0% found this document useful (0 votes)

72 views478 pages

Fixed Income Securities Concepts and Applications 9781547400669 9781547416738 Compress

Uploaded by

moviwojytud3

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

72 views478 pages

Fixed Income Securities Concepts and Applications 9781547400669 9781547416738 Compress

Uploaded by

moviwojytud3

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 478

Sunil Kumar Parameswaran

Fixed Income Securities

Sunil Kumar Parameswaran

Fixed Income
Securities

|
Concepts and Applications
ISBN 978-1-5474-1673-8
e-ISBN (PDF) 978-1-5474-0066-9
e-ISBN (EPUB) 978-1-5474-0068-3

Library of Congress Control Number: 2019948232

Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data are available on the Internet at http://dnb.dnb.de.

© 2020 Sunil Kumar Parameswaran

Published by Walter de Gruyter Inc., Boston/Berlin
Cover image: Freestocker/iStock/Getty Images Plus
Typesetting: VTeX UAB, Lithuania
Printing and binding: CPI books GmbH, Leck

www.degruyter.com
Advance Praise for Fixed Income Securities
This book is a very concise work which collects the major themes of fixed income. For
example, you treat mortgages and their mathematics which are very useful, but which
surprisingly aren’t addressed anywhere else in an accessible way. The abundance of
examples in this book makes it very useful for university professors as a base to teach
fixed income and I will certainly consider adopting this as a book in my university and
recommend it to others.
Prof. Yaacov Kopeliovich, University of Connecticut

Dr. Parameswaran’s new book, Fixed Income Securities: Concepts and Applications,
achieves something that is truly hard—an introduction to the world of fixed income
markets that is both comprehensive and lucid. Each chapter introduces a new con-
cept, and explains the theory through detailed examples. The underlying mathemat-
ics is demystified by working out each step explicitly. The simplicity of the exposition
makes the book suitable not only for a graduate class, but also for an advanced un-
dergraduate class. The accessibility of the book also makes it suitable for self-study.
Prof. Nikunj Kapadia, University of Massachusetts

This book provides comprehensive learning material for fixed income securities and
the related markets. The author does a great job covering both the fundamental con-
cepts and the more advanced applications of fixed income securities. The scope of the
book is broad, yet it also provides the right amount of depth on each topic.
What makes this book stand out is the author’s review of related key finance con-
cepts before diving into the world of fixed income securities.
This book thus can be used as a comprehensive textbook for an undergraduate
course on fixed income securities, an easy-to-absorb self-learning guide for someone
who is new to fixed income securities, and a reference book for more experienced read-
ers who are looking to refresh their knowledge on a specific topic.
Prof. Lingling Wang, University of Connecticut

https://doi.org/10.1515/9781547400669-201
Foreword
I was pleased and honored when Sunil asked me to write a foreword for his upcoming
book, Fixed Income Securities: Concepts and Applications. Sunil and I were graduate
students at Duke University’s Fuqua School of Business in the mid-1980s, completing
our relevant classwork and working for the same professor. While Sunil subsequently
went on to complete his PhD, I left after earning my MBA, and joined Smith Breeden
Associates where I worked for 30 years in the institutional fixed income asset manage-
ment business primarily as a lead portfolio manager. At Smith Breeden, I started in
the research department building complex models to value mortgage backed securi-
ties. After a few years, I was promoted to the portfolio management side. In that role
I managed a billion dollar fixed income portfolio comprised of the same fixed income
securities so well covered in this book by Sunil. Needless to say, I would have been an
avid consumer of this book, if it had been made available a few years earlier. Sunil’s
book covers in adequate detail the theoretical intricacies and formulas that I used reg-
ularly to build models, and focuses on the tools which enabled me to manage highly
complex institutional bond portfolios. Novel investment strategies can blossom from
the most rudimentary concepts in finance with a bit of creativity. An example would
be the S&P 500 plus portfolios I managed which garnered attention in the form of
featured articles from Money magazine, Fortune magazine and other well-read peri-
odicals. I was often asked, “What is a fixed income geek doing managing an equity
product?” The simple answer is that the two product classes are not mutually exclu-
sive. Utilizing the detailed knowledge of various debt financial instruments covered
in this book, Sunil provides the tools and knowledge that professionals like me need
to port or transfer our fixed income expertise over to the equity markets. An in-depth
understanding of the assets and hedging tools covered in this book is the lynchpin to
construct any fixed income portfolio. At Smith Breeden we built portfolios of longer
duration spread assets where we had specific expertise (often Agency MBS) and then
immunized or hedged the portfolio duration to near zero utilizing futures contracts
or interest rate swaps. The remaining negative convexity was hedged utilizing swap-
tions, futures options, or dynamically hedged using futures contracts. The end result
was the creation of a cash like portfolio in terms of duration and convexity risk, but
which retained the asset spread (and risk) of the underlying assets which we desired.
Fixed income geeks like me could now use their expertise in one market to add value
to another market. This opened up a whole new market for us and generated signif-
icant marketing buzz as well. I am of the opinion that the theoretical precision, and
expositional clarity of this book would have been an invaluable asset for me and my
colleagues. This is just one example of how the topics covered in such great detail
here can be utilized by a serious finance student or investment professional. Sunil’s
book covers all these topics and more in the concise and easy to locate manner of a
reference book. I wish I had this book as a resource for all the relevant fixed income

https://doi.org/10.1515/9781547400669-202
viii | Foreword

formulas and Excel commands while I was building high powered fixed income tools
and models. This book is comprehensive, concise, and embellished with the required
expositional clarity. It has the required rigor without being too mathematically heavy
to be out of reach for a reader. Current students and practitioners are lucky to have it
all here.
John Sprow, Chapel Hill, NC, USA
Former Senior Portfolio Manager, Chief Risk Officer, Executive Committee,
Board of Directors at Smith Breeden Associates. Currently retired.
Preface
This book is the end result of 25 years of teaching at business schools across the globe,
and about 15 years of corporate training for some of the leading information technol-
ogy, and financial services companies. The feedback from my students and clients,
has been extremely valuable and I am indebted to them for the inputs and construc-
tive criticism. Nothing is more valuable for teachers and authors, than feedback, both
positive as well as critical.
As a student I was deeply influenced by various books on fixed income products,
in particular, the works of Frank Fabozzi. Later on in the course of my professional
career, I have benefited from the books of various authors such as Kenneth Garbade,
Moorad Choudhary, and Suresh Sundaresan. I owe a tremendous intellectual debt to
these and other authors who have influenced me.
I have over the years made extensive use of Excel while teaching courses and pro-
grams on fixed income products. I am of the opinion that Excel facilitates the exposi-
tion, and aids the comprehension of concepts. In particular, Excel functions make the
computation of complex mathematical expressions extremely easy. This book there-
fore illustrates various Excel functions which are essential for the study of fixed in-
come markets. Each function, and the parameters required to be specified, is dis-
cussed in detail. Goal Seek and Solver, are two tools in Excel that facilitate compu-
tation of solutions for complex non-linear expressions. I have made extensive use of
these and there is an appendix at the end of the book, which explains how to use
them.
This course starts and builds from first principles. All that I expect is that the
reader have a basic knowledge of finance, and an interest in fixed income markets.
The first chapter is on time value of money, which is the fundamental building block
for any study of finance theory. The chapter is detailed, and demonstrates the use of
Excel in elaborate detail to solve problems. There is a detailed discussion of futures
and forwards, options, and swaps toward the end of the book. To help the reader I
have provided a comprehensive primer on derivatives. There are two reasons for this.
The first is to ensure that a reader need not refer to another textbook, if he or she were
to have any queries pertaining to theoretical issues on derivatives. Second, I wanted
to ensure that the material on derivatives is written in my style, so that it seamlessly
blends into my presentations on fixed income products.
This book should be a valuable reference for practitioners and professionals in
financial markets in the English speaking world. Students in MBA programs, and
possibly adventurous students in BBA programs, should also find this book to be an
extremely useful resource. The book has a global focus and perspective, and hence
should cater to readers across geographies.

https://doi.org/10.1515/9781547400669-203
x | Preface

I invite feedback and criticism from readers. Any comments are welcome and I
do not consider anything to be trivial. I hope you get as much pleasure in reading the
book, as I did in writing it.
Sunil Kumar Parameswaran
Director & CEO
Tarheel Consultancy Services
Acknowledgments
I wish to thank the participants who have taken this course at various stages of their
MBA programs, and the executives who have been a part of my corporate training
programs. Their endorsement and criticism of the material, has gone a long way in
making this book acquire its final form. I also owe a deep gratitude to the academics
and practitioners, whose books on fixed income markets have served as invaluable
resources in facilitating my study and understanding of this subject.
I am extremely grateful to Jeff Pepper and Jaya Dalal of De Gruyter for their con-
stant support and encouragement and for ensuring that I stuck to the submission
deadlines set by them. A very special thanks to Nick Wallwork of De Gruyter. I did two
books when Nick was with John Wiley in Singapore. Over the years he has given me
enormous support, and his faith and confidence in me have gone a long way in moti-
vating me to write more for a global audience. Chris Nelson, the editor, has contributed
enormously to make the final product user friendly. He went through the manuscript
with a tooth comb, and his incisive comments and observations have been of immense
help.
I am indebted to Tom Smith, Lingling Wang, Yaacov Kopeliovich, Nikunj Kapadia,
and Jocelyn Evans for their comments and thoughts. Sankarshan Basu, with whom I
have taught this course on a number of occasions, has been of a lot of help, as has
Shrikant Ramamurthy, off whom I have bounced a lot of ideas over the years.
Finally my mother has put up with my moods and tantrums over the past year that
this book has been in progress, and I cannot thank her enough for that.

https://doi.org/10.1515/9781547400669-204
Contents
Advance Praise for Fixed Income Securities | v

Foreword | vii

Preface | ix

Acknowledgments | xi

Chapter 1: A Primer on the Time Value of Money | 1

Nominal and Effective Rates of Interest | 1
Variables and Terms to Be Used and the Corresponding Symbols | 1
The Concept of Simple Interest | 2
The Concept of Compound Interest | 2
Properties of Simple and Compound Interest | 4
Effective Versus Nominal Rates of Interest | 5
A Symbolic Derivation of the Relationship Between Effective and Nominal
Rates of Interest | 5
Computing Effective and Nominal Rates in Excel | 6
Principle of Equivalency of Interest Rates | 7
Continuous Compounding of Interest | 7
Using Excel to Compute the Effective Rate with Continuous
Compounding | 8
Future Value of Cash Flows | 9
Computing the Future Value Using Excel | 9
Present Value of Cash Flows | 11
Computing Present Values of Cash Flows Using Excel | 11
The Internal Rate of Return of an Investment | 12
Pure and Mixed Cash Flows | 13
Descartes’ Rule of Signs and the IRR | 13
A Point About Effective Rates of Interest | 13
Level Annuities | 14
Present Value of a Level Annuity | 15
Future Value of a Level Annuity | 16
Relationship Between PVIFA and FVIFA for a Level Annuity | 16
Level Annuities Due | 17
Present Value of a Level Annuity Due | 17
Computation in Excel of the Present Value of an Annuity Due | 18
Future Value of an Annuity Due | 18
Computation of the Future Value of an Annuity Due in Excel | 19
Relationship Between PVIFA and FVIFA for Annuity Dues | 19
xiv | Contents

Chapter 2: An Introduction to Bonds | 30

Illustration of a Treasury Auction | 49

Chapter 3: Bonds: Advanced Concepts | 65

The Approximate Yield to Maturity Approach | 81

Chapter 4: Yield Curves and the Term Structure | 97

Chapter 5: Duration, Convexity, and Immunization | 118

Appendix 5.4: Convexity of Annuities and Perpetuities | 149

Appendix 5.5: Proof of Single-Period Immunization | 149

Chapter 6: The Money Market | 152

Chapter 7: Floating Rate Bonds | 183

Chapter 8: Mortgage Loans | 203

Features of ARMs | 210

Chapter 9: Mortgage-Backed Securities | 233

Chapter 10: A Primer on Derivatives | 258

Futures and Forwards: Comparisons and Contrasts | 258
The Role of the Clearinghouse in a Futures Trade | 260
Margins for Futures Trades | 261
Contents | xxi

Marking to Market of Futures Contracts | 261

The Exercise Price | 293

Chapter 11: The Valuation of Interest Rate Options | 305

Chapter 12: Interest Rate Forwards and Futures | 341

Chapter 13: Bonds with Embedded Options | 385

Callable Bonds | 385
Yield to Call | 386
xxiv | Contents

Amortizing Swaps | 435

Appendix A | 444
Goal Seek | 444
Solver | 445

Bibliography | 446

Index | 448
Chapter 1
A Primer on the Time Value of Money
All of us have either paid and/or received interest at some point in our lives. Those
of us who have taken educational, housing, or automobile loans have paid interest to
the lending institution. On the other hand, those of us who have deposited funds with
a bank in the form of a savings account or a time deposit have received interest. The
same holds true for people who have bought bonds or debentures.
Interest may be defined as the compensation paid by the borrower of capital to
the lender, for permitting him to use his funds. An economist would define interest as
the rent paid by the borrower of capital to the lender, to compensate him for the loss
of the opportunity to use the funds when it is on loan. After all when we decide not
to live in an apartment or house owned by us, we typically let it out to a tenant. The
tenant will pay us a monthly rental because as long as he is occupying our property,
we are deprived of an opportunity to use it ourselves. The same principle is involved
when it comes to a loan of funds. The difference is that the compensation in the case
of property is termed as rent, whereas when it comes to capital, we term it as interest.

Nominal and Effective Rates of Interest

The quoted rate of interest per period is called the nominal rate of interest. The nom-
inal rate is usually quoted on a per annum basis. The effective rate of interest may
be defined as the interest that a unit of currency invested at the beginning of a year
would have earned by the end of the year. Quite obviously the effective rate will be
equal to the quoted or nominal rate if interest is compounded once per annum. How-
ever, if the interest is compounded more frequently, then the effective rate will exceed
the nominal rate of interest. The term “effective” connotes that compounding at the
stated frequency using the quoted nominal rate, is equivalent to compounding once a
year at the effective rate of interest.

Variables and Terms to Be Used and the Corresponding Symbols

P ≡ amount of principal that is invested at the outset
N ≡ number of periods for which the investment is being made. It may be in terms of
years, or in terms of smaller intervals of time, such as a half-year or a quarter.
r ≡ nominal rate of interest per annum
i ≡ effective rate of interest per annum
m ≡ number of times interest is compounded per annum
P.V. ≡ present value of a stream of cash flows
F.V. ≡ future value of a stream of cash flows

https://doi.org/10.1515/9781547400669-001
2 | Chapter 1: A Primer on the Time Value of Money

The Concept of Simple Interest

Consider the case of an investment of $P that has been made for N years. According
to the principle of simple interest, the interest that will be earned every period is a
constant. In other words, interest is computed every period and credited only on the
original principal, and no interest is payable on any interest that has been accumu-
lated at an intermediate stage.
If r is the quoted rate of interest per year, then an investment of $P will become
dollars P(1 + r) after one year. In the second year, interest will be paid only on P and
not on P(1 + r). Consequently the accumulated value after two years will be P(1 + 2r).
Extending the logic, the terminal balance will be P(1 + rN). N need not be an integer:
that is, investments may be made for a fraction of a year.

Example 1.1. Maureen Chen deposited $10,000 with First National Bank for a period of three years. If
the deposit earns simple interest at the rate of 10% per annum, how much will she have at the end?
An investment of $10,000 will become

10,000 × 1.10 = $11,000

after one year. During the second year, only the original principal of $10,000 will earn interest and not
the accumulated value of $11,000. Consequently the accumulated value after two years will be

10,000 × 1.10 + 1,000 = $12,000

By the same logic the terminal balance after three years will be $13,000.

13,000 = 10,000 × (1 + 0.10 × 3) ≡ P(1 + rN)

Example 1.2. Andrew Gordon deposited $10,000 with ABC Bank five years and six months ago, and
wants to withdraw the balance now. If the bank pays 8% interest per annum on a simple interest basis,
how much is he eligible to withdraw?

P(1 + rN) = 10,000 × (1 + 0.08 × 5.5) = $14,400

Thus Andrew can withdraw $14,400.

The Concept of Compound Interest

Consider the case of an investment of $P that has been made for N years. Assume that
the interest is compounded once per annum: that is, the quoted rate is equal to the
effective rate.
According to the principle of compound interest, every time interest is earned it is
automatically reinvested at the same rate for the next period. Thus, the interest earned
The Concept of Compound Interest | 3

every year will not be a constant like in the case of simple interest, but will steadily
increase. In this case an original investment of $P will become P(1+r) dollars after one
year. The difference is that during the second year the entire amount will earn interest
and consequently the balance at the end of two years will be P(1 + r)2 . Extending the
logic, the balance after N years will be P(1 + r)N . Once again N need not be an integer.

Example 1.3. Assume that Maureen Chen has deposited $10,000 with First National Bank for three
years, and that the bank pays 10% interest per annum compounded annually. How much will she have
at the end?
The investment of $10,000 will become

10,000 × 1.10 = $11,000

after one year. The difference in this case as opposed to the earlier example where simple interest was
considered, is that the entire accumulated value of $11,000 will earn interest during the second year.
Thus, the accumulated value after two years will be

11,000 × 1.10 = $12,100

By the same logic the balance after three years will be $13,310:

12,100 × 1.10 = $13,310 = 10,000 × (1.10)3 ≡ P(1 + r)N

Example 1.4. Assume that Andrew Gordon deposited $10,000 with ABC Bank five years and six months
ago and that the bank has been paying interest at the rate of 8% per annum on a compound interest
basis. The question is, how much can he withdraw?

P(1 + r)N = 10,000 × (1.08)5.5 = $15,269.71

As can be seen from the examples, for a given time period, compounding can yield
substantially more than the return given by simple interest. And since the rate of in-
terest is taken to the power of N, the larger the value of N, the greater will be the impact
of compounding. In other words, the earlier that one starts investing, the greater will
be the return.

Example 1.5. Jesus was born 2019 years ago. Assume that an investment of $1 was made in that year
in a bank which has been paying 1% interest per annum since then, compounded annually. What will
be the accumulated balance at the end of 2019?

1 × (1.01)2019 = $530,705,596

Thus the terminal amount will be in excess of 530 million.

4 | Chapter 1: A Primer on the Time Value of Money

Properties of Simple and Compound Interest

1. If N = 1 (that is, an investment is made for one year) then both the simple and the
compound interest techniques will give the same accumulated value.
2. If N < 1 (that is, the investment is made for less than a year), the accumulated
value using simple interest will be higher. That is

(1 + rN) > (1 + r)N if N < 1

A lot of people are not aware of this. Thus, if a bank were to quote an interest rate
of r% per annum compounded annually, and you were to deposit for nine months,
the payoff would be greater if the bank were to compute the interest on a simple
interest basis. Using similar logic, if a bank were to quote a rate of r% per annum
with semiannual compounding, and you were to deposit for three months, the
payoff would be greater if the bank were to use simple interest.
3. If N > 1 (that is, the investment is made for more than a year), the accumulated
value using compound interest will always be greater. That is

(1 + rN) < (1 + r)N if N > 1

Simple interest is usually used for short-term transactions, that is, for investments for a
period of one year or less. Consequently, simple interest is the norm for money market
calculations.1 However, in the case of capital market securities, that is, medium to
long-term debt securities and equities, we use the compound interest principle. Simple
interest is at times used as an approximation for compound interest over fractional
periods.

Example 1.6. Take the case of Andrew Gordon who has deposited $10,000 with ABC Bank for five years
and six months. Assume that the bank pays compound interest at the rate of 8% per annum for the
first five years and simple interest for the last six months. How much will he be eligible to withdraw?
The balance at the end of five years will be

10,000 × (1.08)5 = 14,693.28

The terminal balance will be

14,693.28 × (1 + 0.08 × 0.5) = $15,281.01

In the earlier example when interest was compounded for five years and six months, the accumulated
value was $15,269.71. The reason why we get a higher value in the second case is that for a fractional
period, which is the last six months in this case, simple interest will give a greater return than com-
pound interest.

1 The money market is the market for debt securities with a time to maturity at the time of issue of one
year or less.
Effective Versus Nominal Rates of Interest | 5

Effective Versus Nominal Rates of Interest

We will first illustrate the difference between nominal rates and effective rates using a
numerical illustration and then derive a relationship between the two symbolically.

Example 1.7. ABC Bank is quoting a rate of 9% per annum compounded annually on deposits placed
with it, whereas XYZ Bank is quoting 8.75% per annum compounded quarterly on funds deposited
with it. From the definitions given at the beginning of the chapter, we can see that the nominal rate
of interest being offered by ABC Bank is 9% per annum, and the nominal rate being offered by XYZ
Bank is 8.75% per annum. In the case of ABC Bank the effective rate of interest is also 9% per annum
because interest is being compounded annually. However, the effective rate of interest being offered
by XYZ Bank will obviously be higher than the rate being quoted by it, since it is compounding on a
quarterly basis. The question is, what is the effective rate being offered by XYZ Bank?
8.75% per annum corresponds to 8.75 4
= 2.1875% per quarter. Consequently, a deposit of $1 for
one year, or four quarters, would accumulate to

1 × (1.021875)4 = 1.090413

Thus, a rate of 8.75% per annum compounded quarterly is equivalent to a rate of 9.0413% with an-
nual compounding. And so, the effective rate of interest being offered by XYZ Bank is 9.0413%. So,
when the frequencies of compounding are different, comparisons between alternative investments
ought to be based on the effective rates of interest and not on the nominal rates. In this case, although
ABC Bank is offering a higher nominal rate, the investor would obviously be better off by investing with
XYZ Bank.

It must be remembered that the distinction between nominal and effective rates is of
relevance only when compound interest is being paid. The concept is of no conse-
quence if simple interest is being paid.

A Symbolic Derivation of the Relationship Between Effective and Nominal Rates of

Interest

Assume that a borrower is offering a nominal rate of r% per annum, and that interest
is being compounded m times per annum. The effective rate of interest i is therefore
given by

r m
(1 + i) = (1 + ) (1.1)
m

We can also derive the equivalent nominal rate if the effective rate is given:
1
r = m[(1 + i) m − 1] (1.2)
6 | Chapter 1: A Primer on the Time Value of Money

Example 1.8. ABC Bank is offering 10% per annum compounded quarterly. If an investor deposits
$10,000 with the bank, how much will he have after one year?
The terminal value will be

0.10 4
10,000 × (1 + ) = $11,038.13
4

The effective annual rate is therefore

(11,038.13 − 10,000)
≡ 10.3813%
10,000

Now assume that ABC Bank wants to offer an effective annual rate of 10% per annum with quar-
terly compounding. The question is, what nominal rate of interest should it quote?
1
We know that r = m[(1 + i) m − 1]. Therefore,

r = 4[(1.10)0.25 − 1] = 0.096455 ≡ 9.6455%

Thus ABC Bank should quote a rate of 9.6455% per annum if it wants to offer an effective annual rate
of 10% per annum.

Computing Effective and Nominal Rates in Excel

Let’s revisit Example 1.8. The nominal annual rate is 10% per annum, and the fre-
quency of compounding is four times per annum. To calculate the effective annual
rate, we use an Excel function called EFFECT. The parameters are:
– Nominal_rate: This is the nominal rate of interest per annum.
– Npery: This is the frequency of compounding per annum.

The nominal rate is 10% or 0.10 in this case. The frequency of compounding per annum
is four. Using the function, we get the effective annual rate of 10.3813%:

EFFECT(0.10, 4) = 10.3813%.

If we are given the effective rate, we can compute the equivalent nominal rate
using the NOMINAL function in Excel. The parameters are
– Effect_rate: This is the effective rate of interest per annum.
– Npery: This is the frequency of compounding per annum.

In this case the effective rate is 10%, and the frequency of compounding is four. Thus,

NOMINAL(0.10, 4) = 9.6455%
Computing Effective and Nominal Rates in Excel | 7

Principle of Equivalency of Interest Rates

Two nominal rates of interest compounded at different intervals of time are said to
be equivalent if a given principal invested for the same total length of time at each of
the two rates, produces the same accumulated value at the end. In other words, two
nominal rates compounded at different intervals of time may be said to be equivalent
if they yield the same effective interest rate.
Assume that ABC Bank is offering 9% per annum with semiannual compounding.
What should be the equivalent rate offered by XYZ Bank if it intends to compound
interest quarterly?
The first step is to calculate the effective annual rate being offered by ABC Bank:
0.09 2
(1 + i) = (1 + ) = 1.092025
2
The next step is to calculate the nominal annual rate that gives the same effective an-
nual rate with quarterly compounding:

r = 4[(1.092025)0.25 − 1] = 0.0890 ≡ 8.90%

Thus 9% per annum with semiannual compounding is equivalent to 8.90% per annum
with quarterly compounding. That is, given a choice between a deposit yielding 9%
per annum with semiannual compounding, and an alternate deposit yielding 8.90%
with quarterly compounding, an investor will be indifferent.

Continuous Compounding of Interest

Consider an amount of $P that is invested for N years at a nominal annual rate of r%

per period. If interest is compounded m times per year, we know that the accumulated
value after N years will be
r mN
P(1 + )
m
In the limit as m → ∞
r mN
P(1 + ) → PerN
m
where e = 2.71828. This is called the case of continuous compounding. If r is the nomi-
nal annual rate, then the effective annual rate with continuous compounding is er − 1.

Example 1.9. Norah Roberts has deposited $10,000 with COZY Bank for a period of five years at 10%
per annum compounded continuously. What will be the final balance in the account?

10,000 × e0.10×5 = 10,000 × 1.6487 = $16,487

8 | Chapter 1: A Primer on the Time Value of Money

Example 1.10. COZY Bank is quoting an interest rate of 10% per annum with quarterly compounding.
What is the equivalent rate with continuous compounding?
We know that two nominal rates are equivalent if they give the same effective annual rate. Let r
be the nominal rate with continuous compounding and k the equivalent nominal rate with quarterly
compounding. We require that

k m
er = (1 + )
m
k
⇒ r = m × ln (1 + ) (1.3)
m
where ln is the natural logarithm of the expression in parentheses.
k r
⇒ (1 + ) = em
m
r
⇒ k = m[e m − 1] (1.4)

In this case,
0.10
r = 4 × ln (1 + ) = 0.09877 ≡ 9.877%
4

Using Excel to Compute the Effective Rate with Continuous Compounding

Let’s first revisit Example 1.9. To compute the terminal value using continuous com-
pounding, we can use the EXP function in Excel. In this case,

10,000 × EXP(0.10 × 5) = 16,487

An alternative, using Excel, is to compute the effective annual rate, by specifying a

high compounding frequency of, say, 500,000. Compounding 500,000 per annum, in
a year that is assumed to consist of 360 days, is equivalent to compounding 58 times
per hour, which is essentially the same as continuous compounding.
EFFECT(0.1,500,000) = 10.5171%

10,000 × (1.105171)5 = $16,487

Now turn to Example 1.10. Given a rate of 10% per annum with quarterly compound-
ing, we seek to find the equivalent rate with continuous compounding.
First let’s find the effective annual rate:

EFFECT(0.1, 4) = 10.3813%

Let’s then convert this effective rate to an equivalent nominal rate with continuous
compounding:

NOMINAL(0.103813, 500 000) = 9.877%

Thus 10% per annum with quarterly compounding, is equivalent to a rate of

9.877% per annum with continuous compounding.
Future Value of Cash Flows | 9

Future Value of Cash Flows

We have already encountered the concept of future value in this chapter although we
did not term it as such. When an amount is deposited for a certain time period at a
given rate of interest, the amount that is accrued at the end of the designated period
of time is called the future value of the original investment. So if $P is invested for N
periods at a periodic interest rate of r%, then the future value of the investment after
N periods is given by

F.V. = P(1 + r)N (1.5)

The expression (1 + r)N is the amount to which an investment of $1 will grow at

the end of N periods, if it is invested at a rate r. It is called the FVIF (future value
interest factor). It is a function of r and N. The advantage of knowing the FVIF is that
we can find the future value of any principal amount, for given values of the interest
rate and the time period, by simply multiplying the principal by the factor. The process
of finding the future value given an initial investment is called compounding.

Example 1.11. Sonia Smith has deposited $10,000 for five years in an account that pays interest at the
rate of 10% per annum compounded annually. What is the future value of her investment?

FVIF (10, 5) = (1.10)5 = 1.6105

Thus,

F .V . = $10,000 × 1.6105 = $16,105

Computing the Future Value Using Excel

To compute the future value using Excel, we need to use the FV function.2 The param-
eters are
– Rate: Rate is the periodic interest rate, which in this case is 10% or 0.10.
– Nper: Nper is the number of periods which is 5 in this case.
– Pmt: Pmt stands for the periodic payment, and is not applicable in this case, be-
cause there are no periodic cash flows. Thus, we can either put a zero, or else an
extra comma in lieu.

2 We are using F.V. to denote the future value of a stream of cash flows, and FV to denote the Future
Value function in Excel. Similarly, we are using P.V. to denote the present value of a stream of cash
flows, while PV is being used to denote the Present Value function in Excel.
10 | Chapter 1: A Primer on the Time Value of Money

– Pv: Pv stands for the present value, or the initial investment, which in this case is
$10,000. We input it as −10,000 in order to ensure that the answer is positive. In
many Excel functions, cash flows in one direction are positive, while those in the
opposite direction are negative. Thus if the investment is positive, the subsequent
inflow is negative and vice versa. In this case, if we specify a negative number for
the present value, we get the future value with a positive sign. If, however, the
present value were to be given with a positive sign, the future value, although it
would have the same magnitude, would have a negative sign.
– Type: This is a binary variable which is either 0 or 1. It is not required at this stage,
and we can just leave it blank.3

We will invoke the function as, FV(0.1, 5, , −10,000), and the answer is $16,105. In this
function we are inputting an extra comma lieu of the value for Pmt. As an alternative
we could have given the value as zero.

Example 1.12. Sharon Peters has deposited $10,000 for four years in an account that pays a nomi-
nal annual interest of 10% per annum with semiannual compounding. What is the future value of her
investment?
10% per annum for four years, with semiannual compounding, is equivalent to 5% per half-year
for eight half-yearly periods.

FVIF (5, 8) = (1.05)8 = 1.4775

Thus,

F .V . = $10,000 × 1.4775 = $14,775

In Excel, the parameters would be: Rate = 0.05, Nper = 8, and Pv = −10, 000.

FV(0.05, 8, , −10,000) = $14,775

Example 1.13. Let’s assume that Sonia deposits $10,000 for five years and six months at a nominal
annual rate of 10% with annual compounding. What will be the future value?

F .V . = 10,000 × (1.10)5.5
⇒ ln(F .V .) = ln(10,000) + 5.5 × ln(1.10) = 9.2103 + 0.5242 = 9.7345
⇒ F .V . = e9.7345 = $16,891

Once again ln stands for the natural logarithm. In Excel, of course, it is simple:

FV(0.10,5.5, , −10,000) = $16,891.

3 We will use it later while dealing with annuities and annuities due.
Present Value of Cash Flows | 11

Present Value of Cash Flows

When calculating the future value of an investment, we seek to determine the termi-
nal value that will be earned, given a rate of interest for a specified period. Now let’s
look at an investment from another angle. Suppose we want our investment to yield a
given terminal value. How much should we invest at the outset, if the applicable rate
of interest is r% and the number of periods is N. That is, instead of computing the ter-
minal value of a given principal, we seek to compute the principal that corresponds
to a given terminal value. The principal amount that is obtained is called the present
value of the terminal amount.
Assume that an investment pays an interest rate of r% per period on a compound
interest basis. If we want to have $F after N periods, how much should we deposit
today? Quite obviously
F
P.V. = (1.6)
(1 + r)N

Example 1.14. Paula wants to deposit an amount of $P with her bank in order to ensure that she has
$15,000 at the end of three years. If the bank pays 10% interest per annum compounded annually,
how much does she have to deposit today?

15,000
P= = $11,269.7220
(1.10)3

The expression 1 N is the amount that must be invested today if we are to have $1
(1+r)
at the end of N periods, in the case of an investment that pays interest at the rate of
r% per period. It is called the PVIF (present value interest factor). It is a function of
r and N. The advantage of knowing the PVIF is that we can find the present value of
any terminal amount, for given values of the interest rate and time period, by simply
multiplying the future amount by the factor. The process of finding the principal cor-
responding to a given future amount is called discounting and the interest rate that is
used is called the discount rate.
There is a relationship between the PVIF and the FVIF for given values of r and N.
One is simply a reciprocal of the other.

Computing Present Values of Cash Flows Using Excel

The required function in Excel is PV. The parameters are:

– Rate
– Nper
– Pmt
– Fv
– Type
12 | Chapter 1: A Primer on the Time Value of Money

Fv stands for the future value. The other parameters have the same meaning as spec-
ified for the FV function. Let’s revisit Example 1.14.

PV(0.10, 3, , −15,000) = $11,269.72.

As before, the extra comma is in lieu of the input for Pmt.

The Internal Rate of Return of an Investment

Suppose we are told that an initial investment of $18,000 will entitle us to receive the
cash flows depicted in Table 1.1. The question is, what is the rate of return that we are
being offered?4

Table 1.1: Vector of cash flows.

Year Cash Flow

0 (18,000)
1 2,500
2 4,000
3 5,000
4 7,500
5 10,000

The rate of return r is obviously the solution to the following equation:

2,500 4,000 5,000 7,500 10,000

18,000 = + + + +
(1 + r) (1 + r)2 (1 + r)3 (1 + r)4 (1 + r)5

The solution to this equation is termed as the internal rate of return (IRR). It can be ob-
tained using the IRR function in Excel.5 The parameters are Values and Guess, though
Guess is usually not required for cash flow streams like this with a single sign change.
We will have more to say about this shortly. If we assume that the data in Table 1.1 is
given in columns A and B of the Excel sheet, we would specify the values as B1:B6,
and invoke the function as IRR(B1:B6). In this case the solution is 14.5189%.

4 Numbers in parentheses denote cash outflows. Thus the parentheses are the equivalent of a negative
sign.
5 We are using the acronym IRR to denote both the internal rate of return and its corresponding func-
tion in Excel.
A Point About Effective Rates of Interest | 13

Pure and Mixed Cash Flows

A vector of cash flows with one sign change is referred to as a pure cash flow, whereas
a vector with more than one sign change is referred to as a mixed cash flow. Table 1.2
provides an illustration of pure and mixed cash flows.

Table 1.2: Pure and mixed cash flows.

Year Pure Mixed

Cash Flow Cash Flow

0 (20,000) (20,000)
1 2,500 2,500
2 5,000 5,000
3 8,000 (8,000)
4 8,000 12,500
5 6,000 6,000

Column two in Table 1.2 represents a pure cash flow, for the sign changes only once,
that is, between Year 0 and Year 1. However in the third column the sign changes three
times. It first changes between Year 0 and Year 1, then between Year 2 and Year 3, and
finally between Year 3 and Year 4. Thus column three of Table 1.2 represents a mixed
cash flow.

Descartes’ Rule of Signs and the IRR

The IRR is a solution to a polynomial of degree N and consequently has N roots.

Descartes’ rule of signs states that the number of real positive roots is either equal to
the number of sign changes, or less than it by an even number. Multiple roots of the
same value are counted separately. The reason why the number of real positive roots
is less than the maximum number by a multiple of two, is because imaginary roots
always come in conjugate pairs.
Thus in Table 1.2, the pure cash flow can have only one real positive IRR. However,
the mixed cash flow may have a maximum of three real positive IRRs.

A Point About Effective Rates of Interest

Suppose that we are asked to calculate the present value or future value of a series of
cash flows arising every six months, and are given a rate of interest in annual terms,
without specifying the frequency of compounding. The normal practice is to divide
the annual rate by two to determine the periodic interest rate for discounting or com-
pounding. That is, the quoted rate is treated as the nominal annual rate and not as
14 | Chapter 1: A Primer on the Time Value of Money

the effective annual rate. In general, if we are given an annual rate of interest and the
cash flows arise at intervals of time equal to n years, then the periodic rate is taken as
r
1 . For instance, if the cash flows arise on a quarterly basis, then n = 0.25 years, and
n
the periodic rate will be 4r .

Example 1.15. Consider the series of cash flows depicted in Table 1.3. Assume that the annual rate of
interest is given to be 10%.

Table 1.3: Vector of cash flows.

Period Cash Flow

6 Months 1,000
12 Months 2,000
18 Months 3,000
24 Months 4,000

The present value will be calculated as

1,000 2,000 3,000 4,000

P.V . = + + + = $8,648.7627
(1.05) (1.05)2 (1.05)3 (1.05)4

Similarly the future value will be

F .V . = 1,000 × (1.05)3 + 2,000 × (1.05)2 + 3,000 × 1.05 + 4,000 = $10,512.625

However, if it were to be explicitly stated that the effective annual rate is 10%, then the calcula-
tions would change. The present value will then be given by

1,000 2,000 3,000 4,000

P.V . = + + + = $8,677.782
0.5
(1.10) (1.10) (1.10)1.5 (1.10)2

Similarly the future value will then be given by

F .V . = 1,000 × (1.10)1.5 + 2,000 × (1.10) + 3,000 × (1.10)0.5 + 4,000

= $10,500.1162

The present value is higher when we use an effective annual rate of 10% for discounting. This is be-
cause the lower the discount rate, the higher the present value, and an effective rate of 10% per annum
is obviously lower than a nominal annual rate of 10% with half-yearly compounding. By the same logic
the future value is lower when we use an effective annual rate of 10% to compound.
Similarly, if we were to calculate the IRR for this cash flow stream, we would get a semiannual
rate of return. We would then multiply it by two to get the annual rate of return.

Level Annuities
A level annuity is a series of identical payments made at equally spaced intervals of
time. Examples include house rent (until the rent is revised), monthly salary (until the
Level Annuities | 15

salary is revised), insurance premiums, and monthly installments on housing loans

and automobile loans.
In the case of an ordinary annuity, the first payment will be made one period from
now. The interval between successive payments is called the payment period. We as-
sume that the payment period is the same as the time period over which interest is
compounded. That is, if the annuity pays annually, we assume annual compounding,
whereas if it pays semiannually, we assume half-yearly compounding and so on.6
An annuity which makes periodic payments of $A for N periods is depicted in
Figure 1.1, where today is depicted as time 0.

A A A A A

0 1 2 3 N −1 N

Figure 1.1: A level annuity.

If the applicable rate of interest is r%, then we can calculate the present and future
values as explained in the following two sections.

Present Value of a Level Annuity

The present value of a level annuity is given by the following equation:

A A A A
P.V. = + + +−−−−+
(1 + r) (1 + r)2 (1 + r)3 (1 + r)N
A A A
Therefore, P.V.(1 + r) = A + + +−−−−+
(1 + r) (1 + r)2 (1 + r)N−1
A
⇒ P.V.[(1 + r) − 1] = A −
(1 + r)N
A 1
⇒ P.V. = [1 − ] (1.7)
r (1 + r)N
1
r
[1 − 1 N ] is called the present value interest factor annuity (PVIFA). PVIFA(r, N)
(1+r)
is the present value of an annuity that pays $1 at periodic intervals for N periods,
computed using a discount rate of r%. The present value of any annuity that pays
$A per period can therefore be computed by multiplying A by the appropriate value of
PVIFA.

6 We can handle more complex cases. That is, payments may come annually and interest may be com-
pounded quarterly, or else payments may come quarterly and interest may be compounded annually.
16 | Chapter 1: A Primer on the Time Value of Money

Example 1.16. Apex Corporation is offering an instrument that promises to pay $1,000 per year for 20
years, beginning one year from now. If the annual rate of interest is 5%, what is the present value of
the annuity?
The present value of the annuity using a discount rate of 5% is

1,000 × PVIFA(5, 20) = 1,000 × 12.4622 = $12,462.20

The same can be computed using Excel. The parameters are: Rate = 0.05, Nper = 20, Pmt = −1,000.
There is no need to input parameters for Fv and Type. This is because there is no lump sum terminal
cash flow, and so there is no need to input a value for the future value. Type needs to be input only for
annuities due, as we shall demonstrate shortly.

PV(0.05, 20, −1000) = $12,462.21

Future Value of a Level Annuity

The future value of a level annuity is given by the following equation:

F.V. = A(1 + r)N−1 + A(1 + r)N−2 + A(1 + r)N−3 + − − − − + A

Therefore, F.V.(1 + r) = A(1 + r)N + A(1 + r)N−1
+ A(1 + r)N−2 + − − − − + A(1 + r)
⇒ F.V.[(1 + r) − 1] = A(1 + r)N − A
A
⇒ F.V. = [(1 + r)N − 1] (1.8)
r
1
r
+ r)N − 1] is called the future value interest factor annuity (FVIFA). It is the
[(1
future value of an annuity that pays $1 per period for N periods, where interest is
compounded at the rate of r% per period. Once again, if we know the factor, we can
calculate the future value of any annuity that pays $A per period.

Example 1.17. Patricia Beck expects to receive $10,000 per year for the next five years, starting one
year from now. If the cash flows can be invested at 10% per annum, how much will she have after five
years?

F .V . = 10,000 × FVIFA(10, 5) = 10,000 × 6.1051 = $61,051

In Excel FV(0.10, 5, −10,000) = $61,051.

Relationship Between PVIFA and FVIFA for a Level Annuity

There is a relationship between the present value interest factor and the future value
interest factor for an annuity.
1 1 (1 + r)N − 1
PVIFA = [1 − ] =
r (1 + r)N r(1 + r)N
Level Annuities Due | 17

(1 + r)N − 1 1
= ×
r (1 + r)N
= FVIFA × PVIF (1.9)

Conversely

PVIFA
FVIFA = = PVIFA × FVIF (1.10)
PVIF

Level Annuities Due

The difference between an annuity and an annuity due is that in the case of an annuity
due the cash flows occur at the beginning of the period. An N period annuity due that
makes periodic payments of $A may be depicted as shown in Figure 1.2. The first cash
flow occurs at time 0, and the last cash flow occurs at time N − 1.

A A A A A

0 1 2 3 N −1 N

Figure 1.2: A level annuity due.

Present Value of a Level Annuity Due

The present value of a level annuity due is given by the following equation:

A A A
P.V. = A + + +−−−−+
(1 + r) (1 + r)2 (1 + r)N−1
A A
Therefore, P.V.(1 + r) = A(1 + r) + A + +−−−−+
(1 + r) (1 + r)N−2
A
⇒ P.V.[(1 + r) − 1] = A(1 + r) −
(1 + r)N−1
A 1
⇒ P.V. = [1 − ] (1 + r)
r (1 + r)N
Hence PVIFAAD (r, N) = PVIFA(r, N) × (1 + r) (1.11)

The present value of an annuity due that makes N payments is obviously greater
than that of a corresponding annuity that makes N payments, because in the case of
the annuity due, each of the cash flows has to be discounted for one period less.
An obvious example of an annuity due is an insurance policy, because the first
premium is due as soon as the policy is purchased.
18 | Chapter 1: A Primer on the Time Value of Money

Example 1.18. David Mathew has just bought an insurance policy. The annual premium is $12,000,
and he is required to make 25 payments. What is the present value of this annuity due if the discount
rate is 10% per annum:

12,000 1
P.V . = [1 − ] × 1.10
0.10 (1.10)25
= $119,816.93

Computation in Excel of the Present Value of an Annuity Due

The present value of the annuity due considered in Example 1.18 can be computed
using Excel as follows: Rate = 0.10, Nper = 25, Pmt = −12,000, Type = 1

PV(0.10, 25, −12,000, , 1) = $119,816.93.

The extra comma is to account for the fact that there is no terminal lump sum cash
flow, and consequently no FV.
Alternately, we can find the present value of an ordinary annuity that makes 24
payments and add the initial cash flow. That is:

PV(0.10, 24, −12,000) + 12,000 = $119,816.93

The third approach is to find the value of an ordinary annuity that makes 25 pay-
ments and then multiply the result by one plus the interest rate:

PV(0.10, 25, −12,000) × 1.10 = $119,816.93

Future Value of an Annuity Due

The future value of a level annuity due is given by the following equation.

F.V. = A(1 + r)N + A(1 + r)N−1 + A(1 + r)N−2 + − − − − + A(1 + r).

Therefore, F.V(1 + r) = A(1 + r)N+1 + A(1 + r)N + A(1 + r)N−1
+ − − − − + A(1 + r)2
⇒ F.V.[(1 + r) − 1] = A(1 + r)N+1 − A(1 + r)
A
⇒ F.V. = [(1 + r)N − 1](1 + r)
r
Hence FVIFAAD (r, N) = FVIFA(r, N) × (1 + r) (1.12)

The future value of an annuity due that makes N payments is higher than that of a
corresponding annuity that makes N payments if the future values in both cases are
Perpetuities | 19

computed at the end of N periods. This is because, in the first case, each cash flow has
to be compounded for one period more.

Example 1.19. In the case of Mathew’s insurance policy, the cash value at the end of 25 years can be
calculated as follows:

(1.10)25 − 1
F .V . = 12,000 × [ ] × 1.10
.10
= $1,298,181.19

Computation of the Future Value of an Annuity Due in Excel

Consider the data given in Example 1.19.

Rate = 0.10, Nper = 25, Pmt = −12,000, Type = 1

FV(0.10, 25, −12,000, , 1) = $1,298,181.19.

The second approach is to find the future value of an ordinary annuity that makes
25 payments and then multiply the result by one plus the interest rate:

FV(0.10, 25, −12000) × 1.10 = $1,298,181.19

Relationship Between PVIFA and FVIFA for Annuity Dues

Like in the case of an annuity, there is a relationship between the present value interest
factor and the future value interest factor for an annuity due.

PVIFAAD (r, N) = PVIFA(r, N) × (1 + r) = FVIFA(r, N) × PVIF × (1 + r)

= FVIFAAD (r, N) × PVIF (1.13)
Conversely, FVIFAAD (r, N) = FVIFA(r, N) × (1 + r) = PVIFA(r, N) × FVIF × (1 + r)
= PVIFAAD (r, N) × FVIF (1.14)

Perpetuities
An annuity that pays forever is called a perpetuity. The future value of a perpetuity is
obviously infinite. But it turns out that a perpetuity has a finite present value.
The present value of an annuity that pays for N periods is

A 1
P.V = [1 − ]
r (1 + r)N
20 | Chapter 1: A Primer on the Time Value of Money

The present value of the perpetuity can be found by letting N tend to infinity. As N →
∞, 1 N → 0. The present value of a perpetuity is therefore Ar .
(1+r)

Example 1.20. Consider a financial instrument that promises to pay $1,000 per year forever. If you
require a 20% rate of return, how much should you be willing to pay for it?
1,000
P.V . = = $5,000
0.20

A perpetuity due pays an additional cash flow of $A at time 0, that is, right at the
outset. Thus, its present value is
A A(1 + r)
+A= (1.15)
r r
Thus the present value of a perpetuity due is equal to the present value of a perpetuity
multiplied by one plus the rate of interest. This is not surprising since a perpetuity is
a special case of an annuity.

The Amortization Method of Loan Repayment

The amortization process refers to the process of repaying a loan by means of equal in-
stallment payments at periodic intervals. It is obvious that the installment payments
form an annuity whose present value is equal to the original loan amount. Each in-
stallment will consist of a partial repayment of principal, and payment of interest on
the outstanding loan balance. An amortization schedule is a table that shows the di-
vision of each payment into a principal component and interest component, together
with the outstanding loan balance after each payment is made.
Consider a loan which is repaid in N installments of $A each. We will denote the
original loan amount by L and the periodic interest rate by r.
A 1
L = A × PVIFA(r,N) = [1 − ] (1.16)
r (1 + r)N
The interest component of the first installment:
A 1
=r× [1 − ]
r (1 + r)N
1
= A [1 − ] (1.17)
(1 + r)N
The principal component:
1
= A − A [1 − ]
(1 + r)N
A
= (1.18)
(1 + r)N
The Amortization Method of Loan Repayment | 21

The outstanding balance at the end of the first payment

A 1 A
= [1 − ]−
r (1 + r) N
(1 + r)N
A 1
= [1 − ] (1.19)
r (1 + r)(N−1)

In general, the interest component of an installment is

1
A [1 − ]
(1 + r)N−t+1

where t represents the number of time periods from time zero; the principal compo-
nent of an installment is
A
;
(1 + r)N−t+1

and the outstanding balance after making the principal payment is

A 1
[1 − ]
r (1 + r)N−t

Example 1.21. Sharon has borrowed $10,000 from Royal Syndicate Bank and has to pay it back in five
equal annual installments. If the interest rate is 10% per annum on the outstanding balance, what is
the installment amount? Draw up an amortization schedule.
Let’s denote the unknown installment amount by A. We know that

A 1
10,000 = × [1 − ]
0.10 (1.10)5
= A × 3.7908 ⇒ A = 2,637.97

We will analyze the first few entries in Table 1.4 in order to clarify the principles involved. At time
0, the outstanding principal is $10,000. After one period, a payment of $2,637.97 is made. The inter-
est due for the first period is 10% of $10,000, which is $1,000. So obviously, the excess payment of

Table 1.4: Amortization schedule.

Year Payment Interest Principal Outstanding

Repayment Principal

0 – – – 10000.00
1 2637.97 1000.00 1637.97 8362.03
2 2637.97 836.20 1801.77 6560.25
3 2637.97 656.03 1981.95 4578.30
4 2637.97 457.83 2180.14 2398.16
5 2637.97 239.82 2398.16 0.00
22 | Chapter 1: A Primer on the Time Value of Money

$1,637.97 represents a partial repayment of principal. Once this amount is adjusted, the outstanding
balance at the end of the first period becomes $8,362.03. At the end of the second period, another in-
stallment of $2,637.97 is paid. The interest due for this period is 10% of $8,362.03, which is $836.20.
The balance, which is $1,801.77, constitutes a partial repayment of principal.
The value of the outstanding principal at the end is zero. As can be seen, the outstanding prin-
cipal declines after each payment. Since the payments are constant, the interest component steadily
declines while the principal component steadily increases. Thus, if a loan is repaid in this fashion, the
initial periodic payments will largely consist of interest payments, and the payments towards the end
will largely consist of principal repayments.

Obtaining the Amortization Schedule Using Excel

Consider, once again, the loan of $10,000, which has to be paid back in five equal
periodic installments. The periodic interest rate is 10%.
The periodic installment can be computed using the PMT function in Excel. The
parameters are:
– Rate
– Nper
– Pv
– FV
– Type

There are two ways in which the PMT function can be invoked for periods other than
the first. For the first period, PMT(0.10, 5, −10,000) = $2,637.97. Now consider the sec-
ond period. We can specify the Nper as 4, and the PV as the outstanding principal,
which is 8,362.03. Thus,

PMT(0.10, 5, −10,000) = PMT(0.10, 4, −8,362.03) = $2,637.97

Similarly,

PMT(0.10, 5, −10,000) = PMT(0.10, 3, −6,560.25) = $2,637.97

Now consider the interest and principal components of each installment. We can
use a function in Excel called IPMT to compute the interest component of an install-
ment and another function called PPMT to compute the principal component of the
installment. The parameters, for both, are
– Rate: This is the periodic interest rate.
– Per: This stands for period and will be explained shortly.
– Nper: This represents the total number of periods.
– Pv: This is the present value.
– Fv: This is the future value.
– Type: This has the usual meaning.
Amortization with a Balloon Payment | 23

Consider the interest and principal components of the first installment.

IPMT(0.10, 1, 5, −10,000) = $1,000. While computing the interest component of
the second installment, we can invoke IPMT as IPMT(0.10, 2, 5, −10,000) or as
IPMT(0.10, 1, 4, −8,362.03). Both will return a value of $836.20. Similarly, the prin-
cipal component of the first installment is PPMT(0.10, 1, 5, −10,000) = $1,637.97. For
the second period, PPMT(0.10, 2, 5, −10,000) = PPMT(0.10, 1, 4, −8,362.03) = $1,801.77.

The Rationale for Why IPMT and PPMT Can Be Used with Two Different Sets of
Parameters

IPMT and PPMT can be used with two sets of parameters. We can keep the total num-
ber of periods at the initial value, specify the present value as the initial loan amount
and keep changing Per to compute the interest and principal components. For the first
installment, Per = 1, and for the nth installment, it is equal to n. The alternative is to re-
amortize the outstanding at the beginning of each period over the remaining number
of periods. Remember that each time we re-amortize, we are back to the first period.
Thus, IPMT(0.10, 2, 5, −10,000) = IPMT(0.10, 1, 4, −8,362.03). After every payment, we
are back to the first period of a loan whose life is equal to the remaining time to matu-
rity, and whose principal amount is equal to the remaining outstanding balance.

Amortization with a Balloon Payment

Victoria Tolkin has taken a loan of $100,000 from ABC Bank. The loan requires her to
pay in five equal annual installments along with a terminal payment of $25,000. This
terminal payment, that, has to be made over and above the scheduled installment
in year five, is called a balloon payment. The interest rate is 10% per annum on the
outstanding principal. The annual installment may be calculated as follows:

A 1 25,000
100,000 = × [1 − ]+
0.10 (1.10)5
(1.10)5
⇒ A = 22,284.81

Obviously, the larger the balloon, the smaller the periodic installment payment is for a
given loan amount. The amortization schedule may be depicted as shown in Table 1.5.

Handling the Balloon Using Excel

The only difference is that in the functions PMT, IPMT, and PPMT, we specify the FV as
25,000, which represents the balloon. Remember that PV and FV must have opposite
signs. In the end, instead of going to zero, the outstanding balance will go to 25,000.
24 | Chapter 1: A Primer on the Time Value of Money

Table 1.5: Amortization schedule with a balloon payment.

Year Payment Interest Principal Outstanding

Repayment Principal

0 – – – 100,000.00
1 22,284.81 10,000.00 12,284.81 87,715.19
2 22,284.81 8,771.52 13,513.29 74,201.90
3 22,284.81 7,420.19 14,864.62 59,337.28
4 22,284.81 5,933.73 16,351.08 42,986.20
5 47,284.81 4,298.62 42,986.19 0.01

When this is paid out in the form of a lump sum, the outstanding balance will be zero.
Once again,

IPMT(0.10, 2, 5, −100,000, 25,000) = IPMT(0.10, 1, 4, −87715.19, 25000) = $8,771.52

A Growing Annuity
In a growing annuity, the cash flows are not constant, but increase over time. Thus
every period they grow at i%. The periodic discount rate is r%.
Consider an annuity that makes a payment of A(1 + i) after one period, and whose
cash flows grow at a rate of i% every period. The discount rate is r%. The cash flows
are as shown in Figure 1.3, where today is depicted as time 0.

A(1 + i) A(1 + i)2 A(1 + i)3 A(1 + i)N−1 A(1 + i)N

0 1 2 3 N −1 N

Figure 1.3: A growing annuity.

Present Value of a Growing Annuity

The present value of a growing annuity is given by the following equation:

A(1 + i) A(1 + i)2 A(1 + i)3 A(1 + i)N

P.V. = + + +−−−−+
(1 + r) (1 + r)2
(1 + r)3
(1 + r)N
(1 + r) A(1 + i) A(1 + i)2 A(1 + i)N−1
Therefore, P.V. =A+ + + − − − − +
(1 + i) (1 + r) (1 + r)2 (1 + r)N−1
(1 + r) A(1 + i)N
⇒ P.V. [ − 1] = A −
(1 + i) (1 + r)N
A Growing Annuity | 25

A(1 + i) (1 + i)N
⇒ P.V. = [1 − ] (1.20)
(r − i) (1 + r)N

Example 1.22. Alpha Corporation is offering an instrument that promises to pay $ 1,000 × 1.05 after
one year. The payments will increase by 5% every year, and the annuity will in all make 20 payments.
If the annual discount rate is 8%, what is the present value of the annuity?
The present value of the annuity using a discount rate of 8% is

1,000 × 1.05 (1.05)20

[1 − ] = $15,075.89
0.08 − 0.05 (1.08)20

The same can be computed using Excel. The effective discount rate is

(1 + r) (0.08 − 0.05)
−1= = 2.8571%
(1 + i) 1.05
PV(0.028571, 20, −1,000) = $15,075.89.

In the preceding illustration, the growth rate was less than the discount rate. This
need not necessarily be the case, and the former may be higher than the latter. The PV
function can be used in Excel even in this case. However, the effective discount rate
will be negative.

Future Value of a Growing Annuity

The future value of a growing annuity is given by the following equation:

F.V. = A(1 + i)(1 + r)N−1 + A(1 + i)2 (1 + r)N−2 + A(1 + i)3 (1 + r)N−3
+ − − − − + A(1 + i)N
(1 + r)
Therefore, F.V. = A(1 + r)N + A(1 + i)(1 + r)N−1
(1 + i)
+ A(1 + i)2 (1 + r)N−2 + − − − − + A(1 + i)N−1 (1 + r)
(1 + r)
⇒ F.V.[ − 1] = A(1 + r)N − A(1 + i)N
(1 + i)
A(1 + i)
⇒ F.V. = [(1 + r)N − (1 + i)N ] (1.21)
(r − i)

As can be seen, the future value is equal to the present value multiplied by (1 + r)N .

Example 1.23. Beta Corporation is offering an annuity that promises to pay $ 2,500 × 1.05 after one
year. The payments will increase by 5% every year, and the annuity will in all make 20 payments. If the
annual rate of interest is 8%, what is the future value of the annuity?
The future value of the annuity using an interest rate of 8% is
2,500 × 1.05
[(1.08)20 − (1.05)20 ] = $175,670.20
0.08 − 0.05
26 | Chapter 1: A Primer on the Time Value of Money

Growing Perpetuity

We can calculate the present value of a growing perpetuity only if the discount rate is
greater than the growth rate of the cash flows. If it isn’t the series will not converge.
The present value as N → ∞ is given by A(1+i)
r−i
.

Example 1.24. Assume that the annuity in Example 1.23 is a perpetuity. The present value will be:

2,500 × (1.05)
= $87,500
0.03

Growing Annuity Due

Consider an annuity that makes a payment of A immediately. The cash flows grow at a
rate of i% every period. The periodic interest rate is r%. The last cash flow is received
after N − 1 periods. The cash flows can be depicted as shown in Figure 1.4, where today
is depicted as time 0.

A A(1 + i) A(1 + i)2 A(1 + i)3 A(1 + i)N−1

0 1 2 3 N −1 N

Figure 1.4: A growing annuity due.

Present Value of a Growing Annuity Due

The equation for the present value of a growing annuity due is the following.

A(1 + i) A(1 + i)2 A(1 + i)3 A(1 + i)N−1

P.V. = A + + + +−−−−+
(1 + r) (1 + r)2
(1 + r)3
(1 + r)N−1
(1 + r) (1 + r) A(1 + i) A(1 + i)2 A(1 + i)N−2
Therefore, P.V. =A +A+ + +−−−−+
(1 + i) (1 + i) (1 + r) (1 + r)2
(1 + r)N−2
(1 + r) (1 + r) A(1 + i)N−1
⇒ P.V.[ − 1] = A −
(1 + i) (1 + i) (1 + r)N−1
A (1 + i)N
⇒ P.V. = [(1 + r) − ]
(r − i) (1 + r)N−1
A (1 + i)N
= [1 − ] (1 + r) (1.22)
(r − i) (1 + r)N
Thus the present value of the growing annuity due is equal to the present value of
the growing annuity multiplied by (1+r)
(1+i)
. If the discount rate is greater than the growth
Growing Annuity Due | 27

rate, the annuity due will have a higher present value than a corresponding annuity.
However, if the discount rate is less than the growth rate, the annuity due will have
a lower present value. The rationale is as follows: As compared to a growing annuity,
in the case of a growing annuity due, every cash flow grows for one period less, and
is discounted for one period less. Consequently, if the discount rate is greater than
the growth rate, the annuity due will have a higher present value as compared to the
annuity. But if the discount rate were to be lower, then it would have a lower present
value.

Example 1.25. Consider a growing annuity, with a value of $1,000 for A, that makes a total of 20 pay-
ments. The growth rate is 5%, and the discount rate is 8%. The present value, as computed in Exam-
ple 1.22, is $15,075.89. Had it been a 20 period annuity due, the present value would be

1.08
15,075.89 × = $15,506.63
1.05

Now consider another annuity with the same cash flows, but with a discount rate of 5%, and a
growth rate of 8%. The present value of the annuity would be:

1,000 × 1.08 (1.08)20

[1 − ] = $27,239.97
0.05 − 0.08 (1.05)20

Were it an annuity due, the present value would be:

1.05
27,239.97 × = $26,483.30
1.08

The results can be verified using Excel. The effective discount rate in the first case is:

(1 + r) (0.08 − 0.05)
−1= = 2.8571%
(1 + i) 1.05
PV(0.028571, 20, −1,000) = $15,075.89

For the annuity due, we need to make the Type 1:

PV(0.028571, 20, −1,000, , 1) = $15,506.63

The effective discount rate in the second case is:

(0.05 − 0.08)
= −2.7778%
1.08
PV(−0.027778, 20, −1,000) = $27,239.97

Had it been an annuity due,

PV(−0.027778, 20, −1,000, , 1) = $26,483.30.

28 | Chapter 1: A Primer on the Time Value of Money

Future Value of a Growing Annuity Due

The future value of a growing annuity due is given by the following equation.

F.V. = A(1 + r)N + A(1 + i)(1 + r)N−1 + A(1 + i)2 (1 + r)N−2 + − − −−

+ A(1 + i)N−1 (1 + r)
(1 + r) (1 + r)N+1
Therefore, F.V. =A + A(1 + r)N
(1 + i) (1 + i)
+ A(1 + i)(1 + r)N−1 + − − − − + A(1 + i)N−2 (1 + r)2
(1 + r) (1 + r)N+1
⇒ F.V.[ − 1] = A − A(1 + i)N−1 (1 + r)
(1 + i) (1 + i)
A
⇒ F.V. = [(1 + r)N − (1 + i)N ](1 + r) (1.23)
(r − i)

As can be seen, the future value is equal to the present value multiplied by (1 + r)N .

Example 1.26. Beta Corporation is offering an annuity that promises to pay $1, 000 immediately. The
payments will increase by 5% every year, and the annuity will in all make 20 payments. If the annual
rate of interest is 8%, what is the future value of the annuity?
The future value of the annuity using an interest rate of 8% is

1,000
[(1.08)20 − (1.05)20 ] × 1.08 = $72,275.74
0.08 − 0.05

The present value is $15,506.63:

15,506.63 × (1.08)20 = $72,275.74

Growing Perpetuity Due

We can calculate the present value of a growing perpetuity due only if the discount
rate is greater than the growth rate of the cash flows. The present value as N → ∞ is
given by A(1+r)
r−i
.

Example 1.27. Assume that the annuity in Example 1.25 is a perpetuity. The present value is:

1,000 × (1.08)
= $36,000
0.03

The present value of a perpetuity due is equal to the present value of a perpetuity,
which was derived earlier, multiplied by (1+r)
(1+i)
.
Chapter Summary | 29

The present value, had it been a perpetuity, would have been:

1,000 × (1.05)
= $35,000
0.03
1.08
36,000 = 35,000 ×
1.05

Chapter Summary
In this chapter, we studied the fundamentals of the time value of money, such as the
present and future values of a cash flow stream. We introduced the concepts of simple
and compound interest. We also examined the difference between the nominal rate of
interest and the effective rate of interest, and the importance of the latter when interest
is compounded more than once per annum. We studied level annuities and annuities
due, and the corresponding perpetuities. We saw that although the future value of a
perpetuity cannot be computed, it will have a finite present value. We discussed the
internal rate of return, the related issues of pure and mixed cash flows, and Descartes
rule of signs. We studied the structure of amortized loans, which are commonly used
in the context of real estate transactions. Finally, we studied growing annuities and
annuities due, and the corresponding perpetuities. Whenever possible, we used Excel
to illustrate the concepts applicable.
In the next chapter, we study the fundamentals of bonds and bond markets.
Chapter 2
An Introduction to Bonds
Bonds are a source of funds for corporations and governments. For companies, they
represent a supplementary source of funds, in addition to equity shares. For a govern-
ment, however, they represent the sole means of raising funds, by way of issuance of
securities. For neither a federal, state, nor local government, can issue equity shares.
Companies must necessarily issue equity shares. They have the option of issuing
bonds to augment this source of funds if they so desire. Of course a government has
alternative means of fundraising such as levy of taxes, and the printing of currency.
There exists a related type of security termed as a debenture. In some markets the
words bond and debenture are used interchangeably. But in markets like the United
States, a bond is a secured debt security, whereas a debenture represents an unsecured
debt security. The word “secured” means that the issue is backed by specified collat-
eral. So in the event of the borrower being unable to repay, the debt holders can have
the collateral liquidated in order to receive what is due to them. Unsecured bonds are
not backed by specified collateral. Holders can only hope that the issuer will have ad-
equate resources when the time comes for repayment. Government debt, in all coun-
tries, is always referred to as bonds.
Bonds and debentures are referred to as fixed income securities. The reason is that
once the rate of interest is set at the onset of the period for which it is due to be paid, it
is not a function of the profitability of the firm. If a firm issues bonds with an interest
rate of 5% per annum, it cannot subsequently ask holders to accept a lesser amount
on account of poor performance. The flip side is that in the event of extra-ordinary
performance, the bond holders cannot demand more than 5%. This is unlike the case
of equity shares, where the percentage of dividends declared by the company can fluc-
tuate from year to year. Interest payments are therefore contractual obligations, and
failure to pay what was promised at the start of an interest computation period is tan-
tamount to default. In contrast a firm is under no obligation to declare a dividend in
a given year. The decision to pay a dividend is taken by the board of directors, which
also decides the magnitude of the dividend. Dissident shareholders, if they are in a
majority, may be able to force a company to overturn its dividend decision.
Why do firms issue bonds? It is because bonds allow the firm to raise additional
capital without diluting the stake of the shareholders. Unlike a shareholder, a bond
holder is a stake holder in a business but is not a part owner of the business. A bond
holder is entitled only to the interest that was promised and to the repayment of prin-
cipal at the time of maturity of the security, and does not partake in the profits of the
firm. Bonds have two other important properties, namely they give the firm a tax shield
and provide leverage.

https://doi.org/10.1515/9781547400669-002
The Leverage Effect | 31

The Leverage Effect

Bonds provide equity shareholders with leverage. Table 2.1 provides a detailed il-
lustration. Company XEON is entirely equity financed and has issued shares worth
$5,000,000. Company BORA has raised the same amount of capital, half in the form of
debt and the other half in the form of equity. The debt carries interest at the rate of 8%
per annum. Let’s consider two situations: in the first, both companies make a profit
of $500,000 from operations, and in the second, they both take a loss of $500,000. To
keep matters simple, we will assume that the firms do not have to pay tax.

Table 2.1: Illustration of leverage.

Case-A Case-B
Firms Make a Profit Firms Take a Loss
XEON BORA XEON BORA

Profits from Operations 500,000 500,000 (500,000) (500,000)

Less Interest 0 (200,000) 0 (200,000)
PAT 500,000 300,000 (500,000) (700,000)
ROI 10% 12% −10% −28%
PAT stands for profit after tax and represents what the shareholders are entitled to.
ROI is the return on investment for shareholders.

As we can see from Table 2.1, the presence of debt in the capital structure creates lever-
age. The profit for the shareholders is magnified from 10% to 12%, when a firm is fi-
nanced 50% with debt. On the other hand, a loss is also magnified, in this case, from
−10% to −28%. As they say, leverage is a double edged sword. It magnifies returns,
irrespective of whether the returns are positive or negative.1 We can also see, from the
case of BORA, that incurring a loss does not give the flexibility to the firm to avoid or
postpone the interest due to bond holders. Interest on bonds is indeed a contractual
obligation.

Tax Shield Due to Interest Payments

Unlike dividends, which are paid out of post-tax profits, the interest paid on bonds
and debentures is deductible as an expense when computing the net income. Thus the
effective interest cost for the issuer is lower than the stated rate. We say that interest
provides a tax shield. By allowing the firm to deduct the interest as an expense, the

1 In finance, we are concerned with profits and losses in relation to the investment made, and not in
absolute terms. Thus a loss of $25,000 on an investment of $50,000 is considered more serious than
an identical loss on an investment of $1,000,000.
32 | Chapter 2: An Introduction to Bonds

government collects a lower amount by way of taxes. In a sense, therefore, it is offering

a subsidy to the issuing entity.
Consider the illustration in Table 2.2. Both firms are assumed to have a profit of
$250,000, and the applicable tax rate is assumed to be 30%.

Table 2.2: Illustration of a tax shield.

Company XEON BORA

Profit from Operations 250,000 250,000

Interest NIL 30,000
PBT 250,000 220,000
Tax (75,000) (66,000)
PAT 175,000 154,000
PBT stands for profits before tax
PAT stands for profits after tax

Let’s analyze the last row of Table 2.2. If company BORA had been a zero debt com-
pany like XEON, its shareholders would have been entitled to a cash flow of $175,000.
However, because it has paid interest, the shareholders are entitled to only $154,000,
which is $21,000 less. Thus, from the standpoint of the shareholders of BORA, they
have effectively paid interest of $21,000. So what explains the missing $9,000, for af-
ter all, we know that BORA did pay $30,000 to its bond holders? The answer is that by
allowing the firm to deduct the interest paid as an expense, prior to the computation
of tax, the government has foregone taxes to the extent of $9,000. The tax shield, as
we term it, is equal to the product of the tax rate and the interest paid. In this case it
is 0.30 × 30,000 = $9,000.
If the rules were to be amended and interest on debt were no longer to be tax
deductible, BORA would have to pay tax on $250,000, and the profit after tax, which
is what belongs to the shareholders, would be only $145,000. In this situation, the
shareholders will feel the full burden of the interest paid by the firm.
Debt securities may be negotiable or non-negotiable. A negotiable security is one
that can be traded in the secondary market, whereas a non-negotiable security cannot
be signed over by the holder in favor of another investor. Bank accounts are classic
examples of non-negotiable investments, for if an investor were to open a term deposit
with a commercial bank, although he can always withdraw the investment and pay
a third party, he cannot transfer the ownership of the deposit. Most corporate and
government bonds can be traded prior to maturity, either directly on an organized
exchange, or via a dealer in an over-the-counter (OTC) trade.
The most basic form of a debt security is referred to as a plain vanilla bond. Bonds
are referred to as IOUs, an acronym which stands for “I owe you.” A plain vanilla bond
is an IOU that promises to pay a fixed rate of interest every period, which is usually
Variables Influencing the Bond Price | 33

every six months, and to repay the principal at maturity. Any bond with additional
features, is said to have bells and whistles attached. For instance, in the case of float-
ing rate bonds, the interest rate does not remain constant from period to period, but
fluctuates with changes in the benchmark to which it is linked.
There are also bonds with embedded options. For example, Convertible bonds can
be converted to shares of stock by the investor. Callable bonds can be prematurely
retired by the issuing company, and putable bonds can be prematurely surrendered
by the bond holders in return for the repayment of the principal.

Variables Influencing the Bond Price

There are four variables that influence the price of a plain vanilla bond. These are:
– Face value
– Term to maturity
– Coupon
– Yield to maturity

Face Value

The face value—also known as par value, redemption value, maturity value, or princi-
pal value—is the principal amount underlying the bond. It is the amount raised by the
issuer from the first holder2 and the amount repayable by the issuer to the last holder.
We denote it by the symbol M.

Term to Maturity

Term to maturity is the time remaining in the life of the bond as measured at the point
of valuation. You can understand it as the length of time after which the debt ceases
to exist and the borrower redeems the issue by repaying the holder. Equivalently, you
can understand it as the length of time for which the borrower has to service the debt
in the form of periodic interest payments. The words maturity, term, tenor, tenure, and
term to maturity are used interchangeably. We will assume that we are at time zero
and denote the point of maturity by T. Thus the number of periods until maturity is T,
which is normally measured in years.
In the debt market, we make a distinction between the original term to maturity
of a security, and its actual term to maturity. The original term to maturity is the term

2 In most cases, bonds are issued at the face value, although there are instances where they are issued
at higher or lower values, as you will see later.
34 | Chapter 2: An Introduction to Bonds

to expiration as measured at the time of issue. However, the actual term to maturity is
the term to maturity as computed at the current point in time. As time elapses, the ac-
tual term to maturity (ATM) continues declining, whereas the original term to maturity
(OTM) remains constant.

Coupon

The contractual interest payment made by the issuer is called a coupon payment. The
name came about because in earlier days bonds were issued with a booklet of post-
dated coupons. On an interest payment date, the holder was expected to detach the
relevant coupon and claim his payment. Even today, bearer bonds come with a booklet
of coupons. These are bonds where no record of ownership is maintained, unlike in
the case of conventional or registered bonds. Thus the holder of a bearer bond needs
to produce the coupon to claim the interest that is due.
The coupon may be denoted as a rate or as a dollar value. We will denote the
coupon rate by c. The dollar value, C, is therefore given by c × M. Most bonds pay in-
terest on a semiannual basis, and consequently the semiannual cash flow is c × M/2.
Semiannual coupons are the norm in the UK, US, Japan, and Australia. However, in
certain regions like the European Union, bonds typically pay annual coupons.
Consider a bond with a face value of $1,000 that pays a coupon of 8% per an-
num on a semiannual basis. The annual coupon rate is 0.08. The semiannual coupon
payment is 0.08 × 1,000/2 = $40.

Yield to Maturity

The yield to maturity, like the coupon rate, is also an interest rate. The difference is that
whereas the coupon rate is the rate of interest paid by the issuer, the yield to maturity,
commonly referred to as the YTM, is the rate of return required by the market. At a
given point in time, the yield may be greater than, equal to, or less than the coupon
rate. The YTM is denoted by y and is the rate of return that a buyer gets when acquiring
the bond at the prevailing price and holding it to maturity.

Valuation of a Bond
To value a bond, we first assume that we are standing on a coupon payment date. That
is, we assume that a coupon has just been received, and consequently the next coupon
is exactly six months or one period away. If T is the term to maturity in years, we have
N coupons remaining where N = 2T. We receive N coupon payments during the life
Par, Premium, and Discount Bonds | 35

of the bond, where each payment or cash flow is equal to C/2. This payment stream
obviously constitutes an annuity. The present value of this annuity is
c×M
2 1
y [1 − N
]
2 (1 + y2 )

The terminal payment of the face value is a lump sum payment. Because we are
discounting the cash flows from the annuity on a semiannual basis, this payment
needs to be discounted on a similar basis. Thus the present value of this cash flow is:

M
N
(1 + y2 )

Hence the price of the bond is given by:

c×M
2 1 M
y [1 − ]+
y N N
2 (1 + 2
) (1 + y2 )

As you can be see, the YTM is that discount rate that makes the present value of
the cash flows from the bond equal to its price. The price of a bond is like the initial
investment in a project. The remaining cash flows are similar to the inflows from the
project. Consequently the YTM is exactly analogous to the IRR, or internal rate of re-
turn, a concept used in capital budgeting.

Example 2.1. Omega has issued bonds with 10 years to maturity and a face value of $1,000. The
coupon rate is 8% per annum payable on a semiannual basis. If the required yield in the market is
equal to 10%, what should be the price of the bond? The periodic cash flow is 0.08 × 1,000/2 = $40.
Thus the present value of all the coupons is:

0.08×1,000
2 1
0.10
[1 − ] = $498.4884
2
(1.05)20

The present value of the face value is

1,000
= $376.8895
(1.05)20

Thus the price of the bond is 498.4884 + 376.8895 = $875.3779

Par, Premium, and Discount Bonds

In Example 2.1, the price of the bond is less than its face value. Such a bond is said
to be trading at a discount to the par value and is therefore referred to as a discount
bond. If the price of the bond is greater than its face value, then it is said to be trading
36 | Chapter 2: An Introduction to Bonds

at a premium to its face value and consequently is referred to as a premium bond. If

the price is equal to the face value, then the bond is said to be trading at par. As you
see from the pricing equation, price and yield are inversely related. If the rate of return
offered by the issuer is equal to that demanded by the market, then the bond obviously
trades at par. However, if the market demands a higher rate of return, it then brings
the price down until the desired yield is obtained. On the other hand, if the desired
yield is lower than the coupon, then traders will be willing to pay a price that is higher
than the face value of the bond.

Influence of Variables on the Bond Price

Let’s first differentiate the price with respect to the face value, keeping other variables
constant:
c
𝜕P 2 1 1
= y [1 − ]+ >0 (2.1)
y N N
𝜕M 2 (1 + 2
) (1 + y2 )

Thus the partial derivative of the price with respect to the face value is positive. That
is, keeping other variables constant, the higher the face value, the higher the bond
price.
Now consider the derivative with respect to the coupon rate:
M
𝜕P 2 1
= y [1 − N
]>0 (2.2)
𝜕c 2 (1 + y2 )

Thus the partial derivative with respect to the coupon rate is positive. That is, keeping
other variables constant, the higher the coupon rate is, the higher the bond price.
Let’s now consider the derivative of the price with respect to the time to maturity.
Unlike other variables, the number of coupons can only increase in increments of 1.
When there are N coupons remaining until maturity, the price is given by:

Mc 1 M
P0 = × [1 − ]+ (2.3)
y y
(1 + 2 )
N
(1 + y2 )
N

When there are N + 1 coupons remaining until maturity, the price is given by:

Mc 1 M
P1 = × [1 − ]+ (2.4)
y y (N+1)
(1 + 2 ) (1 + y2 )
(N+1)

M c y
⇒ P1 − P0 = △P = ×[ − ] (2.5)
y N+1
(1 + 2 ) 2 2

For a par bond, the price change is zero. Thus as we go from one coupon date to
the next, the price of a par bond continues to be equal to its face value. For a premium
The Pull to Par Effect | 37

bond the price change is positive. Thus, holding other parameters constant, as the
time to maturity of a premium bond increases, its price also increases. Finally, for a
discount bond, the price change is negative. Thus if the time to maturity increases, a
discount bond declines in price.
Finally, let’s consider the derivative of the price with respect to the yield to matu-
rity:

𝜕P −cM 1 + { y2 × (N + 1)} NM 1
= 2 [1 − ]− × (2.6)
𝜕y y y
(1 + 2 )
N+1 2 (1 + y2 )
N+1

y
Let’s define 2
as r . From the expression for a Maclaurin series we know that:3

N(N + 1)r 2
(1 + r)N+1 = 1 + r(N + 1) + + h. o. t. (2.7)
2

Thus

1 + r(N + 1) < (1 + r)N+1

y y N+1
⇒ 1 + { × (N + 1)} < (1 + )
2 2
1 + { y2 × (N + 1)}
⇒ N+1
<1
(1 + y2 )
𝜕P
Thus <0
𝜕y

Thus, as the YTM increases, the price of the bond declines.

The Pull to Par Effect

The price of a premium bond is greater than the face value, whereas that of a discount
bond is less than the face value. At maturity all bonds trade at par. As we have demon-
strated, when the number of coupons is increased, the price change, is zero for a par
bond. Thus a par bond continues to trade at par on every successive coupon date if
the YTM remains constant. However, for premium bonds, the price increases with an
increasing time to maturity, keeping the YTM constant. Thus as we approach maturity,
the price of a premium bond steadily declines and approaches par. Finally, for a dis-
count bond, the price change is negative if the time until maturity increases. Hence as
we approach maturity, the price of a discount bond steadily increases and approaches
par.

3 h.o.t stands for higher order terms.

38 | Chapter 2: An Introduction to Bonds

This phenomenon is known as the pull to par effect. The rationale is as follows:
When we move from one coupon date to the next coupon date, we have one coupon
less. Thus the contribution of coupons to the price goes down, leading to a decline in
the price. However, the present value of the face value increases because the face value
is discounted for one period less. This therefore leads to an increase in the bond price.
For premium bonds, the first effect dominates, and the price steadily declines as we go
from one coupon date to the next. In the case of discount bonds, the second effect is
more significant, and the bond price steadily increases. For par bonds, the two effects
neutralize each other, and there is no change in the bond price, which continues to
remain at par.

Example 2.2. Consider two bonds with three years to maturity and a face value of $1,000. Assume that
the coupon for the first bond is 8% per annum and the YTM is 10% per annum. The bond will obviously
sell at a discount. In the case of the second bond, let’s assume that the coupon is 10% per annum and
the YTM is 8% per annum. This bond will obviously trade at a premium. Let’s trace the evolution of the
price for these bonds as we approach maturity. The results are summarized in Table 2.3.

Table 2.3: The pull to par effect.

Time to Price of Price of

Maturity Discount Premium
in Periods Bond Bond
6 949.24 1,052.42
5 956.71 1,044.52
4 964.54 1,036.30
3 972.77 1,027.75
2 981.41 1,018.86
1 990.48 1,009.62
0 1,000.00 1,000.00

An Interesting Result about Bond Prices

Let’s consider a situation where the coupon and the YTM are both changed by the
same percentage. Thus the new coupon is equal to ac, where c is the original coupon
rate, and a is the scaling factor. Similarly, the new YTM is ay, where a may be greater
than one or less than one. The original price is given by
c×M
2 1 M
P= y [1 − ]+
y N N
2 (1 + 2
) (1 + y2 )
The new price P ∗ is given by
ac×M
2 1 M
ay [1 − ]+
ay N ay N
2 (1 + 2
) (1 + 2
)
An Interesting Result about Bond Prices | 39

c×M
2 1 M
= y [1 − ]+
ay N ay N
2 (1 + 2
) (1 + 2
)
1 1 c
⇒ P∗ − P = M × [ − ] × [1 − ] (2.8)
(1 + ay N
) (1 + y N
) y
2 2
1 1
If a > 1, <
ay N N
(1 + 2
) (1 + y2 )
1 1
⇒M×[ − ]<0
ay N N
(1 + 2
) (1 + y2 )
c
If c > y, 1 − <0
y

Thus for a premium bond, △P > 0. Hence if we increase the coupon and the YTM of a
bond by the same percentage, a premium bond will increase in price.
For a discount bond:
c
1− >0 and hence △P <0
y

Thus, for a discount bond, if we increase the yield and the coupon by the same per-
centage, the price will decline.
Now let’s consider the case where a < 1. That is, the yield and coupon are de-
creased by the same percentage.
1 1
>
ay N N
(1 + 2
) (1 + y2 )
1 1
⇒M×[ − ]>0
ay N N
(1 + 2
) (1 + y2 )

In this case, for a premium bond, △P < 0. Hence if we decrease the coupon and the
YTM of a bond by the same percentage, a premium bond will decline in price. For a
discount bond, △P > 0. Thus, for a discount bond, if we decrease the yield and the
coupon by the same percentage, the price will increase.

Example 2.3. Consider two bonds, both with a face value of $1,000 and five years to maturity. Bond
A pays a coupon of 8% per annum on a semiannual basis, whereas Bond B pays a coupon of 10% per
annum, also on a semiannual basis. Bond A has a YTM of 10% per annum, and Bond B has a YTM of
8% per annum. Thus Bond A is a discount bond, and, Bond B is a premium bond.
The initial price of Bond A is $922.7827, whereas the initial price of Bond B is $1,081.1090. Let’s
increase both the coupons and yields by 25%. For Bond A, the new coupon is 10% per annum, and the
YTM is 12.50% per annum. For Bond B, the new coupon is 12.50% per annum, and the new YTM is 10%
per annum. The new price of Bond A is $909.0789, and the price of Bond B is $1,096.5217. Thus the
premium bond has increased in price, whereas the discount bond has declined in price.
Now let’s instead reduce both the coupon and the yield of both the bonds by 25%. For Bond A the
new coupon is 6% per annum, and the new YTM is 7.50% per annum. For Bond B the new coupon is
40 | Chapter 2: An Introduction to Bonds

7.50% per annum, and the YTM is 6% per annum. The new price of Bond A is $938.4041, and that of
Bond B is $1,063.9765. In this case, the discount bond has increased in price, whereas the premium
bond has decreased in price.

Eurobonds and Foreign Bonds

There are three categories of bonds depending on the currency of issue, and the na-
tionality of the issuer. If the currency in which a bond is denominated is different from
the currency of the country of issue we term it a Eurobond. This is not related to the
location of issue. That is, Eurobonds can be issued outside Europe. Thus bonds de-
nominated in U.S. dollars that are sold in Europe are classified as Eurobonds. The na-
tionality of the issuing entity is irrelevant. Thus if Microsoft were to issue U.S. dollar-
denominated bonds in Europe, we would term them Eurobonds. The same would be
the case if Mercedes were to issue dollar-denominated bonds in Europe.
If a bond is denominated in the currency of the country in which the issue is taking
place, but the issuer is from a different country, then we classify the debt security as a
foreign bond. Thus if Microsoft were to issue bonds denominated in euros in Frankfurt,
then it would constitute a foreign bond issue. Foreign bonds are known by nicknames.
Bonds sold in the U.S. are called Yankee bonds; those sold in Japan are called Samurai
bonds; and those sold in the U.K. are called Bulldog bonds.
Finally if the issuing entity belongs to the country of issue and the currency is also
local, then the bond is categorized as a domestic bond. Thus if Mercedes were to issue
bonds denominated in euros in Germany, then we would term it a domestic bond.
Thus if the issuer and the currency are both local, then it is a domestic bond. If
the currency is local but the issuer is not, then it is a foreign bond. And finally if the
currency is not local, it is a Eurobond, irrespective of the nationality of the issuer.
The Eurobond market has grown much more rapidly than the foreign bond mar-
ket. This is due to several reasons. Eurobond issues are not subject to the regulations
of the country in whose currency they are denominated. Consequently, they can be
brought to the market quickly and with less disclosure. This gives the issuer of the
bonds greater flexibility to take advantage of favorable market conditions. On the other
hand, domestic securities issues are regulated by the market regulator of the country,
like the SEC in the U.S.
The origin of the Eurobond market was fueled by the imposition of a cess4 called
the Interest Equalization Tax, which was imposed by the U.S. government in 1963, on
the interest received by American investors from Yankee bonds. What had happened
was that, due to the low interest rate ceiling in the U.S., domestic institutions were
unable to pay a high rate of interest to investors. Foreigners sought to take advan-
tage of this situation by issuing Yankee bonds with relatively higher rates of return.

4 It is a form of tax.
Coupon Dates and Coupon Frequencies | 41

The objective of the Interest Equalization Tax was to ensure that American investors
did not perceive Yankee bonds to be attractive, despite their higher interest rates. The
motivation for this measure was to arrest the perceived flight of capital from the U.S.
As a consequence, foreigners were forced to relocate their dollar borrowings outside
the U.S.
Eurobonds offer favorable tax status, for they are usually issued in bearer form.
That is, the name and address of the owner are not mentioned on the bond certifi-
cate. Thus, in the case of bearer securities, physical possession is the sole evidence
of ownership. Such securities are easier to transfer and offer investors the potential
freedom to avoid and evade taxes. Thus, holders who desire anonymity can receive
interest payments from such securities without revealing their identities. Also, inter-
est on Eurobonds is generally not subject to withholding taxes, or tax deduction at
source. Because of their unique features, investors are willing to accept a lower yield
from Eurobonds than from other securities of comparable risk that lack the favorable
tax status.
Eurobonds are not usually registered with any particular regulatory agency, but
may be listed on a stock exchange, typically London or Luxembourg. Listing is done
not just for the purpose of facilitating trading, but to circumvent restrictions imposed
on certain institutional investors like pension funds that are prohibited from purchas-
ing unlisted securities. Most of the trading in Eurobonds takes place over the counter
or OTC, which refers to a market that is a network of broker-dealers.

Coupon Dates and Coupon Frequencies

Bonds in the U.S., U.K., Japan, and Australia typically pay coupons on a semiannual
basis. However, Eurobonds, and bonds issued in the EU generally, pay coupons on an
annual basis. In the case of bonds paying semiannual coupons, there are six monthly
combinations possible—January/July, February/August, March/September, April/Oc-
tober, May/November, and June/December. In each case, the coupon date is usually
the 1st or the 15th of the month. Thus the bond dealers’ quote sheets usually state the
following codes:
– JJ-01 or JJ-15
– FA-01 or FA-15
– MS-01 or MS-15
– AO-01 or AO-15
– MN-01 or MN-15
– JD-01 or JD-15

It should also be noted that in practice we are highly unlikely to find a bond with the
last day of February as the coupon date because this creates practical complications,
due to the fact that February has an extra day once in four years.
42 | Chapter 2: An Introduction to Bonds

Zero Coupon Bonds

Unlike a plain vanilla bond that pays coupons at periodic intervals, zero coupon
bonds, also known as deep discount bonds do not pay any interest. Such instruments
are always traded at a discount to the face value, and at maturity, the holder receives
the face value. For instance, consider a zero coupon bond with a face value of $1,000
and 10 years to maturity. Assume that the required yield is 8% per annum. The price
may be computed as follows:

1,000
P= = $456.39
(1.04)20

Notice that we have chosen to discount at a rate of 4% for 20 half-year periods,

and not at 8% for 10 annual periods. The reason is that in practice a potential investor
has a choice between plain vanilla bonds and zero coupon bonds. Comparisons be-
tween bonds, are made on the basis of yields, and to draw meaningful inferences it is
imperative that the discounting technique be common across bond types. Because the
cash flows from plain vanilla bonds are usually discounted on a semiannual basis, we
choose to do the same for zero coupon bonds.
A zero coupon bond never sells at a premium. It always trades at a discount except
at the time of maturity when it trades at the face value. That does not mean that a buyer
of such a bond always experiences a capital gain. A buyer who purchases, and holds
it to maturity obviously will have a capital gain. But if a buyer chooses to sell it prior
to maturity, the result may well end up being a capital loss as we shall demonstrate.

Example 2.4. Alex bought a zero coupon bond when there were 10 years to maturity. The prevailing
yield was 8% per annum. Today, a year later, the YTM is 10% per annum. The purchase price was
$456.39. The price at the time of sale is $415.52. Thus in this case the investor has a capital loss of
$40.87.

Creating a Synthetic Zero Coupon Bond

A zero coupon bond can be synthesized using two plain vanilla bonds each of which
pays a coupon at a different rate. Consider two bonds, both with a face value of M and
with N coupons remaining till maturity. The first bond has a coupon of c1 , and the
second has a coupon rate of c2 . Assume we buy w units of the first bond and (1 − w)
units of the second.
The payment every semiannual period will be

c1 c
w×M× + (1 − w) × M × 2
2 2
Zero Coupon Bonds | 43

Let’s choose w, such that the cash flow from the portfolio is zero

c1 c2 c
w(− )=− 2
2 2 2
c2
⇒w=−
c1 − c2

The cash flow at maturity will be

c1 c
w × M × (1 + ) + (1 − w) × M × (1 + 2 ) = M
2 2

Hence we have created a zero coupon bond with a face value of M, and N semiannual
periods to maturity. Here is a numerical example.

Example 2.5. Consider a bond with a coupon of 8% per annum, paid semiannually, and another with
a coupon of 10% per annum, payable half-yearly. Both the bonds have N half-year periods until expi-
ration and a face value of $1,000.

0.10 0.10
w=− =− = 5. Thus (1 − w) = −4
0.08 − 0.10 −0.02

Hence we need to go long in five units of the first bond and short in four units of the second. The
periodic cash flow is

5 × 40 − 4 × 50 = 0

The terminal cash flow is:

5 × 1,040 − 4 × 1,050 = $1,000

The No-arbitrage Price of the Zero Coupon Bond

As we have seen a zero coupon bond can be synthesized using two coupon paying
bonds. To rule out arbitrage, the price of the zero should be a weighted average of the
prices of the coupon paying bonds. Consider the data in Example 2.5. Assume the two
bonds have a time to maturity of five years, and that the bond with an 8% coupon has
a YTM of 10% per annum, while the bond with a coupon of 10% has a YTM of 11% per
annum. Thus the prices of the two bonds will be $922.7827 and $962.3119. To create
the synthetic zero we need to go long in five units of the former, and short in four units
of the latter. Thus the price of the synthetic zero should be 5 × 922.7827 − 4 × 962.3119 =
$764.6659. Thus

1,000
10
= 764.6659
(1 + y2 )
⇒ y = 5.4390%
44 | Chapter 2: An Introduction to Bonds

Price Quotes for Bonds

Bond prices are always quoted as a percentage of par, or per 100 currency units of
face value. The reason is the following. Consider two bonds, one with a face value of
$1,000 and another with a par value of $2,000. Assume that both securities are quoting
at $1,500. The first is obviously a premium bond and the second a discount bond.
However this cannot be discerned from the observed prices unless the corresponding
face values are also specified. On the other hand if the price were to be expressed for a
par value of $100, any value less than 100 would signify a discount bond while a value
more than 100 would connote a premium bond. In the US market prices are quoted as
a percentage of par plus 32nds. In other words the bond market is not decimalized and
prices vary in increments of 321 . For instance consider a quote of 97-08. This does not
mean a price of 97 dollars and 8 cents per dollar 100 of face value. What it means is
that the price per dollar 100 of face value is 97 + 8/32 or $97.25. So if we were to buy
bonds with a face value of $1MM, the amount payable will be 1,000,000 × 97.25/100 =
1
$972,500. In some cases the price will have a trailing + sign. This indicates that a 64
needs to be added. For instance a quote of 97-08+ would connote a price of 97 + 8/32 +
1/64 or 97.265625 per 100 dollars of face value.

Computation of the Bond Price Using Excel

If we are on a coupon date, the PV function in Excel may be used to determine the
price of a bond. The required parameters for the function are:
– Rate: This is the periodic yield to maturity. Thus if a bond pays coupons semian-
nually, the rate is half the applicable YTM.
– Nper: This is the is the number of coupon periods remaining till maturity. Thus,
for a bond paying semiannual coupons, it will be twice the number of years left.
– Pmt: This is the periodic coupon payment. Thus if a bond with a face value of
$1,000 were to pay a coupon of 8% per annum on a semiannual basis, Pmt will
be $40.
– Fv: This is the face value of the bond, which is typically $1,000.

In the PV Excel function, cash flows in one direction are positive while those in the
other direction are negative. Thus if we want the price, which represents our invest-
ment, to be positive, then the cash flows received from the bond must be negative.
Thus Pmt in the above case will be stated as −40, and Fv as −1,000.

Example 2.6. Consider a bond with a face value of $1,000 and five years to maturity. The coupon
is 7.5% per annum and the YTM is 10% per annum. Thus the inputs to the PV function will be
(0.05, 10, −37.50, −1,000). The answer is $903.48.
Different Bond Types | 45

If we know the price, we can compute the YTM using the RATE function in Excel. The
required parameters are Nper, Pmt, Pv, and Fv. For obvious reasons, if Pmt and Fv are
entered with positive signs, then Pv must be entered with a negative sign. The answer
will be in terms of the rate per period. Thus if the bond pays semiannual coupons, the
rate that is obtained must be multiplied by two to annualize it. Here is an illustration.

Example 2.7. Consider a bond with a face value of $1,000 and five years to maturity. The coupon
is 7.5% per annum and the price is $925. What is the YTM? The inputs to the RATE function are
(10, −37.50, 925, −1,000). The answer is 4.7075%. This is a semiannual rate. Thus the annual YTM is
9.4151%.

Different Bond Types

We have already discussed some alternatives to plain vanilla bonds, what we have
termed as bonds with bells and whistles. These include zero coupon bonds and bonds
with built-in options such as callable, putable, and convertible bonds. We will discuss
these bonds in more detail subsequently. For the moment, let’s analyze certain other
types of bonds with bells and whistles.

Amortizing Bonds

A plain vanilla bond is also known as a bullet bond, because the entire principal is
repaid in one installment at the time of maturity. In the case of an amortizing bond,
however, the principal is repaid in installments. Let’s consider a three-year amortizing
bond with a face value of $1,000 and coupon of 8% per annum payable semiannually.
The face value is repaid in four equal installments starting with the time of payment
of the third coupon. Thus the cash flows from the bond are as depicted in Table 2.4.
The first two cash flows represent coupon payments of $40 each. The third con-
sists of a principal payment of $250 and a coupon payment that is 4% of 1,000. The
fourth cash flow consists of a principal payment of $250 and a coupon of 4% on a

Table 2.4: Cash flows from an amortizing bond.

Time Period Cash Flow

1 40
2 40
3 290
4 280
5 270
6 260
46 | Chapter 2: An Introduction to Bonds

principal of $750. The penultimate cash flow consists of a principal payment of $250
and interest at a rate of 4% on a principal of $500. The final cash flow consists of a
principal of $250 and interest at the rate of 4% on a principal of the same amount.
Some issuers issue such bonds because their asset base also has a similar cash
flow profile. These bonds generally can be issued with a lower coupon as compared
to plain vanilla bonds that are similar in all other respects. This is because in the case
of the latter, the entire principal repayment is concentrated at the end. Consequently
there is a greater risk of default for the holder, as opposed to an investor in an amor-
tized bond, who recovers a percentage of the face value prior to maturity. Table 2.5
compares the cash flows from a plain vanilla bond, with those from an amortizing
bond.

Comparison of a Plain Vanilla Bond with an Amortizing Bond

Table 2.5: YTM of a plain vanilla bond versus that of an amortizing bond.

Time Cash Flow Cash Flow Cash Flow Cash Flow Cash Flow Cash Flow

0 −1,000 −1,000 −925 −925 −1,025 −1,025

1 40 40 40 40 40 40
2 40 40 40 40 40 40
3 40 290 40 290 40 290
4 40 280 40 280 40 280
5 40 270 40 270 40 270
6 1,040 260 1,040 260 1,040 260
YTM 8.0000% 8.0000% 11.0028% 11.9620% 7.0608% 6.7780%

As can be seen in the table, if a plain vanilla bond and an amortizing bond, with the
same coupon rate, time to maturity, and face value, are trading at par, then both of
them will have the same YTM, which is obviously equal to the coupon rate. However,
if the two are trading at a discount, with the same price, the amortizing bond will have
a higher YTM. On the other hand, if the two are trading at a premium, once again with
the same price, the amortizing bond will have a lower YTM. We will see a similar result
when we analyze mortgage-backed securities later in the book.
The rationale is as follows. Compared to an amortizing bond, a plain vanilla bond
has a concentrated cash flow at the end. The impact of a given yield change, is more
on long-term cash flows than on shorter-term cash flows. If the yield is more than the
coupon, a plain vanilla bond will trade at a discount. If an amortizing bond were to
have the same yield, it would trade at a higher price. Consequently, if they are trading
at the same price, the amortized bond must have a higher YTM. On the other hand,
if the yield is less than the coupon, the plain vanilla bond will trade at a premium.
If an amortizing bond were to have the same yield, it would trade at a lower price.
Different Bond Types | 47

Consequently, if they are trading at the same price, the amortized bond must have a
lower YTM.

Bonds with Step-Up Coupons and Step-Down Coupons

Take the case of a newly established company, or a company with a relatively virgin
product or service. The market may expect such entities to offer a higher coupon. How-
ever, such firms are unlikely to have a high level of earnings, because revenues are
likely to peak only after the product or service has acquired sufficient market share.
As these companies cannot afford to pay a high coupon at the outset, they may go in
for a step-up coupon bond, where the coupon increases as the bond becomes more
seasoned. For instance, take a company that is in a position to issue a five-year bond
with a coupon of 8% per annum. It may instead decide to issue a bond with a coupon
of 6% per annum for the first three years, and 10% per annum for the last two years.
Such a bond gives the issuer elbow room in the earlier years.
In practice these bonds are usually callable on each coupon date. These bonds
may be of the one-step type or the multi-step variety. In the case of the former, the
coupon resets once during the life of the bond, whereas in the case of the latter it
resets more than once. As the coupon rate on such bonds rises over comparable rates,
the bond is likely to be called. Thus in practice the investors may not receive the higher
payments due after the coupon is scheduled to reset. As compensation for the right to
redeem early, the issuers are required to offer a higher coupon. It may also be the case
that the scheduled coupon increases do not keep pace with the market rates. Thus,
despite the step-up feature, the enhancement in coupon may not be adequate.
In the case of step-down coupon bonds, the security initially offers a high coupon,
and the coupon rate reduces with the passage of time. These may also be callable. A
company which is required to pay a high coupon because of the perceived risk level,
but is confident that its image will improve over a period of time, may issue such bonds
with a higher coupon in the initial years, and lower coupons in subsequent years.

