real-time market data and exploit potentially governors; formulates broad supervisory standards and
profitable trading opportunities. recommends statements of best banking practice.
American depository receipt A US exchange-traded Basis The spot price minus the futures price.
security that has a fixed number of foreign stock shares
Basis point (or bp) One-hundredth (1/100th) of
as the underlying.
1 percent.
American option An option that can be exercised anytime
Bearish trade A trade benefiting if security prices decrease
during its life.
in value.
Annual Percentage Rate (APR) The restatement of an
interest rate (e.g., continuously compounded) in terms of Bid price The price at which a dealer offers to buy
an annualized simple interest rate. securities from a customer.
Arbitrage A trading strategy that has a chance to make Binomial stock price evolution A model of a stock price
riskless profits without requiring a net investment of evolution using a tree in which from any point in time
funds. onward, the stock price takes only one of two possible
values.
Arbitrageur Trader who attempts to make arbitrage profits
from price discrepancies. Block trade Stock trade involving ten thousand shares or
Arbitration A system of mediating disputes before going more.
to the courts. Bond or fixed-income security The debt issued by a
Ask (or offer) price The price at which a dealer offers to borrower who promises both to pay interest on the
sell securities to a customer. borrowing and to repay the principal borrowed at some
future date (see consols for an exception).
Asset-backed security (ABS) Bond issued by a special-
ized purpose vehicle (SPV) that is backed by assets as Bond-equivalent yield The holding period return earned
collateral. on a bond if held until maturity.
At-the-money (option) An option whose underlying Bootstrap method The technique of creating a zero-
asset’s price is close to the exercise price. coupon bond price term structure from market data
(including computation of any missing prices by inter-
Audit trail A detailed information record of a financial
polation).
security transaction.
Broker A financial intermediary who matches buyers and
sellers and earns commissions for this service.
Backward induction A mathematical technique for solv-
ing a problem where one starts at the last date in a model Bullish trade A trade benefiting if security prices increase
and obtains the solution by working backward through in value.
time. Butterfly spread An option trading strategy created by
Backwardation A forward or futures market where the trading four options of the same type (all calls or all
forward or futures price is lower than the spot price, puts) with three different strike prices: two options with
respectively. extreme strike prices are purchased (written) and two
options are written (purchased) with middle strike prices.
Banging (or marking) the close Manipulating prices by
trading a large security position leading up to the market Buying hedge (or long hedge) Taking a long position
close followed by offsetting the position before the end of in a forward or a futures contract to hedge the price risk
trading. from a spot market purchase.
Banker’s discount yield The rate used for quoting Trea- Buy–write (option) Simultaneously buying the underly-
sury bills. ing asset and selling the call.
693
694 GLOSSARY
Calibration A mathematical technique that estimates a different strike prices) so that the resultant portfolio has
derivative pricing model’s parameters by equating the truncated payoffs on both the up and down sides. (Market
model’s price to the market price. regulation) A trading restriction that prohibits the buying
Call option A financial security that gives the owner the or selling of stocks when the stock price moves by more
right to buy a specified quantity of a financial or real asset than a certain number of points.
on or before a fixed future date by paying an exercise price Collateralized debt obligation (CDO) An asset-backed
agreed on when the contract is written. security whose collateral pool consists of ABS bonds, as
Callable bond A bond whose issuer has the option to distinct from mortgage loans. CDO bonds are partitioned
redeem the bond before maturity. into seniority tranches that specify the interest and prin-
cipal payments the bonds receive.
Cap A portfolio of caplets.
Collateralized debt obligation squared (CDOˆ2) An
Capital gain or loss (Gain) Realized by selling an asset at
asset-backed security whose collateral pool consists of
a price higher than the purchase price. (Loss) Realized by
CDO bonds.
selling an asset at a lower price than the buying price.
Combination strategy A portfolio of options of different
Caplet A European call option on an interest rate.
types on the same underlying asset with the same expi-
Cash settlement Ending a derivatives contract by making ration date where the options are either all purchased or
a cash payment instead of exchanging physical securities. written.
Cash-and-carry argument An argument used to deter- Commercial paper A promissory note, issued by a com-
mine the forward price where one purchases the under- pany, that matures in 270 days or less and that is sold as a
lying commodity with borrowed cash and holds the zero-coupon bond.
commodity until the forward contract’s maturity date.
Commercial trader (futures) Futures trader such as a
Central limit theorem (CLT) A key theorem in prob- farmer, manufacturer, or commodities dealer who has
ability that states that the distribution for the aver- legitimate hedging needs because of her commercial
age of a collection of independent random variables, activity with the underlying commodity.
when “standardized” (by subtracting from the average the
Commitment of Traders (COT) A weekly report pub-
common mean and dividing the expression by the stan-
lished by the Commodity Futures Trading Commission
dard deviation of the average), approaches a standard
that shows aggregate commercial and noncommercial
normal distribution as the sample size increases.
trader positions in certain futures and options markets.
Cheapest-to-deliver bond An embedded option available
Commodity (US futures contract) The CFTC Act of
to the seller of a Treasury bond futures contract by which
1974 defines a commodity to include any “goods and
one can deliver any one of several Treasury bonds at
articles . . . and all services, rights, and interests” on which
expiration.
contracts for future delivery can be written.
Clearing (a trade) When buy and sell orders are matched,
Commodity Futures Trading Commission (CFTC)
the trade is officially recognized, and the trade is recorded
The US federal government agency regulating futures and
by the exchange’s clearinghouse.
futures on option markets.
Clearing member Exchange member (broker and dealer)
Commodity indexed note A derivative whose holder
who clears trades.
receives a payment tied to the return on a physical
Clearinghouse An entity associated with or part of an commodity or a commodity index.
exchange that clears trades.
Commodity pool (or commodity fund) Investment
Close (futures market) The end of a day’s trading session. fund investing in commodity futures.
Closed-end fund Investment company that takes cash for Commodity swap A swap in which one counterparty
the issuance of shares at inception, but no additional shares makes a floating payment that is tied to the changes in
are issued thereafter. a commodity price or the return on a commodity price
Closing transaction (or a reversing trade) A trade that index.
cancels an outstanding open position. Competitive market An idealized description of a market
Collar (Option) A portfolio that combines a long position populated by a large number of traders who are too
in the underlying asset with a long put and short call (with insignificant to influence prices.
GLOSSARY 695
Complete market A market where sufficient securities Cum-dividend The price of a stock before it goes
trade so that investors can obtain any desired future ex-dividend.
payoffs. Currency swap A swap involving an exchange of payments
Condor spread An option portfolio created by trading four in one currency for that in another. A plain vanilla
options of the same type (all calls or all puts) with four currency swap (or a cross-currency swap) between two
different strike prices: two options with extreme strike counterparties involves (1) an exchange of equivalent
prices are purchased (written) and two options are written amounts in two different currencies on the start date, (2)
(purchased) with middle strike prices. an exchange of interest payments on two currency loans
at intermediate dates, and (3) a repayment of the principal
Consol A bond that never pays back the principal but pays
amounts on the loans at the ending date.
interest at a promised rate forever.
Consumer Price Index (CPI) An index measuring the Current assets These are cash or cash-equivalent assets
cost of a typical household’s goods and purchased services. held for collection, sale, or consumption within the enter-
prise’s normal twelve-month operating cycle; all other
Contango A forward or a futures market where the forward assets are noncurrent.
or futures price is higher than the spot price.
Current liabilities The liabilities settled within the enter-
Convenience yield A benefit that accrues to the holder of prise’s normal twelve-month operating cycle, or those
a physical commodity. held for trading, or those for which the entity does not
Convertible bond A bond whose holder has the option have an unconditional right to defer payment beyond
to convert the bond into a fixed number of shares of the twelve months; other liabilities are noncurrent.
borrowing company’s stock.
Cost-of-carry model A model that determines a forward Daily settlement (futures) Crediting/debiting of the day’s
price of a commodity under the assumption that buying gains/losses to a futures contract’s margin account.
and carrying the commodity to the future has interest as Dark pool (or dark pool of liquidity) Secretive trade
the only cost. matching networks for institutional traders that do not
Counterparties The traders (buyer and seller) in a financial send orders directly to an exchange or that are not
transaction. displayed in a limit order book.
Coupon A periodic interest payment made on a bond. Day order A limit order that gets cancelled if it is not
executed by the end of the trading day during which it
Coupon bond A bond that pays interest (coupons) peri-
is placed.
odically to the bondholder.
Day trader Trader who tries to profit from daily price
Covered (option trade) An option trade that is accompa- movements, opening trading positions in the morning
nied by a position in the underlying asset that reduces the and closing them out at night.
overall risk; the special case of a cash-secured put (a short
put trade in which the writer deposits an amount equal Dealer A financial intermediary who posts prices at which
to the strike price with the broker) is also considered a she buys (wholesale or bid price) or sells (retail or ask
covered trade. price) securities.
Credit default swap A derivative security that provides Default A failure to pay a promised payment on a financial
insurance against the event that the issuer of a bond might contract at the promised time.
default and the bond declines in value. Delivery date (or maturity date) A predetermined date
Credit swap A swap in which both payments are based on on which a seller of a futures contract agrees to deliver
individual stock returns in the same industry but with the underlying commodity to the buyer in exchange for
different credit ratings. the futures price.
Cross-hedge (imperfect hedge) A hedge in which the Delivery period (futures) Time period during which
spot and futures positions do not exactly offset each other. delivery of the underlying commodity, as stipulated in a
futures contract, can occur.
Cross-trade An offsetting or noncompetitive match of a
buy and sell order; normally banned, but permitted in Delta The sensitivity of an option price to changes in the
special cases. underlying asset price.
696 GLOSSARY
Delta-neutral A portfolio whose delta is zero. Euribor (Euro Interbank Offered Rate) An interest rate
Derivative (or a derivative security) A financial contract index computed on Euro Interbank term deposits.
that derives its value from an underlying variable such as a Eurobond An international bond that is denominated in a
stock price, a commodity price, or even an interest rate. currency not native to the issuing country.
Diagonal spread A portfolio obtained by buying an option Eurocurrency Time deposits in a bank denominated in a
with a particular expiration date and strike price and currency outside the borders of the country where the
simultaneously selling an option with a different expi- bank is located.
ration date and different strike price, where the options
Eurodollar Dollar-denominated time deposits held in a
(both calls or both puts) have the same underlying asset.
bank outside the United States.
Discriminatory auction An auction where the successful
European option An option that can be exercised only at
bidders pay their bid prices to purchase the item being
the expiration date.
sold.
Diversifiable risk A risk that can be eliminated in a Exchange A physical or electronic location where buyers
portfolio of securities via diversification. and sellers meet to trade a standardized commodity under
a given set of rules.
Dividend Payment in the form of cash or stocks made to
existing stockholders. Exchange of physicals (EFP) Terminating a futures con-
tract with a spot trade in which the buyer and the seller
Dividend yield A stock’s dividend payment expressed as a
privately negotiate the delivery of the commodity at a
fraction of the stock price.
location different from those in the futures contract.
Dollar return The value of an invested dollar on a future
Exchange rate The price of one currency in terms of
date.
another.
Dual trading Allowing a market participant to act as a
broker and dealer on the same day, though not on the Exchange-traded fund (ETF) Investment company that
same transaction. gives shareholders fractional ownership rights over a bas-
ket of securities.
Duration The weighted average of time to payment of a
bond’s cash f lows where the weights correspond to the Exchange-traded note (ETN) Bond issued by an under-
present value of the cash f lows received on the respective writing bank, which promises to pay a return (minus
dates. any fees) based on the performance of a market bench-
mark (like a commodity price index) or some investment
Dynamic hedging A hedge that is regularly adjusted over strategy.
time to maintain a riskless position.
Ex-dividend date The date after which the stock owner
does not get the last announced dividend.
Efficient market A market in which security prices “fully
Execution A trade is executed when the counterparties
and accurately” reflect all relevant information.
agree on the terms, conditions, price, and quantity and
Electronic communications network (ECN) An elec- commit to transact.
tronic system that automatically matches publicly dis-
played buy and sell orders at specified limit order prices. Exercise price (or strike price) The predetermined price
that an option holder pays to exercise the option.
Employee stock options Call options on the stock that
are granted by a company to employees as a form of non- Exotic (derivative) Complex derivatives that are not ordi-
cash compensation. nary call or put options.
Equilibrium model Model that determines prices by Expectations hypothesis The hypothesis that zero-
requiring the equality of supply and demand. coupon bonds of different maturities are perfect
substitutes.
Equity option Option that has common stock as the
underlying. Expense ratio (mutual fund) A mutual fund’s annual
operating expenses expressed as a percentage of average
Equity swap A swap in which one side makes a floating
net assets.
payment tied to the return on a stock or a stock index.
Expiration date (option) The last day that an option
Equity-linked note (ELN) A combination of a zero-
holder gets to exercise the option.
coupon (or a low-coupon) bond with a return based on
the performance of a single stock, a basket of stocks, or a Extendible bond A bond whose holder has the option to
stock index. extend the bond’s life.
GLOSSARY 697
Fail to deliver In a repurchase agreement, when a coun- Forward rate curve The graph of forward rates as a
terparty fails to deliver the securities, the loan is extended function of the time to maturity.
by a day at no extra interest. Front running Trading based on nonpublic information
Federal funds market A borrowing or lending market for about an impending transaction by another person.
cash among US banks. Fundamental analysis Security analysis that involves read-
Federal Reserve System (the Federal Reserve or ing accounting and financial information about a com-
simply “the Fed”) The central bank of the United pany to determine whether the share is overvalued or
States. undervalued.
Financial Accounting Standards Board (FASB) A pri- Futures commission merchant (FCM) A futures trade
vate, not-for-profit organization whose primary purpose facilitator who provides a one-stop service for all aspects
is to establish financial accounting standards that govern of futures trading.
the preparation of financial reports by nongovernmental Futures contract A futures contract is an exchange-traded
entities. agreement between a buyer and a seller to trade some
Financial assets Paper assets such as stocks, bonds, and commodity at a fixed price at a later date.
currencies that represent claims to real assets.
Financial engineering The science of applying engineer-
Gamma The sensitivity of the delta to changes in the
ing tools to develop financial contracts to meet the needs
underlying asset price.
of an enterprise.
Gilt security Bond issued by the UK government.
Financial futures (or financials) Futures contracts based
on financial assets (like foreign currencies) or financial Good-till-canceled (GTC) order A limit order that stays
variables (like interest rates). open indefinitely until it is executed or closed by the
trader who placed the order.
Financial Industry Regulatory Authority (FINRA)
A self-regulatory organization created by combining Great Recession A commonly used term for 2007–9,
the self-regulatory functions of the New York Stock when many countries suffered a decline in output and
Exchange and the National Association of Securities experienced large unemployment.
Dealers. Gross domestic product The market value of all goods
Fixed price offering An offer to sell securities where the and services produced in a year in a country.
price, the interest rate, and the maturity are announced
beforehand.
Hedge Protecting one’s investment against loss by making
Floor A portfolio of floorlets. balancing or compensating contracts or transactions.
Floor broker A broker working on an exchange floor. Hedge fund A private actively managed investment fund
Floorlet A European put option on an interest rate. that utilizes sophisticated strategies to generate returns for
Forex swap (or FX swap) A swap in which (1) the high-net-worth individuals.
contract starts with an exchange of one currency for Hedge ratio The number of shares of the stock to trade
another and (2) it ends with a reverse exchange of these for each option to remove all of the price risk from the
cash payments, with interest, at a subsequent date. resulting portfolio.
Forward contract A binding agreement between a buyer Hedger Trader whose aim is to reduce preexisting risk by
and a seller to trade some commodity at a fixed price at a trading derivatives and other securities.
later date. High-frequency trading A computerized trading strategy
Forward price (or delivery price) The fixed price in pursuing fleeting profit opportunities by trading at the
a forward contract in which the counterparties agree to blink of an eye.
trade a commodity on a future delivery date.
Horizontal spread (time or calendar spread) A port-
Forward rate The rate one can contract today on riskless folio obtained by buying an option with a particular
borrowing or lending over a future time period. expiration date and simultaneously selling an option with
Forward rate agreement A derivatives contract that pays a different maturity date, where the options (both calls or
off based on the interest rate for borrowing (or lending) both puts) have the same underlying asset and the same
over some future time period. strike price.
698 GLOSSARY
Hybrid debt (or structured note) A derivative security Intracommodity spread (futures) A spread established
created by combining a bond and equity or a derivative by simultaneously buying and selling futures contracts on
(a forward, an option, or a swap). the same commodity but with different maturities.
Intrinsic value (Asset) The present value of future cash
Implied volatility An option’s underlying asset’s return flows, also known as fundamental value. (Option) The
volatility computed by equating the option’s market price larger of the cash flow obtained by immediately exercising
to the option’s model price. an option or zero.
Index (stock market) A measure computed by taking the
average of a portfolio of stock prices. Law of one price The condition that the same future cash
Index arbitrage An arbitrage trading strategy based on a flows, no matter how they are created, have the same
discrepancy between a stock market index and a futures market price.
contract on the index. LEAPS (Long-Term Equity Anticipation Securities)
Index mutual fund (or stock index fund) A mutual Exchange-traded long maturity options that have matu-
fund that holds a basket of securities in the same propor- rities up to three years.
tion as an index it tries to replicate. Leverage The borrowing used to purchase a security or
Indexing A popular name for investing in index portfolios. portfolio of securities.
Indication schedule (swap) A list of the various rates at Liability A financial obligation that requires the holder to
which a swap dealer bank is willing to trade against six- make payments at some future date.
month bbalibor. Libor (London Interbank Offered Rate) and libid An
Inflation A rise in a general consumer price level. annualized interest rate at which a bank offers to lend
surplus funds to another bank for a fixed time period in
Initial margin (or performance margin) Before initiat- the London market. Libid (London Interbank Bid Rate)
ing a trade, a trader must keep this amount (in the form of is a similar rate at which a bank bids funds.
cash or high-quality, low-risk securities) as collateral with
the broker. Limit order A trade given to a broker that must be filled
at the stated or a better price or not traded at all.
Insurance A financial contract that requires the buyer to
make periodic payments to an insurance company for Libor index rate The daily rate determined by the
protection against a loss event. Intercontinental Exchange by collecting libor quotes on
Eurodollar deposits (or time deposits in nine other cur-
Intercommodity spread (futures) A spread established rencies) with maturities ranging from overnight to a year
by simultaneously buying and selling futures contracts on from a panel of major banks and computing a trimmed
different commodities with the same or different delivery average.
months.
Limit order book A dealer’s book (or an electronic record
Interest rate The rate of return on cash borrowed or lent. system) that records orders waiting to be filled.
International Money Market A division of the Chicago Listing requirements Criteria that a company must satisfy
Mercantile Exchange (now part of the CME Group), to trade its stock on an exchange.
where futures on foreign currencies trade.
Long The buyer of a security.
International Organization of Securities Commis-
sions (IOSCO) An international cooperative body of
securities regulators. Maintenance margin The minimum amount that a trader
International Swaps and Derivatives Association must keep in a margin account to keep a trading position
(ISDA; previously International Swap Dealers open.
Association) An industry body consisting of financial Margin Collateral kept by a trader with a broker to cover
institutions, businesses, governmental entities, and other the risk of losses on a transaction.
end users that trade over-the-counter derivatives. Margin call A directive by a broker to increase the cash in
In-the-money (option) An option having a positive value a margin account when a trade position loses value and
when immediately exercised. declines below the maintenance margin.
GLOSSARY 699
Marked-to-market (futures) A futures position is Modified duration Duration divided by one plus the
marked-to-market at the end of each trading day when bond’s yield.
gains and losses incurred on the futures contract are added Money market account An investment account that earns
to the trader’s margin account. the risk-free rate each period.
Market close (or the closing call) The last minute of Mortgage A loan that requires regular repayment of interest
trading during which a closing price or a closing range is and principal where the loan is collateralized by a home
established. or building.
Market corner A market in which a manipulator has con- Municipal bond (muni) Bonds issued by state, local, or
trol over the deliverable supply of a security or commodity municipal governments whose holders typically enjoy a
so that short sellers are forced to pay the manipulator a variety of tax-exemption benefits on interest earned.
high price to fulfill their contract obligations.
Municipal swap A financial swap that has a state, local, or
Market imperfections (or market frictions) Impedi- municipal government as one of the counterparties.
ments or costs to a trade such as transaction costs (broker-
age commissions and bid–ask prices), trading restrictions, Mutual fund Open-end investment company that takes
or taxes incurred when transacting in a security. new investments and allows investors to redeem their
shares at any time.
Market maker Specially designated dealer on many
exchanges who posts ask and bid prices and stands ready
to trade at those prices. Naked trading strategy Option trade involving a single
Market manipulation Trading securities using market option with no corresponding position in the underlying
power or other means to influence a securities price to asset.
one’s own profit advantage. National Association of Securities Dealers (NASD)
Market order A trade that must be immediately transacted A self-regulatory organization that was created under the
at the best available price. provisions of the US Securities Exchange Act of 1934 for
regulating the US securities industry.