Payment-In-Kind (PIK) Bonds

Payment-in-kind bonds do not pay coupons in the form of cash. Such bonds require
the issuer to offer additional securities to the investors, without any monetary consid-
eration, when the coupon falls due. Again, like in the case of step-up coupons, this
offers the issuer time to prepare for enhanced cash outlays. The security used to pay
interest is usually identical to the underlying bond, but in principle may be different.
Compared to investors in plain vanilla bonds, investors in PIK bonds take more risk
but are likely to get higher returns.
48 | Chapter 2: An Introduction to Bonds

Treasury Securities
Government securities in the U.S. are termed Treasury securities. The Department of
the Treasury issues three categories of marketable debt securities: T-bills, T-notes, and
T-bonds. T-bills are zero coupon securities with a maximum time to maturity of 52
weeks. T-notes have maturities ranging from 1–10 years, and T-bonds have maturities
in excess of 10 years, going up to 30 years in practice. Both T-notes and T-bonds pay
coupons. For a given maturity, a Treasury security is considered to be the safest type
of debt security from the perspective of default risk. This is because the federal gov-
ernment of a country has two powers that other institutions do not. It can print more
currency whenever it wants, and it also can increase the rates of taxes as and when it
deems fit. Consequently, for a given maturity, the YTM of a Treasury security is usu-
ally lower than the yields of all other securities with the same maturity. In practice
there are other factors at work. For instance, Municipal bonds provide holders with
exemption from paying Federal income tax, as a consequence of which, holders may
be prepared to accept a lower yield.
Whereas corporations have a choice between equity and debt, for raising funds,
governments as discussed earlier, can issue only debt securities. There are many rea-
sons why a government might choose to issue debt. It may have a deficit in the current
financial year because expenses are in excess of projected revenues. Or it may have
issued debt in the past, and needs funds to pay interest on the same. Finally, there
could be a situation where debt issued earlier is maturing, and funds are required to
make the redemption payment. The U.S. Treasuries market is the largest bond market
in the world, as well as the most liquid market.
Sometimes the Treasury may follow up an issue with a subsequent issue with the
same remaining term to maturity and coupon rate. The issuance of such a tranche is
termed as a reopening of an existing issue. For instance, assume that six months ago,
the Treasury issued a 10-year note with a coupon of 4% per annum. If it were to now
issue additional 9 1/2-year notes with a coupon of 4%, we would term it as a reopening,
for we are deepening the pool of securities that are already trading in the market.

Treasury Auctions
The interest rate of a Treasury security is determined by the market, and is not set by
the government. The Treasury sells bills, notes, and bonds, by way of an auction pro-
cess. There are two types of bids–competitive and non-competitive. The auctions may
be price based or yield based. Auctions for new securities are yield based. Competi-
tive bidders indicate the minimum yield they require as well as the quantity sought.
Re-openings of existing securities are price based. Competitive bidders indicate the
maximum price that they are prepared to pay as well as the quantity sought. In either
Treasury Auctions | 49

price-based or yield-based auctions, non-competitive bidders indicate only the quan-

tity sought. The implicit statement is that they are prepared to accept any price/yield
that is determined at the auction.
Smaller investors generally submit non-competitive bids. There is an upper limit
of $5 million for such bidders in the U.S. In some countries there is a separate quota
for such investors. In a price-based auction, bids are arranged in descending order of
price, for priority is given to a party willing to pay more. In a yield-based auction, the
bids are arranged in ascending order of yield, for a party willing to accept a lower rate
is given preference.5 In the U.S. all non-competitive bids are accepted in toto. The total
amount of such bids is subtracted from the amount being offered, and the balance is
offered to the competitive bidders. A competitive bidder is also subject to a restriction
in the U.S. that it cannot be allotted more than 35% of the amount being auctioned,
no matter how aggressively it bids.
There are two ways in which securities can be allotted, the multiple price/yield
auction known as a French auction, and the uniform price/yield auction known as a
Dutch auction. We will illustrate both methods by way of an example.

Illustration of a Treasury Auction

The Treasury is offering $35 billion worth of T-bonds. $5 billion worth of non-compe-
titive bids have been received. These are fully accepted, and therefore the amount
available for competitive bidders is $30 billion. There are eight competitive bids which
have been arranged in ascending order of yield. Had it been a price-based auction, we
would have arranged in descending order of price.

Table 2.6: A Treasury auction.

Bidder Bid Bid Aggregate

Yield Amount Demand
Dealer-1 2.372 3 bn 3 bn
Dealer-2 2.375 7 bn 10 bn
Dealer-3 2.380 5 bn 15 bn
Dealer-4 2.385 10 bn 25 bn
Dealer-5 2.388 10 bn 35 bn
Dealer-6 2.388 15 bn 50 bn
Dealer-7 2.400 5 bn 55 bn
Dealer-8 2.405 15 bn 70 bn

5 Price and yield are inversely related.

50 | Chapter 2: An Introduction to Bonds

Dealers 1–4 will have their bids accepted in full. This is because the total quantity
sought by the four of them is $25 billion, which is less than the available amount of
$30 billion. After this $5 billion is left to be allocated, but the total demand at the
next yield level is $25 billion. Both Dealer-5 and Dealer-6 have bid 2.388. Because only
$5 billion is left at this stage, there is pro-rata allocation. Their total demand is $25
billion, of which Dealer-5 has asked for $10 billion which is 40% of the amount, and
Dealer-6 has asked for $15 billion which is 60%. Thus 40% of the remaining amount
of $5 billion, which is $2 billion will be given to Dealer-5, and the balance $3 billion to
Dealer-6. Dealers 7 and 8 are not allocated any securities and are said to be shut out of
the auction. However a party who is shut out can always go to the secondary market
and acquire securities after the issue process.
In the U.S. the market clearing yield, also known as the stop-yield or high-yield,
which in this case is 2.388%, is rounded down to the nearest multiple of 0.125 to de-
termine the coupon. In this illustration, the coupon is set equal to 2.375%. In a Dutch
auction, Dealers 1–6 pay a price corresponding to a YTM of 2.388%. Obviously parties
who are prepared to accept a lower yield will have no objection to accepting a higher
yield. Because the coupon is the nearest lower multiple of 0.125%, the securities are
issued at par or at a discount. In a French auction, however, successful bidders pay a
price corresponding to their bid. Thus the price paid steadily decreases as we go from
Dealer-1 to Dealer-6. In either case, there is pro-rata allocation at the high-yield, and
bidders bidding higher than the market clearing yield are shut out.
The ratio of bids received, the total of competitive as well as non-competitive bids,
to the amount awarded is termed the bid-to-cover ratio. The higher the ratio, the more
the perceived success of the auction. Assume that the Treasury is issuing $25 billion of
bonds. If $150 billion worth of bids are received, the bid to cover ratio is 6:1. However
if $200 billion worth of bids are received, the ratio is 8:1, and the auction is deemed to
be more successful.

Security Identification
In the U.S. and Canada, securities are identified on the basis of a CUSIP number.
CUSIP stands for Committee on Uniform Securities Identification Procedures and is a
9-character alpha-numeric code. Most countries use ISIN, which stands for Interna-
tional Securities Identification Number. An ISIN is a 12-character alpha-numeric code.
To convert a CUSIP to a corresponding ISIN, a 2-character country code is added at
the beginning and a check digit is added at the end. For U.S. securities the code is US,
whereas for Canadian securities it is CA.
While specifying the details of a bond prior to a trade, it is essential to state the
issuer, the coupon, and the year and month of issue. In certain cases the currency
of issue also needs to be specified. In contrast, a stock usually can be identified by
just the name of the issuer. This is because most companies issue only one category
Coupon Strips | 51

of shares.6 However if we take an issuer like IBM, a bond maturing in 2030 may ma-
ture in January or in October. If it were to mature in October the coupon may be 4%
or 5%. Thus merely stating that the issue is an IBM bond maturing in 2030 is not ad-
equate. Both the month of expiration and the coupon rate have to be specified. Some
issuers have securities identical in terms of year and month of expiration as well as
the coupon, but differ with respect to the currency of issue. In such cases the currency
of issue also needs to be spelled out.

Coupon Strips
Coupon stripping is a technique for deriving zero-coupon bonds based on the under-
lying coupon-paying bonds. What used to happen in the 1980s was the following: An
investment bank would buy a large lot of a Treasury bond and transfer it to a spe-
cial purpose vehicle (SPV). The SPV would then issue synthetic zero-coupon securi-
ties, backed by the underlying coupon-paying T-bonds. The SPV was a single-purpose
dedicated trust.7 The SPV was empowered to hold the mother bonds, but it could not
sell them or lend them, pledge them as collateral, or write options on them. It was
entitled to issue zeroes, where each type of security represented the entitlement to a
single cash flow from the mother bond. This process of creating synthetic zeroes from
a conventional mother bond is coupon stripping. If the mother bond has N years to
maturity, we can create 2N + 1 types of zeroes because the mother bond will pay 2N
coupons, and there will be a terminal face value payment at the end. Thus the essence
of coupon stripping is that the entitlement to each cash flow from the underlying bond,
is sold separately. Thus while one investor may buy a security corresponding to the
first coupon, another may buy a security entitling him to a subsequent coupon or face
value. Assume that a T-bond with a face value of $100 million, annual coupon of 10%
paid semiannually, and 10 years to maturity, is acquired by an investment bank. There
will be 21 cash flows, that is, 20 coupons and one face value. Thus 21 types of securi-
ties can be issued. Each coupon payment would be for $5 million. If we assume that
the synthetic securities have a face value of $1,000, 5,000 securities corresponding
to each of the coupons can be issued. 100,000 zeroes corresponding to the principal
cash flow can also be issued. The motivation to issue these securities is that if the sum
of the prices of the baby bonds differs from that of the mother bond, then there is an
arbitrage opportunity. Coupon stripping is a type of financial engineering, whereby
securities that do not otherwise exist, are created.
The investment banks came up with exotic names for the synthetic zeroes. Mer-
rill Lynch, which was the pioneer, offered what it called TIGRS (Treasury Investment

6 Sometimes there could be multiple categories due to non-voting shares, or shares with differential
voting rights (DVR).
7 See Smith [56].
52 | Chapter 2: An Introduction to Bonds

Growth Receipts). Salomon came up with CATS (Certificates of Accrual on Treasury Se-
curities), and Lehman came up with LIONS (Lehman Investment Opportunity Notes).
Consequently these securities came to be known as “Animal Products,” and this seg-
ment of the capital market was known as the “Zoo.”
Trademarks are no longer being issued, because in 1985 the Treasury itself got into
the act to facilitate coupon stripping. An investor cannot buy a coupon or principal
strip directly from the Treasury. What happens is that a dealer who owns a T-bond
can ask the regional Federal Reserve Bank (FRB) where it is held, to replace it with
an equivalent set of STRIPS, or Separate Trading of Registered Interest and Principal
of Securities. There are two categories of STRIPS: C-STRIPS, which come from coupon
payments, and P-STRIPS, which come from principal payments. In 1987 the Treasury
started to allow the reverse process known as a Reconstitution. That is, a trader who
owns all the C-STRIPS and the P-STRIP corresponding to a T-bond, can request that
the FRB replace it with the mother bond. Every C-STRIP and P-STRIP emanating from
a mother bond is assigned a CUSIP. The mother bond is also assigned a CUSIP.
A C-STRIP maturing on a given date, say 15 May 2022, will have the same CUSIP
as another C-STRIP maturing on that day, although the two may have been generated
from different mother bonds. Thus the supply of C-STRIPS scheduled to mature on a
given date increases over time. For instance, take a 20-year T-bond issued on 15 May
2015 with a coupon of 4%. It will pay a coupon on 15 May 2022. Another 20-year T-
bond issued on 15 May 2018 with a coupon of 3%, will also pay a coupon on 15 May
2022. The C-STRIPS corresponding to the two mother bonds will have the same CUSIP.
We say that they are fungible. However, P-STRIPS are not fungible. Every P-STRIP cor-
responding to a mother bond has a unique CUSIP. Thus unlike C-STRIPS whose sup-
ply increases over time, the supply of P-STRIPS is fixed at the time of issuance of the
mother bond.8 In practice, a C-STRIP and a P-STRIP maturing on the same day usu-
ally do not have the same price. One reason is that the P-STRIPS are more liquid due
to greater availability. For instance, a T-bond with a face value of $100 million and
a coupon of 10% and 10 years to maturity, will generate 5,000 C-STRIPS correspond-
ing to each coupon date, and it will generate 100,000 P-STRIPS. There is yet another
reason for the popularity of P-STRIPS. Possession of them facilitates reconstitution.9
Assume that the sum of the STRIPS is cheaper than the mother bond. If a dealer has
the P-STRIP, he only has to acquire the C-STRIPS, which are relatively easier to ob-
tain because of fungibility. However, if he were to possess only some of the C-STRIPS,
he would have to buy the P-STRIP and the remaining C-STRIPS. Thus P-STRIPS carry
a higher price, or a lower yield, than C-STRIPS maturing on the same date. There is
however a category of investors termed as buy-and-hold investors, who as the name
suggests will buy and not trade prior to maturity. Such investors are likely to prefer

8 See Smith [56].

9 See Smith [56].
When Issued (WI) Trading | 53

C-STRIPS, as they offer a higher YTM. Long-term investors, like pension funds and in-
surance companies, prefer such strategies because their liabilities are long term in
nature. Consequently, by investing in such securities, they can lock in a rate of return
without worrying about issues such as the reinvestment of intermediate cash flows.
STRIPS are primarily acquired by institutional investors. However they are acces-
sible to individuals, as well. Many mutual funds invest in these securities and offer an
alternative avenue of investment to retail investors.

When Issued (WI) Trading

WI stands for when, as, and if issued.10 It is a market for forward trading in bonds.
That is, although the issue of these bonds has been announced, they are yet to be is-
sued. Trades commence on the date of announcement from the Treasury, and end on
the actual issue date. The WI market helps investors to gauge the market’s interest
in a security prior to the actual auction. Traders can take both long and short posi-
tions, which will settle on the scheduled issue date. The activity in the WI market has
a bearing on the outcome of a Treasury auction. Bidders who take long positions in
the WI market enter the auction with long positions, whereas those who are short in
the WI market do so with short positions. Thus the strategy employed by a trader is
influenced by activities in the WI market. The presence of the WI market facilitates
price discovery. On or before the date of the auction, securities trade on a yield basis
in the WI market. This is because the coupon has not been set, and consequently the
price cannot be determined. From the day after the auction, securities can be traded
on a price basis because the coupon has been established.

Coupon Rolls
In a coupon roll, a dealer acquires the most recently issued security for a maturity from
a client and simultaneously sells the same amount of the recently announced security.
Thus a roll has two legs. The first leg settles on a T + 1 basis, whereas the second leg
is essentially a forward contract. There also exists what is called a reverse roll. In this
case, the dealer sells the most recently issued security to a client and simultaneously
buys the same amount of the newly announced security. The forward leg in rolls and
the reverse rolls is a when issued (WI) trade.
The issue of interest is the spread between the yield on the new security and that
on the outstanding security. If a dealer “gives” 10 basis points (bp) in a roll, it means
that the WI security has a YTM that is higher by 10 bp compared to the existing security.

10 See Liaw [43].

54 | Chapter 2: An Introduction to Bonds

On the contrary, if a dealer “takes” 10 bp in a roll, it means that the WI security has a
YTM that is lower by 10 bp compared to the outstanding issue.

Accounting for Bonds

Bonds are typically issued at par, but may be issued at a premium or at a discount.
Accounting for a par bond issue is fairly simple. Assume Alpha Corporation issues
1,000 JD-15 bonds on 15 June 2018. The face value is $1,000, the coupon is 10% per
annum paid semiannually, and the time to maturity is three years. When the bonds
are issued, there is an inflow of cash. Thus cash, which is an asset, is debited with
$1,000,000, and bonds payable, a liability is credited with $1,000,000. On the first
coupon date which is 15 December 2018, the total interest payable is $50,000. Bond
interest expenses is debited with $50,000. If the payout is immediate, cash is credited
with $50,000. Otherwise, interest payable is credited with $50,000.

Issue of Discount Bonds

Now assume that while the coupon continues to remain at 10% per annum, the YTM
at the time of issue is 12%. The bond will obviously trade at a discount. The price will
be $950.8268. If 1,000 bonds are issued, the cash inflow will be $950,827. Thus cash
will be debited with this amount. Bonds payable will be credited with $1,000,000.
Unamortized bond discount will be debited with the difference, which is $49,173. This
discount has to be amortized, or written off, during the life of the bond. Thus the dis-
count will steadily decline, and the carrying value, which is the difference between
the face value and the unamortized discount, will steadily increase. At the time of ma-
turity of the bond, the unamortized discount will be zero, and the carrying value will
be equal to the face value. When a company issues a bond at a discount, the effective
interest paid is greater than the coupon rate. This is because at the outset, the issuer
receives an amount less than the face value, whereas at maturity it has to pay the en-
tire face value. Thus the discount must be added to the total coupon paid, in order to
arrive at the effective interest. In our illustration, the total coupon paid over three years
is $300,000. The discount is $49,173. Thus the total interest cost is $349,173. The dis-
count is allocated over the life of the security. Thus every period, the interest expense
will be equal to the coupon interest paid, plus the amortized discount. This procedure
is called amortization of bond discount. There are two approaches, the straight line
method (SLM) and the effective interest method (EIM). We discuss both in the follow-
ing sections.11

11 See Needles and Powers [50].

Accounting for Bonds | 55

The Straight Line Method for Amortizing the Bond Discount

SLM equalizes the amortized discount for every period. In our case $50,000 interest is
paid out twice a year for three years. As per the SLM, the interest expense per period is
349,173
6
= $58,195.50. Thus, while the cash interest paid is $50,000, the interest expense
is higher by $8,195.50. In the books of accounts, we debit bond interest expense with
$58,195.50, and credit cash with $50,000. Unamortized bond discount is credited with
the difference, which is $8,195.50. At maturity, the unamortized bond discount will be
zero.
This method is simple, but has an inherent weakness. Every period the carrying
value increases, whereas the interest expense remains constant. Thus the rate of in-
terest declines over time.

The Effective Interest Method for Amortizing the Bond Discount

In this method, the YTM prevailing at the time of issue is multiplied by the carrying
value of the bond at the beginning of the period. The amount amortized each period
is the difference between the interest computed using the YTM and the actual interest
paid. In our case the original YTM is 12%, or 6% per half-yearly period. The discount
is $49,173, and the initial carrying value is $950,827.
The interest expense for the first half-year is 0.06 × 950,827 = $57,050. The ac-
tual interest paid is $50,000. Thus the amortized discount is $7,050. The unamortized
discount at the end of the first half-yearly period is 49,173 − 7,050 = $42,123. This will
decline at the end of every period. Thus the carrying value will steadily increase. In ac-
counting terms, we would do the following. Debit interest expense with $57,050. Credit
cash with $50,000 and credit unamortized bond discount with $7,050. Similarly the
interest expense for the next semiannual period is 0.06 × (950,827 + 7,050) = $57,473.
The actual interest paid is $50,000, and the amortized discount is $7,473. The unamor-
tized discount is 42,123 − 7,473 = $34,650. The carrying value at the end of the second
half-yearly period is $965,349. The sequence of entries over the life of the bond is sum-
marized in Table 2.7.

Table 2.7: Accounting for a discount bond.

Period Beginning Interest Coupon Amortized Unamortized Ending

Carrying Value Expense Discount Discount Carrying Value

0 49,173 950,827
1 950,827 57,050 50,000 7,050 42,123 957,877
2 957,877 57,473 50,000 7,473 34,651 965,349
3 965,349 57,921 50,000 7,921 26,730 973,270
4 973,270 58,396 50,000 8,396 18,334 981,666
5 981,666 58,900 50,000 8,900 9,434 990,566
6 990,566 59,434 50,000 9,434 0 1,000,000
56 | Chapter 2: An Introduction to Bonds

As can be seen, the carrying value gradually increases and approaches the face value
at maturity. The interest expense is a fixed percentage of the carrying value, and there-
fore it too increases every period. Because the coupon payments are constant, the
amortized discount steadily increases.

Issue of Premium Bonds

Let’s consider the same data for the issue of premium bonds. 1,000 bonds are issued
with a face value of $1,000 each and three years to maturity. The coupon is 10% per
annum payable semiannually, and the YTM is 8% per annum, not 12%, as assumed
earlier.
The initial cash inflow from the bonds is $1,052,421. Thus there is an unamortized
bond premium of $52,421. This too needs to be amortized over the life of the securities.
The premium represents an amount that need not be paid by the issuer at maturity.
Thus it may be considered an advance reduction of the total interest paid over the life
of the bond. The total interest over three years is $300,000. As the premium is $52,421,
the total interest cost is $247,579. Once again this may be amortized using the SLM or
the EIM.

Straight Line Method for Amortizing the Bond Premium

SLM equalizes the amortized premium for every period. In our case $50,000 interest
is paid out twice a year for three years. As per the SLM the interest expense per pe-
riod is 247,579
6
= $41,263. Thus, while the cash interest paid is $50,000, the interest
expense is lower by $8,737. In the books of accounts, we debit bond interest expense
with $41,263. Cash is credited with $50,000. Unamortized bond premium is debited
with the difference, which is $8,737. At maturity, the unamortized bond premium will
be zero.
This method is simple, but has the same inherent weakness, as the SLM for bond
discounts. Every period the carrying value decreases, whereas the interest expense
remains constant. Thus the rate of interest increases over time.

The Effective Interest Method for Amortizing the Bond Premium

declines at the end of every period. Thus the carrying value steadily declines. In ac-
counting terms we would do the following: Debit interest expense with $42,097, credit
cash with $50,000, and debit unamortized bond premium with $7,903. Similarly the
interest expense for the next semiannual period is 0.04 × (1,052,421 − 7,903) = $41,781.
The actual interest paid is $50,000, and the amortized premium is $8,219. The unamor-
tized premium is 44,518 − 8,219 = $36,299. The carrying value at the end of the second
half-yearly period is $1,036,299. The sequence of entries over the life of the bond is
summarized in Table 2.8.

Table 2.8: Accounting for a premium bond.

Period Beginning Interest Coupon Amortized Unamortized Ending

Carrying Value Expense Premium Premium Carrying Value

0 52,421 1,052,421
1 1,052,421 42,097 50,000 7,903 44,518 1,044,518
2 1,044,518 41,781 50,000 8,219 36,299 1,036,299
3 1,036,299 41,452 50,000 8,548 27,751 1,027,751
4 1,027,751 41,110 50,000 8,890 18,861 1,018,861
5 1,018,861 40,754 50,000 9,246 9,615 1,009,615
6 1,009,615 40,385 50,000 9,615 0 1,000,000

Risks Inherent in Bonds

Before we go on to study the various types of risks that impact a bond holder, let’s first
define risk. The term risk refers to the possibility of loss arising due to uncertainty re-
garding the outcome of a transaction. Thus there are two conditions, for a transaction
to be classified as risky. First, there must be more than one possible outcome. And
second, at least one of the probable outcomes should lead to a loss. All bonds expose
bond holders to multiple sources of risk.

Credit Risk

Credit risk, also known as default risk, refers to the probability of default by the bor-
rower. That is, it is the risk that coupon payments and/or principal payments may not
be made as promised. Treasury securities or federal government securities are backed
by the full faith and credit of the federal government. Consequently they are virtually
devoid of default risk.
At the time of a bond issue, the issuer provides an offer document or prospectus
that gives information about its financial soundness and credit-worthiness. But every
potential investor cannot be expected to decipher the intricacies of such a document.
58 | Chapter 2: An Introduction to Bonds

Thus in practice, we have credit rating agencies. Rating agencies, such Standard &
Poor’s Corporation (S&P), Moody’s Investors Service, and Fitch Ratings, specialize in
evaluating the credit quality of an issue at the time of issue. Subsequently they con-
stantly monitor the financial health of the issue and re-rate or modify their recommen-
dations if it found to be necessary.
Top quality issues are classified as investment grade, whereas issues of a lower
quality are termed as speculative grade, or junk bonds.

Moody’s Ratings Scale

Moody’s uses the following rating scale for long-term instruments:

– Aaa: The issue is of the highest quality and carries minimal risk.
– Aa: The issue is of high quality and has very low credit risk.
– A: The security is of upper medium grade and has low credit risk.
– Baa: The security is of medium grade with moderate credit risk.
– Ba: The security has a speculative element and has substantial credit risk.
– B: The security is speculative and has high credit risk.
– Caa: The security is of poor quality and has very high credit risk.
– Ca: The security is highly speculative and is likely to default. However, there is
some possibility of recovery of principal and interest.
– C: This is the lowest rating; such bonds are typically in default. There is very little
possibility of recovery of principal and interest.

Moody’s appends modifiers of 1, 2, and 3 to grades Aa to Caa. A modifier of 1 indicates

that the security is at the higher end of the category. On the other hand, a modifier
of 2 indicates that the security is at the middle of the category. Finally a modifier of 3
indicates that the security is at the lower end of the category. Aaa to Baa are considered
to be of investment grade, whereas the rest are of speculative grade.

S&P’s Rating Scale

Standard and Poor’s rates issues on a scale from AAA to D. Country risk and currency
of repayment are factored into the credit analysis and reflected in the issue rating.
– AAA: These securities are of the highest quality. The issuer’s capacity to meet its
financial commitments is extremely strong.
– AA: These securities are slightly below those rated AAA. The issuer’s capacity to
meet its financial commitments is very strong.
– A: These are somewhat more vulnerable to adverse changes in economic condi-
tions. The issuer’s capacity to meet its financial commitments is still strong.
Risks Inherent in Bonds | 59

– BBB: These securities display adequate protection parameters. But adverse eco-
nomic conditions are more likely to weaken the issuer’s capacity to meet its finan-
cial commitments.
– BB: These are speculative grade securities. However, they are less risky compared
to other securities which are in this category. They are vulnerable to adverse eco-
nomic conditions, which could seriously impact the issuers’ capacity to meet its
financial commitments.
– B: These are more vulnerable to non-payment compared to BB rated securities.
But the issuer currently has the capacity to meet its financial commitments.
– CCC: These securities are currently vulnerable to non-payment. Favourable mar-
ket conditions are required, if the issuer is to meet its commitment. If economic
conditions become adverse, the issuer is unlikely to have the capacity to meet its
financial commitments.
– CC: These securities are highly vulnerable to non-payment. Default has not yet
occurred, but is considered to be a virtual certainty.
– C: These are extremely vulnerable to nonpayment and the chance of recovery is
lower compared to higher rated securities.
– D: These securities are in default or the issuer has filed for bankruptcy.

Grades from AA to CCC may be modified by a + or a − sign, to indicate relative standing.

Bonds rated AAA to BBB are considered to be of investment grade, while those with
lower rankings are categorized as speculative.

Fitch’s Rating Scale

Fitch assigns ratings ranging from AAA at one end to D at the other extreme. Grades
AAA to BBB are investment grade. The rest are speculative grade.
– AAA: Lowest expectation of default risk. It indicates that the issuer is exception-
ally strong from the standpoint of meeting financial commitments.
– AA: Very high credit quality. It indicates an expectation of very low default risk.
It indicates that the issuer is very strong from the standpoint of meeting financial
commitments.
– A: High credit quality. It indicates an expectation of low default risk. It indicates
that the issuer is strong from the standpoint of meeting financial commitments.
– BBB: Good credit quality. It indicates that the expectation of default risk is cur-
rently low. The issuer’s capacity to meet financial commitments is adequate.
– BB: Speculative. There is a greater risk of default. However, the issuer has financial
flexibility that could facilitate the fulfillment of financial obligations.
– B: Highly speculative. There is material default risk, but also a limited margin of
safety. Although financial obligations are currently being met, the future is uncer-
tain.
60 | Chapter 2: An Introduction to Bonds

– CCC: Substantial credit risk. Default is a real possibility.

– CC: Very high level of credit risk. Default appears probable.
– C: Near default. A default or default-like process has begun.
– RD: Restricted default. There has been a default but the issuer has not filed for
bankruptcy, and continues to operate.
– D: Default. The issuer has ceased business and entered into bankruptcy filing or
liquidation.

Non-investment grade bonds are also known as junk bonds, or high-yield bonds. There
are two types of junk bonds. Original issue junk bonds carry a lower rating from the time
of issue. Fallen angels on the other hand were originally classified as investment grade,
but have been subsequently downgraded to junk status. Thus bond ratings can change
over time. If a rating agency is planning to change its rating, it will signal its intention.
S&P places the security on Credit Watch. Moody’s places it on a list termed as Under
Review, whereas Fitch places it on a list termed as Rating Watch. The appearance of a
security on one or more of these lists is a signal that a change in ratings is on the way.
Rating agencies take various facts and factors into account when assigning a rat-
ing. The primary task is to evaluate the ability of issuers to meet their commitments on
time. Factors considered include the issuer’s financial condition and quality of man-
agement; the features of the security being issued; and the nature of revenue sources
backing the issue. The last issue is very important for bonds used to generate funds for
projects that are projected to yield revenues directly. For instance, a municipality may
be planning the construction of an airport with a projected footfall of five million pas-
sengers a year. The rating agency will recalculate the projections using a lower figure
of say three million passengers per annum. The standard policy is to scale down the
projected revenue and scale up the projected expenses, and see whether the project is
still financially viable.
In case you are wondering about the accuracy of these ratings, here are some
statistics.12 Moody’s did a survey using data for an 81-year period from 1920 to 2001.
They found that not a single Aaa-rated bond defaulted in the first year following issue,
during this 81-year period. However, about 7% of B-rated bonds defaulted in the first
year. When they considered a 10-year period following issue, they found that fewer
than 1% of Aaa bonds defaulted, whereas about 44% of B-rated bonds did default.

Bond Insurance
A potential bond issuer can approach an insurance company to insure its bonds to
enhance the credit quality. The rating assigned to the bond depends on the rating of
the insurance company. Thus the issuer needs to obtain insurance from a company

12 See Liaw [43].

Risks Inherent in Bonds | 61

whose rating is superior to its own. The insurer charges a premium, but this can be
passed on to the bond investors in the form of a lower coupon. In the case of such
bonds the insurance company guarantees the timely payment of the principal and
interest. Rating agencies evaluate the issue based on the insurer’s capital adequacy
and claims paying ability.
In the U.S. an investor can have his personal bond portfolio insured by an insur-
ance company. However, this feature is not available in all countries. In the case of
insured bonds, the guarantee provided is unconditional and irrevocable. This means
that if the rating agencies subsequently downgrade the issue, the insurance will not
be revoked. From the standpoint of an investor, an insured bond carries the assurance
of extensive credit analysis and due diligence by two parties, namely the rating agency
and the insurance company.

Liquidity Risk

A market is said to be illiquid or thin when there is an insufficient number of buyers

and sellers. In a liquid market, trades take place at prices close to the true or fair value
of the asset. In an illiquid market, buyers may have to pay a large premium to buy, or
sellers may have to accept a significant discount to sell. Illiquid markets are charac-
terized by large bid-ask spreads because trades will be few and far between. In other
words, if the volume of activity is low, the margin for the dealer must be high, and
consequently the spread will be higher. The impact cost of a trade is another measure
of the liquidity of a market. Consider the limit order book for a security, as depicted in
Table 2.9.

Table 2.9: Illustration of a limit order book.

Quantity Price Price Quantity

100 99.00 100.00 200

200 98.75 100.50 300
300 98.00 101.00 500
500 97.50 102.00 250
1000 97.00 102.50 500

The best bid is $99 and the best ask is $100. Thus no buyer is willing to pay more than
$99, while no seller is willing to accept less than $100. The truth is obviously in be-
tween, and our best guess would be a simple average. Consequently we define the fair
price (FP) as the average of the two prices, which in this case is $99.50. Consider a buy
order for 400 shares; 200 shares would be bought at a price of $100, and another 200
at $100.50. Thus the weighted average price (WAP) is $100.25. We define the impact
62 | Chapter 2: An Introduction to Bonds

cost of the trade as

WAP − FP
FP
100.25−99.50
In this case it is 99.50
= 0.7538%. For a sell order we would define the cost as

FP − WAP
FP
For buy orders the weighted average price is greater than the fair price. However,
for sell orders, the weighted average price is less than the fair price. Thus by definition
the impact cost is always positive. The impact cost depends on the size of the trade as
well as the direction. That is, a buy order for N shares does not usually have the same
impact cost as a sell order for the same quantity. For an order of a given size, the lower
the impact cost, the greater the perceived liquidity.

Interest Rate Risk

Interest rates or yields are the key variable of interest in debt markets. The term inter-
est rate risk refers to the possibility of rates moving in an adverse direction from the
standpoint of the bond holder. Unlike other sources of risk, interest rate risk affects
bonds in not one, but in two ways. The first source of risk is what is termed as reinvest-
ment risk. All bonds, other than zero-coupon bonds, make payments periodically in
the form of coupon payments. These inflows have to be reinvested. Reinvestment risk
refers to the possibility that rates may decline by the time the coupon is received. If so,
the cash inflow has to be invested at a lower than expected rate. This risk impacts all
bonds that make payments prior to maturity. Thus zero-coupon bonds are not subject
to this risk, and that explains their popularity among certain classes of investors.
The second source of interest rate risk is what is termed as market risk or price
risk. An investor who holds the bond until maturity gets the face value. However, if
the bond is sold prior to maturity, the trader receives the market price prevailing at
that point in time, which would be inversely related to the prevailing yield. If interest
rates increase, the bonds have to be sold at a lower price. This is termed market risk
or price risk. The two risks work in opposite directions. Reinvestment risk affects the
investor if rates decline, whereas price risk impacts him if rates increase.

Inflation Risk

Inflation refers to the erosion of the purchasing power of money. Most bonds, including
Treasury bonds, promise cash flows in dollar terms. There is no assurance about the
basket of goods and services that can be purchased with the cash received. Inflation
Chapter Summary | 63

risk is the risk that the actual inflation may be higher than the expected inflation.
When a buyer acquires a bond, the market has a perceived rate of inflation, on the
basis of which the nominal or dollar rate of interest is set. However, the actual inflation
at the time the coupon is received may be higher, which means that the investor has
to settle for a smaller basket of goods and services. The rate of return, as measured by
the ability to buy goods and services, is termed as the real rate of return. Thus, if the
ex-post rate of inflation is higher than the ex-ante rate, the ex-post real rate of return
will be lower than the ex-ante rate.13 This risk is termed as inflation risk.
There exists a category of bonds known as indexed bonds. In the case of such
bonds the coupon is linked to a price index. The higher the inflation is, the higher
the value of the price index and the larger the coupon. Thus these bonds offer higher
cash flows when the inflation rate is high and lower cash flows when it is low, thereby
ensuring that the real return is kept at a virtually constant level.14

Timing Risk

In the case of a plain vanilla bond, there is no uncertainty about when the cash flows
will be received, unless of course there is a default. But consider a callable bond. The
investor has no idea when this bond may be recalled in the future. That is, he is unsure
about the number of coupons he will receive and the point of receipt of the face value.
This is termed as timing risk.

Foreign Exchange Risk

The rate of conversion of one currency into another is known as the exchange rate. Like
stock prices and interest rates, exchange rates are also random variables. Take the case
of an American investor who buys a bond whose cash flows are denominated in euros.
The bond is scheduled to pay a coupon of 10 euros every six months. Assume the rate at
the beginning is 1.2000 U.S. dollars per euro. However, the dollar appreciates, and the
rate after six months is 1.0500 U.S. dollars per euro. Thus while the anticipated cash
inflow is $12, the actual inflow will be only $10.50. This is called foreign exchange risk.

Chapter Summary
This chapter began by illustrating the benefits of a bond issue to the issuer, due to
issues such as leverage and tax shields. The variables influencing the price of a plain

13 Ex-ante means before the event and ex-post means after the event.
14 We examine such bonds in Chapter 7.
64 | Chapter 2: An Introduction to Bonds

vanilla bond were analyzed in detail, and the fundamental bond valuation equation
was derived. In this context the pull to par effect and its inherent reasons were studied
in detail. The topics of Eurobonds and foreign bonds were elaborated upon. We then
looked at zero-coupon bonds and the synthesis of such bonds using two plain vanilla
bonds. Subsequently we studied certain bond types with bells and whistles, such as
amortizing bonds, bonds with step-up and step-down coupons, and payment-in-kind
bonds. The Treasury securities market was the next focus of attention, and we ana-
lyzed a hypothetical Treasury auction in detail. The related issue of coupon stripping
was presented in depth. In this context we looked at trademarks, as well as their suc-
cessors, namely STRIPS. While bonds are often issued at par, they may at times be
issued at a discount or at a premium. We looked at the accounting implications for
such issues. The chapter concluded by studying the various types of risks inherent in
bonds. In this context we presented the rating scales of Moody’s, Fitch, and S&P in
detail.
In the next chapter, we examine the valuation of bonds between coupon dates,
and the related concepts of day-count conventions and accrued interest. We examine
various yield measures, such as the yield to maturity, the realized compound yield,
and the horizon yield. We also introduce the real-life issue of income tax and examine
its implication for the realized compound yield, as well as for the yield to maturity.
At the end of the chapter, we examine the case of bonds with sinking fund provisions
and introduce the concepts of the yield to average life and the yield to equivalent life.
Chapter 3
Bonds: Advanced Concepts
Let us now examine the process of valuing a bond between coupon dates. While valu-
ing a bond between coupon dates, the difference is that the next coupon is not exactly
one period away, but is k periods away where k < 1. Assume that we have a bond with
a face value of M, paying a coupon of C/2 every six months, and with N coupons left
prior to maturity. The semiannual yield to maturity is y/2.

Required Symbols for the Variables

– y ≡ Yield to maturity.
– N ≡ Number of coupons left in the life of the bond.
– k ≡ Time until the receipt of the first coupon expressed as a fraction of six months.
If we are standing on a coupon date, then k = 1 else k < 1.
– Pi,t ≡ Clean price of bond i in the spot market at time t.
– Pd ≡ Dirty price.
– AI i,t1 ,t ≡ Accrued interest on bond i from t1 (the last coupon date) till t (the settle-
ment date).

We first value the bond at time 1, or the next coupon date, and then discount the value
so obtained, back to time 0, or the actual settlement date.
At time 1 we receive a coupon of C2 . There are N − 1 coupons left subsequently,
whose value is
C
2 1
y [1 − N−1
]
2 (1 + y2 )

M
The present value of the face value is N−1 Thus the value of the bond at time 1 is
(1+ y2 )

C
C 2 1 M
+ y [1 − ]+
2 2 (1 + y2 )
N−1
(1 + y2 )
N−1

N
C
2
(1 + y2 ) − 1 M
= y [ N−1
] + N−1
2 (1 + y2 ) (1 + y2 )

Discounting back to time 0, we get a value of

N
c×M
2
[(1 + y2 ) − 1] M
[ × ]+
y k y N−1 y N−1+k
[ (1 + 2
) ( y2 )(1 + 2
) ] (1 + 2
)
https://doi.org/10.1515/9781547400669-003
66 | Chapter 3: Bonds: Advanced Concepts

Day-Count Conventions
To value the bond using the preceding approach, we need to estimate k, that is, the
length of the fractional first period. Unfortunately, there is no universal approach for
computing k. Different markets use different approaches, and at times, even within a
market, the convention varies from product to product. Consequently, we need to be
familiar with various conventions known as day-count conventions. We will first look
at the Actual/Actual method which is used for Treasury bonds in the U.S.

The Actual/Actual Approach

Assume that we are standing on 25 July 2018. There is a T-bond maturing on 15 Novem-
ber 2035 that pays a coupon of 10% per annum on 15 May and 15 November every
year. The face value is $1,000 and the YTM is 12.50% per annum. In the Actual/Ac-
tual method, we need to calculate the actual number of days between the settlement
date and the date of the next coupon. Let’s call this period as N1 . In our case, the next
coupon date is 113 days away as can be determined from Table 3.1.
We have to then compute the number of days between the last coupon date and
the next coupon date. We denote the period as N2 . In our example, the number of
days between the last coupon date and the next coupon date is 184 days as can be
determined from Table 3.2.

Table 3.1: Number of days between the settlement date and the next coupon date.

Month No. of Days

July 6
August 31
September 30
October 31
November 15
Total 113

Table 3.2: Number of days between the two coupon dates.

Month No. of Days

May 16
June 30
July 31
August 31
September 30
October 31
November 15
Total 184
Day-Count Conventions | 67

Hence the next coupon is k = NN1 = 184

113
= 0.6141 semiannual periods away.
2
The Actual/Actual method, often denoted by ACT/ACT and pronounced ack/ack
by Wall Street traders, is one of many day-count conventions that are used in the world
of finance. We will discuss other methods shortly. The denominator under this conven-
tion varies from 181 to 184 days, depending on the coupon payment dates of the bond
that is under consideration. Note that while calculating the numerator and denomi-
nator for computing the fractional period, we can either include the starting date or
the ending date, but not both. We have chosen to include the ending date.
Having determined the value of k, the price of the Treasury bond may be deter-
mined in one of two ways: the market method or the Treasury method.

The Market Method for Bond Valuation

Wall Street traders compute T-bond prices using the expression derived earlier in this
chapter.
N
c×M
2
[(1 + y2 ) − 1] M
Pd = [ × ]+ (3.1)
y k y y N−1 y N−1+k
[ (1 + 2 ) ( 2 )(1 + 2 ) ] (1 + 2 )
Using the example from the preceding section, the price as determined by this
method is
0.125 35
0.10×1,000 [(1 + 2
) − 1] 1,000
2
× +
0.125 0.6141 0.125 35−1 0.125 35−1+0.6141
(1 + 2 ) ( 0.125
2
)(1 + 2
) (1 + 2
)
= $843.4655

In Excel, we would use the PRICE function, in conjunction with the ACCRINT func-
tion, to compute the dirty price using the market method, as we shall shortly demon-
strate. However, the price can be computed even without invoking the PRICE function.
We can use the PV function to first compute the bond’s price on the next coupon date.
We can then add that day’s coupon to it, and discount the whole expression back to
time 0, using the fractional period that we have computed.
In our case, the present value = PV(0.125/2, 34, −50, −1000) = $825.4593. To that
we add the coupon of 50, which adds up to $875.4593. This can be discounted for 0.6141
periods to get the final answer of $843.4655.

The Treasury Method for Bond Valuation

The difference between the market’s approach and that of the Treasury is that the Trea-
sury uses simple interest to discount for the fractional first period. Hence, the Treasury
68 | Chapter 3: Bonds: Advanced Concepts

calculates the bond price as follows:1

N
c×M
2
[(1 + y2 ) − 1] M
Pd = [ × ]+ (3.2)
(1 + k × y2 ) ( y )(1 + y )N−1 y y N−1
[ 2 2 ] (1 + k × 2 )(1 + 2 )

Using the same example, the price as determined by this method is

0.125 35
0.10×1,000 [(1 + 2
) − 1]
2
×
+ 0.6141×0.125 0.125 35−1
(1 2
) ( 0.125
2
)(1 + 2
)
1,000
+
0.6141×0.125 0.125 35−1
(1 + 2
)(1 + 2
)
= $843.1000

This can be easily computed using the PV function. From earlier, we know that
the value on the next coupon date, is $875.4593.

875.4593
0.125×0.6141
= $843.1000
(1 + 2
)

For a given value of k, the Treasury’s approach always gives a lower price, be-
cause when a fractional period is involved, the discount factor using simple interest
is higher than the factor obtained with compound interest. Consequently the present
value, which is nothing but the bond price, is lower.

Accrued Interest
Consider the period between two coupon dates t1 and t2 . Let’s denote the ex-dividend
date for the bond by td , where t1 < td < t2 . Between t1 and td , the bond trades cum
dividend, which means that the person who buys the bond at any point in time during
this period is entitled to the coupon that is paid on t2 .2 This facet is captured by the
bond pricing equation (3.1), because the price includes the present value of the next
coupon.
The price that is obtained as the present value of all the remaining cash flows is
termed as the dirty price. It includes the compensation due to the seller for parting
with the entire next coupon payment, although he has held the bond for a part of the

1 See Garbade [33].

2 The term ex-dividend is a misnomer, for bonds pay coupons and not dividends, unlike stocks. How-
ever the implication is the same as that for a stock.
Accrued Interest | 69

current coupon period. This compensation is called accrued interest and is calculated
as follows:
c×M (t − t1 )
Accrued interest = AI = × (3.3)
2 (t2 − t1 )
where t is the settlement date. Using the previous example, the accrued interest as of
25 July, 2018 is
0.10 × 1000 71
× = $19.2935
2 184
However the bond prices that are quoted in practice are called clean prices and are
computed by subtracting the accrued interest from the dirty price. Therefore
Pi,t = Pd − AI i,t1,t (3.4)
The rationale for computing clean prices is the following. Consider the bond price
after three days, that is, on 28 July. The dirty price on this day is $844.2995. The pas-
sage of three days time leads to an increase of $0.8340 in the dirty price, even though
the YTM has remained unchanged. We know that the price will change if the YTM
changes. However, even if the YTM remains constant, the dirty price will change with
the sheer passage of time. But the clean price will stay relatively constant for short
periods, unless there is a change in the YTM. In this example, the clean price on 25
July is $824.1721, whereas on 28 July it is $824.1908. The difference in the two prices is
$0.0187.
For a bond market analyst, it is important to monitor the changes in the yields
prevalent in the market. If the price data were to consist of dirty prices, it would be
difficult to separate the effect of yield changes from the impact of the accrued interest.
However if clean prices were used for the analysis, any price changes in the short run,
would be induced primarily by yield changes. It is for this reason that the reported
bond prices are invariably clean prices.
One question that may strike you is, whether the accrued interest can be negative.
That is, can there be cases where the seller of the bond has to pay accrued interest
to the buyer? The answer is yes. In some bond markets, the bonds begin to trade ex-
dividend after a certain date. That is, from this date onwards, the sale of a bond results
in the next coupon going to the seller rather than to the buyer. On the ex-dividend date,
the dirty price falls by the present value of the next coupon, and the dirty price is less
than the clean price. We can illustrate this with the help of Example 3.1.

Example 3.1. Consider the bond that matures on 15 November 2035. We will assume that we are stand-
ing on 10 November 2018, which happens to be the ex-dividend date.
The cum-dividend price of the bond is
0.125 35
0.10×1,000 [(1 + 2
) − 1] 1,000
2
0.0272
× +
0.125 35−1 0.125 35−1+0.0272
(1 + 0.125
2
) ( 0.125
2
)(1 + 2
) (1 + 2
)
= $874.0168
70 | Chapter 3: Bonds: Advanced Concepts

The moment the bond goes ex-dividend, the dirty price falls by the present value of the forthcoming
coupon, because the buyer is no longer be entitled to it. Thus, the ex-dividend dirty price is

0.10×1,000
2
874.0168 −
0.125 0.0272
(1 + 2 )
= $824.0992

This is the amount payable by the person who buys the bond an instant after it goes ex-dividend.
The accrued interest an instant before the bond goes ex-dividend is

0.10 × 1,000 179

× = $48.6413
2 184

Thus the clean price at the time of the bond going ex-dividend is

874.0168 − 48.6413 = $825.3755

As you can see the clean price is greater than the ex-dividend dirty price. This negative accrued interest
represents the fact that the seller of the bond has to compensate the buyer because although the
buyer is entitled to receive his share of the next coupon, the entire amount is received by the seller.
The fraction of the next coupon that is payable to the buyer is

0.10 × 1,000 5
× = $1.3587
2 184

Hence the buyer has to pay 825.3755 − 1.3587 = $824.0168, which is close to the ex-dividend dirty
price.

All bonds do not exhibit the ex-dividend phenomenon. Some trade cum-dividend till
the coupon date. Consequently the accrued interest in such cases is always positive.

The Impact of Time on the Dirty Price

To understand the impact of time on the price, let us derive the partial derivative of
the dirty price of the bond, with respect to time.

y N
1 { c × M [(1 + 2 ) − 1] M }
Pd = × { 2 × +
y k y y N−1 y N−1 }
(1 + 2 ) { ( 2 )(1 + 2 ) (1 + 2 ) }
𝜕Pd y
= −Pd × ln (1 + )
𝜕k 2

As k decreases, the price increases. Hence, the dirty price steadily increases as we go
from one coupon date to the next, keeping the other parameters constant.
Computation of Price and Accrued Interest Using Excel | 71

Computation of Price and Accrued Interest Using Excel

When we price a bond on a coupon date, we can use the PV function in Excel to com-
pute the clean price. This is the same as the dirty price on that day, since the accrued
interest is zero.3 However, when we value a bond between coupon dates, we need to
use the Excel function called PRICE to compute the clean price. The accrued inter-
est has to be separately computed using the ACCRINT function. The two can then be
added together determine the dirty price.
Consider the PRICE function. The parameters are
– Settlement is the settlement date, and has to be entered using the DATE function
in a YYYY,MM,DD format.
– Maturity refers to the maturity date that has to be entered in a similar fashion
using the DATE function.
– Rate is the annual coupon rate.
– Yld is the annual YTM.
– Redemption represents the percentage of par that is payable when the bond ma-
tures and in most cases is 100 or 100%. It is not the face value, which in most cases
is $1,000.
– Frequency is the frequency of coupon payments. For bonds paying coupons on a
semiannual basis, the frequency is 2.
– Basis denotes the day-count convention. For the PRICE function, we need to spec-
ify one of the values given in Table 3.3.

Table 3.3: Basis values in Excel for different day-count conventions.

Basis Day-Count
Convention

0 30/360 NASD
1 Actual/Actual
2 Actual/360
3 Actual/365
4 Actual/360 E

In our case we invoke the PRICE function using the following parameters:
DATE(2018,07,25),DATE(2035,11,15),0.10,0.125,100,2,1. Since the face value is $1,000,
the price has to be multiplied by 10. The value obtained is $824.1705.
To compute the accrued interest, we need to invoke the ACCRINT function.

3 Obviously, on a coupon date, the accrued interest is zero.

72 | Chapter 3: Bonds: Advanced Concepts

The parameters are

– Issue: The issue date of the bond, to be input using the DATE function.
– First_interest: The next coupon date, to be input using the DATE function.
– Settlement: The settlement date, to be input using the DATE function.
– Rate: The annual coupon rate in decimals.
– Par: We can give the full par value, whatever it may be. Thus if the bond has a face
value of $1,000, we will give the par as 1,000.
– Frequency: The frequency of coupon payments in a year, which is usually 2.
– Basis: The day-count convention.
– Calc_method: This is required when we specify the actual issue date for the Issue
parameter. If however we were to input the previous coupon date as the date of
issue, this variable is not required. A more detailed explanation is given in the
paragraph that follows.

If we happen to know the issue date, we can specify it using the DATE function. How-
ever, if we do not know when the bond was issued, the issue date should be entered as
the previous coupon date. In our case, we specify it as DATE(2018,05,15). The First_in-
terest date is the next coupon date. In our case it is DATE(2018,11,15). Settlement is
the settlement date, which for us is DATE(2018,07,25). Rate is the annual coupon rate,
that is, 0.10. Par is the actual par value which is 1,000. Frequency is 2 because the
bond pays semiannual coupons. Basis is 1 because it is a Treasury bond, for which the
Actual/Actual convention is applicable. Calc method enters the picture when the ac-
tual issue date is specified and not the previous coupon date. If calc method is given
as 0 (or omitted), the accrued interest is calculated from the previous coupon date to
the settlement date. However, if the calc method is given as 1, the cumulative coupon
is computed from the issue date till the settlement date. If we choose to enter the is-
sue date as the previous coupon date, as we have done, the calc method can either be
omitted, which is tantamount to a value of 0, or it can be entered as 1.

Accrued interest
= ACCRINT(DATE(2018,05,15), DATE(2018,11,15), DATE(2018,07,25), 0.10, 1000, 2, 1)
= $19.2935

The dirty price is 824.1705 + 19.2925 = $843.4640.

If we had given the issue date as 15 May, 2017 keeping other parameters the same,
and input a value of 1 for the calc method, we would have gotten a value of $119.2935,
which is essentially one year’s coupon (May 2017 to May 2018), plus the accrued inter-
est from 15 May 2018 until 25 July 2018.

Accrued interest
= ACCRINT(DATE(2017,05,15), DATE(2018,11,15), DATE(2018,07,25), 0.10, 1000, 2, 1)
= $119.2935
Computation of the YTM Between Coupon Dates | 73

Computation of the YTM Between Coupon Dates

As mentioned earlier, the RATE function can be used to compute the YTM when we are
on a coupon date. However, when we are between coupon dates, since a fractional first
period is involved, we need to use the YIELD function. Assume that we are standing on
25 August 2018 and have a bond which is maturing on 15 November 2035. The coupon
is 10% per annum paid semiannually. The day-count convention is Actual/Actual. The
quoted price is 98-12 which translates to $98.375 per $100 of face value.
The parameters are
– Settlement or the settlement date: In this case it is DATE(2018,08,25).
– Maturity or the maturity date: In this case it is DATE(2035,11,15).
– Rate or the annual coupon rate: In this case it is 0.10.
– Pr or the Clean Price per $100 of face value: In this case it is $98.375. If we have the
dirty price, we need to compute and subtract the accrued interest, before invoking
this function.
– Redemption or redemption value: This is invariably 100.
– Frequency: In this case it is 2 because the bond pays coupons semiannually.
– Basis or the day-count convention: In this case it is 1 because we are following an
Actual/Actual convention.

The YTM = YIELD(DATE(2018,08,25),DATE(2035,11,15),0.10,98.375,100,2,1) = 10.1984%

Because we are giving a frequency of 2 in the function, it automatically returns the
annual YTM. Thus there is no need to multiply the answer by 2.

Other Day-Count Conventions

The Actual/Actual day-count convention is one of the many conventions used in bond
markets. For products other than Treasury bonds, the convention may be different,
even in the U.S. We discuss other conventions in the following sections.

The 30/360 NASD Approach

30/360 NASD is the method used for corporate bonds in the U.S. In this method, the
denominator is taken to be 180. That is, the length between successive coupon dates is
always taken to be 180. Each month is therefore considered to consist of 30 days. The
numerator is then calculated by defining the start and end dates, D1 and D2 as follows:

D1 = (year 1 , month1 , day1 )

D2 = (year 2 , month2 , day2 )
74 | Chapter 3: Bonds: Advanced Concepts

The numerator is then calculated as follows:

360(year 2 − year 1 ) + 30(month2 − month1 ) + (day2 − day1 )

The following additional rules are involved:

1. If day1 = 31, then set day1 = 30.
2. If day1 is the last day of February, then set day1 = 30.
3. If day1 = 30 or has been set equal to 30, on the basis of Rule 1, then if day2 = 31,
set day2 = 30.

The following illustrations calculate the number of days, using the 30/360 NASD ap-
proach, for various start dates and end dates:
– Illustration 1
Fractional period from 15 August 2018 to 15 November 2018:
N1 = 360 × (2018 − 2018) + 30 × (11 − 08) + (15 − 15) = 90
– Illustration 2
Fractional period from 31 August 2018 to 15 November 2018:
N1 = 360 × (2018 − 2018) + 30 × (11 − 08) + (15 − 30) = 75
– Illustration 3
Fractional period from 31 August 2018 to 31 December 2018:
N1 = 360 × (2018 − 2018) + 30 × (12 − 08) + (30 − 30) = 120
– Illustration 4
Fractional period from 30 August 2018 to 30 December 2018:
N1 = 360 × (2018 − 2018) + 30 × (12 − 08) + (30 − 30) = 120
– Illustration 5
Fractional period from 30 August 2018 to 31 December 2018:
N1 = 360 × (2018 − 2018) + 30 × (12 − 08) + (30 − 30) = 120
– Illustration 6
Fractional period from 29 August 2018 to 30 December 2018:
N1 = 360 × (2018 − 2018) + 30 × (12 − 08) + (30 − 29) = 121
– Illustration 7
Fractional period from 29 August 2018 to 31 December 2018:
N1 = 360 × (2018 − 2018) + 30 × (12 − 08) + (31 − 29) = 122
– Illustration 8
Fractional period from 28 February 2018 to 29 July 2018:
N1 = 360 × (2018 − 2018) + 30 × (07 − 02) + (29 − 30) = 149
– Illustration 9
Fractional period from 28 February 2018 to 31 July 2018:
N1 = 360 × (2018 − 2018) + 30 × (07 − 02) + (31 − 30) = 151
Other Day-Count Conventions | 75

Valuation of a Corporate Bond Between Coupon Dates

Consider the case of a corporate bond that pays interest on 15 May and 15 November
every year. The coupon is 8% per annum, and the YTM is 10% per annum. Assume
that we are standing on 15 July of 2018 and the bond matures on 15 November 2037.
What is the dirty price?

120
k= = 0.6667
180
0.10 39
80 [(1 + 2
) − 1] 1,000
2
P=[ × ]+
0.10 0.6667 0.10 39−1 0.10 39−1+0.6667
[ (1 + 2
) ( 0.10
2
) (1 + 2
) ] (1 + 2
)
= $843.4344

30/360 European Convention

In the 30/360 European (E) convention, if the settlement date, day2 , is equal to 31, then
it is always set equal to 30. The additional rules may therefore be stated as follows:
1. If day1 = 31, set day1 = 30
2. If day2 = 31, set day2 = 30

The second rule effectively means that if day2 = 31, then set it equal to 30, no matter
what the value of day1 is. However, in the case of the 30/360 NASD rule, the condition
is that if day1 = 30, or has been set equal to 30, then if day2 = 31, set it equal to 30.

Example 3.2. Consider the fractional period from 29 March 2018 to 31 July 2018.
If we use the 30/360 NASD convention, we compute the numerator as

N1 = 360 × (2018 − 2018) + 30 × (07 − 03) + (31 − 29) = 122

However, as per the 30/360 E convention:

N1 = 360 × (2018 − 2018) + 30 × (07 − 03) + (30 − 29) = 121

Actual/365 Convention

The difference between the Actual/365 and the Actual/Actual methods is that the de-
nominator consists of 365 days even in leap years. For instance, consider a 10% coupon
paying bond that pays interest on 15 November and 15 May every year. Assume that we
are standing on 25 July 2018. The accrued interest can be calculated as follows:

71
AI = 1,000 × 0.10 × = 19.4521
365
76 | Chapter 3: Bonds: Advanced Concepts

Notice that while calculating the accrued interest we multiply M by c and not by c2 .
This is because the denominator of the day count fraction represents the number of
days in an entire year and not in a coupon period.
An Actual/365 bond that pays periodic interest usually accrues interest at a rate
such that the interest accrued over a full coupon period does not equal the periodic
coupon payment. Thus, for an Actual/365 issue, day count functions and interest ac-
cruals over a full coupon period need not generate the actual coupon payment for the
period. Example 3.3 is an illustration from Stigum and Robinson [59].

Example 3.3. Consider the period 15 November 1991 to 15 May 1992. The number of days is 182. The
day count fraction for the full coupon period is

182
= .499
365

and not .500. Because the semiannual coupon payment must equal exactly half the coupon, the inter-
est accrued on a given Actual/365 security over a full coupon period need not equal the exact coupon
payment at the end of the period. This anomalous result cannot occur for an Actual/Actual security.

Actual/360 Convention

Actual/360 is a simple variant of Actual/365. This is the interest payment convention

used for money market instruments in many countries, including the U.S.

Comparison of Day-Count Conventions

Consider a bond with a face value of $1,000 paying a coupon of 8% per annum on 25
July and 25 January every year. Assume we are on 31 August of a year. Calculate the
accrued interest using the following day-count conventions:
a) Actual/Actual
b) 30/360 NASD
c) 30/360 European
d) Actual/360
e) Actual/365

Actual/Actual:

37
AI = 40 × = 8.0435
184

ACCRINT(DATE(2018,07,25),DATE(2019,01,25),DATE(2018,08,31),0.08,1000,2,1) =
8.0435
Additional Coupon-Related Excel Functions | 77

30/360 NASD:

36
AI = 40 × = 8.0000
180

ACCRINT(DATE(2018,07,25),DATE(2019,01,25),DATE(2018,08,31),0.08,1000,2,0) =
8.0
30/360 E:

35
AI = 40 × = 7.7778
180

ACCRINT(DATE(2018,07,25),DATE(2019,01,25),DATE(2018,08,31),0.08,1000,2,4) =
7.7778
Actual/360:

37
AI = 80 × = 8.2222
360

ACCRINT(DATE(2018,07,25),DATE(2019,01,25),DATE(2018,08,31),0.08,1000,2,2) =
8.2222
Actual/365:

37
AI = 80 × = 8.1096
365

ACCRINT(DATE(2018,07,25),DATE(2019,01,25),DATE(2018,08,31),0.08,1000,2,3) =
8.1096

Additional Coupon-Related Excel Functions

The functions that we describe below compute various statistics pertaining to a
coupon period.
– COUPDAYBS: Returns the number of days between the previous coupon date and
the settlement date.
– COUPDAYS: Returns the number of days in the coupon period.
– COUPDAYSNC: Returns the number of days between the settlement date and the
next coupon date.
– COUPNUM: Returns the number of coupons between the settlement date and the
maturity date.

The common parameters for all these functions are:

– Settlement
– Maturity
– Frequency
– Basis
78 | Chapter 3: Bonds: Advanced Concepts

Example 3.4. Consider a bond paying semiannual coupons. The day-count convention is 30/360
NASD. The settlement date is 10 June 2018 and the maturity date is 15 August 2030.
– COUPDAYBS(DATE(2018,06,10),DATE(2030,08,15),2) = 115
360(2018 − 2018) + 30(06 − 02) + (10 − 15) = 115
This function computes the number of days between the previous coupon date and the settlement
date.
– COUPDAYS(DATE(2018,06,10),DATE(2030,08,15),2) = 180
360(2018 − 2018) + 30(08 − 02) + (15 − 15) = 180
This function computes the number of days in the coupon period.
– COUPDAYSNC(DATE(2018,06,10),DATE(2030,08,15),2) = 65
360(2018 − 2018) + 30(08 − 06) + (15 − 10) = 65
This function computes the number of days between the settlement date and the next coupon
date.
– COUPNUM(DATE(2018,06,10),DATE(2030,08,15),2) = 25
This function computes the number of coupon periods in a time interval. In this case, one coupon
on 15 August 2018 and two coupons every year for the next 12 years.

Valuing a Bond in the Final Coupon Period

A plain vanilla bond, with one coupon left, is effectively a zero coupon bond. Because
the time to maturity is less than six months, it is usually treated as a short-term debt
security or a money market security. Thus the cash flow is discounted using simple
interest.

Example 3.5. A bond has a face value of $1,000 and a coupon of 10% per annum payable semiannually.
Today is 25 August 2018 and the maturity date is 15 November 2018. The quoted price is 99-12. The
day-count convention is Actual/Actual.
The number of days until maturity is 82, and the number of days in the coupon period is 184. Thus
the fractional period remaining is 0.4457. The accrued interest is 50 × (1 − 0.4457) = $27.715. Thus the
dirty price is:

99.375 × 10 + 27.715 = $1,021.465

Using the dirty price, we compute the YTM as follows:

1,050
1,021.465 =
(1 + i × 0.4457)
⇒ i = 0.0627 ≡ 6.27%

The annual YTM is 12.54%.

Had we used the compound interest principle for discounting, we would have gotten a value of
12.75%. This is because for a given yield, the discount factor is greater when the simple interest prin-
ciple is adopted. Thus the simple interest approach would give us a lower present value for a given
interest rate. Hence, if the present value is to be the same, the yield should be lower in the case of
simple interest.
Yield Measures: An Introduction | 79

Yield Measures: An Introduction

Once we know the market’s required rate of return from a bond, we can calculate its
price by calculating the present values of the cash flows generated by it. On the other
hand, if we are given information about the price of a bond, we can find the rate of
return that equates the present value of the cash flows to the price. This is nothing but
the internal rate of return (IRR) of the bond. The IRR of a bond is termed as its yield to
maturity (YTM) or its redemption yield. For a bond that pays N coupons, the YTM is a
solution to a polynomial of degree N. The easiest way to compute the YTM of a bond,
on a coupon date, is by using the RATE function in Excel. Alternately, one can use the
approximate yield to maturity demonstrated later on in this chapter as a starting point
and obtain a more precise value using linear interpolation. If we are between coupon
dates, we need to use the YIELD function in Excel to compute the YTM.
In practice an investor has a choice between a wide variety of bonds with differ-
ent coupons and terms to maturity. Besides, the creditworthiness of the issuer varies
from bond to bond. Consequently, although data about bonds may be provided in the
form of prices, it is the yield measure that facilitates comparison between different in-
struments. The yield to maturity is one of the key measures of the yield from a bond.
However, in practice other measures of yield are computed, as well. Let’s now examine
the various yield measures that are used in the market.

Current Yield of a Bond

The current yield measure is perhaps the most unsatisfactory measure. But, it is re-
ported in practice, and you should be acquainted with it. It is also called the flat yield,
interest yield, income yield, or running yield.
The current yield relates the annual coupon payment to the current market price:

Annual Coupon Interest

CY = (3.5)
Price

One of the issues is, whether the price should be the clean price or the dirty price.
The advantage of using the clean price is that the current yield stays constant unless
the yield changes. If the dirty price is used then the current yield will be higher in
the period between the ex-dividend date and the coupon date, because the dirty price
will be lower than the clean price, due to the negative accrued interest phenomenon.
However, the current yield will be lower in the interval between the coupon date and
the ex-dividend date, because the dirty price will be higher than the clean price.4 This
gives rise to a sawtooth pattern. The current yield is more than the coupon rate if the

4 This issue is irrelevant in markets where the bond trades cum-dividend till the coupon date.
80 | Chapter 3: Bonds: Advanced Concepts

bond is trading at a discount, and less than the coupon rate if it is trading at a pre-
mium.
The current yield suffers from two major technical deficiencies. First, it ignores
any capital gain or loss experienced by the bond holder. Second, it fails to consider
the time value of money. Nevertheless, it is used to estimate the cost of, or profit from,
holding a bond. The difference between the current yield and the cost of funding the
bond is known as the net carry. If the short-term funding rate in the market is higher
than the current yield, the bond is said to involve a running cost. This is also known
as negative carry or negative funding.
The current yield computes the interest yield for an investor with a one-year hori-
zon. However, an investor with a one-year horizon experiences a capital gain or loss,
when he sells the bond after one year. Consequently the current yield measure, does
not fully capture all the facets of the return, even for such an investor.

Example 3.6. A bond which pays a coupon of 8% per annum is currently trading at $95. A bond holder
buys the bond by borrowing at the rate of 8.25% per annum. What is the current yield and what is the
net carry?

8
CY = = 0.0842 ≡ 8.42%
95

The cost of funding is 8.25%. So the net carry or return is 0.17%.

Simple Yield to Maturity

The simple yield to maturity measure attempts to rectify the shortcomings of the cur-
rent yield by taking into account capital gains and losses. The assumption made is
that capital gains and losses accrue evenly over the life of the bond, or in other words
on a straight line basis.
The formula is:
C M−P
Simple YTM of a Bond = + (3.6)
P N ×P
2

The problem with the simple yield to maturity is that it does not account for the fact
that an investor in a bond can earn compound interest. As coupons are paid, they
can be reinvested and hence can earn interest. This increases the overall return from
holding the bond. Besides the assumption that capital gains and losses arise evenly
over the life of the bond is an over simplification.
The simple yield to maturity (SYTM) is also known as the Japanese yield, for it is
the main yield measure that is used in the Japanese Government Bond (JGB) market.
If a bond is selling at a discount, the SYTM is higher than the current yield. However,
if it is trading at a premium, then the SYTM is lower than the current yield.
The Approximate Yield to Maturity Approach | 81

Example 3.7. Assume that a bond with 10 years to maturity and a coupon of 8% per annum is trading
at $95. The face value is $100.

8 (100 − 95)
SYTM = +
95 10 × 95
= 0.0842 + 0.0053 = 0.0895 ≡ 8.95%

Yield to Maturity of a Bond

The YTM is the interest rate that equates the present value of the cash flows from the
bond(assuming that the bond is held to maturity) to the price of the bond. Consider
a bond that makes an annual coupon payment of C on a semiannual basis. The face
value is M, the price is P, and the number of coupons remaining is N. The YTM is the
value of y, that satisfies the following equation:
N C
2 M
P = ∑[ y t
]+ N
(3.7)
t=1 (1 + 2
) (1 + y2 )

The value of y that we compute in this fashion is the nominal annual yield, which is
also called the bond equivalent yield (BEY). From the principles of time value of money,
2
you can understand that the effective annual yield is (1 + y2 ) − 1.

The Approximate Yield to Maturity Approach

The bond valuation equation is a non-linear equation. The issue is how to calculate
the YTM given other variables. One way, of course, is to use the IRR function, or the
RATE function in Excel. We will now demonstrate an alternative procedure, using the
approximate yield to maturity or AYM.
M−P
C+ N/2
AYM = M+P
(3.8)
2

The Rationale for the AYM Approach

The annual coupon interest income is C. The capital gain/loss if the bond is held to
maturity (N/2 years) on a straight line basis, is M−P N/2
. If the bond is a discount bond,
there is a capital gain if it is held to maturity otherwise, there is a capital loss. Thus, the
annual income is C + M−P N/2
. The initial investment is P. An instant before redemption,
the money that is locked up in the bond is M. Thus the average investment is M+P
2
.
The approximate YTM is the annual income divided by the average investment and is
82 | Chapter 3: Bonds: Advanced Concepts

equal to
M−P
C+ N/2
M+P
2

The AYM gives us a starting point. We choose a higher value and a lower value,
such that the higher yield gives a price that is lower than the actual price, and the
lower yield gives a price that is higher than the actual price. Using these prices, we
can compute the yield to maturity using linear interpolation.

Example 3.8. Consider a 10% coupon bond with a face value of $1,000, a price of $860, and 10 years
to maturity. Assume that the bond pays interest on a semiannual basis. What is the yield to maturity?
1,000−860
100 + 10
AYM = 1,000+860
2
100 + 14
= ≡ 12.2581%
930
Consider two rates, 12% and 13%
Price at 12% = $885.3008
This price is above $860
Price at 13% = $834.7224
This price is below $860.
Now let’s interpolate. 12%–13% corresponds to a price difference of 885.3008 − 834.7224. Thus
12% − y ∗ should correspond to a price difference of 885.3008 − 860, where y ∗ is the true YTM.
Take the ratio of the two, and solve for y ∗ :
−0.01 50.5784
=
0.12 − y ∗ 25.3008
25.3008
⇒ y ∗ = 0.12 + 0.01 × = 12.5002%
50.5784
We can verify the accuracy of our approximation using Excel. PV(0.125002/2, 20, −50, −1000) =
$859.48. Thus the approximation is excellent.

As we have discussed, we generally require a computer program to calculate the YTM.

A less precise, but very often an effective method, is the use of the AYM with interpola-
tion. The calculation is fairly simple for a coupon bond with two periods left to matu-
rity, and for zero coupon bonds. The former has only two cash flows, which implies a
quadratic equation for computing the YTM, whereas the latter has a single cash flow.

Example 3.9. A Reliance bond with a face value of $1,000 and a coupon rate of 10% per annum,
payable semiannually, has one year left to maturity. It is currently selling at $900. What is the YTM?
50 50 1,000
900 = + +
(1 + 2y ) (1 + y )2 (1 + y )2
2 2
50 1,050
≡ +
(1 + i) (1 + i)2
y
where, we have denoted 2
by i.
The Approximate Yield to Maturity Approach | 83

Therefore

900(1 + i)2 = 50(1 + i) + 1,050

⇒ 900(1 + i)2 − 50(1 + i) − 1,050 = 0

This is a quadratic equation of the form

ax 2 + bx + c = 0,

where x in this case is (1 + i).

The equation has two roots:

−b ± √(b2 − 4ac)
x=
2a
Therefore

50 ± √[(−50)2 − 4 × (900) × (−1,050)]

(1 + i) =
1,800
50 ± 1,944.87
= = 1.1083 or − 1.0527
1,800
Therefore
y
i= = 0.1083 or − 2.0527
2
We discard the negative root since the yield is inevitably positive.
Thus, the nominal annual YTM = 0.1083 × 2 = 0.2166 ≡ 21.66 %

Example 3.10. Consider a zero coupon bond with a face value of $1,000, maturing five years from now.
The current price is $500. What is the YTM?
The first question is, whether we use
1,000
500 = ,
(1 + y)5
or
1,000
500 = 10
(1 + 2y )

The second approach is preferred, as mentioned earlier, because we may like to compare our results
with coupon paying bonds, which typically pay interest on a semi annual basis.
Therefore,
y 10 1,000
(1 + ) = =2
2 500
That is,
y
= (2)0.1 − 1 = 1.0718 − 1 = 0.0718
2
Thus y = 0.1436 ≡ 14.36%
84 | Chapter 3: Bonds: Advanced Concepts

The YTM calculation takes into account the coupon payments, as well as any capital
gains/losses that accrue to an investor who buys and holds a bond to maturity. Before
we analyze the YTM in detail, let’s consider the various sources that contribute to the
returns received by a bond holder.5
A bond holder can expect to receive income from the following sources:
– Coupon payments, which are typically paid every six months.
– A capital gain/loss that is obtained when the bond matures, is called before ma-
turity, or is sold before maturity. From the YTM equation, you can see that we are
assuming that the bond is held to maturity.
– Reinvestment of coupon payments, from the time each coupon is paid until the
time the bond matures, is sold, or is called. Once again, the YTM calculation as-
sumes that the bond is held to maturity. The reinvestment income is the interest
on interest income.

A satisfactory measure of the yield should take into account all three sources of in-
come. The YTM does take all three sources of income into account. However, it makes
two key assumptions:
– The bond is held until maturity.
– The intermediate coupon payments are reinvested at the YTM itself.

The second assumption is in built into the mathematics of the YTM calculation, as we
shall shortly demonstrate.
The YTM is called a promised yield. Why do we use the word promised? It is a
promise because, in order to realize it, the bond holder has to satisfy the preceding
two conditions. If either of them is violated, the bond holder may not get the promised
yield.6
The assumption that the bond is held to maturity is fairly easy to comprehend.
Let’s now focus on the reinvestment assumption.
Consider a bond with a face value of M, annual coupon of C, and number of
coupons left N. Let the bond pay interest on a semiannual basis, i.e. it pays C/2 ev-
ery six months. Let r be the rate at which one can reinvest the coupon payments until
maturity. r would depend on the prevailing rate of interest when the coupon is re-
ceived, and need not be equal to y, the initial YTM, or c, the coupon rate. For ease of
exposition, we can assume that r is a constant for the life of the bond. In practice, the
rate is likely to vary from period to period.
Thus each coupon payment is reinvested at 2r (for six monthly periods). The
coupon stream is obviously an annuity. The final payoff from reinvesting the coupons
is therefore given by the future value of this annuity, using a rate of 2r for every six
monthly period.

5 See Fabozzi [26].

6 Remember, the IRR calculation also assumes that intermediate cash flows are reinvested at the IRR
itself.
The Approximate Yield to Maturity Approach | 85

The future value is

C
2 r N
r [(1 + ) − 1]
2
2

Note that this amount represents the sum of all coupons that are reinvested (which in
this case is the principal), plus interest earned on reinvesting the coupons.
The total value of the coupons = C2 × N = NC2
Thus, interest on interest is equal to
C
2 r N NC
r [(1 + ) − 1] −
2
2 2

The YTM calculation assumes that 2r = y2

In Example 3.11 we demonstrate that an investor gets a rate of return equal to the
YTM, only if he manages to reinvest all the coupons at the YTM.

Example 3.11. Consider an ABC bond that has 10 years to maturity. The face value is $1,000. It pays a
semiannual coupon at the rate of 10% per annum. The YTM is 12% per annum.
Let’s first calculate the price

P = 50 × PVIFA(6, 20) + 1,000 × PVIF(6, 20)

= 50 × 11.4699 + 1,000 × .3118 = 885.295

We assume that the semiannual interest payments can be reinvested at a six-month rate of 6%,
which corresponds to a nominal annual rate of 12%.
Let’s analyze the sources of income for a bond holder, assuming that he holds the bond to matu-
rity.
1. Total coupon received = 50 × 20 = $1,000.
2. Interest on interest got by reinvesting the coupons:

50[(1.06)20 − 1]
= − 1,000
0.06
= 50 × 36.786 − 1,000 = $839.3

Notice that, if we do not deduct 1,000 (or NC2

) in the preceding equation, we can calculate the
income from both sources together. We have separated them for expositional clarity.
3. Finally, in the end, the bond holder gets back the face value of $1,000.
Thus, the terminal cash flow for the bond holder is 1,000 + 839.3 + 1,000 = $2,839.3.
To get this income the bond holder has to pay $885.295 at the outset, which is the investment.
So what is the rate of return? It is the value of i that satisfies the following equation:

885.295(1 + i)20 = 2,839.3

2,839.3 0.05
⇒ (1 + i) = [ ] = 1.0600
885.295
⇒ i = 0.06 ≡ 6% on a semiannual basis, or 12% on a nominal annual basis, which is exactly the
same as the YTM.
86 | Chapter 3: Bonds: Advanced Concepts

So, how did the bond holder realize the YTM? Only by being able to reinvest the
coupons at a nominal annual rate of 12%, compounded on a semiannual basis. Notice
that the reinvestment rate affects only the interest on interest income. The other two
sources are unaffected. If r > y, the interest on interest would have been higher, and
i would have been greater than y. On the contrary, if r < y, the interest on interest
would have been lower, and i would have been less than y.
So, if an investor buys a bond by paying a price that corresponds to a given YTM,
he realizes that YTM only if both the following are true:
1. He holds the bond to maturity.
2. He is able to reinvest the coupons at the YTM.

In the preceding scenario, the risk that the investor faces is that future reinvestment
rates may be less than the YTM that was in effect at the time he purchased the bond.
This risk is called reinvestment risk. The degree of reinvestment risk depends critically
on two factors, namely, the time to maturity, and the quantum of the coupon. For a
bond with a given YTM and coupon rate, the greater the time to maturity, the more de-
pendent is the bond’s total return on the reinvestment income. Thus, everything else
remaining constant, the longer the time to maturity, the greater is the reinvestment
risk.7 Secondly, for a bond with a given maturity and YTM, the greater the quantum of
the coupon, or in other words, the higher the coupon rate, the more dependent is the
bond’s total return on income from reinvestment. Hence, everything else remaining
constant, the larger the coupon rate is, the greater the reinvestment risk. For bonds
selling at a premium, i.e. c > y, vulnerability to reinvestment risk is higher than that
for a bond selling at par. Correspondingly, discount bonds are less vulnerable than
bonds selling at par.
The important thing to note is that for a zero coupon bond, if it is held to matu-
rity, there is absolutely no reinvestment risk because there are no intermediate coupon
payments. Hence, if a ZCB is held to maturity, the yield actually earned equals the
promised YTM. Thus, in order to earn the YTM of a zero coupon bond, only one con-
dition needs to be satisfied. The bond has to be held to maturity. This explains the
popularity of such securities among long-term investors such as pension funds and
insurance companies. If they buy and hold the security to maturity, they are assured
of earning the YTM without having to take any additional steps.

Reinvestment Assumption Behind the YTM Calculation

We can claim that we have received a YTM of y% per annum, if the compounded semi-
annual return on our initial investment is y2 .

7 See Fabozzi [26].

The Realized Compound Yield | 87

The initial investment is

C
2 1 M
P= y [1 − ]+
y N N
2 (1 + 2
) (1 + y2 )

The compounded value of the investment is

y N
P × (1 + )
2
C
2 y N
= y [(1 + ) − 1] + M
2
2

The terminal cash flow from holding the bond, assuming that each coupon is rein-
vested at 2r per semiannual period, is

C
2 r N
= r [(1 + ) − 1] + M
2
2

Equating the two, we get

r y
=
2 2

Thus in order to get an annual YTM of y%, every intermediate cash flow must be rein-
vested at y2 % per six-month period.

The Realized Compound Yield

To calculate the realized compound yield (RCY), we once again assume that the bond
is held until maturity. But we make an explicit assumption about the rate at which the
coupons can be reinvested. That is, unlike in the case of the YTM, we no longer take it
for granted that intermediate cash flows can be reinvested at the YTM.

Example 3.12. Consider the ABC bond. Assume that intermediate coupons can be reinvested at 7% for
six months, or at a nominal annual rate of 14%.
Compared to the earlier example, the coupon income and the final face value payment remain the
same, but, the reinvestment income changes.

50
Interest on Interest = [(1.07)20 − 1] − 1,000
0.07
= 50 × 40.995 − 1,000 = 1,049.75

So, the final amount received = 1,000 + 1,049.75 + 1,000 = $3,049.75.

The initial investment is $885.295.
88 | Chapter 3: Bonds: Advanced Concepts

Therefore, the rate of return is given by,

885.295(1 + i)20 = 3,049.75

3049.75 0.05
⇒ (1 + i) = [ ] = 1.0638
885.295
⇒ i = 6.38%

This is the return for six months. The nominal annual return is 6.38×2 = 12.76%. 12.76% is greater
than the YTM of 12%.

Thus, the realized compound yield is greater than the YTM if the reinvestment rate is
greater than the YTM. If the reinvestment rate were less than the YTM, the RCY would
be less than the YTM.
The RCY can be an ex-ante measure if one makes an assumption about the rein-
vestment rate. It can also be an ex-post measure, if one uses the rate at which the
investor is actually able to reinvest.

The Horizon or Holding Period Return

Now we can relax both the assumptions that we made to calculate the YTM. First, the
investor may not hold the bond to maturity, and, second, may not be able to reinvest
the coupons at the YTM.
Let’s suppose that the investor has an investment horizon that is less than the time
to maturity. The return depends on three sources, namely, the coupons received, the
reinvestment income, and the price at which the investor expects to sell the bond.8

Example 3.13. An investor has a seven year investment horizon, and wants to buy the ABC bond dis-
cussed in Example 3.12. The current price is $885.295. He gets coupons for seven years or 14 periods.
The total coupon income = 50 × 14 = $700.
The investor believes that coupons can be reinvested at 7% per six-month period. He also believes
that when he is ready to sell the bond, the YTM will be 12% per annum on a nominal annual basis. The
first step is to calculate the expected price at the time of sale. Remember that the bond will have three
years to maturity.

P(after 7 years) = 50PVIFA(6,6) + 1,000PVIF(6,6)

= 50 × 4.9173 + 1,000 × .705 = $950.83
50[(1.07)14 − 1]
Interest on Interest = − 700
.07
= 50 × 22.55 − 700 = 427.5

The terminal value = 700 + 427.5 + 950.865 = $2,078.35.

8 The selling price may not be equal to the face value and depends on the YTM at the time of sale.
The Realized Compound Yield with Taxes | 89

The rate of return is given by

885.295(1 + i)14 = 2,078.35

1
2078.35 14
⇒ (1 + i) = [ ] = 1.0629
885.295
⇒ i = 6.29%

The nominal annual return = 6.29 × 2 = 12.58%.

This is the horizon yield. Again, can be calculated on either an ex-ante or ex-post basis.

The Realized Compound Yield with Taxes

In most countries, both the coupon income from a bond, and the capital gain, if any, is
taxed.9 In practice the rate of tax levied on capital gains may or may not be the same as
the rate applicable for income from coupons. The incidence of tax on coupon income
has certain negative consequences for the bond holder. First, the amount of cash re-
ceived at the end of every coupon period is reduced. For instance, if the coupon due
is $50, and the tax rate is 30%, the post-tax coupon is only $35. Thus the total coupon
income over the life of the bond is reduced. If we assume that the bond under consid-
eration has 20 periods until maturity, the post-tax coupon will be $700, as compared
to the $1,000 that would have been received in the absence of the tax. The other con-
sequence of the tax on coupons is that it brings down the reinvestment income. This
is because, as a consequence of the tax, there is less to reinvest every period. Finally,
the interest earned by reinvesting coupon interest, also comes down because of the
tax rate.
Consider a bond with a face value of $M, and N coupons remaining. The semi-
annual coupon rate is c, and the semiannual reinvestment rate is r. In the absence of
taxes, the terminal value of reinvested coupons would be

c/2 r N
M× [(1 + ) − 1]
r/2 2

In the presence of taxes, the periodic coupon rate is c(1−T)

2
. The effective semiannual
r(1−T)
reinvestment rate is 2 . Thus the post-tax value of re-invested coupons is

c(1 − T)/2 r {1 − T} N
M× [(1 + ) − 1]
r(1 − T)/2 2
c r {1 − T} N
=M× [(1 + ) − 1]
r 2

9 See Smith [56].

90 | Chapter 3: Bonds: Advanced Concepts

The post-tax terminal cash flow, when the face value is repaid, is

M − (M − P) × Capital Gains Tax

Example 3.14 illustrates this.

Example 3.14. Consider the bond with a coupon of 10% per annum and a face value of $1,000, ma-
turing after 10 years. The price is $885.30, and we assume that coupons, paid semiannually, can be
reinvested at 8% per annum. Coupon income is taxed at 30%, and capital gains at 25%. We assume
that the bond is held to maturity.
The cash flow from the coupons as computed at the time of expiration is:

0.05
1,000 × [(1 + .04 × 0.7)20 − 1]
0.04
= $921.5625

The capital gain is 1,000 − 885.30 = $114.70. The post tax cash flow from the face value at the end is
1,000−114.70×0.25 = $971.325. Thus an investment of 885.30 yields an inflow of 921.5625+971.325 =
$1,892.8875 after 20 periods. The realized compound yield is 7.7455% on a per annum basis.

Computing the YTM with Taxes

Let’s revisit Example 3.14. Assume that the price of the bond is $975.50. What is the
post-tax YTM? The post-tax periodic cash flow is 50 × 0.7 = $35. The post-tax cash flow
from the face value is 1,000 − (1,000 − 975.50) × 0.25 = $993.875. We now invoke the
RATE function in Excel.
RATE(20, −35,975.50, −993.875) = 3.6535% ≡ 7.3069% per annum.

The Portfolio Yield for Bonds

Suppose you hold a collection of bonds. A simple way to calculate the yield is as a
weighted average of the yields of the individual bonds in the portfolio. But from a
technically more accurate stand-point, you should first compute the cash flows for the
portfolio, and then find the interest rate that makes the present value of the cash flows
equal to the sum of the prices of the components of the portfolio. This is nothing but
the computation of the portfolio IRR. We illustrate both the techniques in the following
sections.
Peter Burns buys a Raccoon bond and a YKK bond. The Raccoon bond has a time
to maturity of five years, a face value of $1,000, and pays coupons semiannually at the
rate of 10% per annum. The YTM is 12% per annum. The YKK bond has a face value
of $1,000, time to maturity of four years, and pays a coupon of 10% per annum, every
six months. The YTM is 16% per annum. If he buys one of each bond, what is the yield
on his portfolio?
The Portfolio Yield for Bonds | 91

The first step is to calculate the two prices

P(Raccoon) = 50PVIFA(6, 10) + 1,000PVIF(6, 10)

= 50 × 7.3601 + 1,000 × .5584 = 926.405
P(YKK) = 50PVIFA(8, 8) + 1,000PVIF(8, 8)
= 50 × 5.7466 + 1,000 × .5403 = 827.63

The initial investment = 926.405 + 827.63 = $1,754.035

The Weighted Average Approach

The weighted average portfolio yield, is an average of the yields of the component
bonds, where the prices of the bonds serve as the weights.

926.405 827.63
yp = 0.12 × + 0.16 ×
1,754.035 1,754.035
= 0.0634 + 0.0755 = 0.1389 ≡ 13.89%

The IRR Approach

Let’s set up the cash flow table, as shown in Table 3.4. The first seven cash flows are
$100 because the two bonds pay a coupon of $50 each. In the eighth period, one bond
pays a coupon of $50, whereas the other, which is maturing, pays $1,050. In the ninth
period, the remaining bond pays a coupon of $50, and in the tenth period, it too ma-
tures after paying out $1,050.
Calculate the IRR using Excel. The answer comes out to be 13.76% on a nominal
annual basis.

Table 3.4: Cash flows from the bond portfolio.

Period Cash Flow Cash Flow Total Cash

from Raccoon from YKK Flow

0 (926.405) (827.63) (1,754.035)

1 50 50 100
2 50 50 100
3 50 50 100
4 50 50 100
5 50 50 100
6 50 50 100
7 50 50 100
8 50 1050 1,100
9 50 50
10 1050 1,050
92 | Chapter 3: Bonds: Advanced Concepts

The Taxable Equivalent Yield (TEY)

Certain bonds, like municipal bonds, are tax exempt. In the U.S., unlike in many other
countries, both the federal and state governments can levy income tax. Often, an in-
vestor would like to compare the yield of a bond whose income is taxable, with the
yield of another bond whose income is tax free. For this purpose it becomes essential
to compute the taxable equivalent yield, or TEY, of the tax-free bond.
Governments in the U.S. follow the guideline of mutual reciprocity. That is, fed-
eral government bonds are exempt from state income taxes, and state and local gov-
ernment bonds are exempt from federal income tax. There are certain securities that
are exempt from all taxes or in other words, are totally tax-free.
The method of computation of the TEY depends on the applicable tax rate. Con-
sider a municipal bond yielding 6% per annum. Assume that it is exempt from fed-
eral income tax, but is subject to state tax. On the other hand, there is a bond whose
income is subject to both forms of tax. Assume that the federal tax rate is 25%. The
taxable equivalent yield of the bond is given by

6.00
TEY = = 8%
(1 − 0.25)

There is no need to adjust for state income tax because it is applicable for both bonds.
Thus if a taxable bond were to yield more than 8% per annum, if would be preferred
to the municipal bond. Otherwise the municipal bond would be more attractive.
Sometimes a bond may be totally tax free, that is, it is exempt from federal as well
as state taxes. In this case, we compute the TEY as follows. First we need to calculate
the effective tax rate. This, contrary to what the reader may think, is not the sum of the
two rates because the state tax that is paid is a deductible expense when computing
the federal tax.
Assume the federal tax rate is 25% and the state tax rate is 8%. On $100 income,
$8 are payable by way of state tax. The net income, $92, attracts federal tax at the rate
of 25%, which amounts to $23. Thus the post-tax income is $69, and the effective tax
rate is 31%. Consider a municipal bond with a YTM of 6.0375%. The TEY is

6.0375
= 8.75%
(1 − 0.31)

Sinking Fund Provisions

At times, a bond issue may carry a sinking fund provision in its indenture, which re-
quires the issuer to set aside money periodically in order to redeem a percentage of the
bonds at regular intervals. In practice the issuer sets up an account with a custodian,
into which the periodic payments are directed. This account is meant to be used for
Serial Bonds | 93

the sole purpose of retiring the debt. The periodic payments made by the issuer may
be fixed or variable, depending on the terms of the bond indenture.
Sinking funds have several positive features from the standpoint of a bond in-
vestor. First, because bonds are being periodically redeemed, there is less risk of de-
fault for bond holders compared to a situation where the entire issue has to be re-
deemed on a single future date. Second, the redemption in the case of sinking funds
takes place at the sinking fund call price. If yields have gone up after issue, and the
current market price is lower, then this offers a benefit for holders whose bonds are be-
ing redeemed. Third, the act of periodic purchase of bonds, leads to demand, which
has the effect of increasing the liquidity. In an environment where yields are increas-
ing and prices are falling, the buying pressure due to the sinking fund provision may
help prop up the bond price.
However, there are drawbacks to a sinking fund. In a market with declining yields,
a bond holder may not benefit because his bond may be redeemed at the price speci-
fied at the outset. Related to this is the risk that cash flows in a declining rate environ-
ment may have to be reinvested at lower rates of interest. In the case of a plain vanilla
bond, the risk for the holder is that coupons may have to be reinvested at lower rates.
However, in the case of a bond with a sinking fund provision, the risk is that the prin-
cipal itself may be returned prematurely, and consequently may have to be reinvested
at a lower rate.
Despite the pros and cons, bonds with a sinking fund provision usually have a
lower YTM than bonds that do not have this feature, but are otherwise similar. This is
first because there is lower default risk and second because there is downside protec-
tion in a falling price environment.

Serial Bonds
In the case of plain vanilla bonds, all the issued bonds mature on the same future
date. However, in the case of a serial bond issue, a fraction of the outstanding bonds
mature at regular intervals until the entire issue is redeemed. For an investor, such
bonds pose less risk of default as compared to a plain vanilla bullet bond, where it is
necessary to ensure that the issuer has adequate funds on the pre-specified maturity
date to be able to redeem the entire issue. Serial bonds, as well as bonds with sinking
funds, are considered less risky than plain vanilla bonds issued by the same issuer.
In the case, of serial bonds and bonds with a sinking fund, a portion of the issue
is redeemed at periodic intervals, but the mechanics are different. With sinking funds,
the issuer makes periodic payments into a custodial account. The trustee uses these
funds to buy bonds in the secondary market and retire them. The trustee either buys
bonds from whoever is willing to sell in the secondary market, or else selects randomly
drawn serial numbers to retire the corresponding bonds. In the case of a serial bond,
however, it is specified right at the outset when a particular bond will be retired.
94 | Chapter 3: Bonds: Advanced Concepts

Yield to Average Life

The yield to average life measure is applicable for bonds with a sinking fund provision,
where a percentage of outstanding bonds is redeemed every year at a specified price.10
Assume that a company has issued bonds with a face value of $100 million and four
years to maturity. Coupon is paid at the rate of 10% per annum, on an annual basis.
At the end of the first year, 20% of the bonds is redeemed; at the end of the second
year, another 25% is redeemed; and finally at the end of three years, 30% of the issue
is redeemed. The residual 25% is redeemed at maturity. The redemption price varies
from year to year as depicted in Table 3.5.

Table 3.5: Redemption schedule for a bond with a sinking fund.

Year Fraction Redemption

Redeemed Price
1 0.20 102.50
2 0.25 101.75
3 0.30 101.00
4 0.25 100.00

The average life is defined as:

4 4
∑(ti wi Pi ) ÷ ∑(wi Pi )
i=1 i=1

The numerator is:

1 × 0.20 × 102.50 + 2 × 0.25 × 101.75 + 3 × 0.30 × 101 + 4 × 0.25 × 100 = 262.275

The denominator is:

0.20 × 102.50 + 0.25 × 101.75 + 0.30 × 101 + 0.25 × 100 = 101.2375

Thus the average life is = 262.275 ÷ 101.2375 = 2.5907 years.

The yield to average life is based on three assumptions:
1. The bond matures at the time corresponding to the average life.
2. The cash flow at the time of maturity is the average redemption price plus coupon
for one period.

10 See Blake [4].

Yield to Equivalent Life | 95

3. The bond pays coupons at regular intervals prior to the time corresponding to the
average life.

N
C M+C
P=∑ + (3.9)
(1 + y) i
(1 + y)N
∗
i=1

In this equation, M is the average redemption price; N ∗ is the time corresponding to

the average life; N is the number of periodic coupons received before the time corre-
sponding to the average life; and y is the yield to average life. In our example: M =
101.2375; N ∗ = 2.5907; N = 2; and C = 10. If we assume that the bond price is 99-12,
the yield to average life is 12.2162%, as computed using Excel.

Yield to Equivalent Life

The mechanics of computing the yield to equivalent life are similar to the method
adopted for the computation of the yield to average life, except that we consider the
present value of the redemption price and not the price itself. This is done to account
for the fact that redemption prices are received at different points in time.

4 4
∑[ti wi PV(Pi )] ÷ ∑[wi PV(Pi )]
i=1 i=1

It must be noted that the present values are computed using the yield to equivalent
life as the discount rate. Hence, unlike the case of yield to average life, we need to
compute the yield to equivalent life before we can calculate the equivalent life.
The yield to equivalent life is defined as:

N
C + wi Pi
P=∑ (3.10)
i=1 (1 + y)i

where N is the number of remaining coupon periods; wi is the redemption fraction at

time i; Pi is the redemption price corresponding to time i; and y is the yield to equiva-
lent life.
In our illustration: N = 4; C = 10; P = 99.375. The yield to equivalent life using
Excel is 15.2176%. The equivalent life is

[0.2 × 1 × 88.96 + 0.25 × 2 × 76.65 + 0.30 × 3 × 66.03 + 0.25 × 4 × 56.74]

÷ [0.20 × 88.96 + 0.25 × 76.65 + 0.30 × 66.03 + 0.25 × 56.74]
= 172.2909 ÷ 70.9505 = 2.4283

years.
96 | Chapter 3: Bonds: Advanced Concepts

Chapter Summary
In this chapter, we first relaxed the assumption, that the bond is being valued on a
coupon date. In order to measure fractional time periods, we introduced the concept of
day-count conventions. We studied the concepts of clean and dirty prices, and the re-
lated issue of accrued interest, and explained why quoted bond prices are clean prices.
We illustrated the computation of clean prices, accrued interest, and the yield to ma-
turity, using Excel, and compared the results of different day-count conventions for
a given set of data. We then went on to look at yield measures. First we studied the
current yield and the simple yield to maturity, and then went on to look at the more
technically precise yield to maturity. We examined the assumptions behind the yield
to maturity and gradually relaxed them, to compute the realized compound yield and
the horizon yield. Coupons and capital gains from bonds may be taxable. We studied
the implications of taxes for the realized compound yield, as well as the YTM. We then
looked at the technique for computing a precise yield for a portfolio of bonds. In real
life, an investor may have a choice between taxable and tax-free bonds. To make a
comparison, we need to compute the taxable equivalent yield of a tax-free bond, and
we looked at related issues. Finally we introduced the concept of a sinking fund pro-
vision and looked at two related yield measures, namely the yield to average life and
the yield to equivalent life.
In the next chapter, we study the various theories of the term structure of interest
rates, and pay close attention to the concepts of the spot rate and the forward rate.
Chapter 4
Yield Curves and the Term Structure
At any point in time, while taking a decision to invest in debt securities, an investor
typically has access to a large number of bonds with different yields and varying times
to maturity. Consequently, it is a common practice for investors and traders to exam-
ine the relationship between the yields on bonds belonging to a particular risk class.
A plot of the yields of bonds that differ only with respect to their time to maturity,
versus their respective times to maturity, is called a yield curve. The curve is an impor-
tant indicator of the state of the bond market, and provides valuable information, to
traders and analysts.
While constructing the yield curve, it is very important to ensure that the data
pertain to bonds of the same risk class, having comparable degrees of liquidity. For
example, a curve may be constructed for government securities or AAA-rated corpo-
rate bonds, but not for a mixture of both. The primary yield curve in any domestic
capital market is the government bond yield curve, for these instruments are free of
default risk. In the U.S. debt market for instance, the primary yield curve is the U.S.
Treasury yield curve.

Analyzing the Yield Curve

The yield curve is an indicator of where the bond market is trading currently. It also
has implications for what the market thinks will happen in the future.
The yield curve sets the yield for all debt market instruments. First it fixes the price
of money over the maturity structure. In practice, the yields of government bonds from
the shortest maturity instrument to the longest set the benchmark for yields on all
other debt instruments. This means that if a five-year government security is trading
at a yield of 5%, all other five-year bonds, irrespective of the issuer, will be trading at
yields over 5%. The excess over the yield on the corresponding government security
is called the credit spread.
The yield curve also acts as an indicator of future yield levels. It assumes certain
shapes in response to market expectations of future interest rates. Market participants
therefore analyze the current shape of the curve to determine the direction of future
interest rates. The curve is scrutinized to divine information, not just by bond traders
and fund managers, but also by corporate finance personnel. Traders and fund man-
agers want to identify securities they believe are under-valued or over-valued so that
they can take suitable investment decisions. Investment bankers handling new debt
issues for their clients want to know the rate of return required by the market for debt
securities of the risk class represented by a client. This enables them to set the coupon
on the to-be-issued security. Central banks and government treasury departments also

https://doi.org/10.1515/9781547400669-004
98 | Chapter 4: Yield Curves and the Term Structure

analyze the curve for its information content about future interest levels. This infor-
mation is then used to set rates for the economy as a whole.

Spot Rates of Interest

The spot rate of interest for a particular time period is the discount rate that is applica-
ble for a zero coupon instrument maturing at the end of the period. Unless otherwise
stated, we take one period as six months.
For instance, assume that the price of a six-month zero coupon bond with a face
value of $1,000 is $975. The one-period spot rate is given by

1,000
975 = ⇒ s1 = 0.051282 ≡ 5.1282%
(1 + s21 )

Similarly, if a one-year or a two-period zero coupon bond has a price of $910, then
the two-period spot rate is given by

1,000
910 = ⇒ s2 = 0.096569 ≡ 9.6569%
s2 2
(1 + 2
)

The Relationship Between Spot Rates and the YTM

A plain vanilla bond consists of a series of cash flows arising at six-month intervals.
Thus such a bond is equivalent to a portfolio of zero coupon bonds, where each cash
flow represents the face value of a zero coupon bond maturing at that particular in-
stant. The correct way to price a bond is by discounting each cash flow at the spot rate
for the corresponding period. The yield to maturity, is that single rate that makes the
present value of all future cash flows from the bond equal to its dirty price. Thus the
yield to maturity is a complex non-linear average of spot rates.
Let’s take the case of a bond with a face value of $1,000 and one year to maturity.
Assume that it pays a coupon of 7% per annum on a semiannual basis. Using the spot
rates derived in the previous section, we can calculate the price of the bond to be

35 1,035
P= + = $975.9747
(1.025641) (1.048285)2

The yield to maturity of this bond is given by

35 1,035
975.9747 = +
(1 + y2 ) (1 + y )2
2
⇒ y = 9.5764%
Yield Curve versus the Term Structure | 99

The fact that the YTM is a complex average of spot rates need not pose any prob-
lems per se. The problem with the yield to maturity is that it is a function of the coupon
rate for bonds with identical terms to maturity but with different coupons.1
For instance, let’s take a 15% coupon bond with a face value of $1,000 and one
year to maturity. Its price is given by
75 1,075
P= + = $1,051.3746
(1.025641) (1.048285)2

The yield to maturity of this bond is given by

75 1,075
1,051.3746 = y +
(1 + 2 ) (1 + y )2
2
⇒ y = 9.4939%

Why is there a difference in the yields to maturity of the two bonds? After all they both
have one year to maturity. Let’s take the 7% bond first. It has
35
(1.025641)
= 0.034965 ≡ 3.4965%
975.9747
of its value tied up in one-period money and the balance 96.5035% tied up in two-
period money.
However, in the case of the 15% bond,
75
(1.025641)
= 0.069552 ≡ 6.9552%
1,051.3746
of its value is tied up in one-period money, whereas the balance 93.0448% is tied up
in two-period money.
The one-period spot rate is less than the two-period spot rate, which implies that
one-period money is cheaper than two-period money. Because the second bond has
a greater percentage of its value tied up in one-period money, its yield to maturity
is lower. This is a manifestation of what is termed the coupon effect. In other words,
because the yield to maturity is a complex average of spot rates, it tends to vary with
the coupon rate on the bond, when we compare bonds with different coupons, but
with the same time to maturity.

Yield Curve versus the Term Structure

Technically speaking, a “yield curve” is a graph depicting the relationship between
the yield to maturity, which is plotted along the Y-axis, and the time to maturity, which

1 See Fabozzi [28].

100 | Chapter 4: Yield Curves and the Term Structure

is plotted along the X-axis. For the purpose of constructing the yield curve, it is imper-
ative that the bonds being compared belong to the same credit risk class. This is the
most commonly used version of the yield curve for the simple reason that the YTM is
the most commonly used measure of the yield from a bond.
The expression “term structure of interest rates,” on the other hand, refers to a
graph depicting the relationship between spot rates of interest, as shown along the
Y-axis, and the corresponding times to maturity, which are plotted along the X-axis.
Once again, to facilitate meaningful inferences, the data used to construct the graph
should be applicable to bonds of the same risk class. The “term structure of interest
rates” is also referred to as the “zero coupon yield curve” because the YTM of a zero
coupon bond is nothing but the spot rate. The zero coupon yield curve is considered
to be the true term structure of interest rates because there is no reinvestment risk.
This is because such bonds do not give rise to any cash flows prior to maturity, and
consequently there is no risk that cash flows may have to be reinvested at a lower than
anticipated rate.
The yield curve is equivalent to the term structure if the term structure is flat, or in
other words, the spot rates are the same for all maturities. This is because, when the
term structure is flat, the YTM, which is a complex average of spot rates, is equal to
the observed spot rate. However, when the term structure is not flat, as is usually the
case, the YTM is somewhere between the lowest spot rate, and the highest spot rate.

Required Symbols
We will use the following symbols in connection with our analysis of yield curves and
the term structure:
– yi ≡ the yield to maturity for an i period bond.
– si ≡ the i period spot rate, or equivalently the yield to maturity of an i period zero
coupon bond.
– fnm−n ≡ the m − n period implied forward rate for a loan to be made after n periods.
– E0 [n Sm−n ] ≡ the current expectation of the m − n period spot rate that is expected
to prevail n periods from now.

Bootstrapping to Obtain Spot Rates

In practice, we are unlikely to have data for the prices of zero coupon bonds conve-
niently maturing at regularly spaced intervals of time. Bootstrapping is a technique for
determining the term structure of interest rates, given price data for a series of coupon
paying bonds. It works as illustrated in Example 4.1.
Practical Difficulties with Bootstrapping | 101

Example 4.1. Assume that we have the following data for four bonds, each of which matures at the
end of the stated period of time. For ease of exposition, we also assume that the bonds pay coupons
on an annual basis. The face value is $1,000 for all the bonds.

Table 4.1: Inputs for determining the zero coupon yield curve.

Time to Price in Coupon

Maturity Dollars

1 Year 980 6%
2 Years 960 8%
3 Years 940 9%
4 Years 925 10%

The one-year spot rate is given by:

1,060
980 =
(1 + s1 )
⇒ s1 = 8.1633%

Using this information, the two-year spot rate can be determined as follows:

80 1,080
960 = +
(1.081633) (1 + s2 )2
⇒ s2 = 10.4042%

Similarly, the three-year spot rate can be determined as follows:

90 90 1,090
940 = + +
(1.081633) (1.104042)2 (1 + s3 )3
⇒ s3 = 11.6597%

And, finally, using the same logic,

100 100 100 1,100

925 = + + +
(1.081633) (1.104042)2 (1.116597)3 (1 + s4 )4
⇒ s4 = 12.8321%

Practical Difficulties with Bootstrapping

In Example 4.1 we used the prices of four bonds, maturing after one, two, three, and
four years, respectively. In some cases, there may be several bonds of the same risk
class, maturing at a given point in time. Usually each has its own coupon. When esti-
mating spot rates for a given maturity, obviously the coupon rates of the bonds being
used are a factor. One of the other major issues, when estimating the term structure
102 | Chapter 4: Yield Curves and the Term Structure

using bootstrapping, is that a bond may not exist or may not actively trade for a par-
ticular maturity. And it is not necessary that we always have access to a set of bonds
whose maturity dates are conveniently spaced exactly one period apart.
Finally, all traded bonds may not be plain vanilla in nature. Institutions like the
U.S. Treasury have issued bonds that can be recalled after a point in time. This too
has implications for the bootstrapping procedure, for we cannot compare apples with
oranges.

Coupon Yield Curves and Par Bond Yield Curves

One of the other problems with bootstrapping is that in practice we receive data in the
form of prices of bonds with different coupons. One way of getting over this estimation-
related problem is by using data for bonds that all have the same coupon. The “yield
curve” that is so obtained is called the “coupon yield curve.” If we construct such a
curve, we can see that in general higher coupon bonds trade at a discount (have higher
yields) relative to lower coupon bonds. This is due to reinvestment risk.
The “par bond yield curve” on the other hand, is an estimate of the yield curve
obtained from data for bonds that have different coupons, but all of which trade at par.
In this case, the coupon for each of these bonds is nothing but its yield to maturity.
Bootstrapping can then be applied to such a data set to derive the vector of spot rates
as before. Example 4.2 is a numerical illustration of this procedure. We are using the
data given in Table 4.1 but assuming that all the bonds are trading at par.

Example 4.2. The one-year spot rate is obviously 6%. Using this information, the two-year spot rate
can be determined as follows:
80 1,080
1,000 = +
(1.06) (1 + s2 )2
⇒ s2 = 8.08%

Similarly, the three year spot rate can be determined as follows:

90 90 1,090
1,000 = + +
(1.06) (1.0808)2 (1 + s3 )3
⇒ s3 = 9.16%

Table 4.2: The par-bond approach to bootstrapping.

Time to Price in YTM =

Maturity dollars Coupon

1 Year 1,000 6%
2 Years 1,000 8%
3 Years 1,000 9%
4 Years 1,000 10%
Deducing a Par Bond Yield Curve | 103

And, finally, using the same logic,

100 100 100 1,100

1,000 = + + +
(1.06) (1.0808)2 (1.0916)3 (1 + s4 )4
⇒ s4 = 10.30%

The par bond yield curve is not commonly encountered in secondary market trading.
However it is often constructed and used by people in corporate finance departments
and others who are involved with issues in the primary market. Investment bankers
use par bond yield curves to determine the required coupon for a new bond that is
to be issued at par. This is because new issues are typically issued at par and conse-
quently the banker needs to know the coupon that needs to be offered to ensure that
the bonds can be issued at par. In practice, the market uses data from non-par plain
vanilla bonds to first derive the zero coupon yield curve. This information is then used
to deduce the hypothetical par yields that would be observed if traded par bonds were
to be available.

Deducing a Par Bond Yield Curve

The par bond yield curve can be derived using a vector of spot rates. Let us consider
the spot rates that we derived in Example 4.1.

Example 4.3. The yield for a one-year par bond is obviously 8.1633%. The yields for the other bonds
are summarized in Table 4.3.
The yield, or equivalently, the coupon for the two-year par bond can be deduced as follows:

C 1,000 + C
1,000 = +
(1.081633) (1.104042)2
⇒ C = $102.9235 ⇒ c = 10.2924%

Similarly,

C C 1,000 + C
1,000 = + +
(1.081633) (1.104042)2 (1.116597)3
⇒ C = $114.3581 ⇒ c = 11.4358%

Table 4.3: Inputs for inferring a par bond yield curve.

Time to Spot Rate

Maturity

1 Year 8.1633%
2 Years 10.4042%
3 Years 11.6597%
4 Years 12.8321%
104 | Chapter 4: Yield Curves and the Term Structure

and

C C C 1,000 + C
1,000 = + + +
(1.081633) (1.104042)2 (1.116597)3 (1.128321)4
⇒ C = $124.3489 ⇒ c = 12.4349%

Thus a two-year bond of this risk class ought to be issued with a coupon of 10.2924% if it is to be sold
at par. Similarly, three-year and four-year bonds should carry coupons of 11.4358% and 12.4349%,
respectively.

Implied Forward Rates of Interest

Consider an investor who is contemplating investing for two periods. He has the option
of investing in a two-period zero coupon bond yielding s2 . Or he can invest in a one-
period bond yielding a rate of s1 , and buy a forward contract to lock in a roll over rate
f11 . The two options are equivalent to the investor if

(1 + s2 )2 = (1 + s1 )(1 + f11 ) (4.1)

f11 is the one-period forward rate, or the rate as of today, for a forward contract to make
a one-period loan one period from today. It is the implied forward rate that is contained
in the term structure. In general, if we have an n period spot rate and an m period spot
rate, where m > n, then
m−n
(1 + sm )m = (1 + sn )n (1 + fnm−n ) (4.2)

where fnm−n is the m−n period implied forward rate for a loan to be made after n periods.
Forward rates are believed to convey information about the expected future in-
terest rate structure. In fact, one theory called the unbiased expectations hypothesis
holds that such rates are nothing but current expectations of future interest rates.

Example 4.4. The one-year spot rate is 5%, the two-year spot rate is 6%, the three-year spot rate is
8%, and the four-year spot rate is 10%. Using this information what can we deduce about implied
forward rates?

(1 + s2 )2 = (1 + s1 )(1 + f11 ) ⇒ f11 = 0.07

The rate for a forward contract to make a one-period loan after one period is 7% per annum.

2
(1 + s3 )3 = (1 + s1 )(1 + f12 ) ⇒ f12 = 0.0953

The rate for a forward contract to make a two-period loan after one-period is 9.53% per annum.

(1 + s3 )3 = (1 + s1 )(1 + f11 )(1 + f21 ) ⇒ f21 = 0.1212

Fitting the Yield Curve | 105

That is, the rate for a forward contract to make a one-period loan after two periods is 12.12%.
3
(1 + s4 )4 = (1 + s1 )(1 + f13 ) ⇒ f13 = 0.1172

The rate for a forward contract to make a three-period loan after one period is 11.72%.
2
(1 + s4 )4 = (1 + s2 )2 (1 + f22 ) ⇒ f22 = 0.1415

The rate for a forward contract to make a two-period loan after two periods is 14.15%. Finally,

(1 + s4 )4 = (1 + s3 )3 (1 + f31 ) ⇒ f31 = 0.1622

That is, the rate for a forward contract to make a one-period loan after three periods is 16.22%.

Fitting the Yield Curve

When plotting a yield curve, we fit a series of discrete points of yield against the cor-
responding maturities. Similarly, for the term structure of interest rates, we plot spot
rates for a fixed time period against the time period. The yield curve itself, however,
is a smooth curve drawn through these discrete points. We require a method that al-
lows us to fit the curve as accurately as possible. This aspect of yield curve analysis is
known as yield curve modeling or estimating the term structure.
The yield curve is derived from coupon bond prices and yields. In attempting to
model the curve from bond yield data, we need to be aware of two fundamental issues.
First, there is the problem of gaps in the maturity spectrum because in the real world
bonds do not mature at regular intervals of time. Second, the term structure is defined
in terms of spot rates or zero-coupon interest rates. But in many markets there is no
zero coupon bond market.
With all this on mind, we now can study four methods for modeling the yield
curve.2

Interpolation

The simplest method that can be used to fit the yield curve is linear interpolation. For
example, assume that we are given the following data:

s5 = 8% and s10 = 9%

We can then calculate the eight-year spot rate as

(8 − 5)
s8 = 8.00 + × (9.00 − 8.00) = 8.60%
(10 − 5)

2 See Choudhry [11].

106 | Chapter 4: Yield Curves and the Term Structure

Polynomial Models of the Yield Curve

A polynomial of degree k can be expressed as

yi = α + β1 t + β2 t 2 + ⋅ ⋅ ⋅ + βk t k + ui (4.3)

where yi is the YTM of bond i, t is its term to maturity, and ui is the residual error.
To determine the coefficients of the polynomial, we minimize the sum of the squared
residual errors given by
N
∑ ui 2
i=1

where N is the number of bonds used.

Regression Models of the Yield Curve

Regression is a variation of the polynomial approach. This method uses bond prices
as the dependent variable and the coupons and face values of the bonds as the inde-
pendent variables. The standard format is

Pi = β1 C1i + β2 C2i + ⋅ ⋅ ⋅ + βN (CNi + M) + ui (4.4)

In this expression, Pi is the dirty price of bond i, Cji is the coupon of bond i in period j,
and ui is the residual error. The spot rates can be derived from the estimated relation-
ships using the expression
1
βn =
(1 + sn )n

The Nelson-Siegel Model of the Yield Curve

Before we describe the Nelson-Siegel technique, we need to familiarize ourselves with

bond pricing in a continuous time framework.
Take a zero coupon bond that pays $1 after n periods. In a discrete time setting,
we would express the bond price as:
1
P(0, n) = (4.5)
(1 + sn )n
where sn is the n-period spot rate at time 0. If we were to compound interest ‘m’ times
per period where m > 1, then we would express the bond price as:
1
P(0, n) = sn mn
(4.6)
(1 + m
)
Fitting the Yield Curve | 107

For instance, if we were to compound four times every period, then m would be equal
to 4. In the limit m → ∞, we get the case of continuous compounding. In the limit, we
can express the price of the bond as:

P(0, n) = e−sn ×n (4.7)

We know that in the discrete time framework the n-period spot rate can be expressed
as

1
(1 + sn )n = (1 + s1 )(1 + f11 )(1 + f21 ). . .(1 + fn−1 ) (4.8)

In the case of continuous compounding, the equivalent representation is:

sn × n = ∫ fs ds (4.9)
0

where fs is the instantaneous forward rate.

Nelson-Siegel proposed the following representation for the instantaneous for-
ward rate:
s
fs = β0 + β1 e−s/θ + β2 × e−s/θ (4.10)
θ

Integrating this function we get the following expression for the n-period spot rate:

1 − e−n/θ 1 − e−n/θ
sn = β0 + β1 × [ ] + β2 × [ − e−n/θ ] (4.11)
n/θ n/θ

The four parameters β0 , β1 , β2 , and θ have to be empirically estimated.

Interpretation of the Nelson-Siegel Model

We can interpret the parameters that must be estimated, as follows:

Using L’ Hôpital’s rule, we find that

lim f (n) = β0
n→∞

lim f (n) = β0 + β1
n→zero
−n
1−e θ
lim [ n ]=0
n→∞
θ
−n
1−e θ −n
lim [ n −e θ ]=0
n→∞
θ
108 | Chapter 4: Yield Curves and the Term Structure

Thus,

lim s = β0
n→∞ n
−n
1−e θ
lim [ n ]=1
n→zero
θ
−n
1−e θ −n
lim [ n −e θ ]=0
n→zero
θ

And,

lim s = β0 + β1
n→zero n

An Economic Interpretation
β0 is a constant that represents the long-term interest rate level.3 β1 captures the slope
of the curve. If it is positive, the curve slopes downward, whereas if it is negative, the
curve slopes upward. β2 captures the hump or the trough in the curve. If it is posi-
tive, there is a hump, but if it is negative, there is a trough. The higher the absolute
value of β2 the more pronounced the hump or the trough. θ is the shape parameter. It
determines the steepness of the slope and the location of the hump or trough.
When the time to maturity tends to infinity, the slope and curvature components
vanish, and both the long-term spot rate and the forward rate converge to β0 . From an
economic standpoint, we would assume β0 to be close to the empirical long-term spot
rate and positive in sign. When the time to maturity tends to zero, only the curvature
component vanishes, and the forward and spot rates converge to β0 + β1 . β1 measures
the slope of the term structure. The degree of curvature is controlled by β2 . Finally,
θ determines the maximum or minimum as the case may be, depending on whether
there is a hump or a trough.
Svensson extended the model by adding two additional variables, to create what
has been termed as the Nelson-Siegel-Svensson model. This model can be stated as:
−n −n −n
1 − e θ1 1 − e θ1 −n 1 − e θ2 −n
sn = β0 + β1 × [ n ] + β2 × [ n − e ] + β3 × [
θ1
n − e θ2 ] (4.12)
θ1 θ1 θ2

In this case we need to estimate six parameters, the two additional parameters being
β3 and θ2 . β3 and θ2 have an interpretation similar to β2 and θ1 . Thus by including β3 ,
Svensson incorporates an additional hump or trough.
The Nelson-Siegel method for estimating the term structure has a number of ad-
vantages. First, its functional form can handle a variety of shapes of the term structure

3 See Annaert et al. [1].

Theories of the Term Structure | 109

that are observed in the market. Second, the model avoids the need to introduce other
assumptions for interpolation between intermediate points. For instance, the boot-
strapping approach gives us a vector of spot rates spaced six months apart. To value a
bond whose life is not an integer multiple of semiannual periods, we obviously need
to interpolate. On the contrary, using the Nelson-Siegel approach, we can derive the
spot rate at any point in time and not just at certain discrete points.

Theories of the Term Structure

From observing yield curves in different markets at various points in time, an individ-
ual who studies the bond market notices that the yield curve tends to adopt one of the
four basic shapes:
– Upward sloping: In the case of a yield curve with a positive slope, also termed as
a rising yield curve, short term yields are lower than long term yields.
– Downward sloping: In the case of a yield curve with a negative slope, also termed
as an inverted yield curve, long term rates are substantially lower than short term
rates.
– Humped: A humped yield curve is characterized by lower rates at the short end of
the spectrum. The curve then rises, reaching a peak at the middle of the maturity
spectrum, and then gradually slopes downward at longer maturities.
– U-shaped: A U-shaped yield curve is characterized by higher rates at the short
end of the spectrum. The curve then falls, reaching a minimum at the middle of
the maturity spectrum, and then gradually slopes upward at longer maturities.

A great deal of effort is expended by analysts and economists in analyzing and inter-
preting the yield curve, for there is often substantial information that is associated
with the curve at any point in time. Various theories have been advanced that purport
to explain the observed shapes of the curve. However, no theory by itself is able to ex-
plain all aspects of the curves that are observed in practice. So often analysts seek to
explain specific shapes of the curve using a combination of the accepted theories.4

The Pure or Unbiased Expectations Hypothesis

The unbiased expectations hypothesis states that the current implied forward rates are
unbiased estimators of future spot rates. Per this theory, long-term rates are geomet-
ric averages of expected future short-term rates. Thus a positively sloped yield curve
would be consistent with the argument that the market expects spot interest rates to

4 See Choudhry [11] and Taggart [64].

110 | Chapter 4: Yield Curves and the Term Structure

rise. If rates are expected to rise, then investors in long-term bonds become perturbed
for they face the specter of a capital loss. This is because rising interest rates lead to
declining bond prices, and long-term bonds are more sensitive to rising interest rates
than short-term bonds. In such situations investors start selling long-dated securities
and buying short-dated securities. This leads to an increase in yields on long-term
bonds and a decline in yields on short-term bonds. The overall result is an upward
sloping yield curve. On the contrary, an inverted yield curve indicates that the market
expects future spot rates to fall.
The hypothesis can be used to explain any shape of the yield curve. For instance, a
humped yield curve is consistent with the explanation that investors expect short-term
rates to rise and long-term rates to fall. Expectations or views on the future direction
of the market are a function mainly of the expected rate of inflation. If the market
expects inflationary pressures in the future, the yield curve slopes positively, whereas,
if inflation is expected to decline, then the yield curve slopes negatively.

The Liquidity Preference Theory (LPT)

Intuitively, most of us feel that longer maturity instruments are riskier than shorter
maturity ones. An investor lending money for five years usually demands a higher rate
of interest than when lending money to the same entity for a year. This is because the
borrower may not be able to repay the loan over a longer term period. For this reason,
long-dated yields should be higher than short dated yields.
Take the case where the market expects inflation to remain fairly stable over time.
The expectations hypothesis postulates that this scenario would be characterized by
a flat yield curve. However, the liquidity preference theory (LPT) predicts a positively
sloping yield curve. The argument is as follows. Generally a borrower would like to
borrow over a long period, whereas a lender would like to lend over a short period.
Thus lenders have to be suitably rewarded if they are to be induced to lend for longer
periods of time. This compensation can be considered a premium for the loss of liq-
uidity from the standpoint of the lender. The premium can be expected to increase the
farther the investor lends across the term structure. So that the longest dated instru-
ments, all else being equal, have the highest yield.
Per this hypothesis, the yield curve should almost always be upward sloping, re-
flecting the bond holders’ preference for liquidity. However, the theory can explain
inverted yield curves by postulating that interest rates are likely to decline in the fu-
ture, and a consequence of this, despite the liquidity premium, is that long-term rates
may be lower than short-term rates.
Theories of the Term Structure | 111

The Expectations Hypothesis versus the LPT: A Mathematical Analysis

Per the expectations hypothesis, forward rates are unbiased expectations of future
spot rates. Thus

fnm−n = E0 [n Sm−n ] (4.13)

In other words, the rate for a forward contract to make an m − n period loan, n periods
from today, is the current expectation of the m − n period spot rate that is expected to
prevail n periods from now. The expectations hypothesis can explain any shape of the
term structure. For instance, an expectation that future short-term interest rates will
be above the current level leads to an upward sloping term structure.

Example 4.5. Assume that s1 = 5.50%; E[1 s1 ] = 6.0%; E[2 s1 ] = 7.5%, and E[3 s1 ] = 8.5%. If so, then

s2 = [(1.055)(1.06)]0.5 − 1 = 5.75%
s3 = [(1.055)(1.06)(1.075)]0.3333 − 1 = 6.33%
s4 = [(1.055)(1.06)(1.075)(1.085)]0.25 − 1 = 6.87%

According to the expectations hypothesis, investors care only about expected returns
and not about risk. Let’s take the case of an investor who chooses to invest for two
periods. He can buy a two-period bond yielding a rate of s2 . Or he can buy a one-period
bond that yields s1 and then roll over into another one-period bond at maturity. Per this
hypothesis, the investor will be indifferent between the two strategies if the expected
returns in both cases are equal.
In other words, the hypothesis predicts that the market will be in equilibrium if

(1 + s2 )2 = E[(1 + s1 )(1 + 1 s1 )] = (1 + s1 )E[(1 + 1 s1 )] (4.14)

But we know that if arbitrage is to be ruled out, then

(1 + s2 )2 = (1 + s1 )(1 + f11 ) (4.15)

Thus, per the expectations hypothesis

f11 = E(1 s1 )

Now let’s focus on the liquidity preference theory. Take the case of an investor
who has a one-period investment horizon. He can buy a one-period bond and lock in
a rate of s1 . Or he can buy a two-period bond and sell it after one year. In the second
case, the rate of return is uncertain at the outset, for it depends on the one-period rate
that prevails one period from now.
112 | Chapter 4: Yield Curves and the Term Structure

Consider a two-period zero coupon bond with a face value of $1,000. Its current
price is

1,000 1,000
=
(1 + s2 ) 2 (1 + s1 )(1 + f11 )

The expected price of the bond after one period is

1,000 1,000
E[ ]≥
(1 + 1 s1 ) 1 + E(1 s1 )

How do we know that the expected price after one year will be greater than or equal to
the face value discounted by the expected one-period spot rate one period from now.
This deduction arises from a result called Jensen’s Inequality. It states that for a convex
function

E[f (X)] ≥ f [E(X)]

That is, the expectation of the function is greater than or equal to the function of the
expectation. In our case, 1,000
1+1 s1
is a convex function because the second derivative is
positive.
The expected rate of return from the two-period bond over the first year is:
1,000 1,000
E [ (1+ s)
]− (1+s1 )(1+f11 )
1 1
1,000
(1+s1 )(1+f11 )
1,000 1,000
1+E(1 s1 )
− (1+s1 )(1+f11 )
≥ 1,000
(1+s1 )(1+f11 )
1,000 1,000
E [ (1+ s)
]− (1+s1 )(1+f11 )
1 1
⇒ 1,000
(1+s1 )(1+f11 )

(1 + s1 )(1 + f11 )
≥ −1
1 + E(1 s1 )

A sufficient condition for the expected one-period return from the two-period bond
to be greater than the one-period spot rate, s1 , is f11 > E(1 s1 ). Thus the expected one-
period return from a two-period bond is greater than the current one-period spot rate
if the implied forward rate is greater than the current expectation of next period’s one-
period spot rate.
Now an investor with a one-period investment horizon will obviously choose to
hold a two-period bond only if its expected return is greater than the assured return
on a one-period bond. This is because if the investor chooses to hold the two-period
bond, he will have to sell it after one period at a price that is unknown at the outset.
From the previous analysis this implies that the forward rate must be higher than the
Theories of the Term Structure | 113

expected one-period spot rate. Thus if investors are risk averse, which is the normal
assumption made in finance theory, the forward rate will exceed the expected spot
rate by an amount equal to the risk premium or what may be termed as the liquidity
premium.5
We know that

(1 + s2 )(1 + s2 ) = (1 + s1 )(1 + f11 )

as per the LPT,

(1 + s1 )(1 + f11 ) > (1 + s1 )[1 + E(1 s1 )]

Therefore,

(1 + s2 )(1 + s2 ) > (1 + s1 )[1 + E(1 s1 )]

Consider a downward sloping yield curve. That implies that s1 > s2 . Therefore, it must
be the case that E(1 s1 ) is substantially less than s1 . In other words, the market expects
spot rates to decline substantially. For instance, if s1 = 7 % and s2 = 6%, then f11 =
5.01%. Per the expectations hypothesis, E(1 s1 ) = 5.01%. However, per the LPT, E(1 s1 ) <
5.01%. If we assume that the liquidity premium is 0.50%, then E(1 s1 ) = 4.51%.
Now let’s take the case of a flat term structure. Per the expectations hypothesis,

s1 = s2 = f11 = E(1 s1 )

However, according to the LPT, E(1 s1 ) < s1 = s2 . Thus, although the expectations hy-
pothesis implies that the market expects spot rates to remain unchanged, the predic-
tion according to the liquidity preference theory is that the market expects spot rates to
decline. For instance, if s1 = s2 = 7%, then according to the expectations hypothesis,
E(1 s1 ) = 7%. However, according to the LPT, E(1 s1 ) = 7 − 0.50 = 6.50%.
Finally let’s take the case of an upward sloping yield curve. If s1 < s2 , then for
a slightly upward sloping yield curve, the LPT would predict that rates are going to
marginally decline. However, if the curve slopes steeply upward, then the LPT implies
that short term rates are going to rise. For instance, assume that s1 = 7% and s2 = 7.1%.
If so,

f11 = 7.2% ⇒ E(1 s1 ) = 7.20 − 0.50 = 6.70%

However, if s2 = 7.3%, then

f11 = 7.6% ⇒ E(1 s1 ) = 7.60 − 0.50 = 7.10%

In both these cases, however, the expectations hypothesis predicts that spot rates are
likely to rise. In the first scenario, per the expectations hypothesis, E(1 s1 ) = 7.20%,
whereas in the second case E(1 s1 ) = 7.60%.

5 See Bodie et al. [5].

114 | Chapter 4: Yield Curves and the Term Structure

The Money Substitute Hypothesis

According to the money substitute hypothesis, short-term bonds are substitutes for
holding cash. Investors hold only short-dated bonds because they are viewed as hav-
ing low or negligible risk. As a result the yields on short-dated bonds are depressed due
to increased demand, and consequently long-term yields are greater than short-term
yields. Borrowers, on the other hand, prefer to issue debt for long maturities and on as
few occasions as possible to minimize costs. Thus the yields on long-term securities
are driven upward due to increased supply and consequently lower prices.6

The Market Segmentation Hypothesis

The market segmentation hypothesis states that the capital market is made up of a
wide variety of issuers, each with different requirements. Certain classes of investors
prefer short-dated bonds, and others prefer long-dated bonds. The theory argues that
activity is concentrated in certain specific areas of the market, and that there are no in-
terrelationships between these segments of the market. The relative amount of funds
invested in each market segment causes differentials in supply and demand that lead
to humps in the yield curve.
Thus, per this theory, the observed shape of the yield curve is determined by the
supply and demand for specific maturity investments, and the dynamics in a partic-
ular market segment have no relevance for any other part of the curve. For example,
banks concentrate a large part of their activity at the short-end of the curve as a part
of daily cash management known as asset-liability management, and for regulatory
purposes known as liquidity requirements. On the other hand, fund managers such as
pension funds and insurance companies are active at the long-end of the market.7 Few
institutions, however, have a preference for medium-dated bonds. This behavior leads
to high prices and low yields at both the short and long ends of the maturity spectrum
and to high yields in the middle of the term structure.
The theory argues that financial institutions like banks, pension funds, and mu-
tual funds often act as risk minimizers and not as profit maximizers as a theory such
as the unbiased expectations hypothesis would have us believe. In their quest for
minimizing risk, they hedge the risk of fluctuations in prices and yields by balanc-
ing the maturity of their assets with that of their liabilities. For instance, the treasurer
or CFO of a company who has surplus funds for a short period invests only in short-
term money market securities such as commercial paper. On the contrary a pension
fund, whose liabilities are long term, invests only in long-term bonds.

6 See Choudhry [11].

7 See Choudhry [11].
Features of the Debt Market and Theories of the Term Structure | 115

Thus, this theory argues that risk aversion precludes agents from switching from
one market segment to another. This is true no matter how attractive rates may be in
another segment of the market. Thus the yield curve is a collection of sub markets.
Each of these has its own supply and demand dynamics and consequently its own
equilibrium rate of interest. Thus the implied forward rate is unrelated to expected
future spot rates.
One policy implication of this hypothesis is that if the submarkets are isolated,
the central bank can alter the shape of the curve by influencing the supply and de-
mand dynamics in one or more market segments. For instance, if the objective is to
have an upward sloping yield curve, the central bank can flood the market with long-
term bonds and acquire short-term bonds. This causes long-term bond prices to fall
or, equivalently, long-term yields to rise. At the same time the induced demand for
short-term bonds, pushes up prices in this market segment, causing short-term yields
to fall. The net consequence is an upward sloping yield curve.
While the expectations hypothesis argues that securities are perfect substitutes
for each other, this theory goes to the other extreme, by arguing that investors and
issuers are so risk averse that they buy and sell bonds only for maturities matching
their desired time horizons.

The Preferred Habitat Theory

The preferred habitat theory is a slightly modified version of the segmentation hypoth-
esis. This suggests that different market participants have an interest in specified areas
of the yield curve but can be persuaded to hold bonds from other parts of the maturity
spectrum if they are provided with sufficient incentives. Hence, banks, which typically
operate at the short-end of the spectrum, may at times hold long-dated bonds when
the price of these bonds falls to a certain level, thereby ensuring that the returns from
holding such bonds is commensurate with the attendant risk. Similar considerations
may persuade long-term investors to hold short-term debt. So the incentive for an in-
vestor to shift out of his preferred habitat is the inducement offered by way of a higher
rate of interest.

Features of the Debt Market and Theories of the Term Structure

In practice we find that the yield curve is generally upward sloping, and yields tend
to taper off at longer maturities.8 The liquidity preference theory postulates the in-
vestors require a higher rate of return from long-term bonds because they subject the

8 See Smith [56] for a detailed exposition.

116 | Chapter 4: Yield Curves and the Term Structure

holders to greater risk. The market segmentation theory argues that demand for long-
term funds is strong relative to short-term funds while supply of short-term funds is
strong relative to long-term funds. This is consistent with the observation that whereas
borrowers want to lock in rates for longer periods, lenders prefer the flexibility of short-
term investments. Thus the long-term debt capital market is characterized by high de-
mand for funds and lower supply, which translates into a high supply of bonds and low
demand. This means that long-term rates will be high. However, the short-term debt
capital market is characterized by lower demand for funds and higher supply, which
would manifest itself as a low supply of bonds and a high demand for the same. This
implies that short-term rates would be lower. This could explain the fact that the yield
curve is typically upward sloping.
In practice we find that short-term yields are more volatile than longer-term yields.
In other words, the term structure of volatility is downward sloping. Both the expecta-
tions theory and the liquidity preference theory explain this by arguing that because
long-term rates are an average of future short term rates, they are less volatile. This is
because averages are always less volatile. The market segmentation theory says that
there is a lot of activity in the money market, whereas the long-term debt market is
relatively more stable. One reason could be that long maturity bonds attract buy and
hold investors like life insurance companies and pension funds. These entities are in
the game for the long run and consequently long-term bonds are less actively traded.
Investors with a shorter time horizon frequently engage in asset reallocation. While
doing so, the tendency is typically to park funds in the money market until a suitable
alternate investment is identified. This could explain the volatility of the money mar-
ket.
Yet another real-life observation is that both short-term and long-term yields usu-
ally change in the same direction. Thus shifts in the yield curve are usually parallel.
This could be explained as follows. Macroeconomic factors such as inflation, mone-
tary policy, balance of trade, foreign exchange rates, and fiscal policy usually tend to
influence both short-term and long-term rates in a similar fashion. This argument is
consistent with all the postulated theories.
The liquidity premium hypothesis predicts that long-term yields will be higher
than a geometric average of expected short-term rates. This is because most investors
have a short-term horizon and therefore need an inducement, to hold longer-term se-
curities. In other words, such investors confront the specter of market risk or price risk.
On the other hand, consider a scenario where most traders have a long-term horizon.
Such participants need an inducement to hold short-term securities, for such invest-
ment strategies expose them to reinvestment risk. So whether the long-term rates are
higher or lower than a geometric average of expected short-term rates depends on how
investors are distributed across maturities.
In practice, speculative bond traders as well as dealers usually have a short-
term horizon. Speculative traders, taking a view on short-term rates,prefer short-term
Chapter Summary | 117

bonds as vehicles of speculation. Bond dealers too typically seek to rotate their in-
ventory frequently and therefore are short-term holders. Hence they tend to demand
higher yields from long-term bonds because the latter expose them to greater risk.
Both speculators and dealers prefer to hold short-term bonds, and the high demand
for these is consistent with a low yield. These observations perhaps explain why the
yield curve is generally upward sloping.

Chapter Summary
In this chapter, we introduced the concept of the yield curve and the closely related
issue of the term structure. The former refers to the relationship between the yield
to maturity and the term to maturity, whereas the latter describes the relationship
between spot rates of interest and the term to maturity. We analyzed the coupon ef-
fect in detail, which explains why otherwise identical bonds with different coupons,
may have different yields to maturity. The bootstrapping technique for estimating spot
rates was studied in detail, and its practical difficulties were elaborated upon. We in-
troduced the issue of forward rates of interest, and looked at different interpretations
of the forward rate, such as the expectations hypothesis and the liquidity premium
hypothesis. Finally, we attempted to explain real-life observations about yields, using
different theories of the yield curve.
In the next chapter, we examine the issues of duration, convexity, and dispersion
of bonds, and study the application to bond immunization strategies.
Chapter 5
Duration, Convexity, and Immunization
The duration of a plain vanilla bond can be defined as its average life. It is very easy to
define duration in the case of securities that yield a single cash flow, like a zero coupon
bond. In such cases there is no difference between the average time to maturity and the
actual time to maturity, for we need concern ourselves only with the terminal payment.
Consequently, in such cases, the duration of the security is nothing but its stated term
to maturity.
However the definition is not so clearcut in the case of a conventional coupon
paying debt security. In such cases, the asset gives rise to a series of cash flows, usually
on a semiannual basis, as well as a relatively large cash flow at the end that constitutes
the principal repayment. The average life of such a security can be obtained only by
taking cognizance of the times to maturity of the component cash flows. Because the
cash flows occur at different points in time, we also need to factor in the issue of the
time value of money.
Convexity of a bond accounts for the fact that the price-yield relationship of a bond
is convex and not linear. Whereas, duration accounts for a first order approximation
to the price-yield relationship, convexity factors into the fact that the relationship is
indeed convex. Immunization strategies protect a bond or a bond portfolio against in-
terest rate risk. As discussed earlier, interest rate changes impact bonds in two ways.
The higher the interest rate, the more the income from the re-investment of coupons.
However the higher the rate, the lower the sale price of the bond at the end of the
investment horizon. See Figure 5.1. An immunization strategy helps ensure that the
terminal cash flow from a bond portfolio at the end of the investment horizon is ade-
quate to meet the liability of the holder.

Figure 5.1: The price of the bond and its rate are inversely related.

https://doi.org/10.1515/9781547400669-005
A Mathematical Definition of Duration | 119

Frederick Macaulay came up with the concept of what he called duration. Per his def-
inition, duration is the weighted average maturity of the bond’s cash flows, where the
present values of the cash flows serve as the weights. This definition is comprehensive,
for it accounts for all the cash flows emanating from the instrument and factors in the
critical concept of the time value of money.

A Mathematical Definition of Duration

Macaulay’s duration may be stated using the following formula:
N
CF t × t
∑ t
t=1 P × (1 + y2 )
Let’s define the symbols used in the preceding expression:
– CF t ≡ the cash flow arising at time t. If the bond pays a coupon of C2 at semiannual
intervals, all the cash flows from the first to the penultimate are equal to C2 , and
the terminal cash flow is equal to C2 +M, where M is the maturity value of the bond.
– y ≡ the yield to maturity of the bond expressed in annual terms. At times, it means
the semiannual YTM.
– P ≡ the dirty price of the bond.
Since t represents the corresponding semiannual period, the final value arrived at is
in half-years and consequently has to be divided by two in order to annualize it.

Example 5.1. Consider a T-note with exactly five years to maturity. The face value is $1,000, and the
coupon is 6% per annum paid on a semiannual basis. The yield to maturity is 6.50%. The dirty price is
given by
P = 30 × PVIFA(3.25, 10) + 1,000 × PVIF(3.25, 10) = $978.9440
The duration calculation is depicted in Table 5.1.

Table 5.1: Computation of the duration on a coupon date.

Time Cash Flow PVCF t wt wt × t
1 30 29.0557 0.029681 0.029681
2 30 28.1411 0.028746 0.057493
3 30 27.2553 0.027842 0.083525
4 30 26.3974 0.026965 0.107861
5 30 25.5665 0.026116 0.130582
6 30 24.7617 0.025294 0.151766
7 30 23.9823 0.024498 0.171487
8 30 23.2274 0.023727 0.189816
9 30 22.4963 0.022980 0.206821
10 1,030 748.0603 0.764150 7.641503
PVCF stands for present value of the cash flow.
The symbol wt stands for the weight.

The duration is 8.7705 semiannual periods or 4.3853 years.

120 | Chapter 5: Duration, Convexity, and Immunization

A Useful Excel Function

When we want to multiply two vectors of data in Excel and add, a useful function is
SUMPRODUCT. Consider the data in Table 5.1. To compute the duration we need to
multiply the time vector with the weight vector. Assume that the time vector runs from
cells A1 to A10, and the weight vector runs from cells D1 to D10. We invoke the function
as sum total = SUMPRODUCT(A1:A10,D1:D10). The answer is 8.7705.

Duration of a Bond When the Settlement Date Is Between Two

Coupon Dates
Assume that the first coupon is k periods away, where k is calculated using the pre-
scribed day-count convention. The duration formula may be expressed as

N−1
CF t+k × (t + k)
∑ t+k
t=0 P × (1 + y2 )

Example 5.2. Consider the bond that we discussed in Example 5.1. Assume that there are 4.875 years
to maturity, or that the first coupon is 0.75 periods away. The duration can be calculated as illustrated
in Table 5.2.

Table 5.2: Computation of the duration between coupon dates.

Time Cash Flow PVCF t wt wt × t

0.75 30 29.28894 0.029681 0.022260

1.75 30 28.36702 0.028746 0.050306
2.75 30 27.47411 0.027842 0.076564
3.75 30 26.60930 0.026965 0.101119
4.75 30 25.77172 0.026116 0.124053
5.75 30 24.96051 0.025294 0.145442
6.75 30 24.17483 0.024498 0.165362
7.75 30 23.41387 0.023727 0.183884
8.75 30 22.67688 0.022980 0.201076
9.75 1,030 754.06560 0.764150 7.450465
PVCF stands for present value of the cash flow.
The symbol wt stands for the weight.

The dirty price is $986.8028, and the duration is 8.5205 semiannual periods or 4.2603 years.

A Concise Formula for the Duration on a Coupon Date

Let’s redefine the following variables:
y ≡ y2 ≡ the semiannual YTM
Duration of a Level Annuity | 121

c ≡ c2 ≡ the semiannual coupon rate

The duration of a plain vanilla bond, on a coupon date, can be stated as1

(1 + y) (1 + y) + N(c − y)
Duration = −
y c [(1 + y)N − 1] + y

The Case of a Par Bond

For a par bond, c = y. So,

(1 + y) (1 + y)
Duration = −
y y(1 + y)N
(1 + y) 1
= × [1 − ]
y (1 + y)N

Example 5.3. Consider the five-year T-note with a face value of $1,000. c is 3% and y is 3.25%. Thus,

1.0325 (1.0325) + 10(.03 − .0325)

D= −
.0325 .03 [(1.0325)10 − 1] + .0325
= 31.7692 − 22.9987 = 8.7705 semiannual periods

Duration of a Level Annuity

The duration of an annuity that pays a cash flow of $A every period for N periods is
given by2

(1 + y) N
−
y [(1 + y)N − 1]

Example 5.4. Consider an annuity that pays $50 per half-year period for five years. The YTM is 6.50%
per annum. The duration is given by

(1.0325) 10
D= −
.0325 [(1.0325)10 − 1]
= 31.7692 − 26.5326 = 5.2366 semiannual periods

1 See Appendix 5.1.

2 See Appendix 5.2 for a derivation.
122 | Chapter 5: Duration, Convexity, and Immunization

Duration of a Perpetuity

The duration of a perpetuity is given by (1+y)

y
. The derivation is given in Appendix 5.2.

Example 5.5. Consider a perpetuity that pays $50 per half-year forever. Assume that the semiannual
yield is 3.25%. The duration is given by

1.0325
= 31.7692 semiannual periods
0.0325

The Rationale Behind Duration

It was observed many years ago that long-term bonds are more sensitive to changes in
the YTM, compared to relatively shorter-term bonds. This was explained by arguing
that because long-term cash flows are more impacted by a given change in yield, long-
term bonds are more sensitive to such changes, as they have more cash flows coming
later. The reason why a long-term cash flow is more impacted by a yield change is the
CF t
following: The present value of a cash flow is y t . In the denominator 1 plus the
(1+ 2 )
discount rate is raised to the power of time. The larger the value of time, the greater is
the impact of a given yield change, △y.
However, it was also observed that an N year zero coupon bond is more price sen-
sitive than an N year coupon paying bond. This was initially perplexing because both
have a time to maturity of N years. In the case of a zero coupon bond, there is a single
cash flow, and the stated time to maturity is no different from the effective time to ma-
turity. However, in the case of a coupon paying bond, when we say that the maturity
is N years, we are essentially saying that the last cash flow has a maturity of N years.
The first cash flow, however, has a maturity of only six months. Thus Macaulay argued
that the effective term to maturity, or the weighted average of the terms to maturity of
the component cash flows, is less than the term to maturity of a coupon paying bond.
This weighted average time to maturity, which we now understand as duration, is the
appropriate measure of price sensitivity. In the case of a zero coupon bond, there is
no difference between the stated term to maturity and the effective term to maturity
or duration. However, in the case of coupon paying plain vanilla bonds, the effective
term to maturity is always shorter than the stated term to maturity.

Example 5.6. Consider 5-year and 10-year plain vanilla bonds, both with a face value of $1,000; a
coupon of 8% per annum paid semiannually; and a YTM of 8% per annum. Obviously both the bonds
trade at par. Now assume that the YTM of both the bonds changes to 10% per annum. The price of the
5-year bond declines to $922.7827, which corresponds to a percentage change of −7.7217%. On the
other hand, the price of the 10-year bond, declines to $875.3779, which corresponds to a percentage
change of −12.4622%. As can be seen, the 10-year bond is indeed more sensitive to a change in yield.
Now let’s consider 5-year and 10-year zero coupon bonds, both with a face value of $1,000.
When the YTM is 8% per annum, the respective prices are $675.5642 and $456.3869. When the
Factors Influencing Duration | 123

YTM increases to 10% per annum, the corresponding prices are $613.9133 and $376.8895. Thus the
5-year zero coupon bond declines in price by 9.1258%, and the 10-year zero coupon bond declines by
17.4189% in price.
Thus we see that the 10-year plain vanilla bond is more sensitive than the 5-year plain vanilla
bond, and the 10-year zero coupon bond is more sensitive than the 5-year zero coupon bond. How-
ever, the 5-year zero is more sensitive than the 5-year plain vanilla bond, and the 10-year zero is more
sensitive than the 10-year plain vanilla bond. The inference, therefore, is that although the price sen-
sitivity of a bond is related to its time to maturity, the maturity of the bond is not the sole influencing
variable.

Factors Influencing Duration

There are five primary influences on the duration of a plain vanilla bond. These are:
– Term to maturity
– Coupon
– Yield to maturity
– Accrued interest
– Coupon frequency

Term to Maturity

Holding the coupon rate constant, the duration of a bond generally increases with its
time to maturity. For par and premium bonds, duration always increases with the term
to maturity. In the case of bonds trading at a discount, the duration generally increases
with the term to maturity. However, there could be bonds trading at a substantial dis-
count from par, for which the duration can actually decrease with an increase in the
time to maturity.

Proof of the Relationship Between Duration and Time to Maturity

When there are N coupons remaining until maturity, the price of a bond is given by:

Mc 1 M
P0 = × [1 − ]+ (5.1)
y y N
(1 + 2 ) (1 + y2 )
N

When there are N + 1 coupons remaining until maturity, the price is given by:

Mc 1 M
P1 = × [1 − ]+ (5.2)
y y (N+1)
(1 + 2 ) (1 + y2 )
(N+1)

M c y
⇒ P1 − P0 = △P = ×[ − ] (5.3)
y N+1
(1 + 2 ) 2 2
124 | Chapter 5: Duration, Convexity, and Immunization

The formula for the duration of a bond when there are N semiannual periods to ma-
turity is
[Mc/2] × 1 [Mc/2] × 2 [Mc/2] × N M×N
D0 = y + 2
+−−−−−−− N
+ N
(5.4)
P0 (1 + 2 ) y
P0 (1 + 2 ) y
P0 (1 + 2 ) P0 (1 + y2 )
If we increase the number of coupon periods by one, the duration is given by:
c y c
P0 MN [ 2 − 2 ] + M (1 + 2 )
D1 = D0 × + (5.5)
P1 P1 (1 + y2 )
N+1

Substituting for P0 from Equation (5.1), we get

(N − D0 )M c y M (1 + c2 )
D1 = D0 + ×[ − ]+ (5.6)
P1 (1 + y2 )
(N+1) 2 2 P1 (1 + y2 )
(N+1)

Consider Equation (5.6). The first and the third terms are positive. For a plain
vanilla bond, the duration is positive and less than the term to maturity. That is, D0 > 0
and D0 < N. For a par bond the middle term is zero. Thus the duration increases with
the term to maturity. For premium bonds the middle term is positive. Hence the du-
ration once again increases as the time to maturity increases. For discount bonds the
middle term is negative. If the coupon rate is much lower than the yield to maturity,
or in other words the bond is trading at a substantial discount, this component may
result in a declining duration as the term to maturity increases.
The implication of this result is the following. A plain vanilla bond with an infinite
time to maturity is a perpetuity. A perpetuity has a duration of (1+y) y
, where y is the
periodic YTM. Thus for par and premium bonds, the duration steadily increases and
approaches this limit. In the case of discount bonds, if the duration declines with an
increase in the time to maturity, it has to subsequently reverse the direction of change
in order to reach the limiting value.
We have seen, that duration is generally related positively to a bond’s remaining
time to maturity. However, duration increases at a decreasing rate. For zero coupon
bonds, duration increases at a constant rate with respect to the time to maturity. This
is because for a zero coupon bond
𝜕D
D=N⇒ =1
𝜕N
Duration and term to maturity are generally related positively for two key reasons. One
is that the principal repayment is a large contributor to the bond price and is a major
influence on the duration. If this cash flow is postponed, then it pulls the duration
along with it. Second, the additional long-term coupons assist this process. Because,
long maturity bonds receive a considerable number of cash flows after a correspond-
ing shorter duration bond has matured.3

3 See Douglas [22].

Factors Influencing Duration | 125

Example 5.7. Consider three bonds with 5, 10, and 30 years, respectively, to maturity. Assume that
the coupon = yield = 10% per annum. Consider the 10-year bond first. It generates a total cash flow
of 20 × 50 + 1,000 = $2,000. Of this amount, $1,500 is received after 5 years. This is 75% of the total
cash flow. The 30-year bond generates a total cash flow of 60 × 50 + 1,000 = $4,000. Of this amount,
$3,500 is received after 5 years. This is 87.50% of $4,000. $3,000 is received after 10 years, which is
75% of the total cash flow.
Due to these factors, long maturity bonds tend to have a higher duration. However, duration in-
creases at a decreasing rate. This is because long-term cash flows of a fixed nominal amount are as-
signed progressively lower present values.

Coupon

If we keep the maturity and the yield constant, the higher the coupon rate is, the lower
the duration. Consider a bond with a face value of $1,000, 10 years to maturity, and a
YTM of 7%.

Table 5.3: Impact of higher coupons on the duration.

Coupon Duration
Rate

1% 9.341265
2% 8.827750
3% 8.416202
5% 7.797649
8% 7.177614
10% 6.885319
15% 6.386328
20% 6.071655

As you can be see, duration declines as the coupon increases, keeping the maturity
and yield constant. There are two reasons for this. First, higher coupon bonds have a
greater percentage of total cash flows occurring before the maturity date. This reduces
the relative influence of the principal repayment. Take the 10-year 10% coupon bond.
It generates a total cash flow of $2,000, of which 50% arises in the form of coupon pay-
ments. On the other hand, a 10-year bond with a coupon of 15% per annum generates
a total cash flow of $2,500 of which 60% is received in the form of coupon payments.
The second factor is that the discounting process has less effect on the earlier cash
flows, which consist exclusively of coupons, than on the later cash flows, which are a
combination of coupons plus principal. Thus larger earlier cash flows in the form of
coupons are assigned greater weights in present value terms.
126 | Chapter 5: Duration, Convexity, and Immunization

Yield to Maturity

The duration of a bond is inversely related to its yield to maturity. High-yield environ-
ments lead to low duration, whereas low-yield environments create high durations.
Let’s consider a bond with a face value of $1,000 and 30 years to maturity. The coupon
is 7% per annum paid semiannually. As you can see from Table 5.4, duration decreases
as the yield to maturity increases.

Table 5.4: Impact of higher yields on the duration.

Yield to Duration
Maturity

1% 19.00552
2% 17.97389
3% 16.92821
5% 14.85735
8% 12.00854
10% 10.38785
15% 7.409989
20% 5.630574

There are two reasons why duration and YTM are inversely related. First, duration is
based on the present value weights of the cash flows. As the discount rate increases the
impact is higher on later cash flows. Thus we are assigning relatively lower weights to
long-term cash flows and higher weights to short-term cash flows. The consequence
is a decline in the duration. The second factor is that as the discount rate increases
the present value of the terminal principal amount declines disproportionately, which
drives down its relative contribution to the price of the bond.
Also newly issued bonds have a coupon that is close to the prevailing YTM. That
is, high yield environments lead to the issue of high coupon bonds, while low yield
environments result in the issue of low coupon bonds. Thus in a high yield environ-
ment the coupon component of a newly issued bond’s market value is even higher,
which makes it very sensitive to changes in the yield.

Accrued Interest

A bond’s duration is inversely related to the amount of accrued interest present in the
dirty price.4 The duration computation is based on dirty prices and not clean prices.
Thus accrued interest has an impact on duration. Accrued interest is an investment
with a duration of zero. The forthcoming coupon payment reimburses the bond holder

4 See Douglas [22].

Percentage Price Change and Duration | 127

for the accrued interest he pays up front. Thus a bond with a higher accrued inter-
est has a shorter duration than a similar bond with a lower accrued interest. That is,
the accrued interest component of a bond’s dirty price pulls down the duration. On
a coupon date, the accrued interest reverts to zero. Consequently there is a jump in
the duration because the dirty price no longer includes the accrued interest which is
a zero duration component.

Example 5.8. Consider a T-bond with a face value of $1,000, a coupon and yield of 7% per annum, and
20 years to maturity. Let’s assume that the bond has been issued on 15 July 2018 and that we are on
14 January 2019. The duration is 10.5540 years. The next day, 15 January 2019, after the coupon has
been paid, the duration jumps to 10.9205.

Coupon Frequency

As we have seen, duration is inversely related to accrued interest. The impact of ac-
crued interest is more pronounced in the case of bonds which pay coupons less fre-
quently. For instance, a bond that pays coupons annually has a greater buildup of
accrued interest than a bond that pays semiannual coupons. This is because in the
case of the former, a full year’s coupon has to be accrued, whereas in the case of the
latter a maximum of half the annual coupon is accrued.

Example 5.9. Let’s reconsider the T-bond with a coupon and yield of 7%, and 30 years to maturity.
Assume that the bond has been issued on 15 July 2018 and that we are on 14 July 2019. If we also
assume semiannual coupons, the duration is 12.3460 years, whereas if we assume annual coupons,
the duration is 12.2805 years. On the next day, 15 July 2019, the semiannual coupon-paying bond has
a duration of 12.7752 years, and the bond with annual coupons has a duration of 13.1371 years.
As we can see, the jump in duration on the coupon date is significantly higher for the bond paying
annual coupons. Also it has a higher duration on the coupon date than the bond with semiannual
coupons. This is because in the case of the annual coupon-paying bond, the entire year’s coupon is
received at the end of the year, whereas in the case of the bond paying half yearly coupons, half the
coupon is received six months prior.

Percentage Price Change and Duration

Consider a coupon bond that pays coupons m times per year, and assume that it has
T years to maturity. Let the annual YTM be y, the annual coupon be C, and the face
value be M.5

5 We are using the symbol y to denote the periodic, usually semiannual, YTM in certain cases, and
the annual YTM in other instances. To avoid confusion, we are explicitly restating the meaning, every
time we redefine the variable.
128 | Chapter 5: Duration, Convexity, and Immunization

mT C
m M
P=∑ y t
+ (5.7)
y mT
t=1 (1 + m
) (1 + m
)

Therefore,6
dP 1 1 1
× =− × ×D (5.8)
dy P m (1 + my )

Thus, for a bond that pays semiannual coupons

dP 1 1 1
× =− × ×D (5.9)
dy P 2 (1 + y2 )

And
dP 1 1
=− ×[ × D] × dY (5.10)
P 2 (1 + y2 )

Because duration is expressed in semiannual terms, 21 × D is the duration in annual

terms. The annual duration divided by one plus the periodic yield is referred to as the
modified duration. Thus
1 1
Dm = ×[ × D]
2 (1 + y2 )
dP
⇒ = −Dm × dY (5.11)
P
The percentage change in the price of a bond is directly proportional to its modified
duration. dP
P
is the limiting value of the percentage change in price. For a finite yield
change, the percentage change in price is △PP
. As △y → 0, △P
P
→ dPP
.

Duration of Annuities Due and Perpetuities Due

For an annuity due,
1 N
Duration = − (5.12)
r [(1 + r)N − 1]
For a perpetuity due,
1
Duration = (5.13)
r
One may think that a perpetuity due has the same duration as an identical perpe-
tuity because the first cash flow occurs at time 0. However, because of the additional
cash flow at the outset, a perpetuity due has a higher price, and consequently a lower
duration.

6 See Appendix 5.3.

Dollar Duration | 129

Dollar Duration

The product of the modified duration and the price of the bond is referred to as the
dollar duration of the bond. In Example 5.1, the duration was 4.3853 years, the semi-
annual yield was 3.25%, and the price was $978.9440. Thus the dollar duration is

4.3853
× 978.9440 = 4,157.8335
1.0325

Computing Duration and Modified Duration with Excel

Assume we are on 15 May 2018. There is a five-year bond with a face value of $1,000,
which pays a coupon of 6% per annum semiannually, and the YTM is 6.50% per an-
num. To calculate the duration, we use the DURATION function in Excel. The required
parameters are
– Settlement: We need to give it using the DATE function with the YYYY,MM,DD for-
mat. In this example it is DATE(2018,05,15).
– Maturity: We need to give it using the DATE function. We know that it is a
five-year bond, so we give the same date five years later. Thus in our case it is
DATE(2023,05,15).
– Coupon: We need to give the annual coupon. In this case it is 0.06.
– Yield: We need to give the annual YTM. In this case it is 0.065.
– Frequency: Because the bond pays coupons on a semiannual basis, the frequency
is 2.
– Basis: This captures the day-count convention. This bond has an integer number
of years to maturity. Thus the basis can be specified as any value. We give it as 0.

The duration = DURATION(Date(2018,05,15),Date(2023,05,15),0.06,.065,2,0) = 4.3853

Note that since we are giving the frequency of coupons as a variable, both the
coupon rate and the YTM should be stated in annual terms. As a consequence, the
duration is directly derived in terms of the number of years.

Modified Duration

The required function is MDURATION. The parameters are identical. The answer is
4.2472. We can verify it manually:

4.3853
= 4.2472
1.0325
130 | Chapter 5: Duration, Convexity, and Immunization

Approximating Duration
The modified duration of a bond can be approximated using the prices obtained by
changing the annual YTM by △y in both directions. Before we proceed, let’s define
the required symbols:
– P0 ≡ the current dirty price
– P− ≡ the dirty price when the YTM declines by △y
– P+ ≡ the dirty price when the YTM increases by △y

P− −P0
The percentage price change when the YTM declines by △y is P0
.
P −P
Thus (P −×△y)
0
gives us an approximation for the rate of percentage change in price
0
with respect to yield.
P −P+
We obtain a similar expression when we increase the YTM by △y. That is (P 0×△y) .
0
The two expressions are not identical because the impact of a yield decline of x
basis points is not the same as that of a yield increase of the same magnitude. So the
best option is to take a simple arithmetic average:

1 P − P0 P − P+ P − P+
×[ − + 0 ]= −
2 (P0 × △y) (P0 × △y) (2P0 △ y)

△P
P△y
is an estimate of the modified duration.

Example 5.10. Consider a bond with a face value of $1,000 and five years to maturity. The coupon is
8% per annum and the yield is 10% per annum. The coupon is paid on a semiannual basis.
If we use the MDURATION function in Excel, we get a value of 3.9808. Let’s compare it with the
approximate value. Assume the change in the annual YTM is 25 basis points and P0 = $922.7827,
P− = $932.0228, P+ = $913.6554:

932.0228 − 913.6554
(2 × 922.7827 × 0.0025)
= 3.9809

The approximation is excellent.

The Concept of Effective Duration

In the case of a plain vanilla bond, the occurrence of a yield change has no impact
on the cash flows from the bond. But there exist other fixed income securities, such
as bonds with built-in call or put options, and mortgage-backed securities, where a
change in yield may have consequences for the cash flows from the security. The ap-
proximate duration formula that we have derived can be used in these situations and
is termed as the effective duration.
Duration as a Center of Gravity | 131

Duration as a Center of Gravity

The duration of a plain vanilla bond may be interpreted as its center of gravity, as the
following argument demonstrates:7
c
1 N M2 ×t MN
Duration = [∑ + ]
P t=1 (1 + y )t (1 + y )N
2 2
N M c2 × t MN
⇒ DP = [∑ y t
+ N
] (5.14)
t=1 (1 + 2
) (1 + y2 )
N M c2 M
P = [∑ y t
+ N
]
t=1 (1 + 2 ) (1 + y2 )
N Mc ×D
M×D
⇒ DP = [∑ 2 y t + N
] (5.15)
t=1 (1 + 2 ) (1 + y2 )

If we subtract Equation (5.15) from Equation (5.14), we get

N M c2 × (t − D) M(N − D)
0 = [∑ y t
+ N
]
t=1 (1 + 2
) (1 + y2 )

Portfolio Duration
Consider a portfolio of N bonds. To compute the portfolio IRR, we need to project all
the cash flows until the maturity of the bond with the longest time until maturity.
The initial investment is equal to the sum of the prices of the components of the port-
folio. Using the vector of cash flows, we can compute the IRR and consequently the
duration.
The portfolio duration can be approximated in the following way. First we can
find the weights of the bonds, where the weight of bond i is Pi ÷ ∑Ni=1 Pi . We can com-
pute the durations of the component bonds using the respective YTMs, and compute
a weighted average duration. However, if we want to get the precise portfolio duration
in the form of a weighted average, we need to do the following. We need to recompute
the prices of all the components using the portfolio IRR. Let’s term these the portfolio
prices. The weight attached to a bond is P ∗ i ÷ ∑Ni=1 P ∗ i , where P ∗ is the portfolio price.
Using these weights we can compute a weighted average of the duration of the compo-
nents of the portfolio, where the discount rate is the portfolio IRR. In this case, we get
the same duration, as calculated using the portfolio cash flows and the corresponding
IRR. Note that although the portfolio price of a bond is different from its actual price,

7 See Taggart [64].

132 | Chapter 5: Duration, Convexity, and Immunization

the sum of the portfolio prices for all the components of the portfolio is equal to the
sum of the actual prices, which is obviously the initial investment. The portfolio IRR is
a non-linear average of the individual YTMs. When we use the portfolio IRR to reprice
a component of the portfolio, its price increases if its YTM is higher than the portfolio
IRR. However the recomputed price is lower if the YTM of the bond is lower than the
portfolio IRR.
Here is an example.

Example 5.11. Consider three bonds, Alpha, Beta, and Gamma. Alpha has a term to maturity of three
years; Beta has a maturity of four years; and Gamma has five years till maturity. All three bonds have
a face value of $1,000 and pay coupons on a semiannual basis. Alpha carries a coupon of 6% per an-
num and a YTM of 10% per annum; Beta carries a coupon of 8% per annum and a YTM of 12% per an-
num; and Gamma carries a coupon of 10% per annum and a YTM of 8% per annum. Their respective
prices are $898.4862, $875.8041, and $1,081.1090. The durations of the bonds, using their respec-
tive YTMs, as computed using the DURATION function in Excel, are 2.7761 years, 3.4605 years, and
4.0954 years. The weighted average using the bond prices is 3.4855 years or 6.9711 semiannual peri-
ods.
Now let’s compute the portfolio duration directly. The portfolio IRR is 9.77% per annum. Using
this IRR and the vector of cash flows for the portfolio, the duration is 6.9260 half-years. The data is
depicted in Table 5.5. The approximation using the weighted average of the individual durations is
therefore reasonably close.

Table 5.5: Computing portfolio duration.

Time Portfolio PVCF Weight

Cash Flow

0 (2,855.3992) (2,855.3992)
1 120 114.4108 0.0401
2 120 109.082 0.0382
3 120 104.0013 0.0364
4 120 99.1573 0.0347
5 120 94.5389 0.0331
6 1120 841.2657 0.2946
7 90 64.4531 0.0226
8 1090 744.2406 0.2606
9 50 32.5494 0.0114
10 1050 651.7003 0.2282
PVCF stands for present value of the cash flow

Now let’s recompute the prices of the three bonds, using the portfolio IRR as the yield for each. Bond
Alpha has a price of $903.9615; Bond Beta has a price of $942.5222; and Bond Gamma has a price
of $1,008.9155. Let’s call these the portfolio prices. As expected, the two bonds with a YTM greater
than the portfolio IRR have increased in price. However, the third bond, whose YTM was lower than
Bond Convexity | 133

the portfolio IRR, has declined in price. It can be verified that the sum of the original prices is equal to
the sum of the portfolio prices, which in this case is equal to $2,855.3992. The durations of the three
bonds, as computed using the portfolio IRR as the discount rate, are 2.7769, 3.4833, and 4.0587. We
know that as the YTM increases, the duration declines. In the case of Bonds Alpha and Beta, the YTMs
are higher than the portfolio IRR. Consequently the durations computed using the respective YTMs are
lower than the values obtained using the portfolio IRR as the discount rate. However, in the case of
the third bond, the original YTM is lower than the portfolio IRR. Consequently, the duration computed
using the original YTM is greater than the duration obtained by using the portfolio IRR as the discount
rate.

903.9615 942.5222 1,008.9155

× 2.7769 + × 3.4833 + × 4.0587
2,855.3992 2,855.3992 2,855.3992
= 3.4630 years or 6.9260 half-years

Thus, the weighted average of the durations computed, using the portfolio IRR as the discount rate
and the portfolio prices as the weights, is exactly the value obtained by discounting the portfolio cash
flows using the portfolio IRR.

Bond Convexity

Figure 5.2: The slope of the price-yield curve is the dollar duration.

Dollar duration is the first derivative of the price-yield function. Convexity is the
derivative of the dollar duration with respect to the yield, divided by the dirty price of
the bond. Dollar duration is the slope of the price yield curve at a specified point (see
Figure 5.2). Convexity, in approximate terms, measures the gap between the tangent
line and the price-yield curve. Convexity can be viewed as an approximate measure of
the difference between the actual bond price and the price predicted by the tangent
line. Convexity enhances a bond’s performance in both bull and bear markets, but
not in a uniform fashion. For plain vanilla bonds, the convexity is always positive.
134 | Chapter 5: Duration, Convexity, and Immunization

That is, the price-yield curve always lies above the dollar duration tangent line. The
convexity effect becomes greater with larger changes in yield. That is, the gap between
the tangent line and the price yield curve increases with an increase in the magnitude
of the yield. Dollar duration is a good estimate for small yield changes, but loses its
predictive power when yield changes are large. A bond’s duration increases in a bull
market as yields fall. This enhances the price gain. The duration reduces in a bear
market, as yields rise, which mitigates the price decline.
For a plain vanilla bond, we know that the price is given by:

Mc 1 M
P= [1 − ]+ (5.16)
y (1 + y)N (1 + y)N
dP Mc Mc McN NM
=− 2 + + −
dY y 2
y (1 + y)N
y(1 + y)N+1
(1 + y)N+1
d2 P 2Mc 2Mc NMc
2
= 3 − −
dy y 3
y (1 + y) N
y (1 + y)N+1
2

McN McN(N + 1) MN(N + 1)

− N+1
− +
2
y (1 + y) y(1 + y)N+2 (1 + y)N+2
2Mc 1 2NMc N(N + 1) Mc
= 3 [1 − ]− + [M − ] (5.17)
y (1 + y) N 2
y (1 + y)N+1
(1 + y)N+2 y

where y represents the semiannual YTM and c is the semiannual coupon. N is the
number of coupons remaining. The convexity of a bond is defined as

d2 P 1
×
dy2 P

The convexity in terms of the number of years for a bond that pays m coupons per year
is given by PeriodicmConvexity
2 .
CF t ×t
duration of a bond in periodic terms is expressed as ∑Nt=1 P×(1+y)t
.
CF t ×t
modified duration is expressed as ∑Nt=1 P×(1+y)t+1
.
CF ×t
dollar duration is stated as ∑Nt=1 (1+y)t t+1 .
If we differentiate the dollar duration with respect to the yield, and divide by the
dirty price, we get the convexity. Thus convexity is given by
N
CF t × t × (t + 1)
∑ t+2
t=1 P × (1 + y)

Example 5.12. Let’s consider the five-year T-note with a price of 978.9440. The coupon is 6% per an-
num, and the YTM is 6.50% per annum.
We need to compute:

N
CF t × t × (t + 1)
∑
t=1 P × (1 + y)t+2
Approximating the Price Change of a Bond for a Given Change in Yield | 135

Table 5.6: Computation of convexity.

Time Cash Flow wt t(t + 1) t(t + 1) × wt

1 30 0.029681 2 0.059361
2 30 0.028746 6 0.172478
3 30 0.027842 12 0.334098
4 30 0.026965 20 0.539303
5 30 0.026116 30 0.783492
6 30 0.025294 42 1.062362
7 30 0.024498 56 1.371895
8 30 0.023727 72 1.708344
9 30 0.02298 90 2.068213
10 1,030 0.76415 110 84.05653
Sum 92.15608
The symbol wt stands for the weight.

92.15608
The convexity in semiannual terms is = 86.4458. The convexity in annual terms is therefore
(1.0325)2
21.6115.

Approximating the Price Change of a Bond for a Given Change in

Yield
The change in the price of a bond may be expressed as8

𝜕P 1 𝜕2 P
dP = × dY + × (dY)2 + h.o.t (5.18)
𝜕y 2 𝜕y2

Thus,

dP 𝜕P dY 1 𝜕2 P (dY)2 h.o.t
= × + × + (5.19)
P 𝜕y P 2 𝜕y2 P P

The first term of Equation 5.19 captures the approximate percentage price change due
to duration. The second term of Equation 5.19 captures the contribution of convexity
to the price change. This may be measured as

1
× Convexity × (dY)2
2

Example 5.13. Consider a 50 basis points increase in the annual YTM of the five-year T-note,with a
coupon of 6% per annum and a YTM of 6.50% per annum. The face value is $1,000, and coupons are

8 h. o. t. stands for higher order terms.

136 | Chapter 5: Duration, Convexity, and Immunization

paid semiannually. The current price, as we have seen, is $978.9440. The new price will be $958.4170.
The exact price change is

958.4170 − 978.9440 = −20.5270

The price change due to duration is

−4.2473 × 978.9440 × 0.005 = −20.7892

The price change due to convexity is

0.5 × 21.6115 × 978.9440 × (0.005)2 = 0.2645

Thus the approximate price change due to both duration and convexity is:

−20.7892 + 0.2645 = −20.5247

As you can see, the combination of duration and convexity does a much better job of
predicting the price change, compared to duration alone. One may wonder why we
do not use third and higher order terms. The rationale is that if we have a third-order
term, it would be multiplied by (△Y)3 . Thus if the change in yield is 50 basis points,
we multiply by 1.25 × 10−7 . This obviously is trivial from the standpoint of attaining
precision.

Dispersion of a Bond
The dispersion of a bond is defined as9

1 N CF t × t 2 2
Dispersion = ×∑ − (Duration) (5.20)
P t=1 (1 + y)t

The dispersion of a zero coupon bond is equal to zero. This is because it gives rise
to a single cash flow.
We know that convexity is defined as
N
1 CF t × t(t + 1)
Convexity = [∑ ] (5.21)
P(1 + y) 2
t=1 (1 + y)t
N
1 CF t × (t 2 + t)
⇒ Convexity = × [∑ ]
P(1 + y) 2
t=1 (1 + y)t
N
1 CF t × t CF t × t 2
= [∑ t
+ ]
P(1 + y) 2
t=1 (1 + y) (1 + y)t

9 We are using the symbol y to denote the semiannual YTM.

Convexity of a Zero Coupon Bond | 137

1 2
⇒ Convexity = [Dispersion + (Duration) + Duration] (5.22)
(1 + y)2

Thus, for a given level of duration, the lower the dispersion is, the lower the con-
vexity. Hence for a given level of duration, zero coupon bonds have the lowest convex-
ity. For a given level of dispersion, the higher the duration, the greater the convexity.

Example 5.14. Consider a bond with a coupon of 6% per annum, paid semiannually, and a YTM of
6.50% per annum. Assume that there are five years to maturity. The dispersion may be calculated as
illustrated in Table 5.7.

Table 5.7: Computation of dispersion.

Time Cash Flow PVCFt t2 wt wt × t 2

1 30 29.28894 1 0.029681 0.029681

2 30 28.36702 4 0.028746 0.114986
3 30 27.47411 9 0.027842 0.250574
4 30 26.60930 16 0.026965 0.431443
5 30 25.77172 25 0.026116 0.65291
6 30 24.96051 36 0.025294 0.910596
7 30 24.17483 49 0.024498 1.200408
8 30 23.41387 64 0.023727 1.518528
9 30 22.67688 81 0.022980 1.861392
10 1,030 754.06560 100 0.764150 76.41503
PVCF stands for present value of the cash flow.
The symbol wt stands for the weight.

The sum of the last column is 83.38554. If we subtract the square of the duration, in semiannual terms,
we get a value of 6.4633.
Dispersion + Duration2 + Duration = 6.4633 + 76.9217 + 8.7705 = 92.1561.

92.1561
= 86.4458
(1.0325)2

which is the convexity in periodic terms.

Convexity of a Zero Coupon Bond

A zero coupon bond has a duration of N, in periodic terms, and a dispersion of zero.
Thus the convexity of a zero coupon bond in semiannual terms is:

N(N + 1)
2
where y is the annual YTM
(1 + y2 )
138 | Chapter 5: Duration, Convexity, and Immunization

Example 5.15. Consider a zero coupon bond with five years to maturity. The YTM is 8% per annum. The
convexity in periodic terms is
10 × 11
= 101.7012
(1.04)2
The convexity in annual terms is
101.7012
= 25.4253
4

Dispersion as an Expected Value

A bond’s dispersion is the variance of the times to receipt of all of its cash flows,
whereas its duration is the mean of the times to receipt of all the cash flows. For a
variable X, the mean is E(X), while the variance is E[X 2 ] − [E(X)]2 . Thus duration is a
measure of the mean, whereas dispersion is a measure of the variance.
The dispersion of a plain vanilla bond measures how spread out in time the pay-
ments are relative to the duration. The more spread out the cash flows are relative to
the mean time to payment (the duration), the greater the dispersion. For a given dura-
tion, the higher the dispersion is the greater the convexity. For plain vanilla bonds, the
duration and the dispersion are non negative. Consequently the convexity is positive.

Portfolio Convexity and Dispersion

Let’s consider the three-bond portfolio that we examined in Example 5.11. In periodic
terms, the portfolio convexity is 55.4231 and the dispersion is 6.0747. We have used
the portfolio IRR to compute these statistics. Using the portfolio IRR as the discount
rate, Bond Alpha has a price of $903.9616, a convexity of 34.4230, and a dispersion of
1.4694; Bond Beta has a price of $942.5222, a convexity of 54.3085, and a dispersion
of 4.2431; and Bond Gamma has a price of $1008.9155, a convexity of 75.2801 and a
dispersion of 8.8040.
The weighted average convexity is
903.9616 942.5222 1,008.9150
× 34.4230 + × 54.3085 + × 75.2801 = 55.4231
2855.3992 2855.3992 2855.3992
The weighted average dispersion is
903.9616 942.5222 1,008.9150
× 1.4694 + × 4.2431 + × 8.8040 = 4.9765
2855.3992 2855.3992 2855.3992
As we can see, the portfolio convexity is equal to a weighted average of the convexities
of the component bonds. However, unlike duration and convexity, portfolio dispersion
is not a weighted average of the dispersions of the component bonds. This is consistent
with the statistical result that although the mean of a sum of variables is equal to the
sum of the individual means, the variance is not a sum of the individual variances.
Properties of Convexity | 139

Properties of Convexity
From Equation (5.22) we can see that the primary factors that influence a bond’s con-
vexity are the duration and the dispersion of the cash flows. Thus, keeping duration
constant, the greater the dispersion is, the higher the convexity.

The Impact of Duration

Convexity is positively related to the duration of the bond. Long duration bonds carry
higher convexities than bonds with a shorter duration. The convexity of a bond is not
only positively related to its duration, it is also an increasing function of the duration.
That is, a bond with twice the duration of another has more than double the convexity.
1 2
Convexity = 2
[Dispersion + (Duration) + Duration]
(1 + y)
𝜕Convexity 1
= [2 × Duration + 1]
𝜕Duration (1 + y)2
As the YTM of a bond increases, its duration decreases. Because duration and con-
vexity are positively related, the convexity of the bond decreases. Thus plain vanilla
bonds are said to exhibit positive convexity. Similarly, as the coupon increases, the du-
ration decreases and so does the convexity. And, the higher the accrued interest, the
lower the duration and consequently the lower the convexity. The duration of a bond
generally increases with its time to maturity, with the exception of certain bonds trad-
ing at a deep discount. Hence the convexity of a bond also generally increases with its
time to maturity.

The Irrelevance of the Face Value

The duration, convexity, and dispersion of a bond are independent of the face value.
This is because the face value appears in every term of the numerator of these three
measures, and it also appears in the denominator in the form of the bond price. For
the same reason, the durations of annuities and perpetuities, are also independent of
the size of the periodic cash flow.

Dollar Convexity
A bond’s convexity multiplied by its dirty price is termed its dollar convexity. We know
that the price change attributable to convexity is given by
1
× Convexity × P × (dY)2
2
140 | Chapter 5: Duration, Convexity, and Immunization

1
= × Dollar Convexity × (dY)2
2
Thus dollar convexity is a measure of the second derivative of the price-yield relation-
ship.

Approximate Convexity
1 𝜕2 P
We know that convexity may be expressed as P
× 𝜕y2

△P− P− − P0 △P+ P0 − P+
= and =
△y △y △y △y
P− +P+ −2P0
Thus the rate of change of △P
with respect to the yield is given by .
△y △y2
If we divide this by P0 , we get the expression for the approximate convexity:
P− + P+ − 2P0
Convexity = (5.23)
P0 △ y 2

Example 5.16. Consider a bond with five years to maturity and a face value of $1,000, paying semian-
nual coupons. The annual coupon rate is 8% per annum and the YTM is 10% per annum. For a 25-basis
point change in the annual YTM

P0 = $ 922.7827; P− = $ 932.0228; P+ = $913.6554

P− + P+ − 2P0 = 0.1128; P0 △ y 2 = 0.0057673

Thus the approximate convexity is 19.5585. The exact value is 19.5736.

Convexity of Annuities and Perpetuities

Let r be the periodic (usually semiannual) yield, and A the periodic cash flow. The
convexity of an annuity is
2 2N N(N + 1)
− −
r 2 r(1 + r) [(1 + r)N − 1] (1 + r)2 [(1 + r)N − 1]

and that of an annuity due is10

2 2N N(N + 1)
− −
r 2 (1+ r) r(1 + r) [(1 + r) − 1] (1 + r)2 [(1 + r)N − 1]
2 N

In Example 5.17 we compute the convexity using the discounted cash flow approach
as well as the concise formulas.

10 See Appendix 5.3 for both derivations.

Convexity of Annuities and Perpetuities | 141

Example 5.17. Consider an annuity and an annuity due, both of which pay a cash flow of $50 for four
periods. The periodic discount rate is 4%. From the concise formulas, the convexity of the annuity is
2 2×4 4×5
− −
0.04 × 0.04 0.04 × 1.04[(1.04)4 − 1] (1.04)2 [(1.04)4 − 1]
= 1,250.0000 − 1,132.1643 − 108.8620 = 8.9744

The convexity of the annuity due is

2 8
= −
0.04 × 0.04 × 1.04 0.04 × (1.04)2 [(1.04)4 − 1]
4×5
−
(1.04)2 [(1.04)4 − 1]
= 1,201.9231 − 1,088.6189 − 108.8620 = 4.4423

The calculations, cash flow by cash flow, are demonstrated in Tables 5.8 and 5.9.

Table 5.8: Computation of convexity for an annuity.

Time C. Flow PVCF Weight wt × t × (t + 1)

1 50 48.0769 0.2649 0.5298

2 50 46.2278 0.2547 1.5282
3 50 44.4498 0.2449 2.9389
4 50 42.7402 0.2355 4.7098
SUM 181.4948 SUM 9.7067
C. Flow stands for cash flow
PVCF stands for present value of the cash flow
wt stands for the weight

9.7067
Convexity = = 8.9744
(1.04)2
Now let’s consider the annuity due.

Table 5.9: Computation of convexity for an annuity due.

Time C. Flow PVCF Weight wt × t × (t + 1)

0 50 50.0000 0.2649 0.0000

1 50 48.0769 0.2547 0.5094
2 50 46.2278 0.2449 1.4695
3 50 44.4498 0.2355 2.8259
SUM 188.7546 SUM 4.8048
C. Flow stands for cash flow
PVCF stands for present value of the cash flow
wt stands for the weight

4.8048
Convexity = = 4.4423
(1.04)2
142 | Chapter 5: Duration, Convexity, and Immunization

Perpetuities

The convexity of a perpetuity is

2 2
= = 1,250
r 2 0.04 × 0.04

The convexity of a perpetuity due is

2 2
= = 1,201.923
(r 2 )(1 + r) 0.04 × 0.04 × 1.04

Thus a perpetuity due has a lower convexity than a perpetuity.

Immunization of a Bond Portfolio

A pension fund has promised to pay a return of 8.40% per annum compounded an-
nually on an initial investment of $5 million, eight years from the date of investment.
If it were to deposit the corpus in a fixed income security, it is exposed to two types
of risk. The first is reinvestment risk, or the risk that cash flows received at intermedi-
ate stages may have to be invested at lower rates of interest. The second risk, which is
termed price risk or market risk, is the risk that rates could increase, causing the price
of the bond to decline at the end of the investment horizon.
The two risks obviously work in opposite directions. The issue is therefore, can
we find a security that ensures that the terminal cash flow is adequate to satisfy the
liability, irrespective of whether rates decline or increase. The process of protecting a
bond portfolio against a change in the interest rate is termed bond immunization. Let’s
consider a simple immunization strategy where we have to immunize a portfolio that
is held in order to satisfy a single liability. It turns out that there are four conditions
that need to be satisfied in such cases. The first is that the present value of the liability
should be equal to the amount invested in the fixed income security at the outset. Sec-
ond, the duration of the security should be equal to the investment horizon.11 Third,
there should be a parallel shift in the yield curve. Finally, if there is a yield change,
then it should be a one-time change, at the very outset.

Example 5.18. Consider a bond with a face value of $1,000 and 12 years to maturity. Assume that the
coupon, paid on an annual basis, is equal to the market yield of 8.40% per annum. The bond obviously
trades at par and can be shown to have a duration of 8 years. Consider the impact of a one-time change
in the interest rate right at the outset. Also consider increments and decrements in multiples of 20
basis points from the prevailing rate of interest. Because the liability is $5 million, and the price of the

11 See Appendix 5.4 for a derivation.

Immunization of a Bond Portfolio | 143

bond is $1,000, we need to invest in 5,000 bonds. The terminal cash flow after 8 years, corresponding
to each of the assumed interest rates, is summarized in Table 5.10.

Table 5.10: Illustration of single-period immunization.

Interest Future Selling Total Cash CF for 5,000 Amount Deviation

Rate Value of Price Flow Bonds Due
Coupons

0.094 939.92 967.89 1,907.81 9,539,032.94 9,532,444 6,588.95

0.092 933.13 974.20 1,907.33 9,536,629.29 9,532,444 4,185.30
0.090 926.39 980.56 1,906.95 9,534,767.40 9,532,444 2,323.41
0.088 919.71 986.98 1,906.69 9,533,448.42 9,532,444 1,004.43
0.086 913.07 993.46 1,906.53 9,532,673.53 9,532,444 229.54
0.084 906.49 1,000 1,906.49 9,532,444.00 9,532,444 0.00
0.082 899.96 1,006.59 1,906.55 9,532,761.07 9,532,444 317.09
0.080 893.48 1,013.25 1,906.73 9,533,626.14 9,532,444 1,182.16
0.078 887.05 1,019.96 1,907.01 9,535,040.58 9,532,444 2,596.60
0.076 880.67 1,026.73 1,907.40 9,537,005.85 9,532,444 4,561.86
0.074 874.34 1,033.57 1,907.90 9,539,523.44 9,532,444 7,079.45

As you can be see, the income from reinvested coupons steadily increases with the
interest rate whereas the sale price steadily decreases with the interest rate. When the
rate does not change, the terminal cash flow is exactly adequate to meet the liability. In
all other cases, there is a surplus. In other words, the terminal cash flow from the bond
is adequate to meet the liability, irrespective of the change in the interest rate. Thus if
the pension fund invests in an asset whose duration is equal to the time to maturity
of the liability, then the terminal inflow is always adequate to meet the contractual
outflow.
To immunize the portfolio as we have just demonstrated, the following conditions
need to be satisfied:
1. The amount invested in the bond must be equal to the present value of the liability.
2. The duration of the bond must be equal to the maturity of the liability.
3. There must be a one-time change in interest rates, right at the outset.
4. There must be a parallel shift in the yield curve.

A Point on Long-Dated Treasury Bonds

Treasury bonds are issued with a maturity of up to 30 years. These bonds have high du-
rations and high convexities. Consequently, they are sought after by traders in search
of convexity. Pension fund managers, who have long term liabilities, need bonds with
high durations to construct immunized portfolios. Therefore, long-dated treasuries are
popular with such market participants.
144 | Chapter 5: Duration, Convexity, and Immunization

Chapter Summary
In this chapter, we introduced the concept of duration, and showed that the price
change of a bond due to a change in its YTM is proportional to its duration. Thus dura-
tion, and not the time to maturity, is an appropriate measure of interest rate sensitivity.
We showed how to compute the duration by discounting each cash flow, by using a
concise formula, and by using the DURATION function in Excel. We derived the du-
ration of annuities, annuities due, perpetuities, and perpetuities due. We studied the
properties of duration and examined the impact of the variables that influence the
price of a bond on its duration. We also demonstrated the computation of the approx-
imate or effective duration of a bond. We then looked at the concepts of convexity and
dispersion. We studied the properties of convexity, and examined the impact of the
variables that influence the price of a bond on its convexity. Then, we derived closed-
form expressions for the convexities of annuities, annnuities due, perpetuities, and
perpetuities due. We also examined the duration, convexity, and dispersion of a port-
folio of bonds. Finally, we examined the the immunization of a bond portfolio. In this
context, we stated the required conditions to be satisfied for single-period immuniza-
tion to be a success.
In the next chapter, we study the market for short-term debt securities, which are
termed money market securities.

Appendix 5.1: Derivation of a Concise Formula for Duration

The price of a bond on a coupon date can be expressed as
C
2 1 M
P= y [1 − ]+
y N N
2 (1 + 2
) (1 + y2 )
Let’s redefine the following variables:
y ≡ y2 ≡ the semiannual YTM
c ≡ c2 ≡ the semiannual coupon rate
The preceding expression for the price can then be written as
Mc 1 M
P= [1 − ]+
y (1 + y) N
(1 + y)N
Mc Mc M
= − +
y y(1 + y) N
(1 + y)N
M[c{(1 + y)N − 1} + y]
=
y(1 + y)N
We know that duration is defined as
N
CF t × t
∑ y t
t=1 P × (1 + 2 )
Appendix 5.1: Derivation of a Concise Formula for Duration | 145

In terms of our redefined variables, it can be expressed as

N
CF t × t
∑
t=1 P × (1 + y)t

We already have an expression for the price. Let’s now derive an expression for

N
CF t × t Mc 2Mc 3Mc NMc NM
∑ t
= + + + ⋅⋅⋅+ +
t=1 (1 + y) (1 + y) (1 + y)2
(1 + y)3
(1 + y)N
(1 + y)N
1 2 N NM
= Mc [ + + ⋅⋅⋅ + ]+
(1 + y) (1 + y)2 (1 + y)N (1 + y)N

Let’s define:

1 2 N
S=[ + + ⋅⋅⋅ + ]
(1 + y) (1 + y)2 (1 + y)N
2 N
S(1 + y) = [1 + + ⋅⋅⋅ + ]
(1 + y) (1 + y)N−1
1 1 1 N
S(1 + y) − S = yS = [1 + + + ⋅⋅⋅ + ]−
(1 + y) (1 + y)2 (1 + y)N−1 (1 + y)N
1 1 1 1 N
⇒S= [1 + + + ⋅⋅⋅ + ]−
y (1 + y) (1 + y)2 (1 + y)N−1 y(1 + y)N

Let’s define:

1 1 1
Z = [1 + + + ⋅⋅⋅ + ]
(1 + y) (1 + y)2 (1 + y)N−1
1 1
Z(1 + y) = [1 + y + 1 + + ⋅⋅⋅ + ]
(1 + y) (1 + y)N−2
(1 + y) 1
⇒Z= −
y y(1 + y)N−1

Thus,

1 (1 + y) 1 N
S= [ − ]−
y y y(1 + y)N−1
y(1 + y)N

Therefore,

N
CF t × t (1 + y) 1 N NM
∑ t
= Mc { − − }+
t=1 (1 + y) y2 y2 (1 + y)N−1 y(1 + y)N (1 + y)N
146 | Chapter 5: Duration, Convexity, and Immunization

Thus,
N
CF t × t
∑ t
t=1 P × (1 + y)
c(1+y)[(1+y)N −1]−Ny(c−y)
y2 (1+y)N
=
c[(1+y)N −1]+y
y(1+y)N
(1 + y) (1 + y) + N(c − y)
= −
y c [(1 + y)N − 1] + y

Appendix 5.2: Duration of Annuities and Perpetuities

The price of an annuity that pays $A every period is

A 1
[1 − ]
r (1 + r)N

where r is the periodic interest rate.

The duration of an annuity is given by

1 N CF t × t
[∑ ]
P t=1 (1 + r)t
1 A 2A 3A NA
= [ + + + ⋅⋅⋅+ ]
P (1 + r) (1 + r)2 (1 + r)3 (1 + r)N
A 1 2 N
= [ + + ⋅⋅⋅ + ]
P (1 + r) (1 + r)2
(1 + r)N

From the result derived in the previous appendix, this may be expressed as:

A (1 + r) 1 N
{ 2 − − }
P r 2
r (1 + r)N−1
r(1 + r)N

If we substitute for P, the cash flow A drops out.

1
{ (1+r)
r2
− − N N }
r 2 (1+r)N−1 r(1+r)
Duration = 1 1
[r − ]
r(1+r)N

(1 + r) 1 N r(1 + r)N
={ − − } ×
r2 r 2 (1 + r)N−1 r(1 + r)N [(1 + r)N − 1]
[(1 + r)N+1 − (1 + r) − Nr]
=
r[(1 + r)N − 1]
(1 + r) N
= −
r [(1 + r)N − 1]
Appendix 5.2: Duration of Annuities and Perpetuities | 147

Consider the expression:

(1 + r) N
D= −
r [(1 + r)N − 1]

Define f (N) = N and g(N) = [(1 + y)N − 1]. As N → ∞, both f (N) and g(N) → ∞. From
L’Hôpital’s rule
df
f (N) dN
LimitN→∞ = LimitN→∞
g(N) dg
dN

In this case
df
=1
dN
and
dG
= (1 + y)N ln(1 + y)
dN
Thus,
df
dN
LimitN→∞ dg
=0
dN

Thus for a perpetuity, the duration is given by (1+r)

r
.
An alternative derivation:
price of perpetuity is given by

A
P=
r
dP A
=− 2
dr r
dP 1
=−
Pdr r
dP (1 + r)
− × (1 + r) =
Pdr r

The price of an annuity due that pays $A every period is:

A 1
P= [1 − ] (1 + r)
r (1 + r)N
A 1 1
= [1 − ] + A × [1 − ]
r (1 + r)N (1 + r)N
dP A 1 AN AN
= − 2 [1 − ]+ +
dr r (1 + r)N
r(1 + r)N+1
(1 + r)N+1
148 | Chapter 5: Duration, Convexity, and Immunization

A 1 AN
2
=−
[1 − ]+ × (1 + r)
r (1 + r)N
r(1 + r)N+1
dP −1 N
= +
Pdr r(1 + r) (1 + r)[(1 + r)N − 1]
dP 1 N
− × (1 + r) = −
Pdr r [(1 + r)N − 1]

1 N
So the duration of an annuity due is r
− . As N → ∞, the second term tends
[(1+r)N −1]
to zero. Hence, the duration of a perpetuity due is 1r .

Appendix 5.3: Duration and Interest Rate Sensitivity

Let’s state the expression for the dirty price of a plain vanilla bond, that pays coupons
m times per year.
C C C
m m m M
P= y + +−−−−−−−−−−+ +
y 2 mT y mT
(1 + m
) (1 + m
) (1 + my ) (1 + m
)

Let’s differentiate P with respect to y:

C C
dP 1 m 1 m
=− × − ×2 × − − − − − − − −−
dy m (1 + y )2 m (1 + my )
3
m
C
1 m 1 M
− ×mT × − ×mT ×
m (1 + y mT+1
) m (1 + my )
mT+1
m
C C
1 1 1× m 2× m
=− × [ + +−−−−−− −
m (1 + my ) (1 + my ) (1 + y )2
m
C
mT × m mT × M
+ + ]
y mT y mT
(1 + m
) (1 + m
)

Therefore
C C
dP 1 1 1 1× m
2× m
× =− × y {[ y + +−−−−−−−
dy P m (1 + m ) (1 + m
) (1 + y 2
)
m
C
mT × m mT × M 1
+ + ] }
(1 + y mT
) (1 + y mT
) P
m m

The term in parentheses is nothing but the duration of the bond. Therefore,

dP 1 1 1
× =− × ×D
dy P m (1 + my )
Appendix 5.4: Convexity of Annuities and Perpetuities | 149

Appendix 5.4: Convexity of Annuities and Perpetuities

The price of an annuity that pays $A every period is
A 1
P= [1 − ]
r (1 + r)N
dP A 1 AN
= − 2 [1 − ]+
dr r (1 + r)N
r(1 + r)N+1
d2 P 2A 1 2AN AN(N + 1)
= 3 [1 − ]− −
dr 2 r (1 + r)N r 2 (1 + r)N+1 r(1 + r)N+2
d2 P 2 2N N(N + 1)
= − −
Pdr 2 r 2 r(1 + r) [(1 + r)N − 1] (1 + r)2 [(1 + r)N − 1]

Using L’ Hôpital’s rule, as

2
N → ∞ Convexity →
r2
The convexity of a perpetuity is r22 .
The price of an annuity due is
A 1
P= [1 − ] (1 + r)
r (1 + r)N
A 1 1
= [1 − ] + A [1 − ]
r (1 + r) N
(1 + r)N
dP AN A 1 AN
= − 2 [1 − ]+
dr (1 + r)N+1 r (1 + r)N
r(1 + r)N+1
d2 P AN(N + 1) 2A 1 2AN AN(N + 1)
2
=− + 3 [1 − ]− −
dr (1 + r)N+2 r (1 + r)N 2
r (1 + r)N+1
r(1 + r)N+2
d2 P 2 2N N(N + 1)
= − −
Pdr 2 r 2 (1 + r) r(1 + r)2 [(1 + r)N − 1] (1 + r)2 [(1 + r)N − 1]

Using L’ Hôpital’s rule, as

2
N → ∞ Convexity →
r 2 (1 + r)
2
Thus the convexity of a perpetuity due is r 2 (1+r)
.

Appendix 5.5: Proof of Single-Period Immunization

We now demonstrate that if the duration of the asset is equal to the investment hori-
zon, then the reinvestment risk balances the price risk.12

12 See Taggart [64] for a similar proof.

150 | Chapter 5: Duration, Convexity, and Immunization

Assume that an investor has an investment horizon of H years. The bond has N
annual coupons remaining, where each coupon is for $C. The face value of the bond
is $M. Let’s assume that out of the N coupons, M are received before the investment
horizon and N − M coupons are received after that. The final wealth can be divided
into two components. The first is the future value of the M coupons received before
the horizon date. The second is the liquidation value of the un-matured bond at the
horizon date. Both these components depend on the level of interest rates following
the purchase of the bond. If rates increase, the first component increases while the
second declines. The bond as a whole is an immunized or risk-less investment if a
change in one component is exactly offset by an opposite change in the other. Let’s
assume that rates change only once, and that the change occurs immediately after the
acquisition of the bond.
Assume that each coupon can be reinvested at a rate R. The total proceeds from
the re-invested coupons is given by

I = C(1 + R)H−1 + C(1 + R)H−2 + ⋅ ⋅ ⋅ + C(1 + R)H−M

𝜕I
>0
𝜕R

Thus the higher the reinvestment rate is, the greater the future value of reinvested
coupons.
The liquidation value of the bond at a YTM of R is

C C C C+M
L= + + ⋅⋅⋅ + +
(1 + R)M+1−H (1 + R)M+2−H (1 + R)N−1−H (1 + R)N−H
𝜕L
<0
𝜕R

Thus the higher the YTM is, the lower the sale price.
The investor’s final wealth is W = I + L. We need a condition such that 𝜕W
𝜕R
= 0;
that is, the final wealth is invariant to changes in the interest rate.