Market portfolio A portfolio that holds all the risky assets
in a market in the same proportions as they exist in the National Association of Securities Dealers Auto-
economy. mated Quotations (NASDAQ) Initially a computer
bulletin board system that listed stock price quotes;
Market-value (or capitalization) weighted average A
evolved to become the world’s first electronic stock
stock price index computed by multiplying each com-
exchange.
pany’s share price by a weight that is proportionate to its
market value (or capitalization) and then computing their National Futures Association (NFA) An industry orga-
average. nization that does a host of self-regulatory, training, pro-
Mark-to-model The valuation of assets and liabilities via motional, and other activities for futures contracts.
the use of a model. Net asset value (NAV) A company’s total assets minus its
Married put A portfolio obtained by simultaneously buy- total liabilities.
ing a put and its underlying asset. Net present value Today’s value of a set of future cash
Martingale A stochastic process in which the expected flows, subtracting the initial cost.
value at a future date equals its current value. Noncommercial trader (futures) Trader other than a
Martingale pricing A method of derivatives pricing that commercial trader; usually associated with speculation in
involves computing the expected discounted payoffs to futures contracts.
the derivative using the pseudo-probabilities and dis- Nondiversifiable risk Risk that cannot be eliminated
counting them backward through time with the riskless through diversification in a portfolio of securities.
rate.
Normal backwardation A forward or futures market in
Master Agreement (ISDA) A comprehensive document
which the forward or futures price, respectively, is less
that consists of the legal conditions agreed to when
than the expected future spot price.
trading over-the-counter derivatives.
700 GLOSSARY
Notional principal A principal amount that is used for Payout ratio The ratio of dividends paid on a stock to
computing cash flows in a derivatives transaction but earnings per share.
never changes hands between the counterparties. Physical delivery (forward and futures contracts) The
Notional variable Variable such as interest rate, inflation process of closing out a forward or a futures contract by
rate, and security index that exists as a notion rather than physically delivering the underlying asset or its ownership
as a traded asset. rights.
Pink quote An electronic system that displays quotes from
On-the-run issue vs. off-the-run issue Newly auc- broker dealers for small stocks trading in the over-the-
tioned Treasury security vs. an issue that has been counter markets.
auctioned earlier. Pit (ring) Polygonal area on an exchange floor where
Open interest The total number of outstanding derivatives futures trading takes place.
contracts. Plain vanilla derivatives Market jargon for the simplest
Open-end company Investment company that is open to derivatives contracts.
additional investments. Ponzi scheme An investment scheme that pays early
Opening call (futures) The initial time period for trading investors’ returns from the cash coming from later
a futures contract. investors.
Open-outcry trading A traditional method of trading Portfolio insurance Synthetic put options created by trad-
futures contracts in which the floor broker or dealer has to ing the underlying stock and money market account using
cry out the bids and asks in the trading pit so that others the option’s delta.
may participate. Position limit The maximum number of futures (deriva-
Option class The set of all call options or put options with tives) contracts that a trader can hold on a particular side
a particular underlying asset. of the market.
Options Clearing Corporation (OCC) A clearinghouse Position trader (trend follower) Speculator who main-
for most exchange-traded options in the United States. tains trading positions for longer than a day.
Options on futures (or futures options) Options that Post (trading post) Specified places on an exchange’s
have a futures contract as the underlying asset. trading floor where traders meet to trade.
Options series A set of options with identical terms and Premium (option) The price of an option.
conditions. Price bubble The difference between an asset’s market
OTC Bulletin Board (OTCBB) A regulated quotation price and its fundamental value.
service that displays real-time quotes, last-sale prices, Price discovery The process of discovering the price of a
and volume information on over-the-counter equity commodity through the interactions of buyers and sellers
securities. in a market.
Out trade A trade that cannot be cleared by a clearinghouse Price limit (daily; or the maximum daily price
because of the recording of incorrect information. fluctuation) Maximum price change allowed during a
Out-of-the-money (option) An option that generates a trading day.
negative cash flow when immediately exercised. Price-weighted average (stock index) A stock index
Over-the-counter (OTC) market Anywhere trading computed by taking a simple average of the prices of the
takes place other than an organized exchange. stocks composing the index.
Overwrite (option) Selling a call option after purchasing Primary dealer Big securities firms that actively bid in
the underlying stock. Treasury auctions and with whom the Fed buys and sells
Treasuries to conduct open-market operations.
Par bond A bond for which the coupon rate is the same as Primary market The market in which a security is first
its yield so that the bond’s price is the same as the bond’s sold to the public.
face value. Principal Amount due on a bond at maturity.
Par swap rate The swap rate initially set on a swap to give Program trading Trading that takes place using a com-
it zero initial value. puter program.
GLOSSARY 701
Protective put A portfolio consisting of buying the put to Savings and loans bank A special kind of bank in the
protect the downside risk of the underlying asset that was United States that specializes in home mortgage lending.
previously purchased. Scalper Speculator who trades many times a day with the
Pseudo-probability (martingale probability or risk- hope of picking up small profits from each transaction.
neutral probability) The risk-adjusted probability used Scenario analysis (stress testing) A method for comput-
in martingale pricing for computing derivatives prices. ing losses based on identifying possible future scenarios
Put option A financial security that gives the right to sell for the economy.
a specified quantity of an underlying asset on or before a Seat (in the context of a securities exchange) Mem-
fixed future date by paying an exercise price. bership on an exchange.
Putable bond A bond that gives the holder the option to Securities and Exchange Commission (SEC) The US
end the bond at certain specific dates. federal government agency in charge of regulating equity
Put–call parity A mathematical relation linking prices of a and equity options markets.
call, a put, a stock, and a bond. Securitization The process of combining many individual
loans and debts into a collateral pool and selling claims
Random walk A stochastic process whose changes are against them to third-party investors.
independent and identically distributed random variables. Security A financial contract (such as a stock, bond, or
Ratio spread An option trading strategy where the number derivative) that gives its holder ownership rights over
of calls (or puts) purchased is different from the number some future cash flows.
of calls (or puts) sold. Self-financing trading strategy A trading strategy over
Real asset Asset including land, buildings, machines, or multiple periods that requires no cash inflow or outflow
commodities, which have a physical form. at intermediate dates.
Reinsurance The practice of insurance companies buying Self-regulatory organization (SRO) An organization
insurance policies on the tail risks generated by the made up of members of an industry or a profession that
insurance policies they have issued. exercises regulatory authority over it.
Repurchase agreement (“does a repo”) The borrowing Semiannual bond equivalent basis Computing interest
of cash (and the loaning of securities) done by selling over a half-year period under the assumption that the year
high-quality securities and then agreeing to purchase has 365 days and the interest is computed over the actual
them back at a fixed date (often the next day) at a higher number of days in the six-month period.
price. Separate trading of registered interests and principal
Residential mortgage–backed security (RMBS) An of securities (STRIPS) Zero-coupon bonds that can
asset-backed security whose collateral pool consists of be created by separating out different cash flows from a
residential mortgage loans. Treasury security.
Rho A measure of the change in an option value due a Settlement The last step of a security transaction where
change in the risk-free interest rate. the buyer pays cash and gets the ownership rights over
the security.
Risk premium The excess expected return per unit of risk
on a security. Settlement date (forward rate agreement) The date
on which a forward rate agreement’s reference rate is
Risk-neutral valuation A method for computing a
determined.
derivative’s value by computing its expected payoffs using
the pseudo-probabilities and discounting using the risk- Settlement price (futures) The last trading price or an
free rate. appropriate price determined by an exchange’s settlement
committee, which is used for daily settlement of the
Round lot A trading unit of stocks, usually based on one
futures contract.
hundred shares.
Shareholder’s equity (equity) The assets owned minus
Round trip (commissions for trading futures) No
the liabilities incurred by a company.
commission charged when entering a position but full
commission charged at the time of closing. Short Market jargon to mean the seller of a security.
702 GLOSSARY
Short selling Borrowing and selling a security one does not Swap facilitator (swap bank) Financial intermediary
own. who acts as a broker or performs the role of a dealer in
Simple forward rate The forward rate expressed as an the swap market.
annualized simple interest rate. Swap issuer (initiator) A party that initiates a financial
Special purpose vehicle (SPV) A legal entity set up for swap by seeking out counterparties to complete the trans-
the purpose of issuing liabilities that are backed by a action.
collection of assets in a collateral pool. Swap rate The predetermined interest rate that a financial
Specialist A dealer-broker market maker whose job is to swap counterparty pays in exchange for a well-known
make an orderly market in securities assigned to him. reference rate such as the bbalibor. This rate gives the
swap a zero value at initiation.
Speculation The taking of risky trades in the hope of
obtaining positive returns. Swaption An option that has an interest rate swap as the
underlying asset.
Spin-off The creation of an independent company by sell-
ing shares or distributing as new shares an existing business Synthetic index A portfolio created for the purpose of
or a division of a parent company. replicating the returns on an index.
Spot price (or cash price) The price for an immediate Synthetic option A portfolio of traded securities that
transaction of a commodity or security. replicates an option’s payoff and consequently must have
the same price as the traded option to avoid arbitrage.
Spot rate The yield on the shortest-maturity default-free
zero-coupon bond.
Spread trading Trading that involves buying one or more Tail risk The risk of occurrence of infrequent events that
securities and selling similar securities with the aim lie on the tail of a probability distribution.
of generating arbitrage profits as the price differences Technical analysis A security analysis technique that
converge. involves looking at price patterns and related measures
Stock A limited liability asset that gives the investor own- to predict whether a stock’s price is overvalued or under-
ership rights over the residual value of a company. valued.
Stock price evolution A model that describes how a stock Tenor (maturity or term of a swap) Time duration of
price evolves randomly across time. a financial swap.
Stop-loss order A limit sell order that is placed below Term structure of interest rates The graph of yields or
current market prices or a limit buy order that is placed forward rates versus the time to maturity.
above current prices. Theta A measure of the change in an option’s value given
Storage costs Costs that are incurred for storing a physical a decrease in the option’s time to maturity.
commodity in the context of the cost-of-carry model. Tick size (minimum price fluctuation) The minimum
Straddle (called put-to-call strategy in London) An price move for an exchange-traded security.
option trading strategy that involves simultaneously buy- Ticker (trading) symbol The abbreviation that identifies
ing (or selling) a call and a put on the same underlying an exchange-traded security.
asset with the same strike price and expiration date. Ticker tape The running paper tape or electronic panel
Strangle An option trading strategy that involves simul- displaying a variety of prices, quotes, and relevant news
taneously buying (or selling) a call and a put on the to subscribers located both in and off the exchange.
same underlying asset with the same expiration date but Time spread (calendar spread) A spread established by
different strike prices. simultaneously buying and selling futures contracts on
Subprime lending (also known as B-paper, near- commodities with different delivery months.
prime, or second chance lending) The practice of Time value (option) An option’s premium (or price)
making loans to borrowers who do not qualify for the minus its intrinsic value.
best market interest rates (the prime market) because of a
Total return index An index that assumes that all disburse-
poor credit history.
ments from the companies in the index get reinvested in
Swap (financial swap) An agreement between two coun- a hypothetical index portfolio.
terparties to exchange a series of (usually) semiannual cash
Total return swap A swap where one counterparty pays
flows over the life of the contract.
the total return on a price index in exchange for a fixed
Swap buyer Counterparty to a (plain vanilla) financial swap or floating interest rate payment.
that pays fixed and receives the floating rate.
GLOSSARY 703
Trading at settlement (or TAS) Special futures contracts a particular strike price while simultaneously selling an
in which the counterparties decide at the time of trading option with a different strike price, where the options
that the contract price will be the day’s settlement price (both calls or both puts) have the same underlying asset
plus or minus an agreed differential. and expire on the same date.
Trading volume (or volume) The total number of con- VIX (CBOE Volatility index) A measure of the market’s
tracts or securities traded during a trading day. near-term volatility as implied by the Standard and Poor’s
Treasury bill (T-bill) US government debt sold in the 500 stock index option prices.
form of a zero-coupon bond that has a maximum matu-
rity of one year. Wall Street An expression for investment banking compa-
Treasury inflation protected security (TIPS) Treasury nies, many of which have offices near Wall Street in New
security that adjusts for inflation and pays a guaranteed York’s financial district.
real rate of return to the investors. Warehousing swap A risk management technique used by
Treasury note and bond (T-note and T-bond) Treasury swap providers in which they enter into different swaps
securities that make semiannual coupon payments with on both the buy and sell side of the market and hedge the
original maturity of more than one year to up to ten years residual risk.
in the case of a Treasury note and more than ten years in Wash sale A sell and buy transaction for the same com-
the case of a Treasury bond. modity initiated without the intent to make a bona fide
Treasury security (Treasuries) Debt issued by the US transaction.
federal government. When-issued market A specialized forward market for
Treasury–Eurodollar (TED) spread The difference Treasury securities that opens before the actual securities
between the bbalibor rate for US dollars and Treasury are auctioned.
rates of equivalent maturities. Wild card play (T-bond futures) An embedded option
available to the seller of Treasury bond futures to issue the
Underlying (for a derivative) A reference asset or a notice of intention to deliver for a few hours after the
notional variable whose price or value determines the market closes.
derivative’s payoff. Writer (option) Seller of an option.
Uniform price auction An auction in which all successful
bidders pay the same price (the highest losing bid or the Yield (yield to maturity or internal rate of return)
lowest winning bid). The interest rate that discounts a bond’s cash flows to
Upstairs market A market away from the trading floor in equal its current price.
which large-sized trades are negotiated, with or without Yield curve A graph of bond yields as a function of
the help of brokers. maturity.
Value-at-risk (VaR) The dollar amount that the firm’s
Zero-cost collar A portfolio that combines a long position
losses will exceed with a given probability (such as
in the underlying asset with a long put and a short call so
1 percent or 5 percent) over a prespecified time interval.
that the resultant portfolio has truncated payoffs on both
Variation margin An amount equal to the change in the up and down sides.
futures settlement price from the previous trading day that
Zero-coupon bond A bond that pays no interest but is
is credited or debited to a trader’s account.
sold at a discount from the principal.
Vega A measure of the change in an option’s value given a
Zero-sum game A game in which one trader’s gain is the
change in the underlying asset’s return volatility.
other’s loss.
Vertical spread (money, perpendicular, or price
spread) A portfolio consisting of buying an option with
References
Allen and Overy, 2002. “An Introduction to the Documen- Bhattacharya, Mihir, 1983. “Transactions Data Tests of Effi-
tation of OTC Derivatives.” www.isda.org. ciency of the Chicago Board Options Exchange.” Journal
Alletzhauser, Albert J., 1990. The House of Nomura: The of Financial Economics 12(2), 161–85.
Inside Story of the Legendary Japanese Financial Dynasty. Bhattacharya, Utpal, and Hazem Daouk, 2002. “The World
Harper Perennial, New York. Price of Insider Trading.” Journal of Finance 57, 75–108.
Amemiya, Takeshi, 1994. Introduction to Statistics and Econo- Billingsley, Patrick, 1995. Probability and Measures. 3rd ed.
metrics. Harvard University Press, Cambridge, Mass. John Wiley, New York.
Amin, Kaushik I., and Robert A. Jarrow, 1991. “Pricing For- Black, Fischer S., 1976. “The Pricing of Commodity Con-
eign Currency Options under Stochastic Interest Rates.” tracts.” Journal of Financial Economics 3(1–2), 167–79.
Journal of International Money and Finance 10(3), 310–29. Black, Fischer S., and Myron S. Scholes, 1973. “The Pricing
Arrow, Kenneth J., 1953. “Le rôle des valeurs boursières of Options and Corporate Liabilities.” Journal of Political
pour la répartition la meilleure des risques.” Econometrie Economy 81, 637–59.
11, 41–48. (Trans. 1964, “The Role of Securities in the Black, Fischer S., Emmanuel Derman, and William W. Toy,
Optimal Allocation of Risk-Bearing.” Review of Economic 1990. “A One-Factor Model of Interest Rates and Its
Studies 31(2), 91–96.) Application to Treasury Bond Options.” Financial Analysts
Arrow, Kenneth J., and Gerard Debreu, 1954. “Existence of Journal 46, 33–39.
an Equilibrium for a Competitive Economy.” Econometrica Box, George E. P., and Norman R. Draper, 1987. Empirical
22(3), 265–90 Model-building and Response Surfaces. John Wiley, New
Artzner, Philippe, Freddy Delbaen, Jean-Marc Eber, and York.
David C. Heath, 1999. “Coherent Measures of Risk.” Brace, Alan, Dariusz Gatarek, and Marek Musiela, 1997.
Mathematical Finance 9(3), 203–28. “The Market Model of Interest Rate Dynamics.”
Mathematical Finance 7(2), 127–55.
Bachelier, Louis J.-B. A., 1900. Théorie de la Spéculation. Brealey, Richard A., and Stewart C. Myers, 2003. Principles
Gauthier-Villars, Paris. of Corporate Finance. 7th ed. McGraw-Hill, New York.
Bakshi, Gurdip, Charles Cao, and Zhiwu Chen, 1997. Brennan, Michael J., and Eduardo S. Schwartz, 1979. “A
“Empirical Performance of Alternative Option Pricing Continuous Time Approach to the Pricing of Bonds.”
Models.” Journal of Finance 52(5), 2003–49. Journal of Banking and Finance 3(2), 133–55.
Bank, Peter, and Dietmar Baum, 2004. “Hedging and Brickley, James A., and John J. McConnell, 1987.
Portfolio Optimization in Financial Markets with a Large “Dividend Policy.” In The New Palgrave: A Dictionary
Trader.” Mathematical Finance 14(1), 1–18. of Economics. Edited by John Eatwell, Murray Milgate,
Bank for International Settlements, 2005. OTC Derivative and Peter Newman. W. W. Norton, New York. (Repr.
Market Activity in the First Half of 2005. Monetary and 1989, The New Palgrave: Finance.)
Economic Department, Bank for International Settle-
ments, Basel, Switzerland. Campbell, John Y., Andrew W. Lo, and A. Craig
Baxter, Martin W., and Andrew J. O. Rennie, 1996. MacKinlay, 1996. Econometrics of Financial Markets.
Financial Calculus: An Introduction to Derivatives Pricing. Princeton University Press, Princeton, N.J.
Cambridge University Press, Cambridge. Cantor, Richard, and Frank Packer, 1994. “The Credit
Bernstein, Peter L., 1992. Capital Ideas: The Improbable Rating Industry.” Federal Reserve Bank of New York
Origins of Modern Wall Street. John Wiley, Hoboken, N.J. Quarterly Review 19(2), 1–26.
Bernstein, Peter L., 1998. Against the Gods: The Remarkable Castelli, Charles, 1877. The Theory of “Options” in Stocks
Story of Risk. John Wiley, New York. and Shares. Fred C. Mathieson, London.
704
REFERENCES 705
Çetin, Umut, Robert A. Jarrow, and Philip E. Protter, Dixit, Avinash K., and Robert S. Pindyck, 1994. Investment
2004. “Liquidity Risk and Arbitrage Pricing Theory.” under Uncertainty. Princeton University Press, Princeton,
Finance and Stochastics 8(3), 311–41. N.J.
Chance, Don M., 2008. “A Synthesis of Binomial Option Duffie, Darrell, 1989. Futures Markets. 1st ed. Prentice Hall,
Pricing Models for Lognormally Distributed Assets.” Englewood Cliffs, N.J.
Journal of Applied Finance 18(1), 38–56. Duffie, Darrell, 2001. Dynamic Asset Pricing Theory. 3rd ed.
Chang, Carolyn W., and Jack S. K. Chang, 1990. “Forward Princeton University Press, Princeton, N.J.
and Futures Prices: Evidence from Foreign Exchange Dunn, Kenneth B., and Kenneth J. Singleton, 1986.
Markets.” Journal of Finance 45(4), 1333–36. “Modeling the Term Structure of Interest Rates under
Chatterjea, Arkadev, 1993. Market Manipulation and a Model Nonseparable Utility and Durability of Goods.” Journal of
of the United States Treasury Securities Market. Unpublished Financial Economics 17(1), 27–55.
PhD dissertation, Cornell University, Ithaca, N.Y.
Chatterjea, Arkadev, and Robert A. Jarrow, 1997. “Market Easterbrook, Frank H., 1986. “Monopoly, Manipulation
Manipulation, Price Bubbles, and a Model of the U.S. and the Regulation of Futures Markets.” Journal of
Treasury Securities Auction Market.” Journal of Financial Business 59(2), 103–27.
and Quantitative Analysis, 33(2), 255–89.
Edwards, Franklin R., and Cindy W. Ma, 1992. Futures
Cornell, Bradford, and Marc R. Reinganum, 1981. and Options. McGraw-Hill, New York.
“Forward and Futures Prices: Evidence from Foreign
Exchange Markets.” Journal of Finance 36(12), 1035–45. Ehrlich, Paul R., 1968. The Population Bomb. Ballantine,
New York.
Courtault, Jean-Michel, Yuri Kabanov, Bernard Bru,
Pierre Crépel, Isabelle Lebon, and Arnaud L. Marchand, Eisenberg, Larry, and Robert A. Jarrow, 1994. “Option
2000. “Louis Bachelier on the Centenary of Théorie de la Pricing with Random Volatilities in Complete Markets.”
Spéculation.” Mathematical Finance 10(3), 341–53. Review of Quantitative Finance and Accounting 4(1), 5–17.
(Repr. 1998, Volatility: New Estimation Techniques for
Cox, David R., and Hilton D. Miller, 1990. The Theory of Pricing Derivatives, Risk, London.)
Stochastic Processes. Chapman Hall, London.
Engle, Robert F., and Joshua V. Rosenberg, 2002.
Cox, John C., Stephen A. Ross, and Mark Rubinstein, “Empirical Pricing Kernels.” Journal of Financial Economics
1979. “Option Pricing: A Simplified Approach.” Journal 64(3), 341–72.
of Financial Economics 7(3), 229–63.
Cox, John C., Jonathan E. Ingersoll, and Stephen A. Ross,
1981. “A Re-examination of Traditional Hypotheses Feynman, Richard P., 1998. The Meaning of It All: Thoughts
about the Term Structure of Interest Rates.” Journal of of a Citizen-Scientist. Perseus Books, Reading, Mass.
Finance 36(4), 769–99. Filer, Herbert J., 1959. Understanding Put and Call Options.
Cox, John C., Jonathan E. Ingersoll, and Stephen A. Ross, Popular Library, New York.
1985. “A Theory of the Term Structure of Interest Finucane, Thomas J., 1997. “An Empirical Analysis of
Rates.” Econometrica 53(2), 385–407. Common Stock Call Exercise: A Note.” Journal of
Culbertson, John M., 1957. “The Term Structure of Interest Banking and Finance 21(4), 563–71.
Rates.” Quarterly Journal of Economics 71(4), 485–517. Fischel, Daniel R., and David J. Ross, 1991. “Should
the Law Prohibit ‘Manipulation’ in Financial Markets?”
Dash, Mike, 2001. Tulipomania: The Story of the World’s Most Harvard Law Review 105(2), 503–53.
Coveted Flower and the Extraordinary Passions It Aroused. Fisher, Irving M., 1930. The Theory of Interest. Macmillan,
Three Rivers Press, New York. New York.
DeGroot, Morris H., and Mark J. Schervish, 2011. Probabil- Follmer, Hans, and A. Schied, 2004. Stochastic Finance: An
ity and Statistics. 3rd ed. Pearson Addison Wesley, Boston. Introduction in Discrete Time. 2nd ed. Walter de Gruyter,
Dezhbakhsh, Hashem, 1994. “Foreign Exchange Forward Berlin.
and Futures Prices: Are They Equal?” Journal of Financial French, Kenneth R., 1980. “Stock Returns and the Week-
and Quantitative Analysis 29(1), 75–87. end Effect.” Journal of Financial Economics 8(1), 55–69.
706 REFERENCES
French, Kenneth R., and Richard Roll, 1986. “Stock Ho, Thomas S. Y., and Sang-Bin Lee, 1986. “Term
Return Variances: The Arrival of Information and the Structure Movements and Pricing Interest Rate
Reaction of Traders.” Journal of Financial Economics 17(1), Contingent Claims.” Journal of Finance 41(5), 1011–29.
5–26. Ho, Thomas S. Y., and Sang-Bin Lee, 2004. The Oxford
Froot, Kenneth A., David S. Scharfstein, and Jeremy C. Guide to Financial Modeling: Applications for Capital
Stein, 1994. “A Framework for Risk Management.” Markets, Corporate Finance, Risk Management and Financial
Harvard Business Review 72(6), 91–102. Institutions. Oxford University Press, New York.
Hughston, Lane P., 1996. “The New Interest Rate Mod-
Gastineau, Gary L., and Mark P. Kritzman, 1996. The els.” In The New Interest Rate Models. Risk Books, London.
Dictionary of Financial Risk Management. Frank J. Fabozzi Hull, John C., 2002. Options, Futures, and Other Derivatives.
Associates, New York. 5th ed. Prentice Hall, Upper Saddle River, N.J.
Gelderblom, Oscar, and Joost Jonker, 2005. “Amsterdam
as the Cradle of Modern Futures and Options Trading,
1550–1650.” In The Origins of Value: The Financial Jarrow, Robert A., 1981. “Liquidity Premiums and the
Innovations That Created Modern Capital Markets. Edited Expectations Hypothesis.” Journal of Banking and Finance
by William N. Goetzmann and K. Geert Rouwenhorst. 5(4), 539–46.
Oxford University Press, New York. Jarrow, Robert A., 1992. “Market Manipulation, Bubbles,
Glasserman, Paul, 2003. Monte Carlo Methods in Financial Corners, and Short Squeezes.” Journal of Financial and
Engineering. Springer Science + Business Media, New Quantitative Analysis 27(3), 311–36.
York. Jarrow, Robert A., 1994. “Derivative Security Markets,
Greenbaum, Stuart I., and Anjan V. Thakor, 2007. Market Manipulation and Option Pricing Theory.”
Contemporary Financial Intermediation. 6th ed. Academic, Journal of Financial and Quantitative Analysis 29(2), 241–61.
Burlington, Mass. Jarrow, Robert A., 1999. “In Honor of the Nobel Laureates
Grossman, Sanford J., and Joseph E. Stiglitz, 1980. “On Robert C. Merton and Myron S. Scholes: A Partial
the Impossibility of Informationally Efficient Markets.” Differential Equation That Changed the World.” Journal
American Economic Review 70(3), 393–408. of Economic Perspectives 13(4), 229–48.
Jarrow, Robert A., 2002a. “Put Option Premiums and
Hao, Jia, Avner Kalay, and Stewart Mayhew, 2010. Coherent Risk Measures.” Mathematical Finance 12(2),
“Ex-Dividend Arbitrage in Option Markets.” Review of 135–42.
Financial Studies 23(1), 271–303. Jarrow, Robert A., 2002b. Modeling Fixed Income Securities
Harrison, J. Michael, and David M. Kreps, 1979. and Interest Rate Options. 2nd ed. Stanford University
“Martingales and Arbitrage in Multiperiod Securities Press, Palo Alto, Calif.
Markets.” Journal of Economic Theory 20(3), 381–408. Jarrow, Robert A., 2008. “Operational Risk.” Journal of
Harrison, J. Michael, and Stanley R. Pliska, 1981. Banking and Finance 32(5), 870–79.
“Martingales and Stochastic Integrals in the Theory Jarrow, Robert A., 2009a. “Credit Risk Models.” Annual
of Continuous Trading.” Stochastic Processes and Their Review of Financial Economics 1, 37–68.
Applications 11(3), 215–60. Jarrow, Robert A., 2009b. “The Term Structure of Interest
Heath, David C., and Robert A. Jarrow, 1988. “Ex- Rates.” Annual Review of Financial Economics 1, 1–28.
Dividend Stock Price Behavior and Arbitrage Jarrow, Robert A., 2011. “Risk Management Models: Cons-
Opportunities.” Journal of Business 61(1), 95–108. truction, Testing, Usage.” Journal of Derivatives 18(4), 89–98.
Heath, David C., Robert A. Jarrow, and Andrew J. Morton, Jarrow, Robert A., 2012. “The Role of ABS, CDS and
1992. “Bond Pricing and the Term Structure of Interest CDOs in the Credit Crisis and the Economy.” In
Rates: A New Methodology for Contingent Claims Rethinking the Financial System. Edited by Alan S. Blinder,
Valuation.” Econometrica 60(1), 77–105. Andrew W. Lo, and Robert M. Solow. Russell Sage
Hicks, John R., 1946. Value and Capital. 2nd ed. Clarendon Foundation, New York.
Press, Oxford. Jarrow, Robert A., and George S. Oldfield, 1981. “Forward
Higgins, Leonard R., 1902. The Put-and-Call. Elfingham Contracts and Futures Contracts.” Journal of Financial
Wilson, Royal Exchange, London. Economics 9(4), 373–82.
REFERENCES 707
Jarrow, Robert A., and Philip E. Protter, 2008. “An of a Nineteenth-Century Forerunner, Jules Regnault.”
Introduction to Financial Asset Pricing.” In Handbook European Journal of the History of Economic Thought 8(3),
in OR & MS 15. Edited by John R. Birge and Vadim 332–62.
Linetsky. Elsevier, The Netherlands.
Jarrow, Robert A., and Andrew Rudd, 1982. “Approximate Knoll, Michael S., 2008. “The Ancient Roots of Modern
Option Valuation for Arbitrary Stochastic Processes.” Financial Innovation: The Early History of Regulatory
Journal of Financial Economics 10(3), 347–69. Arbitrage.” Oregon Law Review 87(1), 93–116.
Jarrow, Robert A., and Stuart M. Turnbull, 1992. “Credit Krugman, Paul R., Maurice Obstfeld, and Marc Melitz,
Risk: Drawing the Analogy.” Risk Magazine 5(9). 2011. International Economics: Theory and Policy, 9th ed.,
Prentice-Hall, Upper Saddle River, N.J.
Jarrow, Robert A., and Stuart M. Turnbull, 1995. “Pricing
Derivatives on Financial Securities Subject to Credit
Lamont, Owen A., and Richard H. Thaler, 2003. “Can
Risk.” Journal of Finance 50(1), 53–85.
the Market Add and Subtract? Mispricing in Tech Stock
Jarrow, Robert A., and Stuart M. Turnbull, 2000. Derivative Carve-outs.” Journal of Political Economy 111(2), 227–68.
Securities. 2nd ed. South-Western, Cincinnati, Ohio.
Leland, Hayne, 1994. “Corporate Debt Value, Bond
Jarrow, Robert A., Haitao Li, and Feng Zhao, 2007. Covenants, and Optimal Capital Structure.” Journal of
“Interest Rate Caps ‘Smile’ Too! But Can the LIBOR Finance 49(4), 1213–52.
Market Models Capture the Smile?” Journal of Finance
Lintner, John, 1956. “Distribution of Incomes of
62(1), 345–82.
Corporations among Dividends, Retained Earnings,
Jarrow, Robert A., Jeff Oxman, and Yildiray Yildirim, and Taxes.” American Economic Review, Papers and
2010. “The Cost of Operational Risk Loss Insurance.” Proceedings 46(2), 97–113.
Review of Derivatives Research 13(3), 273–95.
Lintner, John, 1965. “The Valuation of Risk Assets and the
Jarrow, Robert A., Philip E. Protter, and Kazuhiro Shimbo, Selection of Risky Investments in Stock Portfolios and
2010. “Asset Price Bubbles in Incomplete Markets.” Capital Budgets.” Review of Economics and Statistics 47(1),
Mathematical Finance 20(2), 145–85. 13–37.
Jegadeesh, Narasimhan, 1993. “Treasury Auction Bids and Lutz, Friedrich A., 1940. “The Structure of Interest Rates.”
the Salomon Squeeze.” Journal of Finance 48(4), 1403–19. Quarterly Journal of Economics 55(1), 36–63.
Johnson, R. Stafford, 2004. Bond Evaluation, Selection, and Lynch, Peter (with John Rothchild), 1989. One Up on
Management. Blackwell, Cambridge, Mass. Wall Street: How to Use What You Already Know to Make
Johnston, Elizabeth T., and John J. McConnell, 1989. Money in the Market. Simon and Schuster, New York.
“Requiem for a Market: An Analysis of the Rise and
Fall of a Financial Futures Contract.” Review of Financial Macaulay, Frederick R., 1938. Some Theoretical Problems
Studies 2(1), 1–23. Suggested by the Movements of Interest Rates, Bond Yields
Johnston, Jack, and John DiNardo, 1997. Econometric and Stock Prices in the United States since 1856. National
Methods. 4th ed. McGraw-Hill, New York. Bureau of Economic Research, Cambridge, Mass.
Jones, E. Philip, Scott P. Mason, and Eric Rosenfeld, MacKenzie, Donald, and Yuval Millo, 2003. “Constructing
1984. “Contingent Claims Analysis of Corporate Capital a Market, Performing Theory: The Historical Sociology
Structures: An Empirical Investigation.” Journal of Finance of a Financial Derivatives Exchange.” American Journal of
39(3), 611–25. Sociology 109(1), 107–45.
Jordan, Bradford D., and Susan D. Jordan, 1996. “Salomon Malkiel, Burton G., 2003. A Random Walk Down Wall
Brothers and the May (1991) Treasury Auction: Analysis Street. 8th ed. W. W. Norton, New York.
of a Market Corner.” Journal of Banking and Finance 20(1), Malkiel, Burton G., and Richard E. Quandt, 1969. Strategies
25–40. and Rational Decisions in the Securities Options Markets.
Jorion, Philippe, 2000. “Risk Management Lessons from MIT Press, Cambridge, Mass.
Long-Term Capital Management.” European Financial Markowitz, Harry, 1952. “Portfolio Selection.” Journal of
Management 6(3), 277–300. Finance 7(1), 77–91.
Jovanovic, Franck, and Philippe Le Gall, 2001. “Does Marshall, Alfred, 1920. Principles of Economics. Macmillan
God Practice a Random Walk? The ‘Financial Physics’ and Co., London.
708 REFERENCES
Mehrling, Perry, 2005. Fischer Black and the Revolutionary Pan, Jun, 2002. “The Jump-Risk Premia Implicit in
Idea of Finance. John Wiley & Sons, Hoboken, N.J. Options: Evidence from an Integrated Time-Series
Merton, Robert C., 1973a. “Theory of Rational Option Study.” Journal of Financial Economics 63(1), 3–50.
Pricing.” Bell Journal of Economics and Management Science Park, Hun Y., and Andrew H. Chen, 1985. “Differences
4(1), 141–83. Between Futures and Forward Prices: A Further
Merton, Robert C., 1973b. “An Intertemporal Capital Investigation of Marking to Market Effects.” Journal of
Asset Pricing Model.” Econometrica 41(5), 867–87. Futures Markets 5(1), 77–88.
Merton, Robert C., 1974. “On the Pricing of Corporate Park, James A. 1799. A System of the Law of Marine Insurances.
Debt: The Risk Structure of Interest Rates.” Journal of Macmillan and Co., London.
Finance 29(2), 449–70. Park, Yoon S. 1984. “Currency Swaps as a Long-Term
Merton, Robert C., 1977. “An Analytic Derivation of International Financing Technique.” Journal of International
the Cost of Deposit Insurance and Loan Guarantees: An Business Studies 15(3), 47–54.
Application of Modern Option Pricing Theory.” Journal Paulson, Jr., Henry M., 2010. On the Brink: Inside the Race
of Banking and Finance 1(1), 3–11. to Stop the Collapse of the Global Financial System. Business
Merton, Robert C., and Myron S. Scholes, 1995. “Fischer Plus, New York.
Black.” Journal of Finance 50(5), 1359–70.
Miller, Merton H., 1986. “Financial Innovation: The Last Qin, Wang X. and Nick Ronalds, 2005. “China: The
Twenty Years and the Next.” Journal of Financial and Fall and Rise of Chinese Futures, 1990–2005.” Futures
Quantitative Analysis 21(4), 459–71. Industry Inside May/June,www.futuresindustry.org/fi-
Miller, Merton H., 1997. Merton Miller on Derivatives. John magazine-home.asp?a=1038.
Wiley, New York.
Miller, Merton H., and Franco Modigliani, 1961. Rebonato, Riccardo, 2002. Modern Pricing of Interest
“Dividend Policy, Growth, and the Valuation of Shares.” Rate Derivatives: The LIBOR Market Model and Beyond.
Journal of Business 34(4), 411–33. Princeton University Press, Princeton, N.J.
Miltersen, Kristian R., Klaus Sandmann, and Dieter Rendleman, Richard J., and B. Bartter, 1979. “Two State
Sondermann, 1997. “Closed Form Solutions for Term Option Pricing.” Journal of Finance 34(5), 1092–110.
Structure Derivatives with Log-Normal Interest Rates.” Rendleman, Richard J., and Christopher E. Carabini,
Journal of Finance 52(1), 409–30. 1979. “The Efficiency of the Treasury Bill Futures
Modigliani, Franco, and Merton H. Miller, 1958. “The Market.” Journal of Finance 34(4), 895–914.
Cost of Capital, Corporate Finance, and the Theory of Ross, Stephen A., Randolph W. Westerfield, and Jeffrey
Investment.” American Economic Review 48(3), 261–97. Jaffe, 2004. Corporate Finance. 7th ed. McGraw-
Modigliani, Franco, and Richard Sutch, 1966. “Innovations Hill/Irwin, New York.
in Interest Rate Policy.” American Economic Review
56(1/2), 178–97.
Samuelson, Paul A., 1965a. “Rational Theory of Warrant
Mood, Alexander M., Franklin A. Graybill, and Duance
Pricing.” Industrial Management Review 6, 13–39.
C. Boes, 1974. Introduction to the Theory of Statistics.
McGraw-Hill Higher Education, New York. Samuelson, Paul A., 1965b. “Proof That Properly
Anticipated Prices Fluctuate Randomly.” Industrial
Mossin, Jan, 1966. “Equilibrium in a Capital Asset Market.”
Management Review 6, 41–50.
Econometrica 35(4), 768–83.
Samuelson, Paul A., and Robert C. Merton, 1969. “A
Nelson, Samuel A., 1904. The A B C of Options and Complete Model of Warrant Pricing That Maximizes
Arbitrage. S. A. Nelson, New York. Utility.” Industrial Management Review 10(2), 17–46.
Sandmann, Klaus, Dieter Sondermann, and Kristian
Ofek, Eli, Matthew Richardson, and Robert F. Whitelaw, R. Miltersen, 1995. “Closed Form Term Structure
2004. “Limited Arbitrage and Short Sales Restrictions: Derivatives in a Heath Jarrow Morton Model with
Evidence from the Options Markets.” Journal of Financial LogNormal Annually Compounded Interest Rates.”
Economics 74(2), 305–42. Proceedings of the Seventh Annual European Research
Overby, Brian, 2007. The Options Playbook. TradeKing. Symposium, Bonn, September.
REFERENCES 709
Sarnoff, Paul, 1965. Russell Sage, the Money King. Sullivan, Joseph, 1988. The Way It Was: An Oral History of
Obolensky, New York. Finance 1967–1987. William Morrow & Company, Inc.,
Sawyer, Nick, 2007. “What’s the Score.” Risk Magazine, New York.
20(7), 25–29. Summers, Lawrence H., 1999. “Distinguished Lecture on
Scholes, Myron S., 1998. “Derivatives in a Dynamic Economics in Government: Reflections on Managing
Environment.” American Economic Review 88(3), 350–370. Global Integration.” Journal of Economic Perspectives 13(2),
3–18.
Schwager, Jack, 1992. “Bill Lipschutz: The Sultan of
Currencies.” In The New Market Wizards: Conversations Sundaresan, Suresh, 1984. “Consumption and Equilibrium
with America’s Top Traders. Harper Collins, New York. Interest Rates in Stochastic Production Economies.”
Journal of Finance 39(1), 77–92.
Sharpe, William F., 1964. “Capital Asset Prices: A Theory
of Market Equilibrium under Conditions of Risk.”
Journal of Finance 19(3), 425–42. Thorp, Edward O., and Sheen T. Kassouf, 1967. Beat the
Market: A Scientific Stock Market System. Random House,
Sharpe, William F., 1978. Investments. Prentice Hall, New York.
Englewood Cliffs, N.J.
Tumbridge and Co., 1875. Secrets of Success in Wall Street.
Sharpe, William F., 1995. “My Evolution as an Economist.” New York.
Address delivered at Trinity University, San Antonio,
Texas. In Lives of the Laureates: Thirteen Nobel Economists.
Varian, Hal R., 1992. Microeconomic Analysis. 3rd ed.W. W.
Edited by William Breit and Roger W. Spencer. MIT
Norton, New York.
Press, Cambridge, Mass.
Varian, Hal R., 2003. Intermediate Microeconomics: A Modern
Sharpe, William F., Gordon J. Alexander, and Jeffery V.
Approach. 6th ed. W. W. Norton, New York.
Bailey, 1999. Investments. Prentice Hall, Englewood
Cliffs, N.J. Vasicek, Oldrich A., 1977. “An Equilibrium
Characterization of the Term Structure.” Journal of
Shiller, Robert J., 1989. “The Term Structure of Interest
Financial Economics 5(2), 177–88.
Rates.” In Handbook of Monetary Economics. Edited by
Benjamin Friedman and Frank Hahn. North-Holland, Vega, Joseph Penso, de la, 1688. Confusión de Confusiones
Amsterdam. (Confusions and Confusions). Republished, Sonsbeek
Publishers, 1970, Arnhem, Netherlands.