C
W = C(1 + R)H−1 + C(1 + R)H−2 + ⋅ ⋅ ⋅ + C(1 + R)H−M +
(1 + R)M+1−H
C C C+M
+ M+2−H
+ ⋅⋅⋅ + N−1−H
+
(1 + R) (1 + R) (1 + R)N−H

Divide both sides of the equation by (1 + R)H :

W C C C
= + + ⋅⋅⋅+
(1 + R) H (1 + R) (1 + R) 2
(1 + R)M
C C C C+M
+ M+1
+ M+2
+ N−1
+
(1 + R) (1 + R) (1 + R) (1 + R)N
Appendix 5.5: Proof of Single-Period Immunization | 151

The RHS is nothing but the bond price at the outset. Thus,

W
=P
(1 + R)H
⇒ W = P(1 + R)H
𝜕W 𝜕P
= HP(1 + R)H−1 + (1 + R)H
𝜕R 𝜕R
𝜕P HP
= (1 + R)H [ + ]
𝜕R (1 + R)
𝜕P DP
=−
𝜕R (1 + R)
𝜕W DP HP
⇒ = (1 + R)H [− + ]
𝜕R (1 + R) (1 + R)
= (1 + R)H−1 × P[H − D]

Thus ⇒ 𝜕W
𝜕R
= 0, if H = D.
Chapter 6
The Money Market
The market for capital is divided into the money market and the capital market.
Medium-term and long-term bonds and equity securities are issued in the capital
market. This market is an arena where parties raise funds to make long-term invest-
ments. Thus if a manufacturing company wishes to set up a new factory, or a company
seeks to borrow to construct a new building for its headquarters, it usually approaches
the capital market. However, in practice, funds may also be required for managing
liquidity imbalances. Take the case of the federal government. It gets its revenues
primarily by way of taxes. However, tax receipts tend to be lumpy, and inflows do not
occur throughout the year on a uniform basis. On the other hand, expenditures are
usually incurred on a daily basis. Consequently, even if the federal government has
a surplus budget for the year as a whole, which is usually not the case, during most
of the year it will have a deficit. In such a situation, it has no choice but to borrow
for short periods and hence approaches the money market. Similarly, take the case
of a company that needs short-term capital for meeting revenue expenditure. Such
a company may be profitable for the year, but may often have a cash deficit and re-
quire short-term capital for meeting expenditures. In such cases, it too approaches
the money market. Thus, both government and business entities will approach the
money market when they are in need of short-term funds.
What motivates lenders to lend in such markets? For most business and govern-
ment entities, cash inflows and outflows are not perfectly synchronized, and conse-
quently there is a short-term deficit or surplus. While entities with a deficit have no
option but to borrow, parties with a surplus seek to lend because money is a perish-
able asset.
If cash is kept idle, there is an opportunity cost in terms of interest foregone. When
large amounts of funds are involved, the lost income can be substantial. Take for in-
stance the case of an institution that has $100 million available. If we assume that the
interest rate is 3% per annum and the year consists of 360 days, which is the assump-
tion in the U.S. money market, the loss if the funds are kept idle for six days is

6
100,000,000 × 0.03 × = $50,000
360

This lost income cannot be recovered subsequently. One can argue that we might be
able to invest at double the rate in the following week. However, even assuming that
were to be the case, if we invest for two weeks, we get two weeks worth of interest,
whereas if we invest only for the second week, we get only a week’s worth of interest.
Thus we say that money is perishable. Use it or lose it. Similar analogies can be given
from the hotel and airline industries. If the Boston Marriott has 10 unoccupied rooms
on a given day, then the revenue that is lost is lost forever. For, obviously, the hotel

https://doi.org/10.1515/9781547400669-006
Risk Factors in the Money Market | 153

cannot accommodate two guests in a room on a subsequent day. Similarly, if Emirates

flies from New York to Dubai with 10 empty seats then the income that is lost is lost
forever, because on the next flight two passengers cannot be assigned the same seat.
Because fund inflows and outflows are not predictable, no entity is usually a per-
manent borrower or lender in the money market. A business may suddenly require
money to pay income tax and approach the money market to borrow on a short-term
basis. However, two weeks hence, a customer who had been supplied on credit may
make a large payment that would result in an inward surge of cash. In this situation, it
is likely that the business will approach the market once again, this time as a lender.
Also, there are institutions that simultaneously operate on both sides of the market.
Take the case of a large commercial bank such as Citibank. As a bank it needs money
to accommodate the needs of its clients who seek to borrow from it. Consequently, it
issues negotiable certificates of deposit and borrows federal funds, activities which
warrant its classification as a borrower. However, at the same time, it is likely that the
bank is making short-term loans to satisfy the needs of its corporate clients. So the
answer to the question, as to whether Citibank is a borrower or a lender in a money
market, is that it is both. In some transactions it takes the role of a borrower, and in
others, it plays the part of a lender.

Risk Factors in the Money Market

Investors in the money market seek to keep their temporary surpluses gainfully in-
vested, without compromising on credit risk or liquidity. Liquidity needs cannot be
anticipated, and parties seek the freedom to enter as borrowers and/or as lenders at
short notice. Thus it is imperative from the standpoint of such parties that the market
be characterized by the presence of plenty of active buyers and issuers of securities
who can facilitate transactions without a major price impact. In developed nations,
the money markets are extremely liquid and can absorb large volumes of transactions
with only a relatively small impact on security prices and interest rates. Consequently,
it is possible for borrowers to issue large quantities of securities at short notice, often
in a matter of minutes. Safety is also critical because investments are made for short
periods and may need to be liquidated at any time. For this reason money market in-
vestors are especially sensitive to default risk. Even the slightest perception of a de-
cline in credit quality can bring the market to a grinding halt.
Money market securities, like other debt securities, are in principle vulnerable
to interest rate risk. However, the prices of such securities are relatively more stable
over time. Thus, although such securities do not offer prospects for significant cap-
ital gains, they do not raise the specter of significant capital losses either. There are
essentially two reasons for this. First, interest rate movements over relatively short pe-
riods of time tend to be moderate. Second, the price impact due to a given interest rate
change is greater, the longer the term to maturity of the cash flow being impacted. In
154 | Chapter 6: The Money Market

this case, the maximum term to maturity of a security is one year. Similarly, default
risk is minimal in the money market. This is because the ability to borrow in such
markets is restricted to well-established institutions with impeccable credit ratings,
usually from multiple rating agencies. That is, it is usually not adequate for an issuer
to obtain a rating from, say, S&P; it also needs a second rating from Moody’s or Fitch.

Supervision of the Money Market

The money market is an over-the-counter market made up of dealers and brokers. Con-
sequently there is no money market exchange. The money markets are overseen by
the respective central banks of the countries in which they operate. The major central
banks in the world are listed in Table 6.1.

Table 6.1: Major central banks.

The United States Federal Reserve

The European Central Bank
The Bank of England
The Bank of Japan
The Swiss National Bank
The Bank of Canada
The Reserve Bank of Australia
The Reserve Bank of New Zealand

The U.S. Federal Reserve system consists of 12 district Federal Reserve Banks. These
are identified by the cities in which they are located. The 12 banks are listed in Ta-
ble 6.2.

Table 6.2: The District Federal Reserve Banks.

Federal Reserve Bank of Boston

Federal Reserve Bank of New York
Federal Reserve Bank of Philadelphia
Federal Reserve Bank of Cleveland
Federal Reserve Bank of Richmond
Federal Reserve Bank of Atlanta
Federal Reserve Bank of Chicago
Federal Reserve Bank of St. Louis
Federal Reserve Bank of Minneapolis
Federal Reserve Bank of Kansas City
Federal Reserve Bank of Dallas
Federal Reserve Bank of San Francisco
Key Dates in Money Market Transactions | 155

Key Dates in Money Market Transactions

There are three key dates in the lives of money market instruments:
– Transaction date
– Value date
– Maturity date

The transaction date is the date on which the terms and conditions of a financial in-
strument, such as the term to maturity, transaction amount, and price, are agreed
upon. In other words, it is the date on which the two counter-parties enter into a con-
tract with each other.
The value date is the date on which the instrument starts to earn or accrue a return.
The value date may or may not be the same as the transaction date. If the two dates
are the same, then the transaction is said to be for same day value or value today. In
other cases, the value date is the following business day. Transactions with such a
feature are said to be for next-day value or value tomorrow. Finally, there are markets
and transactions where the value date is two business days after the transaction date.
A transaction with such a feature is referred to as a spot transaction and is said to be
for spot value.
The maturity date is the date on which the instrument ceases to accrue a return.
The maturities of money market instruments are often fixed not by agreeing to a spe-
cific date of maturity, but by agreeing to a term to maturity, which is a specified number
of weeks or months after the value date.

Roll Conventions in the Event of Market Holidays

If the maturity date of a transaction is a market holiday, then it needs to be rolled
forward, or at times backward, to a regular working day. The market may follow one
of three possible conventions for the roll-over. These are
– The following business day convention
– The modified following business day convention
– The preceding business day convention

Irrespective of the roll convention, there is one common rule. That is maturity is set for
the same date as the value date. For instance, if the value date is 21 June, the maturity
date for a three-month deposit is 21 September, and for a six month deposit, it is 21
December.
Now let’s consider the modified following business day convention. Per this con-
vention, if the maturity date happens to be a holiday, then it moves to the following
business day. Take the case of a three-month deposit with a value date of 21 June. Un-
der normal circumstances the maturity date is 21 September. However, if 21 September
156 | Chapter 6: The Money Market

falls on a Saturday, for instance, then the maturity date becomes Monday, the 23rd of
September, assuming of course that this day is not a market holiday. However, in the
process of rolling forward, one cannot select a date in the subsequent calendar month.
In such a situation, the maturity date moves back to the last business day of the ma-
turity month. Consider a one-month deposit made on 31 July. The scheduled maturity
date is normally, 31 August, but assume that 31 August is a Sunday. In this case how-
ever, the following business day falls in September which is the next calendar month.
Consequently, the maturity date moves back to the last business day of August, which
is 29 August.
The following business day convention differs slightly, for it says that in the event
of rolling forward, it is acceptable to cross over to the next calendar month. Thus,
if the maturity date is 31 August, which is a market holiday, then it is modified to 1
September.
The preceding business day convention says that if the scheduled maturity date is
a holiday, it is rolled backward to the previous working day. So if 21 June, the scheduled
maturity date, is a Saturday, then it is rescheduled to Friday 20 June.

The End/End Rule

This rule states that if the value date is the last business day of a calendar month, then
the maturity date is the last business day of the corresponding calendar month. For
example, take a one-month deposit with a value date of 31 May. It matures on 30 June,
assuming it is a business day. Similarly, a one-month deposit with a value date of 30
June matures on 31 July, assuming once again that it is a business day.
A one-month deposit with a value date of 31 January matures on 28 or 29 February
depending on whether it is a leap year. A one-month deposit with a value date of 28
or 29 February matures on 31 March. However, if the maturity date per this rule is a
holiday, then the prescribed roll convention applies.

The Interbank Market

The interbank market is a market for large or wholesale loans and deposits. It is an
arena for borrowing and lending deals between commercial banks, with tenors less
than or equal to one year. Why do economies require an interbank market? On any
given day, certain commercial banks have a surplus of funds while others are con-
fronted with the specter of a deficit. For a bank whose inflow on account of deposits
exceeds the outflow on account of loans and investments, there is a surge in its re-
serves. On the contrary, banks whose demand for funds on account of loan requests
by their clients exceeds the availability of funds by way of deposits, face a depletion of
reserves. The term reserves refers to the balance that a commercial bank has on deposit
The Interbank Market | 157

with the regional Federal Reserve Bank. Because money is perishable, it is prudent for
banks with surpluses to lend them to those with a deficit, despite the fact that interest
rates for such transactions tend to be relatively low. Although the loan of funds is a
manifestation of the desire of lending institutions to keep idle resources productively
invested, on the part of the borrowers, it reflects the fact that they have to bridge the
shortfall. Loans made in the inter-bank market are unsecured.

Types of Loans in the Inter-bank Market

Overnight money refers to money that is borrowed/lent on a given banking day and
scheduled to be repaid on the next banking day. Weekend money is the term used to
describe loans that are made on Friday with repayment being scheduled for the fol-
lowing Monday. Interest on weekend money is payable for a period of three days.

Call money refers to deposits for an unspecified term. The lender can call back the
funds at any time and will be repaid on the same day.

Notice money is money that is lent with a short notice of withdrawal, for example with
seven days’ notice.

Term money is money that is lent or deposited for a fixed period, such as a week or a
month.

Intra-day money refers to money that is lent and repaid on the same day. Borrowing
takes place in the morning of the business day, and repayment is scheduled for the
same afternoon. Although many consider an overnight loan to have the shortest ma-
turity among money market transactions, technically intra-day loans have the shortest
maturity.

LIBOR

LIBOR is an acronym for London Interbank Offer Rate. It is defined as the rate at which
a top-rated bank in London is prepared to lend to a similar bank. It is the main bench-
mark rate in the London interbank market. In practice, LIBOR is quoted for a number
of tenors: 1 month, 2 months, 3 months, 6 months, and 12 months. Thus there are sev-
eral LIBOR rates quoted at any point in time, and any quotation must be prefixed by
its term to maturity. Every bank in London quotes its own LIBOR rate for each tenor.
But usually the rates quoted by competing banks are identical, with some minor dif-
ferences being observed occasionally.
Although LIBOR is the globally recognized benchmark for loans granted by a top-
rated depository institution to parties with a high credit rating, highly rated banks and
companies can often borrow short-term at a rate below the prevailing LIBOR. On the
158 | Chapter 6: The Money Market

other hand, institutions with a lower credit rating may have no option but to borrow
at a rate above LIBOR. Indicative LIBOR rates for the London market as a whole are
released daily by the Intercontinental Exchange (ICE).

ICE LIBOR
ICE LIBOR is the most widely used benchmark or reference rate for short-term interest
rates. It is compiled by ICE and released shortly after 11:00 a. m. London time each
day. ICE maintains a reference panel of contributor banks. The objective is to provide
a reference panel that reflects the balance of the market, by country and by type of
institution. Individual banks are selected on the basis of reputation, scale of market
activity, and perceived expertise in the currency concerned. ICE publishes the quotes
of the panel on the screen. The top quartile and bottom quartile of the quotes are disre-
garded, for they may constitute outliers, and the middle two quartiles are averaged to
arrive at a trimmed arithmetic mean known as the ICE LIBOR rate. ICE provides the LI-
BOR for five currencies: the US Dollar, British pound; euro; Swiss franc; and Japanese
yen.

LIBID
LIBID is the acronym for the London Inter-bank Bid Rate. It is the rate that a bank in
London with a good credit rating is prepared to pay for funds deposited with it by
another highly rated London bank. LIBID, just like LIBOR, is quoted for a number of
tenors. However, whereas LIBOR represents the rate that a bank seeking to borrow
in the interbank market has to pay on a loan availed by it, LIBID is the rate that a
bank with surplus funds has to accept on funds it deposits with another bank. LIBID
is lower than LIBOR. Although the size of the spread between the two rates can vary,
the difference is usually just a few basis points for the same tenor.
In an interbank transaction, the rate agreed upon is often somewhere between LI-
BID and LIBOR, frequently the average of the two rates. Some depository institutions,
therefore, use LIMEAN, which is an arithmetic average of the LIBID and the LIBOR, as
the reference rate for their interbank transactions.

Interest Computation Methods

For some money market instruments, interest is payable on the principal value of the
instrument. That is, interest is paid at the end of the loan period, along with the re-
payment of principal to the investor. This is termed as the add-on approach. On the
other hand, other securities like Treasury bills are what we term as discount securities.
That is, they are issued at a discount to their principal value and repay the principal
at maturity. Quite obviously they are analogous to zero coupon bonds. The difference
may be illustrated by way of a numerical example.
Money Market Forward Rates | 159

Example 6.1. Consider a security with a par value of $100 and a quoted yield of 8%. The maturity is
90 days, and let’s assume that the market convention is that the year consists of 360 days.
If the yield is quoted on an add-on basis, the implication is that an investment of $100 is to be
repaid as $102 after 90 days. The quoted rate is 8% per annum, and the actual annualized return is
also 8% per annum.
0.08 × 90
100 [1 + ] = 102
360
(102 − 100) 360
× = 0.08 ≡ 8%
100 90
If, however, the yield is quoted on a discount basis, the investor has to invest $98 in return for $100 af-
ter three months. Whereas the quoted rate is 8% per annum, the actual annualized return is 8.1632%.
0.08 × 90
100 [1 − ] = 98
360
2 360
× = 8.1632%
98 90
Thus, in the case of add-on securities, the quoted yield and the actual return are identical, but in the
case of a discount security, the actual rate of return is always higher than the quoted yield.

Money Market Forward Rates

Assume that s1 is the quoted rate for an A-day loan and s2 is the rate for a B-day loan,
where B > A. If the rate is quoted on an add-on basis, then we can state that
s1 × A f × (B − A) s ×B
[1 + ] × [1 + ] = [1 + 2 ] (6.1)
360 360 360
where f is the forward rate for a (B − A)-day loan, as fixed at time 0.

Example 6.2. Assume that the 72-day rate is 4% and the 180-day rate is 5.25%. The implied forward
rate is given by
0.04 × 72 f × 108 0.0525 × 180
[1 + ] × [1 + ] = [1 + ]
360 360 360
⇒ f = 6.0350%

If the rate is quoted on a discount basis, the relationship can be stated as

s1 × A f × (B − A) s ×B
[1 − ] × [1 − ] = [1 − 2 ] (6.2)
360 360 360

Example 6.3. Assume that the 72-day rate is 4% and the 180-day rate is 5.25%. The implied forward
rate is given by
0.04 × 72 f × 108 0.0525 × 180
[1 − ] × [1 − ] = [1 − ]
360 360 360
⇒ f = 6.1324%
160 | Chapter 6: The Money Market

For interbank loans, interest is computed and paid along with the principal. The
method of calculating the interest differs according to the currency under considera-
tion. Interest on most currencies, including the U.S. dollar and the euro, is calculated
on the assumption that the year has 360 days and is based on the ACT/360 day-count
convention. This means that the interest payable on a loan for T days with a principal
of P and carrying an interest rate of r is
r T
P× ×
100 360
For certain currencies, however, like the British pound, interest is calculated on the
assumption of a 365-day year and is based on an ACT/365 day-count convention. Con-
sequently the formula for computing interest is
r T
P× ×
100 365

Example 6.4. A bank in New York makes a loan of $25 million from 15 July until 15 October at an interest
rate of 6.00% per annum. The number of days is: 16 + 31 + 30 + 15 = 92. Notice that although the loan
is for three months we consider the actual number of days in the period and do not automatically take
the period as consisting of 90 days. This is because the numerator of ACT/360 is the actual number of
days in the period. The interest is given by:
92
25,000,000 × 0.06 × = $383,333.32
360

Example 6.5. A bank in Atlanta makes a loan of $75 million for a period of one year (365) days at an
interest rate of 6.25% per annum. The interest is given by
365
75,000,000 × 0.0625 × = $4,752,604.17
360
Notice that although the actual number of days exceeds 360, we still compute on a simple interest
basis. This is a feature of money markets.

Term Money Market Deposits

A term money market deposit is a short-term deposit, with a maximum maturity of
one year, that is placed with a bank or a security house. Such deposits carry a fixed
rate of interest, and the rate is linked to the prevailing LIBOR for the same term. Inter-
est is usually on an add-on basis and paid along with the principal when the deposit
matures. These deposits are non-negotiable, that is, they cannot be sold to another
party prior to maturity by the depositor. Depositors can usually break or terminate
the deposit prematurely, that is, they have the option to withdraw their money prior
to expiration. However, this feature at times carries an attached penalty. These days
due to the competitive pressures in the banking industry, it may be possible to place a
deposit with no penalty in the event of premature withdrawal.
Federal Funds | 161

Federal Funds
Federal funds are the principal means of making payments in the money market in the
U.S. By definition, the term “federal funds” refers to money that can be immediately
transferred from the buyer of securities to the seller, or from the lending institution
to the borrower. This term is used because the principal way of effecting an immedi-
ate fund transfer is by debiting the reserve account held by the buyer’s bank or the
lending bank at the regional Federal Reserve Bank, and crediting the reserve account
maintained by the seller’s bank or the borrowing bank at its regional Federal Reserve
Bank. Such transactions are instantaneous and can be effected in a matter of seconds
in practice. On the contrary, conventional payments by way of checks, entail the role
of a clearinghouse, which means that same-day credit is usually infeasible. This is not
acceptable for money market participants because, as we mentioned earlier, money is
a perishable asset.
Banks and other depository institutions must hold in a special reserve account liq-
uid assets equal to a fraction of the funds deposited with them. Only vault cash held
within the bank and reserve balances kept with the local Federal Reserve Bank count
in meeting a U.S. bank’s requirement to hold legal reserves. An increase in deposits
consequently leads to an increase in the availability of federal funds, whereas loans
made and securities purchased, manifest themselves as a reduction in the availability
of such funds. Frequently some, usually smaller, banks hold more legal reserves than
what the law requires. These banks tend to keep their surpluses gainfully invested,
by lending to the larger banks, which are invariably short of funds. Most lenders tend
to make overnight loans of federal funds. This is because the availability of excess re-
serves tends to vary daily, and in a fairly unpredictable fashion. The lending of such
funds is termed as a sale, and the borrowing of such funds is referred to as a purchase.
The 12 member banks of the Federal Reserve System have a share in an Inter-district
Settlement Fund that is maintained with the Federal Reserve headquarters in Wash-
ington D. C. Consequently, if a member bank wants to transfer funds to another mem-
ber, all that is required is a debit to the lender’s share of the settlement fund and a
credit to the borrower’s share.

Treasury Bills
Treasury bills (T-bills), are money market instruments issued by the central or federal
government of a country. In the United States they are issued by the federal govern-
ment, and the rates on these securities set the benchmark for the rates on other money
market securities with the same tenor. The reasons why T-bills carry the lowest yield
for a given tenor are the following. First, they are virtually devoid of credit risk because
they are backed by the full faith and credit of the U.S. government. Second, they are
highly liquid. Third, income from such securities is exempt from state income tax. In
162 | Chapter 6: The Money Market

the U.S., unlike in many other countries, state governments are empowered to levy in-
come tax. The governments follow a guideline of mutual reciprocity. That is, federal
securities are exempt from state income tax and vice versa.
By law, T-bills in the U.S. must have an original maturity of one year or less. Reg-
ular series bills are issued routinely every week or month by way of competitive auc-
tions. Four-week, three-month and six-month bills are auctioned every week, and one-
year bills are sold usually once a month. Of the four maturities, the six-month bills
provide the largest amount of revenue for the U.S. Treasury. On the other hand, cash
management bills are issued only when the Treasury has a special need, on account
of low cash balances. Such bills have maturities ranging from as short as a few days to
as long as six months. They give the maximum flexibility to the Treasury because they
can be issued as and when required. The money raised through these issues is used
by the Treasury to meet any temporary shortfalls.
T-bills are sold by an auction process. The Treasury entertains both competitive
and non-competitive bids. Large investors submit competitive bids, wherein they in-
dicate not only the quantity sought, but also the minimum yield that they are prepared
to accept. The yield in this case refers to the discount yield for the bill being auctioned.
Non-competitive tenders, on the other hand, are submitted by small investors who
agree to accept the yield at which the securities are auctioned off, whatever that yield
may be. Thus, such bidders have to indicate only the quantities sought. Generally the
Treasury fills all non-competitive bids.
Besides being traded in the U.S., T-bills issued by the U.S. Treasury are actively
traded in other major financial centers like London and Tokyo. Thus the market for
such securities operates virtually around the clock. The globalization of the market is
characterized by the presence of non-U.S. dealers in the U.S. market, as well as by the
activities of American dealers in markets outside the country.

Re-openings of T-bills

Every T-bill issue is identified with a unique CUSIP number. Some issues, however, are
a reopening of an existing issue. That is, the new issue is identical in all respects to an
issue that is already trading in the secondary market. For instance, the three-month
bill issued three months after the issue of a six-month bill is considered a re-opening
of the six-month bill. The new issue in this case is given the same CUSIP number.
For a given maturity, the most recently issued securities are referred to as on the run,
whereas those issued earlier are referred to as off-the-run. On-the-run securities gen-
erally trade at slightly lower yields because they are more liquid. The reason is that,
for some time after the issue of such bills, there tends to be active trading in the sec-
ondary market. Thereafter most securities pass into the hands of investors who choose
to hold them until maturity. Consequently, off-the-run securities are less liquid, which
explains the higher yield.
Treasury Bills | 163

Cash management bills are issued via a standard auction process. However, they
are irregular with respect to their term to maturity and auction schedule. If a cash man-
agement bill matures on the same day as a regular bill, which is usually a Thursday,
then it is said to be on-cycle. In this case, the issue is considered to be a reopening,
and is consequently allotted the same CUSIP. However, if it matures on a different day,
it is said to be off-cycle, and obviously carries a different CUSIP number.

Discount Rates and T-bill Prices

The quoted yield on a T-bill is a discount rate, which is used to determine the difference
between the price of a T-bill and its face value. For the purpose of calculation in the
U.S. market, the year is treated as if it has 360 days.
If d is the quoted yield for a T-bill, with a face value of $V and having Tm days to
maturity, the dollar discount D is given by

d T
D=V× × m
100 360

Given the discount, the price may be calculated as P = V − D

We illustrate these concepts with the help of Example 6.6.

Example 6.6. A T-bill with 90 days to maturity and a face value of $1 million, has a quoted yield of
4.8%. What is the price in dollars?
90
The discount is 1,000,000 × 0.048 × 360 = $12,000.
The price is given by

P = V − D = 1,000,000 − 12,000 = $988,000

Example 6.7. A one-year bill (364 days) has just been issued at a quoted yield of 5.4%. What is the
corresponding price in dollars?
364
The discount is 1,000,000 × 0.054 × 360 = $54,600.
The price is given by

P = V − D = 1,000,000 − 54,600 = $945,400.

The Money Market Yield of a T-bill

The money market yield is defined as

V − P 360
×
P Tm
164 | Chapter 6: The Money Market

The quoted yield is defined as

V − P 360
×
V Tm

For a discount security, the price is always lower than its face value. Consequently, the
money market yield for such a security is always higher than the quoted yield.

Example 6.8. Consider the T-bill with 90 days to maturity and a face value of $1 million. The quoted
yield is 4.8%. We have already computed the price as $988,000. The money market yield is

12,000 360
× = 0.048583 ≡ 4.8583%
988,000 90

The Bond Equivalent Yield of a T-bill

The objective of calculating the bond equivalent yield, also known as the coupon equiv-
alent yield, is to compute a yield measure that facilitates comparisons between the rate
of return on discount securities like T-bills and capital market debt instruments like
coupon-paying bonds. The procedure that is adopted to compute the BEY depends on
whether the discount instrument under consideration has less than six months left to
maturity or more.

Case A: Tm < 182 days

The bond equivalent yield for a T-bill with less than 182 days to maturity is just the
equivalent rate of return on a simple interest basis, computed under the assumption
that the year has 365 days. As we saw in the case of the money market yield, the bond
equivalent yield for a T-bill with a given discount rate is always greater than the stated
discount rate. The reason once again is that the bond equivalent yield, as is the case
for the money market yield, is based on the market price, whereas the quoted yield
is based on the face value. Since the price is always less, the quoted yield is always
lower.
The BEY in this case is defined as
V − P 365
×
P Tm

It can be directly computed given the quoted yield, using the following equation:

V 365
BEY = ( − 1) × (6.3)
P Tm
d × Tm
P = V × [1 − ] (6.4)
360
Treasury Bills | 165

Substituting we get

d × 365
BEY = (6.5)
(360 − d × Tm )

Example 6.9 gives an illustration for computing the BEY.

Example 6.9. A T-bill with a face value of $1 million and 90 days to maturity, has a quoted yield of 6%.
What is the bond equivalent yield?
The price is given by

90
P = 1,000,000 − 1,000,000 × 0.06 × = $985,000
360

The bond equivalent yield is given by

0.06 × 365
BEY = = 6.1760%
(360 − 0.06 × 90)

In the case of a bill with fewer than 182 days to maturity, the money market yield can be trans-
formed into the bond equivalent yield by simply multiplying the former by 365 and dividing by 360.

Example 6.10. Take the case of the bill in Example 6.8. The money market yield was computed to be
4.8583%. The BEY can be computed as

365
4.8583 × = 4.9258%
360

Case B: Tm > 182 days

A T-bill with fewer than 182 days to maturity can be directly compared with a bond
with fewer than 182 days to maturity, for the latter is also a zero coupon security and
hence is comparable. However, a conventional coupon paying bond with a time to
maturity that is greater than 182 days (half a year) will make a coupon payment be-
fore it matures. It is inappropriate to compare the simple yield on a discount security
with the YTM of a bond, because although the bond yields an interest payment before
maturity, the discount security does not. To facilitate a comparison between the BEY
for a discount instrument, like a T-bill, with more than 182 days to maturity and the
yield to maturity (YTM) for the bond, the T-bill must be treated as if it too would pay
interest after six months, and that interest is paid on this intermediate interest for the
remaining period left to maturity.
If we denote the BEY as y, we can state that it is the solution to

365
y y Tm −
P [(1 + ) {1 + ( 365 2 )}] = V (6.6)
2 2
2
166 | Chapter 6: The Money Market

The logic is as follows. The future value of an investment equal to P, at the end of
six months, is

y
P (1 + )
2

The future value of this expression, as calculated on the date of maturity must equal
the face value. The compounding factor for the remaining period, on a simple interest
basis, is

365
y Tm − 2
[1 + ( 365 )]
2
2

The end result is that we have a quadratic equation in y. This has two roots, and
we discard the negative root and retain the positive. The positive root for y is given by

T 2 2T V
+ 2√( 365 ) − ( 365m − 1) (1 −
−2Tm m
365 P
)
2Tm
365
−1

Holding Period Return for an Investor

Take the case of an investor who buys a bill at a quoted yield of d1 , when there are Tm1
days left to maturity, and sells it at a discount rate of d2 , when there are Tm2 days left
to maturity. Let’s denote the purchase price by P1 and the sale price by P2 . The holding
period return is given by

P2 − P1 365
×
P1 Tm1 − Tm2

P2 − P1 represents the capital gain or loss. The number of days for which the bill has
been held is Tm1 − Tm2 .

Concept of a Tail in a T-bill Transaction

What is a tail? Trading-glossary.com defines a tail as follows: Calculating the yield at

which a future money market security (one available some period hence) is purchased
when that future security is created by buying an existing instrument and financing the
initial portion of its life with a term repo. We have not dealt with repos yet. For the
moment let’s just interpret it as a collateralized loan.
We can illustrate it with the help of an example.
T-bill Related Computations Using Excel | 167

Example 6.11. Marc Anthony & Co., a brokerage house, is acquiring a 144-day bill. It plans to hold it
for 36 days and then sell it. The issue is, what should be the quoted yield at the time of sale if the firm
is to break even on the transaction. Assume the face value of the bill is $1 million and the quoted yield
at the outset is 6%. The purchase price is

144
1,000,000 × (1 − 0.06 × ) = $976,000
360

The funding cost for 36 days, which is either an actual cost or an opportunity cost, assuming a bor-
rowing rate of 6.40% is

36
976,000 × 0.064 × = $6,246.40
360

Thus the effective cost of the bill is 976,000+6,246.40 = 982,246.40. To break even on the transaction,
the bill, which will have 108 days to maturity at the time of the sale, must be sold for this amount. The
corresponding discount rate can be obtained from the following equation:

108
982,246.40 = 1,000,000 × (1 − d × )
360
⇒ d = 5.9179%

If the discount rate at the time of sale were to be lower, there would be a profit for the brokerage house.
However, if it were to be higher, there would be a loss.

T-bill Related Computations Using Excel

To compute the price of a T-bill, we use the TBILLPRICE function.
The parameters are:
– Settlement
– Maturity
– Discount

Example 6.12. Consider a 126-day T-bill with a quoted yield of 6% per annum. What is the price? If
we knew the actual settlement and maturity dates, we could enter them using the DATE function in a
YYYY,MM,DD format. In this case, all we know is that the bill has 126 days to maturity. Consequently,
we can specify any two numbers so that the difference is 126. The easiest approach is to specify a
value of 0 for settlement and 126 for maturity. Note we could have specified 100 and 226 had we so
desired. The discount is 0.06. Excel computes the price per $100 of face value. Consequently, if we
have a bill with a face value of $1 million, we multiply the answer by 10,000.
In our case: The price = TBILLPRICE(0,126,0.06) = $97.90.
We can verify it:

126
P = 100 × (1 − 0.06 × ) = $97.90
360

To compute the money market yield, we use the TBILLYIELD function.

168 | Chapter 6: The Money Market

The parameters are

– Settlement
– Maturity
– Price

Settlement and maturity are defined as for the TBILLPRICE function. Price is the price
per $100 of face value.

Example 6.13. Consider the bill with 126 days to maturity and a quoted yield of 6%. The price is $97.90.
So we compute the yield as:
Money market yield = TBILLYIELD(0,126,97.90) = 6.1287%.
We can verify it:

(100 − 97.9) 360

Money Market Yield = × = 6.1287%
97.90 126

To compute the bond equivalent yield, we need the TBILLEQ function.

The parameters are
– Settlement
– Maturity
– Discount

The parameters are specified as for the TBILLPRICE function. There are two formulas
for the bond equivalent yield, depending on the time to maturity of the bill, and Excel
automatically uses the appropriate method. If we use the TBILLEQ function with val-
ues such that the difference between maturity and settlement is 182 or less, it uses the
first method. However, if the difference is more than 182, it automatically switches to
the second method.

Example 6.14. Let’s reconsider the bill with 126 days to maturity and a quoted yield of 6%. Bond equiv-
alent yield = TBILLEQ(0,126,0.06) = 6.2138%.
We can verify it:

(100 − 97.9) 365

Bond Equivalent Yield = × = 6.2138%.
97.90 126

Finally if we know the maturity of the bill and its price, we can compute the corre-
sponding discount rate. This can be done using a function called DISC.
The parameters are:
– Settlement
– Maturity
– Price
– Redemption
– Basis
Repurchase Agreements | 169

Settlement and maturity have the same meaning as in the case of the Excel functions
invoked earlier. Price is per $100 of face value. Redemption is invariably 100. Basis is
the day-count convention. In the U.S. we use the Actual/360 convention.

Table 6.3: Basis values for day-count conventions.

Basis Day-Count
Value Convention
0 30/360 NASD
1 Actual/Actual
2 Actual/360
3 Actual/365
4 30/360 European

Example 6.15. Once again, let’s take the bill with 126 days. We can compute the discount rate as d =
DISC(0,126,97.90,100,2) = 0.06 ≡ 6%.
We can verify it:

97.90 360
Quoted Yield = [1 − ]× ≡ 6%
100 126

Repurchase Agreements
A repurchase agreement, or repo, is a money market transaction with two legs. In the
first leg, a party agrees to sell securities at a specified price. At the same time, the same
party agrees to buy back the security subsequently at the original price plus interest,
in what constitutes the second leg of the transaction. Thus, a repo is essentially a col-
lateralized loan. Because if the party selling the securities does not buy them back, the
lender has access to the securities to recover what is owed. Such a transaction has two
perspectives. From the borrower’s side, it is termed as a repo. From the lender’s side,
it is known as a reverse repo. Thus every repo transaction is accompanied by a reverse
repo. Most of these transactions are overnight. That is, the money is repaid and the
securities reacquired on the following day. However, there are transactions of longer
duration known as term repos. In some markets, a repo is known as a ready-forward
contract. This is because, although the first leg is immediate, the second is a forward
contract.
The motive for these transactions is as follows. Dealers in financial markets carry
large quantities of securities as a part of their inventory. Their own capital is usually
inadequate to fund these purchases, and the bulk of the securities held are funded
with borrowed money. A securities dealer may buy a security hoping to sell it soon.
170 | Chapter 6: The Money Market

However, if the sale does not materialize, the dealer has securities coming and an ur-
gent need for cash. An obvious way to raise money is by doing a repo. The motive for
a reverse repo is the converse. A dealer may have sold a security thinking that he will
acquire and deliver. But he may be unable to do so. In this case, he has cash coming
his way and needs securities. An obvious solution is a reverse repo. Thus dealers in
search of cash do repos, and those in search of securities do reverse repos.
Repos were first introduced by the Federal Reserve as a tool for regulating the sup-
ply of money in the economy. Although securities dealers undertake such transactions
to fund their positions, the FED uses them as a monetary policy tool. Today repos are a
major constituent of money market operations worldwide, in both developed capital
markets and markets in the emerging economies. The securities used in repo transac-
tions are typically government securities such as T-bills and T-bonds. However, highly
rated corporate debt securities may also be pledged as collateral in some markets.
If the bond, which is pledged as collateral, pays a coupon during the life of the
repo, the lender collects and hands over the coupon to the borrower. In other words,
although a repo constitutes a transfer of ownership, the borrower continues to enjoy
the benefits of the bond in terms of the coupon payments. That is, the borrower con-
tinues to remain the beneficial owner of the security.

Example 6.16. A dealer wants to borrow by pledging government securities with a face value of $5
million. The current market price is 100-24, and the accrued interest is $3.75 per $100 of face value.
Thus the dirty price of the bond is:
24
100 + 32
+ 3.75
5,000,000 × = $5,225,000
100
The loan carries an interest of 3.6% per annum and is repayable after 27 days. Thus the amount payable
at maturity, when the collateral is taken back, is
27
5,225,000 × (1 + 0.036 × ) = $5,239,107.50
360

Repo Rates

Most securities in the repo trade are close substitutes for each other. Thus there is a
common interest rate known as the general collateral rate. However, at times a dealer
may require a security very urgently and find that it is in short supply. On such an oc-
casion, he can agree to lend money at a lower rate of interest if this particular security
is offered as collateral. Such securities are said to be on “special.”

Margins in Repo Transactions

In principle both the parties to a repo face credit risk. If the collateral declines in value,
the lender is at risk because the borrower may refuse to return the cash, and the market
Repurchase Agreements | 171

value of the securities may be inadequate to recover the loan. On the contrary, if the
collateral appreciates in value, the borrower is at risk, for the lender may decide to
retain the security which by assumption is worth more than what was lent initially.
There is no strategy that can simultaneously protect both parties, and more protection
for one side means greater risk for the other.
Lenders can protect themselves by applying what is termed in market parlance
as a “haircut.” For instance, if the current market price is $100, a lender who applies
a haircut of 4%, will lend only $96. A borrower can protect himself by insisting on a
reverse haircut. That is, he can offer a security priced at $100 for a loan of say $104.
In practice, we cannot have a situation where the lender applies a haircut and the
borrower applies a reverse haircut. In the market we have haircuts. The rationale is
that the lender is giving cash that is more liquid than the security that is being offered
in return. Consequently the right to collect a margin is given to the lender and not to
the borrower.
The size of the haircut depends on the following factors:
– The credit quality of the collateral
– The time to maturity of the collateral
– The term to maturity of the repo

The collateral is periodically valued at the prevailing market rate, a procedure that
is known as “marking to market.” In practice the lender sets a threshold level for the
value of the collateral termed as the maintenance margin level. If the securities fall in
value and a situation arises in which their current market value is lower than the main-
tenance level, the borrower must provide additional securities with a market value
that is sufficient to make up the deficit. On the other hand, if the securities rise in
value, the borrower can ask for extra cash or a partial return of collateral.
Example 6.17 shows the two ways in which the haircut can be applied.

Example 6.17. A dealer that is prepared to offer bonds with a face value of $40 million, enters into a
six-day repurchase agreement with a bank. The collateral is T-bonds with a coupon of 6% and a market
price of $95 per $100 of face value. The haircut is 1.25%. Accrued interest is $2.50 per $100 of face
value. The repo rate is 6.00% per annum.
The loan amount can be calculated in either of two ways:
Method I: The loan amount is
95 + 2.50
40,000,000 × × (1 − 0.0125) = $38,512,500
100

The amount repayable at maturity is

6
38,512,500 × (1 + 0.06 × ) = $38,551,012
360

Method II: The loan amount is

95 + 2.50 1
40,000,000 × × = $38,518,518.50
100 (1 + 0.0125)
172 | Chapter 6: The Money Market

The amount repayable at maturity is

6
38,518,518.50 × (1 + 0.06 × ) = $38,557,037
360

Thus a haircut of x% may be applied by either multiplying the dirty price by (1 − x), or else, by
dividing it by (1 + x). We do not get the same loan amount in both cases. However, both are legitimate
methods for applying a haircut.

The Federal Reserve and Repos

The central bank undertakes repos and reverse repos with primary dealers. A primary
dealer is defined as a dealer who is authorized to deal directly with the central bank
of a country. In the U.S. a primary dealer is authorized to deal with the New York Fed.
The Federal Reserve Bank of New York (FRBNY) is first among equals in the central
banking structure of the United States. This is because all decisions pertaining to open
market operations, are implemented by the FRBNY, as New York is the largest money
market in the world.
Although the mechanics of a repo are standard, the motives of security dealers dif-
fer from those of the central bank. Dealers use repos and reverse repos to manage their
inventories. For the central bank, these transactions constitute an important compo-
nent of the monetary policy instruments available to it. A repo represents a collateral-
ized loan made by the Fed to primary dealers. Thus repos are used by dealers to bor-
row funds from the Fed. On the other hand, a reverse repo transaction entails borrow-
ing funds by the Fed from the primary dealers. From a monetary policy standpoint, a
repo temporarily adds reserve balances to the banking system, whereas a reverse repo
transaction temporarily drains such balances from the system. These transactions are
undertaken by the trading desk at the New York Fed, which in turn implements mone-
tary policy decisions taken by the Federal Open Market Committee (FOMC). When the
Fed does a repo, funds are credited to the dealer’s commercial bank. This enhances
the level of reserves in the banking system. At maturity, in return for the amount lent
plus interest, the Fed transfers the collateral back to the dealer concerned. This auto-
matically neutralizes the extra reserves that were created by the original transaction.
Thus, such operations undertaken by the Fed have a short-term, self-reversing effect
on bank reserves. Reverse repos are similar from the standpoint of their impact on
reserves. In such transactions, the Fed sends collateral to the dealer’s clearing bank
in return for the funds. This action reduces the availability of reserves in the banking
system. At the time of maturity of the contract, the dealer returns the collateral to the
Fed, which in turn returns with interest the funds that were initially borrowed. This
automatically restores the reserves that were reduced by the original transaction.
Negotiable Certificates of Deposit (CDs) | 173

Negotiable Certificates of Deposit (CDs)

A certificate of deposit, or CD, is an interest-bearing receipt for funds left with a de-
pository institution for short periods of time. CDs are usually issued at par and pay
interest explicitly, although banks also issue discount CDs that are similar to T-bills in
structure. Payment is made in federal funds on the day of maturity. Banks and other
institutions issue many types of CDs, but true money market CDs are negotiable instru-
ments that can be resold before maturity and are typically issued with a face value of
$1 million. CDs issued to retail investors are non-negotiable and consequently not clas-
sified as money market instruments. CDs issued in the U.S. and eurodollar CDs usually
have an original maturity that is less than or equal to one year and pay interest at ma-
turity. The day-count convention is Actual/360, and the rate of interest is referred to
as the coupon rate. The interest rate on a money market CD is slightly less than that
on a standard time deposit of the same tenor, because the instrument is negotiable.

Required Symbols

We will now derive the expression for determining the price of a CD, given its yield,
and vice versa.
– N is the OTM, that is, the number of days from the time of issue until the time of
maturity.
– Tm is the ATM, that is, the number of days from the settlement date until the ma-
turity date.
– c is the coupon rate.
– y is quoted yield.
– V is the face value or principal amount of the CD.
– P is the dirty price of the CD.

The amount payable at maturity is given by:

N
V × [1 + c × ]
360

If the quoted yield is y, the dirty price is given by:

N
V × [1 + c × 360
]
P= Tm
(6.7)
[1 + y × 360
]

The accrued interest may be computed as:

(N − Tm )
AI = V × c × (6.8)
360
174 | Chapter 6: The Money Market

Given the dirty price, the yield is given by the following equation:
N
V × [1 + c × 360
] −P 360
y={ }× (6.9)
P Tm

Example 6.18. Consider a CD with 144 days to maturity that has just been issued with a face value of
$1 million and a coupon rate of 3.60%. The amount that the holder receives at maturity is equal to

144
1,000,000 × [1 + 0.036 × ] = $1,014,400
360

If the quoted yield is 4.80%, and there are 108 days until maturity, the dirty price is given by

1,014,400
P= 108
= $1,000,000
(1 + 0.048 × 360
)

The accrued interest is:

(144 − 108)
AI = 1,000,000 × 0.036 × = $3,600
360

Thus the clean price is 1,000,000 − 3,600 = $996,400.

Example 6.19. Consider the CD with a coupon of 3.60%, an OTM of 144 days, and 108 days until ma-
turity. The dirty price is $995,200. What is the yield?

144
1,000,000 × [1 + 0.036 × 360
] − 995,200 360
y= × = 6.4309%
995,200 108

Term Certificates of Deposit

Term CDs have a maturity in excess of one year at the time of issue, and usually pay
coupons on a semiannual basis. The price of the instrument is the present value of all
the cash flows to be received from it as of the settlement date, where the discounting
is done using the quoted yield. The procedure is analogous to the valuation of a bond,
where the coupons and the face value are discounted at the YTM, with certain critical
differences.
1. In the case of a bond, the coupon interest for a full coupon period is C2 irrespective
of the actual number of days in the coupon period. In the case of a term CD, the
coupon for a full period depends on the actual number of days in the period. That
N
is, the interest in this case is given by V × c × 360i where Ni is the number of days
in the ith coupon period. Thus the coupon will vary from period to period.
2. The second difference is that for a bond the discount factor for a full coupon period
t
is (1 + y2 ) . In the case of a term CD, the discounting is based on the actual number
of days in the coupon period.
Term Certificates of Deposit | 175

3. The third crucial difference is that for bonds, cash flows are discounted using com-
pound interest, whereas for a term CD we use simple interest for discounting.

We can illustrate the valuation of a term CD with the help of an example.

Example 6.20. A CD with a coupon rate of 3.6% was issued on 1 July 2019. It is scheduled to mature
on 30 June 2021. On 15 November 2019, the quoted yield is 4.8%. What should be its market price?
The first step is to calculate the number of days in each coupon period, as well as the number of
days from the date of settlement to the first coupon. The length of each period is given in Table 6.4.

Table 6.4: Number of days in each period.

Start Date End Date No. of Days

1 July 2019 1 January 2020 184

1 January 2020 1 July 2020 182
1 July 2020 1 January 2021 184
1 January 2021 30 June 2021 180
15 November 2019 1 January 2020 47

To value the security, we start with the cash flow at maturity. This is discounted back to the penulti-
mate coupon date. The coupon payment at this point of time is added to the present value obtained
in the previous step. The sum is then discounted back to the previous coupon date. This procedure is
repeated until we reach the settlement date.
In our example, the cash flow at maturity is

180
1,000,000 × (1 + 0.036 × ) = $1,018,000
360

The present value of this on 1 January 2021 is

1,018,000
180
= $994,140.62
(1 + 0.048 × 360
)

The coupon due on 1 January 2021 is

184
1,000,000 × 0.036 × = $18,400
360

The amount to be discounted to the previous coupon date is

994,140.62 + 18,400 = $1,012,540.60

The present value on 1 July 2020 is

1,012,540.60
184
= $988,294.47
(1 + 0.048 × 360
)
176 | Chapter 6: The Money Market

The coupon due on 1 July 2020 is

182
1,000,000 × 0.036 × = $18,200
360
The amount to be discounted to the previous coupon date is

988,294.47 + 18,200 = $1,006,494.40

The present value on 1 January 2020 is

1,006,494.40
182
= $982,648.85
(1 + 0.048 × 360
)

The coupon due on 1 January 2020 is

184
1,000,000 × 0.036 × = $18,400
360
The amount to be discounted to the settlement date is

982,648.85 + 18,400 = $1,001,048.80

The present value on 15 November 2019 is

1,001,048.80
47
= $994,814.69
(1 + 0.048 × 360
)

This is the dirty price of the CD as of 15 November 2019. The accrued interest as of this day is
184 − 47
1,000,000 × 0.036 × = $13,700
360
Thus the clean price is

994,814.69 − 13,700 = $981,114.69

NCDs vs. Money Market Time Deposits

In the case of a conventional time deposit, an investor deposits a sum of money with
a bank for a stated period of time. The bank pays interest at a specified rate. At the
end of the deposit period, the investor can withdraw the original sum deposited plus
the interest. In the case of an NCD, however, the depositor is typically issued a bearer
security. Although entitled to claim the deposit with interest at the end of the deposit
period, the depositor can always sell the security prior to maturity in the secondary
market. In the process, ownership of the underlying deposit transfers from the seller
to the buyer. Thus NCDs have one major advantage over a conventional money market
deposit, namely liquidity. A money market deposit cannot be easily terminated until it
matures. If the investor wants to withdraw the funds prior to maturity, he may have to
pay a penalty in terms of a lower rate of interest. However, with increasing competition
in many markets, banks are finding it difficult to levy a penalty. In contrast, NCDs can
be liquidated at any time at the prevailing market rate. Thus these instruments are
attractive for investors who seek the high interest rates offered by time deposits but
who are reluctant to commit their money for the full term of the deposit.
Commercial Paper | 177

The Effective Cost of a CD

For the issuing bank, the effective cost of the CD is greater than the quoted rate of
interest. This is because, first, the bank has to keep a certain percentage of the deposit
as a reserve with the central bank, which may or may not earn interest. Even if the
central bank does pay interest, it is below the market rate. The second reason is that
in most countries the deposits are insured up to a certain limit. Example 6.21 illustrates
what we have just described.

Example 6.21. A bank has issued a CD carrying interest at the rate of 4.50% per annum. It is required
to maintain a reserve of 10% with the central bank, on which it earns nil interest. The deposits have to
be insured and the cost is 10 bp. Thus if a deposit of $100 is made, the bank is paying $4.50 of interest
on $90 of usable money. Thus the effective cost is 5%. If we factor in the insurance premium, the cost
is 5.10%.

Commercial Paper
Commercial paper is a term used for short-term unsecured promissory notes issued
by corporations, primarily for funding their working capital requirements. The term
“unsecured” means that no assets are being pledged as collateral. The issuers are gen-
erally financially strong with high credit ratings, which is a prerequisite for successful
issue of such an instrument. The funds raised in this manner are normally used for
current account transactions such as, purchase of raw materials, payment of accrued
taxes, meeting of wage and salary obligations, and other short-term expenditures,
rather than for capital account transactions or in other words long-term investments.
Although most issuers of paper enjoy a high credit rating, they invariably secure a line
of credit at a commercial bank to provide a greater degree of assurance to investors.
But, the line of credit cannot be used to directly guarantee payment if the company
goes bankrupt, and the lender may renege on the credit line if in its perception the
credit quality of the borrower has significantly deteriorated. Consequently, an irrevo-
cable letter of credit opened by a commercial bank offers a greater degree of reassur-
ance to buyers. Such a letter of credit (LC) makes a bank unconditionally responsible
for repayment if the corporation defaults on its paper.1
There are two major types of commercial paper, namely direct paper and dealer
paper.2 As the name suggests, direct paper refers to securities issued directly by the
borrower and does not involve a dealer as an intermediary. The main issuers of direct
paper are large finance companies and bank holding companies that deal directly with
the investors rather than using securities dealers as intermediaries. Such borrowers
have an ongoing need for huge amounts of short-term money, possess top credit rat-

1 We discuss LCs in more detail shortly.

2 See Stigum and Crescenzi [60].
178 | Chapter 6: The Money Market

ings, and have established working relationships with major institutional investors in
order to place new issues regularly. Direct issues involve substantial distribution and
marketing costs, and consequently need to be issued in large volumes. Issuers need
a dedicated in-house marketing division to maintain sustained contact with potential
investors.
The alternative to direct paper is what is termed as dealer paper. Such paper, as the
name suggests, is issued by security dealers on behalf of their corporate customers.
Dealer paper is issued mainly by firms that borrow less frequently than companies
that issue direct paper. The dealer may underwrite the issue, or may agree to sell on
a best efforts basis. In the case of the former, a dealer or syndicate of dealers, buys
the entire issue from the company at a discount and then tries to resell it at the best
available price in the market. If anything remains unsold, then it falls on the dealer’s
lap. In the case of a best efforts deal, the dealer promises to make a best effort to market
the issue. However, if something remains unsold, the dealer is not required to acquire
it. Companies prefer underwritten issues to best efforts issues because in the former,
there is a risk of devolvement. The term “devolvement,” refers to the risk that the issue
may flop and a portion may devolve on the underwriter, who will be required to buy it.
The specter of this makes a dealer market the issue aggressively. In the case of a best
efforts deal, there is no way of verifying whether the dealer has done his best.

Letters of Credit (LCs)

A letter of credit (LC) is a document issued by a bank, called the issuer or opener, at
the behest of a client, referred to as the applicant. The document is drawn in favor
of a stated beneficiary, stating that the opening bank will effect payment of a stated
sum of money for certain specified good or services, against presentation of a prede-
fined set of documents. LC based transactions are very common in international trade
transactions, although they may be occasionally used for domestic deals, as well.
A letter of credit is a bank’s direct undertaking to the beneficiary. That is, when the
shipment of goods occurs under a letter of credit, the issuing bank does not wait for
the buyer to default, or in other words, for the seller to invoke the undertaking, before
making the payment to the beneficiary. Therefore, in the case of an LC, the principal
liability is that of the issuing bank. It must first pay, on submission of the predefined
set of documents, and then collect the amount from its client. Thus an LC substitutes
the bank’s credit-worthiness for that of its client.

Yankee Paper

Paper that is issued in the U.S. by foreign firms is called Yankee paper. Foreign issuers
find that they can often issue Yankee paper at a cheaper rate than what it would cost
Credit Rating | 179

them to borrow outside the U.S. However, foreign borrowers must generally pay higher
interest costs than U.S. companies with comparable credit ratings. That is, Yankee pa-
per generally carries a higher yield than U.S. paper of the same credit quality. This is
because an American investor can demand higher returns due to the difficulty of gath-
ering information on a foreign firm. Samurai paper refers to yen-denominated paper
issued by foreign issuers in Japan. Both Yankee and Samurai paper are examples of
foreign debt securities. That is, although these securities are denominated in the do-
mestic currency, the issuers are foreign entities.

Credit Rating
Short-term debt securities are rated by the three major rating agencies. Their scales,
with the interpretation, are as follows.

Moody’s Rating Scale

Moody’s uses the following rating scale for short-term taxable instruments:
– Prime-1: Superior ability to repay short-term debt obligations.
– Prime-2: Strong ability to repay short-term debt obligations.
– Prime-3: Acceptable ability to repay short-term debt obligations.
– Not Prime: These do not fall within any of the prime rating categories.

Prime-1, Prime-2, and Prime-3 are all investment-grade ratings.

S&P’s Rating Scale

Standard & Poor’s rates issues on a scale from A-1 to D. Within the A-1 category, an issue
can be designated with a plus sign. This indicates that the issuer’s ability to meet its
obligation is extremely strong. Country risk and currency of repayment of the obligor
are factored into the credit analysis and are reflected in the issue rating.
– A-1: Issuer’s capacity to meet its financial commitment on the obligation is strong.
– A-2: The issuer is susceptible to adverse economic conditions. However, the is-
suer’s capacity to meet its financial commitment is satisfactory.
– A-3: Adverse economic conditions are likely to weaken the issuer’s capacity to
meet its financial commitment on the obligation.
– B: The issue has significant speculative characteristics. The issuer currently has
the capacity to meet its financial obligation but faces major ongoing uncertainties
that could impact its financial commitment.
180 | Chapter 6: The Money Market

– C: The issue is currently vulnerable to nonpayment and is dependent upon fa-

vorable business, financial and economic conditions for the obligor to meet its
financial commitment.
– D: The issue is in payment default. The rating is also used upon the filing of a
bankruptcy petition.

A-1, A-2, and A-3 are all investment-grade ratings.

Fitch’s Rating Scale

Fitch’s short-term ratings indicate the potential level of default within a 12-month pe-
riod. A plus or minus sign, may be appended to the F1 rating to denote relative status
within the category.
– F1: Best quality grade. Indicates strong capacity of obligor to meet its financial
commitment.
– F2: Good quality grade. Indicates satisfactory capacity of obligor to meet its finan-
cial commitment.
– F3: Fair quality grade. Adequate capacity of obligor to meet its financial commit-
ment, but near term adverse conditions could impact the obligor’s commitments.
– B: Of speculative nature. Obligor has minimal capacity to meet its commitment
and is vulnerable to short-term adverse changes in financial and economic condi-
tions.
– C: Possibility of default is high. Financial commitment of the obligor is dependent
upon sustained, favorable business and economic conditions.
– D: The obligor is in default as it has failed on its financial commitments.

Short-term debt issues given a top-grade credit rating by both S&P and Moody’s are
referred to as A1/P1 debt. A1/P1 paper sells at the finest rates.

Bills of Exchange
A bill is an unconditional order addressed by the party that is to be paid, to the coun-
terparty that is required to make the payment. The party that makes out the bill is
termed the drawer. The party that is required or directed to pay is termed the drawee.
Thus the bill is drawn by the drawer on the drawee. The latter is required to make the
payment to the former, as per the terms of the document.
There are three categories of bills of exchange: Treasury bills, bank bills, and trade
bills. A trade bill is drawn by one non-bank company on another, typically demanding
payment for a trade debt. A bank bill, on the other hand, is a bill that is drawn on and
payable by a commercial bank. A bank bill is generally considered safer than a trade
Bills of Exchange | 181

bill. Treasury bills are issued by the federal government. They are considered to be
the safest of the bills in the market for a given maturity. Trade bills can be classified as
sight bills, and time, term or usance bills. A sight bill has to be honored on demand,
that is, the drawee is expected to pay on sight. In the case of a term bill, however,
the specified amount is payable on a future date that is mentioned in the bill. For a
bill with a future payment date, the drawee signs its acceptance across the face of the
bill and returns it either to the drawer or to its bank. When such a bill is accepted by
the debtor, it becomes a promise to pay or an IOU. Because a term bill represents an
undertaking to pay a stated amount at a future date, it is a form of short-term finance
for the debtor. The holder of the bill therefore effectively gives credit to the debtor.
When the drawee stamps its acceptance on a term trade bill, it becomes what is
called a trade acceptance. The drawer can hold the accepted bill until maturity and
present it to the drawee for payment, or sell it in the money market at any time prior to
its maturity date. A bill of exchange can thus be transferred by a simple endorsement.
An accepted bill is essentially a zero coupon security. The last holder, at maturity, re-
ceives the stated value from the drawee. Thus a bill can be repeatedly traded in the
secondary market, at a price that is determined by the prevailing yield for such secu-
rities. The ability of a holder of a bill to sell it at a reasonable price depends on the
credit quality of the drawee and on the existence of a liquid secondary market.
In most commercial transactions, one or more banks enter the picture. For in-
stance, if a U.S. company imports goods from Scania, a Swedish company, the foreign
company will require the American firm to have a letter of credit opened in its favor.
When the goods have been shipped, Scania has the bills sent to the U.S. importer’s
bank. If it is a sight bill, the U.S. bank makes the payment immediately. If it is a time
draft, the bank signs its acceptance on the bill. A bill that is accepted by a commercial
bank is termed a bankers’ acceptance or BA. BAs are obviously more marketable than
trade acceptances. The transactions are usually with recourse. That is, assume Scania
sends the bill to Citibank which accepts it. Subsequently the bill is sold by Scania to
Siemens. On the maturity date, Siemens presents the bill to Citibank. If Citibank is not
in a position to pay, the holder, Siemens in this case, can demand that the drawer,
Scania, make the payment.

Example 6.22. Scania has drawn a bill on Midland Bank for $10 million, with a maturity of 180 days.
The bank has accepted it, and the drawer has sold it to HSBC at a discount of 3.60%. Now, 60 days
hence, HSBC has sold the bill to First National Bank at a discount of 2.70%. The rate of return for HSBC
may be computed as follows.
The purchase price is given by

180
10,000,000 × [1 − 0.036 × ] = $9,820,000
360

The sale price is given by

120
10,000,000 × [1 − 0.027 × ] = $9,910,000
360
182 | Chapter 6: The Money Market

The profit is

9,910,000 − 9,820,000 = $90,000

The return on investment on a 365-day year basis is

90,000 365
ROI = × = 5.5753%
9,820,000 60

The yield on bankers’ acceptances is usually only slightly higher than the rate on T-
bills because banks that issue such acceptances are normally large and have a good
reputation from the standpoint of credit risk. The discount rate on an acceptance is
also comparable with the rate on an NCD, because both instruments are unconditional
obligations of the issuing bank.

Chapter Summary
In this chapter, we examined various money market or short-term debt instruments.
We defined the key dates in such transactions and looked at market conventions in
the event of the maturity dates being market holidays. We began by looking at trans-
actions in the interbank market. In this context, we introduced the concepts of LIBOR
and LIBID. In Chapter 4, we looked at the term structure of interest rates, and in this
chapter we extended it to money markets. Treasury bills, which are one of the most
important constituents of the money market, were studied in detail. We looked at con-
cepts such as the money market yield, the bond equivalent yield, the holding period
return, and the tail. We then demonstrated the use of Excel functions to compute these
yield/return measures. Repurchase agreements and reverse repurchase agreements
were the next topic of discussion. We looked at repos from the perspective of a securi-
ties dealer, as well as a central bank. Negotiable certificates of deposit and term CDs
were studied next, followed by commercial paper. The chapter concluded with a look
at bills of exchange, trade acceptances, and bankers’ acceptances. In the following
chapter, we examine issues pertaining to floating rate bonds.
Chapter 7
Floating Rate Bonds
Unlike plain vanilla bonds, whose coupons remain fixed from issue until maturity,
floating rate notes and bonds, also referred to as floaters, are debt securities whose
coupons are reset periodically based on a reference or benchmark rate. Typically the
coupon on such a security is defined as Benchmark Rate + Quoted Margin.
If the benchmark rate is a short-term rate, say 6-M LIBOR, we refer to the bond as a
floating rate bond.1 However, if the benchmark is a long-term rate, the bond is referred
to as a variable or adjustable rate bond. For instance, consider a security whose coupon
is specified as five-year T-note rate + 75 basis points.
In this case, the reference or benchmark rate is the yield on a five-year T-note, and
consequently, the bond is classified as a variable rate bond. It should be noted that the
quoted margin need not always be positive. For instance, a floater can have a coupon
rate specified as 6-M LIBOR – 25 bp.
Usually the quoted margin remains constant for the life of the floater. However,
there are securities, known as stepped spread floaters, where the margin is reset at
intervals. In the case of such securities, the margin may be increased or decreased.

Call and Put Provisions in Floating Rate Bonds

Floating rate notes often come with call and/or put options. A call option may be in-
voked by the issuer if the reference rate were to decline. In such a situation, the is-
suer can recall a floater and issue a fixed rate bond at the prevailing rate, which is
lower by assumption. The other situation in which a floater can be recalled is when
the currently required margin for the security is lower than what was specified earlier.
In such a situation, the floater can be recalled and replaced with a security carrying
a lower quoted margin. As we have seen earlier in the case of fixed coupon securities,
the call option works against the investors. If they receive the principal prematurely,
they have to reinvest it in either a security with a lower coupon, or a security with a
smaller quoted margin.
Floaters can come with put options, wherein the holder can surrender the bond
prior to maturity. If rates rise in the market, and the floater is putable, the holder can
return the bond, recover the principal, and reinvest it in a bond carrying a higher
coupon, or in a floater with a higher quoted margin.

1 By short-term, we mean it is for a period less than or equal to one year.

https://doi.org/10.1515/9781547400669-007
184 | Chapter 7: Floating Rate Bonds

Caps and Floors for the Coupon Rate

A floater can come with a cap and/or a floor. A cap puts an upper limit on the coupon
and protects issuers against rising rates. A floor on the other hand puts a lower limit
and protects investors against falling rates. If we assume that interest rates cannot
become negative, a floater has a natural floor equal to the quoted margin because the
lowest possible value for the reference rate is zero. Some floaters come with both a
cap and a floor, which means that there is a minimum coupon as well as a maximum
coupon. Such a feature is referred to as a collar.

Symbols: An Important Note of Caution

In some of the mathematical results in this chapter, the symbols for interest rates, both
reference rates and yields, represent periodic (semiannual) rates, whereas in others
they represent annual rates. This has been done deliberately to avoid cluttering the
equations. Wherever required, an explanation has been provided. Apologies to the
readers for any confusion.

Valuation of a Floating Rate Bond

In the case of a default risk-free floating rate bond, the price of the security always
resets to par on a coupon date, although in between two coupon dates, it may sell at
a premium or a discount. For instance, consider a floater with a coupon equal to the
6-M LIBOR. Assume that there are two periods to maturity. The price of the bond at the
end of the first coupon period is given by
cT−1
M+M× 2
PT−1 = yT−1 (7.1)
(1 + 2
)

In this equation cT−1 is the coupon on the penultimate coupon date, and yT−1 is the
required yield on that day. On the coupon reset date, the yield is equal to the coupon
because we have assumed that there is no default risk implicit in the security. Conse-
quently any change in the required yield, also is reflected in the coupon that is set on
that day. We know that if the yield is equal to the coupon, then the bond should sell
at par. Thus PT−1 = M. Now let’s go to the previous coupon date, which in this case is
time 0, because the bond by assumption has only two coupons until maturity.
cT−2
PT−1 + M × 2
PT−2 = yT−2
(1 + 2
)
cT−2
M+M× 2
= y (7.2)
(1 + T−2
2
)
Variations on the Floating Rate Feature | 185

Once again at T − 2, cT−2 = yT−2 and consequently PT−2 = M. This logic can be applied
to a bond with any number of coupons remaining to maturity.
In between two coupon dates, however, the price of such a floater may not be equal
to par. Consider the valuation of the note at time T − 2 + k. The price is given by2
cT−2
PT−1 + M × 2
PT−2+k = 1−k
(7.3)
yT−2+k
(1 + 2
)

Although cT−2 was set at time T − 2 and is equal to yT−2 , yT−2+k is determined at time
T − 2 + k and reflects the prevailing conditions at that point in time. In general yT−2+k
need not equal cT−2 and may be higher or lower. Consequently, in between two coupon
dates, a floater may sell at a premium or at a discount.
Interestingly, if yT−2+k = cT−2 , the price is

yT−2+k k
M × (1 + ) >M
2
Thus, if the yield on a date between coupons is equal to the coupon rate that was set
on the preceding coupon date, the bond sells at a premium.
Now let’s consider the case of floaters characterized by default risk. In this case, it
is not necessary that the risk premium required by the market be constant over time.
For instance, assume that when the bond was issued the required return was equal to
the five-year T-note rate + 75 bp. The coupon was set equal to this rate, and the bond
was sold at par. Six months hence, the issue is perceived to be riskier and the required
return in the market is the 5-year T-note rate + 95 bp. However, the coupon resets at
the prevailing T-note rate + 75 bp. Hence, this issue does not reset to par at the next
coupon date.

Variations on the Floating Rate Feature

Having looked at plain vanilla floating rate notes, we now go on to look at securi-
ties with different features. Although these bonds have the floating rate characteristic,
they have additional features inherent in them.

Inverse Floating Rate Bonds

Unlike a floating rate bond, whose coupon increases with an increase in the reference
rate, for an inverse floater, the higher the reference rate is, the lower the coupon and

2 We denote the time that has elapsed since the previous coupon by k. k is stated in terms of the num-
ber of periods. This is different from the convention followed in the earlier discussion on the valuation
of a bond between coupon dates, where we had defined k as the time until the next coupon.
186 | Chapter 7: Floating Rate Bonds

vice versa. Thus, the term inverse floater arises because the coupon and the reference
rate move in opposite directions. The coupon for an inverse floater is specified as

Coupon = K − L × Reference Rate (7.4)

L is termed as the leverage factor, and an inverse floater with L > 1 is termed as a
leveraged inverse floater. Assume K = 4.5% and L = 1.0. Assume the reference rate is
3-M LIBOR. Let’s also assume that there is a lower bound of zero for LIBOR, although
negative interest rates have appeared on the horizon in some markets. In this situa-
tion, the maximum coupon on the inverse floater is 4.50%. This is termed as a cap. An
alternative method of preventing a negative interest rate, is to specify a floor. For in-
stance, the coupon in Equation (7.4) can be specified as subject to a floor of say 1.50%.
Thus the coupon cannot decline below 1.50%, no matter how high the LIBOR. Now
let’s illustrate the impact of the leverage factor with the help of a numerical example.

Example 7.1. Assume that the coupon is specified as

Coupon = 7.50 − 2.0 × 3-M LIBOR

If on a coupon reset date LIBOR = 2.50%, the coupon is 2.50%. If the leverage factor had been 1.0,
the coupon would be 5.00%. If the LIBOR is 3%, the coupon is 1.50%. In the absence of the leveraging
effect, it would be 4.50%. Thus a 50-basis point increase in the reference rate reduces the coupon by
100 basis points. This illustrates that the leverage factor is 2.0.

Deleveraged Floating Rate Bonds

In the case of a deleveraged floater, the leverage factor is less than 1.0. For instance, if
the coupon is specified as

Coupon = 4.50 − 0.75 × 3-M LIBOR

the leverage factor is 0.75. The impact of a change in the reference rate is illustrated in
Example 7.2.

Example 7.2. If on a coupon reset date LIBOR = 2.50%, the coupon is 2.6250%. If the leverage factor
had been 1.0, the coupon would be 2.00%. If the LIBOR is 3%, the coupon is 2.25%. In the absence of
the leveraging effect, it would be 1.50%. Thus a 50-basis point increase in the reference rate reduces
the coupon by 37.50 basis points. This illustrates that the leverage factor is 0.75.
Variations on the Principal Repayment Feature | 187

Dual-Indexed Floating Rate Bonds

The coupon for a dual-indexed floating rate security is a function of the difference
between two reference rates, denoted as RR1 and RR2, respectively. Thus,

Coupon = K + (RR1 − RR2) (7.5)

Range Notes

In this case, the coupon is equal to the reference rate, as long the reference rate lies in
a specified range. If the reference rate goes outside the bound, however, the coupon
becomes zero. Here is an example.

Example 7.3. Coupon = 3-M LIBOR, if LIBOR lies between 2.50% and 4.50%. Otherwise, the coupon
is zero. So if the LIBOR is 3.25%, the coupon also is 3.25%. If the LIBOR is 2.00%, the coupon is zero.
Similarly, if the LIBOR is 5.25%, the coupon again is zero.

Variations on the Principal Repayment Feature

A floater may have a stated time to maturity, like a plain vanilla bond, or else it may be
a perpetuity, where the principal is never repaid. There also exist amortized floating
rate notes, in which the principal is repaid in installments, starting with a coupon date
prior to the maturity date.

Floaters, Inverse Floaters, and Plain Vanilla Bonds

Take the case of a floater that pays LIBOR and an inverse floater, with the same time to
maturity, that pays 8% – LIBOR. Assume that LIBOR is lower than 8% during the life
of the securities. Let’s assume that there are five years to maturity, and the observed
values of LIBOR at each coupon reset date are as shown in Table 7.1.

Table 7.1: Observed LIBOR values.

Time LIBOR

0 3.75%
1 4.50%
2 5.25%
3 6.00%
4 5.75%
188 | Chapter 7: Floating Rate Bonds

Table 7.2: Coupons for the portfolio of a floater and an inverse floater.

Time Coupon from Coupon from Portfolio

Floater Inverse Floater Coupon

1 3.75 4.25 8.00

2 4.50 3.50 8.00
3 5.25 2.75 8.00
4 6.00 2.00 8.00
5 5.75 2.25 8.00

Take a portfolio that consists of one floater and one inverse floater. The total coupon
on the portfolio is depicted in Table 7.2.
If we assume that both the floater and the inverse floater have a face value of
$1,000, the bond represented by the portfolio has a face value of $2,000. A cash flow
of $80 per period on a face value of $2,000 represents a coupon of 4%. Thus the com-
bination of the floater and the inverse floater is equivalent to two plain vanilla bonds
with a coupon of 4% each.
Hence, to rule out arbitrage, the price of the floater plus the price of the inverse
floater should be equal to twice the price of a plain vanilla bond paying a coupon of
4%.

A More General Relationship for an Inverse Floater

Consider an inverse floater whose coupon rate is K −L×r, where L is the leverage factor
and r is the reference rate, such as LIBOR. A floater with the same reference rate has
a coupon of r. If we combine one inverse floater with L floaters, we get a bond with a
cumulative face value of (L + 1) × M and a coupon of M × K. This is equivalent to one
K
fixed rate bond with a face value of M and a coupon rate of (1+L) .
In our illustration, L was equal to 1.0 and K was 8%. Thus the coupon rate of the
equivalent fixed rate bond was 4%.

Duration of a Floating Rate Bond

Consider a default risk-free floater. The price of the bond when it is between time T − N
and T − N + 1 is given by
cT−N
M(1 + 2
)
PT−N+k = 1−k
(7.6)
yT−N+k
(1 + 2
)
c
𝜕PT−N+k (1 − k) M(1 + T−N2
)
yT−N+k = − × (7.7)
𝜕 2 2 y
(1 + T−N+k )
2−k
2
Convexity of a Risk-Free Floating Rate Bond | 189

(1−k)
𝜕PT−N+k 2
⇒ y =− yT−N+k (7.8)
P𝜕 T−N+k
2
(1 + 2
)

We know that

𝜕PT−N+k
y = −DM
P𝜕 T−N+k
2

or the modified duration at that point in time. Thus

yT−N+k
(1−k)
2
Duration = y × (1 + )
(1 + T−N+k
2
) 2
(1 − k)
= (7.9)
2

The duration of the floater is equal to the time left until the next coupon.

Convexity of a Risk-Free Floating Rate Bond

We can differentiate the price of a risk-free bond twice with respect to the yield, in
order to derive an expression for its convexity.

𝜕2 P
(1−k)(2−k)
4
= (7.10)
P𝜕y2 (1 + yT−N+k 2−k
)
2

Comparison with a Zero Coupon Bond

A risk-free floater that is between coupons is equivalent to a zero coupon bond matur-
ing on the next coupon date. The duration of a zero coupon bond with N semiannual
periods to maturity is N/2 in years, and the convexity is N(N+1)
y 2.
4(1+ 2 )
If we set k = 0, the duration of a risk-free floater is 21 a year and the convexity
1×2
is yT−N 2 . This confirms the deduction that a risk-free floater paying semiannual
4(1+ 2
)
coupons is, on a coupon date, equivalent to a zero coupon bond with one semiannual
period to maturity.

Margin Measures for Floaters

A margin measure is essentially a yield spread measure, and such measures are used
to evaluate the returns from a floater. There are various types of margins that can be
computed in practice. Let’s consider each one of them in turn.
190 | Chapter 7: Floating Rate Bonds

– Simple margin
– Modified simple margin
– Adjusted simple margin
– Adjusted total margin
– Discount margin

Simple Margin

If a floater trades at a premium or a discount, then the premium/discount needs to

be factored in while computing the return. If the floater trades at a discount, then we
need to account for the accretion of the discount over the life of the floater. On the
contrary, if the floater trades at a premium, we need to factor in the amortization of
the premium over the life of the security. The formula for the simple margin is

100 × (100 − P)
Simple Margin = [ + Quoted Margin]
Tm

P is the market price (clean price) of the floater, as a percentage of par, and Tm is the
remaining time until maturity. If it is not an integer, we need to invoke a day-count
assumption. The quoted margin must be expressed in basis points and not percentage
terms. There is a related metric called modified simple margin, which is defined as

100 × (100 − P) 100

Modified Simple Margin = [ + Quoted Margin] (7.11)
Tm P

Example 7.4. A floater is trading at a price of $98.75. It has one year and nine months to maturity. If
we assume a 30/360 day-count convention, the time to maturity is 1.75 years. The quoted margin is 75
basis points.

100 × (100 − 98.75)

Simple Margin = [ + 75] = 146.43 basis points
1.75
100 × (100 − 98.75) 100
Modified Simple Margin = [ + 75] ×
1.75 98.75
= 148.28 basis points

Adjusted Simple Margin

Bond dealers fund the bulk of their inventory with borrowed money. The funding
transaction usually entails the execution of a repurchase agreement or repo. The fund-
ing cost must be incorporated to compute what is termed as the adjusted price. It
should be noted that the bond holder earns a return in terms of the coupon and in-
curs a cost in terms of the repo rate. Let A be the number of days from settlement to
Margin Measures for Floaters | 191

the next coupon date, and B the number of days in the coupon period. If we assume
a 30/360 day-count convention and that there are 1.75 years to maturity, A = 90 and
B = 180. Thus the bond must be financed for 90 days; the accrued interest is for 90
days; and on the next coupon date, a half-year’s coupon is paid out. The adjusted price
is defined as
90 180
(P + AI) × r × 360
− c × 100 × 360
PA = (P + AI) + 90
(7.12)
(1 + i × 360
)

In this equation, c is the coupon rate, r is the repo rate, and i is the current discount
rate.3 The numerator of the second term represents the net cash flow at the time of the
next coupon. We have to discount it to find the present value.
100 × (100 − PA ) 100
Adjusted Simple Margin = [ + Quoted Margin] (7.13)
Tm PA

Let’s illustrate the calculation with the help of an example.

Example 7.5. Let’s use the data given in Example 7.4. The clean price is $98.75. Assume the coupon
rate is 6% per annum and the repo rate is 5.40% per annum. Assume that the current 3-M LIBOR is
5.70% per annum. There are 1.75 years until expiration, which means that 90 days of interest has
accrued. The accrued interest is

100 × 0.03 × 0.5 = $1.50

Thus the dirty price is $100.25. The funding cost is 100.25 × 0.054 × 0.25 = $1.3534 The coupon for six
months is $3.00. So, the adjusted price is

1.3534 − 3.00
PA = 98.75 + 1.5 + = $98.6265
(1 + 0.057 × 0.25)

The effective margin can be computed by substituting the adjusted price in the formula for the modified
simple margin:

100 × (100 − 98.6265) 100

Adjusted Simple Margin = [ + 75] ×
1.75 98.6265
= 155.62 basis points

Adjusted Total Margin

This margin measure incorporates the interest earned by investing the difference be-
tween the bond’s face value and the adjusted price as computed in the preceding sec-
tion. The interest is assumed to be based on the prevailing value of the reference rate.
Here is an example.

3 If the benchmark is the 3-M LIBOR, then i is the current value of the 3-M LIBOR.
192 | Chapter 7: Floating Rate Bonds

Example 7.6. Adjusted Total Margin

100 × (100 − PA ) 100
=[ + Quoted Margin + 100 × (100 − PA ) × r]
Tm PA
We have assumed the 3-M LIBOR to be 5.70%. Thus the interest

= 100 × (100 − PA ) × 0.057 = 100 × (100 − 98.6265) × 0.057 = 7.8290

And the adjusted total margin is

100 × (100 − 98.6265) 100
=[ + 75 + 7.8290] ×
1.75 98.6265
= 163.56 basis points

The Discount Margin

The discount margin is based on a set of assumptions and arrived at by a process of

trial and error, although with Excel, it is a fairly easy task. The first assumption is about
the evolution of LIBOR, which we can assume is the reference rate, over the life of the
security. The common practice is to assume that the current LIBOR will be applicable
for all future periods. Based on this assumption, and given the quoted margin, we can
project the cash flows over the life of the security. We then need to assume a margin
for discounting and add it to the current LIBOR. This rate is used to discount all the re-
maining cash flows. Typically the present value that we obtain is either lower or higher
than the actual price. If it is lower, we need to assume a lower discount margin and
repeat the process. On the other hand, if it is higher, we need to repeat the procedure
by assuming a higher value. Obviously, if the security is trading at par on a coupon
date, the discount margin is equal to the quoted margin.
Let’s revisit the data used in Example 7.5.

Example 7.7. The dirty price is $100.25. The 3-M LIBOR is 5.70%, and the quoted margin is 75 bp.
There are 1.75 years to maturity. Let’s go 0.25 years ahead in time. We will get a cash flow of $3.225.
The three remaining cash flows will be 3.225, 3.225, and 103.225. Assume the discount margin is 60
bp. The present value of the cash flows is
3.225 3.225 103.225
3.225 + + +
1.0315 (1.0315)2 (1.0315)3
= $103.4365

The value today is

103.4365
= $101.8449
(1.0315)0.5
The actual price is $100.25. So we need a higher value. Using Goal Seek in Excel, we find the discount
rate to be 7.2844% per annum. Given the current LIBOR of 5.70%, the discount margin is 158.44 basis
points.
Inflation Indexed Bonds | 193

There are obviously certain limitations as far as the discount margin is concerned. It
is based on the major assumption that the reference rate, which is 3-M LIBOR in our
case, will remain constant for the life of the security. If we relax this assumption, the
value changes.

Inflation Indexed Bonds

Many countries issue bonds, whose cash flows are determined by the prevailing infla-
tion rates, such that the cash flow is higher for the holders, when the rate of inflation
is high, and lower when the rate is low. The rationale is to ensure that, in real terms,
as measured by the purchasing power of the cash flows, the rate of return is reason-
ably constant. In the U.S., the Treasury issues TIPS, or Treasury Inflation-Protected
Securities.
There are two ways in which the cash flows from a bond can be adjusted. One way
is by adjusting the coupons, keeping the face value constant. This is similar to the case
of a floating rate bond, where if the inflation is high, the coupon is high, and vice versa.
Such securities are termed Coupon-Linkers or C-Linkers for short. The alternative is to
keep the coupon rate constant and adjust the principal periodically for inflation to get
an adjusted principal. Every period coupons are paid on the adjusted principal. These
securities are termed Principal-Linkers or P-Linkers. In either case, the cash flows per
period are variable and related to the rate of inflation.
Although we generally expect cash flows to increase over time on account of in-
flation, there could be a rare deflationary situation, where inflation rates are actually
negative. In such a situation, the adjusted principal, for a P-Linker, may be less than
the original par value. In these cases, most governments state at the outset that they
will pay back the original par value or the terminal adjusted principal, whichever is
higher. Also, in the case of a C-Linker, if there is a sharp decline in the price level, and
the inflation adjusted rate is negative, most issuers pay a coupon of zero.

Principal Linkers or P-Linkers

Let’s illustrate the cash flows from a P-Linker with the help of an example.

Example 7.8. Consider a P-Linker with a face value of $1,000 and six years to maturity. The coupon is
7.50% per annum, paid annually. The values of the consumer price index (CPI) at various points in time
are as depicted in Table 7.3.
The rate of inflation for a year is computed by dividing the index value at the end of the year by
the corresponding value at the beginning of the year:
CPIt
πt = ( − 1) × 100
CPIt−1

We can now derive the cash flows and the IRR for a P-Linker.
194 | Chapter 7: Floating Rate Bonds

Table 7.3: Price index values & inflation rates.

Time CPI Inflation

Rate

0 100.0000 –
1 110.0000 10.00%
2 115.5000 5.00%
3 122.4300 6.00%
4 112.6356 −8.00%
5 123.8992 10.00%
6 139.3866 12.50%

Table 7.4: Cash flows for a P-Linker.

Time Adjusted Nominal Real

Principal Cash Flow Cash Flow

0 1,000.00 (1,000) (1,000)

1 1,100.00 82.5000 75
2 1,155.00 86.6250 75
3 1,224.30 91.8225 75
4 1,126.3560 84.4767 75
5 1,238.9920 92.9244 75
6 1,393.8660 1,498.4060 1,075
IRR 13.5871% 7.50%

Analysis

Let’s define the index ratio for the year as It /I0 , that is, the ratio of the CPI at the end
of the year and the CPI at the very beginning. The adjusted principal at the end of the
year is the original principal of $1,000 multiplied by the index ratio. For instance, at
the end of the third year, the CPI is 122.43, which means the index ratio is 1.2243. So
the adjusted principal is 1,000 × 122.43
100
= $1,224.30. The nominal cash flow for this year
is 0.075 × 1,224.30 = $91.8225. The real cash flow for this period is the nominal cash
122.43
flow divided by the index ratio. In this case, it is 91.8225 ÷ [ 100.00 ] = $75.
This can be interpreted as follows. The inflation for the third year is 6%. The value
of the cash flow in terms of period-2 prices is 91.8225
1.06
= $86.6250. The value of this in
86.6250
terms of period-1 prices is 1.05 = $82.50. Finally, the value of this in terms of period-
0 prices is 82.50
1.10
= $75. If we denote the inflation for the year i as πi , we need to discount
the cash flow by:

(1 + π1 ) × (1 + π2 ) × (1 + π3 )
I3 I2 I1 I
= × × = 3
I2 I1 I0 I0
Coupon Linkers or C-Linkers | 195

The terminal nominal cash flow is the final adjusted principal of $1,393.866 plus that
year’s coupon, which is 7.50% of this amount. The IRR of the nominal cash flows is
13.5871%. The real cash flows stay constant, and the IRR is equal to the coupon of
7.50% per annum.

Coupon Linkers or C-Linkers

Let’s illustrate the cash flows from a C-Linker with the help of an example.

Example 7.9. Now let’s consider the case of the C-Linker. The coupon every period is the rate an-
nounced at the outset plus the rate of inflation for the year. The coupon is calculated on the original
principal of $1,000. That is, the principal stays constant, unlike in the case of the P-Linker, whereas
the coupon varies from period to period.

Table 7.5: Cash flows for a C-Linker.

Time Inflation Adjusted Nominal Real

Rate Coupon Cash Flow Cash Flow

0 – – (1,000) (1,000)
1 10.00% 17.50% 175.00 159.0909
2 5.00% 12.50% 125.00 108.2251
3 6.00% 13.50% 135.00 110.2671
4 −8.00% 0.00% 0.00 0.00
5 10.00% 17.50% 175.00 141.2438
6 12.50% 20.00% 1,200.00 860.9149
IRR 13.4434% 7.1793%

Valuing a Risky Floater

In the case of a risk-free floater, the price resets to par on every coupon date because
the required margin is always equal to the quoted margin. In the case of a risky floater,
however, the required margin is usually different from the quoted margin, which has
implications for its valuation.
The general formula for the price of a floating rate bond on a coupon date, when
there are N periods until maturity, is4

M(r1 + QM) M(r2 + QM)

P0 = + + − − − − − − −−
(1 + r1 + DM) (1 + r1 + DM)(1 + r2 + DM)

4 ri represents the benchmark rate for period i.

196 | Chapter 7: Floating Rate Bonds

M(rN−1 + QM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)
M + M(rN + QM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)(1 + rN + DM)
(7.14)

If QM = DM, P0 = M. Both r +QM and r +DM should be interpreted in periodic terms. If

LIBORi +QM
the reference rate for a six-monthly period-i is LIBORi per annum, ri + QM = 2
LIBOR +DM
and ri +DM = 2
i
. To value a floating rate bond, we have to make an assumption
about the one period forward rates for the life of the security. In the case of a risk-
free floater, the price on a coupon date is always equal to the par value, irrespective
of whether the term structure is flat, upward sloping, or downward sloping, as we
shall now demonstrate. Consider the LIBOR rates at a point in time, as depicted in
Table 7.6. The first rate is the current one period spot rate, and the rest are one period
forward rates for subsequent periods. Assume that the discount margin is equal to the
quoted margin and is 50 bp. Assume that all rates are per annum and are reset once
per annum.

Table 7.6: Cash flows for a floater trading at par when the yield curve is upward sloping.

Rate LIBOR Cash PV of

Value Flow Cash Flow

Spot rate for year 1 2.50% 30 29.1262

Forward rate for year 1–2 3.50% 40 37.3413
Forward rate for year 2–3 4.50% 50 44.4539
Forward rate for year 3–4 5.50% 60 50.3252
Forward rate for year 4–5 7.00% 1,075 838.7534
Bond Price 1,000.00

Pricing Equation
The price of the bond, given the assumed term structure, may be stated as follows.

30 40 50 60
Price = + + +
(1.03) (1.03)(1.04) (1.03)(1.04)(1.05) (1.03)(1.04)(1.05)(1.06)
1,075
+ = $1,000.00
(1.03)(1.04)(1.05)(1.06)(1.075)

Now assume a flat term structure where LIBOR is equal to 2.50% per annum for
all periods.
Valuing a Risky Floater | 197

Table 7.7: Cash flows for a risk-less floater when the yield curve is flat.

Rate LIBOR Cash PV of

Value Flow Cash Flow

Spot rate for year 1 2.50% 30 29.1262

Forward rate for year 1–2 2.50% 30 28.2779
Forward rate for year 2–3 2.50% 30 27.4542
Forward rate for year 3–4 2.50% 30 26.6546
Forward rate for year 4–5 2.50% 1,030 888.4870
Bond Price 1,000.00

Finally let’s look at inverted term structure.

Table 7.8: Cash flows for a risk-less floater when the yield curve is downward sloping.

Rate LIBOR Cash PV of

Value Flow Cash Flow

Spot rate for year 1 7.00% 75 69.7674

Forward rate for year 1–2 5.50% 60 52.6547
Forward rate for year 2–3 4.50% 50 41.7894
Forward rate for year 3–4 3.50% 40 32.1457
Forward rate for year 4–5 2.50% 1,030 803.6428
Bond Price 1,000.00

Thus when the quoted margin is equal to the discount margin, a floater trades at par
on a coupon date, irrespective of the assumed values for the forward rates.
Now let’s consider a situation where the quoted margin is different from the dis-
count margin:

M(r1 + QM) M(r2 + QM)

P0 = + + − − − − − − −−
(1 + r1 + DM) (1 + r1 + DM)(1 + r2 + DM)
M(rN−1 + QM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)
M + M(rN + QM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)(1 + rN + DM)
(7.15)
M(r1 + DM) M(r2 + DM)
⇒ P0 = + + − − − − − − −−
(1 + r1 + DM) (1 + r1 + DM)(1 + r2 + DM)
M(rN−1 + DM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)
M + M(rN + DM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)(1 + rN + DM)
198 | Chapter 7: Floating Rate Bonds

M(QM − DM) M(QM − DM)

+ + + − − − − − − −−
(1 + r1 + DM) (1 + r1 + DM)(1 + r2 + DM)
M(QM − DM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)
M(QM − DM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)(1 + rN + DM)
(7.16)
M(QM − DM) M(QM − DM)
=M+ + +−−−−−−−
(1 + r1 + DM) (1 + r1 + DM)(1 + r2 + DM)
M(QM − DM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)
M(QM − DM)
+
(1 + r1 + DM)(1 + r2 + DM). . .(1 + rN−2 + DM)(1 + rN−1 + DM)(1 + rN + DM)
(7.17)
In this case, the price will be greater than par if the quoted margin is greater than
the discount margin, and less than par if the former is less than the latter. However,
the price is a function of the interest rate path that is assumed, as we shall now demon-
strate. Let’s first consider a flat term structure, such that all one-period rates are 2.50%
per annum. Assume the quoted margin is 50 bp., and the discount margin is 75 bp. The
details of the price computation are given in Table 7.9.

Table 7.9: Valuation of a floater whose quoted margin is less than its discount margin when the term
structure is flat.

Rate LIBOR Cash Discount PV of

Value Flow Rate Cash Flow
Spot rate for year 1 2.50% 30 3.25% 29.0557
Forward rate for year 1–2 2.50% 30 3.25% 28.1411
Forward rate for year 2–3 2.50% 30 3.25% 27.2553
Forward rate for year 3–4 2.50% 30 3.25% 26.3974
Forward rate for year 4–5 2.50% 1,030 3.25% 877.7825
Bond Price 988.6320

The present value of the annuity is:

2.5 1
× [1 − ] = 11.3680
0.0325 (1.0325)5
1,000 − 11.3680 = 988.6320

Because the term structure by assumption is flat, the yield, as computed by the IRR of
the cash flows, is also 3.25%.
Now let’s consider an upward sloping term structure, with the same values for the
discount margin and the quoted margin.
Duration of a Risky Floater | 199

Table 7.10: Valuation of a floater whose quoted margin is less than its discount margin when the
term structure is upward sloping.

Rate LIBOR Cash Discount PV of

Value Flow Rate Cash Flow

Spot rate for year 1 2.50% 30 3.25 29.0557

Forward rate for year 1–2 3.50% 40 4.25 37.1616
Forward rate for year 2–3 4.50% 50 5.25 44.1349
Forward rate for year 3–4 5.50% 60 6.25 49.8464
Forward rate for year 4–5 7.00% 1,075 7.75 828.8463
Bond Price 989.0449

The present value of the annuity is

1 1 1
2.50 × [ + +
(1.0325) (1.0325)(1.0425) (1.0325)(1.0425)(1.0525)
1 1
+ + ]
(1.0325)(1.0425)(1.0525)(1.0625) (1.0325)(1.0425)(1.0525)(1.0625)(1.0775)
= 10.9551
1,000 − 10.9551 = 989.0449

The yield, using the IRR function, is 5.2427%.

Duration of a Risky Floater

The formula for the duration of a risky floater is

(1 + y) N −1 M(1 + r1 )
D = (1 − k) + [ − ] × [1 − ]
y [(1 + y)(N−1)
− 1] P0 (1 + y)(1−k)

We have derived a general formula for the duration of a floater between coupon dates,
in Appendix 7.1. The price of a risky floater on a coupon date is equal to the price of
a risk-free floater plus the present value of an annuity. The face value of the riskless
floater is M. The periodic cash flow from the annuity is M(QM−DM)
2
, which we denote
as A. y is the semiannual yield of the bond, which can be computed using the IRR
function in Excel; r1 is the semiannual coupon rate for the first period, which is equal
to the benchmark at time 0 plus the quoted margin; and k is the time elapsed since the
previous coupon. The market price P (the dirty price) of the floater on the next coupon
date is expressed as

A 1
P1 = M + [1 − ]
y (1 + y)(N−1)
200 | Chapter 7: Floating Rate Bonds

If A = 0 (that is, the DM = QM), or in other words the bond is riskless, then
M × (1 + r1 ) = P0 × (1 + y)1−k . Hence the duration in this case is (1 − k) semiannual
periods, which is what we obtained earlier.
N−1
(1 − k) ≥ 0 and the second term, [ (1+y)
y
− (N−1) ], also cannot be negative, since
[(1+y) −1]
M(1+r1 )
it represents the duration of an (N − 1) period annuity. However, [1 − ] may be
P0 (1+y)(1−k)
less than zero if P0 is very low, or in other words the quoted margin is much lower than
the discount margin. Consequently, in such a situation, the duration may be negative.
For a perpetual risky floater, the duration is given by

A(1 + y)k
(1 − k) +
P0 y 2

Example 7.10. Consider a floating rate bond with 0.10 half-years to the next coupon. The face value is
$1,000, and the coupon for the first semiannual period is 5.40% per annum. The yield is 9.60% per
annum, and there are 10 coupons remaining in the life of the bond. That is, N = 10. The current market
price is $915.1262. We have assumed that

300
M(QM − DM) 1,000 × 100×100
= = $ 15
2 2

That is, the difference between the quoted margin and the discount margin is 3% or 300 basis points.

1.048 9
(1 − k) = 0.10; − = 4.6884
0.048 [(1.048)9 − 1]
M(1 + r1 ) 1,000(1.027)
[1 − ] = [1 − ] = −0.1170
P0 (1 + y) (1−k)
915.1262(1.048)0.10

Thus the duration is 0.10 + 4.6884 × −0.1170 = −0.4485. In this case, the duration is negative, which
signifies that as the yield declines, the price will also decline.

Chapter Summary
This chapter examined bonds whose coupons change from period to period, what are
termed as floating rate or variable rate bonds. We also studied bonds known as in-
verse floaters and examined the relationship between floaters, inverse floaters, and
plain vanilla bonds. We studied variants of simple floaters, in the form of bonds with
call and put provisions, and bonds with caps, floors, and collars. We then derived for-
mulas for the duration and convexity of riskless floaters. The next focus of attention
was inflation-indexed bonds, both those for which the principal is linked to inflation,
P-Linkers, and those for which the coupon is linked to inflation, C-Linkers. We then
derived the valuation equation for risky floating rate bonds. We showed that while
the value of a riskless floater is independent of the assumption made about the term
structure of interest rates, risky floating rate bonds have a price that is dependent on
Appendix 7.1: Duration of a Risky Floater | 201

the assumption about the term structure. The chapter concluded with the derivation
of formulas for the duration of risky floating rate bonds, for both bonds with a finite
maturity, and for perpetual bonds.

Appendix 7.1: Duration of a Risky Floater

We derive a general formula for the duration of a floater between coupon dates. The
price of a risky floater on a coupon date is equal to the price of a risk-free floater plus
the present value of an annuity. The face value of the riskless floater is M. The periodic
cash flow from the annuity is M(QM−DM)
2
, which we denote as A. y is the semiannual
yield of the bond, which can be computed using the IRR function in Excel; r1 is the
semiannual coupon rate for the first period, which is equal to the benchmark value
at the outset plus the quoted margin; and k is the time elapsed since the previous
coupon. The market price P (the dirty price) of the floater on the next coupon date is
expressed as

A 1
P1 = M + [1 − ]
y (1 + y)(N−1)

The price at time T − N + k is expressed as

M(1 + r1 ) + Ay [1 − 1
]
(1+y)(N−1)
P0 =
(1 + y)(1−k)

A 1
Let’s define Z = y
[1 − ]
(1+y)(N−1)

𝜕P P × (1 + y)(1−k) 𝜕Z 1
= −(1 − k) × 0 + ×
𝜕y (1 + y)(2−k) 𝜕y (1 + y)(1−k)
𝜕P (1 − k) 𝜕Z 1
=− + ×
P𝜕y (1 + y) P𝜕y (1 + y)(1−k)
𝜕P
Duration = − × (1 + y)
P𝜕y

Thus in this case, the duration is

𝜕Z 1
D = (1 − k) − × × (1 + y)
P𝜕y (1 + y)(1−k)

Z is an N − 1 period annuity. Its duration is given by

𝜕Z (1 + y) N −1
− × (1 + y) = −
Z𝜕y y [(1 + y)(N−1) − 1]
202 | Chapter 7: Floating Rate Bonds

Thus

𝜕Z 𝜕Z Z (1 + y) N −1 Z
− × (1 + y) = − × (1 + y) × = [ − ]×
P𝜕y Z𝜕y P y [(1 + y)(N−1) − 1] P

And the duration of the risky floater is

(1 + y) N −1 Z 1
D = (1 − k) + [ − ]× ×
y [(1 + y)(N−1) − 1] P (1 + y)(1−k)
(1 + y) N −1
= (1 − k) + [ − ]
y [(1 + y)(N−1) − 1]
P0 (1 + y)(1−k) − M(1 + r1 ) 1
×[ × ]
P0 (1 + y)(1−k)
(1 + y) N −1 M(1 + r1 )
= (1 − k) + [ − ] × [1 − ]
y [(1 + y)(N−1) − 1] P0 (1 + y)(1−k)

Appendix 7.2: Duration of a Risky Perpetual Floater

The price of a perpetual floater is expressed as
A
M(1 + r1 ) + y
P0 =
(1 + y)1−k
A
𝜕P (1 − k) y2
=− 2−k
× P0 (1 + y)1−k −
𝜕y (1 + y) (1 + y)1−k
𝜕P (1 − k) A
=− −
P𝜕y 1+y P0 y2 (1 + y)1−k
𝜕P A(1 + y)k
Duration = − × (1 + y) = (1 − k) +
P𝜕y P0 y 2
Chapter 8
Mortgage Loans
The biggest aspiration of many people in this world is to own a home of their own.
And in many countries, particularly in the developed world, the right to own a home
is considered to be virtually a fundamental right. However, buying a home is usually
a very expensive proposition, and for most people it is probably the largest capital
expenditure that they will make in their lifetime. Consequently most of the funds re-
quired to buy a home must be borrowed. A loan that is collateralized by real estate is
called a mortgage loan. The lender is called the mortgagee, and the borrower is called
the mortgagor. Thus, in return for providing the capital, the borrower mortgages the
property to the lender. This is termed as a lien.

Important Mortgage-Related Terms

We will now discuss some commonly used terms pertaining to mortgage loans.
– Buy-down: This refers to a situation when the seller pays an amount to the lender
so that the lender can give a loan at a lower rate and with a lower installment, usu-
ally for an initial period in an adjustable rate mortgage (ARM), which is a mortgage
loan in which the rate is reset periodically. The seller may increase the price of the
home to cover the buy-down.
– Conversion clause: The presence of such a clause in a mortgage loan permits the
conversion of the loan from a floating rate to a fixed rate structure. An ARM may
have such a provision to facilitate the conversion of the loan to a fixed rate mort-
gage. This is usually permitted at the end of the first adjustment period. At the
time of conversion, the new fixed rate is set at the prevailing rate for fixed rate
mortgages.
– Equity: This refers to the difference between the current value of the home and the
outstanding balance on the loan.
– Points: A point or a discount point is equal to 1% of the principal. So if the prin-
cipal is $250,000, two points are $5,000. Points may be charged in the case of
both fixed rate and floating rate loans. These are charged to enable the lender to
cover loan origination expenses or provide additional compensation. There could
be situations where the money required to pay the points can be borrowed. This
obviously increases the amount of the loan. Discount points are points that are
voluntarily paid by borrowers in return for a reduced interest rate.