Shreve, Steven E., 2004. Stochastic Calculus for Finance II:
Continuous Time Models. Springer Science + Business
Media, New York. West, Mark D., 2000. “Private Ordering at the World’s
Simon, Julian L., 1981. The Ultimate Resource. Princeton First Futures Exchange.” Michigan Law Review 98(8),
University Press, Princeton, N.J. 2574–615.
Solow, Robert M., 1956. “A Contribution to the Theory Wilmott, Paul, 1998. Derivatives: The Theory and Practice of
of Economic Growth.” Quarterly Journal of Economics Financial Engineering. John Wiley, Chichester, England.
70(1), 65–94.
Sowell, Thomas, 1995. The Vision of the Anointed: Self-
Congratulation as a Basis for Social Policy, Basic Books (a
member of the Perseus Books Group), New York.
Notation
A(t) Value of a money market account that starts with F(t) Futures price at time t.
an investment of $1 and earns the default-free spot g Storage cost (annual rate).
rate each period (chapters 1–25) or the value of a
G Storage cost (lump sum).
firm’s assets (chapter 26).
h The minimum-variance futures hedge ratio.
b(t) Basis (≡ spot price minus futures price) at time t.
i Simple (“domestic” or “dollar”) interest rate per
B(t,T) The price at time t of a zero-coupon bond paying
year; iE is (“European” or “foreign”) interest rate
$1 at time T (sometimes shortened as B(t) or B).
per year.
BE Zero-coupon bond price in a foreign country (in
I Index.
“European terms”).
ifut Futures interest rate.
C Cash flow or a bond’s coupon.
iS Swap rate.
c European call price or caplet price (chapters 1–25)
or the market clearing CDS rate (chapter 26). i(t,T) Simple forward rate considered at time t, which
becomes effective at time T.
cA American call price.
log Natural logarithm.
corr Correlation coefficient.
L Principal or par value of a bond.
cov Covariance.
L(t) Value of a debt issue at time t.
𝛿 Dividend yield (chapters 1–24), or the time
interval for computing a simple forward rate LN Notional principal.
(chapter 25), or the recovery rate (chapter 26). 𝜆 Probability of default per unit time (in reduced-
div Dollar dividend. form model). 𝜆̃ is the risk neutral default
probability.
dur Duration of a bond.
𝜇 Expected return (chapters 1–24) or expected
durM Modified duration of a bond.
percentage change in the simple forward rate
Δ (Delta) Change in a variable’s value. (chapter 25).
e Exponential function. N(x) Cumulative standard normal distribution function
E(t) A firm’s equity value at time t. with mean 0 and standard deviation 1.
E( . ) Computing expectation of the expression inside N ′ (x) Standard normal density function.
parentheses. p European put price (chapters 1–20) or floorlet
E𝜋 ( . ) Computing expectation of the expression inside price (chapters 21–26).
parentheses using the pseudo-probabilities denoted pA American put price.
by 𝜋.
P Bond price.
fFRA Forward rate agreement rate.
Prob Probability.
f (t,T) Forward rate contracted at time t that is effective
Prob𝜋 Probability that a random variable takes on certain
over the time period [T,T + 1].
values when 𝜋 is the underlying probability.
F(t,T) Forward price at time t for a contract maturing at
𝜙 Term for risk adjustment in pseudo-probability.
time T (sometimes shortened as F).
𝜋 Pseudo-probability that a stock price moves up.
FA Forward price of a non-US currency in American
or dollar terms. 𝜋(t,T) Pseudo-probability that a T-maturity zero-coupon
bond moves up at time t.
FE Forward price of a non-US currency in European
terms or in terms of a currency other than the US PV Present value.
dollar. q Actual probability that a stock price moves up.
710
NOTATION 711
q(t) Actual probability that the term structure of zero- SCDIV Stock price cum-dividend.
coupon bond prices moves up at time t. SXDIV Stock price ex-dividend.
r Continuously compounded interest rate per year. sd Standard deviation.
rE Foreign interest rate (interest rate in “European”
𝜎 Standard deviation (volatility) of stock return per
terms).
year.
R(0,T) Dollar return at time T from investing $1 in risk-
𝜃 Stock’s risk premium.
free bonds today (chapters 1–20).
𝜐 Average forward rate volatility.
R(t) Default-free spot rate of interest declared at time t
that applies to the time period [t,t + 1]. U, D Up and down factor, respectively, in a singleperiod
or multiperiod binomial option pricing model.
RE “Dollar return” concept applied to investment in
foreign currencies (in European terms or in terms u(t,T) Up and down factor, respectively, in an HJM
of a currency other than the US dollar). d(t,T) multiperiod binomial interest rate option pricing
Ref Reference rate (bbalibor or euribor in most cases) model.
announced on some future date. var Variance.
S(t) Stock price or spot price of a commodity at VaR𝛼 Value-at-risk for an 𝛼 percent loss probability.
time t. y The convenience yield per year (chapter 12), or
SA Spot exchange rate in American terms. the yield on a coupon-bond (chapters 21–26).
SE Spot exchange rate in European terms. Y Lump sum convenience yield.
Additional Sources and Websites
Journals
Academic journals are the best source for research on derivatives and risk manage-
ment. They include the following.
General Finance Journals
■ Annual Review of Financial Economics, www.annualreviews.org/journal/financial
■ Finance and Stochastics, www.springer.com/mathematics/quantitative+finance/
journal/780
■ Financial Analysts Journal, www.cfapubs.org/loi/faj
■ Financial Management, https://www.fma.org/financial-management
■ Journal of Banking & Finance, www.journals.elsevier.com/journal-of-banking-
and-finance/
■ Journal of Finance, www.afajof.org
■ Journal of Financial and Quantitative Analysis, www.jfqa.org
■ Journal of Financial Economics, http://jfe.rochester.edu
■ Mathematical Finance, https://onlinelibrary.wiley.com/journal/14679965
■ Review of Financial Studies, http://rfs.oxfordjournals.org
Derivatives and Risk Management Journals
■ Journal of Derivatives, https://jod.iijournals.com/
■ Journal of Fixed Income, https://jfi.iijournals.com/
■ Journal of Futures Markets, http://onlinelibrary.wiley.com/journal/10.1002/%28
ISSN%291096-9934
■ Journal of Risk, www. thejournalofrisk.com
■ Review of Derivatives Research, www.springer.com/business+%26+management/finance/
journal/11147
■ Risk Magazine (publishes mathematical finance articles), www.risk.net/risk-
magazine
Economics Journals with Occasional Articles on
Derivatives and Risk Management
■ American Economic Review, www.aeaweb.org/aer/index.php
■ American Journal of Agricultural Economics, http://ajae.oxfordjournals.org/
■ Econometrica, www.wiley.com/WileyCDA/WileyTitle/productCd-ECTA.html
■ Journal of Economic Perspectives, www.aeaweb.org/jep/
ADDITIONAL SOURCES AND WEBSITES 713
■ Journal of Economic Surveys, www.wiley.com/WileyCDA/WileyTitle/productCd-
JOES.html
■ Journal of Political Economy, www.press.uchicago.edu/ucp/journals/journal/jpe.html
■ Review of Economics and Statistics, www.mitpressjournals.org/loi/rest
Mathematics, Statistics, and Management Journals
with Occasional Articles on Derivatives and Risk
Management
■ Annals of Applied Probability, www.imstat.org/aap/
■ Journal of the American Statistical Association, www.amstat.org/publications/jasa.cfm
■ Stochastic Processes and Their Applications, www.journals.elsevier.com/stochastic-
processes-and-their-applications/
Newspapers and Magazines
Business newspapers and magazines frequently write about derivatives and risk
management.
■ Barron’s, http://online.barrons.com
■ Bloomberg, www.bloomberg.com
■ The Economist, www.economist.com
■ Financial Times, www.ft.com
■ Futures and Options World, www.fow.com
■ Harvard Business Review, http://hbr.org/magazine
■ Investor’s Business Daily, www.investors.com
■ New York Times, www.nytimes.com
■ Wall Street Journal, http://online.wsj.com
Government Agencies
Nowadays, many government agencies have websites; examples follow.
■ US Department of the Treasury, www.treasury.gov
■ Bank for International Settlements, www.bis.org
■ Board of Governors of the Federal Reserve System, www.federalreserve.gov
■ Federal Reserve Bank of New York, www.newyorkfed.org
■ International Monetary Fund (IMF), www.imf.org
■ US Securities and Exchange Commission, www.sec.gov
■ US Department of the Treasury, www.treasury.gov
■ World Bank Group, www.worldbank.org
714 ADDITIONAL SOURCES AND WEBSITES
Industrial Associations and Bodies,
Self-Regulatory Organizations
Many industry associations and bodies as well as self-regulatory organizations dissem-
inate information through websites.
■ CFA Institute, www.cfainstitute.org
■ Financial Accounting Standards Board (FASB), www.fasb.org
■ Financial Industry Regulatory Authority (FINRA), www.finra.org
■ Global Association of Risk Professionals (GARP), www.garp.org
■ International Association of Financial Engineers (IAFE), http://iafe.org
■ International Swaps and Derivatives Association Inc. (ISDA), www2.isda.org
■ National Futures Association (NFA), www.nfa.futures.org
■ Securities Industry and Financial Markets Association (SIFMA), www.sifma.org
Exchanges
Most exchanges have websites that include a description of their business, a historic
timeline, contract specifications, price and volume information, and educational and
promotional material.
■ BM&FBovespa, http://www.b3.com.br/en_us/
■ Chicago Board Options Exchange (CBOE), www.cboe.com
■ CME Group, www.cmegroup.com
■ Eurex (includes ISE), www.eurexchange.com and www.ise.com
■ Korea Exchanges (KRX), http://eng.krx.co.kr
■ Multi Commodity Exchange of India (MCX), www.mcxindia.com
■ NASDAQ OMX, www.nasdaqomx.com
■ National Stock Exchange of India (NSE), www.nseindia.com
■ NYSE Euronext, https://nyse.nyx.com
■ Russian Trading Systems Stock Exchange (RTS), www.rts.ru/en/
Financial Institutions
Predictions on market directions and economic conditions, advice on trading
strategies, and risk management–related issues are found in bulletins, newsletters, and
research reports from a wide variety of sources, including banks, brokerage firms,
commodity trading advisors, credit rating agencies, finance companies (including
mutual funds), and hedge fund managers. These can be found on their respective
websites.
Books on Derivatives and Risk Management
Following is a representative list of books on derivatives and risk management.
Abu-Mostafa, Yaser S., Blake LeBaron, Andrew W. Lo, and Chaplin, Geoff, 2010, Credit Derivatives: Trading, Investing,
Andreas S. Weigend (eds.), 2000, Computational Finance and Risk Management.
1999. Choudhry, Moorad, 2010, The Repo Handbook, 2nd edition.
Acharya, Viral V., Thomas F. Cooley, Matthew P. Richard- Chriss, Neil A., 1997, Black–Scholes and Beyond: Option
son, and Ingo Walter (eds.), 2010, Regulating Wall Street: Pricing Models.
The Dodd-Frank Act and the New Architecture of Global
Culp, Christopher L., 2006, Structured Finance and Insurance:
Finance.
The ART of Managing Capital and Risk, 2nd edition.
Andersen, Leif B. G., and Vladimir V. Piterbarg, 2010,
Das, Satyajit, 2005, Credit Derivatives: CDOs and Structured
Interest Rate Modeling, 3 volumes.
Credit Products, 3rd edition.
Augen, Jeffrey, 2011, Microsoft Excel for Stock and Option
DeRosa, David F., 1998, Currency Derivatives: Pricing Theory,
Traders: Build Your Own Analytical Tools for Higher Returns.
Exotic Options, and Hedging Applications.
Back, Kerry, 2010, A Course in Derivative Securities: Introduc-
Dixit, Avinash K., and Robert S. Pindyck, 1994, Investment
tion to Theory and Computation.
under Uncertainty.
Barker, Philip, 2007, Java Methods for Financial Engineering:
Duffie, Darrell, 1989, Futures Markets.
Applications in Finance and Investment.
Duffie, Darrell, and Kenneth J. Singleton, 2003, Credit Risk:
Baxter, Martin, and Andrew Rennie, 1996, Financial Calcu-
Pricing, Measurement, and Management.
lus: An Introduction to Derivatives Pricing.
Bielecki, Tomasz, Damiano Brigo, and Frederic Patras, Edwards, Franklin R., and Cindy W. Ma, 1992, Futures and
2011, Credit Risk Frontiers: Subprime Crisis, Pricing and Options.
Hedging, CVA, MBS, Ratings, and Liquidity. Embrechts, Paul, Claudia Klüppelberg, and Thomas
Bouzoubaa, Mohamed, and Adel Osseiran, 2010, Exotic Mikosch, 1997, Modelling Extremal Events for Insurance and
Options and Hybrids: A Guide to Structuring, Pricing and Finance.
Trading. Fabozzi, Frank J. (ed.), 2005, The Handbook of Mortgage
Boyle, Phelim, and Feidhlim Boyle, 2001, Derivatives: The Backed Securities, 6th edition.
Tools That Changed Finance. Fabozzi, Frank J. (ed.), 2011, The Handbook of Fixed Income
Brandimarte, Paolo, 2006, Numerical Methods in Finance and Securities, 8th edition.
Economics: A MATLAB-Based Introduction, 2nd edition. Fabozzi, Frank J., Roland Fuss, and Dieter G. Kaiser, 2008,
Brigo, Damiano, and Fabio Mercurio, 2006, Interest Rate The Handbook of Commodity Investing.
Models: Theory and Practice: With Smile, Inflation and Credit. Fabozzi, Frank J., and Vinod Kothari, 2008, Introduction to
Butler, Cormac, 2009, Accounting for Financial Instruments. Securitization.
Chacko, George, Vincent Dessain, Peter Hecht, and Anders Figlewski, Stephen, William L. Silber, and Marti G. Subrah-
Sjoman, 2006, Financial Instruments and Markets: A Case- manyan, 1990, Financial Options: From Theory to Practice.
book. Flavell, Richard, 2010, Swaps and Other Derivatives.
Chacko, George, Anders Sjöman, Hideto Motohashi, and Fouque, Jean-P., George Papanicolaou, Ronnie Sircar, and
Vincent Dessain, 2006, Credit Derivatives: A Primer on Knut Sølna, 2011, Multiscale Stochastic Volatility for Equity,
Credit Risk, Modeling, and Instruments. Interest Rate, and Credit Derivatives.
715
716 BOOKS ON DERIVATIVES AND RISK MANAGEMENT
Froot, Kenneth A. (ed.), 1999, The Financing of Catastrophe McLaughlin, Robert M., 1998, Over-the-Counter Derivative
Risk. Products: A Guide to Legal Risk Management and Documen-
Geman, Hélyette, 2005, Commodities and Commodity Deriva- tation.
tives: Modelling and Pricing for Agriculturals, Metals and McMillan, Lawrence G., 2012, Options as a Strategic Invest-
Energy. ment, 5th edition.
Glasserman, Paul, 2003, Monte Carlo Methods in Financial McNeil, Alexander J., Rüdiger Frey, and Paul Embrechts,
Engineering. 2005, Quantitative Risk Management: Concepts, Techniques,
Haug, Espen G., 2006, The Complete Guide to Option Pricing and Tools.
Formulas, 2nd edition. Miller, Merton H., 1997, Merton Miller on Derivatives.
Ho, Thomas S. Y., and Sang-Bin Lee, 2004, The Oxford Natenberg, Sheldon, 1995, Option Volatility and Pricing:
Guide to Financial Modeling: Applications for Capital Mar- Advanced Trading Strategies and Techniques, 2nd edition.
kets, Corporate Finance, Risk Management, and Financial Pachamanova, Dessislava, and Frank J. Fabozzi, 2010, Simu-
Institutions. lation and Optimization in Finance + Website: Modeling with
Hughston, Lane P. (ed.), 1997, Vasicek and Beyond: MATLAB, @Risk, or VBA.
Approaches to Building and Applying Interest Rate Models. Pearson, Neil D., 2002, Risk Budgeting: Portfolio Problem
Hughston, Lane P. (ed.), 2000, The New Interest Rate Models: Solving with Value-at-Risk.
Recent Developments in the Theory and Application of Yield Rebonato, Riccardo, 1999, Volatility and Correlation: In the
Curve Dynamics. Pricing of Equity, FX and Interest-Rate Options.
Hull, John C., 2011, Options, Futures, and Other Derivatives, Rebonato, Riccardo, 2002, Modern Pricing of Interest-Rate
8th edition. Derivatives: The LIBOR Market Model and Beyond.
Jarrow, Robert A. (ed.), 1998, Volatility: New Estimation Rendleman, Richard J., 2002, Applied Derivatives: Options,
Techniques for Pricing Derivatives. Futures and Swaps.
Jarrow, Robert A., 2002, Modeling Fixed Income Securities and Ritchken, Peter, 1996, Derivative Markets: Theory, Strategy,
Interest Rate Options, 2nd edition. and Applications.
Jarrow, Robert A., 2008, Financial Derivatives Pricing: Selected Ronn, Ehud I., 2003, Real Options and Energy Management:
Works of Robert Jarrow. Using Options Methodology to Enhance Capital Budgeting
Jarrow, Robert A., and Andrew Rudd, 1983, Option Pricing. Decisions.
Jarrow, Robert A., and Stuart M. Turnbull, 2000, Derivative Saunders, Anthony, and Linda Allen, 2010, Credit Risk Man-
Securities, 2nd edition. agement In and Out of the Financial Crisis: New Approaches
Jha, Siddhartha, 2011, Interest Rate Markets: A Practical to Value at Risk and Other Paradigms, 3rd edition.
Approach to Fixed Income. Schönbucher, Philipp J., 2003, Credit Derivatives Pricing
Jorion, Philippe, 2006, Value at Risk: The New Benchmark for Models: Model, Pricing and Implementation.
Managing Financial Risk, 3rd edition. Shiller, Robert J., 2004, The New Financial Order: Risk in the
Kat, Harry M., 2001, Structured Equity Derivatives: The 21st Century.
Definitive Guide to Exotic Options and Structured Notes. Shimko, David C. (ed.), 2004, Credit Risk: Models and
Lindsey, Richard R., and Barry Schachter, 2009, How I Management, 2nd edition.
Became a Quant: Insights from 25 of Wall Street’s Elite. Shreve, Steve E., 2005, Stochastic Calculus for Finance I: The
Lipton, Alexander, and Andrew Rennie, 2011, The Oxford Binomial Asset Pricing Model; Stochastic Calculus for Finance
Handbook of Credit Derivatives. II: Continuous-Time Models.
Mason, Scott P., Robert C. Merton, Andre F. Perold, and Siegel, Daniel R., and Diane F. Siegel, 1990, Futures
Peter Tufano, 1995, Cases in Financial Engineering: Applied Markets.
Studies of Financial Innovation. Smithson, Charles W., 1998, Managing Financial Risk: A
McDonald, Robert L., 2009, Derivatives Markets, 3rd Guide to Derivative Products, Financial Engineering, and Value
edition. Maximization, 3rd edition.
BOOKS ON DERIVATIVES AND RISK MANAGEMENT 717
Stigum, Marcia L., and Anthony Crescenzi, 2007, Stigum’s Van Deventer, Donald R., and Kenji Imai, 2003, Credit Risk
Money Market, 4th edition. Models and the Basel Accords.
Sundaresan, Suresh, 2013, Fixed Income Markets and Their Van Deventer, Donald R., Kenji Imai, and Mark Mesler,
Derivatives, 4th edition. 2013, Advanced Financial Risk Management: Tools and Tech-
Tavakoli, Janet M., 2008, Structured Finance and Collateralized niques for Integrated Credit Risk and Interest Rate Risk
Debt Obligations: New Developments in Cash and Synthetic Managements, 2nd edition.
Securitization, 2nd edition. Veronesi, Pietro, 2010, Fixed Income Securities: Valuation,
Tuckman, Bruce, and Angel Serrat, 2011, Fixed Income Risk, and Risk Management.
Securities: Tools for Today’s Markets, 3rd edition.
Van Deventer, Donald R., and Kenji Imai, 2013, Financial
Risk Analytics: A Term Structure Model Approach for Banking,
Insurance and Investment Management.