Risks in Mortgage Lending

Investors who invest in mortgage loans are exposed to four main risks. These are:
– Default risk

https://doi.org/10.1515/9781547400669-008
204 | Chapter 8: Mortgage Loans

– Liquidity risk
– Interest rate risk
– Prepayment risk

Default Risk

Default or credit risk is the risk that the borrower will default. In western countries,
for low-income families, government agencies provide the insurance. For such mort-
gages the risk is minimal because the insuring agencies are government sponsored.
For privately insured mortgages, the risk depends on the credit rating of the insurance
company. For uninsured mortgages, the risk obviously depends on the credit quality
of the borrower.

Liquidity Risk

An asset is said to be liquid if there are plenty of buyers and sellers available. If an asset
is actively traded, relatively large transactions can be put through without a significant
impact on the price. In the absence of liquidity, large buy orders may send the price
shooting up, whereas large sell orders may cause the price to come crashing down.
One of the indicators of liquidity is the size of the bid-ask spread. The higher the degree
of liquidity, the lower the spread is. Mortgage loans tend to be rather illiquid because
they are large and indivisible. That is, one cannot sell a fraction of a loan. Thus, while
an active secondary market exists for such loans, bid-ask spreads are large relative to
other debt instruments.

Interest Rate Risk

The price of a mortgage loan in the secondary market moves inversely with interest
rates, just like any other debt security. The lower the discount rate is, the higher the
present value and vice versa. However, among debt securities, mortgages are more
vulnerable to interest rate movements. This is because the loans are for long terms to
maturity, and consequently the impact of an interest rate change on the price of the
mortgage can be significant.

Prepayment Risk

Most homeowners pay off all or a part of their mortgage balance prior to the maturity
date. As we have seen in the case of amortized loans, every monthly payment con-
Risks in Mortgage Lending | 205

sists partly of a principal repayment, with the difference being the interest payment
on the outstanding balance at the end of the previous month. This principal compo-
nent is referred to as the scheduled principal for the month. However, at times the
homeowner may pay more than what is scheduled. Payments made in excess of the
scheduled principal repayments are called prepayments. Prepayments may occur for
one of several reasons. First, borrowers tend to prepay the entire mortgage when they
sell their home. The sale of the house may be due to
– A change of employment that necessitates moving
– The purchase of a more expensive home
– A divorce in which the settlement requires sale of the marital residence

Second, if market interest rates decline below the loan rate, the borrower may prepay
the loan and refinance at a lower rate. Third, in the case of homeowners who can-
not meet their mortgage obligations, the property will be repossessed and sold. The
proceeds from the sale are used to pay off the mortgage in the case of uninsured mort-
gages. For an insured mortgage the insurance company pays off the balance. Finally,
if the property is destroyed by an act of God, such as natural calamities, the insurance
proceeds are used to pay off the mortgage. The effect of prepayments, irrespective of
the reason, is that the cash flows from the mortgage become unpredictable. This is be-
cause the amount of principal received at the end of a month, depends on how much
the borrower chooses to prepay in that month, which depends on both economic and
non-economic factors.

Example 8.1. Carol Saunders has taken a mortgage loan for $800,000. It is an eight-year mortgage
with an interest rate of 8.40% per annum. Installments are due every year. At the end of the second
year, that is, just after she has paid the second installment, the interest on housing loans drops to
7.20% per annum. The refinancing fee is 2.75% of the amount being refinanced. Her opportunity cost of
funds is 6% per annum. That is, if she were to save a dollar, she could invest it at this rate. Is refinancing
an attractive option for Carol? Assume she has enough funds to pay the refinancing fee, and that she
therefore need not borrow this amount.
The annual installment due at the outset can be computed using the PMT function in Excel:

Installment = PMT(0.084, 8, −800000) = $141,332.19

The principal outstanding after two installments can be computed using the PV function in Excel:

Principal = PV(0.084, 6, −141332.19) = $645,508.52

If this is reamortized over six years using the new rate of 7.20% per annum, the new annual payment
is

Installment = PMT(0.072, 6, −645508.52) = $136,262.24

Thus she will save 141,332.19 − 136,262.24 = $5,069.95 The present value of the savings can be com-
puted using the PV function:

Present Value = PV(0.06, 6, −5069.95) = $24,930.59

206 | Chapter 8: Mortgage Loans

The cost of refinancing is = 0.0275 × 645,508.52 = $17,751.48.

Because the cost of refinancing is less than the present value of the savings, refinancing is an
attractive option. Now let’s pose a related question. What is the threshold interest rate, below which
refinancing becomes attractive? This may be determined as follows.
To be indifferent to refinancing, the present value of savings should be equal to the cost. Thus the
present value of savings ought to be $17,751.48. The periodic saving is = PMT(0.06, 6, −17751.48) =
$3,610. This implies that the corresponding annual installment on the loan should be $141,332.19 −
$3,610 = $137,722.19. The corresponding interest rate is given by

Interest Rate = Rate(6, −137722.19, 645508.52) = 7.5475%

Thus if the mortgage rate were above 7.5475%, Carol would not refinance. At lower rates, it would be
in her interest to refinance.
An alternative perspective is the following. The number of periods required to recover the refi-
nancing cost from the periodic savings is = NPER(0.06, −5069.95, 17751.48) = 4.05 years. Thus if the
borrower plans to remain in the home for at least four more periods, then refinancing is a beneficial
option.

The Role of the Mortgage Rate in Prepayments

As we have seen, if the prevailing mortgage rate is significantly lower than the rate pre-
scribed in the contract, many homeowners will refinance. The greater the difference
between the two rates is, the greater the incentive to refinance. Refinancing is ben-
eficial if the interest saving is greater than the cost of refinancing. The costs include
legal expenses, repayment of origination fees, obtaining of fresh title insurance, and
the value of the time expended in sourcing a fresh loan. The benefit from refinancing
depends on the original loan rate. Assume the contract rate is 4.50% and that the pre-
vailing rate is 3%. If we consider a principal of $250,000, and 30 years to maturity, the
monthly installment declines from $1,266.71 to $1,054.01, a change of −16.79%. How-
ever if the initial rate were 9.50%, and it declined by 150 bp, the monthly installment
would change from $2,102.14 to $1,834.41, which represents a reduction of only 12.74%.
Interest rates also impact prepayments for the following reason. If the mortgage rate is
low, an existing mortgagor may use the opportunity to buy a bigger or more expensive
home or move to an up-market locality. Such circumstances also cause borrowers to
prepay.
The path taken by mortgage rates also has an impact on prepayment behavior.
Let’s consider two possible paths for a given value of the mortgage rate. Assume that
in path-1, the rate declines from 6% per annum to 3% per annum, increases once again
to 5.50%, and then falls to 3%. In path-2, it increases from 6% to 9%, and then falls to
3%. In the case of path-1, many borrowers will refinance when the rate falls from 6%
to 3%. When the rate rebounds and falls again to 3% per annum, there are unlikely
to be significant prepayments, because most borrowers seeking to benefit from lower
rates would have already refinanced. In the case of path-2, however, there are likely to
Negative Amortization | 207

be significant prepayments when the rate declines from 9% to 3%. This phenomenon
is termed as prepayment burnout.

Negative Amortization
In the case of some mortgage structures, there may be negative amortization. The term
refers to a situation where the installment for a period is inadequate to pay the required
interest for the period. Consequently the deficit gets added on to the outstanding bal-
ance, as a consequence of which the latter increases over time, unlike in the normal
case when the outstanding principal steadily decreases. This can be illustrated with
the help of an example.

Example 8.2. Consider a mortgage loan with a principal of $100,000 and a tenor of 12 months. The
rate is 6% per annum. The monthly installment would have been $8606.64 had it been a plain vanilla
mortgage. However, the terms of the loan stipulate that the monthly payment will be $250 for the first
three months. The installment will then be revised and remain constant for the remaining life of the
loan. The amortization schedule may be depicted as shown in Table 8.1.

Table 8.1: Illustration of negative amortization.

Time Installment Interest Principal Outstanding

Component Component Balance

0 100,000.00
1 250.00 500.00 −250.00 100,250.00
2 250.00 501.25 −251.25 100,501.25
3 250.00 502.51 −252.51 100,753.76
4 11,476.59 503.77 10,972.83 89,780.93
5 11,476.59 448.90 11,027.69 78,753.24
6 11,476.59 393.77 11,082.83 67,670.41
7 11,476.59 338.35 11,138.24 56,532.17
8 11,476.59 282.66 11,193.93 45,338.24
9 11,476.59 226.69 11,249.90 34,088.33
10 11,476.59 170.44 11,306.15 22,782.18
11 11,476.59 113.91 11,362.68 11,419.50
12 11,476.59 57.10 11,419.50 0.00

As can be seen, during the first three months, the installment is inadequate to cover
even the interest due. Hence the principal component is negative, and the unpaid prin-
cipal is added on to the outstanding balance, which as a consequence, increases.
A lender may impose a cap on the negative amortization. For instance, assume
that there is cap of 125% of the original loan amount. Thus if the original loan amount
is $200,000, the cap on the outstanding balance is $250,000. If because of negative
208 | Chapter 8: Mortgage Loans

amortization, the balance reaches the limit, the monthly payment resets to fully amor-
tize the loan over the remaining term to maturity. The concept of a payment cap, which
puts a limit on the change in the installment from one period to the next, does not ap-
ply in such a situation.1

Other Mortgage Structures

The level payment mortgage, which entails a constant monthly payment for the life
of the mortgage, is the simplest mortgage structure. In practice, other mortgage struc-
tures, with more complex features are possible. Let’s discuss some of them.

Adjustable Rate Mortgages (ARMs)

In an adjustable rate mortgage, the interest rate is not fixed for the life of the loan, but
is reset periodically. Thus the monthly payments rise if the interest rate at the time of
resetting is higher than what it was previously and fall if the interest rate declines. The
rate on such mortgages is linked to a benchmark such as LIBOR or the rate for Treasury
securities. Example 8.3 illustrates such a mortgage loan, assuming that rates are reset
at the end of every year.

Example 8.3. Consider a four-year loan with an initial principal of $800,000. Payments are made an-
nually, and the interest rate is reset every year. Assume that the benchmark is LIBOR, as is the mort-
gage rate. Assume that the values of LIBOR observed over the next three years are as shown in Ta-
ble 8.2.

Table 8.2: Observed LIBOR rates.

Time LIBOR

0 4.80
1 5.40
2 5.10
3 4.50

The installment for the first year is

= PMT(0.048, 4, −800000) = $224,562.25

1 We study payment caps in detail later in this chapter.

Other Mortgage Structures | 209

The outstanding balance at the end of the first year is

= PV(0.048, 3, −224562.25) = $613,837.75

The installment for the second year is

= PMT(0.054, 3, −613837.75) = $227,097.96

The outstanding balance at the end of the second year is

= PV(0.054, 2, −227097.96) = $419,887.02

The installment for the third year is

= PMT(0.051, 2, −419887.02) = $226,137.31

The outstanding balance at the end of the third year is

= PV(0.051, 1, −226137.21) = $215,163.95

The installment for the fourth year is

= PMT(0.045, 1, −215163.95) = $224,846.33

The outstanding balance at the end of the fourth year is obviously zero. The PMT for the final year is
the principal outstanding at the end of the third year, plus interest for the fourth year:

215,163.95 × 1.045 = $224,846.33

Table 8.3: Cash flows of an ARM.

Time Installment Interest Principal Outstanding

Component Component Balance

0 800,000
1 224,562.25 38,400.00 186,162.25 613,837.75
2 227,097.96 33,147.24 193,950.72 419,887.02
3 226,137.31 21,414.24 204,723.07 215,163.95
4 224,846.33 9,682.38 215,163.95 0.00

In practice, the mortgage rate will be higher than the index rate, by a few percentage
points, known as the margin. For instance, a lender may set the rate as LIBOR + 45
bp. In this case the margin is 45 basis points. Thus if LIBOR is at 3% currently, the
rate is 3.45%. This is termed the fully indexed rate. The margin can vary depending on
the perceived credit quality of the borrower. Thus more credit-worthy borrowers often
pay lower rates in practice. Sometimes the initial loan rate can be set lower than what
the index level warrants. For instance, if the rate is LIBOR + 45 bp, and the LIBOR is
at 2.80%, the loan rate must be 3.25%. However, the lender may set the rate at, say,
2.60%. Such a discounted rate is termed a teaser rate, which as the name suggests,
is an attempt to induce a potential borrower to take a loan. The teaser rate can be in
effect until the first adjustment date, or even longer.
210 | Chapter 8: Mortgage Loans

Option to Change the Maturity

In the case of certain adjustable rate mortgages, the borrower may be given the option
to keep the monthly installment at the initial level and have the maturity date altered
every time the rate is reset. Let’s see the implications of this for the mortgage discussed
in Example 8.3. In the example, the installment for the first year was $224,562.25. As-
sume the rate rises from 4.80% per annum to 8.40% per annum. The payment for the
second year becomes

= PMT(0.084, 3, −613837.75) = $239,910.70

However if we are given an option, we can continue to pay $224,562.25 per annum,
and extend the maturity of the loan. The modified time period can be computed using
the NPER function in Excel:

= NPER(0.084, −224562.25, 613837.75) = 3.2342 years

Features of ARMs
Most lenders charge a lower interest rate for an ARM at the outset, as compared to a
fixed-rate mortgage. Thus the initial financial burden on the new homeowner is lower.
If rates do not change much, or decline sharply, the borrower stands to benefit as com-
pared to a party who has availed of a fixed rate mortgage. However, the flip side is that
the interest rates may rise, and the borrower then has to pay higher periodic install-
ments.
The interest rate and consequently the monthly payment of an ARM resets peri-
odically. The rate may change every month, quarter, year, or even longer periods. The
time period between rate changes is called the adjustment period. For instance, an
ARM may have an adjustment period of 12 months. That is, the rate is reset once per
annum.

Variations on the ARM Structure

Let’s now discuss some alternatives to a simple ARM structure.

Hybrid ARMs: An N/1 ARM is a hybrid of a fixed rate and a variable rate mortgage loan.
The rate remains fixed for the first N years, after which the loan assumes the feature
of an adjustable rate mortgage. For instance, a 5/1 ARM means that the rate is fixed
for the first five years. Thereafter the rate adjusts annually until the loan is paid off. A
30 mortgage loan may also be described as an n/30-n ARM, for instance a 4/26 ARM.
This means that the rate remains fixed for the first four years, after which it becomes
Features of ARMs | 211

adjustable. Once the fixed rate period expires, the rate may be adjusted annually or at
times even more frequently.

Interest-only or I-O ARMs: In the case of such an ARM, the borrower pays only in-
terest for the first few years. This reduces the cash flow burden on the borrower. At
the end of the initial period, the monthly payments increase, even if interest rates do
not change, because the entire principal has to be repaid over a truncated period. The
interest rate may or may not adjust during the initial (I-O) period. The longer the du-
ration of the I-O period, the higher the periodic payment is after it ends, because the
loan has to be repaid during a shorter period. Example 8.4 provides an illustration.

Example 8.4. Consider a 10-year mortgage loan for $800,000. For the first three years, the borrower
has to pay interest at the rate of 5% per annum, on an annual basis. There is no scheduled principal
repayment during this period. The entire principal will be repaid over the remaining seven years, after
the end of the initial 3-year I-O period.
For the first three years, the payment is 800,000 × 0.05 = $40,000. If the interest rate remains
constant at 5% per annum, the annual payment after three years becomes = PMT(0.05, 7, −800000) =
$138,255.85. Had there been no I-O period, the payment after three years would have been =
PMT(0.05, 10, −800000) = $103,603.66. If the I-O period were for five years, the payment after
the initial interest-only period would be = PMT(0.05, 5, −800000) = $184,779.84.

ARMs with Payment Options: An ARM with a payment option allows the borrower to
choose his payment option every month. There typically is a choice of three options:
– A conventional installment, comprised of principal and interest. This leads to a
reducing balance at the end of every month.
– An interest-only option, which requires the payment of interest on the balance
outstanding at the start of the month. With this option, the outstanding balance
remains constant.
– A reduced installment option, wherein the borrower pays less than under a con-
ventional mortgage. If the total payment is less than the scheduled interest for the
period, there is negative amortization.

An ARM with a payment option will have a built-in recalculation period, usually every
five years. At the end of the period, the payments are recalculated or recast based on
the remaining term of the loan. For instance, assume that you have taken a loan of
$250,000 to be repaid over 15 years in monthly installments. At the end of five years,
the monthly installments are recalculated for the remaining 10 years. If there has been
negative amortization in the first five years, there could be a significant increase in
the magnitude of the installment. On each recast date, the new minimum payment
becomes a fully amortizing payment over the remaining term of the loan. In such a
situation, restrictions like payment caps, which put a limit on the increase in the pe-
riodic installment, do not apply.
212 | Chapter 8: Mortgage Loans

For instance, consider a loan of $250,000 to be repaid over 15 years in equal

monthly installments. Assume that the interest rate is 3% per annum. The installment
amount is PMT(0.03/12, 180, −250000) = $1,726.45. If the outstanding balance after
five years is $300,000, due to negative amortization, the recast installment becomes
PMT(0.03/12, 120, −300000) = $2,896.82, assuming that the mortgage rate remains at
3% per annum. This represents an increase of 68%. It is important to understand that
lenders do not allow outstanding balances to increase without limits. In practice, if
the outstanding balance reaches a preset limit of say 125% of the original outstanding,
the loan is automatically recast, irrespective of whether the scheduled recast period
has been reached.

Interest Rate Caps

ARMs may come with interest rate caps. A cap places a limit on the amount that the
interest can increase from one coupon reset date to another. There are two possibilities
– we can have a periodic adjustment cap or a lifetime cap. A periodic adjustment cap
places a limit on the amount that the interest rate can adjust up or down from one
adjustment period to another, after the first adjustment. A lifetime cap puts a limit on
the maximum interest rate over the life of the loan. In some countries, the law may
mandate a lifetime cap.

Example of Interest Rate Caps

Consider a four-year mortgage loan with a principal of $250,000. The rate is equal to
LIBOR + 1%. There is a periodic adjustment cap of 1% and a lifetime cap of 5%. The
initial mortgage rate is 3%. Assume that the values of LIBOR are as shown in Table 8.4.

Table 8.4: Mortgage rates with and without rate caps.

Time LIBOR Rate Rate

without Cap with Cap

0 2.00 3.00 3.00

1 2.50 3.50 3.50
2 3.75 4.75 4.50
3 4.50 5.50 5.00

From period-0 to period-1, the rate increases by 0.50%. This is within the periodic limit
of 1%. Also 3.50% is below the maximum permissible limit of 5%. In the absence of a
periodic cap, the rate in the third year would have been 4.75%. However, because there
Graduated Payment Mortgages | 213

is a periodic cap of 1%, it can be only 4.50%, or in other words, the rate for the previous
year plus 1%. In the final year, the rate should be 5.50%. This is within the periodic
cap of 1%, as compared to the rate for the previous year, which is 4.50%. However,
because there is a lifetime cap of 5%, the rate for the last year is 5%.

Carryovers
In certain cases, if the lender cannot increase the rate because of the presence of a
rate cap, it can carry over the increase that was not imposed to future periods. This
is termed as a carryover. For instance, assume that the initial rate is 5%. The LIBOR
at the end of the first year is 6.25%. However there is a periodic cap of 0.75%, as a
consequence of which the rate for the second year is capped at 5.75%. If the LIBOR
for the third year remains unchanged at 5.75%, the rate under normal circumstances
ought to remain at 5.75%. However, because there is a carryover of 0.50% from the
previous year, which is the uncapped rate of 6.25% less the capped rate of 5.75%, the
rate for the third year is adjusted to 6.25%.

Payment Caps

A payment cap imposes an upper limit on how much the installment can change from
one year to the next. In practice, unlike rate caps, these can lead to negative amorti-
zation. Here is an example.
Consider a loan of $250,000 to be repaid in five equal annual installments. Assume
that the initial LIBOR is 4%. The payment for the first year is = PMT(0.04, 5, −250000) =
$56,156.78. Assume that at the end of the first year, the LIBOR increases to 7.50%.
However there is a payment cap of 5%. The outstanding balance at the end of the
first year is = PV(0.04, 4, −56156.78) = $203,843.22. The uncapped installment for the
second year is = PMT(0.075, 4, −203843.22) = $60,860.96. However, the installment
is capped at 56,156.78 × 1.05 = $58,964.62. The interest due for the second year is
= 203,843.22 × 0.075 = $15,288.24. Because the capped installment is higher, there is
no negative amortization in this illustration.

Graduated Payment Mortgages

Young home buyers may not have the disposable income required to take advantage
of a conventional fixed rate mortgage. However, many people in this age bracket have
the potential to earn substantially more in the coming years. To encourage such par-
ties to take loans, the graduated payment mortgage was designed. Such a mortgage
starts with a level of payment that steadily increases every year up to a point and then
remains steady thereafter. We can illustrate it with the help of an example.
214 | Chapter 8: Mortgage Loans

Example 8.5. Maureen Toohey is being offered a mortgage loan by Tennessee National Bank with the
following terms. She has to repay the loan in eight annual installments. The annual payment increases
by 5% for the first three years and remains constant thereafter. The mortgage rate is 7.20% per annum
for a loan amount of $800,000. What does the payment schedule look like?
Let’s denote the annual payment for the first year by A. The first four payments are A; 1.05A;
(1.05)2 A; and (1.05)3 A. The next four payments are (1.05)3 A, which constitute a four-year annuity. The
loan amount is the present value of these eight payments.
1
A 1.05A 1.1025A 1.157625A 1.157625A [1 − (1.072)4 ]
800,000 = + + + +
1.072 (1.072) 2
(1.072)3 (1.072)4 0.072 (1.072)4
⇒ 800,000 = 6.5738A
⇒ A = $121,695.21

Thus the payments over the eight-year period are as depicted in Table 8.5.

Table 8.5: Installments for the graduated payment mortgage.

Year Cash Flow

1 121,695.21
2 127,779.98
3 134,168.97
4 140,877.42
5 140,877.42
6 140,877.42
7 140,877.42
8 140,877.42

Growing Equity Mortgages

There are certain similarities between a graduated payment mortgage, and a grow-
ing equity mortgage, although there are key differences. In both cases the monthly
payments are scheduled to increase over time. However, in a graduated payment mort-
gage, there could be negative amortization for an initial period. In contrast, the first
period’s payment for a growing equity mortgage is the same as that of a traditional
fixed rate mortgage. Increases in the periodic payments constitute prepayments,
which directly amortize the principal. Thus the term of the mortgage and the total
interest payable decreases.

A Comparison of the Three Mortgage Structures

Let’s consider the following data to make a comparison between the three mortgage
structures. The mortgage rate is 7.20% per annum for a loan amount of $800,000, to
A Comparison of the Three Mortgage Structures | 215

be repaid over a period of eight years. Let’s assume that the annual payment increases
by 5% every year for the graduated payment, as well as the growing equity mortgages.

Table 8.6: Cash flows for a graduated payment mortgage vs. that for a growing equity mortgage.

Yr. Int. for FRM Pri. for FRM Int. for GPM Pri. for GPM Int. for GEM Pri. for GEM

1 57,600.00 77,414.42 57,600.00 57,539.39 57,600.00 77,414.42

2 52,026.16 82,988.26 53,457.16 67,439.19 52,026.16 89,738.98
3 46,051.01 88,963.42 48,601.54 78,339.63 45,564.95 103,288.45
4 39,645.64 95,368.78 42,961.09 90,327.15 38,128.19 118,167.89
5 32,779.09 102,235.33 36,457.53 103,495.11 29,620.10 134,490.78
6 25,418.14 109,596.28 29,005.89 117,944.39 19,936.76 152,379.66
7 17,527.21 117,487.21 20,513.89 133,783.90 8,965.43 124,519.83
8 9,068.13 125,946.29 10,881.45 151,131.23 0.00 0.00
Total 280,115.39 299,478.60 251,841.59
Yr. stands for year.
Int. stands for the interest payment.
Pri. stands for principal payment.
FRM stands for fixed rate mortgage.
GPM stands for graduated payment mortgage.
GEM stands for growing equity mortgage.

Table 8.7: Outstanding balances for the three mortgage structures.

Time Outstanding Outstanding Outstanding

Balance for FRM Balance for GPM Balance for GEM
0 800,000.00 800,000.00 800,000.00
1 722,585.58 742,460.61 722,585.58
2 639,597.31 675,021.42 632,846.59
3 550,633.90 596,681.79 529,558.15
4 455,265.12 506,354.64 411,390.26
5 353,029.78 402,859.53 276,899.49
6 243,433.50 284,915.14 124,519.83
7 125,946.29 151,131.23 0.00
8 0.00 0.00 0.00
FRM stands for fixed rate mortgage.
GPM stands for graduated payment mortgage.
GEM stands for growing equity mortgage.

Let’s do a year-by-year analysis to illustrate the concepts.

216 | Chapter 8: Mortgage Loans

Year 1
Total outstanding at the beginning of the year for all three mortgages = $800,000
Interest component for all three mortgages = 800,000 × 0.072 = $57,600
Total installment for FRM and GEM = $135,014.42
Total installment for GPM = $115,139.39
Principal component for FRM and GEM = 135,014.42 − 57,600 = $77,414.42
Principal component for GPM = $57,539.39
Total outstanding for FRM and GEM at the end of the year = $722,585.58
Total outstanding for GPM at the end of the year = $742,460.61

Year 2
Interest component for FRM and GEM = $52,026.16
This is because the outstanding balance at the end of the previous period was the same
for both.
Interest component for GPM = $53,457.16
Total installment for FRM = $135,014.42, which is obviously a constant for every period.
Total installment for GEM = 135,014.42 × 1.05 = $141,765.14
Total installment for GPM = 115,139,39 × 1.05 = $120,896.36
Principal component for FRM = 135,014.42 − 52,026.16 = $82,988.26
Principal component for GEM == 141,765.14 − 52,026.16 = $89,738.98
Principal component for GPM == 120,896.36 − 53,457.16 = $67,439.20
Total outstanding for FRM at the end of the year = $639,597.31
Total outstanding for GEM at the end of the year = $632,846.59
Total outstanding for GPM at the end of the year = $675,021.42

Year 3
Interest component for FRM = $46,051.01
Interest component for GEM = $45,564.95
Interest component for GPM = $48,601.54
Total installment for FRM = $135,014.42
Total installment for GEM = 141,765.14 × 1.05 = $148,853.40
Total installment for GPM = 120,896.36 × 1.05 = $126,941.18
Principal component for FRM = 135,014.42 − 46,051.01 = $88,963.42
Principal component for GEM == 148,853.40 − 45,564.95 = $103,288.45
Principal component for GPM == 126,941.18 − 48,601.54 = $78,339.63
Total outstanding for FRM at the end of the year = $550,633.90
Total outstanding for GEM at the end of the year = $529,558.15
Total outstanding for GPM at the end of the year = $596,681.79
A Comparison of the Three Mortgage Structures | 217

Year 4
Interest component for FRM = $39,645.64
Interest component for GEM = $38,128.19
Interest component for GPM = $42,961.09
Total installment for FRM = $135,014.42
Total installment for GEM = 148,853.40 × 1.05 = $156,296.07
Total installment for GPM = 126,941.18 × 1.05 = $133,288.23
Principal component for FRM = 135,014.42 − 39,645.64 = $95,368.78
Principal component for GEM == 156,296.07 − 38,128.19 = $118,167.89
Principal component for GPM == 133,228.23 − 42,961.09 = $90,327.15
Total outstanding for FRM at the end of the year = $455,265.12
Total outstanding for GEM at the end of the year = $411,390.26
Total outstanding for GPM at the end of the year = $506,354.64

Year 5
Interest component for FRM = $32,779.09
Interest component for GEM = $29,620.10
Interest component for GPM = $36,457.53
Total installment for FRM = $135,014.42
Total installment for GEM = 156,296.07 × 1.05 = $164,110.88
Total installment for GPM = 133,288.23 × 1.05 = $139,952.65
Principal component for FRM = 135,014.42 − 32,779.09 = $102,235.33
Principal component for GEM == 164,110.88 − 29,620.10 = $134,490.78
Principal component for GPM == 139,952.65 − 36,457.53 = $103,495.11
Total outstanding for FRM at the end of the year = $353,029.78
Total outstanding for GEM at the end of the year = $276,899.49
Total outstanding for GPM at the end of the year = $402,859.53

Year 6
Interest component for FRM = $25,418.14
Interest component for GEM = $19,936.76
Interest component for GPM = $29,005.89
Total installment for FRM = $135,014.42
Total installment for GEM = 164,110.88 × 1.05 = $172,316.42
Total installment for GPM = 139,952.65 × 1.05 = $146,950.28
Principal component for FRM = 135,014.42 − 25,418.14 = $109,596.28
Principal component for GEM == 172,316.42 − 19,936.76 = $152,379.66
Principal component for GPM == 146,950.28 − 29,005.89 = $117,944.39
218 | Chapter 8: Mortgage Loans

Total outstanding for FRM at the end of the year = $243,433.50

Total outstanding for GEM at the end of the year = $124,519.83
Total outstanding for GPM at the end of the year = $284,915.14

Year 7
Interest component for FRM = $17,527.21
Interest component for GEM = $8,965.43
Interest component for GPM = $20,513.89
Total installment for FRM = $135,014.42
Total installment for GEM = $133,485.26
This represents the outstanding of $124,519.83 plus $8,965.43 by way of interest.
Total installment for GPM = 146,950.28 × 1.05 = $154,297.79
Principal component for FRM = 135,014.42 − 17,527.21 = $117,487.21
Principal component for GEM == 133,485.26 − 8,965.43 = $124,519.83
Principal component for GPM == 154,297.79 − 20,513.89 = $133,783.90
Total outstanding for FRM at the end of the year = $125,946.29
Total outstanding for GEM at the end of the year = 0.00
Total outstanding for GPM at the end of the year = $151,131.23

Year 8
Interest component for FRM = $9,068.13
Interest component for GPM = $10,881.45
Total installment for FRM = $135,014.42
Total installment for GPM = 154,297.79 × 1.05 = $162,012.68
Principal component for FRM = 135,014.42 − 9,068.13 = $125,946.29
Principal component for GPM == 162,012.68 − 10,881.45 = $151,131.23
Total outstanding for FRM at the end of the year = 0.00
Total outstanding for GPM at the end of the year = 0.00

Mortgage Servicing
A mortgage loan has to be serviced. Servicing a mortgage loan includes the following
activities:
– Collection of monthly payments and forwarding the proceeds to the current owner
of the loan. Remember, mortgage loans can be sold in the secondary market, and
consequently the current owner need not be the original owner.
– Sending payment notices to mortgagors. Mortgage loans are typically paid in
monthly installments. Consequently a payment notice has to be sent at monthly
intervals.
Mortgage Servicing | 219

– Reminding mortgagors whose payments are overdue. If the monthly installment

is not paid by the due date, an overdue notice has to be sent.
– Maintaining records of principal balances. Every installment consists of an inter-
est component and a principal component. Consequently, at the time of receipt of
each monthly payment, the outstanding principal has to be computed.
– Administering escrow balances for real estate taxes and insurance purposes. An
escrow account is a trust account held in the name of the homeowner to pay statu-
tory levies like property taxes and insurance premiums. The maintenance of such
an account helps ensure that the payments are made when due, for it becomes
the lender’s responsibility to do so. Usually, the borrower makes deposits on a
monthly basis, along with the loan payment, and the payments accrue at the
lender. The amount required to be deposited every month is a function of the cost
of insurance and the tax assessment of the property concerned. Consequently, it
fluctuates from year to year, because both the insurance premium and the prop-
erty tax vary over time. The servicer earns interest on such payments.
– Initiating foreclosure proceedings if necessary. If the borrower defaults on the pay-
ment obligation, the lender can seize the property in order to recover what is due.
This is termed as the right of foreclosure.
– Furnishing tax information to mortgagors when applicable. In most countries the
interest payments on the mortgage loan during the course of a financial year, as
well as a percentage of the principal that is repaid during the year, are eligible
for an income tax benefit. Consequently it is a common practice to provide the
borrower with an annual statement that includes the total interest payment made
during the year and the total principal repaid in that year.

Income for the Servicer

The primary source of income is the servicing fee, which is a fixed percentage of
the outstanding mortgage balance. Because the mortgage loan is an amortized loan,
where the outstanding principal declines with each payment, the revenue from ser-
vicing, declines over time. The typical servicing fee could be in the range of 25 to
50 basis points per annum. The servicer also earns interest on the escrow balances.
Servicing rights can be traded in the market.
Because the servicer gets a servicing fee, the lender in a mortgage loan gets only
a reduced percentage of the interest payment on the mortgage loan. This is termed net
interest.

Example 8.6. Consider a mortgage loan of $100,000 to be repaid in four equal annual installments.
The interest rate is 6% per annum, and the servicing fee is 0.50% per annum. The amortization sched-
ule is depicted in Table 8.8.
220 | Chapter 8: Mortgage Loans

Table 8.8: Amortization schedule with a servicing fee.

Time Inst. Int. Ser. Net. Prin. Outst.

0 100,000.00
1 28,859.15 6,000.00 500.00 5,500.00 22,859.15 77,140.85
2 28,859.15 4,628.45 385.70 4,242.75 24,230.70 52,910.15
3 28,859.15 3,174.61 264.55 2,910.06 25,684.54 27,225.61
4 28,859.15 1,633.54 136.13 1,497.41 27,225.61 0.00
Inst. stands for installment.
Int. stands for interest component.
Ser. stands for servicing fee.
Net. stands for net interest.
Prin. stands for principal component.
Outst. stands for outstanding balance.

Computing the Servicing Fee and the Net Interest Using the IPMT Function
The scheduled outstanding at the beginning of “period t” is

A 1
[1 − ]
r (1 + r)N−t
The interest component is

1
A × [1 − ]
(1 + r)N−t
Let’s denote the servicing fee percentage by s and the net interest percentage by i.
Therefore, r = s + i. If we multiply the outstanding balance at the beginning of the
period by s, we get the servicing fee as As
r
[1 − 1 Ai
N−t ] and the net interest as r [1 −
(1+r)
1
]. The sum of the two is equal to
(1+r)N−t

A(s + i) 1 1
[1 − ] = A × [1 − ]
r (1 + r)N−t
(1 + r)N−t
Thus the gross interest for a period is equal to the sum of the servicing fee and the net
interest.
Now let’s consider the data in Example 8.6. If we compute the servicing fee for the
first period as = IPMT(0.005, 1, 4 − 0, −100000), we get $500.00, and similarly if we
compute the net interest as = IPMT(0.055, 1, 4 − 0, −100000), we get $5,500. The total
of the two is the gross interest of $6,000. Similarly, for the second month the servicing
fee = IPMT(0.005, 1, 4−1, −77140.85) = $385.70, and the net interest = PMT(0.055, 1, 4−
1, −77140.85) = $4,242.75. The sum of the two is equal to the gross interest of $4,628.45.
Consequently if we use the IPMT function to compute the servicing fee and net
interest, we have to specify PER as 1, NPER as the remaining number of periods, and
the PV as the outstanding balance at the end of the previous period.
Mortgage Insurance | 221

Mortgage Insurance
There are two types of mortgage-related insurance. The first type, required by the
lender to insure against default by the borrower, is called mortgage insurance or pri-
vate mortgage insurance. It is usually required by lenders on loans with a loan-to-value
(LTV) ratio that is greater than a specified limit. The amount insured is some percent-
age of the loan and may decline as the LTV declines. The loan to value ratio can be
understood as follows. Every borrower has to make a down payment, which is the dif-
ference between the price of the property and the loan amount. That is, the percentage
of the property value that is funded with borrowed money is less than 100%. The LTV
ratio is obtained by dividing the loan amount by the market value of the property. The
lower the LTV ratio, the greater is the protection for the lender in the event of default
by the borrower.
The second type of mortgage-related insurance is acquired by the borrower, usu-
ally through a life insurance company, and is typically called credit life. This is not
required by the lender. Such policies provide for the continuation of mortgage pay-
ments after the death of the insured person so that survivors can continue to live in
the house.

Sale of Mortgage Loans

There are agencies that buy and pool mortgage loans and issue securities that are
backed by the underlying pool. This phenomenon is termed as securitization and is
examined in detail in the next chapter. Such agencies buy what they call conforming
mortgages. A conforming mortgage is one that meets the underwriting criteria estab-
lished by the agency, from the standpoint of being eligible to be included in a pool for
the purpose of being securitized. A conforming mortgage must satisfy three criteria.
– A maximum payment to income (PTI) ratio. The PTI ratio is the ratio of monthly
payments (the loan payment plus any real estate tax payments) to the borrower’s
monthly income. Obviously, the lower the ratio, the greater is the likelihood of
the borrower being able to pay as scheduled. For the purpose of computing the
PTI ratio, the normal practice is to take the income as the borrower’s take-home
salary. This gives a more conservative estimate, for the cost of the employee to the
company is typically much higher.
– A maximum loan-to-value or LTV ratio.
– A maximum loan amount. Mortgages that are non-conforming because they are
for amounts in excess of the purchasing limit set by these agencies are termed
jumbo mortgages.

Certain mortgages may not meet the underwriting guidelines from the standpoint of
credit quality or loan-to-value ratio. These are termed as subprime mortgages and have
now become world famous for the wrong reasons.
222 | Chapter 8: Mortgage Loans

The Average Life of a Mortgage Loan

Unlike a plain vanilla bond, which returns the entire principal as a bullet payment
at the end, a mortgage loan pays back a portion of the principal as a part of every
monthly installment. Thus the average life of a mortgage loan is less than its stated
life. The average life is defined as

N
∑(t × PCF t )/L
t=1

In this expression, PCF t is the principal component of the cash flows at time t, and L is
the loan amount. In the case of a mortgage loan that is repaid strictly as per schedule,
that is, without any prepayments, the PCF is the principal component of the install-
ment, or what is termed as the scheduled principal. However, if in a period there is a
prepayment, the PCF is the sum of the scheduled principal and the prepayment. Quite
obviously, the larger the prepayments, the shorter the average life becomes.
Let’s illustrate the computation of average life with the help of an example.

Example 8.7. Consider a 15-month mortgage loan for an amount of $250,000. The interest rate is
7.20% per annum, and payments are made in 15 equal monthly installments. The payment schedule
is presented in Table 8.9.

Table 8.9: Average life of a mortgage loan in the absence of pre-payments.

Time Installment Interest Principal Outstanding

Component Component Balance

0 250,000.00
1 17,477.83 1,500.00 15,977.83 234,022.17
2 17,477.83 1,404.13 16,073.70 217,948.47
3 17,477.83 1,307.69 16,170.14 201,778.33
4 17,477.83 1,210.67 16,267.16 185,511.17
5 17,477.83 1,113.07 16,364.76 169,146.40
6 17,477.83 1,014.88 16,462.95 152,683.45
7 17,477.83 916.10 16,561.73 136,121.72
8 17,477.83 816.73 16,661.10 119,460.62
9 17,477.83 716.76 16,761.07 102,699.55
10 17,477.83 616.20 16,861.63 85,837.91
11 17,477.83 515.03 16,962.80 68,875.11
12 17,477.83 413.25 17,064.58 51,810.53
13 17,477.83 310.86 17,166.97 34,643.56
14 17,477.83 207.86 17,269.97 17,373.59
15 17,477.83 104.24 17,373.59 0.00
Prepayments of Principal | 223

Assume the data for column 1 of the table is in cells A2 to A17 of the Excel sheet, and the data for the
principal component of an installment is in cells D3 to D17. The average life can be computed using the
SUMPRODUCT function in Excel.
Average life = SUMPRODUCT(A3:A17,D3:D17) = 8.11.
Thus this loan has an average life of 8.11 months.
The various functions need to be invoked as follows:
– Payment = PMT(.072/12, 15, −250000)
– Interest = IPMT(.072/12, A3, 15, −250000)
– Principal = PPMT(.072/12, A3, 15, −250000)
PER will have a value of 1 when we compute IPMT and PPMT for the first month, 2 when we compute
for the second month, and so on.

Prepayments of Principal
In practice, the cash flows from a mortgage loan are unpredictable, because most
homeowners prepay at one or more stages during the life of the loan. Prepayments
reduce the average life, and the higher the prepayment speed is, the shorter the aver-
age life. Thus, without making an assumption about prepayments, we cannot project
the cash flows from a mortgage loan.

Single Month Mortality (SMM)

The single month mortality can be defined as:

(SP t − Lt )
SMMt =
SP t

SPt is the scheduled outstanding principal at the end of month t. This is what the
outstanding principal will be if there is no prepayment at the end of the month. Lt is
the actual outstanding at the end of the month. The actual outstanding is less than
the scheduled outstanding if there is a prepayment. The difference between the two
constitutes the prepayment. We can rewrite the equation as

Lt = SP t (1 − SMMt )

Example 8.8. Consider a 15-month mortgage loan for an amount of $250,000. The interest rate is
7.20% per annum, and payments are made in 15 monthly installments. Assume that the SMM is 2.50%
for all months. The payment schedule is presented in Table 8.10.
224 | Chapter 8: Mortgage Loans

Table 8.10: Average life of a mortgage loan with prepayments.

Time Inst. Int. Sch. Pr. Sch. Out. Prep. Act. Out.

0 250,000 250,000
1 17,477.83 1,500.00 15,977.83 234,022.17 5,850.55 228,171.61
2 17,040.89 1,369.03 15,671.86 212,499.76 5,312.49 207,187.26
3 16,614.86 1,243.12 15,371.74 191,815.52 4,795.39 187,020.14
4 16,199.49 1,122.12 15,077.37 171,942.76 4,298.57 167,644.20
5 15,794.50 1,005.87 14,788.64 152,855.56 3,821.39 149,034.17
6 15,399.64 894.20 14,505.44 134,528.73 3,363.22 131,165.51
7 15,014.65 786.99 14,227.66 116,937.85 2,923.45 114,014.41
8 14,639.28 684.09 13,955.20 100,059.21 2,501.48 97,557.73
9 14,273.30 585.35 13,687.96 83,869.77 2,096.74 81,773.03
10 13,916.47 490.64 13,425.83 68,347.20 1,708.68 66,638.52
11 13,568.56 399.83 13,168.73 53,469.79 1,336.74 52,133.04
12 13,229.34 312.80 12,916.55 39,216.50 980.41 38,236.08
13 12,898.61 229.42 12,669.19 25,566.89 639.17 24,927.72
14 12,576.15 149.57 12,426.58 12,501.14 312.53 12,188.61
15 12,261.74 73.13 12,188.61 0.00 0.00 0.00
Inst. stands for installment.
Int. stands for interest component.
Sch. Pr. stands for scheduled principal.
Sch. Out. stands for scheduled outstanding.
Prep. stands for prepayment.
Act. Out. stands for actual outstanding.

Analysis of a Loan with Prepayments

Let’s analyze the cash flows for the first two periods in Table 8.10. At the outset, the out-
standing principal was $250,000. The first installment is= PMT(0.072/12, 15, −250000),
which is $17,477.83. The interest is = IPMT(0.072/12, 1, 15, −250000) = $1,500. The
scheduled principal is = PPMT(0.072/12, 1, 15, −250000) = $15,977.83. The sched-
uled outstanding is the previous actual outstanding minus the scheduled princi-
pal payment. In this case, it is $250,000 − $15,977.83 = $234,022.17. The prepay-
ment for the month is the scheduled outstanding times the SMM. In our exam-
ple it is 234,022.17 × 0.025 = $5,850.55. The actual outstanding at the end of the
first month is the scheduled outstanding minus the prepayment. In our case it is
$234,022.17 − $5,850.55 = $228,171.61.
Now let’s consider the second period. The actual outstanding at the end of the first
period is $228,171.61. The installment for the period is = PMT(0.072/12, 14, −228171.61),
which is $17,040.89. The interest is = IPMT(0.072/12, 1, 14, −228171.61) = $1,369.03. The
scheduled principal is = PPMT(0.072/12, 1, 14, −228171.61) = $15,671.86. The scheduled
outstanding is the previous actual outstanding minus the scheduled principal pay-
Relationship between Cash Flows with and without Prepayments | 225

ment. In this case it is $228,171.61 − $15,671.86 = $212,499.76. The prepayment for the
month is the scheduled outstanding times the SMM. In our example it is $212,499.76 ×
0.025 = $5,312.49. The actual outstanding at the end of the second period is the sched-
uled outstanding minus the prepayment. In our case it is $212,499.76 − $5,312.49 =
$207,187.26.
It can be seen that the three functions, PMT, IPMT, and PPMT, are invoked us-
ing the actual outstanding balance at the start of the period as the parameter PV. The
scheduled outstanding balance is computed purely to facilitate the calculation of the
pre-payment amount.
To compute the average life, we first need to compute the total principal repaid in
a month. This is the sum of the scheduled principal for the month and the pre-payment
for the month. Table 8.11 has the data.

Table 8.11: Total monthly principal with prepayments.

Time Total
Principal

1 21,828.39
2 20,984.35
3 20,167.13
4 19,375.94
5 18,610.03
6 17,868.66
7 17,151.10
8 16,456.68
9 15,784.70
10 15,134.51
11 14,505.47
12 13,896.96
13 13,308.37
14 12,739.11
15 12,188.61

Assume that the time period is in columns A2 to A16, and the total principal is in B2 to B16. Average
life = SUMPRODUCT(A2:A16,B2:B16) = 7.23. In this illustration, the average life has come down from
8.11 months to 7.23 months.

Relationship between Cash Flows with and without Prepayments

Let’s define the following variables:
– A is the constant periodic installment in the absence of prepayments.
– At is the installment payment in “period t” in the presence of prepayments.
226 | Chapter 8: Mortgage Loans

– I ∗ t is the interest payment in month t in the absence of prepayments.

– It is the interest payment in month t in the presence of prepayments.
– P ∗ t is the principal payment in month t in the absence of prepayments.
– Pt is the scheduled principal payment in month t in the presence of prepayments.
– SP ∗ t is the outstanding balance at the end of month t in the absence of prepay-
ments.
– SP t is the outstanding balance at the end of month t in the presence of prepay-
ments.
– Lt is the actual outstanding at the end of month t in the event of prepayments.

The relationship between these variables are as follows. Additional details are given
in Appendix 8.1.
t−1
At = A ∏[1 − SMMi ]
i=1

= A × [1 − SMM]t−1 if SMMi = SMM ∀i

t−1
It = It ∗ ∏[1 − SMMi ]
i=1
t−1
Pt = Pt ∗ ∏[1 − SMMi ]
i=1
t−1
SP t = SP t ∗ ∏[1 − SMMi ]
i=1
t
Lt = SP t × [1 − SMMt ] = SP t ∗ ∏[1 − SMMi ]
i=1

Example 8.9. Consider the data in Table 8.9 and Table 8.10. The former depicts the amortization
schedule in the absence of prepayments, while the latter presents the cash flows in the presence of
prepayments. The SMM is 2.50% for all months.
Let’s consider the data for the sixth month to illustrate our arguments:
A = $17,477.83; A6 = $15,399.64

A6 = 17,477.83 × (1 − 0.025)5 = $15,399.64

I∗ 6 = $1,014.88; I6 = $894.20

I6 = 1,014.88 × (1 − 0.025)5 = $894.20

P ∗ 6 = $16,462.95; P6 = $14,505.44

P6 = 16,462.95 × (1 − 0.025)5 = $14,505.44

SP ∗ 6 = $152,683.45; SP 6 = $134,528.73

SP 6 = 152,683.45 × (1 − 0.025)5 = $134,528.73

L6 = 152,683.45 × (1 − 0.025)6 = $131,165.50
Conditional Prepayment Rate (CPR) | 227

Conditional Prepayment Rate (CPR)

Whereas the SMM is a monthly rate, the prepayment rate is often expressed in annual
terms, as what is termed the conditional prepayment rate or the CPR. The mathematical
relationship between the two is given by

CPR = 1 − [1 − SMM]12

Example 8.10. The projected CPR for a mortgage loan is 7.20%. What is the corresponding SMM?

1
SMM = 1 − (1 − 0.072) 12 = 0.006208 ≡ 0.6208%

When we assume a CPR of x% per annum, it amounts to the assumption of a con-

stant prepayment rate for the life of the mortgage. However, this is contrary to what is
observed in reality. Borrowers usually do not sell their homes or refinance their mort-
gage loan when the loan is fairly new. With the passage of time, the propensity to sell
or refinance increases, and so does the CPR, in a fairly linear fashion. After a certain
period, the prepayment speeds flatten out and remain fairly constant. This is termed
as seasoning.
The Public Securities Association, which is known today as SIFMA, created a PSA
model for mortgage loans. The PSA divided the life of a mortgage loan into two distinct
periods. The initial period of 30 months is referred to as the ramp. During this period,
the CPR grows linearly at the rate of 0.2% per month, starting from a value of zero,
until it attains a value of 6% at the end of the 30th month. After the 30th month, the
mortgage is considered to be off the ramp, and the CPR is assumed to remain constant
at 6%. This can be summarized as follows:
Let t be the number of months after the mortgage was originated. If t ≤ 30, CPR =
t
6 × 30 %. If t > 30, CPR = 6%. This is termed as 100 PSA.
1
So for the 12th month, the CPR = 2.40%. The SMM is 1 − (1 − 0.024) 12 = 0.002022 ≡
0.2022%. For months 31-360, the CPR is 6%, and the SMM is 0.5143%.
Faster and slower speeds can be assumed. For a particular month, 75 PSA, means
that the CPR is 75% of what it would have been had the prepayment speed been 100
PSA. Similarly, 125 PSA means that the CPR is 125% of what it would have been under
the assumption of 100 PSA.
So if we assume 75 PSA, the CPR in the 12th month is 1.80%, and for months 31-360,
it is 4.50%. Similarly, if we assume 125 PSA, the CPR in the 12th month is 3%, and that
in months 31-360 is 7.50%.
Note that if we assume a speed of X PSA, it means that the CPR for a month is X
times what it would have been under 100 PSA. It does not mean that the SMM would
be X times the SMM under 100 PSA.
228 | Chapter 8: Mortgage Loans

An Equal Principal Repayment Loan

Sometimes a loan is structured in such a way that the principal is repaid in equal in-
stallments. This is an alternative to an amortized loan. Thus, the principal component
of each installment remains constant. However, like in the case of the amortized loan,
the interest component of each payment declines steadily, on account of the dimin-
ishing loan balance. Therefore, the total magnitude of each payment also declines.
Let’s illustrate the payment stream for a four year loan of $100,000, assuming that
the interest rate is 6% per annum.

Table 8.12: Payment schedule for an equal principal repayment loan.

Time Interest Principal Total Outstanding

Payment Balance

0 100,000
1 6,000 25,000 31,000 75,000
2 4,500 25,000 29,500 50,000
3 3,000 25,000 28,000 25,000
4 1,500 25,000 26,500 0.00

If we denote the number of installments by N, the periodic principal amount is L/N.

The interest for “period t” is

(t − 1)L L(N − t + 1)
r × [L − ]=r×
N N
LT
The outstanding principal at the end of “period t” is L − N
.

Weighted Average Coupon (WAC) and Weighted Average Maturity

(WAM)
Although mortgage loans are extremely illiquid, they can be pooled in order to issue
securities backed by the underlying pool. These are termed mortgage-backed securities
MBS and can be extremely liquid. However, all the mortgages that constitute the pool
do not have the same mortgage rate or time to maturity. Thus it is a common practice in
the market for mortgage-backed securities to account for the weighted average coupon
rate and the weighted average time to maturity. The weighted average coupon (WAC) of
a mortgage pool is determined by weighting the mortgage rate of each loan in the pool
by the principal amount outstanding. That is, the weight for each mortgage rate is the
outstanding amount of that mortgage divided by the cumulative outstanding amount
of all the mortgages in the pool. Similarly, the weighted average maturity (WAM) is
found by weighting the remaining number of months to maturity of each of the loans
Chapter Summary | 229

in the pool by the principal amount outstanding. Let’s illustrate the calculations, for
a hypothetical pool.
Assume that four mortgages are being pooled. Mortgage-A has a life of 350 months
and a coupon of 6.40%; mortgage-B has a life of 320 months and a coupon of 4.80%;
mortgage-C has a life of 280 months and a coupon of 5.40%; and mortgage-D has a life
of 300 months and a coupon of 5.70%. The current outstanding for loan-A is 100,000;
loan-B is 200,000; loan-C is 300,000; and loan-D is 400,000. Thus the weights are
10%, 20%, 30%, and 40%.

WAM = 350 × 0.10 + 320 × 0.20 + 280 × 0.30 + 300 × 0.40 = 303 months
WAC = 6.40 × 0.10 + 4.80 × 0.20 + 5.40 × 0.30 + 5.70 × 0.40 = 5.50%

Chapter Summary
This chapter addressed the issue of mortgage loans. We started with a discussion on
the risks inherent in such loans and the concept of loan refinancing. We studied alter-
natives to plain vanilla mortgages, such as adjustable rate mortgages (ARMs), grad-
uated payment mortgages (GPM), and growing equity mortgages (GEM). In the con-
text of ARMs, we introduced the concepts of rate and payment caps. The concept of
the average life of a mortgage loan was explained. We then brought in the issue of
prepayments, and introduced the concepts of single month mortality (SMM) and the
conditional prepayment rate (CPR). The chapter concluded with a brief introduction
to equal principal repayment loans.

Appendix 8.1
The single month mortality rate for a given month t, SMMt is defined as:

SP t − Lt
SMMt =
SP t

where SP t is the scheduled principal outstanding at the end of month t, and Lt is the
actual principal outstanding at the end of month t.
Thus the actual principal at the end of month t is given by

Lt = SP t [1 − SMMt ]

Consider a mortgage with N months to maturity, and a monthly interest rate of r. The
original loan balance is given by

A1 1
L0 = × [1 − ]
r (1 + r)N
230 | Chapter 8: Mortgage Loans

where A1 is the scheduled monthly installment for month 1. The scheduled principal
at the end of the first month is given by

SP 1 = L0 − [A1 − r × L0 ]
= L0 × (1 + r) − A1
A1 1
= [1 − ]
r (1 + r)(N−1)

The actual balance at time 1 is given by

L1 = SP 1 [1 − SMM1 ]
A2 1
L1 = × [1 − ]
r (1 + r)(N−1)

where A2 is the scheduled monthly installment for month 2.

A2 1
⇒ SP 1 [1 − SMM1 ] = × [1 − ]
r (1 + r)(N−1)
A 1 A 1
⇒ 1 [1 − ] [1 − SMM1 ] = 2 × [1 − ]
r (1 + r) (N−1) r (1 + r)(N−1)
⇒ A2 = A1 × [1 − SMM1 ]

Similarly, it can be shown that

A3 = A2 × [1 − SMM2 ]
= A1 × [1 − SMM1 ] × [1 − SMM2 ]

In general, the scheduled monthly installment in month t is given by

t−1
At = A1 ∏[1 − SMMi ]
i=1

If SMMi = SMM∀i, then

At = A1 × [1 − SMM]t−1

Let’s denote the monthly payment on the original principal in the absence of pre-
payments by A. Then we can state

t−1
At = A ∏[1 − SMMi ]
i=1

= A × [1 − SMM]t−1 if SMMi = SMM ∀i

Appendix 8.1 | 231

In the absence of prepayments, the interest and principal components of “pay-

ment t” and the outstanding principal at the end of t periods are given by

1
It ∗ = A × [1 − ]
(1 + r)N−t+1
A
Pt ∗ =
(1 + r)N−t+1
A 1
SP t ∗ = [1 − ]
r (1 + r)N−t
In the presence of prepayments, the actual principal outstanding at the end of t −1
is given by

At 1
Lt−1 = × [1 − ]
r (1 + r)N−t+1
Thus, in the presence of prepayments, the interest and principal components of
the t th payment are given by

1
It = At × [1 − ]
(1 + r)N−t+1
and
At
Pt =
(1 + r)N−t+1
The scheduled principal at the end of the “period t” is given by

At 1
SP t = [1 − ]
r (1 + r)N−t
We know that
t−1
At = A ∏[1 − SMMi ]
i=1

Thus
t−1
It = It ∗ ∏[1 − SMMi ]
i=1
t−1
Pt = Pt ∗ ∏[1 − SMMi ]
i=1

and
t−1
SP t = SP t ∗ ∏[1 − SMMi ]
i=1
232 | Chapter 8: Mortgage Loans

The actual principal at the end of t periods is given by

Lt = SP t × [1 − SMMt ]
t
⇒ Lt = SP t ∗ ∏[1 − SMMi ]
i=1
Chapter 9
Mortgage-Backed Securities

Figure 9.1: The process of securitization.

Mortgage loans are extremely illiquid. Besides, the lender faces significant exposure
to credit risk as well as prepayment risk. Thus making a secondary market for whole
loans is an extremely difficult proposition. However, it is possible to issue liquid debt
securities that are backed by underlying pools of mortgage loans, known as mortgage-
backed securities (MBS). That is, the cash flows stemming from the underlying loans
are passed through to the holders of these debt securities, and hence the name pass-
through. Each holder of a pass-through security is entitled to a prorata undivided
share of each cash flow that emanates from the underlying pools, as and when the
homeowners make monthly payments. Each monthly payment consists of an interest
component, a principal component, and potentially an additional amount on account
of prepayment. Any amount that constitutes a prepayment is termed an unscheduled
principal, as opposed to the scheduled or expected principal repayment. Such pooling
of multiple loans makes the issuance of a large quantity of mortgage-backed securities
feasible, and consequently improves their liquidity. Also it facilitates the diversifica-
tion of credit risk, and other risks such as geographic risks.
In the U.S. the Government National Mortgage Association (GNMA), which is a
federal government entity, and quasi-government agencies such as the Federal Na-
tional Mortgage Association (FNMA) and the Federal Home Loan Mortgage Corpo-
ration (Freddie Mac), issue mortgage-backed securities. These are termed as Ginnie
Maes, Fannie Maes, and Freddie Macs in the market. There are also non-agency se-
curities termed as private label securities. Agency passthroughs are a unique feature
of the U.S. market, and equivalent examples may not be available in other markets.

https://doi.org/10.1515/9781547400669-009
234 | Chapter 9: Mortgage-Backed Securities

GNMA guarantees the timely payment of principal and interest on securities that are
issued by lenders approved by it. This guarantee enables the corresponding securities
to be priced higher. It must be understood the GNMA does not acquire mortgage loans,
and nor does it issue securities. In practice, private institutions, approved by GNMA,
originate and pool loans and issue securities backed by them. The resulting securi-
ties are guaranteed by GNMA. FNMA purchases mortgage loans from big commercial
banks, whereas Freddie Mac purchases loans from smaller banks and lenders.

Cash Flows for a Pass-Through Security

In practice, the cash flow that is passed through to investors in a given period is not
equal to the payment received from the underlying pool. The monthly cash flow for
a pass-through is less than the monthly cash flow from the underlying loans by an
amount equal to the servicing and guaranteeing fees. Thus, the coupon rate on the
pass-through is less than the mortgage rate of the underlying pool by an amount equal
to the servicing and guaranteeing fees. Let’s assume a servicing fee of 0.6% per annum
and a guaranteeing fee of 0.6% per annum. If the mortgage rate is 9% per annum, the
coupon rate of the pass-through is 7.80%. The examples in this chapter are generic in
nature, and are not intended to illustrate the structure of an agency pass-through in
the U.S.
The monthly cash flow consists of the net interest, which is the gross interest less
the servicing and guaranteeing fees, the scheduled principal payment, and the pre-
payment.
The cash flow for the month is different from the total principal for the month, for
the latter is the sum of the scheduled principal payment and the prepayment for the
month. It does not include the net interest. The average life of a security is based on the
total principal received. However, the price of a pass-through and its bond equivalent
yield, which we define in the next section, are based on the cash flow stream. The
mechanics of a pass-through are illustrated in Example 9.1.

Example 9.1. Consider a home loan with a principal of $500,000 and 15 months to maturity. The mort-
gage rate is 9% per annum. 500 securities backed by these loans have been created. These securities
have been distributed as follows: 200 to Michael, 200 to Scott, and 100 to Maureen. These securities
are identical in all respects, and the holder of a security is entitled to 0.2% of any cash flow that arises
from the underlying loans.
As we mentioned in the chapter introduction, each holder is entitled to a pro rata undivided share
of any cash flow. Undivided may be understood as follows. Michael owns 40% of the securities which
are equivalent to 40% of the underlying collateral that is tantamount to one loan of $200,000. How-
ever, by virtue of the fact that he has 40% of the securities, it cannot be construed that Michael owns
one underlying loan in its entirety. All that we can say is that Michael is entitled to 40% of each cash
flow that emanates from the underlying pool.
Cash Flows for a Pass-Through Security | 235

The total cash flow for the first month, as shown in Table 9.1, is:

3,250 + 31,618.2 + 11,709.6 = $46,577.80

Because 500 securities have been issued, the first month’s cash flow is $93.16 per security. This con-
sists of $6.50 of interest, $63.24 of scheduled principal, and $23.42 of unscheduled principal or pre-
payment. Because Michael and Scott own 200 securities each, they each receive

200 × 93.16 = $18,632

Maureen has 100 securities, so she receives $9,316.

Now assume that at the end of the first month there is an additional pre-payment of $15,000.
Each security is entitled to $30 extra. As Michael and Scott have 200 securities each, they receive an
additional payment of $6,000, whereas Maureen receives an additional payment of $3,000.

Table 9.1: Cash flows for a pass-through security.

Time Inst. SFee GFee NINT SPR SOUT PREP AOUT

0 500,000 500,000
1 35,368.2 250.0 250.0 3,250.0 31,618.2 468,381.8 11,709.6 456,672.3
2 34,484.0 228.3 228.3 2,968.4 31,059.0 425,613.3 10,640.3 414,973
3 33,621.9 207.5 207.5 2,697.3 30,509.6 384,463.4 9,611.6 374,851.8
4 32,781.3 187.4 187.4 2,436.5 29,970.0 344,881.9 8.622.0 336,259.8
5 31,961.8 168.1 168.1 2,185.7 29,439.9 306,819.9 7,670.5 299,149.4
6 31,162.8 149.6 149.6 1,944.5 28,919.1 270,230.3 6,755.8 263,474.5
7 30,383.7 131.7 131.7 1,712.6 28,407.6 235,066.9 5,876.7 229,190.2
8 29,624.1 114.6 114.6 1,489.7 27,905.2 201,285.1 5,032.1 196,252.9
9 28,883.5 98.1 98.1 1,275.6 27,411.6 168,841.3 4,221.0 164,620.3
10 28,161.4 82.3 82.3 1,070.0 26,926.8 137,693.5 3,442.3 134,251.2
11 27,457.4 67.1 67.1 872.6 26,450.5 107,800.7 2,695.0 105,105.7
12 26,770.9 52.6 52.6 683.2 25,982.7 79,123.0 1,978.1 77,145.0
13 26,101.7 38.6 38.6 501.4 25,523.1 51,621.9 1,290.6 50,331.3
14 25,499.1 25.2 25.2 327.2 25,071.6 25,259.7 631.5 24,628.2
15 24,812.9 12.3 12.3 160.1 24,628.2 0 0 0
Inst. is the periodic installment.
SFee is the servicing fee.
GFee is the guaranteeing fee.
NINT is the net interest.
SPR is the scheduled principal repayment.
SOUT is the scheduled outstanding at the end of the month.
PREP is the prepayment for the month.
AOUT is the actual outstanding at the end of the month.

The average life is 7.25 months, assuming the SMM is 2.50%. The monthly IRR of
a security under three different price assumptions is given is Table 9.2.
Table 9.3 depicts the cash flows and the IRR for each of the three cases, assuming
the SMM increases from 2.50% to 5.00%.
236 | Chapter 9: Mortgage-Backed Securities

Table 9.2: Total principal and cash flow for the pass-through security.

Time Total Principal Cash Flow Per Security

Case-A Case-B Case-C

0 0 (975) (1,000) (1,025)

1 43,327.74 93.16 93.16 93.16
2 41,699.28 89.34 89.34 89.34
3 40,121.18 85.64 85.64 85.64
4 38,592.00 82.06 82.06 82.06
5 37,110.36 78.59 78.59 78.59
6 35,674.90 75.24 75.24 75.24
7 34,284.31 71.99 71.99 71.99
8 32,937.30 68.85 68.85 68.85
9 31,632.64 65.82 65.82 65.82
10 30,369.10 62.88 62.88 62.88
11 29,145.51 60.04 60.04 60.04
12 27,960.73 57.29 57.29 57.29
13 26,813.63 54.63 54.63 54.63
14 25,703.14 52.06 52.06 52.06
15 24,628.19 49.58 49.58 49.58
IRR 1.0137% 0.65% 0.30%

Table 9.3: The case of a 5% SMM and cash flows from a pass-through security.

Time Cash Flow Per Security

Case-A Case-B Case-C
0 (975) (1,000) (1,025)
1 116.57 116.57 116.57
2 107.78 107.78 107.78
3 99.55 99.55 99.55
4 91.86 91.86 91.86
5 84.66 84.66 84.66
6 77.94 77.94 77.94
7 71.66 71.66 71.66
8 65.80 65.80 65.80
9 60.33 60.33 60.33
10 55.22 55.22 55.22
11 50.46 50.46 50.46
12 46.02 46.02 46.02
13 41.89 41.89 41.89
14 38.04 38.04 38.04
15 34.46 34.46 34.46
IRR 1.0554% 0.65% 0.2603%

The average life declines to 6.50 months. As should be obvious, the higher the SMM,
the lower is the average life.
Cash Flows for a Pass-Through Security | 237

Cash Flow Yield of a Pass-Through Security

The cash flow yield for a pass-through is the IRR computed using the projected cash
flow stream. The result is a monthly rate, which is usually converted to a bond equiva-
lent yield to facilitate comparisons with conventional debt securities. If we denote the
IRR as im , where im denotes a monthly rate, then the bond equivalent yield may be
expressed as

2[(1 + im )6 − 1]

As you can see from Tables 9.2 and 9.3, the IRR, and consequently the BEY, for a pass-
through is inversely related to the price of the security, as is the case for a conventional
debt security. If the security is priced at par, the cash flow yield is equal to the coupon
rate, irrespective of the rate that is assumed for prepayment.
This can be demonstrated as follows. Let’s assume for ease of exposition that the
SMM is the same for all months. This is a convenient but not essential assumption.
Let’s also assume that there are no servicing or guaranteeing fees.

Symbols Required for the Exposition

We use the following symbols:

– CF t ≡ cash flow from the pass-through in month t.
– It ≡ the interest component of the cash flow in month t.
– Pt ≡ the scheduled principal component of the cash flow in month t.
– St ≡ the prepayment rate in month t.
– A ≡ the monthly installment in the absence of prepayments.
– It ∗ ≡ the interest component of the cash flow in month t in the absence of prepay-
ments.
– Pt ∗ ≡ the principal component of the cash flow in month t in the absence of pre-
payments.
– y ≡ the IRR of the pass-through.
– r ≡ the coupon rate of the underlying mortgage loan.
– SP t ≡ the scheduled outstanding balance at the end of month t.

The cash flow in month t is

CF t = [It + Pt + St × SP t ]
t−1 t−1 t−1
= It ∗ ∏[1 − Si ] + Pt ∗ ∏[1 − Si ] + St × SP t ∗ ∏[1 − Si ]
i=1 i=1 i=1
t−1 t−1
A × St 1
= A ∏[1 − Si ] + × ∏[1 − Si ] × [1 − ]
i=1
r i=1 (1 + r)(N−t)
238 | Chapter 9: Mortgage-Backed Securities

t−1 S
St t−1 r
t t−1
= A [∏[1 − Si ] + ∏[1 − Si ] − ∏ [1 − Si ]]
i=1
r i=1 (1 + r)(N−t) i=1

Now assume that Si = S ∀ i:

N
CF t (1 − S)N A(r + S)
∑ t
= [1 − ]×
i=1 (1 + y) (1 + y) N r(y + S)
A × S(1 + r) 1 (1 − S)N
− [ − ]
r[(y − r) + S(1 + r)] (1 + r)N (1 + y)N
Lr(1 + r)N
A=
(1 + r)N − 1

If r = y then

N
CF t
∑ =L
i=1 (1 + y)t

Thus irrespective of the prepayment speed, if the pass-through is trading at par,

the IRR is equal to the coupon rate.

Collateralized Mortgage Obligations

In the case of a pass-through, there is only one class of security that is issued, and
every security is entitled to the same undivided share of the cash flows from the un-
derlying pool. The feature of such securities is that prepayments have the same impli-
cations for all the security holders.
In real life, however, investors differ with respect to the prepayment exposure and
the average security life sought by them. The securitization process can take these
preferences into account by creating multiple classes of securities, or what are referred
to in MBS parlance as tranches. There are clearly specified rules as to how the cash
flows from the underlying pool are to be directed to the holders of the various tranches.
Such securities, with multiple tranches, are referred to as collateralized mortgage obli-
gations or CMOs. Let’s introduce the concept by studying a very simple CMO structure
known as a sequential pay CMO.

Example 9.2. Consider the mortgage loan with a size of $500,000. The time to maturity is 15 months,
and the SMM is 2.50% for all months. Instead of issuing 500 securities with a face value of $1,000
each, we issue 250 tranche A securities, with a face value of $1,000, 150 tranche B securities with a
face value of $1,000, and 100 tranche C securities with a face value of $1,000.
The cash flows from the underlying mortgage loan are directed as follows. All principal payments
from the underlying loan, both scheduled and unscheduled, first go to tranche A. During this time,
the remaining tranches continue to earn interest on the outstanding principal. Once tranche A is fully
paid off, all subsequent principal payments, scheduled and unscheduled, are directed to the holders
Collateralized Mortgage Obligations | 239

of tranche B. Once again, while tranche B is being paid off, holders of tranche C get interest on their
outstanding principal. Extending the logic, the redemption of tranche B results in further principal
payments being directed to tranche C.
Let’s do an analysis month by month to illustrate the concepts.

Month-1
Total principal = $43,327.74
Net interest = $3,250
Tranche B and C holders receive a total interest of (0.078/12) × 250,000 = $1,625.
Tranche A receives the entire principal and 3,250 − 1,625 = $1,625 by way of interest.

$ 1,625 = (0.078/12) × 250,000

The outstanding balance for tranche A is 250,000 − 43,327.74 = $206,672.26.

Month-2
Total principal = $41,699.28
Net interest = $2,968.40
Tranche B and C holders receive a total interest of $1,625.
Tranche A receives the entire principal and $1,343.40 by way of interest.

$ 1,343.40 = (0.078/12) × 206,672.26

The outstanding balance for tranche A is 206,672.26 − 41,699.28 = $164,973.00.

Month-3
Total principal = $40,121.18
Net interest = $2,697.30
Tranche B and C holders receive a total interest of $1,625.
Tranche A receives the entire principal and $1,072.30 by way of interest.

$ 1,072.30 = (0.078/12) × 164,973.00

The outstanding balance for tranche A is 164,972.98 − 40,121.18 = $124,851.80.

Month-4
Total principal = $38,592.00
Net interest = $2,436.50
Tranche B and C holders receive a total interest of $1,625.
Tranche A receives the entire principal and $811.50 by way of interest.

$ 811.50 = (0.078/12) × 124,851.80

The outstanding balance for tranche A is 124,851.80 − 38,592.00 = $86,259.80.

Month-5
Total principal = $37,110.36
Net interest = $2,185.70
Tranche B and C holders receive a total interest of $1,625.
Tranche A receives the entire principal and $560.70 by way of interest.

$ 560.70 = (0.078/12) × 86,259.80

The outstanding balance for tranche A is 86,259.80 − 37,110.36 = $49,149.44.

240 | Chapter 9: Mortgage-Backed Securities

Month-6
Total principal = $35,674.90
Net interest = $1,944.50
Tranche B and C holders receive a total interest of $1,625.
Tranche A receives the entire principal and $319.50 by way of interest.

$ 319.50 = (0.078/12) × 49,149.44

The outstanding balance for tranche A is 49,149.44 − 35,674.90 = $13,474.54.

Month-7
Total principal = $34,284.31
Net interest = $1,712.60
Tranche B and C holders receive a total interest of $1,625.
Tranche A receives $13,474.54 by way of principal and $87.60 by way of interest.

$1,625.00 + $87.60 = $1,712.60

$ 87.60 = (0.078/12) × 13,474.54

The outstanding balance for tranche A is zero.

Tranche B gets 34,284.31 − 13,474.54 = $20,809.77 by way of principal.
The outstanding balance for tranche B is 150,000 − 20,809.77 = $129.190.23.

Month-8
Total principal = $32,937.30
Net interest = $1,489.70
Tranche C receives (0.078/12) × 100,000 = $650 by way of interest.
Tranche B receives the entire principal and $839.70 by way of interest.

$ 839.70 = (0.078/12) × 129,190.23

The outstanding balance for tranche B is 129,190.23 − 32,937.30 = $96,252.93.

Month-9
Total principal = $31,632.64
Net interest = $1,275.60
Tranche C receives $650 by way of interest.
Tranche B receives the entire principal and $625.60 by way of interest.

$ 625.60 = (0.078/12) × 96,252.93

The outstanding balance for tranche B is 96,252.93 − 31,632.64 = $64,620.29.

Month-10
Total principal = $30,369.10
Net interest = $1,070.00
Tranche C receives $650 by way of interest.
Tranche B receives the entire principal and $420.00 by way of interest.

$ 420.00 = (0.078/12) × 64,620.29

The outstanding balance for tranche B is 64,620.29 − 30,369.10 = $34,251.19.

Collateralized Mortgage Obligations | 241

Month-11
Total principal = $29,145.51
Net interest = $872.60
Tranche C receives $650 by way of interest.
Tranche B receives the entire principal and $222.60 by way of interest.

$ 222.60 = (0.078/12) × 34,251.19

The outstanding balance for tranche B is 34,251.19 − 29,145.51 = $5,105.68.

Month-12
Total principal = $27,960.73
Net interest = $683.20
Tranche C receives $650 by way of interest.
Tranche B receives $5,105.68 by way of principal and $33.20 by way of interest.

$ 33.20 = (0.078/12) × 5,105.68

Tranche C receives 27,960.73 − 5,105.68 = $22,855.05 by way of principal.

The outstanding balance for tranche C is 100,000 − 22,855.05 = $77,144.95.
Month-13
Total principal = $26,813.63
Net interest = $501.40
The entire principal as well as the interest goes to tranche C.

$ 501.40 = (0.078/12) × 77,144.95

The outstanding balance for tranche C is 77,144.95 − 26,813.63 = $50,331.32.

Month-14
Total principal = $25,703.14
Net interest = $327.20
The entire principal as well as the interest goes to tranche C.

$ 327.20 = (0.078/12) × 50,331.32

The outstanding balance for tranche C is 50,331.32 − 25,703.14 = $24,628.18.

Month-15
Total principal = $24,628.19
Net interest = $160.10
The entire principal as well as the interest goes to tranche C.

$ 160.10 = (0.078/12) × 24,628.18

The outstanding balance for tranche C is zero.

Tranche A has an average life of 3.5815 months; tranche B has an average life of 9.1961
months; and tranche C has an average life of 13.5210 months. The mother loan has an
average life of 7.2538 months.
7.2538 = 0.50 × 3.5815 + 0.30 × 9.1961 + 0.20 × 13.5210
Thus the average life of the mother loan is a weighted average of the average lives of
the tranches.
242 | Chapter 9: Mortgage-Backed Securities

Table 9.4: Summary of a Sequential Pay CMO.

T Tranche A Tranche B Tranche C

Int. Prin. Out. Int. Prin. Out. Int. Prin. Out.

0 250,000.0 150,000.0 100,000.0

1 1,625.0 43,327.7 206,672.3 975.0 0 150,000.0 650.0 0 100,000.0
2 1,343.4 41,699.3 164,973.0 975.0 0 150,000.0 650.0 0 100,000.0
3 1,072.3 40,121.2 124,851.8 975.0 0 150,000.0 650.0 0 100,000.0
4 811.5 38,592.0 86,259.8 975.0 0 150,000.0 650.0 0 100,000.0
5 560.7 37,110.4 49,149.4 975.0 0 150,000.0 650.0 0 100,000.0
6 319.5 35,674.9 13,474.5 975.0 0 150,000.0 650.0 0 100,000.0
7 87.6 13,474.5 0 975.0 20,809.8 129,190.2 650.0 0 100,000.0
8 0 0 0 839.7 32,937.3 96,252.9 650.0 0 100,000.0
9 0 0 0 625.6 31,632.6 64,620.3 650.0 0 100,000.0
10 0 0 0 420.0 30,369.1 34,251.2 650.0 0 100,000.0
11 0 0 0 222.6 29,145.5 5,105.7 650.0 0 100,000.0
12 0 0 0 33.2 5,105.7 0 650.0 22,855.1 77,144.9
13 0 0 0 0 0 0 501.4 26,813.6 50,331.3
14 0 0 0 0 0 0 327.2 25,703.1 24,628.2
15 0 0 0 0 0 0 160.1 24,628.2 0
T is the time.
Int. is the interest received in the month.
Prin. is the principal payment received in the month.
Out. is the outstanding balance at the end of the month.

Extension Risk and Contraction Risk for Mortgage-backed

Securities
Pass-through securities and CMOs expose investors to prepayment risk. In a declining
interest rate environment, prepayments impact security holders in two ways. As they
are debt securities, we would expect the price of such securities to rise in a scenario
where rates are declining. However, in the case of mortgage-backed securities, holders
of underlying bonds are likely to prepay and refinance at a lower rate when market
rates go down. The overall result is price compression; that is, the rise in price is not
as large as in the case of plain vanilla bonds. The second adverse consequence for
holders of such securities is, of course, reinvestment risk, as is the case for all debt
holders. That is, cash flows that are received in a declining interest rate environment
obviously have to be reinvested at lower rates of interest. The risk due to these two
adverse consequences is termed as contraction risk.
Now let’s consider a rising interest rate environment. We would obviously expect
the prices of mortgage-backed securities to fall. However, unlike the case of plain
vanilla bonds, principal payments are likely to slow down because holders of underly-
ing mortgage loans are less likely to refinance when mortgage rates are rising. This ex-
acerbates the anticipated price decline. In such a situation, the security holders want
prepayments to be high to afford them an opportunity to reinvest larger amounts at
Accrual Bonds | 243

high rates of interest. The adverse impact for mortgage-backed securities in a rising
interest rate environment is termed extension risk.

Accrual Bonds
In the case of the sequential pay CMO that we just studied, every tranche receives inter-
est every month, based on the principal outstanding for that particular tranche at the
beginning of the month. However, many sequential pay CMO structures do not require
that every tranche be paid interest every month based on the outstanding principal.
In the case of a sequential pay CMO with an accrual bond, also termed as a Z-bond,
the Z bond does not receive any monthly interest until the previous tranches are fully
retired. Therefore, the interest for the Z-bond accrues and is added to its original prin-
cipal, leading to negative amortization, until the other classes of the CMO have been
paid off. In the months prior to the retirement of the penultimate tranche, the interest
that would otherwise have been paid to the Z-bond is directed to the tranche that is
receiving principal at that point in time and thus helps to speed up the repayment of
principal.
Let’s consider a CMO in which tranches A and B are as specified in Example 9.2.
However, the last tranche is an accrual bond. Quite obviously, the inclusion of the Z-
bond reduces the maturity as well as the average life of each of the two other tranches.

Example 9.3. Let’s do an analysis month by month to illustrate the concepts.

Month-1
Total principal = $43,327.74
Net interest = $3,250
Tranche B holders receive a total interest of (0.078/12) × 150,000 = $975.
Tranche A holders receive an interest of (0.078/12) × 250,000 = $1,625.
Tranche A holders receive a total principal of 43,327.74 + (3,250 − 975 − 1,625) = $43,977.74.
The unpaid interest on the Z-bond is (0.078/12) × 100,000 = $650.
The outstanding balance for tranche A is 250,000 − 43,977.74 = $206,022.26.
The outstanding balance for tranche B is $150,000.
The outstanding balance for the Z-bond is 100,000 + 650 = $100,650.

Month-2
Total principal = $41,699.28
Net interest = $2,968.40
Tranche B holders receive a total interest of (0.078/12) × 150,000 = $975.
Tranche A holders receive an interest of (0.078/12) × 206,022.26 = $1,339.14.
Tranche A holders receive a total principal of 41,699.28 + (2,968.4 − 975 − 1,339.14) = $42,353.54.
The unpaid interest on the Z-bond is (0.078/12) × 100,650 = $654.23.
The outstanding balance for tranche A is 206,022.26 − 42,353.54 = $163,668.72.
The outstanding balance for tranche B is $150,000.
The outstanding balance for the Z-bond is 100,650 + 654.23 = $101,304.23.

Month-3
Total principal = $40,121.18
244 | Chapter 9: Mortgage-Backed Securities

Net interest = $2,697.30

Tranche B holders receive a total interest of (0.078/12) × 150,000 = $975.
Tranche A holders receive an interest of (0.078/12) × 163,668.72 = $1,063.85.
Tranche A holders receive a total principal of 40,121.18 + (2,697.3 − 975 − 1,063.85) = $40,779.63.
The unpaid interest on the Z-bond is (0.078/12) × 101,304.23 = $658.48.
The outstanding balance for tranche A is 163,668.72 − 40,779.634 = $122,889.09.
The outstanding balance for tranche B is $150,000.
The outstanding balance for the Z-bond is 101,304.23 + 658.48 = $101,962.71.

Month-4
Total principal = $38,592.00
Net interest = $2,436.50
Tranche B holders receive a total interest of (0.078/12) × 150,000 = $975.
Tranche A holders receive an interest of (0.078/12) × 122,889.09 = $798.78.
Tranche A holders receive a total principal of 38,592 + (2,436.50 − 975 − 798.78) = $39,254.72.
The unpaid interest on the Z-bond is (0.078/12) × 101,962.71 = $662.76.
The outstanding balance for tranche A is 122,889.09 − 39,254.72 = $83,634.37.
The outstanding balance for tranche B is $150,000.
The outstanding balance for the Z-bond is 101,962.71 + 662.76 = $102,625.57.

Month-5
Total principal = $37,110.36
Net interest = $2,185.70
Tranche B holders receive a total interest of (0.078/12) × 150,000 = $975.
Tranche A holders receive an interest of (0.078/12) × 83,634.37 = $543.62.
Tranche A holders receive a total principal of 37,110.36 + (2,185.70 − 975 − 543.62) = $37,777.44.
The unpaid interest on the Z-bond is (0.078/12) × 102,625.57 = $667.07.
The outstanding balance for tranche A is 83,634.37 − 37,777.44 = $45,856.93.
The outstanding balance for tranche B is $150,000.
The outstanding balance for the Z-bond is 102,625.57 + 667.07 = $103,292.64.

Month-6
Total principal = $35,674.90
Net interest = $1,944.50
Tranche B holders receive a total interest of (0.078/12) × 150,000 = $975.
Tranche A holders receive an interest of (0.078/12) × 45,856.93 = $298.07.
Tranche A holders receive a total principal of 35,674.90 + (1,944.50 − 975 − 298.07) = $36,346.33.
The unpaid interest on the Z-bond is (0.078/12) × 103,292.64 = $671.40.
The outstanding balance for tranche A is 45,856.93 − 36,346.33 = $9,510.60.
The outstanding balance for tranche B is $150,000.
The outstanding balance for the Z-bond is 103,292.64 + 671.40 = $103,964.04.

Month-7
Total principal = $34,284.31
Net interest = $1,712.60
Tranche B holders receive a total interest of (0.078/12) × 150,000 = $975.
Tranche A holders receive an interest of (0.078/12) × 9,510.60 = $61.82.
Tranche A holders receive a total principal of $9,510.60.
The unpaid interest on the Z-bond is (0.078/12) × 103,964.04 = $675.77.
Tranche B holders receive a principal payment of (34,284.31 − 9,510.60) + (1,712.60 − 975 − 61.82) =
$25,449.50.
Accrual Bonds | 245

The outstanding balance for tranche A is zero.

The outstanding balance for tranche B is 150,000 − 25,449.50 = $124,550.50.
The outstanding balance for the Z-bond is 103,964.04 + 675.77 = $104,639.81.

Month-8
Total principal = $32,937.30
Net interest = $1,489.70
The unpaid interest on the Z-bond is (0.078/12) × 104,639.81 = $680.16.
Tranche B holders receive an interest of (0.078/12) × 124,550.50 = $809.58.
Tranche B holders receive a principal payment of (32,937.30 + 1,489.70 − 809.58) = $33,617.42.
The outstanding balance for tranche B is 124,550.50 − 33,617.42 = $90,933.08.
The outstanding balance for the Z-bond is 104,639.81 + 680.16 = $105,319.97.

Month-9
Total principal = $31,632.64
Net interest = $1,275.60
The unpaid interest on the Z-bond is (0.078/12) × 105,319.97 = $684.58.
Tranche B holders receive an interest of (0.078/12) × 90,933.08 = $ 591.07.
Tranche B holders receive a principal payment of (31,632.64 + 1,275.60 − 591.07) = $32,317.17.
The outstanding balance for tranche B is 90,933.08 − 32,317.17 = $58,615.91.
The outstanding balance for the Z-bond is 105,319.97 + 684.58 = $106,004.55.

Month-10
Total principal = $30,369.10
Net interest = $1,070.00
The unpaid interest on the Z-bond is (0.078/12) × 106,004.55 = $689.03.
Tranche B holders receive (0.078/12) × 58,615.91 = $381 by way of interest.
Tranche B holders receive a principal payment of (30369.1 + 1,070 − 381) = $31,058.10.
The outstanding balance for tranche B is (58,615.91 − 31,058.10) = $27,557.81.
The outstanding balance for the Z-bond is 106,004.55 + 689.03 = $106,693.58.

Month-11
Total principal = $29,145.51
Net interest = $872.60
Tranche B holders receive (0.078/12) × 27,557.81 = $179.13 by way of interest.
Tranche B holders receive a principal payment of $27,557.81.
The outstanding balance for tranche B is zero.
The interest for the Z-bond is (0.078/12) × 106,693.58 = $693.51.
The Z-bond holders receive a principal payment of (29,145.51 + 872.60 − 179.13 − 27,557.81 − 693.51) =
$1,587.66.
The outstanding balance on the Z-bonds is (106,693.58 − 1,587.66) = $105,105.90.

Month-12
Total principal = $27,960.73
Net interest = $683.20
Principal for the Z-bond is $27,960.73.
Interest for the Z-bond is $683.20. 0.078
12
× 105,105.90 = $683.20.
The outstanding balance on the Z-bonds is (105,105.90 − 27,960.73) = $77,145.20.

Month-13
Total principal = $26,813.63
Net interest = $501.40
246 | Chapter 9: Mortgage-Backed Securities

Principal for the Z-bond is $26,813.63.

Interest for the Z-bond is $501.40.
The interest for the Z-bond is (0.078/12) × 77,145.20 = $501.40.
The outstanding balance on the Z-bonds is (77,145.20 − 26,813.63) = $50,331.60.
Month-14
Total principal = $25,703.14
Net interest = $327.20
Principal for the Z-bond is $25,703.14.
Interest for the Z-bond is $327.20.
The interest for the Z-bond is (0.078/12) × 50,331.608 = $327.20.
The outstanding balance on the Z-bonds is (50,331.60 − 25,703.14) = $24,628.45.
Month-15
Total principal = $24,628.19
Net interest = $160.10
Principal for the Z-bond is $24,628.19.
Interest for the Z-bond is $160.10.
The interest for the Z-bond is (0.078/12) × 24,628.45 = $160.10.
The outstanding balance on the Z-bonds is zero.

Table 9.5: Summary of a CMO with an accrual bond.

T Tranche A Tranche B Z-Bond

Int. Prin. Out. Int. Prin. Out. Int. Prin. Out.
0 250,000.0 150,000.0 100,000.0
1 1,625.0 43,977.7 206,022.3 975.0 0 150,000.0 0 0 100,650.0
2 1,339.1 42,353.5 163,668.8 975.0 0 150,000.0 0 0 101,304.2
3 1,063.8 40,779.7 122,889.1 975.0 0 150,000.0 0 0 101,962.7
4 798.8 39,254.8 83,634.3 975.0 0 150,000.0 0 0 102,625.5
5 543.6 37,777.4 45,856.9 975.0 0 150,000.0 0 0 103,292.5
6 298.1 36,346.3 9,510.6 975.0 0 150,000.0 0 0 103,963.9
7 61.8 9,510.6 0 975.0 25,449.5 124,550.5 0 0 104,639.7
8 0 0 0 809.6 33,617.5 90,933.1 0 0 105,319.9
9 0 0 0 591.1 32,317.2 58,615.9 0 0 106,004.4
10 0 0 0 381.0 31,058.1 27,557.7 0 0 106,693.5
11 0 0 0 179.1 27,557.7 0.0 693.5 1,587.80 105,105.7
12 0 0 0 0 0 0 683.2 27,960.7 77,145.0
13 0 0 0 0 0 0 501.4 26,813.6 50,331.3
14 0 0 0 0 0 0 327.2 25,703.1 24,628.2
15 0 0 0 0 0 0 160.1 24,628.2 0
T is the time.
Int. is the interest received in the month.
Prin. is the principal payment received in the month.
Out. is the outstanding balance at the end of the month.

The average life of tranche A is 3.5263 months, and that of tranche B is 9.0110. As we
can see, the average life of each of the tranches has reduced due to the presence of the
Creating Floating Rate Tranches | 247

Z-bond. Z-bonds have appeal to investors who are primarily concerned with the specter
of reinvestment risk. Because such bonds do not entail the receipt of any cash flows
until all the other classes are fully retired, reinvestment risk is totally eliminated until
the Z-bond starts receiving its first cash flow. Accrual bonds are popular with pension
funds and life insurance companies, since they are securities with high durations,
but which offer higher yields than other high duration securities such as long-term
Treasuries.

Creating Floating Rate Tranches

Floating rate bonds can be created from a fixed rate tranche. Let’s illustrate this using
tranche A of the sequential pay CMO, which has a par value of $250,000 and a 7.80%
coupon. In practice, to create a floater, we have to simultaneously create an inverse
floater. The cumulative principal of the two tranches in this case is $250,000. There is
no rule as to what the principal values of the respective tranches should be as long as
their combined par value is $250,000.
Let’s assume that the benchmark for the two tranches is the 3-month LIBOR and
that we want to create a floater with a par value of $150,000 and an inverse floater with
a par value of $100,000. The coupon on the floater is LIBOR + m basis points, and that
for the inverse floater is C − L × LIBOR.
It is a common practice to specify the coupon for an inverse floater as C−L× Bench-
mark. C is the cap or maximum interest on the inverse floater, whereas L is the coupon
leverage. The concept of a cap is easily understood as the maximum possible coupon.
The lower limit for LIBOR is zero. Consequently the upper limit for the coupon on the
inverse floater is C. The coupon leverage measures the rate of change of the coupon of
the inverse floater with respect to the benchmark. We know that the total annual inter-
est available is 7.80% of $250,000, which is $19,500. This is also the maximum interest
available for the floating tranche. Thus the cap on the floater is 19,500÷150,000 = 13%.
The ratio of floaters to inverses is 3:2, which implies that the coupon leverage for the
inverse floater is 1.50. If we assume that the coupon on the floater is LIBOR + 125 bp,
the minimum coupon on the floater is 1.25%, which is the case if LIBOR is zero. Hence,
the minimum interest on the floater is 150,000 × 0.0125 = $1,875. Thus the maximum
interest for the inverse floater is 19,500 − 1,875 = $17,625. The cap on the inverse is
therefore 17,625 ÷ 100,000 = 0.17625 or 17.625%. We can express the coupons of the
two tranches as
Floater Coupon: LIBOR + 1.25%
Inverse Floater Coupon: 17.625% − 1.50× LIBOR
The weighted average coupon is:

0.60 × (LIBOR + 1.25) + 0.40 × (17.625 − 1.50 × LIBOR) = 7.80%

248 | Chapter 9: Mortgage-Backed Securities

Notional Interest-Only Tranches

Let’s go back to the sequential pay CMO that we studied earlier. We assumed that the
underlying mortgage had a coupon rate of 7.80% and that each of the three tranches
that we created also had a coupon of 7.80%. In practice, each tranche of a CMO has a
different coupon rate, and there is an excess representing the difference between the
interest on the underlying collateral and the coupons on the various tranches. There-
fore, it is a common practice to create a tranche that is entitled to receive only the
excess coupon interest. Such tranches are referred to as notional interest-only securi-
ties.
Let’s now assume that the underlying collateral has a coupon of 7.8%, and the
three tranches created from it have coupon rates as shown in Table 9.6.

Table 9.6: Coupon rates for a sequential pay CMO.

Tranche Initial Principal Coupon Rate

A 250,000 6.60%
B 150,000 7.20%
C 100,000 7.50%

Consider tranche A. It pays a coupon of 6.60%, and the coupon on the underlying loan
is 7.80%. Thus there is excess interest of 1.20%. Let’s assume that we want to create
a notional interest-only security with a coupon of 8.0%. To find the notional amount
on which the interest payable to this security is calculated, we proceed as follows. An
excess interest of 1.20% on $250,000 worth of principal can be used to pay an inter-
est of 8.0% on a principal amount of $37,500. Similarly, in the case of tranche B, an
excess interest of 0.60% on a principal amount of $150,000 can be used to pay an in-
terest of 8.0% on a principal of $11,250. Tranche C, using the same logic, generates
excess interest to support $3,750 worth of principal. Thus the excess interest gener-
ated can service a bond with a principal of 37,500 + 11,250 + 3,750 = $52,500. This is
the notional principal for the tranche that is scheduled to receive the excess interest.
The term notional connotes that this principal amount is used merely to compute the
interest payable to this class for the month, and that the principal per se is not repaid.
The contribution of a tranche to the notional principal is in general given by the
following formula:

Notional amount for an x% interest only (IO) class

Par Value of the Tranche × Excess Interest
= x
100
Interest-Only and Principal-Only Strips | 249

Interest-Only and Principal-Only Strips

Stripped mortgage-backed securities can be created by either directing all the prin-
cipal or all the interest to a particular class. The security that is scheduled to receive
only the principal is referred to as the principal-only or PO class. The other security is
obviously termed the interest-only or IO class.
Let’s first consider the PO class. The yield obtained by the security holder depends
on the speed with which the underlying pool generates prepayments. The faster the
speed with which prepayments are received, the greater the rate of return for holders
of such securities is. In the case of the IO class holders, however, the desire is to have
slow prepayments. Such securities do not receive any principal payments, and the
income, which is exclusively by way of interest payments will be higher, the larger the
outstanding balance on the underlying pool. This corresponds to a slower prepayment
rate on the underlying pool. In the case of a PO security, we know the amount of cash
to be received but not the timing of the cash flows. However, in the case of IO securities
we know neither the magnitude of the cash flows nor the timing.

Planned Amortization Class (PAC) Bonds

In the case of planned amortization class (PAC) bonds, if the prepayment pattern falls
within a prespecified range, then the cash flows received by the security are known.
PAC bonds are not created in isolation. They are accompanied by a category of bonds
called support bonds. It is this category that absorbs the pre-payment risk, thereby
protecting the PAC bonds against both extension risk and contraction risk. To create
a PAC bond, we need to specify a range of prepayment speeds. These are referred to
as the PAC collars. The PAC schedule lists the principal amount payable to the PAC
bondholders if the prepayment speed is within the range provided by the lower PAC
collar and the upper PAC collar. The schedule is obtained by taking the lower of the
principal cash flows generated by the lower collar and the upper collar.

Example 9.4. Consider a home loan with a principal of $250,000 and 15 months to maturity. The mort-
gage rate is 7.2% per annum. There are no servicing or guaranteeing fees. A PAC bond and a support
bond are created, where the lower PAC collar is an SMM of 2.50% and the upper collar is an SMM of
12.50%. Table 9.7 lists the principal payments from the underlying loan assuming three different pre-
payment speeds between the lower and upper collars, namely 2.50%, 7.50%, and 12.50%. As before,
let’s do a month-by-month analysis.
The scheduled principal for month t is
t−1 t−1
A
Pt = Pt ∏[1 − SMMi ] = ∏[1 − SMMi ]
∗

i=1 (1 + r)N−t+1 i=1

The prepayment for the month is

t−1
A 1
SP t × SMMi = × [1 − ] ∏[1 − SMMi ] × SMMi
r (1 + r) N−t
i=1
250 | Chapter 9: Mortgage-Backed Securities

The total principal that is received in a month is

t−1 t−1
A A 1
PCF t = ∏[1 − SMMi ] + × [1 − ] ∏[1 − SMMi ] × SMMi
(1 + r) N−t+1
i=1
r (1 + r) N−t
i=1

If we assume a constant SMM for all months,

A A 1
PCF t = [1 − SMM]t−1 + × [1 − ] [1 − SMM]t−1 × SMM
(1 + r) N−t+1 r (1 + r)N−t

Table 9.7: Scheduled principal payments and prepayments under various prepayment speeds.

Time SMM = 2.50% SMM = 7.50% SMM = 12.50%

Sch. Pr Prep. Sch. Pr Prep. Sch. Pr Prep.

1 15,977.83 5,850.55 15,977.83 17,551.66 15,977.83 29,252.77

2 15,671.86 5,312.49 14,868.17 15,120.18 14,064.49 23,838.11
3 15,371.74 4,795.39 13,835.58 12,948.49 12,380.26 19,310.82
4 15,077.37 4,298.57 12,874.70 11,011.75 10,897.73 15,534.75
5 14,788.64 3,821.39 11,980.55 9,287.33 9,592.73 12,393.81
6 14,505.44 3,363.22 11,148.50 7,754.64 8,444.00 9,789.09
7 14,227.66 2,923.45 10,374.24 6,394.98 7,432.83 7,636.35
8 13,955.20 2,501.48 9,653.75 5,191.32 6,542.75 5,863.96
9 13,687.96 2,096.74 8,983.29 4,128.23 5,759.25 4,411.06
10 13,425.83 1,708.68 8,359.40 3,191.66 5,069.58 3,225.98
11 13,168.73 1,336.75 7,778.84 2,368.87 4,462.50 2,264.92
12 12,916.55 980.41 7,238.60 1,648.31 3,928.12 1,490.79
13 12,669.19 639.17 6,735.88 1,019.49 3,457.72 872.23
14 12,426.58 312.53 6,268.07 472.93 3043.66 382.74
15 12,188.61 0 5,832.76 0 2,679.18 0
Sch. Pr stands for scheduled principal.
Prep. stands for prepayment.

The maximum total principal for the PAC bond can be $187,358.11, as can be seen from
Table 9.8. Let’s round down to $187,000. The principal for the last month is 2,679.18 −
(187,358.11 − 187,000) = $2,321.08. The support bond has a principal of 250,000 −
187,000 = $63,000.
The average life for the PAC bond is 5.67 months, and that for the support bond is
11.88 months, when the SMM is 2.50%.
However, when the SMM is 12.50%, the average life for the PAC bond is 5.67
months, and that for the support bond is 2.29 months. We can see that, although the
average life of the PAC bond stays the same as in the case of an SMM of 2.50%, the
average life of the support bond reduces drastically from 11.88 months to 2.29 months.
Thus the stability accorded to the PAC bond is at the cost of the support bond.
Planned Amortization Class (PAC) Bonds | 251

Table 9.8: Total principal under various pre-payment speeds and the total principal for a PAC bond.