Name Index
Alexander, Gordon J., 67 Fama, Eugene, 423 Kreps, David M., 370, 371
Alletzhauser, Albert J., 155 Feynman, Richard, 368 Kritzman, Mark P., 134
Amemiya, Takeshi, 281n6 Fischel, Daniel R., 203
Amin, Kaushik I., 296 Fisher, Irving M., 504 Lamont, Owen A., 343
Arrow, Kenneth, 8 Fisk, James, 202 Lee, Sang-Bin, 565n2, 595
Artzner, Philippe, 667 Follmer, Holmer, 667 Leeson, Nick, 260n1
Froot, Kenneth A., 259 Leland, Hayne, 672
Bachelier, Louis, 369, 370 Lintner, John, 8
Bailey, Jeffery V., 67 Gastineau, Gary L., 134 Lipschutz, Bill, 311
Bank, Peter, 440 Gelderblom, Oscar, 291 Lutz, Friedrich A., 504
Barings, Francis, 260n1 Glasserman, Paul, 94, 448 Lynch, Peter, 10
Bartter, Brit J., 369, 370, 371, 407 Gould, Jay, 202
Bakshi, Gurdip, 479 Greenbaum, Stuart I., 540 Ma, Cindy W., 83
Baum, Dietmar, 440 Greenspan, Alan, 10, 295 Macaulay, Frederick R., 507
Baxter, Martin W., 371n1, 564 Gross, David, 10 MacKenzie, Donald, 305
Bernstein, Peter L., 656 Madoff, Bernard, 201
Black, Fischer S., 8, 12, 312, 368, 370, Harrison, J. Michael, 370, 371 Malkiel, Burton G., 291, 305
383, 423, 424, 429, 620–22, 647 Harte, John, 169 Malthus, Thomas, 168
Blanchard, Olivier, 20 Heath, David C., 131n1, 360n3, Markowitz, Harry, 8
Box, George E. P., 679 565, 621 Mellon, Andrew William, 23
Brace, Alan, 565, 622, 648 Hicks, John, 8, 504 Meriwether, John, 677
Brealey, Richard, 507 Higgins, Leonard R., 337 Merton, Robert C., 8, 12, 368, 370,
Buffett, Warren, 10, 204, 312, 356 Ho, Thomas S. Y., 595 371, 383, 423, 426, 669, 670
Holden, Craig W., 490 Miller, Hilton D., 453
Çetin, Umut, 677 Holdren, John, 169 Miller, Merton, 7, 8, 10, 342
Chatterjea, Arkadev, 311, 678 Honkapohja, Seppo, 476n3 Millo, Yuval, 305
Coase, Ronald, 7–9 Hughston, Lane P., 565n2 Miltersen, Kristian R., 565, 622, 648
Cox, David R., 453 Modigliani, Franco, 8, 504
Cox, John C., 369, 370, 371, 407, 504 James, William, 311n2 Mood, Alexander M., 453
Culbertson, John M., 504 Jarrow, Robert A., 131n1, 185, 215, Morgenbesser, Sidney, 320
244n2, 272, 296, 311, 357n2, Morton, Andrew J., 565, 621
Debreu, Gerard, 8 360n3, 361, 369–71, 375, 384, 407, Mossin, Jan, 8
DiNardo, John, 281n6 425, 428, 440, 449, 457, 473, 504, Myers, Stewart, 507
Dixit, Avinash K., 673 565, 584, 593, 606, 607, 614, 615,
Draper, Norman R., 679 621, 626, 636, 647, 649, 667–69, Nelson, Charles R., 474n2
Drew, Daniel, 314 673, 675, 676, 678, 679 Nelson, Samuel A., 428, 463
Duffie, Darrell, 94, 272, 448 Jegadeesh, Narasimhan, 204 Nocera, Joe, 656
Dunn, Kenneth B., 504 Johnson, R. Stafford, 540
Johnston, Jack, 281n6 Oppenheimer, Ernest, 196
Easterbrook, Frank H., 203 Jones, E. Philip, 672 Overby, Brian, 305
Edelman, Asher B., 114 Jonker, Joost, 291
Edwards, Franklin R., 83 Jordan, Bradford D., 204 Pan, Jun, 479
Ehrlich, Paul R., 168, 168–69 Jordan, Susan D., 204 Pindyck, Robert S., 673
Eisenberg, Larry, 473 Pliska, Stanley R., 370, 371
Engle, Robert F., 312, 480 Keynes, John Maynard, 8, 188 Ponzi, Charles, 201
Evans, George W., 476n3 Knoll, Michael, 342 Protter, Philip E., 357n2, 425
718
NAME INDEX 719
Qin, Wang X., 86 Sandmann, Klaus, 565, 622, 648 Thakor, Anjan V., 540
Quandt, Richard E., 291 Sargent, Thomas, J., 476n3 Thaler, Richard H., 343
Sarnoff, Paul, 292 Thales of Miletus, 292
Rebonato, Riccardo, 636 Schied, Alexander, 667 Tobin, James, 8
Rendleman, Richard J., 369, 370, Scholes, Myron S., 8, 12, 368, Turnbull, Stuart M., 185, 215, 272, 370,
371, 407 370, 383, 423, 429 371, 407, 669, 673
Rennie, Andrew J. O., 371n1, 564 Schwager, Jack, 311n1
Ronalds, Nick, 86 Sharpe, William F., 8, 67, Varian, Hal R., 193
Rosenberg, Joshua V., 480 370, 371 Vasicek, Oldrich A., 371, 565, 595
Ross, David J., 203 Shimbo, Kazuhiro, 425
Ross, Stephen A., 369, 672 Singleton, Kenneth J., 504 Weatherstone, Dennis, 662
Rubinstein, Mark, 369, 423 Sowell, Thomas, 268 Wessel, David, 5
Rudd, Andrew, 369, 370, 371, 407 Sullivan, Joseph, 306 West, Mark D., 155
Summers, Lawrence H., 289, 312 Wilmott, Paul, 94, 448
Sage, Russell, 292, 337 Sundaresan, Suresh, 504
Samuelson, Paul A., 8, 370, 423, 426 Sutch, Richard, 504 Zhu Rongji, 289, 312
Subject Index
ABS CDSs (asset-backed security credit Arbitrage (arb), 112–27; see also for forwards with market
default swaps), 148 No-arbitrage principle imperfections, 250
ABSs, see Asset-backed securities and cost-of-carry model, 216 and options price restrictions, 346,
Active management, duration, 512 defined, 6 349–52, 355
Actual probabilities examples of, 116 and put–call parity, 340
and binomial HJM model, 576 index, 121–23 for zero-value portfolios, 218
and binomial options pricing model, for indexes with dividends, 238 Arbitrageurs, 57
382–83, 399 and market efficiency, 118–19 Arbitration, 56
and BSM model, 433–34, 457–58 tax (regulatory), 342 Ask price
and Eurodollar futures rates, 609–10 in trading strategies, 119–23 and banker’s discount yield, 46–47
and HJM libor model, 637 Arbitrage-free conditions of stock, 56
pseudo- and, 382–83, 385–87, binomial trees under, 378, 573 of T-bills, 45–47
433–34, 457–58, 576, 638 futures prices under, 608–9 Asset allocation, portfolio risk, 14
and risk premium, 385–88 futures rates under, 609–10 Asset-backed securities (ABSs), 38, 498
Actuaries, 327 for single-period binomial HJM Asset-backed securities credit default
Adjustable-rate mortgage (ARM), 498 model, 570–73 swaps (ABS CDSs), 148
Airline fuel costs, hedging of, 146, 147, for two-period binomial HJM model, Assets, 14
260, 261 597–99 cash flows and values of, 131–32
Airline miles, arbitrage with, 116 Arbitrage opportunities, 63, 113–16 consumption, 214
Alternative trading systems (ATSs), with American puts, 353 with continuously compounded
61–62 and binomial HJM model, 566 dividend yield, 64
American options and binomial options pricing models, current, 658
arbitrage opportunities with, 353–54 375 financial, 7
binomial pricing model for, 410–12 with caplets, 579 intrinsic value of, 215, 375, 424
Black–Scholes–Merton model for, with European puts, 352–53 investment, 214
447–48 and HJM libor model, 623 long-term, 658
early exercise of, 357–62 illegal, 124–26 production, 214
exercise of, 94 Arbitrage profits real, 7
HJM libor model for, 647 in binomial options pricing model, Assignment, in swap closing, 544
price bounds for, 99–100, 104, 372–73 Associated persons (APs), 173
346–48, 352 for commodities with storage Asymmetric information, 677
profit diagrams for, 313 costs/convenience yields, 244 At-the-money (term), 95, 101
put–call parity for, 344–46 for foreign currency forwards, 242 Auctions, Treasuries, 38–39
value of European vs., 347 and forward prices, 220 Australian Options Market, 295
American Stock Exchange, 295 for forwards with fixed dollar Automated trading, 58, 295
American terms, currency quotes, dividends, 233 Average forward rate volatility, 632–36
138 and law of one price, 212–13
Amortized swaps, 549 and put–call parity, 341 Backward induction, 396–97, 601
Amsterdam, Netherlands, 157 for synthetic indexes, 238 Backwardation, 246–49
Amsterdam Stock Exchange, 295 Arbitrage tables Balance sheet, 15, 16, 658, 659, 670
Annual Percentage Rate (APR), 25 for commodities with storage Banging the close, 204–5
Annualized rate of return, 24–25, 45, 46 costs/convenience yields, 245 Bankers, financial engineering by, 135
Annualized volatility, 443 for cost-of-carry model, 220–21 Banker’s discount yield, 46, 47
Appreciation, US dollar, 138 for currency forward price, 243 Bankers Trust, 539, 540
APR (Annual Percentage Rate), 25 for forward price of dividend-paying Bankhaus Herstatt, 669
APs (associated persons), 173 stock, 235 Barings Bank, 156, 260n1
720
SUBJECT INDEX 721
Barrier options, 449 stock price path in, 393–94 Eurobonds, 8
Basel Committee on Banking and stock price tree, 395–96 extendible, 130
Supervision, 12–13, 17, 668–78 two-period, 396–402 inflation-indexed, 43
Basis, 185, 187, 264 Binomial process assumption, 376, 567, modified duration for, 508, 509
Basis point (bp), 37 593 Treasury, 43–45, 530
Basis risk, 264–67 Binomial stock price evaluation, 393–94 Boomsters, 168–69
BBA, see British Bankers’ Association Binomial theorem, 687–88 BOP (binomial option pricing) data,
libor vs. libor rate index, 47–48 Binomial trees 394–95
Bear raids, 124 arbitrage-free conditions for, 573, 597 Borrowing, role of derivatives in, 5
Bear spreads, 321–23 in Microsoft Excel, 413 Borrowing rates, 251
Bearish trades, 65, 93 in options pricing model, 377–78 Bottom straddle, 330
BEP (breakeven point), 316 Black Monday, 316 BOVESPA, 295
Bid/ask (bid/offer) spread, 56, 251 Black’s model of caplet pricing, 620, 621 Bp (basis point), 37
Bid price, 18, 46, 56 Black–Scholes–Merton (BSM) model, 8, Breakeven point (BEP), 313
Binomial (term), 369, 687 12, 94, 422–59, 462–87, 621 Bretton Woods system, 6
Binomial coefficients, 403, 688 actual vs. pseudo-probabilities for, Broker loan rate, 67
Binomial Heath–Jarrow–Morton 433–34 Brokerage commissions, 249, 251
(binomial HJM) model, 558–89, for American options, 447–48 Brokers
592–615 assumptions with, 424–28 in futures market, 173–74
assumptions with, 565–70, 593–96 and binomial options pricing model, introducing, 173
caplet pricing with, 620–22 369, 407–8 in stock market, 56–57
equal pseudo-probability condition calibration of, 474–80 of swaps, 135
for, 575–76, 586–87 continuous compounding in, 34 Broker’s call, 67
for forwards, 607–8 derivations of, 454–56, 486–87 BSM model, see Black–Scholes–Merton
for futures, 608–10 with dividends, 444–45 model
and history of interest rate derivative for European options, 486–87 Budget deficit, 38
models, 564–65 for exotic options, 448–49 Build and break approach, 130–31
in history of options pricing, 369–71 for foreign currency options, 446–47 Bull spread, 321–22, 324, 326
and HJM libor model, 613–15 formula, 429–32 Bullish trades, 65, 93
and interest rate derivatives, 560–64 Greeks in, 434–40 Butterfly spread, 325, 326
multiperiod, 607 hedging limitation of, 480–81 Buyers
single-period, 570–85 and hedging of Greeks, 463–69, options, 93, 95–98
for swaptions, 611–12 482–87 swap, 545
Treasuries in, 37 and hedging portfolios, 469–70 Buying (long) hedges, 89, 266
two-period, 597–606 in history of options pricing/trading, Buying the spread, 187
Binomial interest rate derivatives pricing 291, 370–71 Buyout, swap, 544
model, 384 HJM libor model vs., 630–31 Buy–write options, 317, 318
Binomial option pricing (BOP) data, inputs for, 440–44 BXM (CBOE S&P 500 BuyWrite
394–95 and lognormal assumption for stock Index), 317
Binomial options pricing model, prices, 451–59
367–88, 392–418 martingales in, 432–33 Calendar spread, see Time spread
for American options, 410–12 and pricing/hedging argument, Calibration, 444, 474–80
applications and uses of, 368–69 428–29 defined, 636
assumptions with, 374–76 risk-neutral valuation in, 433 and hedging errors, 476–77
BOP data in, 394–95 and vega hedging, 470–73, 485–86 and implied volatilities, 477–80
example, 372–74 Block trades, 60 and theoretical vs. econometric
and history of options pricing, Bond-equivalent yield, T-bill, 45–47 models, 474–75
369–71 Bonds; see also Zero-coupon bonds using, 475–77
multiperiod, 403–9 (zeroes) Call backspread, 327n5
and no-arbitrage argument, 384–88 callable, 131 Call loan rate, 67
for options with dividends, 409–10 cheapest-to-deliver, 530 Call options (calls), 95–100; see also
single-period, 376–83 convertible, 130 Put–call parity
spreadsheet programs for coupon, 43, 519–20 American, 99–100, 344–46, 360–62
implementing, 413–17 defined, 7 arbitrage opportunities with, 350
722 SUBJECT INDEX
binomial pricing model for, 372–74, Caplets, 559–61, 605 CFTC, see Commodity Futures Trading
400, 401 delta hedging of, 643–45 Commission
BSM model for, 431 delta of, 642 Chase Manhattan Bank, 540
on commodity futures, 296 gamma hedging of, 646–47 Cheapest-to-deliver bonds, 530
covered, 313, 317–20 gamma of, 642 Chicago Board of Trade (CBOT; CBT)
on currency, 446–47 hedge ratio for, 581 futures regulation at, 196–97
defined, 93 maturity of, 577 goal of, 12, 259
delta and gamma for, 437, 438 payoffs for, 562 history of, 157–59
delta hedging errors with, 465, 468–69 risk-neutral valuation of, 581, 602–4, interest rate derivatives of, 6, 520–21
down-and-out, 449 637–38 trading at, 83
early exercise of, 360–62 Caps, 559, 560, 638–41 Chicago Board Options Exchange
European, 347, 352, 372–74, 400, hedging interest rate risk with, 562–63 (CBOE), 105, 290
415–16, 431, 438, 446, 447 HJM libor model for, 638–41 call option example, 105–6
exchange-traded, 105–6 payoffs for, 563 opening of, 293
on indexes, 445–46 speculation with, 627–28 trading on, 296
on interest rates, see Caplets on swaps, 549 and VIX, 294, 480
intrinsic and time values of, 98–99 Carrying charges, 210 Chicago Mercantile Exchange (CME;
long, 298, 315–16, 318, 319, Cartels, 196 the Merc)
344–45, 467 Cash-and-carry argument, 210 futures trading at, 83
payoffs and profit diagrams for, 95–98 Cash and margin accounts, see Margin history of, 158
price bounds for, 99–100, 347–49 accounts IMM division, 6, 159
pricing, with Microsoft Excel, 414–16 Cash dividends, 62 interest rate derivatives on, 521
risk-neutral pricing for, 401 Cash flows China, futures trading in, 86
short, 314–18, 344–45 and asset values, 131–32 China Financial Futures Exchange, 86
and strike price, 350, 352 compounding and discounting of, Clearance fees, 166
synthetic, 372, 397–400 31–37 Clearing (of trades)
Call ratio spread, 327n5 in currency swaps, 139, 142–43 for call options, 105–6
Call spreads for dividend-paying stock, 234 for futures, 84
bearish, 323 of FRAs, 522 for securities, 59, 61
bull, 321–22, 324, 326 in interest rate swaps, 136–38 Clearing members, 59
one-by-two, 325 intermediate, 215, 375, 424, 566, 623 Clearinghouses, 83, 84, 158, 166–67
and reinsurance, 327 in margin accounts, 179–80 Closed-end companies, 120
Callable bonds, 131 present value of, 35 Closed-end funds, 119–20
Callable swaps, 550 Cash prices, see Spot prices Closed-form solutions, 94n1, 641
Cancel orders, 108 Cash-secured puts, 316 Closing (of trades), 205
Cancel replace orders, 108 Cash-settled futures contracts, 160 for call options, 105–6
Cancellable swaps, 550 Cash settlement, 74, 175 early, 181
Capital asset pricing model, 8 Cause-and-effect relationships, in for futures, 85, 181
Capital budgeting, 506, 659 theoretical models, 474 for swaps, 544
Capital gains, 7, 24 CBOE, see Chicago Board Options Closing call, 166
Capital losses, 7 Exchange Closing price (closing range), 166
Caplet and floorlet pricing (CFP) data, CBOE S&P 500 BuyWrite Index Closing transactions, 174
627 (BXM), 317 CLT (central limit theorem), 453–54
Caplet pricing, 619–20 CBOE Volatility index (VIX), 294, 480 CME, see Chicago Mercantile Exchange
with Black’s model, 621–22 CBOT, see Chicago Board of Trade CME ClearPort, 162
history of, 620–22 CBT, see Chicago Board of Trade CME Group, 195, 197
with HJM libor model, 627–31 CDS spread, 148 CMSs (constant maturity swaps), 550
with single-period binomial HJM CEA (Commodity Exchange Act), 197 Code of SWAPS, 542
model, 576–80 Central limit theorem (CLT), 453–54 Coherent risk measures, 667
with two-period binomial HJM CFMA (Commodity Futures Collars, option, 323–34, 564
model, 601–4 Modernization Act of 2000), Combination strategies, options trading,
Caplet–floorlet parity, 639–41 198–99 329–31
and floorlet pricing, 583, 605–6 CFP (caplet and floorlet pricing) data, COMEX, see Commodity Exchange
proof of, 588–89 627 Commercial traders, futures, 205
SUBJECT INDEX 723
Committee on Uniform Security Confirmations, of Master Agreements, law of one price in, 210–14, 217
Identification Procedures (CUSIP) 542 and market imperfections, 249–51
number, 44 Congestion, 126 and nothing comes from nothing
Commodities, 161n1 Consols, 143 principle, 213–14
Black’s model for pricing options on, Constant maturity forward rates, 633, setup of, 216
621–22 634 with storage costs and convenience
hedging for spot commodity positions, Constant maturity swaps (CMSs), 550 yields, 243–46
269–70 Constant parameters, hedging, 470–72 Cost of construction, options, 372–73,
intercommodity spreads, 185, 188 Consumer price index (CPI), 23 380
intracommodity spreads, 185, 187, 188 Consumption assets, cost-of-carry model Cost of exercising (term), 436
with storage costs and convenience for, 214 Costs-of-carry, 210
yields, 245 Contango, 246–49 Counterparties, 60
in supply chains, 86 Contingent claims models, 669–71 Counterparty risk, see Credit risk
Commodity Exchange (COMEX), 161, Contingent orders, 107, 108 Coupon bonds, 43, 503, 519–20; see also
162, 164 Continuous compounding, 687 Zero-coupon bonds (zeroes)
Commodity Exchange Act (CEA), 197 Continuous trading assumption, 426, Coupons, 7, 43
Commodity Exchange Commission, 197 427, 626 Covered calls, 313, 317
Commodity funds, 174 Continuously compounded dividend Covered interest rate parity, 241
Commodity futures, 161, 296 yield, 64 Covered puts, 319–20
Commodity Futures Modernization Act Continuously compounded forward Covered trading strategies, for options,
of 2000 (CFMA), 198–99 rates, 625, 626 314, 317–20
Commodity Futures Trading Continuously compounded interest Cox–Ross–Rubinstein binomial model,
Commission (CFTC), 81, 197 rates, 27, 29–34, 37, 218 407
enforcement actions by, 201 Continuously compounded return, 30, CPI (consumer price index), 23
listing requirements of, 55 451, 452 CPOs (commodity pool operators),
registration with, 173 Contract months, for Eurodollar futures, 174
regulation by, 198–200 527 Credit default swaps (CDSs), 147–48,
on short squeezes, 126 Contract size, 162 676
Commodity Futures Trading Convenience yield (convenience value), Credit events, credit default swaps and,
Commission Act, 160, 197 244–45 147–48
Commodity index traders, 205 Convergence of basis, 185 Credit risk, 13, 272–73, 660
Commodity-indexed notes, 134 Conversions, 292 and binomial HJM model, 566
Commodity markets, 205–6 Convertible bonds, 130–31 and binomial options pricing models,
Commodity pool operators (CPOs), 174 Convexity hedging, 512 375
Commodity pools, 174 COP (common options price) data, and BSM model, 424
Commodity price indexes, 167–69 313–14 and cost-of-carry model, 214
Commodity price risk, 15 Cornering the market, 125–26, 197 hedging, 148
Commodity swaps, 146–47 Corporate financial managers, 497 and HJM libor model, 622
Commodity trading advisors, 174 Corporate financial risk management, options pricing with, 371
Common options price (COP) data, 14–17, 658–60 reduced-form models of, 673–76
313–14 Corporate hedging, 259–64 structural models of, 669–70
Competitive markets, 193 Correlations, in econometric models, Credit spread, 321
and binomial HJM model, 566 474 Credit support, for swaps, 542