Time SMM = 2.50% SMM = 7.50% SMM = 12.50% Minimum

Tot. Pr Tot. Pr Tot. Pr Tot. Pr.

1 21,828.39 33,529.49 45,230.60 21,828.39

2 20,984.35 29,988.35 37,902.60 20,984.35
3 20,167.13 26,784.07 31,691.08 20,167.13
4 19,375.94 23,886.45 26,432.48 19,375.94
5 18,610.03 21,267.88 21,986.54 18,610.03
6 17,868.66 18,903.14 18,233.08 17,868.66
7 17,151.10 16,769.21 15,069.18 15,069.18
8 16,456.68 14,845.07 12,406.71 12,406.71
9 15,784.70 13,111.52 10,170.31 10,170.31
10 15,134.51 11,551.06 8,295.56 8,295.56
11 14,505.47 10,147.71 6,727.42 6,727.42
12 13,896.96 8,886.91 5,418.91 5,418.91
13 13,308.37 7,755.38 4,329.95 4,329.95
14 12,739.11 6,741.00 3,426.40 3,426.40
15 12,188.61 5,832.76 2,679.18 2,679.18
Total 250,000 250,000 250,000 187,358.11
Tot. Pr. stands for total principal.

Table 9.9: Total principal and principal for the PAC bond and the support bond under 2.50% SMM.

Time SMM = 2.50%

Tot. Pr Prin. for Prin. for
PAC Bond Support Bond

1 21,828.39 21,828.39 0.00

2 20,984.35 20,984.35 0.00
3 20,167.13 20,167.13 0.00
4 19,375.94 19,375.94 0.00
5 18,610.03 18,610.03 0.00
6 17,868.66 17,868.66 0.00
7 17,151.10 15,069.18 2,081.93
8 16,456.68 12,406.71 4,049.97
9 15,784.70 10,170.31 5,614.39
10 15,134.51 8,295.56 6,838.95
11 14,505.47 6,727.42 7,778.05
12 13,896.96 5,418.91 8,478.05
13 13,308.37 4,329.95 8,978.42
14 12,739.11 3,426.40 9,312.71
15 12,188.61 2,321.08 9,867.53
Total 250,000 187,000 63,000
Prin. stands for principal.
Tot. Pr. stands for total principal.
252 | Chapter 9: Mortgage-Backed Securities

Table 9.10: Total principal and principal for the PAC bond and the support bond under 12.50% SMM.

Time SMM = 12.50%

Tot. Pr Prin. for Prin. for
PAC Bond Support Bond
1 45,230.60 21,828.39 23,402.22
2 37,902.60 20,984.35 16,918.25
3 31,691.08 20,167.13 11,523.95
4 26,432.48 19,375.94 7,056.54
5 21,986.54 18,610.03 3,376.51
6 18,233.08 17,868.66 364.43
7 15,069.18 15,069.18 0.00
8 12,406.71 12,406.71 0.00
9 10,170.31 10,170.31 0.00
10 8,295.56 8,295.56 0.00
11 6,727.42 6,727.42 0.00
12 5,418.91 5,418.91 0.00
13 4,329.95 4,329.95 0.00
14 3,426.40 3,426.40 0.00
15 2,679.18 2,321.08 358.11
Total 250,000 187,000 63,000
Prin. stands for principal.
Tot. Pr. stands for total principal.

Table 9.11: Total principal and principal for the PAC bond and the support bond under an SMM of
zero.

Time SMM = 0.00%

Tot. Pr Prin. for Out. Bal. for Prin. for
PAC Bond PAC Bond Sup. Bond
1 15,977.83 15,977.83 171,022.17 0.00
2 16,073.70 16,073.70 154,948.47 0.00
3 16,170.14 16,170.14 138,778.33 0.00
4 16,267.16 16,267.16 122,511.17 0.00
5 16,364.76 16,364.76 106,146.40 0.00
6 16,462.95 16,462.95 89,683.45 0.00
7 16,561.73 16,561.73 73,121.72 0.00
8 16,661.10 16,661.10 56,460.62 0.00
9 16,761.07 16,761.07 39,699.55 0.00
10 16,861.63 16,861.63 22,837.91 0.00
11 16,962.80 16,962.80 5,875.11 0.00
12 17,064.58 5,875.11 0.00 11,189.47
13 17,166.97 0.00 0.00 17,166.97
14 17,269.97 0.00 0.00 17,269.97
15 17,373.59 0.00 0.00 17,373.59
Total 250,000 187,000 63,000.00
Tot. Pr. stands for total principal.
Prin. stands for principal.
Out. Bal. stands for outstanding balance.
Sup. stands for support.
Planned Amortization Class (PAC) Bonds | 253

When the SMM is zero, for months 1–6, the total principal received is less than what
is due for the PAC bond. Hence the entire principal is directed to the PAC bond. For
months 7–11, the entire principal received is directed to the PAC bond. In these months,
the cash flow directed to the PAC bond is more than the payment required as per the
PAC schedule. However, because there is a deficit in the earlier months, the entire prin-
cipal is directed to the PAC bond. At the end of the 11th month, the outstanding prin-
cipal is $5,875.11. The total principal that is received in the 12th month is $17,064.58.
Thus $5,875.11 is directed to the PAC bond, which as a consequence is fully paid off,
and $11,189.47 is directed to the support bond. For months 13–15, the entire principal
is directed to the support bond.
The average life for the PAC bond is 6.25 months, and that for the support bond is
13.65 months.

Table 9.12: Total principal and principal for the PAC bond and the support bond under 15.00% SMM.

Time SMM = 15.00%

Tot. Pr Prin. for Out. Bal. for Prin. for
PAC Bond PAC Bond Sup. Bond

1 51,081.16 21,828.39 165,171.61 29,252.77

2 41,451.07 20,984.35 144,187.26 20,466.72
3 33,550.65 20,270.15 123,917.12 13,280.50
4 27,079.13 27,079.13 96,837.99 0.00
5 21,786.83 21,786.83 75,051.16 0.00
6 17,466.67 17,466.67 57,584.49 0.00
7 13,946.98 13,946.98 43,637.50 0.00
8 11,085.62 11,085.62 32,551.89 0.00
9 8,764.93 8,764.93 23,786.96 0.00
10 6,887.67 6,887.67 16,899.29 0.00
11 5,373.50 5,373.50 11,525.78 0.00
12 4,156.16 4,156.16 7,369.62 0.00
13 3,181.02 3,181.02 4,188.60 0.00
14 2,403.12 2,403.12 1,785.48 0.00
15 1,785.48 1,785.48 0.00 0.00
Total 250,000 187,000 63,000.00
Tot. Pr. stands for total principal.
Prin. stands for principal.
Out. Bal. stands for outstanding balance.
Sup. stands for support.

Analysis of the SMM = 15% Scenario

In the first month, the amount due for the PAC bond is $21,828.39, whereas the to-
tal principal received is $51,081.16. The difference of $29,252.77 is paid to the support
bond. The outstanding for the support bond is 63,000 − 29,252.77 = $33,747.23.
254 | Chapter 9: Mortgage-Backed Securities

In the second month, the amount due for the PAC bond is $20,984.35, whereas the
total principal received is $41,451.07. The difference of $20,466.72 is paid to the support
bond. The outstanding for the support bond is 33,747.23 − 20,466.72 = $13,280.51.
In the third month, the total principal received is $33,550.65. The amount paid
to the support bond is $13,280.51. The outstanding for the support bond is zero. The
balance of $20,270.15 is paid to the PAC bond.
Because the support bond has been fully paid off at the end of the third month,
the entire principal received for the remaining months is directed to the PAC bond,
which is fully paid off at the end of the 15th month.
The average life of the PAC bond is 5.30 months, and the average life of the support
bond is 1.75 months.

Chapter Summary
The chapter looked at mortgage-backed securities. We first analyzed the cash flows to
a pass-through security and examined the concept of a cash flow yield. We then con-
sidered two types of collateralized mortgage obligations in detail, namely sequential
pay CMOs, and CMOs with an accrual bond. The issues of extension risk and contrac-
tion risk in the context of mortgage-backed securities was studied. We examined how
to create a floating rate tranche from a sequential pay CMO and also looked at notional
interest-only tranches, as well as interest-only (IO) and principal-only (PO) strips. Fi-
nally, we examined in detail a class of securities known as planned amortization class
(PAC) bonds.
The chapter contains an appendix, in which the relationship between the SMM
and the coupon rate for the scheduled principal component is studied. We also look
at the relationship between the pre-payment for a month and the time of receipt, for
a given value of the SMM. The appendix concludes with a study of the relationship
between scheduled principal payments and prepayments, and the SMM for a given
value of the time period.

Appendix 9.1
We now derive a relationship between the SMM and the coupon rate, and the sched-
uled principal component. The scheduled principal for month t is

t−1 t−1
A
Pt = Pt ∗ ∏[1 − SMMi ] = ∏[1 − SMMi ]
i=1 (1 + r)N−t+1 i=1
t−1
Lr (1 − s)
= ×
[1 − 1
] (1 + r)N−t+1
(1+r)N
Appendix 9.1 | 255

Lr
= × (1 + r)t−1 × (1 − s)t−1
(1 + r)N − 1
Let’s call this expression Z.
Lr
× (1 + r)t−1 × (1 − s)t−1 [ln(1 + r) + ln(1 − s)]
𝜕Z
=
𝜕t (1 + r)N − 1
= Z [ln(1 + r)(1 − s)]
𝜕Z
If (1 + r)(1 − s) = 1 then =0
𝜕t
and the scheduled principal for the month is a constant.
s 𝜕Z
Thus if r = then =0
1−s 𝜕t
s 𝜕Z s 𝜕Z
However, if r > then > 0 and if r < then <0
1−s 𝜕t 1−s 𝜕t
s
Let’s analyze this result using some data. Assume s = 0.01. Thus 1−s = 0.010101.
Consider three values for r, namely 1.0101% per month, 1.125% per month, and 0.875%
per month. The results are summarized in Table 9.13.

Table 9.13: Scheduled principal and the relationship between the interest rate and the SMM.

Time r = 1.0101% r = 0.8750% r = 1.25%

1 15,519.79 15,669.54 15,256.62

2 15,519.79 15,648.58 15,292.85
3 15,519.79 15,627.65 15,329.17
4 15,519.79 15,606.75 15,365.58
5 15,519.79 15,585.88 15,402.07
6 15,519.79 15,565.03 15,438.65
7 15,519.79 15,544.21 15,475.32
8 15,519.79 15,523.42 15,512.07
9 15,519.79 15,502.66 15,548.91
10 15,519.79 15,481.93 15,585.84
11 15,519.79 15,461.22 15,622.86
12 15,519.79 15,440.54 15,659.96
13 15,519.79 15,419.89 15,697.15
14 15,519.79 15,399.26 15,734.43
15 15,519.79 15,378.67 15,771.80

Now let’s study the relationship between the prepayment and time. Given a constant
SMM, we expect the prepayment to steadily decline with time because the scheduled
outstanding balance steadily declines. The prepayment for the month is
t−1
A 1
SP t × SMMi = × [1 − ] ∏ [1 − SMMi ] × SMMi
r (1 + r)N−t i=1
256 | Chapter 9: Mortgage-Backed Securities

A 1
= SP t × s = × [1 − ] (1 − s)t−1 × s
r (1 + r)N−t
L
= N
× [(1 + r)N − (1 + r)t ] × s × (1 − s)t−1
[(1 + r) − 1]
L L
= × s × (1 + r)N × (1 − s)t−1 −
[(1 + r)N − 1] [(1 + r)N − 1]
× s × (1 + r)t (1 − s)t−1

Let’s call this Z.

L
× s × (1 + r)N × (1 − s)t−1 × ln(1 − s)
𝜕Z
=
𝜕t [(1 + r)N − 1]
L
− × s × (1 + r)t (1 − s)t−1 × ln(1 − s)
[(1 + r)N − 1]
L
− N
× s × (1 + r)t (1 − s)t−1 × ln(1 + r)
[(1 + r) − 1]
L
= × s × (1 − s)t−1 [(1 + r)N ln(1 − s) − (1 + r)t {ln(1 − s) + ln(1 + r)}]
[(1 + r)N − 1]
L
= × s × (1 − s)t−1 [(1 + r)N ln(1 − s) − (1 + r)t {ln(1 − s)(1 + r)}]
[(1 + r)N − 1]
L
N
× s × (1 − s)t−1 > 0
[(1 + r) − 1]
Consider

[(1 + r)N ln(1 − s) − (1 + r)t {ln(1 − s)(1 + r)}]

If it has to be positive

(1 + r)N ln(1 − s) > (1 + r)t {ln(1 − s)(1 + r)} ⇒ (1 + r)N−t ln(1 − s) > ln(1 − s)(1 + r)
⇒ ln(1 − s) [(1 + r)N−t − 1] > ln(1 + r)

This cannot be true, because

ln(1 − s) < 0; ln(1 + r) > 0 and [(1 + r)N−t − 1] > 0

Thus
𝜕Z
<0
𝜕t
which implies that pre-payments will steadily decline over time.
Now let’s study the relationship between the scheduled principal and the SMM
for a given value of t.
Lr
Pt = × (1 + r)t−1 × (1 − s)t−1
(1 + r)N − 1
Appendix 9.1 | 257

Lr(t − 1)
× (1 + r)t−1 × (1 − s)t−2 < 0
𝜕Pt
=−
𝜕s (1 + r)N − 1

Thus, as the SMM increases, the scheduled principal payment declines.

Then, let’s consider the relationship between the pre-payment and the SMM. The
pre-payment for a given value of t is

L L
SP t × s = × s × (1 + r)N × (1 − s)t−1 − × s × (1 + r)t (1 − s)t−1
[(1 + r)N − 1] [(1 + r)N − 1]

Let’s call this Z.

L L(t − 1)
× (1 + r)N × (1 − s)t−1 − × s × (1 + r)N × (1 − s)t−2
𝜕Z
= N
𝜕s [(1 + r) − 1] [(1 + r)N − 1]
L L(t − 1)
− N
× (1 + r)t (1 − s)t−1 + N
× s × (1 + r)t (1 − s)t−2
[(1 + r) − 1] [(1 + r) − 1]
L N t−1 s(t − 1)
= × (1 + r) × (1 − s) [1 − ]
[(1 + r)N − 1] 1−s
L s(t − 1)
− × (1 + r)t (1 − s)t−1 [1 − ]
N
[(1 + r) − 1] 1−s
L s(t − 1)
= × (1 − s)t−1 [1 − ] × [(1 + r)N − (1 + r)t ]
N
[(1 + r) − 1] 1−s
L
× (1 − s)t−1 ≥ 0
[(1 + r)N − 1]
[(1 + r)N − (1 + r)t ] ≥ 0

However, [1 − s(t−1)
1−s
] may be positive or negative. Thus the pre-payment may increase
or decrease with an increase in the SMM. Because the total principal is the sum of the
scheduled principal and the pre-payment, the total principal may increase or decrease
with an increase in the SMM.
Chapter 10
A Primer on Derivatives
To understand the nature of derivatives, we first need to consider the nature of a typical
cash or spot transaction. In a cash or a spot transaction, as soon as a deal is struck
between the buyer and the seller, the buyer has to hand over the payment for the asset
to the seller, who in turn has to transfer the rights to the asset to the buyer.
In the case of a forward or a futures contract, however, the actual transaction does
not take place when an agreement is reached between a buyer and a seller. In such
cases, at the time of negotiating the deal, the two parties merely agree on the terms at
which they will transact at a future point in time, including the price to be paid per
unit of the underlying asset. Thus, the actual transaction per se occurs only at a future
date that is decided at the outset. Consequently, unlike the case of a cash transaction,
no money changes hands when two parties enter into a forward or futures contract.
But both of them have an obligation to go ahead with the transaction on the scheduled
date. Failure to perform by either party is construed as default. Consequently, forward
and futures contracts are termed commitment contracts.

Example 10.1. Mitosa Inc. has entered into a forward contract with Hudson Bank to acquire £250,000
after 120 days at an exchange rate of $1.60 per pound. Four months hence, the company is required to
pay $400,000 to the bank and accept the pounds in lieu. The bank, as per the contract, has to accept
the dollars, and deliver the British currency.

In the case of both forward and futures contracts, there obviously has to be a buyer and
a seller. The party that agrees to buy the underlying asset in such contracts is known
as the long and is said to assume a long position. That party is also termed the buyer.
On the other hand, the counter party that agrees to sell the underlying asset as per
the contract is known as the short and is said to assume a short position. That party is
also termed as the seller. Thus the long agrees to take delivery of the underlying asset
on a future date, and the short agrees to make delivery on that date. In Example 10.1,
Mitosa is the long, and Hudson Bank is the short.

Futures and Forwards: Comparisons and Contrasts

Although both forward and futures contracts contain certain essential similarities,
they differ in certain crucial respects. Forward contracts and futures contracts are sim-
ilar in the sense that both require the long to acquire the asset on a future date and the
short to deliver the asset on that date. Both types of contracts impose an obligation on
the long as well as the short. However, there is one major difference between the two
types of contracts. Futures contracts are standardized, whereas forward contracts are
customized.

https://doi.org/10.1515/9781547400669-010
Futures and Forwards: Comparisons and Contrasts | 259

Let’s clarify the meaning of “standardization” and “customization.” In any con-

tract of this nature, certain terms and conditions need to be clearly defined. The major
terms that should be made explicit are the following:
– How many units of the underlying asset is the long required to acquire, or put
differently, how many units of the asset does the short have to deliver?
– What is the acceptable grade, or in certain cases, what are the acceptable grades
of the underlying asset that are allowable for delivery?
– Where should delivery be made? Can delivery be made only at a particular loca-
tion, or do one or both parties have a choice of locations?
– When can delivery be made? Is it possible only on a particular day, or is there a
specified period during which it can occur?

In a customized contract, the preceding terms and conditions have to be negotiated

between the buyer and the seller of the contract. Consequently, the two parties are free
to incorporate any features that they can mutually agree upon. However, in a standard-
ized contract, there is a third party that specifies the allowable terms and conditions.
The long and the short have the freedom to design a contract within the boundaries
specified by such a party. However, they cannot incorporate features other than those
that are specifically allowed. The third party in the case of futures contracts is the fu-
tures exchange. A futures exchange is similar to a stock exchange and is an arena where
trading in futures contracts takes place. We can illustrate the difference between cus-
tomization and standardization with the help of an example.

Example 10.2. Consider the wheat futures contract that is listed for trading on the Kansas City Board of
Trade. According to the terms specified by the exchange, each futures contract requires the delivery of
5,000 bushels of wheat. A bushel is a unit of measure for agricultural commodities. A bushel of wheat
is equivalent to 60 pounds or 27.216 kilograms. The allowable grades are No. 1, No. 2, and No. 3. The
allowable locations for delivery are Kansas City and Hutchinson. The specifications state that delivery
can be made at any time during the expiration month. Both the allowable grades and the permissible
delivery locations, are specified by the futures exchange.
Now take the case of Ronald Peters, a wholesale dealer, who wants to acquire 5,000 bushels of
No. 1 wheat in Kansas City during the last week of the month. Assume that there is another party, Mike
Smith, a farmer, who is interested in delivering 5,000 bushels of No. 1 wheat in Kansas City during the
last week of the month. In this case, the futures contracts that are listed on the exchange are obviously
suitable for both the parties. Consequently, if they were to meet on the floor of the exchange at the
same time, a trade could be executed for one futures contract, at a price of say $2.75 per bushel. Notice
that the price that is agreed upon for the underlying asset is one feature that is not specified by the
exchange. This has to be negotiated between the two parties who are entering into the contract and
is a function of demand and supply conditions.
Let’s now consider a slightly different scenario. Assume that Ronald wants to acquire 4,750
bushels of No. 1 quality wheat in Topeka during the last week of the month and that Mike is looking
to sell the same quantity of wheat in Topeka during that period. The terms of the contract that are
being sought by the two parties are not within the framework that has been specified by the futures
exchange in Kansas City. There are two reasons for this. First, 4,750 represents a quantity that is less
than the size of one contract, and fractional contracts cannot be bought or sold. Second, Topeka is
260 | Chapter 10: A Primer on Derivatives

not a permissible location for delivery. Consequently, neither party can enter into a futures contract
to fulfill its objectives. However, nothing prevents the two men from getting together to negotiate an
agreement that incorporates the features that they desire. Such an agreement would be a customized
agreement that is tailor made to their needs. This kind of an agreement is called a forward contract.
Thus futures contracts are exchange-traded products just like common stocks and bonds, and forward
contracts are private contracts.

One of the key issues in the case of futures contracts that permit delivery of more than
one specified grade or at multiple locations is who gets to decide where and what to
deliver. Traditionally, the right to choose the location and the grade has always been
given to the short. Also, the right to initiate the process of delivery has traditionally
been given to the short. A person with a long position, therefore, cannot demand deliv-
ery. What this also means is that, in practice, investors with a long position and no de-
sire to take delivery exit the market prior to the commencement of the delivery period
by taking an opposite or offsetting position. For, after the delivery period commences,
they can always be called upon to take delivery without having the right to refuse.

The Role of the Clearinghouse in a Futures Trade

A clearinghouse is an entity that is associated with a futures exchange. The clearing-
house essentially guarantees that neither the long nor the short, needs to worry about
the possibility of the other party defaulting. It does so by positioning itself as the effec-
tive counter-party for each of the two original parties to the trade when a futures deal is
struck. That is, the clearinghouse becomes the effective buyer for every seller, and the
effective seller for every buyer. Thus, each party to a transaction needs to worry only
about the financial strength and integrity of the clearinghouse, and not of the other
party with which it has traded. It must be remembered that neither the long nor the
short trades with the clearinghouse. The clearinghouse enters the picture only after
an agreement is reached between the two parties.
What happens in practice is the following. Assume that Ronald goes long in one
futures contract and that Mike goes short. The moment this trade is executed, it is
replaced by two equivalent trades. The first equivalent trade has Ronald going long
and the clearinghouse going short, and the second has the clearinghouse going long
and Mike going short. The clearinghouse is hence termed the central counter-party
(CCP), and this process is termed novation.
The need for a clearinghouse arises for the following reason. A futures contract
imposes an obligation on both the parties. On the expiration date of the contract, de-
pending on the movement of prices in the interim, it is in the interest of one of the two
parties to the agreement, either the long or the short, to go through with the transac-
tion. However, a price move in favor of one party would clearly translate into a loss
for the other. Consequently, given an opportunity, one of the two parties would like to
default on the expiration date. Let’s illustrate this with the help of an example.
Margins for Futures Trades | 261

Example 10.3. Consider two people, Peter and Keith. Assume that Peter has gone long in a futures
contract to buy an asset five days hence at a price of $40 and that Keith has taken the opposite side of
the transaction. Let’s first take the case where the spot price of the asset five days later is $42.50. If
Keith already has the asset, he is obliged to deliver it for $40, thereby foregoing an opportunity to sell
it in the spot market at $42.50. Otherwise, if he does not have the asset, he is required to acquire it by
paying $42.50, and then subsequently deliver it to Peter for $40. Quite obviously, Keith will choose to
default unless he has an impeccable conscience and character.
Now let’s consider a second situation where the price of the asset five days hence is $37.50. If
Keith already has the asset, he would be delighted to deliver it for $40, for the alternative is to sell it
in the spot market for $37.50. Even if he were not to have the asset, he is more than happy to acquire
it for $37.50 in the spot market and deliver it to Peter. The problem here is that Peter will refuse to pay
$40 for the asset if he can get away with it. There are two ways of looking at it. If he does not want the
asset, taking delivery at $40 would entail a subsequent sale at $37.50 and therefore a loss of $2.50.
On the other hand, even if he were to require the asset, he would be better off buying it in the spot
market for $37.50.

The purpose of having a clearinghouse is to ensure that defaults do not occur. A clear-
inghouse ensures protection for both the parties to the trade by requiring them to post
a performance bond or collateral called a margin. The amount of collateral is adjusted
daily to reflect any profit or loss for each party, as compared to the previous day, based
on the price movement during the day. By doing so, the clearinghouse effectively takes
away the incentive for a party to default as you shall shortly see.

Margins for Futures Trades

As we have just seen, whenever two parties enter into an agreement to trade at a fu-
ture date, there is always an element of default risk. In other words, there is always a
possibility that one of the parties does not carry out its part of the deal as required by
the contract.
In the case of futures contracts, compliance is ensured, by requiring both the long
and the short to deposit collateral with their broker, in an account known as the margin
account. This margin deposit is therefore a performance guarantee.
The amount of collateral is related to the potential loss that each party is likely to
incur. In the case of a futures contract, because both the parties have an obligation,
it is necessary to collect collateral from both of them. After such potential losses are
collected, the incentive to default is effectively taken away. For even if the party that
ends up on the losing side were to fail to perform its obligation, the collateral collected
from it would be adequate to take care of the interests of the other party.

Marking to Market of Futures Contracts

The reason for collecting margins is to protect both the parties against default by the
other. The potential for default, to reiterate once again, arises because a position once
262 | Chapter 10: A Primer on Derivatives

opened can and will invariably lead to a loss for one of the two parties if it complies
with the terms of the contract.
This loss, however, does not arise all of a sudden at the time of expiration of the
futures contract. As the futures price fluctuates in the market from trade to trade, one
of the two parties to an existing futures position experiences a gain, and the other
experiences a loss. Thus, the total loss or gain from the time of getting into a futures
position until the time the contract expires or is offset by taking a counter-position,
whichever happens first, is the sum of these small losses and profits corresponding to
each observed price in the interim.
The term marking to market refers to the process of calculating the loss for one
party, or equivalently the corresponding gain for the other, at specified points in time,
with reference to the futures price that prevailed at the time the contract was pre-
viously marked to market. In practice, when a futures contract is entered into, it is
marked to market for the first time at the end of the day. Subsequently, it is marked
to market every day until the position is either offset or the contract itself expires. The
party that has incurred a profit has the amount credited to its margin account, and the
other party, which has incurred an identical loss, has its margin account debited.
Let’s illustrate how profits and losses arise in the process of marking to market and
highlight the corresponding changes to the margin accounts of the respective parties.
Let’s take the case of Peter who has gone long in a futures contract, expiring five days
hence, with Keith at a futures price of $40. Assume that the price at the end of five
days is $42.50 and the prices at the end of each day prior to expiration are as depicted
in Table 10.1.

Table 10.1: End of the day futures prices.

Day Futures Price

0 40.00
1 40.50
2 39.50
3 38.00
4 40.50
5 42.50
Day 0 denotes the time the contract was entered into, and the corresponding price is the futures price
at which the deal was struck.
Day t represents the end of that particular day, and the corresponding price is the prevailing futures
price at that instant.

Let’s assume that per the contract, Peter is committed to buying 100 units of the asset,
and that at the time of entering into the contract, both the parties had to deposit $500
as collateral in their margin accounts. The amount of collateral that is deposited when
a contract is first entered into is called the initial margin.
Marking to Market of Futures Contracts | 263

At the end of the first day the futures price is $40.50. This means that the price
per unit of the underlying asset for a futures contract being entered into at the end of
the day is $40.50. If Peter were to offset the position that he had entered into in the
morning, he would have to do so by agreeing to sell 100 units at $40.50 per unit. If so,
he would earn a profit of $0.50 per unit, or $50 in all. While marking Peter’s position to
market, the broker behaves as though it were offsetting. That is, the broker calculates
Peter’s profit as $50 and credits it to his margin account. However, because Peter has
not expressed a desire to actually offset, the broker acts as if it were reentering into a
long position at the prevailing futures price of $40.50.
At the end of the second day, the prevailing futures price is $39.50. When the con-
tract is marked to market, Peter takes a loss of $100. Remember that his contract was
reestablished the previous evening at a price of $40.50, and if the broker now behaves
as if it were offsetting at $39.50, the loss is $1 per unit, or $100 in all. Once again a new
long position is automatically established, this time at a price of $39.50.
This process continues either until the delivery date, when Peter actually takes
possession of the asset, or until the day that he chooses to offset his position if that
happens earlier. As you can see from this illustration, rising futures prices lead to prof-
its for the long, whereas falling futures prices lead to losses.
Now let’s consider the situation from Keith’s perspective. At the end of the first
day, when the futures price is $40.50, marking to market means a loss of $50 for him.
That is, his earlier contract to sell at $40 is effectively offset by making him buy at
$40.50, and a new short position is established for him at $40.50. By the same logic,
at the end of the second day, his margin account is credited with a profit of $100. As
you can see, shorts lose when futures prices rise and gain when the prices fall.
Thus, the profit or loss for the long is identical to the loss or profit for the short. It
is for this reason that futures contracts are called zero sum games. One man’s gain is
another man’s loss.
As you can see, by the time the contract expires, the loss incurred by one of the two
parties, in this case the short, has been totally recovered. In our illustration, Peter’s
account has been credited with $250 by the time the contract expires. This amount
represents the difference between the terminal futures price and the initial futures
price, multiplied by the number of units of the underlying asset. These funds come
from Keith’s account which is debited. Now, if Keith were to refuse to deliver the asset
at expiration, Peter would not be at a disadvantage. For, because he has already real-
ized a profit of $250, he can take delivery in the spot market at the terminal spot price
of $42.50 per unit in lieu of taking delivery under the futures contract. Effectively, he
gets the asset at a price of $40 per unit, which is what he contracted for in the first
place.
Forward contracts, unlike futures contracts, are not marked to market. Conse-
quently, both the parties to the contract are exposed to credit risk, which is the risk that
the other party may default. Thus, in practice, the parties to a forward contract tend to
be large and well known, such as banks, financial institutions, corporate houses, and
264 | Chapter 10: A Primer on Derivatives

brokerage firms. Such parties find it easier to enter into forward contracts, compared
to individuals, because their credit worthiness is easier to appraise.
Both longs and shorts in the case of futures contracts therefore have to deposit
a performance bond, known as the initial margin, with their brokers as soon as they
enter into a futures contract. If the markets subsequently move in favor of a party to the
contract, the balance in the margin account increases, or if the market moves against
the party, the balance is depleted.
Now, the broker has to ensure that a client always has adequate funds in its margin
account. Otherwise the entire purpose of requiring clients to maintain margins is de-
feated. Consequently, the broker specifies a threshold balance called the maintenance
margin, which is less than the initial margin. If the balance in the margin account de-
clines below the level of the maintenance margin, due to adverse price movements
the client is immediately asked to deposit additional funds to take the account bal-
ance back to the level of the initial margin. In futures market parlance, we say that
the broker has issued a margin call to the client. A margin call is always bad news,
for it is an indication, that a client has suffered major losses since opening the margin
account. The additional funds deposited by a client when a margin call is complied
with are referred to as a variation margin.

The Settlement Price for Futures Contracts

The price that is used to compute the daily gains and losses for the longs and the
shorts, when the futures contracts are marked to market at the end of each day, is
called the settlement price. In many cases, futures exchanges adopt the practice of
setting the settlement price equal to the observed closing price for the day. Some ex-
changes believe that there is often heavy trading towards the close of the day and
consequently set the settlement price equal to a volume-weighted average of the ob-
served futures prices, in the last half hour or hour of trading. At the other extreme, if
there are no trades at the end of the day, the exchange may set the settlement price
equal to the average of the observed bid and ask quotes.

Movements in the Margin Account

Let’s now look at a detailed example to illustrate the concepts that we have just dis-
cussed. Let’s consider the case of Patty, who went long in a contract for 100 units of
the asset at a price of $60 per unit, and deposited $2,000 as collateral for the same.
Let’s assume that the broker fixes a maintenance margin of $1,250 for the contract, the
contract lasts for a period of five days, and the futures prices on the subsequent days
are as shown in the second column of Table 10.2. The effect on the margin account is
summarized in the same table.
Offsetting of Futures Contracts | 265

Table 10.2: Changes in the margin account over the course of time.

Day Futures Daily Cumulative Account Margin

Price Gain/Loss Gain/Loss Balance Call

0 60.00 2,000
1 62.50 250 250 2,250
2 56.50 (600) (350) 1,650
3 50.00 (650) (1,000) 1,000 1,000
4 58.50 850 (150) 2,850
5 62.50 400 250 3,250
Numbers in parentheses denote losses.

Let’s analyze in detail a few of the entries in Table 10.2 to illustrate the concepts. Con-
sider the second row. As compared to the time the contract was entered into, the price
has increased by $2.50 per unit or $250 for 100 units. Thus, Patty, who entered into a
long position, has gained $250, which is required to be credited to her margin account.
After this amount is credited, the margin account, which had an opening balance of
$2,000, has an end-of-the day balance of $2,250.
The settlement price at the end of the second day is $56.50. Thus, as compared to
the position at the end of the previous day, Patty has suffered a loss of $6 per unit or
$600 for 100 units. When this loss is debited to her margin account, the balance in the
account falls to $1,650. Since this is above the maintenance level of $1,250, there is no
margin call. The settlement price at the end of the next day is $50, which implies that
Patty has suffered a further loss of $650. When this loss is factored into the margin
account by debiting it with $650, the balance in the account falls to $1,000, which is
less than the maintenance margin requirement of $1,250. Therefore, at this point in
time, a margin call will be issued for $1,000, which is the amount required to take
the balance back to the initial margin level of $2,000. In response to the call, Patty is
expected to pay a variation margin of $1,000. The accuracy of the computation may
be viewed as follows. She initially deposits $2,000 and then subsequently deposits
another $1,000. Her cumulative gain/loss is $250. Hence the terminal balance in the
account is 2,000 + 1,000 + 250 = $3,250, which is what we observe.

Offsetting of Futures Contracts

Futures contracts are easier to offset than forward contracts. Offsetting essentially
means taking a counterposition. It means that if a party has originally gone long, it
should subsequently go short and vice versa. The effect of offsetting is to cancel an
existing long or short position in a contract. Remember, that a forward contract is a
customized private contract between two parties. Thus, if a party to a forward contract
266 | Chapter 10: A Primer on Derivatives

wants to cancel the original agreement, it must seek out the counterparty with which
it had entered into a deal and have the agreement canceled.
However, canceling a futures contract is a lot simpler. For example, a futures con-
tract between two parties, say Jacob and Victor, to transact in wheat at the end of a
particular month, will be identical to a similar contract between two other parties,
say Kimberly and Patricia. This is because both the contracts would have been de-
signed according to the features specified by the exchange. So if Jacob, who entered
into a long position, wants to get out of his position, he need not seek out Victor, the
party with whom he had originally traded. All he has to do is to go back to the floor
of the exchange and offer to take a short position in a similar contract. This time the
opposite position may be taken by a new party, say Robert. Thus, by taking a long po-
sition initially with Victor, and a short position subsequently with Robert, Jacob can
ensure that he is effectively out of the market and has no further obligations. This is
the meaning of offsetting.
The profit or loss for an investor who takes a position in a futures contract and sub-
sequently offsets it is equal to the difference between the futures price that prevailed
at the time the original position was taken and the price at the time the position is
offset. For a long, the profit is FT − F0 , where F0 is the initial futures price and FT is
the terminal futures price. Thus, if the futures price increases, the long has a profit;
otherwise, he has a loss. For the short, the profit is F0 − FT . Hence a price decline leads
to a profit, whereas a price rise amounts to a loss.

Spot-Futures Convergence of Prices

At the time of expiration of the futures contract, the futures price must be the same as
the cash or spot market price. For, after all, what is a futures contract? It is a contract
to transact at a future point in time. At the expiration date of the contract, any futures
contract must lead to an immediate transaction because the contract is scheduled to
expire immediately and hence is valid only for an instant. Thus a person who enters
into a futures contract at the time of expiration is effectively entering into a spot market
transaction. Consequently, if the futures price at expiration is different from the spot
price, there will be arbitrage opportunities.
Let’s denote the futures price at expiration by FT and the spot price at that point
in time by ST . It must be the case that FT = ST . Let’s examine the consequences of FT
being greater than ST , or FT being less than ST .
– FT > ST
This situation can be exploited by an arbitrageur as follows. The investor can ac-
quire the asset in the spot market at a price of ST and simultaneously go short in
a futures contract. Because the contract is scheduled to expire immediately, the
investor can at once deliver for a price of FT . Thus FT − ST , which by assumption
is positive, represents an arbitrage profit for such an individual.
Delivery in the Case of Futures Contracts | 267

Example 10.4. Assume that the futures price of an asset at the time of expiration is $425, whereas the
spot price is $422. An arbitrageur immediately acquires 100,000 units of the underlying asset in the
spot market at $422 per unit and simultaneously goes short in a futures contract. Because the contract
is expiring he immediately delivers at $425, thereby making a costless riskless profit of $300,000.

– FT < ST
An arbitrageur exploits this condition by going long in a futures contract. Because
it is about to expire, the investor can take immediate delivery by paying FT and
then sell the asset in the spot market for ST . In this case, ST − FT , which by as-
sumption is positive, represents an arbitrage profit.

Example 10.5. Assume that the futures price of an asset at the time of expiration is $422, whereas
the spot price is $425. An arbitrageur takes a long position in a futures contract, which entails taking
immediate delivery at $422 per unit. The asset can then be immediately sold in the spot market for
$425 per unit. Thus, once again, the arbitrageur is able to lock in a costless, riskless profit of $300,000
for a transaction involving 100,000 units.

Delivery in the Case of Futures Contracts

Although both forward and futures contracts call for delivery at the time of expira-
tion, there are fundamental differences between them in this context. First, in prac-
tice, most forward contracts are settled by delivery. However, only a small fraction of
the futures contracts that are entered into (in some markets the figure is as low as
2%), result in actual delivery. The remainder are offset prior to expiration by taking a
counterposition.
Second, a forward contract is a customized agreement between two parties, so
unless it is canceled, it usually results in the short delivering to the original party who
had gone long. In the case of futures, however, the link between the long and the short
is broken by the clearinghouse, once a contract is entered into. In a futures trade,
one or both parties may offset and exit the market. Hence, when a short expresses
the desire to deliver, it is not necessarily the original long that takes delivery of the
asset, for that party may no longer have an open long position. Thus, in the case of
futures contracts, the exchange decides to whom the short should deliver. In practice,
the person with the oldest outstanding long position is usually called upon to take
delivery.
An OTC derivative such as a forward contract can also be assigned to another
party. The process can be tricky and can vary. Usually the non-broker party finds a
Wall Street firm to “step in” as it steps out. Brokers are generally happy to face each
other so they would not generally see this as increasing counterparty risk. All parties
typically need to agree on terms and price, the new broker who is stepping in and the
268 | Chapter 10: A Primer on Derivatives

original non-broker party. The original broker and the incoming Wall Street firm now
face each other with respect to counterparty credit risk.
Finally, under a forward contract, the price that is paid by the long at the time of
delivery, is different from the amount paid to take delivery under a futures contract
with the same features and on the same underlying asset. A forward contract, it must
be remembered, is not marked to market at intermediate points in time. Consequently,
at expiration, the long has to pay the price that was agreed upon at the outset, to take
delivery. However, in the case of a futures contract, the contract is marked to market
on every business day during its lifetime. Hence, in order to ensure that the long gets
to acquire the asset at the price that is agreed upon at the outset, he is asked to pay
the prevailing futures price at expiration, which as you have seen earlier, is the same
as the prevailing spot price at expiration. We can illustrate the above arguments using
symbols, and with the help of a numerical example.
Consider a futures contract that was entered into on day 0 at a price F0 and expires
on day T. We denote the price at expiration by FT . Such a contract is marked to market
on days 1, 2, 3... up to day T. The cumulative profit for the long due to marking to
market is

(FT − FT−1 ) + (FT−1 − FT−2 ) + (FT−2 − FT−3 ) + .... + (F2 − F1 ) + (F1 − F0 ) = (FT − F0 )

To acquire the asset at the original price of F0 , the long must be asked to pay a price P
at the time of delivery, such that

P − (FT − F0 ) = F0
⇒ P = FT = ST

Thus the price paid by the long at the time of delivery must equal the prevailing
futures price at expiration, or equivalently, the prevailing spot price at expiration. In
the case of a forward contract, however, is no marking to market and, hence, no inter-
mediate cash flows. Consequently, at the time of delivery, the price P paid by the long,
must be the same as the price that was agreed upon originally. That is

P = F0

Example 10.6. Let’s now illustrate these arguments using a numerical example. Consider a futures
contract on wheat that was entered into at a price of $2.50 per bushel. Assume that the contract lasts
for a period of 5 days and the movement in the futures price on subsequent days is as depicted in
Table 10.3.
In this case, F0 = 2.5 and FT = 3.5. A person who went long in a futures contract at a time when
the futures price was $2.50, has to pay $3.50 at the time of delivery. Taking into account the profit
of $1.00 due to marking to market, the long effectively gets the asset for $2.50, which is nothing but
the initial futures price. On the other hand, a person who went long in a forward contract at a price of
$2.50, has to pay $2.50 at the time of taking delivery.
Cash Settlement of Futures Contracts | 269

Table 10.3: Marking a contract to market.

Day Futures Profit from

Price Marking to Market

0 2.5
1 2.4 (0.10)
2 2.2 (0.20)
3 2.5 0.30
4 2.8 0.30
5 3.5 0.70
Total 1.00

Cash Settlement of Futures Contracts

There are certain futures contracts that do not allow for physical delivery of the un-
derlying asset. In such cases, the contract is marked to market until the last day, and
subsequently all positions are declared closed. Under such circumstances, both the
long and the short exit the market, with their cumulative profit (which could also be
a loss) since the inception of the contract. However, the short does not deliver the un-
derlying asset at the end.

Valuation of Futures and Forwards

A forward contract entails an obligation on the part of the short to make delivery of
the asset on a future date, and an equivalent obligation on the part of the long to take
delivery.
From the perspective of the short, if the difference between the forward price and
the prevailing spot price exceeds the cost of carrying the asset until delivery, then
clearly there is an arbitrage opportunity. For instance, in the case of an asset that pays
no income before the maturity of the forward contract, the cost of carrying the asset
is rSt , where r is the riskless rate of interest and St is the prevailing spot price. Conse-
quently, if

Ft − St > rSt

then a person could exploit the situation by borrowing and buying the asset, and si-
multaneously going short in a forward contract to deliver on a future date.
Such an arbitrage strategy is called cash-and-carry arbitrage. Hence, to rule it out,
we require that

Ft − St ≤ rSt ⇒ Ft ≤ St (1 + r)
270 | Chapter 10: A Primer on Derivatives

Example 10.7. Cash-and-carry arbitrage can be illustrated with the help of an example.
Assume that IBM is currently selling for $100 per share, and is not expected to pay any dividends
for the next six months. The price of a forward contract for one share of IBM to be delivered after six
months is $106.
Consider the case of an investor who can borrow funds at the rate of 5% per six-month period.
Such an individual can borrow $100 and acquire one share of IBM, and simultaneously go short in a
forward contract to deliver the share after six months for $106. Thus the rate of return on the invest-
ment is

(106 − 100)
= 0.06 ≡ 6%
100

whereas the borrowing cost is only 5%.

Consequently, cash-and-carry arbitrage is a profitable proposition under such circumstances.
This is because

Ft > St (1 + r)

or, in other words, the forward contract is overpriced.

The rate of return obtained from a cash-and-carry strategy is called the implied repo
rate. Thus, a cash-and-carry strategy is profitable if the implied repo rate exceeds the
borrowing rate. The two statements mean the same. That is, if cash-and-carry arbi-
trage is possible, the futures contract is overpriced and the implied repo rate is higher
than the borrowing rate. Conversely, if the implied repo rate is higher than the borrow-
ing rate, then cash-and-carry arbitrage is profitable, for the contract is overpriced.
By engaging in a cash-and-carry strategy, the investor has ensured a payoff of $106
after six months for an initial investment of $100. It is as if the investor has bought a
zero coupon debt instrument with a face value of $106 for a price of $100. Hence, a
combination of a long position in the stock and a short position in a forward contract is
equivalent to a long position in a zero coupon bond. Such a deep discount instrument
is referred to as a synthetic T-bill. Hence we can express the relationship as

Spot-Forward = Synthetic T-bill.

The implication is that, while a long spot position, as well as a short forward position,
are risky in isolation, their combination is risk-less.
A negative sign indicates a short position in that particular asset. Thus, if we own
any two of the three assets, we can artificially create the third.
Cash and carry arbitrage requires a short position in a forward contract and arises
if the contract is overpriced. However, if Ft were to be less than St (1 + r), then such a
situation too would represent an arbitrage opportunity, this time for the long. Under
such circumstances, an investor could short sell the asset and invest the proceeds at
the risk-less rate, and simultaneously go long in a forward contract to reacquire the
asset at a future date.
Valuation of Futures and Forwards | 271

This kind of an arbitrage strategy is called reverse cash-and-carry arbitrage. In

order to rule out such profit opportunities, we require that

Ft ≥ St (1 + r)

Example 10.8. We can illustrate reverse cash-and-carry arbitrage with the help of a numerical exam-
ple.
Assume once again that IBM is selling for $100 per share and the company is not expected to pay
any dividends for the next six months. Let the price of a forward contract for one share of IBM to be
delivered after six months be $104.
Consider the case of an arbitrageur who can lend money at the rate of 5% per six monthly period.
Such an individual can short sell a share of IBM and invest the proceeds at 5% interest for six months,
and simultaneously go long in a forward contract to acquire the share after six months for $104. We
are assuming that the arbitrageur can lend the proceeds from the short sale. In practice, the amount
has to be deposited with the broker, who, of course, can invest it to earn interest. In a competitive
market, brokers may pass on a part of the interest income to the client who is short selling. This is
called a short interest rebate. However, the effective rate of return earned by the short-seller is lower
than the prevailing market rate. In practice, although institutional traders have the clout to demand a
short interest rebate, individual traders do not.
The effective borrowing cost is

(104 − 100)
= 0.04 ≡ 4%
100

which is less than the lending rate of 5%.

Consequently there is a profit to be made by employing a reverse cash-and-carry strategy under
such circumstances. This is because

Ft < St (1 + r)

or, in other words, the forward contract is underpriced.

The cost of borrowing funds under a reverse cash-and-carry strategy is called the im-
plied reverse repo rate. Thus reverse cash-and-carry arbitrage is profitable only if the
implied reverse repo rate is less than the lending rate. Once again, if the implied re-
verse repo rate is less than the lending rate, then reverse cash-and-carry arbitrage is
profitable, for the contract is underpriced. Equivalently, if the futures contract is un-
derpriced, it means that the implied repo rate is less than the lending rate, and con-
sequently, reverse cash-and-carry arbitrage is profitable.
Cash-and-carry arbitrage is ruled out if Ft ≤ St (1 + r), whereas reverse cash-and-
carry arbitrage is ruled out if Ft ≥ St (1 + r). Thus, in order to rule out both forms of
arbitrage, we require that

Ft = St (1 + r) (10.1)

This is termed the cost of carry relationship.

272 | Chapter 10: A Primer on Derivatives

The Case of Assets Making Payouts

If a person who is holding an asset receives income from it, then such an inflow obvi-
ously reduces the carrying cost. The carrying cost can now be defined as rSt −I where I
is the future value of the income as calculated at the time of expiration of the forward
contract.1 Consequently, in order to rule out cash-and-carry arbitrage, we require that

Ft − St ≤ rSt − I ⇒ Ft ≤ St (1 + r) − I

Similarly, from a short-seller’s perspective, the effective income obtained by investing

the proceeds from the short sale are reduced by the amount of payouts from the as-
set because the short-seller is required to compensate the lender of the asset for the
payouts.
Thus to rule out reverse cash-and-carry arbitrage, we require that

Ft − St ≥ rSt − I ⇒ Ft ≥ St (1 + r) − I

Therefore, to preclude both forms of arbitrage, it must be the case that

Ft = St (1 + r) − I (10.2)

Now let’s illustrate cash-and-carry arbitrage in the case of assets making payouts
with the help of a numerical example. The extension to reverse cash-and-carry arbi-
trage is straightforward.

Example 10.9. Let’s go back to the case of the IBM share. Assume that the share is selling for $100
and the stock is expected to pay a dividend of $5 after three months and another $5 after six months.
Forward contracts with a time to expiration of six months are available at a price of $96 per share.
Let’s assume that the second dividend payment will occur just an instant before the forward contract
matures.
Consider the case of an investor who can borrow at the rate of 10% per annum. Such an individual
can borrow $100, buy a share of IBM, and simultaneously go short in a forward contract to sell the share
after six months for $96. After three months, the investor gets a dividend of $5 which can be invested
for the remaining three months at a rate of 10% per annum. And finally, just prior to delivering the
share under the forward contract, the individual receives a second dividend of $5.
Thus, at the time of delivery of the share, the total cash inflow for the investor is
0.10
96 + 5 × [1 + ] + 5 = $106.125
4
Hence the rate of return on the synthetic T-bill is
(106.125 − 100)
= 0.06125 ≡ 6.125%
100
which is greater than the borrowing rate of 5% for six months.

1 The reason why we take the future value of income, is because the interest cost is computed at the
point of expiration of the futures contract. So in order to be consistent with the principles of the time
value of money, the income too should be computed at the same point in time.
Conversion Factors When There Are Multiple Deliverable Grades | 273

Consequently, cash-and-carry arbitrage is profitable. This is because

Ft + I > St (1 + r)

where I is the future value of the payouts from the asset as calculated at the point of expiration of the
forward contract.

Conversion Factors When There Are Multiple Deliverable Grades

Certain futures contracts give the short the flexibility to deliver more than one grade of
the underlying asset. Thus the short can choose the grade to deliver. The exchange in
such cases designates one grade as the par grade. If the short delivers the par grade, he
will receive the prevailing futures price at expiration, FT , for reasons that we have al-
ready explained. However, delivering a more valuable grade results in a premium, and
when delivering a less valuable grade, the short does so at a discount. There are two
ways of incorporating the premium or discount for grades other than the par grade:
multiplicative adjustment and additive adjustment. Let’s examine each of these tech-
niques.

Multiplicative Adjustment of the Futures Price

In the case of contracts that permit more than one grade of the underlying asset to be
delivered and specify a multiplicative system of price adjustment, the short receives
an amount equal to ai FT at the time of delivery of grade i. For premium grades, ai is
greater than 1.0, whereas for discount grades, it is less than 1.0.
We denote the spot price of grade i at expiration by Si,T . Hence, the profit for the
short when delivering grade i is

ai FT − Si,T (10.3)

Grade i is preferred to another grade j if

ai FT − Si,T > aj FT − Sj,T (10.4)

At expiration, in order to preclude arbitrage, the profit from delivering the most pre-
ferred grade must be zero. If we denote this grade as grade i, it must be the case that

ai FT − Si,T = 0
Si,T
⇒ FT = (10.5)
ai
274 | Chapter 10: A Primer on Derivatives

In the case of all other grades, it must be the case that

aj FT − Sj,T < 0
Sj,T
⇒ FT < ∀j (10.6)
aj

Thus, the grade chosen for delivery obviously will be the one for which aS is the low-
est. Such a grade is called the cheapest to deliver grade, and aS is called the delivery
adjusted spot price. Thus, the cheapest to deliver grade, is the one with the lowest de-
livery adjusted spot price. At expiration therefore, the futures price must converge to
the delivery-adjusted spot price of the cheapest to deliver grade.
Gold futures contracts allow for the delivery of gold within a certain weight range
and with varying degrees of fineness.2 The price received by the short = weight × fine-
ness × futures price. Thus, per ounce of gold delivered, the short receives fineness ×F.
The conversion factor in this case is the fineness, and the par grade is 100% fine.
Assume that gold is available with either 99% fineness or with 100% fineness. Let
the spot price of 99% fine gold be $495 per ounce and that of 100% fine gold be $505
495
per ounce. The delivery adjusted spot prices are 0.99 = 500, and 505
1
= 505 respectively.
Thus the 99% fine gold is the cheapest to deliver grade.

Additive Adjustment of the Futures Price

Now let’s consider the case of contracts that permit more than one grade to be deliv-
ered but use an additive system of price adjustment. In the case of such contracts, the
short receives FT + ai for the delivery of grade i. For a premium grade, ai is positive,
whereas for a discount grade, it is negative.
The profit from delivering grade i is

FT + ai − Si,T (10.7)

and grade i is preferred to another grade j if

FT + ai − Si,T > FT + aj − Sj,T

⇒ Si,T − ai < Sj,T − aj (10.8)

Hence, the cheapest to deliver grade is the one for which S − a is the lowest. That is,
grade i is the cheapest to deliver grade if

Si,T − ai < Sj,T − aj ∀j (10.9)

2 See Siegel and Siegel [55].

Hedging Using Futures Contracts | 275

To rule out arbitrage, the profit from delivering the cheapest to deliver grade must be
zero. That is

FT + ai − Si,T = 0
⇒ FT = Si,T − ai (10.10)

In this case S − a is the delivery-adjusted spot price, and once again, the futures price
converges to the delivery-adjusted spot price of the cheapest to deliver grade.
It must be noted that irrespective of whether the multiplicative or the additive sys-
tem is used, the cheapest to deliver grade need not be the one with the lowest spot
price. For example, consider the following data for corn.

Table 10.4: An illustration of additive price adjustment.

Grade Spot Price Conversion Delivery-Adjusted

Factor Spot Price

No. 1 6.51 0.015 6.495

No. 2 6.50 0 6.500
No. 3 6.49 −0.015 6.505

The par grade is obviously No. 2. But the cheapest to deliver grade is No. 1, which in-
cidentally has the highest spot price.

Hedging Using Futures Contracts

A hedger, by definition, is a person who wants to protect himself against an unfavor-
able movement in the price of the underlying asset. Quite obviously, a person who
seeks to hedge has already assumed a position in the underlying asset. If the hedger
already owns the asset, or in other words has a long position, the worry is that the price
of the asset may fall subsequently. On the other hand, if the hedger has made a com-
mitment to buy, or in other words has taken a short position, the worry is that the price
of the asset may rise subsequently. In either case, the reason to hedge is to avoid risk.
Notice that we have defined a short position in a broader sense then in the case of
a short sale involving an asset. In the case of a short sale, the short-seller borrows an
asset from a broker in order to sell and therefore has a commitment to buy it back at a
future date and return it. Now take the case of a German company that has imported
goods from the U.S and been given 90 days credit by the American supplier. The com-
pany therefore has a commitment to buy dollars after 90 days. Just like in the case
of the short-seller, this company also worries that the price of the asset, in this case
the U.S. dollar, may go up by the time it is procured. Thus in a more general sense, a
short position connotes a commitment to buy an asset at a future date. An example in
276 | Chapter 10: A Primer on Derivatives

the case of physical commodities would be the case of a wheat mill that knows it will
have to procure wheat after the harvest, which we can assume is one month away. Its
worry consequently would be that the harvest may be less plentiful than anticipated,
and therefore the price of the wheat in the spot market may turn out to be higher than
what is currently expected. Such an entity also may exhibit a desire to hedge.
Futures contracts can help people to hedge, irrespective of whether they have a
long or a short position in the underlying asset. Consider a person who owns an asset.
Such a person can hedge by taking a short position in a futures contract. If the price
of the underlying asset falls subsequently, that person can still sell at the original fu-
tures price because the other party is under an obligation to buy at this price. We can
illustrate this with the help of an example.

Example 10.10. Greg owns 50,000 shares of IBM. His worry is that the spot price of the shares may
decline substantially during the next three months.
Futures contracts on IBM are available with a time to maturity of three months, and each con-
tract is for 100 shares. If the current futures price is $75 per share, then by going short in 500 futures
contracts, Greg can ensure that he can sell the shares three months hence for $3,750,000.
This amount of $3,750,000 is guaranteed irrespective of what the actual spot price at the end
of three months turns out to be. Thus, by locking in this amount, Greg can protect himself against a
decline in the price below the contracted value of $75 per share. However, the flip side is that he is
unable to benefit if the spot price at expiration turns out to be greater than $75, for he is obliged to
deliver at this price.

Now consider the issue from the perspective of a person, Tom, who has a short position
in the underlying asset. His worry is that the price may rise by the time he acquires it in
the cash market. He can hedge by taking a long position in the futures market. For, if
the spot price rises subsequently, he can still buy at the original futures price because
the other party is under an obligation to sell at this price.

Example 10.11. Vincent has imported goods worth £312,500 from London, and is required to pay after
one month. He is worried the dollar may depreciate by then, or in other words, that the dollar price of
the pound may go up. A futures contract expiring after one month is available, and the contract size is
62,500 pounds. Let the futures price be $1.75 per pound.
So if Vincent takes a long position in five futures contracts, he can lock in a dollar value of
$546,875 for the pounds. Once again this is true irrespective of the spot exchange rate one month
later. So although Vincent can protect himself against a depreciating dollar, or in other words an ex-
change rate greater than $1.75 per pound, he is precluded from taking advantage of an appreciating
dollar, which would manifest itself as an exchange rate below $1.75 per pound.

Hedging and Ex-Post Regret

A hedger who uses futures contracts may end up regretting a decision ex-post because
the result would have been better without the hedge. Take the case of Vincent who
Hedging and the Case of Cash-Settled Contracts | 277

went long in futures to acquire £312,500 at $1.75 per pound. If the spot exchange rate
at the end turns out to be greater than $1.75 per pound, or in other words, if the dol-
lar depreciates, then his decision to hedge would certainly be perceived as wise and
sound.
But, if the spot exchange rate at expiration were to be less than $1.75 per pound, or
in other words if the dollar appreciates, Vincent would end up looking a little foolish.
For he could have bought the pounds at a lower cost in terms of dollars had he decided
not to hedge.
The problem is that, a priori, Vincent cannot be expected to be certain as to
whether the dollar would depreciate or appreciate. So if he is a risk-averse individual,
then he may very well decide to hedge his risk, notwithstanding the possibility that
he may end up looking silly ex-post.
Thus, an investor will hedge when uncomfortable leaving open exposure to price
risk, where price risk refers to the risk that the spot price of the asset may end up mov-
ing in an adverse direction from the investor’s perspective. There is no guarantee that
ex-post the decision will be vindicated. Hindsight as they say is a perfect science. A
normal individual cannot be expected to be prescient, and a prescient investor cer-
tainly would not need derivatives in order to hedge.3

Hedging and the Case of Cash-Settled Contracts

The effective price at which to sell if a hedger takes a short futures position, or buy in
the case of a long futures position, is equal to the futures price prevailing at the time
of initiation of the futures position, irrespective of whether the contract is delivery
settled or cash settled.
Let’s first take the case of a hedger who goes short in futures contracts. If the con-
tract is cash settled, the total profit from marking to market is F0 − FT . The hedger has
to sell the asset in the spot market at a price of ST , and the overall cash inflow will be

ST + (F0 − FT ) = F0

because by the no-arbitrage condition, FT = ST . Notice that the profit from the futures
position has to be added to determine the effective inflow. If it is a positive number,
that is, it actually is a profit, then it leads to a higher effective inflow. If it is a negative
number, that is, it is a loss, then it leads to a lower effective inflow.

Example 10.12. Let’s take the case of Greg who went short in 500 futures contracts at a price of $75
per share, where each contract is for 100 shares. Assume that the futures price or equivalently the
spot price at expiration is $78.50 per share.

3 See Parameswaran and Hegde [36].

278 | Chapter 10: A Primer on Derivatives

Greg’s cumulative profit from marking to market is

500 × 100 × (75.00 − 78.50) = −$175,000

He can sell the shares in the spot market for $78.50 per share or $3,925,000 in all. The effective
amount received for 50,000 of shares is

$ 3,925,000 − $175,000 = $3,750,000

which amounts to $75 per share.

The same is true for a hedger who goes long in futures. If the contract is cash settled,
the cumulative profit from marking to market is FT − F0 . The hedger then has to buy
the asset in the spot market by paying ST . The overall cash outflow is4

−ST + (FT − F0 ) = −F0

Notice that the profit from the futures position is added to the outflow in the spot mar-
ket to determine the effective outflow. Thus a profit, or inflow from the futures market,
reduces the effective outflow, whereas a loss, or outflow from the futures market, in-
creases the effective outflow.

Example 10.13. Take the case of Vincent who went long in a futures contract at a price of $1.75 per
pound. Assume that the terminal spot or equivalently futures price is $1.72 per dollar. His cumulative
profit from marking to market is

312,500 × (1.72 − 1.75) = −$9,375

He can acquire £312,500 in the spot market by paying $537,500. His effective outflow is therefore
$546,875 which translates into an exchange rate of $1.75 per pound, which is the initial futures price.

Perfect Hedges Using Futures Contracts

A hedge using futures contracts that can lock in, with absolute certainty, a selling price
for an investor with a long position in the spot market, or a buying price for an investor
with a short position in the spot market, is called a perfect hedge. Both the examples
given earlier are illustrations of a perfect hedge. In the first case, Greg was assured of a
selling price of $75 per share, and in the second case, Vincent was assured of a buying
price of $1.75 per pound.
If we are to ensure that the hedge is perfect, then the following conditions must
necessarily hold:

4 The negative sign denotes an outflow.

Perfect Hedges Using Futures Contracts | 279

1. Futures contracts must be available on the commodity that the hedger is seeking
to buy or to sell. If this is not the case, then one has to use a contract on a related
commodity. This is termed cross-hedging. The criterion for a close relationship, is
that the prices of the two should be highly positively correlated.
2. The date on which the hedger wishes to buy or sell the underlying asset must
coincide with the date on which the futures contract being used is scheduled to
expire.
3. The number of units of the underlying asset, which is sought to be bought or sold
by the hedger, must be an integer multiple of the size of the futures contract.

The Importance of Terminating the Hedge on the Expiration Date

It important that the date on which the asset is bought or sold be the same as the
expiration date of the futures contract, to ensure that the hedge is perfect.
Take the case of a hedger who has gone short in futures contracts. We can denote
the initial futures price by F0 and the terminal spot and futures prices by ST and FT ,
respectively. Now assume that the hedger wishes to sell the goods in his possession on
day t ∗ , where 0 < t ∗ < T. That is, the hedger wishes to sell the goods prior to the date
of expiration of the futures contract. Doing so requires offsetting the futures position
on that day by taking a counterposition.
The cumulative profit from the futures market due to marking to market is F0 −Ft ∗ .
The proceeds from the sale of the good in the spot market are St ∗ . Thus the effective
sale proceeds are

St ∗ + (F0 − Ft ∗ ) = F0 + (St ∗ − Ft ∗ )

Now if t ∗ were the same as T, then St ∗ would be equal to Ft ∗ and the effective price
received would be F0 . In other words, the hedge would be perfect, for the initial fu-
tures price would have been locked in. But, in general, when t ∗ is a date prior to the
expiration of the futures contract, Ft ∗ will not be equal to St ∗ . Thus, the effective price
received ultimately depends on both Ft ∗ and St ∗ . Because these are unknown vari-
ables until the end, there is always uncertainty regarding the effective price that will
ultimately be received by the hedger.
A similar argument can be advanced for the case where the hedger takes a long
position in a futures contract. If the asset is bought on day t ∗ and the futures position
taken earlier is offset, the effective outflow on account of the asset is

−St ∗ + (Ft ∗ − F0 ) = −F0 − (St ∗ − Ft ∗ ) (10.11)

Once again there will be uncertainty about the effective price at which the asset will
be bought.
280 | Chapter 10: A Primer on Derivatives

Thus, because spot-futures convergence is assured only at the time of expiration

of the futures contract, it is necessary that the date on which the transaction in the
underlying asset takes place be the same as the expiration date of the futures contract,
if we are to ensure that the hedge is perfect.

The Importance of Hedging an Integer Multiple of the Contract Size

It is essential that the number of units that the hedger is seeking to buy or sell be an
integer multiple of the size of the futures contract, if we are to ensure that the hedge
is perfect.
Assume that a farmer has 1,050 units of a commodity that he wishes to sell after
one month. Futures contracts on the commodity expiring after one month are avail-
able, and each contract is for 100 units. So this farmer theoretically needs to go short
in 10.5 futures contracts, but in practice, must go short in 10 contracts or in 11. In either
case, the effective price received per unit is uncertain.

Case-A: The farmer uses 10 contracts

The profit from marking to market is
10 × 100 × (F0 − FT )
The proceeds when the goods are sold in the spot market are 1,050ST . The effective
price received per unit of the good is
1,050ST + 1,000(F0 − FT )
1,050
50ST + 1,000F0
=
1,050
Since ST is a random variable whose value is known only at the end, the effective price
that will be received per unit is uncertain.

Case-B: The farmer uses 11 contracts

The profit from marking to market is
11 × 100 × (F0 − FT )
The proceeds from the sale of goods are 1,050ST . The effective price received per unit
of the good is
1,050ST + 1,100(F0 − FT )
1,050
1,100F0 − 50ST
=
1,050
Once again there is uncertainty regarding the effective price that will be received.
Choosing an Expiration Month for Hedging | 281

Choosing an Expiration Month for Hedging

Let’s assume that futures contracts are available on the commodity whose price we
want to hedge and that the transaction size is an integer multiple of the size of the
futures contract. However, the scheduled transaction date rarely coincides with the
expiration date of the futures contract. The only option under such circumstances is
to choose a futures contract that expires after the date on which we want to transact.
For, if we use a futures contract that expires earlier, then subsequently we will have
an open or uncovered position, which means that we will not be hedged or protected
until the transaction date.
If we use a contract that expires after the transaction date, then such a contract
has to be offset on the transaction date. For a short hedger, the effective price received
under such circumstances is

F0 + (St ∗ − Ft ∗ )

The uncertainty in this case arises because St ∗ need not equal Ft ∗ . The same holds true
for a hedger who takes a long position in futures contracts.
At any point in time t, St − Ft is called the basis and is denoted by bt . At the time of
expiration of the futures contract, ST is guaranteed to equal FT , and we can be sure that
the basis will be zero. However, prior to expiration we cannot make such an assertion.
Consequently, a hedger who is forced to offset the futures contract prior to expiration
faces uncertainty regarding the basis, or what is called basis risk.
We have defined the basis as St −Ft .5 For a short hedger, the effective price received
is F0 + bt . Thus the higher the value of the basis the better it is for the hedger, who
benefits from a rising basis. However, for a long hedger, the effective outflow is F0 + bt .
Thus the lower the value of the basis, the smaller the outflow is. We know that a person
with a long futures position benefits from a rising futures price, whereas an investor
with a short futures position benefits from a declining futures price. The basis is a
synthetic price, for it is the difference of two prices. We have seen that a rising basis
benefits the shorts, and a falling basis benefits the longs. Consequently we can say
that a short hedger is long the basis, and a long hedger is short the basis.

Speculation Using Futures Contracts

Unlike a hedger, whose objective is to avoid risk, a speculator is a person who con-
sciously seeks to take risk, hoping to profit from subsequent price movements. Such a
person is either betting that the price will rise, and he is said to be bullish, or is hoping
that it will fall, and he is said to be bearish.

5 Some authors prefer to define it as the futures price minus the spot price.
282 | Chapter 10: A Primer on Derivatives

From the standpoint of finance theory, speculation and gambling are two different
phenomena. A speculator is a person who evaluates the risk of an investment and the
anticipated return from it prior to committing. Such a trader therefore takes a position
only if he feels that the anticipated return is adequate considering the risk being taken.
In other words, the trader may be said to be taking a calculated risk.
A gambler on the other hand is someone who takes a risk purely for the thrill of
taking it. The expected return from the strategy is of no consequence for such a person
when taking a decision to gamble.
Active speculation adds depth to a market and makes it more liquid. A market
characterized solely by hedgers does not have the kind of volumes required to make it
efficient. In practice, when a hedger seeks to take a position, very often the opposite
side of the transaction is taken by a speculator, although this need not always be the
case. There could be situations where both traders are speculators and the long is a
bull who is anticipating a price rise, and the short is a bear who believes that prices
are likely to decline. It is also possible that both traders are hedgers. The long may be a
consumer who is worried about rising prices, and the short may be a producer, who is
perturbed by the specter of declining prices. Divergence of views, and a desire to take
positions based on those views is a sin qua non for making the free market system a
success. Thus speculators, along with hedgers and arbitrageurs, play a pivotal role in
financial markets.
Consider an investor who is of the view that the price of an asset is going to rise.
One way that a person can take a speculative position is by buying the asset in the
spot market and holding on to it, in the hope of offloading it subsequently at a higher
price.
However, buying the asset in the spot market entails incurring substantial costs.
In addition, in the case of a physical asset, the investor has to face the hassle of storing
and insuring it.
All this can be avoided if futures contracts are used for speculation. When taking
a long futures position, the investor can lock in a price at which to acquire the asset
subsequently. If the hunch is true and the spot price at the time of expiration of the
futures contract is higher, then the investor can take delivery at the initial futures price
as per the contract and sell it at the prevailing spot price, thereby making a profit.
The advantage of using futures is that the entire value of the asset need not be paid
at the outset. All that is required is a small margin. In other words, futures contracts
provide leverage.

Example 10.14. Futures contracts on IBM with three months to expiration are available at a price of
$75 per share. Alex is of the opinion that the spot price of IBM three months hence will be at least $78
per share. Therefore, he chooses to speculate by going long in 100 futures contracts, each of which is
for 100 shares.
Introduction to Options | 283

If his hunch is right and the spot price after three months is $80 per share, then Alex makes a
profit of

100 × 100 × (80 − 75) = $50,000

However, there is always a possibility that Alex is wrong. Let’s assume that he read the market
incorrectly and the price at expiration turns out to be $72 per share. If so, he would have to acquire
the shares at $75 per share and sell them in the spot market at $72 per share, thereby making a loss
of

100 × 100 × (72 − 75) = −$30,000

Thus speculation using futures can give rise to substantial gains if one is right, but can lead to signif-
icant losses if one misjudges the market.

Now let’s take the case of a bear, who is of the opinion that the market is going to
fall. He too can speculate, but by going short in a futures contract. If his hunch turns
out to be right, and the market price at the time of expiration of the futures contract is
indeed lower than what the futures price was at the outset, he can buy at the prevailing
market price and sell it at the contract price, thereby making a profit.

Example 10.15. Nina, like Alex, observes that the futures price for a three-month contract on IBM is $75
per share. However, unlike Alex, she is of the opinion that in three months time, IBM will be selling
for $72 or less per share in the spot market. Assume that she takes a short position in 100 futures
contracts.
If her hunch is right and the price of IBM after three months is $70 per share, Nina makes a profit
of

100 × 100 × (75 − 70) = $50,000

However, if the market rises to $78 after three months, Nina incurs a loss of

100 × 100 × (75 − 78) = −$30,000

So, like bulls, bears can use futures to speculate, but in their quest for substantial gains, there is
always a risk of substantial losses.

Introduction to Options
We have covered two types of derivative securities, namely futures and forward con-
tracts, in detail. In a futures contract no money changes hands when the contract is
negotiated, nor does the title to the goods. The important point to note is that once
a futures contract is negotiated, both the long and the short have an obligation at a
future date. Because both parties have an obligation, there is always a possibility that
the party with a loss at the expiration of the contract may default. As discussed previ-
ously, compliance is ensured by requiring both the parties to deposit good faith money
284 | Chapter 10: A Primer on Derivatives

or collateral called margins and by adjusting the profits and losses on a daily basis by
a procedure known as marking to market.
Options contracts, which are the focus now, are derivative contracts, but by design
are different from futures contracts. In an options contract, the buyer of the contract,
also called the holder or the long, has a right, and the seller, also known as the writer
or the short, has an obligation. Thus, in the case of an options contract, the long has
the freedom to decide whether to go through with the transaction, whereas the short
has no choice but to carry out the seller’s part of the agreement if and when the holder
chooses to exercise his right. Therefore, unlike in the case of futures contracts, both
the parties need not deposit collateral. For, if a person has a right, there is no fear
of noncompliance because he will exercise his right if it is in his interest and need
not otherwise. Consequently, in the case of options contracts, only the shorts have to
deposit margins.
The buyer of an option has either the right to buy the underlying asset or the right
to sell it, depending on the terms of the agreement. Therefore, there are two types of
options contracts, calls and puts. A call option gives the long the right to acquire the
underlying asset, whereas a put option gives the long the right to sell the underlying
asset. An option may give the long the right to transact only at a future point in time,
namely the expiration date of the option, or else it may give the flexibility of exercising
the option at any point in time, up to and including the expiration date. Consequently,
both calls and puts can be of two types, European and American. A European option
gives the holder the freedom to exercise the right only at the time of expiration of the
contract, whereas an American option can be exercised by the long at any point in
time on or before expiration. It must be noted that the terms European and American
have nothing to do with geographical locations and that in practice most exchange
traded contracts are American. Most textbooks, however, begin with an analysis of
European options because they are easier to value, as we have to take into account
the possibility of exercise at only a single point in time. Given the same features in all
other respects, an American option is more valuable than the corresponding European
option. This is because the holder of an American option has the flexibility to exercise
early, whereas the holder of a European option does not have the freedom to exercise
prior to expiration.

Common Terms Associated with Options

Let’s define the various terms that are used in connection with options contracts.

Exercise Price

This is the price the holder of a call option has to pay to the writer, per unit of the
underlying asset, if exercising the option. In the case of puts, it is the price the holder
Common Terms Associated with Options | 285

of a put option receives per unit of the underlying asset, if the option is exercised. The
exercise price is also known as the strike price.
The exercise price enters the picture only if the holder chooses to exercise the
option. The holder has a right and not an obligation, and may or may not decide to
transact, which means that the exercise price may or may not be paid/received.

Expiration Date

This is the point in time after which the contract becomes void. It is the only point in
time at which a European option can be exercised and the last point in time at which
an American option can be exercised. The expiration date is also known as the exercise
date, strike date or maturity date.

Option Premium

This is the price that the holder has to pay to the writer at the outset, in order to ac-
quire the right to exercise. The option premium is a sunk cost. If the holder were not
to exercise prior to expiration, the premium cannot be recovered.
Why does the buyer of an option have to pay a premium at the outset? Options
contracts entail the payment of a premium at the outset, because the buyer is acquiring
a right from the writer, who is taking on an obligation to perform if the buyer exercises
the right. Rights, it must be understood, are never free, and one always has to pay a
price to acquire them.
Futures and forward contracts, in contrast, do not entail the payment of a pre-
mium by the long. Such contracts impose an equivalent obligation on both the long
and the short. The futures price, which is the price at which the long will acquire the
asset at the time of expiration, is set in such a way that from the standpoints of both
the long and the short, the value of the contract at inception is zero. In other words,
the two equivalent and opposite obligations ensure that neither party has to pay the
other at the outset.
Another way of appreciating the difference between a commitment contract, such
a futures contract, and a contingent contract, such as an options contract, is as fol-
lows. When a futures contract is sealed, either the long or the short may have to coun-
tenance a loss. However, after a party buys an option from a writer, thereafter he will
not have to witness the specter of a cash outflow. If he exercises he will get a positive
cash flow, or else if he does not there will be no consequence. On the contrary, the
writer may have to face a cash outflow subsequently.
286 | Chapter 10: A Primer on Derivatives

Notation
We use the following symbols to depict the various variables:
– t ≡ today, a point in time before the expiration of the options contract.
– T ≡ the point of expiration of the options contract.
– St ≡ the stock price at time t.
– ST ≡ the stock price at the point of expiration of the options contract.
– X ≡ the exercise price of the option.
– Ct ≡ a general symbol for the premium of a call option at time t, when we do not
wish to make a distinction between European and American options.
– Pt ≡ a general symbol for the premium of a put option at time t, when we do not
wish to make a distinction between European and American options.
– CE,t ≡ the premium of a European call option at time t.
– PE,t ≡ the premium of a European put option at time t.
– CA,t ≡ the premium of an American call option at time t.
– PA,t ≡ the premium of an American put option at time t.
– r ≡ the riskless rate of interest per annum.

Exercising Call and Put Options

Let’s now take a look at the cash flows involved when an option is exercised. To keep
matters simple, we focus on European options, although the underlying logic is the
same in the case of American options. We use stock options, that is contracts based on
the common stock of a company, for the purpose of illustration. In principle, however,
an options contract can be written on any asset.

Example 10.16. Consider a person who buys a call option on IBM that expires in December 2019 and
has an exercise price of $125. Let’s assume that the option premium is $7.50 per unit of the underlying
asset, which in this case is a share of IBM. Option premia are always quoted on a per share basis. The
contract size for stock options, or in other words, the number of shares that the option holder can buy
or sell per contract, is kept fixed at 100 in the U.S. So, in this case, as soon as the deal is struck, the
buyer has to pay $7.50 × 100 = $750 to the writer. In exchange, the buyer gets the right to buy 100
shares of IBM on the expiration date at a price of $125 per share.
When will the holder choose to exercise his right? Quite obviously, if the stock price at the time
of expiration of the option is greater than $125, then it makes sense to exercise the option and buy
the shares at $125 each. Otherwise, it is best to let the option expire worthless. Readers encountering
options for the first time may feel that it makes sense to exercise only if the stock price at expiration is
greater than (125 + 7.50) or $132.50. This viewpoint is erroneous. Assume that the terminal stock price
is $127.50. If the option is exercised, the holder can buy 100 shares for $125 each and immediately
sell them for $127.50 each. We use the symbol π to denote profits and indicate losses by putting the
numbers in parentheses after the corresponding currency symbol. After taking into account the option
premium that was paid at the outset, the total profit is

π = (127.50 − 125) × 100 − 750 = $(500)

Exercising Call and Put Options | 287

Thus, the holder who exercises the option, suffers a loss of $500, but allowing the contract to ex-
pire worthless means losing the entire initial premium of $750. This argument is an illustration of the
maxim that sunk costs are irrelevant when taking investment decisions.

Example 10.17. Let’s reconsider the previous example, but assume that the options under considera-
tion are put options. The question is, when will the holder choose to exercise his right? Exercise is a
profitable proposition if the terminal stock price is less than $125. Otherwise, it is best to let the option
expire worthless. Assume that the terminal stock price is $115 and the premium paid for the option is
$3.00 per share. The profit is

π = (125 − 115) × 100 − 300 = $700

Notice, that in either case, when it is in the interest of the option holder to exercise,
the circumstances are not in favor of the option writer. You should now be able to
appreciate why an options contract is a right for the holder, but an obligation for the
writer. If both parties were to have rights, then the writer would refuse to transact when
conditions are in favor of the holder and vice versa. Thus, we can impose obligations
on both the parties like in the case of forward and futures contracts, or else give one
party a right and impose an obligation on the other like in the case of options contracts.

Payoffs and Profits: A Symbolic Representation

The payoff for a call holder at expiration is equal to ST − X, if ST > X, or 0 if ST ≤ X.

The payoff can therefore be represented as Max(0, ST − X).
The profit can be represented as Max(0, ST − X) − Ct , where Ct is the premium paid
for the call when it was acquired.
The payoff for the call writer is −Max(0, ST − X) = Min(0, X − ST ), and the profit is
Min(0, X − ST ) + Ct .
In the case of call options, therefore, the maximum profit for the holder is un-
limited because the stock price at the time of exercise has no theoretical upper limit.
The maximum loss, however, is restricted to the initial premium that is paid. From the
standpoint of the call writer, the situation is just the reverse. The maximum loss is un-
limited, whereas the maximum profit is the option premium that the writer receives at
the outset.
Similarly, the payoff for a put holder is equal to X − ST , if ST < X, or 0 if ST ≥ X.
The payoff can therefore be represented as Max(0, X − ST ).
The profit can be represented as Max(0, X − ST ) − Pt , where Pt is the premium paid
for the put when it was acquired.
The payoff for the put writer is −Max(0, X − ST ) = Min(0, ST − X), and the profit is
Min(0, ST − X) + Pt .
288 | Chapter 10: A Primer on Derivatives

For a put holder, therefore, the maximum profit is equal to the exercise price less
the option premium. Because, stocks have limited liability, the price cannot go below
zero. The holder’s maximum loss is once again the option premium. For the put writer,
the maximum profit is equal to the premium received, whereas the maximum loss is
equal to the exercise price minus the premium. Thus, irrespective of whether it is a
call or a put, an option is a zero sum game. The holder’s profit is equal to the writer’s
loss and vice versa.
As can be seen, an option holder can never lose more than the premium, whereas
the profit for an option writer can never exceed the premium. Call option buyers face
the possibility of finite losses and theoretically infinite profits. However, both the prof-
its and losses are capped for put holders. For call writers, profits are finite, but losses
are theoretically infinite. Put writers, however, face finite profits as well as losses.

Moneyness of the Option

The moneyness of an option refers to the relationship between the prevailing spot
price of the underlying asset, and the exercise price of the option.

Call Options

– If St > X, the call option is said to be in-the-money.

– If St = X, the call option is said to be at-the-money.
– If St < X, the call option is said to be out-of-the-money.

Example 10.18. Consider a stock that is currently trading at $500. A call option with an exercise price
of $500 is said to be at the money. A call with a lower exercise price, say $450, is said to be in the
money, whereas one with a higher exercise price, say $550, is considered to be out of the money.
Obviously, a call option is exercised only if it happens to be in the money.

Put Options

– If St > X, the put option is said to be out-of-the-money.

– If St = X, the put option is said to be at-the-money.
– If St < X, the put option is said to be in-the-money.

Example 10.19. Consider a stock that is currently trading at $500. A put option with an exercise price
of $500 is said to be at the money. A put with a lower exercise price, say $450, is considered to be out of
Intrinsic Value and Time Value of Options | 289

the money, whereas one with a higher exercise price, say $550, is said to be in the money. Obviously,
a put option, too is exercised only if it happens to be in the money.

Intrinsic Value and Time Value of Options

The intrinsic value of an option is equal to the amount by which it is in the money, if it is
in the money; otherwise, it is equal to zero. Therefore, the intrinsic value of a call option
is

I.V. = Max[(St − X), 0] (10.12)

and that of a put option is

I.V. = Max[(X − St ), 0] (10.13)

The difference between an option’s price and its intrinsic value is called the time value
or the speculative value of the option. By definition, no option, European or American,
call or put, can have a negative intrinsic value. However, some options may have a
negative time value.

Example 10.20. Assume that the price of a stock is $100 and the exercise price of a call option is
$97.50. If the call premium is $3.50, then

100.00 − 97.50 = $2.50

is the intrinsic value of the call option, and

3.50 − 2.50 = $1.00

is its time value.

Example 10.21. Assume that the price of a stock is $100 and the exercise price of a put option is
$97.50. The put premium is $1.25. In this case, the intrinsic value is zero because the put is out of
the money. Therefore, the entire premium of $1.25 is the time value of the option.

The Absence of Arbitrage and Its Implications for Option Prices

Let’s now state certain results pertaining to option prices. All option prices must sat-
isfy these properties, or else there will be arbitrage opportunities. What do we mean
by an arbitrage opportunity? Arbitrage refers to the ability of a trader to make cost-less
and risk-less profits. The phrase of significance here is cost less and risk-less. A costly
but risky investment should yield a risk-adjusted expected rate of return. A strategy
290 | Chapter 10: A Primer on Derivatives

that entails an investment, but is devoid of risk, should yield the risk-less rate of re-
turn. The ability of an individual to earn a return without making an investment, in an
environment devoid of risk, is referred to as arbitrage. In other words, a strategy that
yields a cash inflow at certain points in time and a zero cash flow at other points in
time can be termed an arbitrage strategy. This is because it leads to non-negative re-
turns for the investor without requiring an investment (which would have manifested
itself as a cash outflow) at any point in time. As should be obvious, an arbitrage oppor-
tunity will be exploited by anyone who perceives it to the maximum possible extent
until it is eliminated.
These conditions must hold if arbitrage is to be ruled out. They are, however, inde-
pendent of the way the option prices are determined. In other words, these conditions
are not specific to a particular option valuation model.

Non-Negative Option Premia

An option cannot have a negative price. The existence of a negative price would mean
that the writer is prepared to pay the holder to buy the option. If so, the holder can ac-
quire the option, pocket the payment, and simply forget about the contract. The reason
why the holder need not worry about further adverse consequences, or in other words,
subsequent negative cash flows, is because the option is a right and not an obligation.
Obviously, the presence of negative option premia would be a classic example of an
arbitrage opportunity.

Non-Negative Time Value of American Options

Consider American calls. We use the symbol CA,t to denote the price of an American
call option at time t. CA,t , should be greater than or equal to Max[(St − X),0].

Proof
If St − X is less than zero, we can say that CA,t ≥ 0 because an option cannot have a
negative premium. However, if St − X > 0, CA,t must be worth at least St − X. The proof
can be demonstrated as follows. Consider a case where St − X > 0 and CA,t < St − X.
If so, you can buy the option by paying CA,t and immediately exercise it. The profit of
St − X − CA,t that you get is clearly an arbitrage profit. Therefore, to preclude arbitrage,
we require that

CA,t ≥ Max[(St − X), 0]

In other words, an American call must be worth at least its intrinsic value. That is, an
American call cannot have a negative time value.
Lower Bound for Call Options | 291

Example 10.22. Let St = $105 and X = $100. Assume that CA,t = $4.50. An arbitrageur can buy the
call and immediately exercise it. This yields an arbitrage profit of

(105 − 100) − 4.50 = $0.50

per share.

A similar argument demonstrates that the put premium for an American option, PA,t ,
should be greater than or equal to Max[(X − St ),0]. That is, an American put, too, is
worth at least the intrinsic value and cannot have a negative time value.

Lower Bound for Call Options

X
Ct ≥ Max [0, St − ]
(1 + r)T−t

If St − X T−t > 0, the call premium will be greater. This automatically implies that Ct >
(1+r)
St −X. This result is common for European as well as American calls. The implication is
that an American call will never be exercised early, because its price is always greater
than its intrinsic value. A party who seeks to encash, will rather offset than exercise.
Consequently both European and American calls on a non-dividend paying stock will
have the same premium.
The rationale is the following. Keeping other variables constant, which obviously
implies a constant intrinsic value, the longer the time to expiration is, the greater the
interest earned on the exercise price by the call holder. Thus the longer the time to
expiration, the greater the time value is. Thus European and American calls on a non-
dividend paying stock, always have a non-negative time value except at the point of
expiration, at which time all options must have a premium equal to the intrinsic value,
as we shall shortly see.

Lower Bound for Put Options

X
PE,t ≥ Max [0, − St ]
(1 + r)T−t

If X T−t − St > 0, the put premium will be greater for European options. However,
(1+r)
this result does not necessarily imply that PE,t > X − St . Thus a European put on a
non-dividend paying stock may have a negative time value.
Keeping other variables constant, which obviously implies a constant intrinsic
value, the longer the time to expiration is, the greater the interest lost on the exer-
cise price by the put holder. Thus there could be a situation where the time value is
292 | Chapter 10: A Primer on Derivatives

negative. For instance, if the stock price is zero, the put holder will want to exercise
immediately. However, because the option is European, the holder is compelled to
wait, which explains the possible negative time value. In such a situation, an Amer-
ican put may be exercised. Consequently, unlike American calls, American puts on
non-dividend paying stocks may be exercised early.

Put-Call Parity for European Options

The put-call parity theorem states that
X
CE,t − PE,t = St − (10.14)
(1 + r)T−t
where CE,t represents the price of a European call with an exercise price of X and time
to expiration equal to T − t periods, and PE,t represents the price of a European put
with the same exercise price and expiration date.
Now let’s extend the result to a stock that pays a known dividend D at time t ∗ ,
which is a point in time before the expiration of the contract.
X D
CE,t − PE,t = St − T−t
− (10.15)
(1 + r) (1 + r)t
∗ −t

Option Premia at Expiration

American options and European calls on non-dividend paying stocks have a non-
negative time value, as we have seen earlier, whereas European puts on non-dividend
paying stocks may have either a positive or a negative time value depending on the
extent to which the option is in the money. At the time of expiration, the time value
of an option must be zero, that is, the option premium must be equal to the intrinsic
value. This is because, the holder does not have the freedom of waiting any further,
and consequently will not pay for the option to wait.

Proof

Consider a call option that is in the money at expiration. Assume that the time value is
positive, that is, CT is greater than ST −X. If so, an arbitrageur will immediately sell the
call. If it is exercised, the net cash flow is CT −(ST −X), which by assumption is positive.
The same holds true if the option is out of the money and the time value is positive,
because in this case the writer need not worry about exercise. On the other hand, if the
time value is negative, that is, CT is less than (ST −X), an arbitrageur will buy the option
and immediately exercise it. The net cash flow is ST − X − CT , which is guaranteed to
be positive. Thus, to preclude arbitrage, an option must sell for its intrinsic value at
expiration. The same rationale holds for put options, whether European or American.
Variables of Interest for Option Valuation | 293

Variables of Interest for Option Valuation

The values of call and put options, depend on the following variables.

The Current Stock Price

The prevailing stock price is obviously a major factor in determining the value of an
option. Everything else remaining constant, the higher the current stock price is, the
greater the value of a call option and the lower the value of a put option.

The Exercise Price

As one would expect, the higher the exercise price is for a given set of values for the
other variables, the lower the value of a call option and the higher the value of a put
option.

Dividends

A dividend payout leads to a drop in the stock price as the stock goes ex-dividend. Con-
sequently, dividends paid out during the life of the option lead to a reduction in call
values and an increase in put values. Exchange traded options are usually not payout
protected from the standpoint of cash dividends. That is, the terms of the original op-
tion contract are not modified if the underlying stock pays a dividend subsequently.6
The terms of the agreement, however, do change in the event of stock splits or stock
dividends.

Volatility

Modern finance theory is based on the assumption that all investors are risk averse.
Consequently, increases in the volatility, as measured by the variance of the rate of re-
turn of the stock, are perceived negatively and therefore lead to a higher risk premium
being demanded.
In the case of an options contract, however, the holder is protected on one side,
because the maximum loss is limited to the initial premium paid by the holder. There-
fore, an increase in the volatility is perceived positively, although it signals a greater

6 Some exchanges make an adjustment if the dividend is perceived to be extraordinary. In such cases,
the exchange has to define the meaning of what it terms as extraordinary.
294 | Chapter 10: A Primer on Derivatives

probability of higher stock prices as well as lower stock prices. Hence an increase in
the volatility of the rate of return on the underlying asset leads to an increase in the
value of both call as well as put options.

Time to Maturity

American options and European calls on non-dividend paying stocks have a non-
negative time value, whereas European puts on such stocks may have either a positive
or a negative time value depending on the extent to which the option is in the money.
As we have seen, the time value of an option must be zero at the time of expiration;
that is, the option premium must be equal to the intrinsic value. Therefore, everything
else remaining constant, the value of an option generally declines with the passage
of time. We use the word “generally,” because certain European puts that are deep in
the money may have a negative time value that approaches zero at expiration. Hence,
options are also known as wasting assets, because they experience a decline in value
with the passage of time.

Riskless Rate of Interest

Let’s take the case of an investor who is contemplating the purchase of a stock. One
alternative is to buy a call instead, which can be subsequently exercised. The price of a
call must be less than the prevailing stock price. This is because if a stock is desirable,
it would make sense to acquire it at the current stock price, rather than pay a premium
that, by assumption, is greater and subsequently pay the exercise price in the event of
exercise.
So a person who has an amount equal to the value of the stock with him, can
decide to buy the call instead of the stock and then invest the difference at the riskless
rate of interest. Everything else being the same, the higher the rate of interest, the
more attractive the strategy is of buying the call and investing the surplus. Therefore,
it follows that the higher the interest rate is, the higher the price of the call.
What about puts? Consider the case of a person who already owns the stock and
is contemplating selling it. One alternative is to buy a put, which ensures that the
buyer will receive a minimum price of $X for it subsequently. The higher the interest
rate is, the more attractive the alternative of selling immediately in the spot market
rather than buying a put. Consequently, everything else being the same, the higher the
interest rate, the less attractive the put option is. Hence, put prices decline as interest
rates rise.
The Binomial Model of Option Valuation | 295

The Binomial Model of Option Valuation

Let’s now go on to study a theoretical pricing model for call and put options called the
binomial option pricing model (BOPM).
The binomial model assumes that given the current stock price, during the next
period the price can either change by X% or Y%, where X and Y are specified. Be-
cause the stock price can take on only two possible values at the end of the period, the
model is called binomial. A realistic portrayal of the behavior of a stock requires the
specification of Y as a negative number.

The One Period Binomial Model

In the one period case, we consider only two points in time. These are, the present
time (T-1) and the expiration time of the option T.
Let the current stock price be St . The stock price at the end of the period can take
on only the following values:
X
ST = St (1 + ) = uSt (10.16)
100
X
where u = (1 + 100
) and stands for the up state, or
Y
ST = St (1 + ) = dSt (10.17)
100
Y
where d = (1 + 100 ) and stands for the down state. Y, in this case, is a negative number.
The price tree may be depicted as shown in Figure 10.1.

Figure 10.1: Stock price tree for the one period case.

Now, the call value at expiration will be Max[0, uSt − E] if the up state is reached, or
Max[0, dSt −E] if the down state is reached. We are going to use the symbol E to denote
296 | Chapter 10: A Primer on Derivatives

the exercise price because the symbol X has already been used to denote the extent of
an up movement.
Let Cu = Max[0, uSt − E] and Cd = Max[0, dSt − E].
Cu is the call value at expiration if the up state is reached, and Cd is the call value
at expiration if the down state is reached.
Our objective is to find the value of the call today, that is Ct .
Consider the following strategy. Let’s buy α shares of stock and write one call op-
tion. The current value of this portfolio is αSt − Ct .
If the up state is reached, the portfolio will have a value = αuSt − Cu , whereas if
the down state is reached, it will have a value = αdSt − Cd .
Now let’s make this portfolio riskless by choosing α, such that

αuSt − Cu = αdSt − Cd
⇒ αSt (u − d) = Cu − Cd
Cu − Cd
⇒α= (10.18)
St (u − d)

α is known as the hedge ratio.

This portfolio is riskless because the payoff is independent of the state of nature at
time T. Since our portfolio by design is riskless, it must earn the riskless rate of return
to preclude arbitrage. Therefore,

αuSt − Cu = αdSt − Cd = (αSt − Ct )r

where r = 1+ riskless rate of interest. The symbol r is often used to denote the riskless
rate of interest per period. In the binomial model, r represents one plus the riskless
rate per period, and is defined as one plus the periodic interest rate.
We typically assume that the parameters u and d, as well as the riskless rate of
return r, are constant. However, although this is done in order to simplify the calcula-
tions, this is not a necessary condition. The model merely requires that these variables
be deterministic, that is they are known with certainty to all investors. It can also be
shown that u should be greater than r, which in turn should be greater than d, to pre-
clude arbitrage. Thus,

αuSt − Cu = (αSt − Ct )r
Cu − Cd C − Cd
⇒ [ ] × uSt − Cu = ( u ) × rSt − Ct r
St (u − d) St (u − d)
C − Cd
⇒ ( u ) × (u − r) − Cu = −Ct r
u−d
r−d u−r
Cu ( u−d ) + Cd ( u−d )
⇒ Ct = ( ) (10.19)
r
The Binomial Model of Option Valuation | 297

r−d
Let = p. Therefore,
u−d
u−r
=1−p
u−d
pC + (1 − p)Cd
⇒ Ct = ( u ) (10.20)
r
This is the one-period binomial call option pricing formula. p and (1 − p) are referred
to as risk-neutral probabilities.

Example 10.23. Let St = 100, and X = 100. u = 1.2, d = 0.80, r = 1.05

Cu = Max[0, 1.2 × 100 − 100] = 20

Cd = Max[0, 0.8 × 100 − 100] = 0
r−d 1.05 − 0.80
p= = = 0.625
u−d 1.2 − 0.80
1 − p = 0.375
C −C 20−0
The hedge ratio = S u(u−d)
d
= 100(1.2−0.80) = 0.5
t
That is, we have to buy 0.5 shares for every call that we write, in order to form a riskless portfolio.
0.625 × 20 + 0.375 × 0
Ct = = $11.9048
1.05

The Two-Period Case

Thus far, we have assumed that the option has only one period left to expiration. In
general, the stock price will move many times between the date of valuation and the
expiration date. In the multi-period case also, the same arguments hold.
Let the stock price two periods before the expiration date be St . The stock price
tree can then be depicted as shown in Figure 10.2.
At T-1, there is only one period left to expiration, and so we an apply the one-period
model to get Cu and Cd at T − 1:
pCuu + (1 − p)Cud
Cu =
r
pCud + (1 − p)Cdd
Cd =
r
We know Cuu , Cud , and Cdd because they represent terminal values of the option.
Cuu is the option price at T if there are two upticks in the stock price. Cud is the
option price at T if there is an uptick followed by a downtick or vice versa. Cdd is the
option price at T if there are two downticks in the stock price.
After we find Cu and Cd , we can work backwards to find Ct :
pCu + (1 − p)Cd
Ct =
r
p2 Cuu + 2p(1 − p)Cud + (1 − p)2 Cdd
= (10.21)
r2
298 | Chapter 10: A Primer on Derivatives

Figure 10.2: Stock price tree for the two-period case.

Using the same iterative process we can work backwards, moving one period at a time,
to solve the N-period case.

Example 10.24. Let’s use the same data as in the earlier example. St = X = 100, u = 1.2, d = 0.8,
r = 1.05.
First, let’s draw the price tree.

Figure 10.3: Stock price tree for the two-period example.

p = 0.625, (1 − p) = 0.375
Cuu = Max[0, 144 − 100] = 44
Cud = Max[0, 96 − 100] = 0
Cdd = Max[0, 64 − 100] = 0
Using the Binomial Model: The Case of European vs. American Puts | 299

0.625 × 44 + 0.375 × 0
Cu = = $26.1905
1.05
0.625 × 0 + 0.375 × 0
Cd = =0
1.05
0.625 × 26.1905 + 0.375 × 0
Ct = = $15.5896
1.05
Thus the two-period call is priced higher than the one-period call. This is to be expected because Eu-
ropean calls on a non-dividend paying stock will have a positive time value.

The Binomial Model for European Puts

Consider the one-period case. The put value at expiration is Max[0, E − uSt ] if the up
state is reached, or Max[0, E − dSt ] if the down state is reached. Therefore,
Pu = Max[0, E − uSt ]
and Pd = Max[0, E − dSt ]
Using arguments similar to what we used for the call, it can be shown that
pPu + (1 − p)Pd
Pt = (10.22)
r
where, p and (1 − p) are as defined before.

Example 10.25. Let’s use the same data that we used for the call. That is, St = X = 100, u = 1.2,
d = 0.8, r = 1.05.

Pu = Max[0, 100 − 120] = 0

Pd = Max[0, 100 − 80] = 20
0.625 × 0 + 0.375 × 20
Pt = = $7.1429
1.05

Using the Binomial Model: The Case of European vs. American Puts
Consider the data used in the earlier examples. Now, assume that the stock pays no
dividends. If so, American and European calls have the same premium. However, an
American put may be priced higher than a European put. Let’s look at a put option
with three periods to expiration. The stock price tree can be depicted as shown in Fig-
ure 10.4.

Valuing a European Put Using the Binomial Model

To value the option, we proceed backwards as follows:

0.625 × 0 + 0.375 × 0
Puu = =0
1.05
300 | Chapter 10: A Primer on Derivatives

Figure 10.4: Stock price tree for the three-period case.

0.625 × 0 + 0.375 × 23.20

Pud = = $8.2857
1.05
0.625 × 23.2 + 0.375 × 48.8
Pdd = = $31.2381
1.05
0.625 × 0 + 0.375 × 8.2857
Pu = = $2.9592
1.05
0.625 × 8.2857 + 0.375 × 31.2381
Pd = = $16.0884
1.05
0.625 × 2.9592 + 0.375 × 16.0884
Pt = = $7.5073
1.05

Valuing an American Put Using the Binomial Model

For American puts, we have to, compare, at each node, the value obtained from the
model with the intrinsic value. If the intrinsic value is greater, we have to use it for
subsequent calculations. The rationale is the following. If the intrinsic value is higher
than the model value, a trader who wants to exercise will obviously do so. A trader
who wants to wait would rather exercise, collect the intrinsic value, and recreate the
option at the model value. Thus if the intrinsic value at a node is greater than the
model value, early exercise takes place.
At uuSt , the value according to our previous calculations is 0. The intrinsic value
is also 0.
The Black-Scholes Formula for Valuing Options | 301

At udSt , the model price is $8.2857, and the intrinsic value is 4, which is less.
Thus Pu has the same value for an American put.
But at ddSt , the model price is $31.2381, and the intrinsic value is 36. So we calcu-
late Pd as follows:

0.625 × 8.2857 + 0.375 × 36

Pd = = $17.7891
1.05
The intrinsic value at the node corresponding to a stock price of dSt is 20, which
is greater than $17.7891.
So we calculate Pt as follows:

0.625 × 2.9592 + 0.375 × 20

Pt = = $8.9043
1.05
This is not the final step, however. You must compare $8.9043 with the intrinsic value
at T-3 which in this case is 0. Thus the option will sell for $8.9043.
As you can see, the American put is priced higher than the European put. This is
because early exercise is optimal at an intermediate stage.

The Black-Scholes Formula for Valuing Options

Black and Scholes obtained exact formulae for valuing call and put options on non-
dividend paying stocks, by assuming that stock prices follow a lognormal process.
This assumption implies that logs of prices are normally distributed. The formulae
obtained by them are:

CE,t = St N(d1 ) − Xe−r(T−t) N(d2 ) (10.23)

and

PE,t = Xe−r(T−t) N(−d2 ) − St N(−d1 ) (10.24)

where
S σ2
ln ( Xt ) + (r + 2
) (T − t)
d1 = (10.25)
σ √(T − t)

and

d2 = d1 − σ √(T − t) (10.26)

N(X) is the cumulative probability distribution function for a standard normal vari-
able, and σ is the standard deviation of the rate of return on the stock. N(X) may be
computed in Excel by using the NORMSDIST function.
302 | Chapter 10: A Primer on Derivatives

Example 10.26. Consider a stock which is currently selling for $100. Call and put options are available
with X = 100 and time to expiration = 6 months.
The riskless rate of interest = 10% per annum, and the volatility is 30% per annum. So, St = X =
100; T − t = 0.5 years; r = 10% ≡ 0.10; σ = 30% ≡ 0.30.
Let’s consider the call first:

100 (0.30)2
ln ( 100 ) + [0.10 + 2
] 0.5
d1 =
0.3√0.5
0.0725
= = 0.3418
0.2121
d2 = 0.3418 − 0.2121 = 0.1296

N(0.3418) = 0.6337 and N(0.1296) = 0.5516.

CE,t = 100 × 0.6337 − 100e−0.10×0.5 × 0.5516

= 10.9065
PE,t = 100e−0.10×0.5 N(−0.1296) − 100N(−0.3418)

Now, N(−X ) = 1 − N(X ). This is because the normal distribution is symmetrical. So, N(−0.1296) =
1 − 0.5516 = 0.4484, and N(−0.3418) = 0.3663. Thus, PE,t = $6.0294.

Put-Call Parity and Option Pricing Models

Put-call parity is a more general result than either the binomial or the Black-Scholes
models. This is because put-call parity merely requires the absence of arbitrage,
whereas option pricing models require an additional assumption about the evolu-
tion of stock prices. In the following section, we show that the Black-Scholes formula
satisfies put-call parity.

CE,t = St N(d1 ) − Xe−r(T−t) N(d2 )

= St [1 − N(−d1 )] − Xe−r(T−t) [1 − N(−d2 )]
= Xe−r(T−t) N(−d2 ) − St N(−d1 ) + St − Xe−r(T−t)
= PE,t + St − Xe−r(T−t)

Consider the data given earlier when we illustrated the one-period binomial model.
The one period call was priced at $11.9048, while the one period put was priced at
$7.1429. The difference is $4.7619. The stock price and the exercise price were both $100
and the risk-less rate per period was 5%. Thus

X 100
St − = 100 − = $4.7619
(1 + r) 1.05

Hence put-call parity is satisfied as expected.

Chapter Summary | 303

Interpretation of N(d1 ) and N(d2 )

The Black-Scholes formula for call options states that

CE,t = St N(d1 ) − Xe−r(T−t) N(d2 )

The option premium is invariant to the risk preferences of investors and is conse-
quently valid in a world characterized by risk neutrality. A risk-neutral investor would
value the option as the discounted value of its expected payoff. Consequently, because

CE,t = e−r(T−t) [St er(T−t) N(d1 ) − XN(d2 )]

it is obvious that

[St er(T−t) N(d1 ) − XN(d2 )]

is the expected payoff from the option from the perspective of a risk-neutral investor.
St er(T−t) N(d1 ) is the expected value of a variable in a risk-neutral world, that is
equal to ST if the option is exercised, and is equal to zero otherwise. N(d2 ) is the prob-
ability that the option will be exercised in a risk-neutral world. If the option is exer-
cised, there is an outflow of X, otherwise the outflow is zero. Consequently, XN(d2 ) is
the expected outflow on account of the exercise price.
The formula for puts states that

PE,t = Xe−r(T−t) N(−d2 ) − St N(−d1 )

⇒ = e−r(T−t) [XN(−d2 ) − St er(T−t) N(−d1 )]

If N(d2 ) is the probability that a call option with an exercise price of X is exercised in a
risk-neutral world, then 1 − N(d2 ) or N(−d2 ) is the probability that a put with the same
exercise price will be exercised. Thus XN(−d2 ) is the expected inflow on account of the
exercise price. St er(T−t) N(−d1 ) is the expected value of a variable that is equal to ST if
the option is exercised or equal to zero otherwise.
N(d1 ) is known as delta, which is the partial derivative of the call price with respect
to the stock price. For put options, the delta is −N(−d1 ). For calls delta is between zero
and one. That is, an option that is deep out-of-the-money has a delta close to zero,
whereas a deep in-the-money call option has a delta close to one. For puts, delta is be-
tween −1 and zero. Deep out-of-the-money options have a delta close to zero, whereas
deep-in-the-money options have a delta equal to minus one.

Chapter Summary
The chapter provided an introduction to forward and futures contracts, as well as op-
tions contracts. We go on to study interest rate forwards, futures, and options, for
304 | Chapter 10: A Primer on Derivatives

which this chapter should prove to be an adequate introduction to readers who are
not well-acquainted with financial derivatives. The chapter commenced with a study
of forward contracts and futures contracts, and presented their similarities and differ-
ences. In this context, key issues such as margining, and marking to market, were in-
troduced. We then presented no-arbitrage arguments to value forward contracts. The
concept of the cheapest to deliver grade, when multiple grades are allowed for de-
livery, was examined. The use of futures contracts for hedging and speculation was
studied in detail. The focus then turned to options contracts. Properties of European
and American options, both calls and puts, were studied. The binomial model, which
is used later to study forward rate agreements, and bond options, was examined in
detail. The chapter concluded with a study of the Black-Scholes model. The use of this
model from a pricing standpoint was illustrated, and the key terms of the formula were
interpreted.
Chapter 11
The Valuation of Interest Rate Options
In this chapter we study models for the evolution of interest rates over time, and their
use in the valuation of options on interest rates. We briefly discuss equilibrium models
of the term structure, although the bulk of the discussion is on no-arbitrage models.
The primary focus is on the Ho-Lee model and the Black-Derman-Toy (BDT) models.

Short Rates
A short rate of interest is a future spot rate of interest that might rise over time. Usually
it is denoted as a single-period rate of interest and defined as the rate for the shortest
period of time considered by a model. In a fixed income context, we take the shortest
period of time as six months, as most bonds generate cash flows on a semiannual
basis.
At any point in time, if we consider a series of spot rates corresponding to different
lengths of time, we can derive a unique vector of forward rates. However, for any future
point in time, there are an infinite number of short rates that may arise. Each possible
value has a varying probability of occurrence. At the initial point in time, the one-
period spot rate is equal to the one-period forward rate, which is equal to the one-
period short rate. However, if we look at a time horizon beyond the first period, these
interest rate measures in general are not equal to each other.

Issues in the Valuation of Interest Rate Derivatives

A Black-Scholes approach or a standard binomial model, which works well in the case
of stock options, is inappropriate for interest rate dependent financial claims and their
derivatives. For instance, let us take the case of a bond maturing after N years. Al-
though the price before maturity is expected to evolve in a random fashion, as the
time to expiration approaches zero, the price of the underlying security should tend
towards its par value. Consequently, option pricing models based on lognormal or
binomial outcomes for the price of the underlying asset are unsuitable for valuing a
derivative like an option on an interest rate dependent asset such as a bond.
There is no standard model for pricing options on debt securities. However, almost
all of the models that have been postulated are based on the following three-step ap-
proach:1
1. The random character of interest rate movements is first modeled.

1 See Rendleman [53].

https://doi.org/10.1515/9781547400669-011
306 | Chapter 11: The Valuation of Interest Rate Options

2. The interest rate process is used to infer the distribution of prices of the underlying
debt securities.
3. The distribution of prices of the underlying asset is used to value the option.

Thus, the difference between competing models lies in the way they model the interest
rate process.

Equilibrium Models of the Term Structure

Such models derive a stochastic process for the evolution of the short rate by making
a set of assumptions about economic variables. In a one-factor model, the stochastic
process for the short rate is the only source of uncertainty. In general, short rates are
assumed to follow an Ito process of the form:2

dr = μ(r)dt + σ(r)dZ

That is, the drift μ, and the variance rate σ 2 are assumed to be functions of the short
rate, but are considered to be independent of time. There is, however, a critical prob-
lem with this approach to the pricing of interest rate derivatives. In practice, it is usu-
ally the case that the prices of the underlying debt securities, which are inferred from
the postulated interest rate process, are not equal to their observed market prices. It is
obvious that if we are unable to correctly price the underlying security, it is impossible
for us to have faith in the derivative prices that are obtained from the same.

Arbitrage-Free Term Structure Models

This class of option pricing models takes the observed term structure of interest rates
as an input, unlike the earlier class of models that generate the term structure as an
output. The stochastic process for the short rate that is implied by the observed term
structure is deduced, and from it the values of the underlying debt securities are in-
ferred. Finally, the price of the derivative security is obtained. The appealing feature
of this approach is that the prices of the debt securities that are derived in this fashion
are entirely consistent with the observed term structure.
Let’s now look at alternative models of the term structure.

2 dZ = ϵ√△t, where ϵ ∼ N(0, 1); that is, it is drawn from a standard normal distribution, which has a
mean of zero and a variance of 1.
The Binomial Tree Approach to the Term Structure | 307

The Ho-Lee Model

Per the Ho-Lee model, the evolution of the short rate may be specified as

drt = μt dt + σdZ (11.1)

In this model, the drift varies over time, and the variance remains constant. A short-
coming of this model is that interest rates can become negative, although this is not
as outrageous today as it was a decade ago. The other drawback of this model is that
it does not incorporate mean reversion, which refers to the tendency of a variable to
return to its long-term average, after moving up and down in the short-run. Interest
rates and security prices are obvious candidates. In practice mean reversion can be
self-fulfilling. If a variable moves to an abnormally low or high value, it can attract
players who do not subscribe to the herd mentality and take a counter-approach. If
a sufficient segment of the market were to think along the lines of the heretics, the
market could move back to a long-term average.

The Hull-White Model

The Hull-White model specifies the evolution of the short rate as

drt = [μt − αrt ]dt + σdZ (11.2)

This model builds in mean reversion. Like in the case of the Ho-Lee model, interest
rates can become negative.

The Black-Derman-Toy Model

The Black-Derman-Toy (BDT) model may be stated as

d ln rt = μt dt + σt dZ (11.3)

The BDT model, like the other two, is a single-factor model. That is, it assumes that
there is a single stochastic factor. The model was the first to combine the lognor-
mal distribution with mean reversion. Unlike the Ho-Lee and Hull-White models, the
volatility is not a constant. Thus, in order to use the model, we need to take cognizance
of both the term structure of interest rates and the term structure of volatility.

The Binomial Tree Approach to the Term Structure

Now let’s illustrate no-arbitrage models of the term structure using a discrete time
framework. Before we proceed we need to define certain variables.
308 | Chapter 11: The Valuation of Interest Rate Options

We denote the length of a time period in the binomial tree as △t, where time is
measured in years. Interest rates consequently are specified as annual rates. sn is the
n-period spot rate at time 0, quoted as a percentage per annum, but compounded at
an interval equal to △t. Thus if a dollar is invested at time 0 for n periods, the future
value is

F.V. = (1 + sn △ t)n

We define rn as the short rate that will prevail n periods later. r0 is obviously equal to
the one-period spot rate s1 . The time step in the binomial tree is taken as six months
for the purpose of illustration. Consequently, △t = 0.50.
Consider a bond that pays $1 after two periods. Its present value is

1
s2 2
(1 + 2
)

The short rate after one period is either ru or rd . Thus the value of the bond after one
period is either

1
Pu = r
(1 + 2u )

or
1
Pd = r
(1 + 2d )

The present value of the one-period bond is therefore

Pu
(1 + s21 )

or
Pd
(1 + s21 )

Let q be the probability of an up move and (1 − q) be the probability of a down move.

Thus
1 q × Pu (1 − q) × Pd
= +
(1 + s2 2
) (1 + s21 ) (1 + s21 )
2
1 q 1−q
⇒ = r + s r
(1 + s2 2
) (1 + s21 ) (1 + 2u ) (1 + 21 ) (1 + 2d )
2
ru r
(1 + ) (1 + 2d ) (1 + s21 ) r ru
⇒ 2
= q (1 + d ) + (1 − q) (1 + )
(1 s2 2
+ 2) 2 2
The Binomial Tree Approach to the Term Structure | 309

ru rd s1
(1 + ) (1 + ) (1 + ) rd r
⇒ 2 2 2
− q (1 + ) − (1 − q) (1 + u ) = 0 (11.4)
(1 + s2 2
) 2 2
2

This is the fundamental equation that is used in no-arbitrage models to determine the
short rates. To apply these models in practice, one has to model ru and rd . The models
differ in their approaches to modeling these parameters.

Some Insights into the Ho-Lee Model

Let’s depict a four-period interest rate tree. Each node is denoted as rk,j , where k repre-
sents the point in time that we are at and j represents the state. The tree can be modeled
as shown in Figure 11.1.

Figure 11.1: Interest rate tree.

Thus an up move from node rk,j is a move to node rk+1,j , while a down move is a move to
node rk+1,j+1 . Notice that the tree in Figure 11.1 is recombining; that is, rud is the same as
rdu . The recombination condition not only reduces the number of nodes at each point
in time and consequently makes computation relatively easier, it also has implications
for σ, as we demonstrate shortly.
310 | Chapter 11: The Valuation of Interest Rate Options

Interest rates span time whereas prices do not. Consequently the interest rate tree
can be used to price bonds with up to four periods to maturity. That is, if we have a
bond that pays $1,000 at time t4 , the price tree corresponding to the preceding interest
rate tree can be depicted as shown in Figure 11.2.

Figure 11.2: Price tree.

Consider time point tk , where tk = k △ t, and let rk be the short rate at this point in
time. For the process

dr = μt dt + σt dZ

we can approximate dr as rk+1,j − rk,j or as rk+1,j+1 − rk,j , and write

rk+1,j = rk,j + μk △ t + σk √△t, that is ϵ = +1,

and

rk+1,j+1 = rk,j + μk △ t − σk √△t, that is ϵ = −1

Calibrating the Ho-Lee Model | 311

Thus

r1,1 = r0,1 + μ0 △ t + σ0 √△t

and

r1,2 = r0,1 + μ0 △ t − σ0 √△t

⇒ r1,1 − r1,2 = 2σ0 √△t

Hence the spread between the upstate and the down state is a function of the volatility
parameter. Similarly

r2,1 = r1,1 + μ1 △ t + σ1 √△t

r2,2 = r1,1 + μ1 △ t − σ1 √△t = r1,2 + μ1 △ t + σ1 √△t
r2,3 = r1,2 + μ1 △ t − σ1 √△t

Notice that r2,2 can be expressed as a down move from r1,1 or as an up move from r1,2
because of the recombination requirement. Thus

r1,1 + μ1 △ t − σ1 √△t = r1,2 + μ1 △ t + σ1 √△t

⇒ 2σ1 √△t = r1,1 − r1,2
r1,1 − r1,2 2σ0 √△t
⇒ σ1 = = = σ0
2√△t 2√△t
Therefore if the interest rate tree is to recombine, then it must be the case that σ is a
constant for all points in time. In other words, σk = σ∀k.

Calibrating the Ho-Lee Model

Let’s now illustrate how the Ho-Lee model can be fitted to the observed term structure
using the binomial tree approach. The parameters of the model are chosen in a way
such that the branches of the tree recombine. This, as you have already seen, implies
that the standard deviation, σ, is constant for all points in time. In practice the value
of σ is chosen arbitrarily. It is also assumed that given a state of nature, there is a 50%
probability of an up move and, therefore, obviously a 50% probability of a down move.
The movements in the short rate are expressed as follows:

rt+1 = rt + μt △ t + σϵt+1 √△t

where ϵt+1 = +1 with a probability of 0.50 and is equal to −1 with a probability of 0.50.
Thus

rk+1,j = rk,j + μk △ t + σ √△t

312 | Chapter 11: The Valuation of Interest Rate Options

and

rk+1,j+1 = rk,j + μk △ t − σ √△t

The values of μ are determined to generate the term structure that is observed in prac-
tice. Before we proceed to illustrate the calibration technique, let’s first introduce the
concept of an Arrow-Debreu security, a concept that we use subsequently to value in-
terest rate dependent securities.

Arrow-Debreu Securities

Arrow-Debreu securities, also known as pure securities pay $1 if a particular state of

nature occurs. However, in all other states of nature, the payoff is zero. Obviously each
node of a binomial tree represents a particular state of nature. The value at time t0 of
a pure security that pays off $1 in state (k, j) is denoted by qk,j .
Let’s first consider a one-period binomial tree, shown in Figure 11.3. We assume
that the one-period interest rate is 8% per annum. Each period represents a time in-
terval of six months.

Figure 11.3: One period interest rate and state price tree.

Thus the price of a security that pays off $1 in the up state is given by

1 0.50
q1,1 = 0.50 × = = $0.4808
(1 + 0.08 × 0.5) 1.04

q1,2 or the price of a security that pays $1 in the down state, is obviously the same. The
values of q that are calculated in this fashion are referred to as state prices.
A riskless security in such an environment is obviously one that pays off $1 irre-
spective of the state of nature after one period. Its price is obviously given by

q1,1 + q1,2 = 0.4808 + 0.4808 = $0.9616

Calibrating the Ho-Lee Model | 313

The riskless rate of interest is therefore

1 − 0.9616
≡ 4%
0.9616
per half-yearly period, or 8% per annum, which is what we expect.
Now let’s extend the model by a period, as shown in Figure 11.4. Assume that if
the up state is reached after a period, the rate of interest will be 8.75% per annum,
whereas if the down state is reached, it will be 7.25% per annum.

Figure 11.4: Two-period interest rate and state price tree.

At the end of two periods, there obviously are three states of nature. We can calculate
the prices of the Arrow-Debreu securities as follows:
0.5 × 0.5
q2,1 =
(1.04) × (1 + 0.0875
2
)
0.5 0.5
= q1,1 × = 0.4808 × = $0.2303
1.04375 1.04375
Let’s analyze this expression. The value of an Arrow-Debreu security that pays off $1
0.5
at state (2,1) is 1.04375 at state (1,1). We know that the value of a pure security that pays
off $1 in state (1,1) is $0.4808 as calculated at state (0,1). Consequently, the value of a
pure security that pays off $1 in state (2,1), as calculated at state (0,1), is
0.5
0.4808 × = $0.2303
1.04375
Now let’s consider a pure security that pays off $1 if state (2,2) occurs. This state
can be attained via two paths. Hence, using the same logic
0.5 0.5
q2,2 = q1,1 × 0.0875
+ q1,2 × 0.0725
(1 + 2
) (1 + 2
)
314 | Chapter 11: The Valuation of Interest Rate Options

0.5 0.5
= 0.4808 × + 0.4808 × = $0.4623
1.04375 1.03625
Similarly,
0.5
q2,3 = q1,2 × = $0.2320
1.03625
Thus this approach enables us to price pure securities at various states of nature by
working our way forward through the tree.
Now consider the term structure in Table 11.1.

Table 11.1: Vector of spot rates.

Period Spot Rate

0 7.50%
1 7.00%
2 6.25%
3 6.75%

Assume that σ = √1 . Thus σ × √△t = 1.0. r0,1 is obviously 7.50%. q0,1 = 1

△t
We know that

r1,1 = r0,1 + μ0 △ t + 1.00 = 8.50% + μ0 × 0.50

and

r1,2 = r0,1 + μ0 × 0.50 − 1.00 = 6.50% + μ0 × 0.50

0.50
q1,1 = q1,2 = = $0.481928
(1 + 0.075
2
)

The prices of the pure securities for the next period may be computed as follows:
0.50 0.481928
q2,1 = q1,1 × =
(1 +
0.085+μ0 ×0.50
) 2.085 + μ0 × 0.50
2
0.50 0.50
q2,2 = q1,1 × 0.085+μ0 ×0.50
+ q1,2 × 0.065+μ0 ×0.50
(1 + 2
) (1 + 2
)
0.481928 0.481928
= +
2.085 + μ0 × 0.50 2.065 + μ0 × 0.50
0.50 0.481928
q2,3 = q1,2 × =
(1 +
0.065+μ0 ×0.50
) 2.065 + μ0 × 0.50
2

Obviously,
1 1
q2,1 + q2,2 + q2,3 = 2
= = $0.933511
(1 + 0.5 × s2 ) (1.035)2
Calibrating the Ho-Lee Model | 315

Thus,
2 2 0.933511
+ = = 1.937034
2.085 + μ0 × 0.50 2.065 + μ0 × 0.50 0.481928
⇒ μ0 = −1.9874%

Thus r1,1 = 0.085 − 0.019874 × 0.50 = 0.075063 ≡ 7.5063% and r1,2 = 0.065 − 0.019874 ×
0.50 = 0.055063 ≡ 5.5063% The prices of the pure securities are

q2,1 = $0.232247
q2,2 = $0.466755
q2,3 = $0.234508

Now let’s proceed to the next point in time:

r2,1 = r1,1 + μ1 × 0.50 + 1.0 = 7.5063 + 1.0 + μ1 × 0.50 = 8.5063 + μ1 × 0.50

r2,2 = r1,1 + μ1 × 0.50 − 1.0 = 7.5063 − 1.0 + μ1 × 0.50 = 6.5063 + μ1 × 0.50
r2,3 = r1,2 + μ1 × 0.50 − 1.0 = 5.5063 − 1.0 + μ1 × 0.50 = 4.5063 + μ1 × 0.50

The prices of the pure securities for the next point in time can be expressed as
follows:
0.50 0.232247
q3,1 = q2,1 × =
(1 +
0.085063+μ1 ×0.50
) 2.085063 + μ1 × 0.50
2
0.50 0.50
q3,2 = q2,1 × 0.085063+μ1 ×0.50
+ q2,2 × 0.065063+μ1 ×0.50
(1 + 2
) (1 + 2
)
0.232247 0.466755
= +
2.085063 + μ1 × 0.50 2.065063 + μ1 × 0.50
0.50 0.50
q3,3 = q2,2 × 0.065063+μ1 ×0.50
+ q2,3 × 0.045063+μ1 ×0.50
(1 + 2
) (1 + 2
)
0.466755 0.234508
= +
2.065063 + μ1 × 0.50 2.045063 + μ1 × 0.50
0.50 0.234508
q3,4 = q2,3 × =
(1 +
0.045063+μ1 ×0.50
) 2.045063 + μ1 × 0.50
2

Obviously,
1 1
q3,1 + q3,2 + q3,3 + q3,4 = 3
= = $0.911818
(1 + 0.5 × s3 ) (1.03125)3
Thus,
0.232247 0.466755 0.234508
2× +2× +2×
2.085063 + μ1 × 0.50 2.065063 + μ1 × 0.50 2.045063 + μ1 × 0.50
= 0.911818 ⇒ μ1 = −0.034677 = −3.4677%
316 | Chapter 11: The Valuation of Interest Rate Options

Therefore,

r2,1 = 8.5063 − 3.4677 × 0.50 = 6.7725%

r2,2 = 6.5063 − 3.4677 × 0.50 = 4.7725%
r2,3 = 4.5063 − 3.4677 × 0.50 = 2.7725%

The prices of the pure securities are

q3,1 = $0.112320
q3,2 = $0.340258
q3,3 = $0.343589
q3,4 = $0.115651

Finally, let’s proceed to the penultimate point in time:

r3,1 = r2,1 + μ2 × 0.50 + 1 = 6.7725 + 1 + μ2 × 0.50 = 7.7725 + μ2 × 0.50

r3,2 = r2,1 + μ2 × 0.50 − 1 = 6.7725 − 1 + μ2 × 0.50 = 5.7725 + μ2 × 0.50
r3,3 = r2,2 + μ2 × 0.50 − 1 = 4.7725 − 1 + μ2 × 0.50 = 3.7725 + μ2 × 0.50
r3,4 = r2,3 + μ2 × 0.50 − 1 = 2.7725 − 1 + μ2 × 0.50 = 1.7725 + μ2 × 0.50

The prices of the pure securities at this point in time can be expressed as
0.50 0.112320
q4,1 = q3,1 × =
(1 +
0.077725+μ2 ×0.50
) 2.077725 + μ2 × 0.50
2
0.50 0.50
q4,2 = q3,1 × 0.077725+μ2 ×0.50
+ q3,2 × 0.057725+μ2 ×0.50
(1 + 2
) (1 + 2
)
0.112320 0.340258
= +
2.077725 + μ2 × 0.50 2.057725 + μ2 × 0.50
0.50 0.50
q4,3 = q3,2 × 0.057725+μ2 ×0.50
+ q3,3 × 0.037725+μ2 ×0.50
(1 + 2
) (1 + 2
)
0.340258 0.343589
= +
2.057725 + μ2 × 0.50 2.037725 + μ2 × 0.50
0.50 0.50
q4,4 = q3,3 × 0.037725+μ2 ×0.50
+ q3,4 × 0.017725+μ2 ×0.50
(1 + 2
) (1 + 2
)
0.343589 0.115651
= +
2.037725 + μ2 × 0.50 2.017725 + μ2 × 0.50
0.50 0.115651
q4,5 = q3,4 × =
(1 +
0.017725+μ2 ×0.50
) 2.017725 + μ2 × 0.50
2

Obviously,
1 1
q4,1 + q4,2 + q4,3 + q4,4 + q4,5 = 4
= = $0.875665
(1 + 0.5 × s4 ) (1.03375)4
Calibrating the Ho-Lee Model | 317

Thus,

0.112320 0.340258
2× +2×
2.077725 + μ2 × 0.50 2.057725 + μ2 × 0.50
0.343589 0.115651
+2× +2× = 0.875665
2.037725 + μ2 × 0.50 2.017725 + μ2 × 0.50
⇒ μ2 = 0.070277 = 7.0277%

Therefore,

7.0277
r3,1 = 7.7725 + = 11.2863%
2
7.0277
r3,2 = 5.7725 + = 9.2863%
2
7.0277
r3,3 = 3.7725 + = 7.2863%
2
7.0277
r3,4 = 1.7725 + = 5.2863%
2

The prices of the pure securities are

q4,1 = $0.053160
q4,2 = $0.215741
q4,3 = $0.328336
q4,4 = $0.222092
q4,5 = $0.056336

Table 11.2: Summary of the Ho-Lee model-based interest rates.

Period Node Interest

Rate

1 r1,1 7.5063%
1 r1,2 5.5063%
2 r2,1 6.7725%
2 r2,2 4.7725%
2 r2,3 2.7725%
3 r3,1 11.2863%
3 r3,2 9.2863%
3 r3,3 7.2863%
3 r3,4 5.2863%

The no-arbitrage interest rate tree we derived is shown in Figure 11.5.

The state price tree is shown in Figure 11.6.
318 | Chapter 11: The Valuation of Interest Rate Options

Figure 11.5: No-arbitrage interest rate tree.

Note: Readers will observe that in the figures depicting the short rates, the nodes are
linked by straight lines. This is because interest rates span a period of time. However,
in the figures depicting the state prices, that follow, there are no links between the
nodes. This is because state prices represent a point in time. This mode of depiction is
intentional and should not be construed as an error.

t0 t1 t2 t3 t4

0.053160

0.112320

0.232247 0.215741

0.481928 0.340258

1.0 0.466755 0.328336

0.481928 0.343589

0.234508 0.222092

0.115651

0.056336
Figure 11.6: State price tree.
Calibrating the Black-Derman-Toy Model | 319

Calibrating the Black-Derman-Toy Model

Let’s now illustrate how the BDT model can be fitted to the observed term structure
using the binomial tree approach. Table 11.3 gives the term structure of interest rates
and also the local volatility, or the volatility of short rates at each point in time.

Table 11.3: Vector of spot rates and local volatilities.

Period Spot Rate σ √△t

0 7.50%
1 7.00% 6.00%
2 6.25% 4.50%
3 6.75% 3.00%

The stochastic process for the BDT model is

d ln r = μt dt + σt dZ

Consider time point tk , where tk = k △ t, and let rk be the short rate at this point in
time. For the process

d ln r = μt dt + σt dZ

define z(r) = ln(r) We can approximate d ln(r) as ln(rk+1,j ) − ln(rk,j ) or as ln(rk+1,j+1 ) −

ln(rk,j ), and write

ln(rk+1,j ) − ln(rk,j ) = μk △ t + σk+1 √△t and

ln(rk+1,j+1 ) − ln(rk,j ) = μk △ t − σk+1 √△t
⇒ z1,1 = z0,1 + μ0 △ t + σ1 √△t and
⇒ z1,2 = z0,1 + μ0 △ t − σ1 √△t and
⇒ z1,2 = z1,1 − 2σ1 √△t

Similarly,

z2,1 = z1,1 + μ1 △ t + σ2 √△t and

z2,2 = z1,1 + μ1 △ t − σ2 √△t = z1,2 + μ1 △ t + σ2 √△t
⇒ z2,2 = z2,1 − 2σ2 √△t

Similarly,

z2,3 = z1,2 + μ1 △ t − σ2 √△t

320 | Chapter 11: The Valuation of Interest Rate Options

⇒ z2,3 = z2,2 − 2σ2 √△t

= z2,1 − 4σ2 √△t

Similarly,

z3,1 = z2,1 + μ2 △ t + σ3 √△t and

z3,2 = z2,1 + μ2 △ t − σ3 √△t = z2,2 + μ2 △ t + σ3 √△t
⇒ z3,2 = z3,1 − 2σ3 √△t

Similarly,

z3,3 = z2,2 + μ2 △ t − σ3 √△t = z2,3 + μ2 △ t + σ3 √△t

⇒ z3,3 = z3,2 − 2σ3 √△t
⇒ z3,3 = z3,1 − 4σ3 √△t
z3,4 = z2,3 + μ2 △ t − σ3 √△t
⇒ z3,4 = z3,3 − 2σ3 √△t
z3,4 = z3,1 − 6σ3 √△t

In general,

zk+1,j+1 = zk+1,1 − 2 × jσk+1 √△t

rk+1,j
⇒ ln ( ) = μk △ t + σk+1 √△t and
rk,j
rk+1,j+1
ln ( ) = μk △ t − σk+1 √△t
rk,j
rk+1,j
⇒ ( ) = exp[μk △ t + σk+1 √△t]
rk,j
⇒ rk+1,j = rk,j × exp[μk △ t + σk+1 √△t] and
⇒ rk+1,j+1 = rk,j × exp[μk △ t − σk+1 √△t]

An Issue with Recombination

Consider state (2,2). There are two paths, namely an up move from (0,1), followed by a
down move from (1,1). Or a down move from (0,1), followed by an up move from (1,2).

z1,1 = z0,1 + μ0 △ t + σ1 √△t

z2,2 = z1,1 + μ1 △ t − σ2 √△t
⇒ z2,2 = z0,1 + [μ0 + μ1 ] △ t + [σ1 − σ2 ]√△t
An Issue with Recombination | 321

z1,2 = z0,1 + μ0 △ t − σ1 √△t

z2,2 = z1,2 + μ1 △ t + σ2 √△t
⇒ z2,2 = z0,1 + [μ0 + μ1 ] △ t − [σ1 − σ2 ]√△t

For the tree to recombine, we require that

z0,1 + [μ0 + μ1 ] △ t + [σ1 − σ2 ]√△t = z0,1 + [μ0 + μ1 ] △ t − [σ1 − σ2 ]√△t

⇒ σ1 = σ2

This is inconsistent with our assumption of time varying volatility. The issue can be
resolved by noticing that

zk+1,j+1 = zk+1,1 − 2 × jσk+1 √△t

Thus all we need to do is estimate zk+1,1 , and we can derive all the other rates at time
(k + 1) using the given value of sigma. Thus instead of estimating the µs as we did in
the case of the Ho-Lee model, we estimate zk+1,1 , or equivalently rk+1,1 , and derive the
rates at the other nodes at that point in time from the estimate.
The one-period spot rate is 7.50%. The corresponding volatility at time one is
6/△t%. The variable to be estimated is r1,1 .

r1,2 = r1,1 × exp[−2σ1 √△t]

= r1,1 × exp(−2 × 0.06) = r1,1 × exp(−0.12)
0.50
q1,1 = q1,2 = = 0.481928
(1 + 0.075
2
)

The prices of the pure securities for the next period are computed as follows:

0.50 0.481928
q2,1 = q1,1 × r =
(1 + 1,1 ) (2 + r1,1 )
2
0.50 0.50
q2,2 = q1,1 × r1,1 + q1,2 × r
(1 + 2
) (1 + 1,2
2
)
0.481928 0.481928
= +
(2 + r1,1 ) [2 + r1,1 × exp(−0.12)]
0.50
q2,3 = q1,2 × r
(1 + 1,22
)
1
= 0.481928 ×
[2 + r1,1 × exp(−0.12)]

Obviously,

0.481928 0.481928
q2,1 + q2,2 + q2,3 = 0.933511 = +
(2 + r1,1 ) (2 + r1,1 )
322 | Chapter 11: The Valuation of Interest Rate Options

0.481928 0.481928
+ +
[2 + r1,1 × exp(−0.12)] [2 + r1,1 × exp(−0.12)]
2 × 0.481928 2 × 0.481928
= +
(2 + r1,1 ) [2 + r1,1 × exp(−.12)]

The solution is r1,1 = 6.8915%.

Thus,

r1,2 = 6.8915 × exp(−0.12) = 6.1122%

The state prices are

q2,1 = $0.232938
q2,2 = $0.466756
q2,3 = $0.233818

In the next period, the variable to be estimated is r2,1 :

r2,2 = r2,1 × exp[−2σ2 √△t]

= r2,1 × exp(−2 × 0.045) = r2,1 × exp(−0.09)
r2,3 = r2,1 × exp[−4σ2 √△t]
= r2,1 × exp(−4 × 0.045) = r2,1 × exp(−0.18)

The prices of pure securities for the next point in time may be expressed as follows:

0.50 0.232938
q3,1 = q2,1 × r =
(1 + 2,1 ) (2 + r2,1 )
2
0.50 0.50
q3,2 = q2,1 × r2,1 + q2,2 × r2,2
(1 + 2
) (1 + 2
)
0.232938 0.466756
= +
(2 + r2,1 ) [2 + r2,1 × exp(−0.09)]
0.50 0.50
q3,3 = q2,2 × r + q2,3 × r
(1 + 2,2 2
) (1 + 2,3
2
)
0.466756 0.233818
= +
[2 + r2,1 × exp(−0.09)] [2 + r2,1 × exp(−0.18)]
0.50 0.233818
q3,4 = q2,3 × r2,3 =
(1 + ) [2 + r2,1 × exp(−0.18)]
2

Obviously,

1 1
q3,1 + q3,2 + q3,3 + q3,4 = 3
= == $0.911818
(1 + 0.5 × s3 ) (1.03125)3
An Issue with Recombination | 323

Thus,
0.232938 0.466756 0.233818
2× +2× +2× = 0.911818
(2 + r2,1 ) [2 + r2,1 × exp(−0.09)] [2 + r2,1 × exp(−0.18)]
⇒ r2,1 = 5.1969%

Therefore,

r2,2 = 4.7497%
r2,3 = 4.3409%

The prices of the pure securities are

q3,1 = $0.113519
q3,2 = $0.341483
q3,3 = $0.342390
q3,4 = $0.114425

Finally, let’s proceed to the penultimate point in time. The variable to be estimated
is r3,1 :

r3,2 = r3,1 × exp[−2σ3 √△t]

= r3,1 × exp(−2 × 0.03) = r3,1 × exp(−0.06)
r3,3 = r3,1 × exp[−4σ3 √△t]
= r3,1 × exp(−4 × 0.03) = r3,1 × exp(−0.12)
r3,4 = r3,1 × exp[−6σ3 √△t]
= r3,1 × exp(−6 × 0.03) = r3,1 × exp(−0.18)

The prices of pure securities for the next point in time may be expressed as follows:
0.50 0.113519
q4,1 = q3,1 × r3,1 =
(1 + ) (2 + r3,1 )
2
0.50 0.50
q4,2 = q3,1 × r3,1 + q3,2 × r
(1 + 2
) (1 + 3,2
2
)
0.113519 0.341483
= +
(2 + r3,1 ) [2 + r3,1 × exp(−0.06)]
0.50 0.50
q4,3 = q3,2 × r3,2 + q3,3 × r
(1 + 2 ) (1 + 3,32
)
0.341483 0.342390
= +
[2 + r3,1 × exp(−0.06)] [2 + r3,1 × exp(−0.12)]
0.50 0.50
q4,4 = q3,3 × r + q3,4 × r
(1 + 3,32
) (1 + 3,42
)
324 | Chapter 11: The Valuation of Interest Rate Options

0.342390 0.114425
= +
[2 + r3,1 × exp(−0.12)] [2 + r3,1 × exp(−0.18)]
0.114425
q4,5 =
[2 + r3,1 × exp(−0.18)]

Obviously,

1 1
q4,1 + q4,2 + q4,3 + q4,4 + q4,5 = 4
= = $0.875665
(1 + 0.5 × s4 ) (1.03375)4

Thus,
0.113519 0.341483
2× +2×
(2 + r3,1 ) (2 + r3,1 × exp(−0.06))
0.342390 0.114425
+2× +2× = 0.875665
(2 + r3,1 × exp(−0.12)) (2 + r3,1 × exp(−0.18))
⇒ r3,1 = 9.0245%

Therefore,

r3,2 = 8.4990%
r3,3 = 8.0040%
r3,4 = 7.5379%

The prices of the pure securities are

q4,1 = $0.054309
q4,2 = $0.218091
q4,3 = $0.328389
q4,4 = $0.219742
q4,5 = $0.055135

The no-arbitrage interest rate tree we derive is shown in Figure 11.7.

The state price tree is shown in Figure 11.8.

An Issue with Calibration

If we assume that the local volatilities are given in order to calibrate the model, it is
necessary that all the forward rates be positive to ensure that the short rates are not
negative. In our case, all the forward rates are positive, and consequently there is no
issue on this score.
At time zero, the forward rates are 6.5012% for a one-period loan after one period;
5.6278% for a two-period loan after one period; and 6.5006% for a three-period loan
An Issue with Calibration | 325

Table 11.4: Summary of BDT model-based interest rates.

Period Node Interest

Rate

1 r1,1 6.8915%
1 r1,2 6.1122%
2 r2,1 5.1969%
2 r2,2 4.7497%
2 r2,3 4.3409%
3 r3,1 9.0245%
3 r3,2 8.4990%
3 r3,3 8.0040%
3 r3,4 7.5379%

Figure 11.7: No-arbitrage interest rate tree for the BDT model.

after one period, where all the rates are annualized. The forward rate for a one-period
loan after two periods is 4.7581% and 6.5003% for a two-period loan after two periods.
Finally, the rate for a one-period loan after three periods is 8.2573% per annum. Thus,
it is not surprising that all the short rates that we have generated are positive.
326 | Chapter 11: The Valuation of Interest Rate Options

t0 t1 t2 t3 t4

0.054309

0.113519

0.232938 0.218091

0.481928 0.341483

1.0 0.466756 0.328389

0.481928 0.342390

0.233818 0.219742

0.114425

0.055135

Figure 11.8: State price tree for the BDT model.

Valuation of a Plain Vanilla Bond

1, 040 × (0.053160 + 0.215741 + 0.328336 + 0.222092 + 0.056336)

+ 40 × (0.112320 + 0.340258 + 0.343589 + 0.115651)
+ 40 × (0.232247 + 0.466755 + 0.234508) + 40 × (0.481928 + 0.481928)
= 1, 040 × 0.875665 + 40 × 0.911818 + 40 × 0.933511 + 40 × 0.963856
= $1,023.0589

The value will be the same if we use the state prices obtained from the Black-Derman-
Toy model. This is because at a point in time, the sum of the state prices is the same
irrespective of the model being used, as the computation is based on the same vector
of spot rates. Another way to understand this is as follows. The bond price is given by

40 40
+
(1 + s21 ) (1 + s2 )2
2
40 1,040
+ +
s3 3 s 4
(1 + 2 ) (1 + 24 )
= 38.5542 + 37.3404 + 36.4727 + 910.6914 = $1,023.0587

Because both the models have been calibrated using the same vector of spot rates,
they give the same bond price.
Valuation of a Zero Coupon Bond | 327

Valuation of a Zero Coupon Bond

Consider a zero coupon bond that pays $1,000 at time t4 . Its value at various nodes of
the tree can be computed by working backwards. Let’s first consider the Ho-Lee-based
tree.
At t3 , its value at node (3,1) is given by

1,000
0.112863
= $946.5829
(1 + 2
)

At node (3,2), the value is given by

1,000
0.092863
= $955.6287
(1 + 2
)

Similarly at node (3,3), the value is

1,000
0.072863
= $964.8491
(1 + 2
)

Finally at node (3,4), the value is

1,000
0.052863
= $974.2491
(1 + 2
)

Working backwards we can calculate the prices at time t2 as follows. At node (2,1) the
value is

946.5829 955.6287
0.5 × 0.067725
+ 0.5 × 0.067725
= $919.9539
(1 + 2
) (1 + 2
)

Similarly at node (2,2), the value is

955.6287 964.8491
0.5 × 0.047725
+ 0.5 × 0.047725
= $937.8592
(1 + 2
) (1 + 2
)

Finally at node (2,3), the value is

964.8491 974.2491
0.5 × 0.027725
+ 0.5 × 0.027725
= $956.2925
(1 + 2
) (1 + 2
)

Using the same logic, the value at (1,1) is

919.9539 937.8592
0.5 × 0.075063
+ 0.5 × 0.075063
= $895.3044
(1 + 2
) (1 + 2
)
328 | Chapter 11: The Valuation of Interest Rate Options

and the value at (1,2) is

937.8592 956.2925
0.5 × 0.055063
+ 0.5 × 0.055063
= $921.7001
(1 + 2
) (1 + 2
)

At the starting node, the value is therefore

895.3044 921.7001
0.5 × 0.075
+ 0.5 × 0.075
= $875.6648
(1 + 2
) (1 + 2
)

The bond prices may be depicted in a tree format as in Figure 11.9.

t0 t1 t2 t3

946.5829

919.9539

895.3044 955.6287

875.6648 937.8592

921.7001 964.8491

956.2925

974.2491

Figure 11.9: Evolution of the price of a zero coupon bond: The Ho-Lee tree.

Now reconsider a zero coupon bond that pays $1,000 at time t4 , and consider the BDT
tree. At t3 , its value at node (3,1) is given by
1,000
0.090245
= $956.8256
(1 + 2
)

At node (3,2) the value is given by

1,000
0.084990
= $959.2372
(1 + 2
)

Similarly at node (3,3), the value is

1,000
0.08004
= $961.5200
(1 + 2
)

Finally at node (3,4), the value is

1,000
0.075379
= $963.6794
(1 + 2
)
Valuation of a Zero Coupon Bond | 329

Working backwards we can calculate the prices at time t2 as follows. At node (2,1) the
value is
956.8256 959.2372
0.5 × 0.051969
+ 0.5 × 0.051969
= $933.7679
(1 + 2
) (1 + 2
)

Similarly at node (2,2), the value is

959.2372 961.5200
0.5 × 0.047497
+ 0.5 × 0.047497
= $938.1001
(1 + 2
) (1 + 2
)

Finally at node (2,3), the value is

961.5200 963.6794
0.5 × 0.043409
+ 0.5 × 0.043409
= $942.1508
(1 + 2
) (1 + 2
)

Using the same logic, the value at (1,1) is

933.7679 938.1001
0.5 × 0.068915
+ 0.5 × 0.068915
= $904.7583
(1 + 2
) (1 + 2
)

while the value at (1,2) is

938.1001 942.1508
0.5 × 0.061122
+ 0.5 × 0.061122
= $912.2463
(1 + 2
) (1 + 2
)

At the starting node, the value is therefore

904.7583 912.2463
0.5 × 0.075
+ 0.5 × 0.075
= $875.6649
(1 + 2
) (1 + 2
)

The bond prices may be depicted in a tree format as in Figure 11.10.

t0 t1 t2 t3

956.8256
933.7679
904.7583 959.2372
875.6649 938.1001
912.2463 961.5200
942.1508
963.6794

Figure 11.10: Evolution of the price of a zero coupon bond: The BDT tree.
330 | Chapter 11: The Valuation of Interest Rate Options

The price of the bond is the same at time t0 , although it differs from model to model
at subsequent nodes. This is because the price at t0 is

1,000 1,000
= = $875.6649
s4 4 (1.03375)4
(1 + 2
)

Valuing a European Call

Consider a two-period European call on the zero coupon bond with an exercise price
of $925. Let’s first value it using the Ho-Lee model. The payoff is zero in node (2,1),
$12.8592 at node (2,2), and $31.2925 at node (2,3). The value of the option at t0 is

0.466755 × 12.8592 + 0.234508 × 31.2925 = $13.3404

Now let’s value it using the BDT model. The payoff is 8.7679 in node (2,1), $13.1001
at node (2,2), and $17.1508 at node (2,3). The value of the option at t0 is

0.232938 × 8.7679 + 0.466756 × 13.1001 + 0.233818 × 17.1508 = $12.1671

Valuing an American Put

Valuing an American option is more complex because we have to check for the pos-
sibility of early exercise at every node. Consider a two-period put on the zero coupon
bond with an exercise price of $930. Let’s look at the Ho-Lee model first. The payoffs
from the option at time t2 are $10.0461 at node (2,1) and zero at nodes (2,2) and (2,3).
The model value at (1,1) is

10.0461
.5 × 0.075063
= $4.8413
(1 + 2
)

The payoff from early exercise is $34.6956. Thus the option will be exercised early.
The model value at node (1,2) is zero. The payoff from early exercise is $8.2999.
Thus the option will be exercised early.
The model value at (0,1) is

34.6956 8.2999
.5 × 0.075
+ .5 × 0.075
= $20.7207
(1 + 2
) (1 + 2
)

The payoff from early exercise is $54.3352. So the option will be exercised right at
the outset, that is, at time t0 .
Now let’s turn to the BDT model. The payoffs from the option at time t2 is zero at
all three nodes.
Caps, Floors, and Collars | 331

The model value at (1,1) is zero. The payoff from early exercise is $25.2417. Thus the
option will be exercised early.
The model value at node (1,2) is zero. The payoff from early exercise is $17.7537.
Thus the option will be exercised early.
The model value at (0,1) is

25.2417 17.7537
.5 × 0.075
+ .5 × 0.075
= $20.7207
(1 + 2
) (1 + 2
)

The payoff from early exercise is $54.3351. So the option will be exercised right at the
outset, that is, at time t0 .
Let us take a closer look at this result. In the case of both the models, the zero
coupon bond price at each node is less than the prices of the two nodes at the next
point in time, from which it is derived. Thus the intrinsic value at every node is greater
than or equal to the model value that is obtained from the subsequent point in time.
Thus the American put will always be exercised at time zero if it is in the money, that is,
the exercise price is higher than the price at that time. If however the exercise price is
lower than the price at time zero, then the intrinsic and the model values at all points
in time are zero, and consequently the option has a zero premium at the outset.

Caps, Floors, and Collars

Before we go on to analyze caps and floors, let’s first define caplets and floorlets.
A caplet is a call option on an interest rate, whereas a floorlet is a put option on an in-
terest rate. To evaluate these options, we first reproduce a segment of the no-arbitrage
interest rate tree that we derived earlier.

Figure 11.11: Segment of the Ho-Lee no-arbitrage interest rate tree.

332 | Chapter 11: The Valuation of Interest Rate Options

Consider a one-period caplet with an exercise price of 5%. Let the underlying principal
be $1,000. Each time period is assumed to be of six months duration, and we take it
as 0.5 years in order to avoid issues regarding day-count conventions.3
At time t1 if the node (1,1) is attained, the call is in the money because the interest
rate of 7.5063% is greater than the exercise price of 5%. The payoff is

1,000 × (0.075063 − 0.05) × 0.5 = $12.5315

In the case of interest rate options, the payoff occurs not when the option expires, but
at a point in time when the next interest payment is due. In the preceding case, the rate
as determined at t1 is applicable for computing the interest due at t2 . Consequently the
caplet pays off at t2 . Thus the value at t1 is the present value of the payoff. In this case,
the value is
12.5315
0.075063
= $12.0782
(1 + 2
)

If we attain node (1,2) at time t1 , the caplet once again is in the money. The payoff is

1,000 × (0.055063 − 0.05) × 0.5 = $2.5315

The value at t1 is the present value of the payoff. In this case, the value is
2.5315
0.055063
= $2.4637
(1 + 2
)

Thus the value of the caplet at time t0 is computed as

12.5315 2.5315
0.5 × 0.075063
+ 0.5 × 0.055063
(1.0375) (1 + 2
) (1.0375) (1 + 2
)
= 5.8208 + 1.1873 = $7.0081

Now consider a one-period floorlet with an exercise price of 7% and a principal of

$1,000. At node (1,1), the option is out of the money because the interest rate is greater
than the exercise price. However, at node (1,2), the option is in the money. The payoff
is

1,000 × (0.07 − 0.055063) × 0.5 = $7.4685

Thus the value of the floorlet at t0 is computed as

7.4685
0.5 × 0.055063
= $3.5028
(1.0375) (1 + 2
)

3 In practice, the length of the time period has to be calculated per the day-count convention that is
applicable in the market in question.
Caps, Floors, and Collars | 333

Caps and Floors: A Detailed Perspective

A cap is a portfolio of caplets that can be acquired by a borrower who has taken advan-
tage of a floating rate loan, to protect against an increase in interest rates. Similarly, a
floor is a portfolio of floorlets that can be acquired by a lender who has made a loan
on a floating rate basis to protect against a decline in interest rates. Caps and floors
are cash settled. If a caplet or floorlet is in the money, the writer makes a payment to
the holder. The principal that is used to compute the payoff is termed as a notional
principal, for it is specified purely to facilitate the computation of the payoff and is not
intended to be paid or received.

Required Symbols
– P ≡ notional principal.
– K ≡ contract rate or the exercise price.
– R ≡ reference rate used to compute the payoff, like the LIBOR.
– Ni ≡ the number of days in the time period from the exercise date until the pay-
ment date.
– N ≡ the number of days in the calendar year. In the US and the EU, it is taken as
360, whereas in the UK, it is taken as 365.

The payment, if the caplet or floorlet is in the money, is usually made in arrears. Thus
with reference to Figure 11.9, if the caplet or floorlet is in the money at a node at time
t1 , the payment will occur at time t2 , which we assume, is six months hence. Conse-
quently, there is no need to discount the payoff, as would be required if the payment
were made at time t1 itself.

Example 11.1. Consider a notional principal of $1,000,000. The reference rate is the 6-M LIBOR, and
the contract rate is 5%. The day-count convention is Actual/360. Assume that the number of days in
the next semiannual period is 184. If the reference rate is 6.50%, the payoff after six months is

(6.50 − 5.00) 184

1,000,000 × ×
100 360
= $7,666.67

Example 11.2. Consider the same data as in the previous example. Consider a floor with a contract
rate of 7.50%. If reference rate is 5%, the payoff after six months is

(7.50 − 5.00) 184

1,000,000 × ×
100 360
= $12,777.78

Using the interest rate tree that we derived earlier (Figure 11.5), let’s price a three-
period cap and a three-period floor. The cap consists of three caplets, expiring after
334 | Chapter 11: The Valuation of Interest Rate Options

one, two, and three periods respectively. The value of the one-period caplet is $7.0081.
The value of the two-period caplet is calculated as follows.
The payoff if state (2,1) is attained is

1,000 × [0.067725 − 0.05] × 0.5 = $8.8625

The payoff in states (2,2) and (2,3) is zero. The value at time t0 of the payoff is
1 1 8.8625
0.5 × 0.5 × × ×
1.0375 (1 + 0.075063 ) (1 + 0.067725 )
2 2
= $1.9909

Thus the value of the caplet = $1.9909.

Similarly, the value of a three-period caplet is given by
1 1
0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.075063
)
2
1 1
× 0.067725
× 0.112863
× 1,000 × [0.112863 − 0.05] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.075063 )
2
1 1
× 0.067725
× 0.092863
× 1,000 × [0.092863 − 0.05] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.075063 )
2
1 1
× 0.047725
× 0.092863
× 1,000 × [0.092863 − 0.05] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.055063 )
2
1 1
× 0.047725
× 0.092863
× 1,000 × [0.092863 − 0.05] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.075063 )
2
1 1
× 0.047725
× 0.072863
× 1,000 × [0.072863 − 0.05] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.055063 )
2
1 1
× 0.047725
× 0.072863
× 1,000 × [0.072863 − 0.05] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.055063 )
2
Caps, Floors, and Collars | 335

1 1
× 0.027725
× 0.072863
× 1,000 × [0.072863 − 0.05] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.055063 )
2
1 1
× 0.027725
× 0.052863
× 1,000 × [0.052863 − 0.05] × 0.5
(1 + 2
) (1 + 2
)
= 3.3418 + 2.3004 + 2.3228 + 2.3455 + 1.2510 + 1.2631 + 1.2756 + 0.1613 = $14.2615

Thus the value of a three-period cap is

$7.0081 + $1.9909 + $14.2615 = $23.2605

The value of a three-period floor is computed in a similar fashion. Let’s consider a

one-period floorlet with an exercise price of 6.4620% and an underlying principal of
$1,000. The payoff is zero in state (1,1). However, there is a positive payoff at state (1,2),
that is given by

1,000 × [0.06462 − 0.055063] × 0.5 = $4.7785

This payoff obviously occurs at time t2 . Consequently, the value of the floorlet at t0 is

4.7785
0.5 × 0.055063
= $2.2412
(1.0375) (1 + 2
)

The value of the two-period floorlet is computed as follows. The payoff in state
(2,1) is zero. The payoff in state (2,2) is

1,000 × [0.06462 − 0.047725] × 0.5 = $8.4475

and the payoff in state (2,3) is

1,000 × [0.06462 − 0.027725] × 0.5 = $18.4475

The floorlet is valued as

1 1
0.5 × 0.5 × ×
1.0375 (1 + 0.075063 )
2
1
× 0.047725
× 1,000 × [0.06462 − 0.047725] × 0.5
(1 + 2
)
1 1
+ 0.5 × 0.5 × ×
1.0375 (1 + 0.055063 )
2
1
× 0.047725
× 1,000 × [0.06462 − 0.047725] × 0.5
(1 + 2
)
336 | Chapter 11: The Valuation of Interest Rate Options

1 1
+ 0.5 × 0.5 × ×
1.0375 (1 + 0.055063
)
2
1
× 0.027725
× 1,000 × [0.06462 − 0.027725] × 0.5
(1 + 2
)
= 1.9162 + 1.9348 + 4.2669 = $8.1179

Similarly, the value of a three-period floorlet is given by

1 1
0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.055063
)
2
1 1
× 0.027725
× 0.052863
× 1,000 × [0.06462 − 0.052863] × 0.5
(1 + 2
) (1 + 2
)
= $0.6623

Thus the value of a three-period floor is

$2.2412 + $8.1179 + $0.6623 = $11.0215

Interest Rate Collars

An interest rate collar is a combination of a cap and a floor. It requires the investor to
take a long position in a cap and a short position in a floor, or vice versa. A borrower
takes a long position in the collar by buying the cap and selling the floor. This puts
a maximum and a minimum on the borrowing rate. A lender takes a short position
in a collar by selling the cap and buying the floor. This too puts a maximum and a
minimum on the lending rate.
Let us first consider the case of a borrower. If the rates rise, the borrower exercises
the cap and the counterparty does not exercise the floor. Thus there is an upper limit
on the borrowing rate. However, if rates decline, the borrower does not exercise the
cap, and the counterparty exercises the floor. Hence, although the rates have moved
in the borrower’s favor, he still has to pay a minimum rate that is equivalent to the
contract rate of the floor.
Now consider the case of a lender. If rates rise, the lender does not exercise the
floor, whereas the counterparty does exercise the cap. This puts an upper limit on the
lending rate, equivalent to the contract rate of the cap, which means that the lender
cannot take full benefit of the higher rate. However, if rates decline, the lender exer-
cises the floor, and the counterparty does not exercise the cap. Consequently there is
a lower limit on the lending rate.

Example 11.3. Consider a long collar with a notional principal of $1,000,000. The contract rate is 6%
for the cap and 4.50% for the floor. The number of days in the semiannual period is 184, and the day-
count convention is Actual/360.
Caps, Floors, and Collars | 337

If the LIBOR on the exercise date is 7.50%, the holder of the collar exercises the cap. The writer
of the collar, who is the holder of the floor, does not exercise. The payoff is

(7.50 − 6.00) 184

1,000,000 × × = $7,666.67
100 360

The actual interest paid is

184
1,000,000 × 0.075 × = $38,333.33
360

The effective interest paid is 38,333.33 − 7,666.67 = $30,666.67. This translates to a rate of

30,666.67 360
× × 100 = 6.00%
1,000,000 184

If the LIBOR on the exercise date is 3%, the counterparty exercises the floor. The holder of the
cap obviously does not exercise it. The payoff is

(4.50 − 3.00) 184

−1,000,000 × × = −$7,666.67
100 360

The actual interest paid is

184
1,000,000 × 0.03 × = $15,333.33
360

The effective interest paid is 15,333.33 + 7,666.67 = $23,000.00. This translates to a rate of

23,000.00 360
× × 100 = 4.50%
1,000,000 184

Thus the maximum interest payable is 6%, which is the exercise price of the cap, and the minimum
interest payable is 4.50%, which is the exercise price of the floor. Thus, although protected if rates rise
above 6%, the holder of the collar cannot take advantage of the situation if the rate declines below
4.50%.
If the rate is between 6% and 4.5%, neither the cap nor the floor are exercised, and the borrower
pays the prevailing market rate. For instance, if the rate is 5.25% per annum, the interest paid is

184
1,000,000 × 0.0525 × = $26,833.33
360

which is between the upper bound of $30,666.67 and the lower bound of $23,000.00.

Valuation of a Zero Cost Collar

A zero cost collar is a position where the premium of the cap is equal to that of the floor,
and as a consequence the premium paid for the collar is zero and hence the name.
Using the interest rate tree that we derived previously (Figure 11.5), let’s price a
three-period cap and a three-period floor. The cap consists of three caplets, expiring
after one, two, and three periods, respectively, and the floor consists of three floor-
lets. Assume the exercise price for the caplets is 6.50% per annum and the underlying
principal is $1,000.
338 | Chapter 11: The Valuation of Interest Rate Options

At time t1 , if the node (1,1) is attained, the call is in the money, and the payoff is

1,000 × (0.075063 − 0.065) × 0.5 = $5.0315

If we attain node (1,2) at time t1 , the caplet is out of the money. Thus the value of the
caplet at time t0 is computed as

5.0315
0.5 × 0.075063
= $ 2.3371
(1.0375) (1 + 2
)

The value of the two-period caplet is calculated as follows. The payoff if state (2,1) is
attained is

1,000 × [0.067725 − 0.065] × 0.5 = $1.3625

The payoff in states (2,2) and (2,3) is zero. The value at time t0 of the payoff is

1 1 1.3625
0.5 × 0.5 × × ×
1.0375 (1 + 0.075063 ) (1 + 0.067725 )
2 2
= $0.3061

Similarly, the value of a three-period caplet is given by

1 1
0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.075063 )
2
1 1
× 0.067725
× 0.112863
× 1,000 × [0.112863 − 0.065] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.075063 )
2
1 1
× 0.067725
× 0.092863
× 1,000 × [0.092863 − 0.065] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.075063 )
2
1 1
× 0.047725
× 0.092863
× 1,000 × [0.092863 − 0.065] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.055063 )
2
1 1
× 0.047725
× 0.092863
× 1,000 × [0.092863 − 0.065] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.075063 )
2
Captions and Floortions | 339

1 1
× 0.047725
× 0.072863
× 1,000 × [0.072863 − 0.065] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.055063 )
2
1 1
× 0.047725
× 0.072863
× 1,000 × [0.072863 − 0.065] × 0.5
(1 + 2
) (1 + 2
)
1 1
+ 0.5 × 0.5 × 0.5 × ×
1.0375 (1 + 0.055063 )
2
1 1
× 0.027725
× 0.072863
× 1,000 × [0.072863 − 0.065] × 0.5
(1 + 2
) (1 + 2
)
= 2.5444 + 1.4954 + 1.5100 + 1.5247 + 0.4302 + 0.4344 + 0.4387 = $8.3777

Thus the value of a three-period cap is

$2.3371 + $0.3061 + $8.3777 = $11.0209

The premium of the three-period floor with a contract rate of 6.4620% and a princi-
pal of $1,000 is $11.0215. If this collar is bought, the investor has to pay a maximum
rate of 6.50% per annum and a minimum rate of 6.4620% per annum, which virtually
amounts to a fixed rate loan because the difference is only 3.8 basis points. The cost
of the strategy is 11.0215 − 11.0209 = $0.0006, which is essentially a cost of zero.

Captions and Floortions

A caption is a financial instrument that gives the holder the right to buy a cap at a
future point in time. Take the case of a firm that may borrow funds at a later date. If it
decides to avail itself of the loan, it becomes concerned about the prospect of a rising
interest rate and consequently may decide to acquire the cap. However, if it decides
not to avail of the loan, there is no need to acquire the cap. The caption permits it to
set the contract rate in advance for the cap, while giving it the freedom to refrain from
exercise if the cap is not required. If the contract rate in the option is X and prevailing
contract rate in the future is K, the caption is exercised if K > X. Otherwise, the buyer
can allow the caption to expire and acquire a cap at the prevailing contract rate if
required.
Similarly, a floortion gives the holder the right to buy a floor at a future point in
time. An investor who may make a loan at a later date would be interested in such
an instrument. If a party decides to extend the loan, it will be perturbed about the
prospect of declining interest rates and consequently may exercise the floortion. How-
ever, if the party decides not to extend the loan, the option can be allowed to expire. If
the contract rate in the option is X, and prevailing contract rate in the future is K, the
floortion is exercised if X > K. Otherwise, the buyer can allow the floortion to expire
and acquire a floor at the prevailing contract rate if required.
340 | Chapter 11: The Valuation of Interest Rate Options

Chapter Summary
In this chapter we looked at interest rate options. The issues involved in building
models of the term structure were briefly discussed and the focus then shifted to
no-arbitrage models. We examined in detail both the Ho-Lee model and the Black-
Derman-Toy model. In this context, the concept of Arrow-Debreu securities was ex-
plained. We calibrated both the models and derived the short-rate tree. The trees
were used to value coupon paying bonds, zero coupon bonds, and options such as
caps, floors, and collars. The chapter concludes with a very brief introduction to cap-
tions and floortions. In the next chapter, we look at interest rate forward and futures
contracts.
Chapter 12
Interest Rate Forwards and Futures
In this chapter we first discuss forward contracts on interest rates, known as forward
rate agreements or FRAs. Subsequently we go on to look at short-term interest rate
(STIR) or money market futures, primarily Eurodollar futures, and long-term interest
rate futures, namely T-bond and T-note futures.

Forward Rate Agreements (FRAs)

Forward rate agreements are over-the-counter (OTC) derivatives, based on a short-term
interest rate such as LIBOR. They represent an agreement between two counter-parties
to fix a future interest rate. There is no physical delivery of cash. The contract specifies
a notional principal on the basis of which interest is computed, and the profit/loss is
settled at the end of the contract period.
On the contract settlement date, if the contract rate specified in the FRA differs
from the reference rate such as LIBOR, which is usually the case, one party makes a
cash payment to the other. For a potential borrower the FRA locks in a borrowing rate,
whereas for a potential lender it locks in a lending rate. It should be noted that there
is no obligation to borrow or lend at a subsequent date, and such a contract may be
used purely as a speculative tool.
The fixed interest rate in the contract, or what is termed the contract rate, is the
rate that is compared subsequently with the prevailing reference rate. The party who
agrees to pay the contract rate is termed the buyer, and the party who agrees to receive
the contract rate is termed the seller. As we mentioned, FRAs are used for hedging or
for speculation. Traders who are bullish about interest rates buy FRAs. If the reference
rate on the settlement date is higher than the contract rate, then they receive a cash
inflow. However, if the reference rate is lower, then they have to make a cash payment.
On the other hand, traders who are bearish about interest rates sell FRAs. If the ref-
erence rate on the settlement date is lower than the contract rate, then they receive
a cash inflow. However, if the reference rate is higher, then they have to make a cash
payment. Here is an example.

Example 12.1. Consider an FRA with a contract rate of 6% per annum. Assume that the reference rate
is 6-M LIBOR. If on the settlement date the prevailing LIBOR is 4.50%, the buyer makes a payment to
the seller. However if it is 7.50%, the seller makes a payment to the buyer.
Assume that the notional principal is $10,000,000 and that the day-count is 30/360. Thus six
months is 180 days. If LIBOR is 4.50%, the buyer of the FRA pays

180 6.00 − 4.50

10,000,000 × × = $75,000
360 100

https://doi.org/10.1515/9781547400669-012
342 | Chapter 12: Interest Rate Forwards and Futures

to the seller of the FRA. On the other hand if the LIBOR is 7.50%, the seller of the FRA pays

180 7.50 − 6.00

10,000,000 × × = $75,000
360 100

to the buyer.

Example 12.2. Now let’s consider a situation where a FRA is used as a hedge.
Metro Rail has borrowed on a floating rate basis and is consequently worried about the possibility
of rising interest rates. The interest payable is LIBOR + 50 bp, and interest is determined in advance
and paid in arrears. To lock in a borrowing rate, the firm buys a FRA with a contract rate of 4.75%.
Assume that the LIBOR on the settlement date is 5.50%. The firm borrows at 6.00% in the market.
However, because the reference rate is higher than the contract rate, the counter-party pays the cash
equivalent of the difference in rates, namely 5.50 − 4.75 = 0.75%, based on the specified principal
amount and the prescribed day-count convention. Thus the effective borrowing cost for the firm is
6.00 − 0.75 = 5.25%, which is the contract rate of 4.75% plus the spread of 50 bp.
However, what if the LIBOR on the settlement rate is 3.50%. The firm borrows at 4.00% in the
market. In this case, because the contract rate is higher than the reference rate, it has to pay the cash
equivalent of the difference in rates, namely 4.75−3.50 = 1.25% to the counter-party. Thus the effective
borrowing cost for the firm is 4.00 + 1.25 = 5.25%, which once again is the contract rate of 4.75% plus
the spread of 50 bp. Thus this is a perfect hedge, for it locks in a borrowing rate will absolute certainty.
In the next example, we will consider a situation where a party wants to hedge a lending rate.

Example 12.3. Jyske Bank in Copenhagen plans to lend at LIBOR + 50 bp after six months. However, it
is worried about a declining interest rate. It therefore sells a FRA with a contract rate of 6.50%. If the
LIBOR on the settlement date is 4%, the bank lends at 4.50%. However, because the reference rate is
less than the contract rate, the counter-party, who in this case is the buyer of the FRA, pays the cash
equivalent of 6.50 − 4.00 = 2.50%. Thus the effective lending rate for the bank is 2.50 + 4.50 = 7.00%,
which is the contract rate plus the spread. On the other hand, if the LIBOR on the settlement date
is 7.50%, the bank lends at 8%. In this case, it has to pay the equivalent of 7.50 − 6.50 = 1.00% to
the counter-party. The effective rate of interest is once again 7%, which is the contract rate plus the
spread.
Thus borrowers can lock in a borrowing rate using an FRA, whereas lenders can lock in a lending
rate. Because an FRA is a commitment contract, the hedgers cannot take advantage of the situation if
the market moves in their favor. In Example 12.2, if the rate falls to 3.50%, in the absence of the FRA
Metro Rail could have borrowed at 4% per annum. However, because it has an obligation under the
FRA, it is committed to borrow at 5.25%. Similarly, if the LIBOR were to rise to 7.50%, Jyske Bank could
have lent at 8% in the absence of the FRA. However, because of its prior obligation, it is committed
to lend at 7%. Thus FRAs lock in borrowing and lending rates with absolute certainty. However, there
is no assurance that the outcome with hedging will be superior to the outcome without hedging. In
other words, there is always a risk of ex-post regret. Hedgers use FRAs to eliminate risk, despite the
fact that their decision may have to be regretted subsequently.

An FRA is quoted using two numbers. An A×B FRA implies that settlement is A months
from today for a loan or deposit of B − A months. The first date is referred to as the
settlement date, and the second is termed the final maturity date. Standard FRAs have
maturities of 3, 6, or 12 months. However, because they are OTC products, contracts
Forward Rate Agreements (FRAs) | 343

with nonstandard maturities, can also be bought and sold. Standard specifications
are like 1 × 4, 3 × 6, 6 × 9, 1 × 7, 3 × 9, and 6 × 12.
Let’s consider a 6 × 12 FRA. If today is the transaction date, the spot settlement
date is T + 2. The settlement date is six months from T + 2. The reference rate is fixed
two days before the settlement date, or six months from the trade date. The maturity
date is six months from the settlement date. There are two possibilities; that is, the
cash payment may be made on the settlement date, which is termed an in arrears FRA,
or the payment may occur on the maturity date, which is termed a delayed settlement
FRA.1 We will now give a detailed example of both types of FRAs. In the market, in
arrears contracts are more common than delayed settlement contracts.
Let’s consider a 3 × 9 FRA with a contract rate of 6%. Assume that the LIBOR on
the reference fixing date is 7.50%. Also assume that the notional principal is $18 mil-
lion, and that the day-count convention is Actual/360. In our case, we assume that
six months corresponds to 184 days.2 Because the reference rate is higher than the
contract rate, the seller has to pay the buyer.
The cash amount is
184
18,000,000 × (0.075 − 0.06) × = $138,000
360
Had the FRA been a delayed settlement contract, the cash flow would occur six months
from the settlement date, and the value on the settlement date would be the present
value of this amount, determined by discounting at the LIBOR prevailing on the refer-
ence fixing date. In our case, the present value would be
138,000
184
= $132,905.30
(1 + 0.075 × 360
)

Determining the Contract Rate

If we are given rates for money market deposits, we can easily determine the contract
rate. Consider a 3 × 9 FRA. Assume the three-month LIBOR is s1 % per annum and
the nine-month LIBOR is s2 % per annum. Investing for nine months at the 9-M LI-
BOR is equivalent to investing for three months at the 3-M LIBOR and rolling over for
six months using the FRA. To keep the mathematics simple, let’s assume a 30/360
day-count convention, which implies that three months consist of 90 days and nine
months consist of 270 days. Let s1 = 6% and s2 = 7.50%. To rule out arbitrage3
90 180 270
[1 + s1 × ] × [1 + k × ] = [1 + s2 × ]
360 360 360

1 See Chance [8].

2 For contracts in USD and EUR, the day-count convention is Actual/360, as is the practice in the
money market. For transactions in GBP, AUD, CAD, or JPY, the convention is Actual/365.
3 We are using k as a symbol for the FRA rate.
344 | Chapter 12: Interest Rate Forwards and Futures

90 180 270
⇒ [1 + 0.06 × ] × [1 + k × ] = [1 + 0.075 × ]
360 360 360
⇒ k = 8.1281%

If we are given the borrowing/lending rates in the money market, we can deter-
mine an upper as well as a lower bound for the contract rate. Consider the following
quote in the market:
3-M: 0.05000 − 0.05125 and 9-M: 0.0600 − 0.0625.
That is, the dealer is prepared to borrow at 5% for three months and lend at 5.125%
for the same period. In the case of a nine-month transaction, the same dealer is pre-
pared to borrow at 6% and lend at 6.25%. Assume a 30/360 day-count convention. A
dealer can borrow for three months, roll over the loan using a FRA, and lend for nine
months. Quite obviously a dealer does so only if there is a profit to be made. Thus we
require that

90 180 270
⇒ [1 + 0.05 × ] × [1 + k × ] ≤ [1 + 0.0625 × ]
360 360 360
⇒ k ≤ 6.7901%

Now consider a situation where the dealer borrows for nine months, lends for three
months and rolls over using a FRA. In this case, for the transaction to be profitable,
we require that

90 180 270
⇒ [1 + 0.05125 × ] × [1 + k × ] ≥ [1 + 0.06 × ]
360 360 360
⇒ k ≥ 6.3561%

Thus the bid should be greater than or equal to 6.3561%, and the ask should be less
than or equal to 6.7901%. The same result can be obtained using a no-arbitrage ar-
gument. First, let’s consider the case where the arbitrageur borrows for three months
and lends for nine months. He can borrow at the dealer’s ask, which is 5.1250%, and
lend at the dealer’s bid, which is 0.06%. To rule out arbitrage

90 180 270
⇒ [1 + 0.05125 × ] × [1 + k × ] ≥ [1 + 0.06 × ]
360 360 360
⇒ k ≥ 6.3561%

Now consider a situation where the arbitrageur borrows for nine months and lends for
three months:

90 180 270
⇒ [1 + 0.05 × ] × [1 + k × ] ≤ [1 + 0.0625 × ]
360 360 360
⇒ k ≤ 6.7901%
Using Short Rates to Determine the FRA Rate | 345

Using Short Rates to Determine the FRA Rate

We can use the short-rate trees that we obtained in Chapter 11, using the Ho-Lee and
BDT models, to compute the FRA rates. Let’s first consider the Ho-Lee model-based
tree. We will reproduce the short rates and state prices for all the nodes to facilitate
the argument.

Table 12.1: Summary of the Ho-Lee model-based interest rates.

Period Node Interest Rate

1 r1,1 7.5063%
1 r1,2 5.5063%
2 r2,1 6.7725%
2 r2,2 4.7725%
2 r2,3 2.7725%
3 r3,1 11.2863%
3 r3,2 9.2863%
3 r3,3 7.2863%
3 r3,4 5.2863%

t0 t1 t2 t3 t4

0.053160

0.112320

0.232247 0.215741

0.481928 0.340258

1.0 0.466755 0.328336

0.481928 0.343589

0.234508 0.222092

0.115651
0.056336

Figure 12.1: State price tree for the Ho-Lee model.

Consider a one-period in arrears FRA. In six months the short rate can be 7.5063% or
5.5063%. The probability of reaching the respective nodes is 0.50. We know that the
value of a forward or a futures contract at the outset must be zero. Thus if the unknown
contract rate is k

0.481928 × (7.5063 − k) + 0.481928 × (5.5063 − k) = 0

⇒ k = 6.5063%
346 | Chapter 12: Interest Rate Forwards and Futures

Now consider a two-period in arrears FRA. There are three nodes respectively. Thus if
the unknown contract rate is k

0.232247 × ×(6.7725 − k) + 0.466755 × (4.7725 − k) + 0.234508 × (2.7725 − k) = 0

⇒ k = 4.7677%

Now let’s consider a one-period delayed settlement FRA. The payoffs will be received
one period hence and thus need to be discounted. Once again if the unknown contract
rate is k
(7.5063 − k) (5.5063 − k)
0.481928 × 0.075063
+ 0.481928 × 0.055063
=0
(1 + 2
) (1 + 2
)
⇒ k = 6.5015%

Now consider a two-period delayed settlement FRA. Using a similar approach,

(6.7725 − k) (4.7725 − k)
0.232247 × × 0.067725
+ 0.466755 × 0.047725
(1 + 2
) (1 + 2
)
(2.7725 − k)
+ 0.234508 × 0.027725
=0
(1 + 2
)
⇒ k = 4.7579%

Let’s repeat the computations using the BDT tree.

Table 12.2: Summary of BDT model-based interest rates.

Period Node Interest Rate

1 r1,1 6.8915%
1 r1,2 6.1122%
2 r2,1 5.1969%
2 r2,2 4.7497%
2 r2,3 4.3409%
3 r3,1 9.0245%
3 r3,2 8.4990%
3 r3,3 8.0040%
3 r3,4 7.5379%

Consider a one-period in arrears FRA. In six months the short rate can be 6.8915%
or 6.1122%. Thus if the unknown contract rate is k

0.481928 × (6.8915 − k) + 0.481928 × (6.1122 − k) = 0

⇒ k = 6.5019%
Using Short Rates to Determine the FRA Rate | 347

t0 t1 t2 t3 t4

0.054309

0.113519

0.232938 0.218091

0.481928 0.341483

1.0 0.466756 0.328389

0.481928 0.342390

0.233818 0.219742

0.114425
0.055135

Figure 12.2: State price tree for the BDT model.

Now consider a two-period in arrears FRA:

0.232938 × ×(5.1969 − k) + 0.466756 × (4.7497 − k) + 0.233818 × (4.3409 − k) = 0

⇒ k = 4.7589%

Next let’s consider one-period delayed settlement FRA:

(6.8915 − k) (6.1122 − k)
0.481928 × 0.068915
+ 0.481928 × 0.061122
=0
(1 + 2
) (1 + 2
)
⇒ k = 6.5011%

Finally, consider a two-period delayed settlement FRA. Using a similar approach,

(5.1969 − k) (4.7497 − k)
0.232938 × × 0.051969
+ 0.466756 × 0.047497
(1 + 2
) (1 + 2
)
(4.3409 − k)
+ 0.233818 × 0.043409
=0
(1 + 2
)
⇒ k = 4.7584%

The contract rates for the delayed settlement FRAs can be confirmed as follows.
The contract rate for the one-period FRA is given by
s2 2
(1 + 2
) (1.035)2
[ s1 − 1] × 2 = [ − 1] × 2 = 6.5012%
(1 + 2) (1.0375)

Similarly, the contract rate for the two-period FRA is given by

s3 3
(1 + 2
) (1.03125)3
[ − 1] × 2 = [ − 1] × 2 = 4.7581%
s2 2 (1.0350)2
(1 + 2
)
348 | Chapter 12: Interest Rate Forwards and Futures

This is consistent with the results from the Ho-Lee and BDT models. The minor
difference is due to rounding errors.

Eurodollar Futures
Eurodollar (ED) futures contracts trade on the CME in Chicago. The underlying interest
rate, is the London Inter Bank Offer Rate (LIBOR). Each futures contract is for a time
deposit with a principal of $1 million and three months to maturity. The quarterly cycle
for Eurodollar contracts is March, June, September, and December. On the CME, a to-
tal of 40 quarterly futures contracts spanning 10 years are listed at any point in time.
In addition, the four nearest serial months are also listed. Let’s assume that we are
standing on 30 June 2018. The four available serial months are July, August, October,
and November 2018. The remaining available months are September and December
2018; March, June, September, and December 2019–2027; March and June 2028. Eu-
rodollar futures contracts expire at 11:00 a. m. London time, on the second London
bank business day before the third Wednesday of the contract month. The contracts
are cash settled to the 3-M LIBOR. The futures price is quoted in terms of an index and
implies an interest rate.
Quoted ED Futures Price = 100.00 – Implicit Interest Rate
The implicit interest rate in this case is an actual or add-on interest rate and not a
discount rate as in the case of T-bills.

Calculating Profits and Losses on ED Futures

Let’s suppose that the futures price is 96. This represents a yearly interest rate of 4%
or 1% per quarter. Thus the implied quarterly interest payment on a time deposit of $1
million is

0.01 × 1,000,000 = $10,000

If the futures price were to fall to 95 at the end of the day, then it would represent a
quarterly interest payment of $12,500 on a deposit of $1 million.
Consider a person who goes long in an ED futures contract at 96. Such a person
is agreeing to lend money at the equivalent interest rate, namely 1% per quarter.4 If
interest rates rise subsequently to 5% per annum, that is, the futures price goes down

4 The logic is the same as that for underlying futures in the case of a debt security such as a T-bill. In
the case of bills, a long position means that you are willing to buy bills at the expiration of the contract.
In the case of ED futures, a long position means that you are willing to make a term deposit of three
months. In either case you are a lender.
Eurodollar Futures | 349

to 95 and the contract is marked to market, then the long loses $2,500. The logic is
that when the contract is marked to market, it is as if the person is offsetting by going
short, which in this case, means agreeing to borrow at 1.25% per quarter. Thus when
the interest rates rise, the longs lose. The reverse is true for the shorts; that is, when
the interest rates fall, they lose.
You should by now be able to see the logic behind quoting futures prices in terms
of an index, rather than in terms of interest rates. If futures prices are quoted in terms
of interest rates, then the longs gain when the futures prices fall, and the shorts gain
when the futures prices rise. But in all the other markets, we observe that the longs
gain when futures prices rise, whereas the shorts gain if futures prices fall. Thus to
make money market futures consistent with other futures markets, we do not quote
futures prices in terms of interest rates, but do so in terms of an index. When the index
rises, the longs gain, and when it falls, the shorts gain.
Yet another reason for quoting futures prices in terms of an index rather than in
terms of interest rates is to ensure that the bid prices are lower than the ask prices.
Remember that as an investor, you typically face borrowing rates that are higher than
lending rates. Thus the rates underlying a short position will be greater than the rates
underlying a long position. To be consistent with the principle that bid prices are al-
ways lower than the ask prices, it is necessary to convert the rates to equivalent index
values.
Now consider the case where the ED index changes from F0 to F1 . The profit for a
long is

(100 − F0 ) 90 (100 − F1 ) 90
1,000,000 × × − 1,000,000 × ×
100 360 100 360
(F − F0 ) 90
= 1,000,000 × 1 ×
100 360
The cash flow for the short is
(F0 − F1 ) 90
1,000,000 × ×
100 360
A change of one basis point, corresponds to a price change of

0.01 90
1,000,000 × ×
100 360
= $25

Locking in a Borrowing Rate

Let’s first illustrate how borrowing and lending rates can be locked in using ED futures
before deriving the fair price of a contract using cash-and-carry arbitrage and reverse
cash-and-carry arbitrage arguments.
350 | Chapter 12: Interest Rate Forwards and Futures

Assume that today is 15 July 2018. Rolodex Inc. is planning to borrow $1 million
on 14 September for a period of 90 days. The company is confident that given its rat-
ing, it can borrow at LIBOR. It is however worried that interest rates may rise before
14 September, which is the last day of trading for the September futures contract. The
current September futures price is 94, and the current LIBOR for a 90 day loan is 5.85%.
Because the company is planning to borrow money, it requires a short hedge. The
rationale is that a short hedge will yield a profit if rates rise, which is precisely the
situation where a borrower needs a positive payoff. Consequently borrowers need to
go short in ED futures, whereas lenders need to go long. Assume that Rolodex goes
short in one September futures contract, and let’s consider two different scenarios on
14 September, one where the LIBOR is higher as compared to the rate implicit in the
futures price, and the other where it is lower.

Case A: LIBOR = 4%
The interest payable on the loan of $1 million is
90
0.04 × 1,000,000 ×
360
= $10,000

Gain/loss from the futures market is

(F0 − F1 ) 90
1,000,000 × ×
100 360
(94 − 96) 90
= 1,000,000 × ×
100 360
= ($5,000)

Therefore, the effective interest paid is5

10,000 + 5,000 = $15,000

Case B: LIBOR = 7%
The interest payable is
90
0.07 × 1,000,000 ×
360
= $17,500

Profit/loss from the futures position is

(94 − 93) 90
1,000,000 × ×
100 360
= $2,500

5 The terminal futures price is 100 − 4.00 = 96.

Eurodollar Futures | 351

Thus, the effective interest paid is

17,500 − 2,500 = $15,000

Thus the company can lock in an interest payable of $15,000, irrespective of the
prevailing LIBOR on 14 September. This amount corresponds to a rate of:

15,000 360
× ≡ 6%
1,000,000 90

which is nothing but the rate implicit in the initial futures price. We should note a
couple of points. One, the fact that the 3-M LIBOR at the outset is 5.85% is of no con-
sequence for our analysis. Second, we have assumed that the transaction date is the
same as the expiration date of the futures contract. The significance of this is that the
expiration date is the point in time at which spot and futures prices will converge, a
consequence of which is that the basis risk is zero.
It is also important to take cognizance of the following fact. Because the ED futures
contract is cash settled, the profit/loss from the futures contract will be received/paid
on 14 September when the contract expires. However, the company is required to pay
interest on the loan only 90 days later. Thus, if there is a profit from the futures po-
sition, it can be reinvested. On the contrary, a loss from the futures market has to be
financed. If adjustments were made for the interest on such profits/losses, the effective
interest paid on the loan would be higher than 6% in case A and lower than 6% in case
B. In our study of ED futures contracts, we will ignore such interest on profits/losses.

Locking in a Lending Rate

Assume that today is 15 September 2018. Microsoft is planning to lend $1 million on 14

December for a period of 90 days. The company is of the opinion that it can lend at the
prevailing LIBOR plus 75 bp and wants to lock in a rate using ED futures. 14 December
is the last day of trading for the December contract, and the current futures price is 96.
Because the company wishes to lock-in a lending rate, it needs to go long in futures
contracts. As before, let’s consider two scenarios for the terminal value of LIBOR.

Case A: LIBOR = 5.25%

The interest receivable on the loan is
90
0.06 × 1,000,000 × = $15,000
360
The profit/loss from the futures position is

(94.75 − 96) 90
1,000,000 × × = ($3,125)
100 360
352 | Chapter 12: Interest Rate Forwards and Futures

Thus the effective interest received is

15,000 − 3,125 = $11,875

Case B: LIBOR = 3.25%

The interest receivable on the loan is
90
0.04 × 1,000,000 × = $10,000
360
The profit/loss from the futures position is
(96.75 − 96) 90
1,000,000 × × = $1,875
100 360
So the effective interest received is

10,000 + 1,875 = $11,875

Thus, irrespective of the scenario at expiration, the company is assured of a receivable

of $11,875. The equivalent interest rate is
11,875 360
× × 100 = 4.75%
1,000,000 90
This represents the rate implicit in the initial futures price, plus the assumed spread
of 75 bp. Now let’s go on to analyze the pricing of ED futures.

Cash-and-Carry Arbitrage

Let’s assume that an arbitrageur is confronted with the following situation on 15 Au-
gust 20XX. Futures contracts expiring on 18 September are priced at 95. The 90-day
ED deposit made on 18 September will mature on 17 December. The interest rate for an
ED deposit between 15 August and 17 December is 6.12%. The rate for a loan between
15 August and 18 September is 3%.
Consider the following strategy. Borrow $1 million for 34 days. Simultaneously
go short in a futures contract to borrow the maturity amount for another 90 days.
Invest the borrowed funds in a 124-day deposit.6 The amount due after 34 days is
34
1,000,000 (1 + 0.03 × 360 ). The futures contract locks in a rate of 5% when this amount
is rolled over for another 90 days. Thus, the amount payable after 124 days is
34 90
1,000,000 (1 + 0.03 × ) × (1 + 0.05 × ) = $1,015,368.7
360 360

6 34 is the number of days between 15 August and 18 September, and 124 is the number of days be-
tween 15 August and 17 December.
Eurodollar Futures | 353

The amount received when the 124-day investment matures is

124
1,000,000 × (1 + 0.0612 × ) = $1,021,080
360

Thus there is an arbitrage profit of

1,021,080 − 1,015,368.70 = $5,711.30

Reverse Cash-and-Carry Arbitrage

Let’s suppose that all the other variables have the same values as in the preceding
section, except for the futures price, which we assume is 91.50. Consider the following
strategy. Borrow $1 million for 124 days. Invest it for 34 days, and go long in a futures
contract to roll over the maturity amount for a further period of 90 days. The amount
repayable after 124 days is

124
1,000,000 × (1 + 0.0612 × ) = $1,021,080
360
34
The investment in the 34-day deposit yields 1,000,000 (1 + 0.03 × 360 ). The futures
contract locks in a rate of 8.50% for this amount. Thus, the amount receivable after
124 days is

34 90
1,000,000 (1 + 0.03 × ) × (1 + 0.085 × ) = $1,024,143.50
360 360

and there is an arbitrage profit of

1,024,143.50 − 1,021,080 = $3,063.50

The No-Arbitrage Pricing Equation

We now derive an expression for the futures price F, which precludes both cash-and-
carry and reverse cash-and-carry arbitrage for a given set of interest rates. We denote
the day on which we are standing as day t. The futures contract is assumed to expire at
T. We denote the Eurodollar rate for a T − t day loan by s1 and the borrowing/lending
rate for a T + 90 − t day loan by s2 . In order to rule out both forms of arbitrage, we
require that

(100 − F) 90 s T −t
1,000,000 × [1 + × ] × [1 + 1 × ]
100 360 100 360
s T + 90 − t
= 1,000,000 × [1 + 2 × ]
100 360
354 | Chapter 12: Interest Rate Forwards and Futures

(100 − F) 90 s T −t s T + 90 − t
⇒ [1 + × ][1 + 1 × ] = [1 + 2 × ] (12.1)
100 360 100 360 100 360
In our case, s1 = 3%, and s2 = 6.12%. Therefore
(100 − F) 90
[1 + × ] = 1.018195
100 360
⇒ F = 92.7220

Thus 95 was too high, which is why cash-and-carry arbitrage was profitable. However,
91.50 was too low, a consequence of which reverse cash-and-carry arbitrage was prof-
itable.

Hedging an N-day Loan Using ED Futures

We have illustrated how ED futures can be used to lock in borrowing/lending rates for
90-day loans. These contracts can be used to lock in rates for an N-day loan, where
N is close to 90. The necessary condition for such a hedge is that the interest rate for
the N-day loan should move closely with the rate for a 90-day loan. Before giving an
illustration, let’s first derive the required hedge ratio.
Assume that dN = α + d90 + ϵ, where dN is the annualized rate for an N-day loan
and d90 is the annualized rate for a 90-day loan. We assume7 that ϵ = 0. If so

△dN = △d90

Take the case of a party that is raising an N-day loan for $Q million. The change in the
interest payable due to a rate change is
△dN N
Q × 1,000,000 × ×
100 360
The profit/loss per contract from the futures market is
△d90 90
1,000,000 × ×
100 360
The number of futures contracts, Qf , ought to be chosen in a way such that:

△dN N △d90 90
Q × 1,000,000 × × = Qf × 1,000,000 × × (12.2)
100 360 100 360
That is, the magnitude of the change in the amount payable/receivable should be
equal to the profit/loss from the futures contract. Because we have assumed that
△dN = △d90 , we get the result that:

Q × N = Qf × 90

7 The significance of this assumption is that there is no basis risk.

Eurodollar Futures | 355

Qf N
⇒ = the hedge ratio =
Q 90

Let’s illustrate how this hedge will perform in practice, with the help of Exam-
ple 12.4.

Example 12.4. Once again assume that we are standing on 15 July 20XX and that Rolodex borrows $10
million on 14 September, for a period of 117 days. The current September futures price is 95.75. The
firm can borrow at the prevailing LIBOR on 14 September. Because the firm is borrowing, it needs a
short position in ED futures. Assume α = 0. In other words, the 90-day LIBOR is equal to the 117-day
LIBOR.

117
Qf = 10 × = 13
90

Let’s examine the performance of this hedge. First consider a situation where the LIBOR on 14 Septem-
ber is 4%. The actual interest paid by Rolodex is

117
0.04 × 10,000,000 × = $130,000
360

The profit/loss from the futures position is

(95.75 − 96) 90
13 × 1,000,000 × ×
100 360
= ($8,125)

Thus, the effective interest paid by the company is $138,125.

Now consider a situation where the LIBOR on 14 September is 4.5%. The actual interest paid by
Rolodex is

117
0.045 × 10,000,000 × = $146,250
360

Profit/loss from the futures position is

(95.75 − 95.50) 90
13 × 1,000,000 × ×
100 360
= $8,125

The effective interest paid by the company is $138,125.

Hence, irrespective of the prevailing LIBOR on 14 September, the company has locked in an in-
terest expense of $138,125. This amount corresponds to an interest i such that

117
10,000,000 [1 + i × ] = $10,138,125
360

i is therefore equal to 4.25%, which is nothing but the rate implicit in the initial futures price. Had we
assumed a non-zero value for α, we would have locked in the rate implicit in the futures price plus α.
For instance, if we had assumed that α = 0.50%, we would have locked in 4.75%.
356 | Chapter 12: Interest Rate Forwards and Futures

A More General Argument

Assume that the 90-day LIBOR and the N-day LIBOR are related by the following rela-
tionship: dN = α + β × d90 + ϵ. Once again, if we assume ϵ = 0,

△dN = β × △d90

Take the case of a party that is raising an N-day loan for $Q million. The change in the
interest payable due to a rate change is

△dN N
Q × 1,000,000 × ×
100 360
The profit/loss per contract, from the futures market is

△d90 90
1,000,000 × ×
100 360
The number of futures contracts, Qf , ought to be chosen in a way such that

△dN N △d90 90
Q × 1,000,000 × × = Qf × 1,000,000 × × (12.3)
100 360 100 360
Because we have assumed that △dN = β × △d90 , we get the result

Q × β × N = Qf × 90
Qf β×N
⇒ = the hedge ratio =
Q 90
Now let’s illustrate how this hedge will perform in practice, with the help of Exam-
ple 12.5.

Example 12.5. Once again assume that we are standing on 15 July 20XX and that Rolodex borrows $10
million on 14 September, for a period of 108 days. The current September futures price is 95.75. The
firm can borrow at the prevailing LIBOR plus 50 bp on 14 September. Because the firm is borrowing, it
requires a short position in ED futures. Assume d108 = 0.25 + 1.25 × d90 + ϵ. Using the same argument

108
Qf = 10 × 1.25 × = 15
90

let’s examine the performance of this hedge. Consider a situation where the 90-day LIBOR on 14
September is 4%. Thus the 108-day LIBOR is 0.25 + 1.25 × 4 = 5.25%. The actual interest paid by
Rolodex is
108
0.0575 × 10,000,000 × = $172,500
360

The profit/loss from the futures position is

(95.75 − 96) 90
15 × 1,000,000 × ×
100 360
Using ED Futures to Create a Fixed Rate Loan | 357

= ($9,375)

Thus, the effective interest paid by the company is 172,500 + 9,375 = $181,875. The corresponding
rate is
181,875 360
× ≡ 6.0625%
10,000,000 108
6.0625 = 0.25 + 1.25 × 4.25 + 0.50 ≡ α + β × 4.25 + 50 bp

Using ED Futures to Create a Fixed Rate Loan

Assume that Jyske Bank is able to borrow money at LIBOR for three monthly periods.
Consequently the interest payable on its liability is variable. The bank, however, has
a potential client who wishes to borrow at a fixed rate for a period of one year. Thus
the interest receivable from the proposed asset is fixed in nature. The bank would like
to use futures contracts to mitigate the risk of borrowing at a floating rate, but lending
at a fixed rate. It turns out that ED futures contracts can be used to hedge the funding
risk and determine a suitable rate that can be quoted by the bank while negotiating a
fixed rate loan. We can illustrate the principle involved with the help of Example 12.6.

Example 12.6. IT20, a company based in London, wants a loan for $100 million for a period of one year
from 15 September 20XX, at a fixed interest rate. Let’s assume that the 90-day LIBOR on 15 September
is 4.8% and that December, March, and June contracts are available at 96.1, 95.6, and 96 respectively.
For ease of exposition, we assume that the dates on which the bank will roll over its three-month bor-
rowings, namely 15 December, 15 March, and 15 June, are the same as the dates on which the futures
contracts for those months are scheduled to expire. Once again, the objective of making such an as-
sumption, is to ensure that there is no basis risk. Each ED futures contract is on $1 million and the bank
is borrowing at a floating rate. Consequently, it needs a short position in 100 each of the December,
March, and June futures contracts.
For the first quarter, the interest expense for the bank is

90
100,000,000 × 0.048 × = $1,200,000
360
There is no uncertainty about this amount because it is based on the current 3M LIBOR. The short
position in December contracts, locks in a rate of 3.9% for a period of 90 days from December to
March. This corresponds to an interest expense of

90
100,000,000 × 0.039 × = $975,000
360
In a similar fashion, the short position in March contracts, locks in

90
100,000,000 × 0.044 × = $1,100,000
360
for the 90-day period from March. The June contracts lock in the following interest amount for the last
quarter:

90
100,000,000 × 0.04 × = $1,000,000
360
358 | Chapter 12: Interest Rate Forwards and Futures

Hence, the total interest payable by the bank for the 12-month period is

1,200,000 + 975,000 + 1,100,000 + 1,000,000 = $4,275,000.

This corresponds to an annualized interest rate of

4,275,000
× 100 = 4.275%
100,000,000

The bank can now quote a fixed rate to the client, based on this effective cost of funding, after
factoring in hedging costs and a suitable profit margin.

Stack and Strip Hedges

We assumed in the preceding illustration that the bank would use 100 futures con-
tracts for each of the expiration months. Such a hedge, where the same number of
contracts are used for each expiration month right from the outset, is called a strip
hedge. However, if the maturity of a futures contract is far away, the contract may not
be actively traded. Such illiquidity could be a deterrent for a potential hedger, who
would seek the ability to enter and exit the market at a price that is close to the true or
fair value of the asset. In this case, when the hedge is initiated in September, it is con-
ceivable that the June futures may be illiquid. In such a situation, rather than starting
with an equal number of December, March, and June contracts, the bank may initiate
the hedge using an unequal number of December and March contracts, without taking
a position in the June contract. Subsequently, when the December contracts expire, it
will take a position in the June contract. Such a hedge, where the number of contracts
for each maturity is not equal at the outset, is known as a stack hedge. Let’s illustrate
the mechanics of such a hedge with the help of Example 12.7.

Example 12.7. Assume that the June futures contracts are perceived to be illiquid when the hedge is
initiated. The bank therefore decides to hedge using 100 December contracts and 200 March contracts.
The position in December futures is obviously intended to lock in a rate for that month’s borrowing.
Out of the 200 March contracts, 100 are intended to lock in a rate for March, and the remaining 100 are
meant for hedging the June exposure. The inherent assumption is that on 15 December, the bank will
partially offset its March position and go short in 100 June contracts, assuming that they have begun
to be actively traded by then. The relative performance of a stack hedge vis a vis a strip hedge depends
on the movement of interest rates between September and December.

Case A: The Stack Hedge Is Equivalent to the Strip

Assume that the March futures price moves from 95.6 to 95 between September and December, and the
June futures price moves from 96 to 95.4 during the same period. When the 100 extra March contracts
are liquidated in December, the profit from the position is

95.60 − 95 90
100,000,000 × ( )×
100 360
Stack and Strip Hedges | 359

= $150,000

The 100 June contracts that the bank enters into at 95.4 lock in an interest expense of

90
100,000,000 × 0.046 × = $1,150,000
360
for the last quarter. The effective interest expense for this quarter is therefore:

1,150,000 − 150,000 = $1,000,000

which is the same amount that was locked in by the strip hedge.
Thus, if the three-month ED rate, as contained in the March futures price, changes by the same
magnitude and direction as the yield contained in the June futures price, then the strip and stack
hedges will be equivalent. In interest rate parlance, we would say that there has been a parallel shift
in the yield curve.