and binomial options pricing models, Cost-of-carry model, 209–25, 229–51 Critical assumptions, of models, 679
375 arbitrage table approach, 220–21 Cross-currency settlement risk, 669
and BSM model, 424 assumptions with, 214–16, 230–32 Cross-currency swaps, see Currency
and cost-of-carry model, 215 with backwardation/contango, swaps
and HJM libor model, 623 246–49 Cross-hedges, 264, 269–70
Complete markets, 156, 290, 399, 425 and dividend-paying stocks, 232–43 Cross-rates, currency, 138
Compound interest rates, 27–30 with dividend yield, 235–36 Crude oil, 204
Compounding with dollar dividends, 232–35 Cum-dividend (term), 62, 63
cash flow, 31–37 and forward maturities, 224–25 Cumulative standard normal
continuous, 687 for forwards at intermediate dates, distribution, 688–90
Condor spreads, 325, 326 221–24 Currency forwards, 78–79
724 SUBJECT INDEX
Currency risk, for businesses, 15 Delta- and vega-neutral portfolios, and put–call parity, 344
Currency swaps 485–86 synthetic indexes with, 236–39
fixed-for-fixed, 139–45 Delta hedging, 463–66 Documentation, swap, 543–44
plain vanilla, 94–95, 139–40, 145 and constant parameters, 470–72 Dodd–Frank Wall Street Reform and
Current assets, 658 with HJM libor model, 641–47 Consumer Protection Act, 199
CUSIP number, 44 of long calls, 467 Dojima Rice Exchange, 155, 157
of portfolios, 469 Dollar, US, 6, 138
Daily price limit (maximum daily price Delta-neutral portfolios, 483, 643–44 Dollar dividends
fluctuation), 162 Depreciation, US dollar, 138 and binomial options pricing model,
Daily return variance, 663 Derivative transaction execution 410
Daily settlement, futures, 84, 176–80 facilities, 198 and BSM model, 444–45
Dalian Commodity Exchange, 86 Derivatives (derivative securities), 3–12 and cost-of carry model, 232–35
Dark pools of liquidity, 61–62 acceptance of pricing models for, and put–call parity, 344
DARPA (Defense Advanced Research 475–76 Dollar return, 31–37, 378
Projects Agency), 193 application and uses of, 10–11 Doomsters, 168–69
Day orders, 108 build and break approach for, 130–31 Double privilege, 329
Day traders, 57 financial innovation with, 5–7 Dow Jones Industrial Average (DJIA),
DCMs, see Designated contract markets future of, 679–80 121
De Beers Group, 196 investors’ views of, 9–10 Dow Jones–AIG Commodity Index
De Moivre-Laplace theorem, 455 pricing of, 12, 118 Family, 167
Dealers terminology for, 7, 9 Down-and-out call options, 449
futures, 173–74 valuation of, 12 Down factors, 377, 378, 408
market maker, 56 Derivatives market Duration, 507–10
stock, 56–57 expansion of, 5–6 in active management, 512
swap, 205, 541 global, 4 and modified duration hedging,
Dealing rooms, OTC market, 79 history of, 3–5 510–12
Debit spread, 321 Designated contract markets (DCMs), and price change in bonds, 508–10
Decay, options, 313 198 Duration-neutral portfolios, 512
Default, 259 Diagonal spreads, 321 Dynamic hedging, 272
Default-free coupon bonds, yield for, Diamond cartel, 195–96 Dynamic market completion, 399
503 Digital options, 448–49
Default risk, see Credit risk Direct terms, for currency quotes, 138 EAR (effective annual interest rate),
Defense Advanced Research Projects Discounting, cash flow, 31–37 30n3
Agency (DARPA), 193 Discrete trading, 427 Early exercise, of American options,
Deferred swaps, 550 Diversifiable risk, 14 357–62
Deliverable supply, 125 Diversification, value-at-risk and, ECN (electronic communications
Delivery (maturity) date 665–66 network), 59, 62
of forwards, 73, 78 Dividend date, 62, 410, 445 Econometric models, 474–75
of futures, 162 Dividend rates, 410, 445 Econometrics, 281
and interest rate caplets, 627–28 Dividend yield, 63–64 Economic terms, of swap transactions,
of options, 296–97, 395–96, 413 cost-of carry model with, 235–36 542
Delivery period, futures, 162 forward price of index with, 237–39 Effective annual interest rate (EAR),
Delivery price, forward, see Forward put–call parity for options with, 344 30n3
price Dividends, 62–64 Efficient market hypothesis (EMH),
Delta (hedge ratio), 428, 643 and binomial options pricing model, 118–19
and BSM model, 463 409 Efficient markets, 118–19
of caplets, 579, 581, 643–44 and BSM model, 444–46 EFP (exchange for physicals), 175
interpreting, 478 cash, 62 Electronic communications network
of options, 435–39 defined, 7 (ECN), 59, 62
in options pricing, 379–80 463, 482 dollar, 232–35, 344, 410, 444–45 ELNs (equity-linked notes), 134
of portfolios, 469 and early exercise, 360–62 EMH (efficient market hypothesis),
sensitivity of, see Gamma forwards on stocks with, 232–43 118–19
Delta- and gamma-neutral portfolios, options adjustments for, 299–300 Employee stock options, calls against,
484–85 options with, 409–10, 444–46 318–19
SUBJECT INDEX 725
Equal pseudo-probability condition, 572 Exchange-traded funds (ETFs), 168, Financial distress costs, 259
for binomial HJM model, 575–76, 235–36 Financial engineering, 18, 131–34; see
597 Exchange-traded notes (ETNs), 168 also specific instruments, e.g.: Swaps
proof of, 586–87 Exchange-traded options, 105–6, Financial futures (financials), 155–56,
Equilibrium models, 387 289–90 161
Equity Exchanges; see also specific exchanges, e.g.: Financial innovation, 5–7
shareholder’s, 15, 658 New York Stock Exchange (NYSE) Financial institutions
value of, 659–60, 670 futures, 81, 83–88, 161, 166–67, 201 interest rate derivatives at, 497
Equity carve-outs, 343 options, 295–96 risk management from perspective of,
Equity-linked notes (ELNs), 134 securities, 55–57 17
Equity options trading, 289; see also stock, 60–61 Financial managers, corporate, 497
Options trading Exclusions, insurance policy, 327 Financial risk, 4
Equity swaps, 146 Execution (of trades) Firms, hedging by individuals vs.,
Estimated trading volume, on price for futures, 83–85 257–58
quotes, 166 for securities, 58 First delivery day, 162
ETFs (exchange-traded funds), 168, for stock, 60 First notice day, futures, 174
235–36 Exempt boards of trade, 198 Fixed-for-fixed currency swaps, 139–45
ETNs (exchange-traded notes), 168 Exercise by exception (ex-by-ex) Fixed-income securities, 23; see also
Euribor (Euro Interbank Offered Rate), procedure, 301 Bonds
603–4 Exercise prices, see Strike prices Fixed-rate loans, 136–38, 545–47
Eurobonds, 8 Exercise value, of options, 98, 103 Fixed-rate mortgage (FRM), 498
Eurodollar futures, 160, 524–30, 560, Exercising, of options, 93, 300 FLEX options, 294
561, 608–10 Exotic derivatives, 94 Floating-rate loans, 136–38, 545–47
Eurodollar futures rate, 610–11 Exotic options, 448–49 Floor trading, illegal arbitrage in, 124–26
Eurodollar rate, 48, 499 Exotic swaps, 94 Floorlets, 559–61; see also Caplet–floorlet
Eurodollars, 8, 499 Expectations hypothesis, 502, 504–6 parity
development of, 8 Expected return, value-at-risk and, 664 binomial HJM model for, 583, 605–6
in London interbank market, 47–48 Expiration date, option, 93, 95, 348, 352 caplet and floorlet pricing data, 627
in synthetic interest rate swaps, 550–52 Expiration month, option, 296 HJM libor model for, 639
Treasury–Eurodollar spread, 48, 500, Expiration process, option, 301 maturity of, 577
554 Exponents, 685, 686 payoffs for, 562
yield and swap curve for, 554 Extendible bonds, 130 Floors
Europe, medieval, 156 defined, 559, 560, 638
European options; see also specific pricing Face value, of coupon bonds, 43 HJM libor model for, 638–41
models, e.g.: Black–Scholes–Merton Factorials, 686 payoffs for, 562
(BSM) model Factors (term), 584 Foreign currency forwards, 239–43
American vs., 347 Failure to deliver, 41 Foreign currency options, 446–47
arbitrage opportunities with, 352–53 FASB 133 (Accounting for Derivative Forex (foreign exchange) markets,
on currency, 447 Instruments and Hedging 78–80, 138
delta and gamma for, 435, 438 Activities), 16 Forex (FX) swaps, 138–39
exercise of, 94 FCM (futures commission merchant), Forward contracts, see Forwards
on indexes, 445–46 84, 173 Forward exchange rates, 240–43
price bounds for, 347, 348, 352 February cycle, 297 Forward market model, 622; see also
profit diagrams for, 312–13 Fed, see Federal Reserve System Heath–Jarrow–Morton (HJM) libor
put–call parity for, 338–42, 344 Federal debt, 38–39 model
strike price of, 352 Federal funds market, 48 Forward price (F), 73, 76; see also
European Options Exchange, 295 Federal Reserve System, 6, 67 Cost-of-carry model
European terms, for currency quotes, Fees, option, 93 and arbitrage profit, 219–20
138 Final settlement, of Eurodollar futures, defined, 11
Even lot, 162 527 future vs., 180–85
Ex-dividend (term), 62, 234 Finance, 8 and maturity, 183–85, 224–25
Excess margin, 179 Financial assets, 7 and spot price, 246–49
Exchange for physicals (EFP), 175 Financial crisis, see Great Recession Forward rate agreements (FRAs),
Exchange rate, 138 (2007-2009) 521–23, 560, 561
726 SUBJECT INDEX
and forward rates for interest rate cash-settled, 160 Gearing, see Leverage
derivatives, 530–31 Eurodollar, 160, 524–29, 561, 563, Geometric Brownian motion
rates for Eurodollar futures vs., 610–11 608–10 assumption, 426
in synthetic interest rate swaps, 550–52 features of, 161–64 Gilt securities, 143
two-period binomial HJM model for, forwards vs., 81–83, 180–85, 272–73 GNMA (Government National
607 hedging with, 88–89, 194, 264–68, Mortgage Association), 160, 520
Forward rate agreements rate 272–73 Gold futures, 162–64
Eurodollar futures rate vs., 610–11 HJM libor model for, 647 basis risk with, 265–66
for HJM libor model, 631–32 interest rate, 524–30 daily settlement of, 177–78
Forward rate curve, 516 price quotes for, 164–66 price quotes for, 164–66
Forward rates, 513–20 types of, 159–61 trade records, 88
average volatility of, 632–36 Futures accounts, 176 Gold market
computing, 613–14, 623–25 Futures commission merchant (FCM), bonds indexed to prices in, 132–33
constant maturity, 633, 634 84, 173 manipulation in, 202
continuously compounded, 625 Futures contracts, see Futures Goldman Sachs and Company, 312
defined, 514 Futures market, 192–206 Goldman Sachs Group, 656
for discrete time intervals, 623 brokers and dealers in, 173–74 Good-until-canceled (GTC) orders, 108
for future period, 566 and commodity markets, 205–6 Government
futures vs., 610 and commodity price indexes, 167–69 interest rate derivatives for, 497
in HJM libor model, 623, 637 exchanges and clearinghouses in, Government National Mortgage
for interest rate derivatives, 530–31 166–67 Association (GNMA), 160, 520
interpretations of, 516–17 history of, 156–61 Grade specifications, in futures trading,
limit distribution of, 614 key conditions of, 193–94 163
and martingales, 637 manipulation in, 202–5 Grain Futures Act, 197, 198
from swap rates, 553 regulation of US, 196–201 Grain Futures Administration, 197, 198
Forward swaps, 550 speculation in, 195–96 Grantors, see Writers, option
Forwards (forward contracts), 73–78; see Futures options, 160, 295–96 Great Moderation, 6
also Forward rate agreements Futures prices, 524, 525 Great Recession (2007–2009)
(FRAs); Interest rate swaps arbitrage-free, 608–9 credit default swaps in, 540
application and uses of, 155–56 forwards vs., 180–85 derivatives in, 3, 5, 6
binomial HJM model for, 607–8 and maturity, 183–85 model misuse in, 463
defined, 11 price discovery, 194 scenario analysis by banks in, 666
on dividend-paying stocks, 232–43 Futures rates Greeks
futures vs., 81–83, 180–85, 272–73 arbitrage-free, 609–10 in Black–Scholes–Merton model,
hedging with, 88–89, 272–73 forward vs., 610–11 434–39
at intermediate dates, 221–24 Futures trading, 172–89 hedging, 463–69, 482–85, 642–47
maturities of, 224–25 on exchanges, 83–88 Gross domestic product, 17
terminology for, 73–78 margin accounts and daily settlement GTC (good-until-canceled) orders, 108
timeline for, 75 for, 176–80
Forwards market spreads in, 185–88 Heath–Jarrow–Morton (HJM) libor
history of, 156–57 trading futures, 174–76 model, 12, 618–49
over-the-counter, 78–80 FX (forex) swaps, 138–39 and actual vs. pseudo-probabilities,
Fractional stock splits, 299–300 638
Free-market environmentalism, 168 Gambling, with derivatives, 5 for American options, 647
Frictions, see Market imperfections Gamma assumptions with, 622–27
FRM (fixed-rate mortgage), 498 of call options, 437–39 and binomial HJM model, 613–15
Front running, 124 of caplets, 642 Black–Scholes–Merton model vs., 630
Fundamental analysis, 119 of European options, 435–39 and caplets, 619–22, 627–31
Fundamental value interpreting, 434–35 for caps and floors, 638–41
and cost-of-carry model, 215 of portfolios, 469–70 delta hedging in, 643–45
and price bubbles, 375, 424 of put options, 437–39 forward rates and martingales in, 637
Futures (futures contracts), 81–89 Gamma hedging, 466–69, 483–85, for futures, 647
application and uses of, 155–56 641–42, 646–47 hedging Greeks with, 640–47
binomial HJM model for, 608–10 Gasoline futures, manipulation of, 204 inputs for, 631–36
SUBJECT INDEX 727
risk-neutral valuation in, 637–38 Heath–Jarrow–Morton (HJM) libor Interbank market, 55, 78; see also
for swaptions, 647 model Over-the-counter (OTC) markets
Heath–Jarrow–Morton (HJM) model, Horizontal spreads, 185, 321 Intercommodity spread, futures, 185,
see Binomial Heath–Jarrow–Morton Housing, residential, see Residential 188
(binomial HJM) model housing market Interest, in cost-of-carry model, 218–20
Heating oil futures, manipulation of, Hunt Brothers, market manipulation of, Interest rate collars, 549
204–5 202–3 Interest rate derivatives (IRDs),
Hedge ratio, see Delta Hurricane risk, hedging, 327–28 496–534; see also Interest rate swaps
Hedgers, 57 Hybrid debt, 134 about, 497–99
Hedging, 256–81; see also Delta hedging and binomial HJM model, 560–64
of airline fuel costs, 146, 147 IB (introducing broker), 173 forward rate agreements, 521–23
and binomial options pricing model, IBM (International Business Machines), forward rates for, 513–20, 530–31
399 297, 299–304, 539–40 FRA rates for, 530–31
and BSM model, 480–81 Identical distribution, of stock returns, history of, 520–21, 564–65
calibration and errors in, 476–77 452 interest rate futures, 524–30
with caps, 562–63 iid property, of stock returns, 452 interest rate risk management with,
computing optimal hedge ratio, Illegal arbitrage opportunities, 124–26 507–13
273–75 IMM (International Monetary Market) and Treasuries, 43
of constant parameters, 470–72 division, 6, 159 and yields, 499–507
convexity, 512 Imperfect hedges, 264, 269–71 Interest rate futures, 524–30
corporate, 259–64 Implied volatility, 441 Interest rate parity, covered, 241
with currency forwards, 79 average forward rate, 636 Interest rate risk, 15, 562, 641
of default risk, 148 and calibration, 477–80 Interest rate risk management, 507–13
defined, 3 computing, 478–79 Interest rate swaps, 136–38, 538–55
with derivatives, 5, 10–11 historical vs., 632 and forward rate agreements, 550–54
duration, 512–13 In-the-money (term), 95, 101 history of, 539–42
dynamic, 272 Independence, of stock returns, 452 institutional features of, 542–44
with Eurodollar futures, 526 Index arbitrage, 121–23 plain vanilla, 94, 136–38
by firms vs. individuals, 257–58 Index futures, 160 valuation of, 544–49
with forwards, 88–89, 272–73 Index mutual fund, 235 variations of, 549–50
with futures, 88–89, 272–73, 264–68, Indexes; see also Stock indexes and yield–swap curve relation, 552,
272–73 European options on, 445–46 554
gamma, 466–69, 483–85, 646–47 for interest rate derivatives, 560 Interest rates, 22–50
of hurricane risk, 327–28 market-cap weighted, 123 about, 23–24
of interest rate risk, 641 synthetic, 236–39 compound, 25–31
modified duration, 510–12 total return, 236, 238 continuously compounded, 27–30,
in options trading, 317–21 India 36–37, 219
of portfolios, 469–70 futures trading in, 86–87 discounting and compounding cash
pricing and hedging argument, Indirect terms, for currency quotes, 138 flows with, 31–37
428–29 Individuals libor vs. libor rate index, 47–48
rho, 439 hedging by firms vs., 257–58 and measuring time, 26
risk-minimization, 268–72 interest rate derivatives for, 497 and option value, see Rho
of risks, 16 Induction, backward, 396–97, 601 and rate of return, 24–25
and synthetic calls, 372 Inflation, 23 and rounding/reporting of numbers,
vega, 439, 470–73, 480, 485–86, 641 Informationally efficient markets, 118 25
Herstatt risk, 669 Initial margin, 67, 163, 176 simple, 27, 27–31
High-frequency trading, 425; see also Initial public offering (ipo), 54 term structure of, 371, 502, 567–68,
Algorithmic trading (algo) Input price risk, 88–89 594, 647
High price, futures, 166 Insiders, 119 and Treasuries, 37–47
Historical data, pricing models and, 476 Insolvent firms, 659 uncertainty about, 375–76, 425
Historical volatility, 441–44, 633–36 Institutional investors, in Treasuries, 40 Intermediaries, forward transaction, 80
HJM models, see Binomial Institutional Networks (Instinet), 62 Intermediate cash flows, 215, 375, 424,
Heath–Jarrow–Morton (binomial Insurance, 5, 290 566, 623
HJM) model; Integer multiple stock splits, 299 Intermediate dates, forwards at, 221–24
728 SUBJECT INDEX
Internal rate of return (IRR), see Yield Last updated, on price quotes, 166 Lognormal assumption (for stock prices),
Internalization, of orders, 59 Law of large numbers, 327 426, 451–59
International Business Machines, see Law of one price, 114, 117 Lognormal distribution, for simple
IBM cost-of-carry model from, 207–11, forward rate, 626
International Commercial Exchange, 216–21 London Interbank Bid Rate (libid), 48
159 and no-arbitrage principle, 117 London Interbank Offered Rate (libor),
International Monetary Market (IMM) LE (local expectations) hypothesis, 504, 47–48, 50, 499
division, 6, 159 600 London Stock Exchange, 59, 157
International Organization for LEAPS (Long-Term Equity AnticiPation London Stock Exchange Committee,
Standardization (ISO), 79n3 Securities), 294 291
International Organization of Securities Least squares method, 281 London Traded Options Market, 295
Commissions (IOSCO), 12, 13 Legal risk, 13, 678 Long calls
International Securities Exchange, 295 Lending, role of derivatives in, 5 in covered trading strategies, 318, 320
International Swap Dealers Association, Lending rates, 251 delta and gamma hedging of, 467
138, 541 Leverage margin requirements for, 298
International Swaps and Derivatives with derivatives, 5, 9 profit diagrams for, 315–16
Association (ISDA), 138, 541 margin for, 66 put–call parity for, 344–46
Intracommodity spread, futures, 185, and put–call parity, 342 Long hedges, 88, 266
187, 188 Liabilities, 14, 658 Long positions (longs)
Intrinsic value Libid (London Interbank Bid Rate), 48 in covered trading strategies, 317–20
and cost-of-carry model, 215 Libor (London Interbank Offered Rate), in derivatives, 74
of options, 98–99, 103 47–48, 50, 499 in 1:2 hedges, 320
and price bubbles, 375, 424–25 Libor market model, 622; see also on options, 94
Introducing broker (IB), 173 Heath–Jarrow–Morton (HJM) libor in options market, 107
Investment assets, cost-of-carry model model payoffs for forwards with, 77–78
for, 214 Libor vs. libor rate index, 47–48 and put–call parity, 339
Investment companies, portfolio risk Lifetime highs, futures, 166 in stock, 314
management by, 14 Lifetime lows, futures, 166 Long puts
IOSCO (International Organization of Limit buy orders, 108 in covered trading strategies, 319
Securities Commissions), 12, 13 Limit distribution, forward rate, 614 margin requirements for, 298
IPO (initial public offering), 54 Limit order book, 62 profit diagrams for, 315, 316
IRDs, see Interest rate derivatives Limit orders, 62, 105, 108 put–call parity for, 339, 345
IRR (internal rate of return), see Yield Linear regression models, 281 Long straddles, 330–31
ISDA, see International Swaps and Liquidation, payoffs at, 132, 133 Long strangles, 330–31
Derivatives Association Liquidation-only trading, 203 Long-term assets, 658
ISDA Agreement Structure, 542 Liquidity Long-Term Capital Management
ISO (International Organization for dark pools of, 61–62 (LTCM), 677
Standardization), 79n3 in securities market, 57 Long-Term Equity AnticiPation
Issuers, see Writers, option Liquidity preference hypothesis, 504 Securities (LEAPS), 294
Ito’s lemma, 487 Liquidity risk, 13, 660, 668, 677–78 Loss distribution, 660–62
ITT Corporation, 540 Listing requirements, stock Low price, futures, 166
exchange, 55 LTCM (Long-Term Capital
January cycle, 297 Loans Management), 677
Jarrow–Rudd binomial model, 407 fixed-rate, 136–38, 545–46
Jensen’s inequality, 610 floating-rate, 136–38, 545–46 Maintenance margin, 67, 163, 178
J. P. Morgan, 540–41, 662 interest rate swaps with, 136–38 Malthusian doctrine, 168
J. P. Morgan Chase, 498 secured, 67 Managing books (term), 18, 541
Local expectations (LE) hypothesis, Margin
Kansas City Board of Trade, 160 504–5, 600 for futures, 163
Locked-in transactions, 60 in history of futures, 158
Last delivery day, 162 log(x), 685 for options trading, 298–99
Last notice day, 174 Logarithmic (continuously for portfolio of futures, 182
Last trade price, 165 compounded) return, 30, 451–52 Margin accounts, 66–68, 176–80
Last trading day, futures, 162, 174, 527 Logarithms, 685–86 Margin calls, 67
SUBJECT INDEX 729
Mark-to-model accounting, 221 and futures prices, 183–85 no-arbitrage principle for, 383
Marked-to-market, futures, 176, 272 and option prices, 355 stock prices in, 393–94
Market buy orders, 107 prices at, 395–96, 415 and two-period binomial options
Market-cap weighted index, 123 and prices of forwards, 224–25 pricing model, 403–5
Market capitalization, 122 spot price at, 183 Multiplicative binomial model, 377–78
Market close (closing call), 166 of swaps, 135 Mutual funds, 120
Market corners, 125–26, 202 Maturity date, see Delivery date
Market imperfections (market frictions) Maximum daily price fluctuation, 162 n-factorial, 403, 686
arbitrage opportunities from, 115 MCX (Multi Commodity Exchange), 87 N-period binomial options pricing
and binomial HJM model, 565 Mean, of constant maturity forward model, 405–9
and binomial options pricing models, rates, 634 Naked short selling, 77
374 Mean-variance approach, 268–71 Naked trading strategies, 314–16
and cost-of-carry model, 214, 247, Merc, see Chicago Mercantile Exchange NASDAQ, see National Association of
250 Microsoft Excel; see also Spreadsheet Securities Dealers Automated
defined, 95 software Quotations
and futures/spot prices, 183 computing optimal hedge ratio in, National Association of Securities
and HJM libor model, 622 273–75 Dealers Automated Quotations
and put–call parity, 342–46 multiperiod binomial options pricing (NASDAQ), 61, 62
Market maker dealers, 56 models in, 413–17 National Commodity and Derivatives
Market manipulation option pricing project, 491–93 Exchange Limited, 87
futures, 202–5 Miller and Modigliani (M&M) National Futures Association (NFA), 81,
and option pricing, 439–40 propositions, 8, 117, 257 160, 197
options, 305–6 Minimum price fluctuation, 162, 190, National Multi Commodity Exchange
Market model, 565, 622; see also 527 of India Ltd., 86–87
Heath–Jarrow–Morton (HJM) libor Minimum-variance hedge ratio (h), National Securities Clearing
model 276–81 Corporation (NSCC), 60
Market orders, 60, 105, 107 Mizuho Securities Company, 678 Natural logarithms, 685–86
Market price, calibration and, 476 Model risk, 679–80 NAV (net asset value), 120
Market risk, 13, 660, 668 Models definition of, 214 Necessary conditions, for no arbitrage,
Market segmentation hypothesis, 504 Modified duration, 507–10 570–71, 586–87, 597–98
Marking the close, 204 Modified-duration hedge ratio, 511 Net asset value (NAV), 120
Marking-to-market (term), 84 Modified duration hedging, 510–12 Net present value, risk premiumand, 386
Married puts, 318, 319 Money market account (mma), 372 Netherlands, 59, 73, 204, 291, 305
Martingale pricing dollar return from, 31 New York Mercantile Exchange
and binomial options pricing model, moving cash flows across time with, (NYMEX), 84, 158
369 34–35 New York Stock Exchange (NYSE), 60
and BSM model, 456 and spot rate, 565, 595 NFA, see National Futures Association
defined, 379 Money market yield basis, 549 No-arbitrage principle, 117
Martingales Monotonicity, of risk measures, 667 in binomial options pricing model,
in Black–Scholes–Merton model, Montreal Stock Exchange, 295 376–77
432–33 Mortgage-backed securities, 498 and index arbitrage, 123
and forward rates, 637 Mortgages, 498 proof of, 384–88
in HJM libor model, 637 Multi Commodity Exchange (MCX), 87 No-dividend stocks, American options
and single-period binomial HJM Multifactor models on, 358
model, 573–75 single-period binomial HJM, 584–85 Noncommercial traders, futures, 205
in single-period binomial options two-period binomial HJM, 606 Nondiversifiable risk, 14
pricing model, 378–79 Multiperiod binomial HJM model, 607 Noneconomic terms, in Master
and two-period binomial HJM model, Multiperiod binomial options pricing Agreements, 542
599–600 model, 403–9 Nonfinancial companies, risk
and zero-coupon bond prices, binomial coefficients and management for, 17
573–75, 599–600 pseudo-probabilities in, 403 Nonlinear statistical model, BSM as, 479
Master Agreement, ISDA, 542 and BSM model, 407–9 Nonstandard derivatives, 94
Maturity features of, 399 Normal backwardation, 248, 249
of caplets and floorlets, 577 general form of, 405–7 Normal contango, 247, 248
730 SUBJECT INDEX
Normal distribution, 688–90 Options Clearing Corporation (OCC), Pacific Exchange, 295
Notes 105, 293, 299, 300 Palm Inc., 337, 343
commodity-indexed, 134 Options Industry Council, 319 Par bond yield, 506, 552
equity-linked, 134 Options market Par swaps, 145, 547
exchange-traded, 168 history of, 290–96 Par value, see Principal value
Treasury, 43, 44, 530 longs and shorts in, 107 Partial differential equation proof, of
Nothing comes from nothing principle, regulation and manipulation of, 305–6 BSM model, 429, 458–59
117–18 traders in, 311–12 Payer swaptions, 560
Notional principal, in interest rate swaps, Options on futures (futures options), Payment dates
136 160, 296 of floorlets, 639
Notional variables, 7 Options on spot (spot options), 160, 295 of forward rate agreements, 523
NSCC (National Securities Clearing Options prices, 336–63; see also specific of interest rate caplets, 629
Corporation), 60 pricing models, e.g.: Binomial options Payoff diagrams
Numbers, reporting, 26 pricing model for bonds indexed to gold, 133
Numerical methods, of pricing options, COP data, 313 for call options, 95–97
369, 371, 448 and early exercise of American for caps and caplets, 563
NYMEX, see New York Mercantile options, 357–62 for cost-of-carry model, 212, 213
Exchange history of options pricing, 369–71 for put options, 101
NYSE, see New York Stock Exchange and market manipulation, 439–40 put–call parity on, 337–41
put–call parity, 337–42 Payoffs
OCC, see Options Clearing Corporation quotes for, 301–4 for caplets, 577–79, 627–28
Off-the-run securities, 40 restrictions on, 346–56 for floors and floorlets, 563–64
Offer price, see Ask price Options series, 301 for futures with early closes, 181
Offset, in swap closing, 544 Options trading, 300–1, 310–32 at liquidation, 132
On-the-run securities, 40 combination strategies, 329–31 for long vs. short forwards, 77–78
One-by-two call spread, 325 COP data for, 313 for zero-coupon bonds with forward
One-factor models, 584 early exercise of American options, rate, 517
One price, law of, see Law of one price 357–62 PEPS (premium equity participating
1:2 hedge, 320 hedged strategies, 317–20 securities), 134
OPEC, see Organization of the history of, 312–13 Percentage dividend rates, 409
Petroleum Exporting Countries naked strategies, 314–16 Perfect hedges, 264, 268
Open-end companies, 119–20 profit diagrams for, 312–13 Perfect markets, 95
Open interest, 87–88, 165 spread strategies, 320–29 Perfect substitutes hypothesis, 504
Open repos, 42 Optiver Holding, 204–5 Performance margin, see Initial margin
Opening call period, for futures, 165 Order placement Philadelphia Stock Exchange, 295
Opening high price, 165 for futures, 83–84 Physical delivery, forwards and futures,
Opening low price, 165 for options, 105 74, 76, 83, 174
Opening price (opening range), 165 for stock, 59, 60 Pink quote, 61
Operational risk, 13, 661, 668, 678 Organization of the Petroleum Plain vanilla (standard) derivatives, 94
Optimal hedge ratio (h), 273–75, 280 Exporting Countries (OPEC), 194 Plain vanilla swaps, 135, 543
Option class, 301 OTC Bulletin Board (OTCBB), 61 currency, 140–41, 145
Option contracts, see Options OTC markets, see Over-the-counter interest rate, 94, 136–38, 544–49
Options (option contracts), 92–109, (OTC) markets Policy Analysis Market, 193–95
288–300; see also specific types, e.g.: OTC transactions, stock, 55 Ponzi schemes, 201
American options Out-of-the-money (term), 95, 101, 305 Portfolio risk, 14
decay of, 313 Output price risk, 89 Portfolio risk management, 14
with dividends, 409–10, 444–46 Over-the-counter (OTC) markets, 3 Portfolio theory, 8
exchange-traded, 105–6, 289–90 forwards, 78–80 Portfolios
features of, 296–300 options, 289–90 delta- and gamma-neutral, 484–85
order placement strategies, 107–8 stock, 55 delta and vega-neutral portfolios,
outright purchases of, 314 trading securities on, 61 485–86
price quotes for, 301–4 Overnight repos, 42 delta-neutral, 483, 643, 644
quarterly cycles of, 297 Overwrite, option, 317, 318 duration-neutral, 512
terminology for, 93–94 Owners, see Buyers futures, 182
SUBJECT INDEX 731
hedging of, 469–70 Production assets, cost-of-carry model price bounds for, 104, 348, 349, 352
top-down construction of, 14 for, 214 pricing, with Microsoft Excel, 414–16
Position limits, 163, 299, 528 Profit diagrams protective, 319
Position size, futures, 162 for call options, 96–98, 101, 103 short, 315, 316, 318, 319, 345–46
Position traders (trend followers), 57 for long vs. short forwards, 77–78 and strike price, 352
Positive homogeneity, of risk measures, for options price restrictions, 346–48 Put spreads, 313, 321, 322
667 for options trading, 312–13 Put-to-call strategies, see Straddles
Preferred habitat theory, 504 Profit margin, 66 Putable swaps, 550
Premium Profits; see also Arbitrage profits Put–call parity, 337–46
option, 93, 94, 290 of hedged vs. unhedged firms, 263 and BSM model for pricing puts,
Premium equity participating securities and rate of return, 24–25 432–3
(PEPS), 134 and zero-profit point, 313 for European options, 338–42
Present value, 34–35 Promised rate of return, 25 and market imperfections, 342–46
Price bounds Protective put, 318, 319 on payoff diagrams, 337–39
for American call options, 99–100 Pseudo-probabilities, 379; see also pricing put options with, 401
for American put options, 103–4 Martingales and single-period binomial options
for forwards with market actual and, 385, 433–34, 457–58, pricing model, 374
imperfections, 249–50 576, 638 PWI (price-weighted index), 121
Price bubbles, 305 and binomial HJM model, 576
and cost-of-carry model, 215 and binomial options pricing model, Quality specifications, in futures trading,
and options pricing, 375, 424, 425 403, 408 163
Price-contingent orders, 108 and BSM model, 408, 433–34, 457–58 Quarterly cycles, options, 297
Price discovery, 156, 194 computing, with Microsoft Excel, 413 Quarterly options (quarterlies), 297
Price limit, daily, 162 and HJM libor model, 638
Price quotes and no-arbitrage principle, 384 Random walk processes, 452
futures, 162, 164–66 risk-neutral valuation with, 381, 582 Rate-capped swaps, 549
options, 301–4 and risk premium, 385–88 Rate of return, 24–25; see also Yield
Price relatives, stock, 441 Put and Call Brokers and Dealers annualized, 25, 45, 48
Price restrictions, options, 346–56 Association, 292 promised vs. realized, 25
Price risks Put options (puts), 100–4; see also Ratio log, 633n2
commodity, 15 Put–call parity Ratio spread, 324–26, 327n5
input, 88–89 American, 104, 344–46, 348, 349, Real assets, 7
output, 89 352, 353 Real options, 672–73
Price-weighted average, 121 arbitrage opportunities with, 354–55 Realized rate of return, 25
Price-weighted index (PWI), 121 in binomial options pricing model, Receiver swaptions, 560
Pricing and hedging argument, 428–29 373, 374 Reduced-form models, 371, 669,
Pricing models; see also specific models, binomial pricing model for, 373, 374 673–76
e.g.: Black–Scholes–Merton model BSM model for, 431–2 Reference rate, FRA, 522
arbitrage in, 113 cash-secured, 316 Regulation(s)
for currency swaps, 143–44 on commodity futures, 296 and financial innovation, 7, 8
for derivatives, 12 covered, 319 of futures market, 196–201
Primary dealers, Treasuries, 39 defined, 93 of options market, 305–6
Primary market, stock, 54 delta and gamma for, 437–38 of stock exchanges and OTC markets,
Principal, 27 delta hedging errors with, 465 56
Principal value, 7 digital, 448 Regulation ATS, 61
of coupon bonds, 43 European, 348, 352–3, 374, 415–16, Regulation Q, 8
of zero-coupon bond, 32 431–32, 439 Regulatory arbitrage, 342
Private placement, 134 exercising, 93 Reinsurance, 327, 328–29
Probabilities, see Actual probabilities; on interest rates, see Floorlets Repo rate, 41
Pseudo-probabilities intrinsic and time values of, 103 Repurchase agreement (repo) market,
Probability distribution, for losses, 660–62 long, 298, 315, 316, 318, 319, 339, 41–42
Procter & Gamble Corporation (P&G) 344–45 Resale market, US Treasury, 40
interest rate swaps with, 539 married, 319 Return-to-maturity expectations
risk management in, 16–17 payoff and profit diagrams for, 100–2 (RME) hypothesis, 505
732 SUBJECT INDEX
Reuters/Jeffries CRB Index (RJ/CRB), Risk-neutral probabilities, 379; see also fixed-income, 23; see also Bonds
167 Pseudo-probabilities gilt, 143
Reverse cash-and-carry, 217 Risk-neutral valuation off- vs. on-the-run, 40
Reverse repo market, 41–42 in binomial HJM model, 581–83 premium equity participating, 134
Reverse stock splits, 300 and binomial options pricing model, riskless vs. risky, 568
Reversing trades, 174 381–82, 400–1 Securities and Exchange Commission
Rho, 436, 439 and BSM model, 433, 456 (SEC), 55
Rho hedging, 439 of caplets, 602–4 on calls against employee stock
Rhône-Poulenc, 130 and HJM libor model, 637–38 options, 318–19
RICI (Rogers International Commodity of zero-coupon bonds, 574, 600–1 on options, 290
index), 167 Risk premiums regulation of options by, 200
Risk; see also Credit risk and BSM model, 457–58 regulation of stock exchanges by, 55
basis, 264–7 and cost-of-carry model, 246–47 Securities markets, 54–57
classification of, 12–14 and expectations hypothesis, 504 Securitization, 38
corporate, 14–15 and pseudo-/actual probabilities, Security futures products, 198
currency, 15 385–87 Self-financing trading strategy, 398
diversifiable, 14 of zero-coupon bonds, 576 Self-regulatory organizations (SROs),
financial, 4 Riskless securities, 568 200
Herstatt, 669 Risky securities, 568 Sellers
hurricane, 327 RJ/CRB (Reuters/Jeffries CRB Index), option, see Writers, option
input price, 88–89 167 swap, 545
interest rate, 15, 562, 641 RME (return-to-maturity expectations) Selling hedges, 89
legal, 13, 678 hypothesis, 505 Selling the spread, 187
liquidity, 13, 660, 668, 677–78 Robust assumptions, 679 Semiannual bond equivalent basis, 548
market, 13, 660, 668–69 Rogers International Commodity index Semistrong-form efficiency, 119
measures of, 667–68 (RICI), 167 Settlement
model, 679–80 Roller-coaster swaps, 550 cash, 74, 76, 175
nondiversifiable, 14 Rolling, of futures indexes, 167 daily, 84, 176–80
nonhedged, 16 Round lot, 60 of exchange-traded call options, 105–6
operational, 13, 661, 668, 678 Round-trip commissions, for trading of exchange-traded futures, 84
output price, 89 futures, 84 final, 528
portfolio, 14 Rounding, conventions for, 26 futures trading at, 163
price, 15, 88–89 Royal Exchange of London, 157 of securities trades, 59, 60–61
settlement, 13, 669 Settlement date, FRA, 522, 523
tail, 329 Sallie Mae (Student Loan Marketing Settlement price, futures, 166
Risk-adjusted interest rate, 382–83 Association), 540 Settlement risk, 13, 669
Risk management, 12–19, 655–81 Salomon Brothers, Inc., 135, 203–4, Shanghai Futures Exchange, 86
and Basel Committee’s four risks, 311, 539, 678 Shareholder/debt holder conflict, 671
668–78 Saving, role of derivatives in, 5 Shareholder’s equity, 15, 659
and classification of risk, 12–14 Savings and loans banks, 135, 498 Shipping certificates, for futures, 174
corporate financial, 14–17, 658–60 Scalpers, 57 Short calls
future of traded derivatives and, Scenario analysis, 666–68 in covered trading strategies, 317–20
679–80 Sealed bid auctions, 38 in 1:2 hedge, 320
with futures, 83 Sealed bids, 38 profit diagrams for, 315, 316
interest rate, 507–13 Seasonal swaps, 550 put–call parity for American, 344–45
loss distribution in, 660–62 SEC, see Securities and Exchange Short hedges, 89, 266
perspectives on, 17–18 Commission Short positions (shorts), 74
portfolio, 14 Secondary markets, 40, 54 in covered trading strategies, 126, 318,
scenario analysis in, 666–68 Secured loans, 67 319
VaR in, 662–66 Securities; see also Derivatives; Treasury on options, 93, 107
Risk Management Guidelines for Derivatives securities (Treasuries) payoffs for forwards with, 77–78
(Basel Committee on Banking), 13, asset-backed, 498 in stock, 314
668 automated trading of, 58 Short puts
Risk-minimization hedging, 268–72 defined, 7 in covered trading strategies, 318, 319
SUBJECT INDEX 733
profit diagrams for, 315 Spot rate, 500, 502, 518, 565, 570, with continuous vs. discrete trading,
put–call parity for, 345–6 595 427
Short selling, 64–66 Spread orders, 108 lognormal assumption for, 426,
margin accounts for, 67, 68 Spread trading, 120–21 451–59
naked, 77 futures, 185–88 and price of call option, 350
proceeds from, 251n4 options, 320–29 Stock privileges, 329
rate of return for, 25n2 Spreads, 18 Stock splits, 299–300
in repo market, 41 bear, 321–23 Stock trading
Short squeeze, 125–26, 203–4 bid/ask, 56, 251 alternative systems for, 61–62
Short straddle, 331 bull, 321–23, 324, 326 on exchanges, 60–61
Silver futures, manipulation of, 202–3 butterfly, 326, 326 and margin, 66–68
Simple forward rate, 613, 624–25, 631 buying, 187 at NYSE, 59–62
Simple interest rates, 27 call, 321, 323, 324, 326, 327 on OTC securities market, 61
Single-period binomial HJM model, call backspread, 327n5 short selling in, 64–66
570–85 call ratio, 327n5 Storage costs, 243–45
and actual vs. pseudo-probabilities, CDS, 148 Straddles, 330–31
576 condor, 325, 326 Straight cancel orders, 108
arbitrage-free conditions for, 570–73 credit, 321 Strangles, 330–31
caplet pricing with, 576–80 debit, 321 Stress testing (scenario analysis),
equal pseudo-probability condition diagonal, 321 666–68
for, 575–76 horizontal (time), 185, 321 Strike prices, 93, 297–98, 350, 352
floorlet valuation with, 582–83 intercommodity, 185, 188 STRIPS (separate trading of registered
hedge ratio in, 581 intracommodity, 185, 187 interests and principal of securities),
and martingales, 573–75 one-by-two call, 325 44–45
multiple factors in, 584–85 put, 313, 321, 322 Strong-form efficiency, 119
risk-neutral valuation with, 581–82 ratio, 324–26, 327n5 Structural models, 371, 669–73
and zero-coupon bond prices, 573–75 selling, 187 Student Loan Marketing Association
Single-period binomial options pricing Treasury–Eurodollar, 48, 500, 554 (Sallie Mae), 540
model, 376–83 vertical, 321–24 Subadditivity, of risk measures, 667
and actual vs. pseudo-probabilities, SROs (self-regulatory organizations), Sufficient conditions, for no arbitrage,
382–83 200 572–73, 587, 597–99
arbitrage-free trees in, 378 Standard & Poor’s 500 Index (S&P 500), Superglue argument, 346, 352–54
binomial trees in, 377–78 63, 121, 123, 160 Supply chains, commodities in, 86
for European calls, 372–75 Standard normal random variable, Swap books, managing, 541
hedge ratio for, 380–81 426–27 Swap buyers, 545
martingales in, 378–79 Standardization, futures market, 193–94 Swap curve, 552, 554
multiperiod vs., 399 Statistical models, 474–75, 636 Swap dealers, 205–6, 541
no-arbitrage principle for, 376–77, Stochastic volatility option models, Swap execution facilities, 199
383 472–73 Swap facilitators (swap banks), 136, 541
risk-neutral valuation in, 381–82 Stock, 59–70 Swap issuers, 542
stock prices in, 378–79, 393–94 defined, 7 Swap market, 540
Southwest Airlines, 260–61 dividend-paying, 234 Swap rate, 547, 553
S&P 500, see Standard & Poor’s 500 dividends of, 62–64, 232–43 Swap sellers, 545
Index forwards on, 232–43 Swaps, 135–48; see also Interest rate
S&P GSCI indexes, 167 margin accounts for trading, 66–68 swaps; Plain vanilla swaps
Speculation, 5, 10–11 short selling, 64–66 accreting, 550
with caps, 561–62 transactions with, 58–59 amortized, 549
in futures market, 195–96 Stock index funds, 235 application and uses of, 135–36
Speculators, 57 Stock indexes, 63, 121–23 asset-backed securities credit default,
Spin-offs, 343 Stock prices 148
Split ratio, 299 in binomial options pricing model, callable, 550
Spot commodity positions, 269–70 378–79, 393–94 cancellable, 550
Spot options, 160, 295 binomial tree for, 395–96 commodity, 146–47
Spot prices, 76, 183, 186, 246–49 in BSM model, 432–33 constant maturity, 550
734 SUBJECT INDEX
credit default, 147–48, 676 measuring, 26 Treasury bonds (T-bonds), 43–47, 552
currency, 139–47 moving cash flows across, 31–37 Treasury futures, 529–30
development of, 8 Time-contingent orders, 108 Treasury inflation protected security
equity, 146–47 Time-of-day orders, 108 (TIPS), 43
exotic, 94 Time spread, 185, 321 Treasury notes (T-notes), 43–47, 552
forex, 138–39 Time value Treasury securities (Treasuries), 37–47
forward, 550 debt, 671–72 characteristics of, 37–38
par, 145, 547 of options, 99, 103 and federal debt auction markets,
putable, 550 TIPS (Treasury inflation protected 38–39
rate-capped, 549 security), 43 futures manipulation of, 203–4
roller-coaster, 550 Top-down portfolio construction, 14 investment methods for, 40–43
seasonal, 550 Top straddle, 331 Treasury–Eurodollar spread, 48
terminology for, 135 Total return indexes, 236, 238 types of, 43–47
yield, 549 Traders yield and swap curve for, 503, 552,
Swaptions, 560, 611–12, 647 commercial, 205 554
Synthetic calls, 372, 397–400 commodity index, 205 Treasury–Eurodollar (TED) spread, 48,
Synthetic caplets, 578–79, 601–2 day, 57 500, 554
Synthetic futures contracts, 121 noncommercial, 205 Trend followers, 57
Synthetic indexes, 236–39 in options market, 311–12 Troubled Asset Relief Program (TARP),
Synthetic interest rate swaps, 550–52 position, 57 289, 368, 666
Synthetic options, 372, 376, 397–400, risk management by, 17 Tulip bulb bubble, 305
429 stock market, 56–57 Two-factor models, 584
Trades Two-period binomial HJM model,
T-bills (Treasury bills), 43–47 bearish, 65, 93 597–606
T-bonds (Treasury bonds), 47–47, 524 block, 60 arbitrage-free conditions for, 597–99
T-notes (Treasury notes), 43, 49, 530 bullish, 65, 93 caplet pricing with, 601–4
Tail risk, 329 clearing of, 59, 60–61, 84, 105–6 for Eurodollar futures, 608–10
TARP, see Troubled Asset Relief reversing, 174 floorlet pricing with, 605–6
Program Trading; see also Futures trading; Options for forward rate agreements, 607–8
TAS (Trading at Settlement) contracts, trading; Stock trading with multiple factors, 606
163 alternative systems for, 61–62 for swaptions, 611–12
Tax arbitrage, 342 arbitrage in, 119–23 and zero-coupon bond
Tax shields, 263 automated, 58, 295 prices/martingales, 599–601
Taxes covered strategies, 314, 317–20 Two-period binomial options pricing
and financial innovation, 7, 8 discrete, 427 model, 396–402
and hedging strategies, 262–64 high-frequency, 425 and backward induction, 396–97
Taylor series, 482–87, 691–92 liquidation-only, 203 and multiperiod binomial options
Technical analysis, 118 naked strategies, 314–16 pricing model, 403–5
TED spread, see Treasury–Eurodollar self-financing strategy, 398, 399 one-step valuation with, 401–2
spread spread, 120–21, 185–88, 320–29 and risk-neutral pricing, 400–1
Tenor, of swaps, 135 Trading at Settlement (TAS) contracts, and synthetic option construction,
Term repos, 42 204 397–400
Term structure, of interest rates, 371, Trading futures, 174–76 Two-period binomial tree, 598–99
502, 567–68, 594, 647 Trading hours, 162, 528
Termination date, FRA, 522 Trading month, 162 UITs (unit investment trusts), 120
Theoretical models, 475–75, 636 Trading rooms, 79 Unbiased expectations (UE) hypothesis,
Theta, 436 Trading (ticker) symbols, 162 505
Third market makers, 59 Trading unit, 162 Uncertainty, lognormal assumption for
35 percent rule, 203 Trading volume, 87–88 stock prices and, 452–53
3Com, 343 Transaction costs, 7, 9 Uncovered trading strategies, for
Tick size, 162, 297, 527 Translation invariance, of risk measures, options, 314, 316
Ticker symbols, 162, 190 667–68 Under-margined accounts, 179
Time Treasury bill futures, 520, 529–30 Underlying, for derivatives, 7, 527
of forwards, 73, 74–75 Treasury bills (T-bills), 43–47 Unit investment trusts (UITs), 120
SUBJECT INDEX 735
United Stock Exchange (USE), 87 Volatility smile, 478 for Treasury securities, 503
US Department of Agriculture (USDA), Volume-weighted average price zero-coupon, 501–2
197 (VWAP), 205 Yield swaps, 549
US Department of Defense, 193–94 VWI (value-weighted index), 122–23 Yield to maturity, see Yield
US dollar, 6, 138
US Mint, 116 Warehouse receipts, for futures, 174 Zero-cost collars, 323, 324, 564
US Treasury, 38–39, 43–47, 368 Warehousing swaps, 541 Zero-coupon bond prices, 32
Up factors, 377, 378, 408 Weak-form efficiency, 118, 119 and binomial HJM model, 573–75,
USDA (US Department of Agriculture), Well-functioning market assumption, 594–95, 599–601
197 215, 375, 424, 566, 623 and forward rate agreements, 631–32
USE (United Stock Exchange), 87 “When, as, and if issued” trading, 40 and forward rates, 517–19
When-issued market, 40 and martingales, 573–75, 599–601
Valuation; see also Risk-neutral valuation Wild card option, for T-bond futures, single-period, 569–70
with binomial options pricing 530 from swap rates, 553
model, 401–2 World Bank, 539–40 term structure of, 567–68, 594
of derivatives, 12 Writers, option, 93 two-period, 594, 595
of interest rate swaps, 544–49 margin requirements for, 298–99 Zero-coupon bonds (zeroes), 7
of plain vanilla currency swaps, 145 payoffs for, 96 compounding and discounting cash
of swaptions, 611–12 profits for, 97–98 flows with, 32–33
Value-at-risk (VaR), 662–66 uncovered, 316 and continuously compounded rate,
Value Line Index, 160 36
Value-weighted index (VWI), 122–23 Yield, 499–507 coupon bonds vs., 43
VaR (value-at-risk), 662–66 banker’s discount, 45–47 in currency swap valuation, 143
Variance, historical volatility and, 443 bond equivalent, 45–47 financial innovation with, 8
Variation margin, 176 computing, 506 moving cash flows across time with,
Vega, 436 convenience, 244–47 34–37
Vega hedging, 439, 470–73, 480, dividend, 63–64, 235–36, 237–39, payoff table for, 517
485–86, 641 344 risk-neutral valuation of, 600, 600–1
Vertical spread, 321–24 and expectations hypothesis, 504–5 risk premiums of, 576
VIX (CBOE Volatility index), 294, 480 and forward rate, 515, 516 in United Kingdom and US, 143
Volatility; see also Vega par bond, 506, 552 yields on, 499–501
annualized, 443 and yield curves, 500–3, 506–7 Zero-coupon yield curve (zero curve),
average forward rate, 632–36 and zero-coupon bonds, 499–500 501–2, 516
of futures prices, 194 Yield auctions, 38 Zero net supply markets, 10, 74
historical, 441–44, 632–36 Yield curves, 500–3, 506–7 Zero-profit point, 313
implied, 444, 477–80, 632, 636 and forward rates, 516 Zero-sum game, 10, 75, 77
and value-at-risk, 664 and swap curves, 552, 554 Zhengzhou Commodity Exchange, 86
You might also like
- (Journal of Portfolio Management, Neuberger) The Log ContractNo ratings yet(Journal of Portfolio Management, Neuberger) The Log Contract7 pages
- Necessary and Sufficient No-Arb Conditions For S3-SSVINo ratings yetNecessary and Sufficient No-Arb Conditions For S3-SSVI13 pages
- Notes On Stochastic Finance-Nicolas PrivautNo ratings yetNotes On Stochastic Finance-Nicolas Privaut564 pages
- Options - 45 Years Since The Publication of The Black-Scholes-Merton Model The Gershon Fintech Center Conference (Zvi Wiener, Alexander Lipton, David Gershon Etc.) (Z-Library)No ratings yetOptions - 45 Years Since The Publication of The Black-Scholes-Merton Model The Gershon Fintech Center Conference (Zvi Wiener, Alexander Lipton, David Gershon Etc.) (Z-Library)554 pages
- Obstfeld, Taylor - Global Capital Markets PDF67% (3)Obstfeld, Taylor - Global Capital Markets PDF374 pages
- The Crisis of Crowding: Quant Copycats, Ugly Models, and the New Crash NormalFrom EverandThe Crisis of Crowding: Quant Copycats, Ugly Models, and the New Crash NormalNo ratings yet
- (Routledge Advances in Risk Management) Kin Keung Lai, Jerome Yen, Shifei Zhou, Hao Wang - Volatility Surface and Term Structure - High-Profit Options Trading Strategies-Routledge (2013)100% (1)(Routledge Advances in Risk Management) Kin Keung Lai, Jerome Yen, Shifei Zhou, Hao Wang - Volatility Surface and Term Structure - High-Profit Options Trading Strategies-Routledge (2013)98 pages
- Munk - 2008 - Financial Asset Pricing Theory100% (2)Munk - 2008 - Financial Asset Pricing Theory360 pages
- Applied Computational Economics and Finance PDFNo ratings yetApplied Computational Economics and Finance PDF521 pages
- New Trends in Energy Derivatives: Alexander Eydeland Morgan StanleyNo ratings yetNew Trends in Energy Derivatives: Alexander Eydeland Morgan Stanley33 pages
- Advance Stochastic Calculus (Abstracts) PDF100% (2)Advance Stochastic Calculus (Abstracts) PDF106 pages
- EXOTIC OPTIONS and HYBRIDS - MOHAMDE BOUZOUBAA, Mohamed Bouzoubaa and Adel Osseiran, - 2010, 2010 - John Wiley & Sons - 9780470688038 - Anna's Archive100% (1)EXOTIC OPTIONS and HYBRIDS - MOHAMDE BOUZOUBAA, Mohamed Bouzoubaa and Adel Osseiran, - 2010, 2010 - John Wiley & Sons - 9780470688038 - Anna's Archive390 pages
- Auction Theory: Pak-Sing Choi Felix Munoz-GarciaNo ratings yetAuction Theory: Pak-Sing Choi Felix Munoz-Garcia304 pages
- Lecture 14 - Mutual Fund Theorem and Covariance Pricing TheoremsNo ratings yetLecture 14 - Mutual Fund Theorem and Covariance Pricing Theorems16 pages
- Algorithmic Trading & Quantitative Strategies Gappy Lecture 2No ratings yetAlgorithmic Trading & Quantitative Strategies Gappy Lecture 225 pages
- Algorithmic Trading & Quantitative Strategies Gappy Lecture 5No ratings yetAlgorithmic Trading & Quantitative Strategies Gappy Lecture 522 pages
- Crowding The Exits - Quantitative Equity Market Portfolio EndogeniNo ratings yetCrowding The Exits - Quantitative Equity Market Portfolio Endogeni43 pages
- Theory of Financial Risks From Statistical Physics To Risk Management Cambridge University Press 2000 PDF100% (1)Theory of Financial Risks From Statistical Physics To Risk Management Cambridge University Press 2000 PDF117 pages
- Oksendal - Stochastic Differential Equations50% (2)Oksendal - Stochastic Differential Equations385 pages
- 0082 A Multi-Factor Model For Energy DerivativesNo ratings yet0082 A Multi-Factor Model For Energy Derivatives20 pages
- Marcus Overhaus, Ana Bermudez, Hans Buehler, Andrew Ferraris, Christopher Jordinson, Aziz Lamnouar - Equity Hybrid Derivatives-Wiley (2007)0% (1)Marcus Overhaus, Ana Bermudez, Hans Buehler, Andrew Ferraris, Christopher Jordinson, Aziz Lamnouar - Equity Hybrid Derivatives-Wiley (2007)334 pages
- An Intuitive Introduction To Finance and Derivatives: Alex Backwell100% (1)An Intuitive Introduction To Finance and Derivatives: Alex Backwell91 pages
- HiddenMarkovModels RobertFreyStonyBrook PDFNo ratings yetHiddenMarkovModels RobertFreyStonyBrook PDF34 pages
- 20 A Long-Run and Short-Run Component Model of Stock Return VolatilityNo ratings yet20 A Long-Run and Short-Run Component Model of Stock Return Volatility23 pages
- Equity and Foreign Exchange Hybrid Models For Pricing Long-Maturity Financial DerivativesnNo ratings yetEquity and Foreign Exchange Hybrid Models For Pricing Long-Maturity Financial Derivativesn184 pages
- Four Derivations of The Black-Scholes FormulaNo ratings yetFour Derivations of The Black-Scholes Formula19 pages
- Options Trader Handbook: Understanding The Relationship Between CME Group FX Options On Futures and OTC OptionsNo ratings yetOptions Trader Handbook: Understanding The Relationship Between CME Group FX Options On Futures and OTC Options24 pages
- Viviana Fanelli (Author) - Financial Modelling in Commodity Markets-Chapman and Hall - CRC (2020)No ratings yetViviana Fanelli (Author) - Financial Modelling in Commodity Markets-Chapman and Hall - CRC (2020)145 pages
- The Swiss Army Knife of Options Analytics: Vola Dynamics100% (1)The Swiss Army Knife of Options Analytics: Vola Dynamics4 pages
- Vdocuments - MX - Synthetic Cdos Modelling Valuation and Risk Management PDFNo ratings yetVdocuments - MX - Synthetic Cdos Modelling Valuation and Risk Management PDF386 pages
- Risk Management Beyond Asset Class Diversification (Page, 2013)No ratings yetRisk Management Beyond Asset Class Diversification (Page, 2013)8 pages
- Hidden Markov Models in Finance Further Developments and Applications100% (1)Hidden Markov Models in Finance Further Developments and Applications280 pages
- Eri WP Predicting Decomposing Risk Data Driven Portfolios 0 PDFNo ratings yetEri WP Predicting Decomposing Risk Data Driven Portfolios 0 PDF46 pages
- Chiarella, C., He, X. Z., Nikitopoulos, C. S. (2015) - Derivative100% (1)Chiarella, C., He, X. Z., Nikitopoulos, C. S. (2015) - Derivative616 pages
- Credit Models and the Crisis: A Journey into CDOs, Copulas, Correlations and Dynamic ModelsFrom EverandCredit Models and the Crisis: A Journey into CDOs, Copulas, Correlations and Dynamic ModelsNo ratings yet
- (Ebook PDF) An Introduction To Derivative Securities, Financial Markets, and Risk Management 2nd Edition Instant DownloadNo ratings yet(Ebook PDF) An Introduction To Derivative Securities, Financial Markets, and Risk Management 2nd Edition Instant Download46 pages
- (Ebook PDF) An Introduction To Derivative Securities, Financial Markets, and Risk Management 2nd Edition PDF Download100% (1)(Ebook PDF) An Introduction To Derivative Securities, Financial Markets, and Risk Management 2nd Edition PDF Download45 pages
- SMM200 Derivatives and Risk Management 2021 QuestionsNo ratings yetSMM200 Derivatives and Risk Management 2021 Questions5 pages
- Hull OFOD10e MultipleChoice Questions and Answers Ch15No ratings yetHull OFOD10e MultipleChoice Questions and Answers Ch156 pages
- New York University Leonard N. Stern School of BusinessNo ratings yetNew York University Leonard N. Stern School of Business22 pages
- Chryss John T. Querol BSBA-2 (Financial Management 1) Journal Critique I. CitationNo ratings yetChryss John T. Querol BSBA-2 (Financial Management 1) Journal Critique I. Citation3 pages
- Sergei Fedotov: 20912 - Introduction To Financial MathematicsNo ratings yetSergei Fedotov: 20912 - Introduction To Financial Mathematics27 pages
- (Ebook PDF) Options Futures and Other Derivatives 9th by John C. Hull PDF Download100% (2)(Ebook PDF) Options Futures and Other Derivatives 9th by John C. Hull PDF Download43 pages
- Group 14 Options Pricing Model Questions Block 3No ratings yetGroup 14 Options Pricing Model Questions Block 39 pages
- Options, Futures, and Other Derivatives: Eleventh Edition (Global) John C. Hull - Ebook PDF Download100% (1)Options, Futures, and Other Derivatives: Eleventh Edition (Global) John C. Hull - Ebook PDF Download50 pages
- Integrated Inventory Valuation in Multi-Echelon Production/distribution SystemsNo ratings yetIntegrated Inventory Valuation in Multi-Echelon Production/distribution Systems16 pages
- Hull: Options, Futures, and Other Derivatives, Tenth Edition Chapter 13: Binomial Trees Multiple Choice Test Bank: Questions With AnswersNo ratings yetHull: Options, Futures, and Other Derivatives, Tenth Edition Chapter 13: Binomial Trees Multiple Choice Test Bank: Questions With Answers8 pages