Case B: The Strip Outperforms

If the increase in the March yield is less than the increase in the June yield, then the strip hedge will
outperform the stack. For instance, assume that the March futures price moves from 95.6 to 95 whereas
the June futures price moves from 96 to 95.20. Because the futures price has declined, there is a profit
when the extra March contracts are offset, which as before is $150,000. However, the interest ex-
penses for the last quarter is higher. This expense is

90
100,000,000 × 0.048 × = $1,200,000
360
The effective interest is

1,200,000 − 150,000 = $1,050,000

which is greater than the amount of $1,000,000 that was locked in by the strip hedge.
The same is the case if the decrease in the March yield is more than the decrease in the June
yield. In this case, there is a loss when the March contracts are offset. However, the interest expense
locked in by the June contracts is lower. For instance, assume that the March futures price moves to
95.9, whereas the June futures price moves to 96.1.
The profit from the March position is

95.6 − 95.9 90
100,000,000 × ( )×
100 360
= ($ 75,000)

The interest expense for the last quarter, as locked in by the June contracts, is

90
100,000,000 × 0.039 × = $975,000
360
The effective interest paid is

975,000 + 75,000 = $1,050,000

which is greater than the amount of $1,000,000 that was locked in by the strip hedge. Clearly the gain
on account of the decline in yield implicit in the June contracts is not adequate to compensate for the
loss incurred when the March contracts are offset.
360 | Chapter 12: Interest Rate Forwards and Futures

Case C: The Stack Outperforms

If the increase in the March yield is more than the increase in the June yield, then the stack hedge will
out perform the strip. For instance, assume that the December futures price moves from 95.60 to 95,
and the June futures price moves from 96 to 95.60. The effective interest rate for the last quarter in
this case is
90
100,000,000 × 0.044 × − 150,000
360
= 1,100,000 − 150,000
= $950,000

This amount is less than the value of $1,000,000 that was locked in by the strip hedge.
The same is true if the decrease in the March yield is less than the decrease in the June yield. In
this case, once again there is a loss when the March contracts are offset. However, this is more than
compensated for by the decline in the rate for the last quarter. For instance assume that the March
futures price moves from 95.60 to 95.90, whereas the June futures price moves from 96 to 96.50.
The loss from the March position is $75,000. The interest expense for the last quarter is

90
100,000,000 × 0.035 × = $875,000
360
The effective interest is

875,000 + 75,000 = $950,000

which is less than the amount of $1,000,000 that was locked in by the strip hedge.

FRAs vs. ED Futures: An Important Point

A bullish speculator who believes that interest rates are going to increase, or a hedger
who wants to lock in a borrowing rate, will buy a FRA. Similarly, a bearish specu-
lator who believes that interest rates are going to decline, or a hedger who wants to
lock in a lending rate, will sell the FRA. However if such speculators and hedgers use
ED futures instead of FRAs, bullish speculators and hedgers seeking to lock in a bor-
rowing rate will go short in ED futures. On the other hand, bearish speculators and
hedgers seeking to lock in a lending rate will go long in ED futures. Thus from a hedg-
ing/speculation standpoint, long positions in FRAs are equivalent to short positions
in ED futures, whereas short positions in FRAs are equivalent to long positions in ED
futures.

Federal Funds
Federal funds are perhaps the most important of all money market instruments be-
cause they are the primary means of making payments in this market. The term refers
to money that is available for immediate payment. Commercial banks use this instru-
ment as the principal way to adjust their legal reserve account at the Federal Reserve
Federal Funds | 361

bank in the district where they are located. Federal funds are same-day money. In con-
trast, the normal payments that are made by check take at least 24 hours for the payee
to receive the funds. As can be appreciated, such transactions, called clearinghouse
funds, are far too slow for money market participants. Banks and other depository
institutions must hold in a reserve account assets equal to a fraction of the funds de-
posited with them. Legal reserve requirements are met by holding vault cash and re-
serve balances with the local Federal Reserve bank. Because these reserves earn little
or no income, banks whose reserve balances exceed statutory requirements lend the
excess reserves in their possession to financial institutions with a deficit. Here is an
illustration.

Example 12.8. Consider the following balance sheets for two commercial banks.

Liabilities Assets

Deposit from Alice $5,000 Reserves at the Fed $2,000

Vault Cash $500

Loan to Maureen $2,500

Figure 12.3: Balance Sheet of Bank One.

Liabilities Assets

Deposit from Paula $2,000 Reserves at the Fed $500

Loan to Rita $1,500

Figure 12.4: Balance Sheet of Bank Two.

362 | Chapter 12: Interest Rate Forwards and Futures

Assume that banks are required to maintain 10% of their liabilities as reserves. Bank One therefore
is required to keep $500 as reserves. However, it has $2,000. Consequently it has excess reserves of
$1,500. Bank Two on the other hand is required to keep a reserve of $200. Thus it has an excess of
$300. Now assume that a customer called Rachel approaches Bank Two seeking a loan of $840. Be-
cause it has excess reserves of $300, the bank needs to borrow an additional $X , to meet the request.
The amount borrowed should be such that

0.9X + 300 = 840 ⇒ X = $600

Bank One is obviously in a position to lend this amount to Bank Two. If it does, the positions of the two
banks are as depicted in Figures 12.5 and 12.6.

Liabilities Assets

Deposit from Alice $5,000 Reserves at the Fed $1,400

Loan to Maureen $2,500

Loan to Bank Two $600

Vault Cash $500

Figure 12.5: Balance sheet of Bank One following the transaction.

Liabilities Assets

Deposit from Paula $2,000 Reserves at the Fed $260

Loan from Bank One $600 Loan to Rita $1,500

Loan to Rachel $840

Figure 12.6: Balance sheet of Bank Two following the transaction.

Bank Two has total reserves of $260 which is exactly 10% of its total liability of $2,600.
The legal reserve requirement of banks is calculated on a daily average basis over
a period known as the reserve computation period. The Federal Reserve calculates the
Fed Funds Futures | 363

daily average of transaction deposits held by each depository institution over this pe-
riod and then multiplies that average by the required reserve percentage to determine
the amount of legal reserves that must be held by each institution. These legal reserves
must on average be equal to the required amount over a period known as the reserve
maintenance period.
Once the borrower and the lender agree on the terms of the loan, the lending insti-
tution contacts the district Federal Reserve bank requesting a wire transfer of federal
funds. The Fed then transfers reserves through its wire network, FEDWIRE, to the Fed-
eral Reserve bank serving the region where the borrowing institution is located. The
transaction is reversed when the loan is repaid. Most federal funds loans are overnight
transactions or continuing contracts that have no specific maturity date and can be
terminated without advance notice by either party. Longer maturity loans are referred
to as term federal funds. Most of the transactions are overnight, because the reserve
position can fluctuate substantially from day to day, a consequence of which banks
with excess reserves prefer to lend on an overnight basis.

Fed Funds Futures

The underlying asset is a 30-day federal funds loan, with a principal value of $5 mil-
lion. Prices are quoted as 100.00 minus the average daily fed funds overnight rate. At
any point, in time contracts for the next 36 consecutive months are available. Con-
tracts expire on the last business day of the delivery month and are cash settled.
A change of one basis point in the quoted futures price is equivalent to

0.01 30
5,000,000 × × = $41.67
100 360

T-Note and T-Bond Futures

Futures contracts in the US are available on 2-year, 5-year, and 10-year T-notes. In the
case of the 2-year T-note contract, the underlying asset is a T-note with a face value of
$200,000. Multiple grades are allowable for delivery. The deliverable grade must have
an original maturity of not more than 5 years and 3 months, and an actual maturity
of not less than 1 year and 9 months from the first day of the delivery month, and not
more than 2 years from the last day of the delivery month. The actual futures price is
subject to a multiplicative adjustment factor called the conversion factor (CF).8
In the case of futures contracts on 5-year T-notes, the underlying asset is a T-note
with a face value of $100,000. The deliverable grade must have an original maturity

8 The conversion factor is a multiplicative adjustment factor. In the “Conversion Factors” section, we
describe in detail the procedure for its computation.
364 | Chapter 12: Interest Rate Forwards and Futures

of not more than 5 years and 3 months, and an actual maturity of not less than 4 years
and 2 months from the first day of the delivery month.
The underlying asset for contracts on 10-year notes is a T-note with a face value
of $100,000. The deliverable grade must have an actual term to maturity of not less
than 6 years and 6 months from the first day of the delivery month and not more than
10 years from that date. There is also a futures contract called an ultra 10-year T-note
futures contract. The underlying asset is a T-note with a face value of $100,000 and
an actual term to maturity of not less than 9 years and 5 months and not greater than
10 years from the first day of the delivery month.

T-Bond Contracts

The underlying asset is a T-bond with a face value of $100,000. Multiple grades are
allowable for delivery. The deliverable grades must have a maturity of at least 15 years
from the first day of the delivery month and less than 25 years from that date. There
is also a futures contract termed as an ultra T-bond futures contract. The underlying
asset is a T-bond with a face value of $100,000 and an actual term to maturity of at
least 25 years as of the first day of the delivery month.
All the preceding contracts, on T-notes and T-bonds, are subject to delivery settle-
ment.

Conversion Factors
As one can see from the contract specifications, a wide variety of notes and bonds
with different coupons and maturity dates are eligible for delivery under any particu-
lar futures contract. The choice as to which bond to deliver is made by the short, and
obviously the price received depends on the bond that it chooses to deliver. If the short
delivers a more valuable bond, it should receive more than what it would, were it to de-
liver a less valuable bond. Thus, in order to facilitate comparisons between bonds, the
exchange specifies a conversion factor for each bond that is eligible for delivery. This
is a multiplicative price adjustment system to facilitate comparisons between different
bonds that are eligible for delivery.
The conversion factor, for a bond is the value of the bond per $1 of face value,
as calculated on the first day of the delivery month, using an annual YTM of 6%
with semiannual compounding. For the purpose of calculation, the life of the bond is
rounded down to the nearest multiple of three months. If after rounding off, the bond
has a life that is an integer multiple of semiannual periods, then the first coupon is
assumed to be paid after six months. However, after rounding off, if the life of the
bond is not equal to an integer multiple of half-yearly periods, then the first coupon
is assumed to be paid after three months and the accrued interest is subtracted. The
following examples illustrate these principles.
Conversion Factors | 365

The invoice price, which is the price received by the short, is calculated as follows:

Invoice Price = Invoice Principal Amount + Accrued Interest

= CF i × F × 100,000 + AI i (12.4)

where CF i is the conversion factor of bond i, F is the quoted futures price per dollar of
face value,9 and AI i is the accrued interest.

Example 12.9. Let’s assume that we are short in a September futures contract and that today is 1
September 2018. Consider a 5% T-Bond that matures on 15 May 2047. This bond is obviously eligi-
ble for delivery under the futures contract. On September 1, this bond has 28 years and 8 21 months to
maturity. When we round down to the nearest multiple of 3 months, we get a figure of 28 years and 6
months.
The first coupon is assumed to be paid after six months. The conversion factor may therefore be
calculated as follows
5
2
PVIFA(3,57) + 100PVIF(3,57)
CF =
100
67.8773 + 18.5472
=
100
= 0.8642

PVIFA stands for present value interest factor annuity and PVIF stands for present value interest factor.

1 1
PVIFA(3,57) = × [1 − ]
0.03 (1.03)57
1
PVIF(3,57) =
(1.03)57

Example 12.10. Instead of the May 2047 bond, consider another bond that is maturing on 15 February
2047, with a coupon of 4.75%. This bond, too, is suitable for delivery. On 1 September 2018, this bond
has 28 years and 5 21 months to maturity. The life of the bond when we round down to the nearest
multiple of 3 months is 28 years and 3 months.
In this case, we assume that the first coupon is paid after three months. The conversion factor CF,
can be calculated in three steps as shown here:
1. First, find the price of the bond three months from today, using a yield of 6% per annum.

4.75 4.75
P= + PVIFA(3,56) + 100PVIF(3,56)
2 2
= 2.375 + 64.0430 + 19.1036
= $85.5216

2. Discount the price gotten in the preceding step for another three months:

85.5216
1
= $84.2669
(1.03) 2

9 T-Bond futures prices are quoted in the same way as the cash market prices, that is, they are clean
prices.
366 | Chapter 12: Interest Rate Forwards and Futures

3. Subtract the accrued interest for three months from the price obtained in the second step:

4.75 1
AI = × = 1.1875
2 2
84.2669 − 1.1875
CF = = 0.8308
100

Why do we adopt two different procedures for calculating the CF? The CF is used to
multiply the quoted futures price, which is a clean price. Hence the CF should not in-
clude any accrued interest. In the first example, the bond has a life that is an integer
multiple of semiannual periods after rounding off. Consequently, we need not be con-
cerned with accrued interest.10 However, in the second example, accrued interest for
a quarter is present in the value we get in the second step. Hence, in this case, we need
to subtract this interest in order to arrive at the conversion factor.

Calculating the Invoice Price for a T-Bond

Assume that on 15 September 2018 the short declares a decision to deliver the 5% bond
maturing on 15 May 2047 under the September futures contract. The actual delivery
will obviously take place two business days later, that is, on 17 September.
The invoice price may be computed as follows. The first step is obviously to cal-
culate the accrued interest. The last coupon would have been paid on 15 May 2018,
and the next is due on 15 November 2018. Between the two coupon dates there are
184 days. Between the last coupon date and the delivery date, there are 125 days. The
accrued interest for a T-bond with a face value of $100,000 is
0.05 125
AI = × × 100,000 = $1,698.3696
2 184

The futures settlement price11 on 15 September is assumed to be 95-12. This corre-

sponds to a decimal futures price per dollar of face value of
12
95 + 32
= $0.9538
100
The conversion factor has already been calculated to be 0.8642.

Invoice Price = 0.9538 × 0.8642 × 100000 + 1,698.3696 = $84,125.77

10 On a coupon date, the accrued interest is zero.

11 Note that the settlement price is based on the day the intention to deliver is declared, whereas
accrued interest is computed as of the delivery day. The reason for this is that after the intention to
deliver is declared, marking to market ceases and hence the futures price payable by the long is the
settlement price as of that day. However, because the long receives the bond only on the delivery day,
accrued interest must be computed until that day.
The Cheapest-to-Deliver (CTD) Bond | 367

The Cheapest-to-Deliver (CTD) Bond

The conversion factor is an example of a multiplicative adjustment, when more than
one grade of the underlying asset is specified for delivery. As we have seen, under such
circumstances the short will choose to deliver the bond that is the cheapest to deliver.
To illustrate this, as well as other concepts that follow, let’s depict a hypothetical sit-
uation.

t−1 t t1 T t2

C/2 C/2 C/2

Figure 12.7: Time line for the CTD bond at expiration.

t represents today. The last coupon of $C/2 was paid at t−1 and the next coupon is due
at time t1 . There is a T-bond futures contract expiring at time T, which is followed by a
coupon at time t2 . When the short delivers a particular bond, bond i, at time T, it will
receive12

(FT × CF i + AI i,t1 ,T ) × 100,000

The cost of acquisition of the bond in the spot market at time T is (Pi,T + AI i,t1 ,T ) ×
100,000. Thus the profit for the short is

(FT × CF i + AI i,t1 ,T ) × 100,000 − (Pi,T + AI i,t1 ,T ) × 100,000

= (FT × CF i − Pi,T ) × 100,000 (12.5)

Pi,T
If FT × CF i − Pi,T is to be maximized, or equivalently, CF i (FT − CF i
) is to be maximized,
P P
we require that CFi,T be minimized. CFi,T is the delivery adjusted spot price in this case.
i i
Thus, as we have seen, the cheapest-to-deliver bond is the one with the lowest delivery-
adjusted spot price. The delivery date spot-futures convergence ensures that FT × CF i −
Pi,T = 0; that is, the quoted futures price at expiration equals the delivery-adjusted
spot price of the cheapest-to-deliver bond.
In practice, T-bond futures contracts give the short a number of delivery options.
Consequently, the quoted futures price at expiration is usually less than the delivery-
adjusted spot price of the cheapest-to-deliver bond.

12 From now on, we denote the clean price per dollar of face value by P, and the futures price per
dollar of face value by F. AI is used to denote the accrued interest per dollar of face value.
368 | Chapter 12: Interest Rate Forwards and Futures

The Cheapest-to-Deliver Bond Prior to Expiration

Consider the following cash-and-carry arbitrage strategy. Buy bond i at time t by bor-
rowing at the rate r. Simultaneously go short in CFi futures contracts expiring at time
T, where CF i is the conversion factor of the bond.
Let Pi,t be the clean spot price of the T-bond at time t and Pi,T be the clean spot
price at time T. We denote the corresponding quoted futures prices by Ft and FT , re-
spectively. The payoff at the time of expiration is given by

(FT × CF i + AI i,t1 ,T ) × 100,000 + CF i (Ft − FT ) × 100,000

= (Ft × CF i + AI i,t1 ,T ) × 100,000 (12.6)

The arbitrageur receives a coupon of C/2 at time t1 . This can be reinvested until
time T to yield an amount Ii , which is the future value of the payout per dollar of face
value. Hence the implied repo rate is

Ft × CF i + AI i,t1 ,T + Ii − (Pi,t + AI i,t−1 ,t )

IRR = (12.7)
(Pi,t + AI i,t−1 ,t )

To preclude cash-and-carry arbitrage, the IRR should be less than the borrowing rate.
But in the case of T-bonds, as in the case of other futures contracts that allow for mul-
tiple deliverable grades, more than one bond is eligible for delivery, and each one has
its own IRR. If the IRR is greater than the borrowing rate, then arbitrage is possible.
Such arbitrage continues, until there is no profit to be made from any of the bonds that
are eligible for delivery. At this point in time, the cheapest-to-deliver bond is the one
that maximizes the IRR, which in an arbitrage free setting just equals the borrowing
rate. If we call this bond i, then

Ft × CF i + AI i,t1 ,T = (Pi,t + AI i,t−1 ,t ) × (1 + r) − Ii

[(Pi,t + AI i,t−1 ,t ) × (1 + r) − Ii − AI i,t1 ,T ]
⇒ Ft = (12.8)
CF i

We define

[(Pi,t + AI i,t−1 ,t ) × (1 + r) − Ii − AI i,t1 ,T ]

as the no-arbitrage futures price for bond i less accrued interest, Fi∗ , or the no-arbitrage
quoted futures price. Therefore,

Fi∗
Ft = (12.9)
CF i
The Cheapest-to-Deliver (CTD) Bond | 369

Prior to expiration, the actual futures price is equal to the delivery-adjusted no-
arbitrage quoted futures price for the cheapest-to-deliver bond. For any other bond j,13

Ft × CF j + AI j,t1 ∗ ,T + Ij − (Pj,t + AI j,t−1 ∗ ,t )

<r
(Pj,t + AI j,t−1 ∗ ,t )
[(Pj,t + AI j,t−1 ∗ ,t ) × (1 + r) − Ij − AI j,t1 ∗ ,T ]
⇒ Ft <
CF j
Fi∗ Fj∗
⇒ Ft = < ∀j (12.10)
CF i CF j

Example 12.11. Let’s illustrate the preceding concepts in detail using a numerical example. Assume
that today is 14 August 2018. September futures contracts expire on 30 September. There are two
bonds that are eligible for delivery. One is a 5% coupon bond maturing on 15 May 2047, and the other
is a 6.5% coupon bond expiring on 15 November 2047. For ease of exposition, we have chosen bonds
that do not pay any coupons between time t, which represents the day on which we are standing, and
time T , which is the expiration date of the futures contract. Hence we do not have to worry about the
future value of payouts I. The conversion factor for the first bond, which we call bond A, is 0.8642,
whereas that for the second, which we call bond B, is 1.0683.
The quoted spot price for bond A is 108-00 and that for bond B is 132-11. These prices correspond
to a YTM of 4.5% per annum.The borrowing rate is 7% per annum. Now if a bond is traded on 14 Au-
gust, the actual settlement takes place on 15 August, which is the next business day and therefore the
relevant day for our calculations.
The market price of bond A can be calculated as follows. The number of days from the last coupon
date, which is 15 May 2018, until 15 August 2018 is 92. The number of days from 15 May 2018 until 15
November 2018, which is the next coupon date, is 184. Hence the accrued interest per $100 of face
value is

5 92
× = $1.25
2 184

The dirty price per $100 of face value is therefore

108 + 1.25 = $109.25

Similarly in the case of bond B, the accrued interest is

6.5 92
× = $1.6250
2 184

and the total market price is

11
(132 + ) + 1.6250 = $133.9688
32

13 t−1 ∗ denotes the last coupon date of bond j, and t1 ∗ denotes the coupon date corresponding to t1
for bond i.
370 | Chapter 12: Interest Rate Forwards and Futures

The next step is to calculate the delivery-adjusted no-arbitrage quoted futures prices for the two bonds.
The number of days between 15 August and 30 September is 46. Hence, for bond A,

46
Pd (1 + r) = 109.25 × [1 + 0.07 × ] = $110.2272
360

The accrued interest from the last coupon date until the expiration date of the futures contract is

5 138
× = $1.8750
2 184

Hence the no-arbitrage futures price of bond A less accrued interest is

110.2272 − 1.8750 = $108.3522

The delivery-adjusted no-arbitrage quoted futures price for bond A is therefore

108.3522
= 125.3786 ≡ 125 − 12
0.8642

Similarly, the no-arbitrage futures price for bond B less accrued interest is

46 6.5 138
133.9688 × [1 + 0.07 × ]− × = 135.1671 − 2.4375
360 2 184
= 132.7295 ≡ 132 − 23

Hence, the delivery-adjusted no-arbitrage quoted futures price for bond B is

132.7295
= 124.2437 ≡ 124 − 08
1.0683

Thus bond B is cheaper to deliver as of 14 August 2018.

Risk in an Arbitrage Strategy Due to Multiple Deliverable Grades

Financing risk and payout risk are inherent in arbitrage strategies. The first refers to
the fact that the arbitrageur may be unable to borrow or lend at a fixed rate for the
entire time horizon, and in practice may have to do so for short periods and repeat-
edly roll forward. Each time the arbitrageur rolls forward in practice, there is a risk
that interest rates may have increased or declined. Payout risk refers to the fact that
the underlying asset may not make payouts as forecasted. This is very important for
an asset such as an equity share, where the actual dividend could be substantially
different from the forecasted value. However, for federal government securities such
as T-notes and T-bonds, that is not an issue. However, there is a third element of risk
as well, which is confronted by arbitrageurs who take long positions in contracts that
specify multiple grades of the underlying asset as suitable for delivery. As we have
seen in the case of such contracts, the option of which grade to deliver is always given
to the short. The short will obviously opt to deliver the grade that is the cheapest at
Risk in an Arbitrage Strategy Due to Multiple Deliverable Grades | 371

the time of delivery. The issue is that this grade need not correspond to the grade that
has been sold short by an arbitrageur at an earlier point in time, as part of a reverse
cash-and-carry arbitrage strategy. If the grade that is delivered by the short is differ-
ent from what the long requires, then the arbitrageur has to acquire the grade that
he originally short sold. The grade received from the short as a part of the futures con-
tract has to be disposed of in the spot market. The net result is that the ex-post implied
reverse repo rate for the arbitrageur could be higher than the ex-ante implied reverse
repo rate and may at times be even higher than the lending rate. If so, the realized
profit may be less than anticipated, or in a worse situation, the arbitrageur may end
up with a net loss. Hence, reverse cash-and-carry arbitrage, under such circumstances
is fraught with danger and more appropriately termed as risk arbitrage.
Before we go on to analyze the additional risk in the case of a reverse cash-and-
carry transaction, let’s first reconsider a cash-and-carry transaction. Normally we as-
sume that the arbitrageur will take a spot-futures position in the ratio of 1:1. That is, if
the futures contract is for 100 units of the underlying asset, the arbtrageur goes long in
100 units for every contract in which he has a short position. However, although this
is appropriate in the case of contracts that specify only one deliverable grade, the situ-
ation is different for cases where multiple deliverable grades have been specified and
a multiplicative price adjustment method is in use. Assume that the arbitrageur goes
long in one unit of grade i of the asset. We assume that multiple grades have been
specified for delivery, and that a multiplicative price adjustment system is in opera-
tion. Let’s denote the required short position as h futures contracts. When delivering
the asset at expiration, the arbitrageur receives ai FT . The profit from marking to mar-
ket is h(Ft − FT ). For the strategy to be riskless, we require that

ai FT + h(Ft − FT ) = ai Ft
⇒ h = ai (12.11)

Thus the appropriate number of futures contracts is equal to the price adjustment fac-
tor of the grade in which the arbitrageur has taken a long position. It should be noted
that if an additive system of price adjustment is used, then the arbitrageur still needs
a spot futures position of 1:1. This can be demonstrated as follows:

FT + ai + h(Ft − FT ) = Ft + ai ⇒ h = 1

The very fact that a cash-and-carry arbitrage strategy is initiated signifies that the
ex-ante implied repo rate is greater than the borrowing rate. That is,

ai Ft − Si,t
>r (12.12)
Si,t

At the point of expiration of the futures contract, there are two possibilities. It
may be the case that grade i is the cheapest-to-deliver grade. If so, the arbitrageur will
372 | Chapter 12: Interest Rate Forwards and Futures

deliver it and have a realized implied repo rate equal to what has been anticipated
from the very outset. However, what if some other grade, j, has become the cheapest
to deliver? If this is the case, the arbitrageur can sell the unit of grade i in his possession
at its prevailing spot price and acquire ai units of grade j to deliver under the futures
contract.
The cash inflow is

Si,T + ai × (aj FT ) + ai (Ft − FT )

This represents the inflow from three sources, namely, the cash flow from the sale of
grade i in the spot market, the proceeds from the delivery of grade j under the futures
contract, and the cumulative profit from marking to market. The cash outflow, on ac-
count of the acquisition of grade j, is

ai Sj,T

The net inflow therefore is

Si,T + ai × (aj FT ) + ai (Ft − FT ) − ai Sj,T

= Si,T + ai Sj,T + ai (Ft − FT ) − ai Sj,T
Si,T Sj,T
= ai Ft + ai [ − ]
ai aj
> ai Ft (12.13)

because if grade j is the cheapest-to-deliver by assumption, then its delivery adjusted

S
spot price aj,T will be lower than that of grade i and equal to the futures price, FT . Thus
j
we conclude that if a cash-and-carry arbitrage strategy is initiated because it looks
profitable at the outset, it can only lead to a greater profit, if not the anticipated profit,
at the time of expiration. In other words, the ex-post implied repo rate is greater than
or equal to the ex-ante implied repo rate, which by assumption is greater than the
borrowing rate. Hence, there is no risk as such.
Now let’s turn our attention to a reverse cash-and-carry strategy. The very fact that
such a strategy is being initiated, implies that

ai Ft − Si,t
<r (12.14)
Si,t

If the short delivers grade i, the grade that has been short sold by the arbitrageur, at
the end, then the anticipated arbitrage profit is realized. However, what if the arbi-
trageur is forced to take delivery of another grade, j, because the short finds that it is
the cheapest to deliver? If so, the arbitrageur has to sell this grade in the spot market
and acquire grade i at its prevailing spot price to cover the short position.
Seller’s Options | 373

The outflow is

ai (aj FT ) + Si,T

This represents the outflow on account of taking delivery under the futures contract
and the cost on account of the covering of the initial short position. The inflow is

ai Sj,T + ai (FT − Ft )

This represents the cash flow on account of the sale of grade j, as well as the cumulative
cash flow due to marking to market.
The net outflow is

ai (aj FT ) + Si,T − ai Sj,T − ai (FT − Ft )

= ai Sj,T + Si,T − ai Sj,T − ai (FT − Ft )
Si,T Sj,T
= ai Ft + ai [ − ]
ai aj
> ai Ft (12.15)

Consequently, the ex-post implied reverse repo rate can be greater than the ex-ante
implied reverse repo rate and perhaps even greater than the lending rate. Therefore, to
reiterate, what looked like a profitable arbitrage opportunity may end up as a reduced
arbitrage profit or perhaps even a loss. Thus arbitrage, under such circumstances, is
accompanied by an additional element of risk.

Seller’s Options
A person who takes a short position in a T-bond futures contract has a number of op-
tions at the time of delivery. These options are referred to as quality options and timing
options. A quality option in one under which the short has the right to select which
bond to deliver. With a timing option, the short can choose the time of delivery. Before
we go on to analyze the options in detail, let’s take a look at the delivery procedure.

The Delivery Process

Delivery in the case of T-bond and T-note futures contracts is a three-day process. The
three-day period begins with what is called the intention day, which is the day on
which the short notifies the clearing corporation of an intention to deliver. The in-
tention day can be any day from two business days prior to the first business day of
the delivery month until two business days before the last business day of the deliv-
ery month. On the next day, which is called the notice day, the clearing corporation
374 | Chapter 12: Interest Rate Forwards and Futures

informs both parties of the other’s intention to make or take delivery. The short has
to then prepare an invoice for the long that gives details about the security being de-
livered and the amount of payment for delivery. Finally, on the next day, which is the
delivery day, the short has to deliver the bonds to the long in exchange for the amount
mentioned in the invoice.
Although delivery continues until the end of the delivery month, trading in the
futures contracts ceases on the seventh business day prior to the last business day of
the delivery month.

The Wild Card Option

Before we examine it in detail, let’s first define the wild card option. It is an option that
gives the right to the short to decide whether to deliver, even after trading has ceased
for the day. The futures market closes at 2:00 p. m. However, the short has until 8:00
p. m. on a given day to declare an intention to deliver. Thus the option allows the short
the right to profit from a favorable price movement in the six-hour interval between
2:00 p. m. and 8:00 p. m.
The settlement price used to calculate the invoice price is the price that is deter-
mined at 2:00 p.m.on the intention day.14 However, the short has until 8:00 p. m. on
that day to notify the exchange of a decision to deliver. Thus the short has an option
to lock in the 2:00 p. m. price by announcing an intention to deliver any time before
8:00 p. m. This means that if interest rates change after 2:00 p. m. then the short can
profit by delivering a bond that is now cheaper to deliver.
In actual practice the short has a bouquet of wild card options, that is, one for
every potential intention day. On the first intention day,15 the short has the wild card
option described in the preceding paragraph. However, if there is no change in prices
between 2:00 p. m. and 8:00 p. m., then the short can simply wait for the next day
hoping that something will happen between 2:00 p. m. and 8 p. m. on that day. This
can go on until the last intention day which is the third to last business day of the
expiration month.
In practice there are the end-of-day wild card options that we just described and
also an end-of-month wild card option. The end-of-month option works as follows.
The final settlement price is determined seven days before the end of the contract
month. However, the short can wait until the end of the contract month to make a
call on whether to deliver.
The wild card option has a timing component as well as a quality component.
Let’s first illustrate the timing option component of the wild card option.

14 This price changes from day to day during the delivery period until the last day of trading. For all
subsequent deliveries, the settlement price is the price as of the last trading day.
15 This is the second to last business day of the month preceding expiration.
Seller’s Options | 375

The Timing Option

Consider the following time line.

t−1 t t1 t1 + 1 t1 + 2 T t2

C/2 C/2

Figure 12.8: The delivery process.

t−1 denotes the time when the last coupon was paid, and t2 denotes the next coupon
date. The contract expires at T.
Let’s consider the case of an investor who gets into a cash-and-carry strategy at
time t, by buying 1 unit of bond i and going short in CF i futures contracts, where CF i
is the conversion factor of the bond that has been bought.
Assume that the investor decides to deliver at time t1 + 2 by declaring an intention
to deliver at time t1 . We assume that bond i, the bond in question, is the cheapest to
deliver at that point in time. Therefore, the futures settlement price at time t1 , Ft1 , is
such that

Pi,t1
F t1 =
CF i

where Pi,t1 is the quoted price of bond i at time t1 .

If there is no timing option, then bond i indeed is delivered, and the proceeds from
delivery under the futures contract are

CF i × (Ft1 × CF i + AI i,t−1 ,t1 +2 ) × 100,000

The proceeds from the sale of the surplus bonds16 are

(1 − CF i )(Pi,t1 + AI i,t−1 ,t1 +1 ) × 100,000

The total proceeds are17

CF i × (Ft1 × CF i + AI i,t−1 ,t1 +2 ) × 100,000 + (1 − CF i )(Pi,t1 + AI i,t−1 ,t1 +1 ) × 100,000

= CF i × (Pi,t1 + AI i,t−1 ,t1 +2 ) × 100,000 + (1 − CF i )(Pi,t1 + AI i,t−1 ,t1 +1 ) × 100,000
= (Pi,t1 + AI i,t−1 ,t1 +1 ) × 100,000 (12.16)

16 If CF i > 1, additional bonds have to be purchased if the short decides to deliver, which means that
there will be an outflow.
17 AI i,t−1 ,t1 +2 ≅ AI i,t−1 ,t1 +1 .
376 | Chapter 12: Interest Rate Forwards and Futures

Now assume that between 2:00 p. m. and 8:00 p. m. on day t1 , the YTM changes.
Let the corresponding spot price of bond i be Pi,t
∗
1
. We assume that bond i continues to
be the cheapest-to-deliver bond, in order for us to be able to focus exclusively on the
timing option.
The proceeds from delivery under the futures contract are the same, namely

CF i × (Pi,t1 + AI i,t−1 ,t1 +2 ) × 100,000

But the proceeds from the sale of the surplus bonds are

(1 − CF i )(Pi,t
∗
1
+ AI i,t−1 ,t1 +1 ) × 100,000

The total proceeds therefore are

CF i × (Pi,t1 + AI i,t−1 ,t1 +2 ) × 100,000 + (1 − CF i )(Pi,t

∗
1
+ AI i,t−1 ,t1 +1 ) × 100,000

The incremental profit from the yield change is

CF i × (Pi,t1 + AI i,t−1 ,t1 +2 ) × 100,000 + (1 − CF i )(Pi,t

∗
1
+ AI i,t−1 ,t1 +1 ) × 100,000
− (Pi,t1 + AI i,t−1 ,t1 +1 ) × 100,000
= (1 − CF i )(Pi,t
∗
1
− Pi,t1 ) × 100,000 (12.17)

If CF i > 1, the incremental profit will be positive if Pi,t

∗
1
< Pi,t1 , that is, if yields rise.
Thus, in the case of a bond whose conversion factor is greater than one, the timing
option is beneficial if yields rise. Conversely, if the conversion factor is less than one,
then the yield has to fall in order to enable the investor to profit from the timing option.
We can illustrate this principle with the help of Example 12.12.

Example 12.12. Assume that there are two bonds that are eligible for delivery on 7 September 2018.
Bond A carries a 5% coupon and matures on 15 May 2047. Bond B carries a 11% coupon and matures
on 15 November 2038. The conversion factor for bond A is 0.8642, and that for bond B is 1.5777. If we
assume that the YTM for both the bonds is 6%, then the quoted spot price for bond A is 86-12, and that
for bond B is 158-02. The delivery-adjusted spot price of bond A is 99-30, and that of bond B is 100-06.
Bond A is cheaper to deliver and the futures price at 2:00 p. m. is equal to its delivery-adjusted spot
price of 99-30.
Let’s assume that an investor has initiated a cash-and-carry strategy on 7 August 2018 with bond
A and suddenly announces an intention to deliver on 7 September 2018. In the absence of a timing
option, the investor would lock in
30 5
99 + 32 2 117
× 0.8642 × 100,000 + × × 100,000
100 100 184
= $87,955.65

Let’s assume that after 2:00 p. m., the YTM suddenly falls to 4.50%. The quoted price of bond A is now
$108. Now, if bond A is delivered, the proceeds are
30 5
99 + 32 2 117
0.8642 × ( × 0.8642 + × ) × 100,000
100 100 184
Seller’s Options | 377

5
108 116
+ (1 − 0.8642) ( + 2 × ) × 100,000
100 100 184
= 76,011.2720 + 14,880.4310
= $90,891.7030

Thus the timing option has clearly paid off.

The Quality Option

Let’s examine the wild card option again. Assume that after 2:00 p. m. on 7 September,
the interest rate suddenly rises to 7.50%. The new spot price of bond A becomes 70-22,
whereas that of bond B is 136-03. The timing option has no value in this case as one
can verify. But let’s consider what happens if we deliver bond B instead of bond A. If
bond A is delivered in the absence of the wild card option, then the payoff is $87,955.65.
However, if bond A is sold at the new price and CF a units of bond B are purchased for
delivery under the futures contract, the proceeds are as follows.
Payoff from delivery under the futures contract:
30
99 + 32 5.5 117
= 0.8642 × ( × 1.5777 + × ) × 100,000
100 100 184
= $139,281.95

Proceeds from the sale of bond A:

22
70 + 32 2.5 116
=( + × ) × 100,000
100 100 184
= $72,263.59

Cost of acquisition of bond B:

03
136 + 32 5.5 116
= 0.8642 ( + × ) × 100,000
100 100 184
= $120,608.70

The net proceeds = 139,281.95 + 72,263.59 − 120,608.70 = $90,936.82, which is greater

than $87,955.65, the amount that would have been received in the absence of the wild
card option. The quality option is clearly valuable in this case.

The End-of-Month Option

We have already examined the wild card option inherent in the end-of-month option.
There is, however, another dimension to this option. Let’s assume that interest rates
378 | Chapter 12: Interest Rate Forwards and Futures

are stable for the days after the last day of trading. Under these conditions, for every
additional day while holding the bond, the short earns accrued interest, but must fi-
nance it for every day it is held. If the coupon rate of the bond exceeds the financing
rate, then the short should deliver on the last day, or else immediately. This timing
option is known as the accrued interest option.

The Pure Quality Option

We have seen that there is an end-of-day wild card option as well as an end-of-month
wild card option, both of which have a quality option component. But even in the ab-
sence of these options, the short can decide which bond to deliver, which itself is a
quality option. This is because the party who has the right to choose the grade stands
to benefit if another grade is cheaper to deliver subsequently. Thus, even in the ab-
sence of the wild card or the end-of-month options, a cash-and-carry strategy gives a
quality option to the short.

Hedging
Let’s first illustrate our arguments using the cheapest-to-deliver bond and then discuss
the case where other bond portfolios have to be hedged.

Hedging the Cheapest-to-Deliver Bond: A Naive Approach

A naive approach for hedging Q bonds is to use Q futures contracts, or in other words, a
hedge ratio of 1:1. This strategy is not satisfactory, as the following example illustrates.

Example 12.13. Let’s go back to our earlier illustration in Example 12.11 regarding a cash-and-carry
strategy initiated on 14 August. We found that the bond that carried a coupon of 6.5% was the
cheapest-to-deliver bond. The quoted spot price was 132-11, and the corresponding futures price was
124-08. The conversion factor was 1.0683. Assume that on 15 September, the settlement price is
82-19. The corresponding quoted spot price is 88-07. We are assuming that this bond continues to be
the cheapest-to-deliver bond.
The bond can be sold in the spot market to yield

7
88 + 32 3.25 123
× 100,000 + × × 100,000
100 100 184
= $90,391.30

The profit from the short futures position is

8 19
(124 + 32
) − (82 + 32
)
× 100,000
100
= $41,656.25
Hedging | 379

The total payoff is $132,047.55. A perfect hedge would lock in the original futures price of 124-08 to
yield
8
124 + 32 3.25 123
× 1.0683 × 100,000 + × × 100,000
100 100 184
= $134,908.82

In this case we are clearly under-hedged.

The Conversion Factor Approach

Let’s assume that the hedge is initiated at time t0 and lifted at time t1 , after delivery
has commenced:
(Pt0 + AI t−1 ,t0 )(1 + r) − AI t−1 ,t1
Ft0 =
CF i
P t1
Ft1 =
CF i
Pt1 − (Pt0 + AI t−1 ,t0 )(1 + r) + AI t−1 ,t1
⇒ Ft1 − Ft0 =
CF i
Pt1 − Pt0
⇒ Ft1 − Ft0 ≅
CF i
△P
⇒ △F ≅ (12.18)
CF i
if we ignore the cost of carry.
Now a perfect hedge should be such that

△P = h∗ △ F
Qf
where h∗ = Q
is the optimal hedge ratio.18 Therefore,

△P
h∗ = = CF i (12.19)
△F
Let’s examine the efficiency of this hedge using the same data as for the example on
the naive hedging strategy.

Example 12.14. The proceeds from the spot market when the cheapest-to-deliver bond is sold are
$90,391.30.
Profit from the futures market = 1.0683 × 41,656.25 = $44,501.37.
The total proceeds = $134,892.67, which is very close to the value of $134,908.82 that is implied
by the original futures price.

18 Q is the exposure in the spot market and Qf is the number of futures contracts.
380 | Chapter 12: Interest Rate Forwards and Futures

Hedging a Portfolio Other Than the CTD Bond

The naive hedging strategy has been discredited even for hedging the price of the CTD
bond. Hence we will not pursue it further. Let’s instead analyze how to extend the
conversion factor approach to hedge a bond other than the CTD bond.
Our hedge ratio h should be such that △P = h △ F. We know that

△PCTD
△F ≅
CF CTD

Therefore

△PCTD △P
△P = h ⇒ h = CF CTD . (12.20)
CF CTD △PCTD

Such hedge ratios are called perturbation hedge ratios.19 To use these ratios, for a given
change in yield, we have to calculate the ratio of the change in the price of the bond
being hedged to the change in the price of the CTD bond. In practice, the problem is
that this ratio depends on the change in the yield, which cannot be predicted exactly.
Hence for hedging a portfolio of bonds other than the CTD bond, the preferred hedging
technique involves the use of the duration of the bond.
The hedge ratio, which is derived in Appendix 12.1, is given by
yCTD
Dh × Ph × (1 + 2
)
CF CTD × yh
DCTD × PCTD × (1 + 2 )

The number of futures contracts required is

Face Value of Spot Exposure

×h
Face Value of the Bond Underlying the Futures Contract

We next illustrate this hedging technique with the help of Example 12.15.

Example 12.15. Assume that today is 14 August 2018. September futures contracts expire on 30
September. The cheapest-to-deliver bond, as we saw earlier, is a 6.5% coupon bond maturing on 15
November 2047. Its quoted price is 132-11, which corresponds to a YTM of 4.5% per annum, and the
conversion factor is 1.0683. The quoted futures price, as calculated before, is 124-08.
Consider a portfolio manager who is holding 10,000 IBM bonds maturing on 15 August 2047.
The face value is $100, the coupon rate is 7.50% per annum, and the YTM is 8.50% per annum. The
manager plans to sell the bonds on 30 September and wants to protect against an increase in the yield
using T-bond futures contracts. The duration of the CTD bond on 14 August is 15.1947 years. The dirty
price is $133.9688. The price of the IBM bonds is $89.2877, and the corresponding duration is 11.3778
years.

19 See Blake [4].

Hedging | 381

The hedge ratio is therefore

(11.3778 × 89.2877 × 1.0225)

1.0683 × = 0.5229
(15.1947 × 133.9688 × 1.0425)

Let’s assume that the YTM of the IBM bonds on 30 September is 10% per annum and that the
YTM of the CTD bond is 6%. The price of the IBM bond therefore is $77.4141, and the futures price is
$100.0045. If the yield remains at 8.50%, the dirty price of IBM is $90.2216.
The loss from the spot market is

10,000 × (90.2216 − 77.4141) = ($128,075.00)

The profit from the futures market is

10,000 × 0.5229 × (124.25 − 100.0045) = $126,779.70

The net loss = 126,779.70 − 128,075.00 = ($1,295.28).

1,295.28
The tracking error = 128,075.00 ≡ 1.01%.

One issue that arises in the context of the duration-based hedging approach is whether
we should use input values as calculated at the inception of the hedge or as deter-
mined at the point of termination. Most authors argue that we should use the expected
values of the variables as of the termination date of the hedge. However, in practice,
forecasting these variables is not always easy, and consequently, the current values are
often used as inputs. The two approaches do not yield substantially different results
unless the instrument being hedged and the CTD bond are significantly different.20

Changing the Duration of a Portfolio of Bonds

Consider a portfolio of bonds that currently has a duration of Dh . Let’s assume that
we want to change its duration to DT , which denotes the target duration, by going
long in futures contracts. The question is, what is the appropriate number of futures
contracts to use? Let’s denote the required hedge ratio by h, the current value per unit
of the bond being hedged by Ph , and the value of the overall portfolio consisting of the
bonds and the futures contracts by V. It can be shown that21
yCTD
(DT − Dh ) × Ph × (1 + 2
)
⇒h= yh × CF CTD (12.21)
DCTD × PCTD × (1 + 2
)

The required number of futures contracts is

Face Value of Spot Exposure
×h
Face Value of the Bond Underlying the Futures Contract

20 See Daigler [18].

21 See Appendix 12.2.
382 | Chapter 12: Interest Rate Forwards and Futures

Here is a numerical illustration.

Example 12.16. Assume that we are on 14 August 2018 and want to increase the duration of the IBM
bonds from 11.3778 to 15. If we are holding IBM bonds with a face value of $100 million we need to go
long in

(15.0 − 11.3778) × 89.2877 × 1.0225 100,000,000

1.0683 × ×
15.1947 × 133.9688 × 1.0425 100,000
= 0.1665 × 1,000 ≃ 167 contracts

If we want to reduce the duration, we have to go short in futures contracts.

Chapter Summary
In this chapter we looked at interest rate forward and futures contracts. We began
by considering forward rate agreements, and their use in hedging and speculation.
The focus then shifted to Eurodollar futures and their uses from the standpoints of
locking in borrowing and lending rates. We also briefly looked at federal funds and
futures on them. The chapter concluded with a detailed study of Treasury bond and
Treasury note futures. We examined issues pertaining to their valuation, the issue of
conversion factors, and the concept of the cheapest-to-deliver (CTD) bond. We looked
at the various options given to the sellers of T-note and T-bond futures. Finally we
studied the use of T-bond futures to hedge corporate bonds and their usefulness in
changing the duration of a bond.

Appendix 12.1: Duration-Based Hedge Ratio

For any bond, we know that
dP −D
= × dy
P (1 + y2 )

where D is the duration of the bond in annual terms.22 Therefore

△P −D
≅ × △y
P (1 + y2 )

If we denote the bond being hedged as bond h, then the optimal hedge ratio is

× Ph × △yh
−Dh
y
△Ph (1+ 2h )
CF CTD = CF CTD ×
× PCTD × △yCTD
△PCTD −DCTD
y
(1+ CTD
2
)

22 We are assuming that the bond pays semiannual coupons.

Appendix 12.2: Required Number of Contracts to Change the Duration | 383

yCTD
Dh × Ph × (1 + 2
) × △yh
= CF CTD × yh (12.22)
DCTD × PCTD × (1 + 2 ) × △yCTD

If we assume that yield curve movements are parallel, that is △yCTD = △yh , then the
hedge ratio is23
yCTD
Dh × Ph × (1 + 2
)
CF CTD × yh
DCTD × PCTD × (1 + 2 )

which simplifies further to

Dh × Ph
CF CTD ×
DCTD × PCTD

if we assume that percentage yield curve movements are parallel.24 The number of
futures contracts required is
Face Value of Spot Exposure
×h
Face Value of the Bond Underlying the Futures Contract

Appendix 12.2: Required Number of Contracts to Change the

Duration
We know that

△ V = △P + h △ F
−DV
⇒ × V × △yV
(1 + y2V )
DCTD
y × PCTD × △yCTD
−Dh (1+ CTD )
= y × Ph × △yh − h × 2

(1 + 2h ) CF CTD

Let’s assume that △yv = △yh = △yCTD . Therefore

DCTD
y × PCTD
−DV −Dh (1+ CTD )
×V = × Ph − h × 2

(1 + y2V ) y
(1 + 2h ) CF CTD

At the inception of the hedge, V = Ph because the value of the futures contracts is zero.
Thus the preceding expression can be written as
DCTD
× PCTD
−(DV − Dh )
y
(1+ CTD )
y × Ph = − h × 2
.
(1 + 2h ) CF CTD

23 The per unit face value of the bond being hedged should be taken to be equal to the face value of
the bond underlying the futures contract.
24 This implies that (1+yCTD ) = (1+yh ) .
△y △y
CTD h
384 | Chapter 12: Interest Rate Forwards and Futures

If our target duration for the overall portfolio is DT , then we should choose h such
that

× Ph
(DT −Dh )
y
(1+ 2h )
h= DCTD
× CF CTD
y × PCTD
(1+ CTD
2
)
yCTD
(DT − Dh ) × Ph × (1 + 2
)
⇒h= yh × CF CTD (12.23)
DCTD × PCTD × (1 + 2
)

The required number of futures contracts is

Face Value of Spot Exposure

×h
Face Value of the Bond Underlying the Futures Contract
Chapter 13
Bonds with Embedded Options
In this chapter we take a detailed look at bonds with built-in options. More specifically,
we examine the intricacies of bonds with built-in call and put options and bonds that
give the holders the option to convert them into shares of stock. Callable bonds, which
have a built-in call option, give the issuer the right to call back the bond prior to the
maturity date. If the interest rates have declined, the issuer can exercise this option,
retire the existing bonds, and issue new bonds with a lower coupon, in accordance
with the prevailing low-interest environment. Putable bonds, which have a built-in
put option, enable the holders to prematurely surrender the bonds in return for the
face value. Quite obviously, if interest rates have increased, the holders can exercise
this option to recover their principal, and re-invest the money in bonds with a higher
coupon, which are likely to be available in the prevailing high-interest environment.
Convertible bonds can be converted into shares of common stock of the entity that is-
sued the bonds. Exchangeable bonds are similar, except that they facilitate conversion
into shares of another company and not the issuing company, although in practice the
two may be related.

Callable Bonds
In the case of a callable bond, the issuer has the option to call back the bond prior
to the scheduled maturity date. As the holder of the option, the issuer has to pay a
premium for it. This shows up in the form of a lower price compared to a plain vanilla
bond, which is identical in all other respects except for the call feature. In practice, a
party who is planning a bond issue has to offer a higher coupon on the callable bond,
as compared to an otherwise similar plain vanilla bond. Subsequently if we compare
a plain vanilla bond and a callable bond that are identical in all respects including the
coupon, the callable will have a lower price or a higher yield to maturity.
Why does the callable bond have a higher yield to maturity? The call option en-
ables the issuer to call the bond back when interest rate are declining. If the holder
gets his money back in such circumstances, he has to re-invest at a lower rate of inter-
est. Thus the call option works in favor of the issuer and against the investor. From the
investors’ perspective, they are perfectly happy receiving a higher coupon, and will
not appreciate being compelled to invest in lower return instruments. Thus the holder
of a callable bond faces uncertainty about the number of coupons he or she is going
to get, and the time when the principal will be repaid. This is referred to as timing risk.
In practice, to compensate for this, the issuer states in advance that it will pay a call
premium, over and above the face value, if the bond is recalled. The premium may be
equal to six month’s coupon, or at times may even be one year’s coupon.

https://doi.org/10.1515/9781547400669-013
386 | Chapter 13: Bonds with Embedded Options

A bond may be discretely callable or continuously callable. The former is callable

only on specific dates, typically coupon dates. Thus if today is a coupon date, and the
bond is not called, it cannot be called for another six months, assuming semiannual
coupons. Continuously callable bonds on the other hand, can be called at any point
in time. In options parlance, discretely callable bonds embody a Bermudan option
and are characterized by a finite number of call dates, whereas continuously callable
bonds have a built-in American option, which permits the issuer to call back at any
date. In practice, if we issue a 20-year bond and say that it can be recalled right away,
it is unlikely to have many takers. Consequently a callable bond may have a call pro-
tection period, during which it cannot be called, no matter how low the prevailing
interest rates may be. Such bonds are referred to as deferred callable bonds. Once the
call protection period expires, the bonds become freely callable.
If we denote the dirty price of the callable bond by Pcal , the dirty price of the cor-
responding plain vanilla bond by Ppv , and the premium for the call option by C, we
can state that:

Pcal = Ppv − C (13.1)

Yield to Call
Consider a bond with a face value of M, N semiannual coupon periods until maturity,
and a coupon rate of c% per annum, which translates to $C/2 every six months. As-
sume that the bond can be called at M∗ where M∗ = M + C or M + C2 . Let’s denote a
call date by N∗, where N∗ < N. We denote the dirty price by Pd . The yield to call is the
value of the discount rate that satisfies the following equation:

N∗ C
2 M∗
Pd = ∑ [ yc t
]+ yc t
(13.2)
t=1 (1 + 2
) (1 + 2
)

Let’s illustrate the computation of the yield to call using Excel.

Example 13.1. Consider a bond with a face value of $1,000 and eight years to maturity. It pays a coupon
of 8% per annum on a semiannual basis. It is callable after five years, just after the tenth coupon is
paid, and if called, one year’s coupon is paid as a call premium. If the yield to maturity is 10% per
annum, what is the yield to call? The first step is to compute the dirty price using the PV function in
Excel. The parameters are RATE, NPER, PMT, and FV. RATE is the semiannual YTM, which is 5%; NPER
is the number of semiannual periods until maturity, which is 16; PMT is the periodic coupon, which is
$40; and FV is the terminal cash flow, which is $1,000:

The price = PV(0.05, 16, −40, −1000) = $891.62

Given this information we can use the RATE function to determine the yield to call. The parameters
are NPER, PMT, PV, and FV. The call price is 1,000 + 80 = $1,080. Thus FV = (1,080). PV is the dirty
Yield to Call | 387

price, which is $891.62. PMT is −40, and NPER is equal to 10, since the bond is callable after five years.
Note that PMT and FV should have the same sign, whereas PV should have the opposite sign. This is
because the first two are inflows for the bond holder, and the third is an outflow. Thus we need to
invoke the RATE function as follows:

= RATE(10, −40, 891.62, −1080) ∗ 2 = 12.1658%

We need to multiply the answer by two because the RATE function gives the semiannual yield to
call. This is because time is being measured in semiannual periods.
For a bond with multiple call dates, there is a yield to call for each such date. In Excel we can easily
compute the yields to call for multiple call dates.

Relationship between the Yield to Call and the Yield to Maturity

Let’s first consider a callable bond without a call premium. If the bond is trading at
a discount, and we use a discount rate equal to the YTM, the price increases as we
reduce the number of periods. As we saw in Chapter 2, this is because of the pull to
par effect. For the price to remain the same, we need to use a higher discount rate.
Thus the YTC will be higher than the YTM. If a call premium is applicable, the YTC will
be even higher.
Now consider a bond trading at par. If there is no call premium, the yield to call
is equal to the yield to maturity. If there is a call premium, however, the yield to call
becomes greater than the YTM because of a larger terminal cash flow.
Finally let’s consider a bond that is trading at a premium. First assume that there
is no call premium. If we use a discount rate equal to the YTM, the price decreases as
we reduce the number of periods. For the price to remain the same, we need to use a
lower discount rate. Thus the YTC will be less than the YTM. However, if we build in
a call premium, the result is ambiguous, for the YTC may be less than or greater than
the YTM because the terminal cash flow used for computing the former is greater than
that for the latter. Let’s illustrate these concepts using Example 13.2.

Example 13.2. Consider a bond with eight years to maturity, a coupon of 8% per annum, paid semi-
annually, and a YTM of 10% per annum. The face value is $1,000. The dirty price as obtained earlier is
$891.62. If the bond is called after five years, and there is no call premium, the YTC is given by

= RATE(10, −40, 891.62, −1000) ∗ 2 = 10.8664%

If a call premium equal to one year’s coupon is built in, the YTC is given by

= RATE(10, −40, 891.62, −1080) ∗ 2 = 12.1658%

Thus in both the cases the YTC is greater than the YTM.
Now assume the YTM is 8% per annum. The price is obviously $1,000. If the bond is called after
five years, and there is no call premium:

= RATE(10, −40, 1000, −1000) ∗ 2 = 8%

388 | Chapter 13: Bonds with Embedded Options

Thus the YTC is equal to the YTM.

If a call premium equal to one year’s coupon is built in, the YTC is given by

= RATE(10, −40, 1000, −1080) ∗ 2 = 9.2932%.

As is to be expected the YTC is greater than the YTM.

Finally assume the YTM is 6% per annum. The dirty price is $1,125.61. If the bond is called after
five years, and there is no call premium, the YTC is given by:

= RATE(10, −40, 1125.61, −1000) ∗ 2 = 5.1206%.

Thus as expected the YTC is less than the YTM.

If a call premium equal to one year’s coupon is built in, the rate is given by:

= RATE(10, −40, 1125.61, −1080) ∗ 2 = 6.4069%.

The YTC is higher than the YTM.

However what if the bond is called after three years.

= RATE(6, −40, 1125.61, −1080) ∗ 2 = 5.8520%.

Hence the YTC for premium bonds may be greater than or less than the YTM, if a call premium is built
in, depending on the call date.

The Approximate Yield to Call Approach

The approximate yield to call may be defined as1

C+ (M ∗ −P)
N ∗ /2
AYC = (13.3)
(M ∗ + P)/2

Example 13.3. Let’s reconsider the data used in Example 13.1. That is, the time to maturity is eight
years, the face value is $1,000, and the coupon is 8% per annum. The price as we demonstrated earlier
is $891.62.
80 + (1,080 − 891.62)/5
AYC = = 11.9370%
(1,080 + 891.62)/2
Let’s consider a higher value, namely 12.50%, and a lower value, 11.50%:
The price using a rate of 11.50% is = PV(0.0575, 10, −40, −1080) = $915.40.
The price using a rate of 12.50% is = PV(0.0625, 10, −40, −1080) = $879.97.
Now let’s interpolate:

0.1150 − 0.1250 ≡ 915.40 − 879.97 and 0.1150 − X ≡ 915.40 − 891.62

23.78
Thus X = 0.1150 + 0.01 × = 12.1712%
35.43
The exact value from the RATE function in Excel is 12.1658%. Thus the approximation is excellent.

1 M ∗ is the call price; N ∗ is the call date; C is the annual coupon, and P is the dirty price.
Reinvestment Assumption | 389

Reinvestment Assumption
The yield to maturity calculation is based on the assumption that the bond is held
until maturity and all intermediate cash flows are reinvested at the yield to maturity.
Similarly the yield to call is calculated based on the assumption that the bond is held
until the call date and all the intermediate cash flows are reinvested at the yield to call.
Example 13.4 demonstrates this.

Example 13.4. Consider the data used in Example 13.1. The dirty price is $891.62, the call price is
$1,080, the periodic coupon is $40, and the call date is 10 semiannual periods away. Assume all inter-
mediate cash flows can be reinvested at 12.1658% per annum. The future value of 10 coupons is given
by:

= FV(0.121658/2, 10, −40) = $529.29

Thus the terminal cash flow is 529.29 + 1,080 = $1,609.29. As the initial investment is 891.62, the rate
of return is given by

891.62 × (1 + i)10 = 1,609.29 ⇒ i = 6.0829%

which corresponds to an annual rate of 12.1658%. Thus in order to earn the yield to call corresponding
to a particular call date, we need to hold the bond until the call date and reinvest all the intermediate
cash flows at the yield to call.

Concept of the Yield to Worst

If a bond has a number of call dates, we can compute the yield to call for each call
date, as well as the yield to maturity. The lowest of these values is termed the Yield to
Worst (YTW). The problem with this measure is that it is a function of the bond price.
Consider a five-year bond that pays annual coupons at the rate of 6% per annum.
The face value is $1,000. We assume that the bond is callable at the end of the first,
second, third, or fourth year. If called after one year, $80 is paid as a call premium.
Every subsequent call date the call premium is assumed to decline by $20. Let’s con-
sider three different price scenarios: a par bond, a discount bond with a dirty price of
$880, and a premium bond with a dirty price of $1,120. The results are summarized in
Table 13.1. All the results have been obtained using the RATE function in Excel.
As we can see, if the bond is trading at a discount, the yield to worst is 9.9025%,
which corresponds to the YTM. The same is the case if the bond is trading at par. How-
ever if the bond is trading at a premium, the yield to worst is 1.7857%, which is the
yield to call corresponding to the first call date.
390 | Chapter 13: Bonds with Embedded Options

Table 13.1: The yields for different price scenarios.

Time to Redemption Dirty Price

Redemption Price 880 Yield 1,000 Yield 1,120 Yield

5 1,000 9.0925% 6.0000% 3.3533%

4 1,020 10.2343% 6.4541% 3.2303%
3 1,040 12.1962% 7.2413% 3.0473%
2 1,060 16.2758% 8.8726% 2.7144%
1 1,080 29.5455% 14.0000% 1.7857%

Valuation of a Callable Bond

Let’s value a callable bond using the Ho-Lee and BDT interest rate trees that we ob-
tained in Chapter 11. Let’s first reproduce the Ho-Lee tree, and the corresponding state
prices.

Figure 13.1: The Ho-Lee no-arbitrage interest rate tree.

The state price tree is shown in Figure 13.2.

Consider a four-period bond with a face value of $1,000 that pays a coupon of $40
every period. Using the state price tree obtained by calibrating the Ho-Lee model, we
can compute the value of this bond as follows:
1,040 × (0.053160 + 0.215741 + 0.328336 + 0.222092 + 0.056336)
Valuation of a Callable Bond | 391

t0 t1 t2 t3 t4

0.053160
0.112320
0.232247 0.215741
0.481928 0.340258
1.0 0.466755 0.328336
0.481928 0.343589
0.234508 0.222092
0.115651
0.056336

Figure 13.2: State price tree.

+ 40 × (0.112320 + 0.340258 + 0.343589 + 0.115651)

+ 40 × (0.232247 + 0.466755 + 0.234508) + 40 × (0.481928 + 0.481928)
= 1,040 × 0.875665 + 40 × 0.911818 + 40 × 0.933511 + 40 × 0.963856
= $1,023.0589

Now assume that this bond is callable, and if called, a call premium of $40 is paid.
At time t4 , the cash flow is 1,000 + 40 = $1,040. The value at t3 node-1 is

1,040
0.112863
+ 40 = 984.4462 + 40 = $1,024.4462
(1 + 2
)

If the bond is called, the issuer has to pay 40 + 40 + 1,000 = $1,080. Thus the bond
will not be recalled.
The value at node-2 is
1,040
0.092863
+ 40 = 993.8539 + 40 = $1,033.8539
(1 + 2
)

If the bond is called, the issuer has to pay $1,080. Thus the bond will not be recalled.
The value at node-3 is
1,040
0.072863
+ 40 = 1,003.4430 + 40 = $1,043.4430
(1 + 2
)

If the bond is called, the issuer has to pay $1,080. Thus the bond will not be recalled.
The value at node-4 is
1,040
0.052863
+ 40 = 1,013.2190 + 40 = $1,053.2190
(1 + 2
)

If the bond is called, the issuer has to pay $1,080. Thus the bond will not be recalled.
392 | Chapter 13: Bonds with Embedded Options

Now let’s move to time t2 . The value at node-1 is

1,024.4462 1,033.8539
0.5 × 0.067725
+ 0.5 × 0.067725
+ 40 = 495.4461 + 499.9958 + 40
(1 + 2
) (1 + 2
)
= $1,035.4419

If the bond is called, the issuer has to pay $1,080. Thus the bond will not be recalled.
The value at node-2 is
1,033.8539 1,043.4430
0.5 × 0.047725
+ 0.5 × 0.047725
+ 40 = 504.8793 + 509.5621 + 40
(1 + 2
) (1 + 2
)
= $1,054.4414

If the bond is called, the issuer has to pay $1,080. Thus the bond will not be recalled.
The value at node-3 is
1,043.4430 1,053.2190
0.5 × 0.027725
+ 0.5 × 0.027725
+ 40 = 514.5880 + 519.4092 + 40
(1 + 2
) (1 + 2
)
= $1,073.9972

If the bond is called, the issuer has to pay $1,080. Thus the bond will not be recalled.
Now let’s move to time t1 . The value at node-1 is

1,035.4419 1,054.4414
0.5 × 0.075063
+ 0.5 × 0.075063
+ 40 = 498.9930 + 508.1491 + 40
(1 + 2
) (1 + 2
)
= $1,047.1421

If the bond is called, the issuer has to pay $1,080. Thus the bond will not be recalled.
The value at node-2 is
1,054.4414 1,073.9972
0.5 × 0.055063
+ 0.5 × 0.055063
+ 40 = 513.0944 + 522.6104 + 40
(1 + 2
) (1 + 2
)
= $1,075.7048

If the bond is called, the issuer has to pay $1,080. Thus the bond will not be recalled.
The value at time t0 is

1,047.1421 1,075.7048
0.5 × 0.075
+ 0.5 × 0.075
= 504.6468 + 518.4120
(1 + 2
) (1 + 2
)
= $1,023.0587

Since the bond is not recalled at any of the subsequent nodes, the value of the
callable bond is equal to that of the plain vanilla bond. In other words, the call option
has a value of zero.
Valuation of a Callable Bond | 393

Now let’s increase the coupon to 10% per annum. The value of the plain vanilla
bond is

1,050 × (0.053160 + 0.215741 + 0.328336 + 0.222092 + 0.056336)

+ 50 × (0.112320 + 0.340258 + 0.343589 + 0.115651)
+ 50 × (0.232247 + 0.466755 + 0.234508) + 50 × (0.481928 + 0.481928)
= 1,050 × 0.875665 + 50 × 0.911818 + 50 × 0.933511 + 50 × 0.963856
= $1,059.9075

Now consider the callable bond. At time t4 , the cash flow is 1,000 + 50 = $1,050.
The value at t3 node-1 is
1,050
0.112863
+ 50 = 993.9121 + 50 = $1,043.9121
(1 + 2
)

If the bond is called, the issuer has to pay 50 + 50 + 1,000 = $1,100. Thus the bond will
not be recalled.
The value at node-2 is
1,050
0.092863
+ 50 = 1,003.4101 + 50 = $1,053.4101
(1 + 2
)

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
The value at node-3 is
1,050
0.072863
+ 50 = 1,013.0915 + 50 = $1,063.0915
(1 + 2
)

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
The value at node-4 is
1,050
0.052863
+ 50 = 1,022.9615 + 50 = $1,072.9615
(1 + 2
)

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
Now let’s move to time t2 . The value at node-1 is
1,043.9121 1,053.4101
0.5 × 0.067725
+ 0.5 × 0.067725
+ 50 = 504.8602 + 509.4538 + 50
(1 + 2
) (1 + 2
)
= $1,064.3140

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
The value at node-2 is
1,053.4101 1,063.0915
0.5 × 0.047725
+ 0.5 × 0.047725
+ 50 = 514.4295 + 519.1574 + 50
(1 + 2
) (1 + 2
)
= $1,083.5869
394 | Chapter 13: Bonds with Embedded Options

If the bond is called, the issuer will have to pay $1,100. Thus the bond will not be
recalled.
The value at node-3 is

1,063.0915 1,072.9615
0.5 × 0.027725
+ 0.5 × 0.027725
+ 50 = 524.2779 + 529.1455 + 50
(1 + 2
) (1 + 2
)
= $1,103.4234

If the bond is called, the issuer has to pay $1,100. Thus the bond will be recalled, and
the value will be $1,100.
Now let’s move to time t1 . The value at node-1 is

1,064.3140 1,083.5869
0.5 × 0.075063
+ 0.5 × 0.075063
+ 50 = 512.9068 + 522.1947 + 50
(1 + 2
) (1 + 2
)
= $1,085.1015

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
The value at node-2 is

1,083.5869 1,100.0000
0.5 × 0.055063
+ 0.5 × 0.055063
+ 50 = 527.2767 + 535.2634 + 50
(1 + 2
) (1 + 2
)
= $1,112.5401

If the bond is called, the issuer will have to pay $1,100. Thus the bond will be recalled.
The value at time t0 is

1,085.1015 1,100.0000
0.5 × 0.075
+ 0.5 × 0.075
= 522.9405 + 530.1205
(1 + 2
) (1 + 2
)
= $1,053.0610

Because the value of the plain vanilla bond is $1,059.9075, the value of the call option
inherent in the callable bond is

1,059.9075 − 1,053.0610 = $6.8465

Putable Bonds
In the case of a putable bond, the holder has the option to turn in the bond prior to
the scheduled maturity date. Being in possession of the option, the holder has to pay a
premium for it. This shows up in the form of a higher price, compared to a plain vanilla
bond, which is identical in all other respects except for the put feature. In practice, a
party that is planning a bond issue can offer a lower coupon on the putable bond,
Valuation of a Putable Bond | 395

as compared to an otherwise similar plain vanilla bond. If we compare a plain vanilla

and a putable bond that are identical in all respects including the coupon, the putable
will have a higher price or a lower yield to maturity.
Why does the putable bond have a higher price? The put option enables the holder
to turn in the bond in a rising interest rate environment. If the issuer gets the bond
back in such circumstances, it has to reissue new bonds with a higher coupon, as the
market rates have increased. Thus the put option works in favor of the investor and
against the issuer. Issuers are perfectly happy paying a lower coupon and do not want
to reissue bonds with a higher coupon rate, which is a consequence of the investors
exercising their put options. Thus issuers of putable bonds face uncertainty about the
number of coupons they have to pay and the time when the principal has to be repaid.
This too is a manifestation of timing risk.
If we denote the dirty price of the putable bond by Pput , the dirty price of the corre-
sponding plain vanilla bond by Ppv , and the premium for the put option by P, we can
state that

Pput = Ppv + P (13.4)

Valuation of a Putable Bond

Consider a four-period bond with a face value of $1,000 that pays a coupon of $40
every period, and let’s use the Ho-Lee tree to price it. The price of a plain vanilla bond,
as shown earlier, is $1,023.0589. Now consider a bond that allows the holder to put
back at par. Obviously a put option is exercised at a node only if the model value is
lower than the price at which it can be put back.
At time t4 , the cash flow is 1,000 + 40 = $1,040. The value at t3 node-1 is

1,040
0.112863
+ 40 = 984.4462 + 40 = $1,024.4462
(1 + 2
)

If the bond is put back, the issuer has to pay $1,040. Thus the bond will be put back
by the investor.
The value at node-2 is
1,040
0.092863
+ 40 = 993.8539 + 40 = $1,033.8539
(1 + 2
)

If the bond is put back, the issuer has to pay $1,040. Thus the bond will be put back
by the investor.
The value at node-3 is
1,040
0.072863
+ 40 = 1,003.4430 + 40 = $1,043.4430
(1 + 2
)
396 | Chapter 13: Bonds with Embedded Options

If the bond is put back, the issuer has to pay $1,040. Thus the bond will not be put
back by the investor.
The value at node-4 is
1,040
0.052863
+ 40 = 1,013.2190 + 40 = $1,053.2190
(1 + 2
)

If the bond is put back, the issuer has to pay $1,040. Thus the bond will not be put
back by the investor.
Now let’s move to time t2 . The value at node-1 is
1,040.0000 1,040.0000
0.5 × 0.067725
+ 0.5 × 0.067725
+ 40 = 502.9682 + 502.9682 + 40
(1 + 2
) (1 + 2
)
= $1,045.9364

If the bond is put back, the issuer has to pay $1,040. Thus the bond will not be put
back.
The value at node-2 is
1,040.0000 1,043.4430
0.5 × 0.047725
+ 0.5 × 0.047725
+ 40 = 507.8807 + 509.5621 + 40
(1 + 2
) (1 + 2
)
= $1,057.4428

If the bond is put back, the issuer has to pay $1,040. Thus the bond will not be put
back.
The value at node-3 is
1,043.443 1,053.2190
0.5 × 0.027725
+ 0.5 × 0.027725
+ 40 = 514.5880 + 519.4092 + 40
(1 + 2
) (1 + 2
)
= $1,073.9972

If the bond is put back, the issuer has to pay $1,040. Thus the bond will not be put
back.
Now let’s move to time t1 . The value at node-1 is
1,045.9364 1,057.4428
0.5 × 0.075063
+ 0.5 × 0.075063
+ 40 = 504.0504 + 509.5955 + 40
(1 + 2
) (1 + 2
)
= $1,053.6459

If the bond is put back, the issuer has to pay $1,040. Thus the bond will not be put
back.
The value at node-2 is
1,057.4428 1,073.9972
0.5 × 0.055063
+ 0.5 × 0.055063
+ 40 = 514.5549 + 522.6104 + 40
(1 + 2
) (1 + 2
)
= $1,077.1653
Pricing the Callable and Putable Bonds Using the BDT Model | 397

If the bond is put back, the issuer has have to pay $1,040. Thus the bond will not be
put back.
The value at time t0 is

1,053.6459 1,077.1653
0.5 × 0.075
+ 0.5 × 0.075
= 507.7812 + 519.1158 = $1,026.8970
(1 + 2
) (1 + 2
)

The price of the plain vanilla bond is $1,023.0589. Thus the value of the put option is

P = 1,026.8970 − 1,023.0589 = $3.8381

Pricing the Callable and Putable Bonds Using the BDT Model
Let’s first reproduce the BDT interest rate tree.

Figure 13.3: No-arbitrage interest rate tree for the BDT model.

Because both the Ho-Lee and BDT models have been calibrated using the same vector
of spot rates, there is no difference in the prices of the plain vanilla bonds. However,
the prices of the callable and putable bonds differ because the option premium is path
dependent.
Let’s first value the callable bond using the BDT tree. We once again consider the
bond paying an annual coupon of 10%.
398 | Chapter 13: Bonds with Embedded Options

The Callable Bond and the BDT Model

At time t4 , the cash flow is 1,000 + 50 = $1,050. The value at t3 node-1 is

1,050
0.090245
+ 50 = 1,004.6669 + 50 = $1,054.6669
(1 + 2
)

If the bond is called, the issuer has to pay 50 + 50 + 1,000 = $1,100. Thus the bond will
not be recalled.
The value at node-2 is
1,050
0.084990
+ 50 = 1,007.1990 + 50 = $1,057.1990
(1 + 2
)

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
The value at node-3 is
1,050
0.080040
+ 50 = 1,009.5959 + 50 = $1,059.5959
(1 + 2
)

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
The value at node-4 is
1,050
0.075379
+ 50 = 1,011.8633 + 50 = $1,061.8633
(1 + 2
)

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
Now let’s move to time t2 . The value at node-1 is

1,054.6669 1,057.1990
0.5 × 0.051969
+ 0.5 × 0.051969
+ 50 = 513.9780 + 515.2120 + 50
(1 + 2
) (1 + 2
)
= $1,079.1900

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
The value at node-2 is
1,057.1990 1,059.5959
0.5 × 0.047497
+ 0.5 × 0.047497
+ 50 = 516.3373 + 517.5079 + 50
(1 + 2
) (1 + 2
)
= $1,083.8452

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
The value at node-3 is
1,059.5959 1,061.8633
0.5 × 0.043409
+ 0.5 × 0.043409
+ 50 = 518.5432 + 519.6528 + 50
(1 + 2
) (1 + 2
)
= $1,088.1960
Pricing the Callable and Putable Bonds Using the BDT Model | 399

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
Now let’s move to time t1 . The value at node-1 is

1,079.1900 1,083.8452
0.5 × 0.068915
+ 0.5 × 0.068915
+ 50 = 521.6212 + 523.8713 + 50
(1 + 2
) (1 + 2
)
= $1,095.4926

If the bond is called, the issuer has to pay $1,100. Thus the bond will not be recalled.
The value at node-2 is
1,083.8452 1,088.1960
0.5 × 0.061122
+ 0.5 × 0.061122
+ 50 = 525.8520 + 527.9629 + 50
(1 + 2
) (1 + 2
)
= $1,103.8149

If the bond is called, the issuer has to pay $1,100. Thus the bond will be recalled.
The value at time t0 is:

1,095.4926 1,100.0000
0.5 × 0.075
+ 0.5 × 0.075
= 527.9484 + 530.1205 = $1,058.0689
(1 + 2
) (1 + 2
)

Because the value of the plain vanilla bond is $1,059.9075, the value of the call option
inherent in the callable bond is

1,059.9075 − 1,058.0689 = $1.8386

Valuation of the Putable Bond

Consider a four-period bond with a face value of $1,000 that pays a coupon of $40
every period, and let’s use the BDT tree to price it. The price of a plain vanilla bond,
as shown earlier, is $1,023.0589.
At time t4 , the cash flow is 1,000 + 40 = $1,040. The value at t3 node-1 is

1,040
0.090245
+ 40 = 995.0987 + 40 = $1,035.0987
(1 + 2
)

If the bond is put, the issuer has to pay $1,040. Thus the bond will be put back by the
investor.
The value at node-2 is
1,040
0.084990
+ 40 = 997.6067 + 40 = $1,037.6067
(1 + 2
)

If the bond is put, the issuer has to pay $1,040. Thus the bond will be put back by the
investor.
400 | Chapter 13: Bonds with Embedded Options

The value at node-3 is

1,040
0.080040
+ 40 = 999.9808 + 40 = $1,039.9808
(1 + 2
)

If the bond is put, the issuer has to pay $1,040. Thus the bond will be put back by the
investor.
The value at node-4 is
1,040
0.075379
+ 40 = 1,002.2265 + 40 = $1,042.2265
(1 + 2
)

If the bond is put, the issuer has to pay $1,040. Thus the bond will not be put back by
the investor.
Now let’s move to time t2 . The value at node-1 is
1,040.0000 1,040.0000
0.5 × 0.051969
+ 0.5 × 0.051969
+ 40 = 506.8303 + 506.8303 + 40
(1 + 2
) (1 + 2
)
= $1,053.6606

If the bond is put, the issuer has to pay $1,040. Thus the bond will not be put back.
The value at node-2 is
1,040.0000 1,040.0000
0.5 × 0.047497
+ 0.5 × 0.047497
+ 40 = 507.9373 + 507.9373 + 40
(1 + 2
) (1 + 2
)
= $1,055.8746

If the bond is put, the issuer has to pay $1,040. Thus the bond will not be put back.
The value at node-3 is
1,040.0000 1,042.2265
0.5 × 0.043409
+ 0.5 × 0.043409
+ 40 = 508.9534 + 510.0430 + 40
(1 + 2
) (1 + 2
)
= $1,058.9964

If the bond is put, the issuer has to pay $1,040. Thus the bond will not be put back.
Now let’s move to time t1 . The value at node-1 is
1,053.6606 1,055.8746
0.5 × 0.068915
+ 0.5 × 0.068915
+ 40 = 509.2817 + 510.3519 + 40
(1 + 2
) (1 + 2
)
= $1,059.6336

If the bond is put, the issuer has to pay $1,040. Thus the bond will not be put back.
The value at node-2 is
1,055.8746 1,058.9964
0.5 × 0.061122
+ 0.5 × 0.061122
+ 40 = 512.2815 + 513.7961 + 40
(1 + 2
) (1 + 2
)
= $1,066.0776
Yield Spreads for Callable Bonds | 401

If the bond is put, the issuer has to pay $1,040. Thus the bond will not be put back.
The value at time t0 is

1,059.6336 1,066.0776
0.5 × 0.075
+ 0.5 × 0.075
= 510.6668 + 513.7723 = $1,024.4391
(1 + 2
) (1 + 2
)

The price of the plain vanilla bond is $1,023.0589. Thus the value of the put option is

P = 1,024.4391 − 1,023.0589 = $1.3802

Yield Spreads for Callable Bonds

There are different ways in which the yield of a callable bond can be compared with
that of an equivalent plain vanilla bond. We study three such approaches in this sec-
tion.

The Traditional Yield Spread

In this approach we compare the YTM of a plain vanilla bond, with that of the callable
bond. Let’s use the prices obtained using the Ho-Lee model. The price of the plain
vanilla bond is $1,059.9075, and that of the callable is $1,053.0610. The periodic cash
flow is $50, and the face value is $1,000. There are four periods until maturity. We use
the RATE function in Excel to compute the YTMs.

= RATE(4, −50, 1059.9075, −1000) = 6.7478%

= RATE(4, −50, 1053.0610, −1000) = 7.1071%

Thus the yield spread is 7.1071 − 6.7478 = 0.3593%.

Similarly, we can find the yield spread for the putable bond, which has a coupon
of $40 per period. The YTM of the equivalent plain vanilla bond is

= RATE(4, −40, 1023.0589, −1000) = 6.7482%

The YTM of the putable is given by

= RATE(4, −40, 1026.8970, −1000) = 6.5434%

Thus the yield spread is 6.5434 − 6.7482 = −0.2048%.

402 | Chapter 13: Bonds with Embedded Options

The Static Spread

We saw earlier that the YTM of a plain vanilla bond is a complex average of the spot
rates. Consequently, it is vulnerable to the coupon effect. In other words, two bonds
with the same maturity, but different coupons, will have different YTMs, although the
cash flows in both cases have been discounted using the same vector of spot rates.
The static spread on the other hand is based on the spot rates. It is obtained by
adding a constant spread s to each element of the spot-rate vector. For a plain vanilla
bond, the static spread is zero. Now let’s compute the static spread for the callable
bond with a periodic coupon of $50.

50 50 50 1,050
+ + + = 1,053.0610
(1 + 0.0750+s
) 0.0700+s 2 0.0625+s 3 0.0675+s 4
2 (1 + 2
) (1 + 2
) (1 + 2
)

Using Solver in Excel, s = 0.3592%, which is exactly what we got for the yield spread.
Now let’s consider the four-period putable bond with a price of $1,026.8970.

40 40 40 1,040
+ + + = 1,026.8970
(1 + 0.0750+s
) 0.0700+s 2 0.0625+s 3 0.0675+s 4
2 (1 + 2
) (1 + 2
) (1 + 2
)

Using Solver in Excel, s = −0.2048%, which is again equal to the yield spread.
As we can see, for both the callable and the putable bond, the yield spread is the
same as the static spread. This is not surprising because the respective YTMs for the
two bonds are obtained from the same vector of spot rates.

The Option-Adjusted Spread

In this approach, we add a constant spread to each interest rate along the various
paths. The spread is adjusted using a trial-and-error process, until the price obtained
is equal to the market price of the bond. Let’s illustrate this approach for the plain
vanilla and the callable bond.

The Plain Vanilla Bond

Consider the bond with four periods to maturity, a face value of $1,000, and a coupon
of 10% per annum. Let’s use the Ho-Lee tree to illustrate our arguments and denote
the unknown spread by s.
At time t4 , the cash flow is 1,000 + 50 = $1,050. The value at t3 node-1 is

1,050
0.112863+s
+ 50
(1 + 2
)
Yield Spreads for Callable Bonds | 403

The value at node-2 is

1,050
0.092863+s
+ 50
(1 + 2
)

The value at node-3 is

1,050
0.072863+s
+ 50
(1 + 2
)

The value at node-4 is

1,050
0.052863+s
+ 50
(1 + 2
)

Now let’s move to time t2 . The value at node-1 is

1,050 1,050
+ 50 + 50
(1+ 0.112863+s ) (1+ 0.092863+s )
0.5 × [ 2
0.067725+s
] + 0.5 × [ 2
0.067725+s
] + 50
(1 + 2
) (1 + 2
)
[ ] [ ]
The value at node-2 is
1,050 1,050
+ 50 + 50
(1+ 0.092863+s ) (1+ 0.072863+s )
0.5 × [ 2
0.047725+s
] + 0.5 × [ 2
0.047725+s
] + 50
(1 + 2
) (1 + 2
)
[ ] [ ]
The value at node-3 is
1,050 1,050
+ 50 + 50
(1+ 0.072863+s ) (1+ 0.052863+s )
0.5 × [ 2
0.027725+s
] + 0.5 × [ 2
0.027725+s
] + 50
(1 + 2
) (1 + 2
)
[ ] [ ]
Now let’s move to time t1 . The value at node-1 is
1,050 1,050
+50 +50
(1+ 0.112863+s (1+ 0.092863+s
0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.067725+s
2
) (1+ 0.067725+s
2
)
0.075063+s
(1 + 2
)
1,050 1,050
+50 +50
(1+ 0.092863+s (1+ 0.072863+s
0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.047725+s ) (1+ 0.047725+s )

+ 50
2 2
+ 0.075063+s
(1 + 2
)

The value at node-2 is

1,050 1,050
+50 +50
(1+ 0.092863+s (1+ 0.072863+s
0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.047725+s
2
) (1+ 0.047725+s
2
)
0.055063+s
(1 + 2
)
404 | Chapter 13: Bonds with Embedded Options

1,050 1,050
+50 +50
(1+ 0.072863+s (1+ 0.052863+s
0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.027725+s ) (1+ 0.027725+s )

+ 50
2 2
+ 0.055063+s
(1 + 2
)

The value at time t0 is

1,050 1,050
+50 +50
(1+ 0.112863+s (1+ 0.092863+s
0.5 × 0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.067725+s
2
) (1+ 0.067725+s
2
)
0.075063+s 0.075+s
(1 + 2
) × (1 + 2
)
1,050 1,050
+50 +50
(1+ 0.092863+s (1+ 0.072863+s
0.5 × 0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.047725+s
2
) (1+ 0.047725+s
2
)
+ 0.075063+s 0.075+s
(1 + 2
) × (1 + 2
)
1,050 1,050
+50 +50
(1+ 0.092863+s (1+ 0.072863+s
0.5 × 0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.047725+s
2
) (1+ 0.047725+s
2
)
+ 0.055063+s 0.075+s
(1 + 2
) × (1 + 2
)
1,050 1,050
+50 +50
(1+ 0.072863+s (1+ 0.052863+s
0.5 × 0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.027725+s
2
) (1+ 0.027725+s
2
)
+ 0.055063+s 0.075+s
(1 + 2
) × (1 + 2
)
0.5 × 50 0.5 × 50
+ 0.075+s
+ 0.075+s
= $1,059.9075
(1 + 2
) (1 + 2
)

Using Solver in Excel, we get a value of zero for s, the option-adjusted spread (OAS).
Now let’s turn to the callable bond.
1,050 1,050
+50 +50
(1+ 0.112863+s (1+ 0.092863+s
0.5 × 0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.067725+s
2
) (1+ 0.067725+s
2
)
0.075063+s 0.075+s
(1 + 2
) × (1 + 2
)
1,050 1,050
+50 +50
(1+ 0.092863+s (1+ 0.072863+s
0.5 × 0.5 × [0.5 × { 2
} + 0.5 × { 2
} + 50]
) )

(1+ 0.047725+s
2
) (1+ 0.047725+s
2
)
+ 0.075063+s 0.075+s
(1 + 2
) × (1 + 2
)
1,100 50
+ 0.5 × 0.075+s
+ 0.5 × 0.075+s
= $1,053.0610
(1 + 2
) (1 + 2
)

Using Solver in Excel, we get a value of zero for s, the OAS.

Analysis of the Option-Adjusted Spread (OAS)

For both the plain vanilla bond and the callable bond, we find that the OAS is zero if the
bonds are fairly priced. The OAS is used to measure the extent to which a bond is mis-
priced. A positive OAS indicates that the security is under-priced whereas a negative
OAS indicates that it is overvalued. However, the issue in practice is that the spread
Convertible Bonds | 405

may be non-zero due to mispricing, as well as due to misspecification of the valuation

model. For instance, assume that the callable bond has been valued using the BDT
model and the price has been determined to be $1,058.0689. If an analyst computes
the OAS using this price and the BDT model, the result is a value of zero for it. However,
what if the analyst chooses to use the Ho-Lee model? The result is a value of −0.5221%.
In this case, the conclusion of a non-zero spread is because of model misspecification
and not due to the mispricing of the bond.

Convertible Bonds
Convertible bonds, or convertibles, can be converted to equity shares at a prespeci-
fied conversion rate. Exchangeable bonds are structurally identical, but the identity
of the bond issuer and that of the company whose shares are offered on conversion is
not the same. Convertibles offer a lower coupon than plain vanilla debt with similar in-
vestment characteristics. The reduced coupon reflects the value of the inherent option
to convert to shares of equity. A convertible bond may contain an event-risk clause. If
such a covenant is present, the bond holders can demand immediate redemption in
the event the issuing entity is taken over by another entity or merged with it. Such
bonds also contain call and put options in practice, which can be used to enforce con-
version.
The right to convert is an option given to the holder. It may be exercisable only
on a particular date, which is akin to a European option; on one of several exercise
dates, which is similar to a Bermudan option; or at any time in a specified interval,
which is equivalent to an American option. The number of shares into which the bond
can be converted is termed as the conversion ratio (CR). The ratio may be constant for
the life of the bond, or else may vary over time. For instance, if the issuing company
feels that its stock price will increase over time, it may issue a bond with a declining
conversion ratio. In this case, the later the conversion option is exercised, the greater
is the price paid per share. For instance, in the case of a bond with a face value of
$1,000, the issuer may state that the bond can be converted into 40 shares if the option
is exercised within three years, into 32 shares if exercise takes place between three and
five years, and 25 shares if the exercise occurs after five years. The price per share at
which the bond can be converted into equity is termed the conversion price (CP). In
most cases the conversion price is higher than the prevailing stock price at the time of
issue. Subsequently, the market price must at least reach this threshold before holders
will contemplate conversion. The relationship between the conversion ratio and the
conversion price may be stated as:
Redemption Value
Conversion Price =
Conversion Ratio
Issuers first set the required conversion price and then determine the corresponding
conversion ratio. For instance, in the case of a bond with a face value of $1,000, the
406 | Chapter 13: Bonds with Embedded Options

issuer may fix the conversion price at $25. This means that the conversion ratio is
1,000
25
= 40.
The current market value of the shares if the bond is converted is known as the
conversion value or the parity value. For instance, if the conversion ratio is 40, and
the prevailing share price is $21.50, the parity value is 40 × 21.50 = $860. The market
price of a plain vanilla bond that is equivalent in all respects except for the conversion
feature is termed the straight value of the bond. At any point in time, the price of the
convertible bond must be greater than or equal to the higher of the parity value and
the straight value to preclude arbitrage. Here is an illustration.

Example 13.5. Consider a convertible bond with a face value of $1,000 and 10 years to maturity. The
coupon is 8% per annum, payable semiannually. The bond is convertible into 40 shares of equity,
which are currently priced at $21.50 each. The YTM of a comparable plain vanilla bond is 10% per an-
num. The parity value is $860. The straight value using Excel is = PV(0.05, 20, −40, −1000) = $875.38.
Because the straight value is higher than the parity value, the bond must trade for at least $875.38 to
preclude arbitrage. Thus if the price of the convertible is $890, there is no arbitrage opportunity. Let’s
examine the consequences if this condition is violated.
Assume the market price of the convertible bond is below the parity value of $860, for instance,
$845. An arbitrageur would buy 1,000 bonds by paying $845,000 and immediately convert them into
40,000 shares, which can be sold at a price of $21.50 each, to yield $860,000. The difference of
$15,000 is clearly an arbitrage profit.
Now consider the situation where the price of the convertible is between the parity value and the
straight value. Let’s assume it is $870. An arbitrageur would buy 1,000 convertible bonds and short
1,000 plain vanilla bonds. The initial cash flow is 875,380 − 870,000 = $5,380. Every six months the
convertible pays a coupon of $40 per bond, which can be used to pay the coupon due on the bond that
has been shorted. At the end, the face value that is received on redemption of the convertible bonds
can be used to fulfill the payment obligation on account of the plain vanilla bonds that were shorted
at the outset. Consequently the initial cash flow of $5,380 does indeed represent an arbitrage profit.

Bonds are usually issued at their face value. At the time of issue, the difference be-
tween the face value of the convertible bond and the parity value is termed the con-
version premium (CP). The premium is usually expressed as a percentage of the parity
value. If the conversion value is $860, the premium is 1,000 − 860 = $140. This is
usually expressed as

140
CP = × 100 ≡ 16.2790%
860

The premium may also be stated as the difference between the conversion price and
the current market price per share, divided by the current share price. In this case the
conversion price is 1,000
40
= 25 and the share price is 21.50. Thus the premium is $3.50.
The premium in percentage terms is:

3.5
× 100 = 16.2790%
21.50
Convertible Bonds | 407

Subsequently the premium is computed as the difference between the prevailing price
of the convertible and the conversion value. This, too, is expressed as a percentage of
the conversion value. For instance, assume that weeks after the issue, the convertible
is trading at $1,040. The premium is

180
1,040 − 860 = $180 ≡ × 100 = 20.93%
860

Changes in the Conversion Ratio

The conversion ratio is changed if the underlying stock undergoes a split or a reverse
split, or the issuer declares a stock dividend. A stock dividend is a dividend that is
paid to the shareholders in the form of shares. That is, the company offers additional
shares to the shareholders without requiring them to pay any money. The issue of such
shares without any monetary consideration entails the transfer of funds from the re-
serves and surplus account on the balance sheet to the share capital account. Such
a fund transfer is known as the capitalization of reserves.2 From a theoretical stand-
point, stock dividends do not create any value for the shareholders. The issue of addi-
tional shares, per se, does not lead to an increase in the asset base of the issuing entity,
nor does it have any implications for the earnings capacity of the firm. Consequently,
following a stock dividend the share price should theoretically decline. For instance,
assume that a firm has issued 250,000 shares and the share price prior to the dividend
announcement is $42 per share. If the firm announces a 20% stock dividend, there is
an issue of an additional 50,000 shares, and there will be 300,000 shares outstanding
after the dividends are paid. Considering that there is no change in the value of the
firm, the ex-dividend price, P, should be such that

42 × 250,000 = P × 300,000
⇒ P = $35

Unlike stock dividends, which entail the capitalization of reserves, stock splits
lead to a decrease in the par value of the shares, accompanied by a simultaneous in-
crease in the number of shares outstanding. An n:1 stock split means that n new shares
are issued to the existing shareholders in lieu of one existing share. Take the case of
an 8:5 split. What this means is that the holder of five shares will have eight shares af-
ter the split. This is exactly analogous to a 60% stock dividend. Thus splits and stock
dividends are mathematically equivalent. However, they are operationally different.
Take the case of a company that has issued 250,000 shares with a par value of $100

2 In some markets, a stock dividend is termed as a bonus share, where the word “bonus” connotes
that shareholders are being offered additional shares without having to make a monetary investment.
408 | Chapter 13: Bonds with Embedded Options

each. If it announces an 8:5 split, the number of shares issued increases to 400,000.
However the par value declines to $62.50. The issued capital remains at $25,000,000.
Because a split is equivalent to a stock dividend, the share price after a split behaves in
the same way as it would after an equivalent stock dividend. Assume that an investor
is holding 10,000 shares worth $42 each. The market value of the shares is $420,000.
If the firm announces an 8:5 split, then the investor will have 16,000 shares after the
split. Because the split is value neutral in theory, the post-split share price ought to be

420,000
= $26.25
16,000

A reverse split or a consolidation is similar to a split, with the difference being the
following. In the case of an n : m reverse split, n will be less than m. Take the case of
an investor who is holding 10,000 shares with a market price of $42 per share. If the
firm were to announce a 5:8 reverse split, the post-reverse split price ought to be

420,000
= $67.20
6,250

Now let’s illustrate how the conversion price and the conversion ratio are adjusted for
corporate actions such as splits/reverse splits and stock dividends.

Example 13.6. Consider a bond with a face value of $1,000 and a conversion ratio of 40. The conver-
sion price is obviously $25. In the case of an n : m split, the conversion ratio is multiplied by n/m, which
implies that the conversion price is multiplied by m/n. For instance, if there is an 8:5 split, the conver-
sion ratio becomes (40 × 8) ÷ 5 = 64, and the corresponding conversion price is (25 × 5) ÷ 8 = $15.625.
On the other hand, if there is a 5:8 reverse split, the conversion ratio becomes (40 × 5) ÷ 8 = 25, and
the corresponding conversion price is (25 × 8) ÷ 5 = $40. In the case of an n : m stock dividend, the
adjustment factor is (n + m) ÷ m. For instance, if there is a 3:2 stock dividend, the adjustment fac-
tor is 5/2 = 2.5. The conversion ratio is multiplied by the factor, and the conversion price is divided
by the factor. In our case the new ratio is 40 × 2.5 = 100, and the corresponding conversion price is
25 ÷ 2.5 = $10.

Convertible bonds often come with built-in call options. If the issuer invokes the op-
tion, it gives the holders some time to decide if they wish to convert. If they do not,
the bonds are redeemed. The call option may be provisional or absolute. In the case
of the former, there is a threshold price level for the share price. If this trigger is hit,
then the call option can be invoked. However, if the call option is absolute, then there
is no attached precondition.
These bonds also may have built-in put options. One such option is a rolling pre-
mium put option. Bonds with such an option, provide a bouquet of put options that
enable the holder to redeem at a premium over the par value at dates spread out over
time. The redemption price increases at each successive option exercise date. The pres-
ence of a call option in a convertible bond warrants a higher coupon. On the other
hand, a rolling premium put, if incorporated, results in a bond with a lower coupon
Convertible Bonds | 409

than a plain convertible. The rationale for an increasing premium is as follows. If the
share price is not increasing at an attractive rate, the holder of a put may choose to re-
deem. However, the increasing premium in a rolling put acts as an incentive to stay in-
vested. A consequence of an increasing premium is that, as each exercise date elapses,
the share price has to rise even more to make conversion an attractive proposition. As-
sume for instance that the bond is issued at par with a conversion ratio of 40. Thus the
original conversion price is $25. Assume that after N years the price at which it can be
put is $1,200. The new conversion price is $30. If after another few years the price at
which it can be put is $1,325, the corresponding conversion price becomes even higher
at $33.1250.

Pros and Cons of a Convertible Issue: The Issuer’s Perspective

Take the case of a company that is seeking to raise capital. If it is already listed, one
alternative is a follow-on public offering (FPO). However, the issuer may or may not
be able to command an attractive premium to the prevailing share price. Another al-
ternative is a rights issue to existing shareholders. But such an issue has to be made
at a discount to the prevailing market price. The issue of a convertible bond offers
a third alternative. The conversion option enables the issue of an instrument with a
lower coupon. And if the conversion price is set above the prevailing share price, the
issuer can expect an attractive price per issued share if the conversion happens sub-
sequently. The third option allows the issuer to take advantage of the tax deductible
feature of interest on debt. Also, if the issuer resorts to a follow-on public offering or
a rights issue, there is a dividend implication almost immediately, for the new share-
holders would expect compensation by way of dividends. In contrast, if convertible
bonds are issued, there would be a dividend implication only if the bonds are con-
verted at a later date, by which time the issuer may be in a more comfortable financial
position.
The negative feature of a convertible is that typically such bonds come with a put
feature. If the share price does not appreciate adequately, holders may choose to ex-
ercise the put option rather than convert. And as can be appreciated, put options are
exercised in a rising interest rate environment. As is the case with all putable bonds,
the exercise of the inherent put option means that the issuer has to reissue fresh debt
at a higher coupon rate.

Concept of Break-Even

The current yield of a convertible bond is usually higher than the dividend yield for
investors who are holding the equity shares of the bond issuer. The break-even period
of the bond is defined as the number of years it would take for the bond holder to
410 | Chapter 13: Bonds with Embedded Options

recover the conversion premium out of the differential between the current yield and
the dividend yield. Here is an example.

Example 13.7. Consider a convertible bond with a face value of $1,000 and a coupon of 8.16% per
annum payable semiannually. The market price of the bond is $1,088. The current yield is

81.60
= 7.50%
1,088

The conversion ratio is 40. The current share price is $25. The dividend per share is $1.25. Thus the
dividend yield is

1.25
= 5.00%
25

The conversion value of the bond is 40 × 25 = $1,000. Thus the conversion premium is $88 or 8.8% of
the conversion value. The break even is therefore:

0.088
= 3.52 years
(0.075 − 0.0500)

Liquid Yield Option Notes (LYONs)

These are essentially zero coupon bonds with a conversion feature. Consequently they
can be redeemed at face value at maturity, assuming of course that they have not been
converted earlier. The terms of conversion are usually set in such a way that there must
be a significant appreciation in the share price to make conversion attractive. Thus
in practice, redemption at maturity is more probable than conversion prior to it, and
hence such bonds offer high yields to induce investors to invest. We can illustrate the
mechanics of such a bond with the help of Example 13.8.

Example 13.8. Consider a zero coupon bond with a face value of $1,000 and eight years to maturity.
Assume the YTM is 6% per annum. The price at the outset is given by

1,000
= $623.1669
(1.03)16

The bond gives the holder the right to put after five years, at a price corresponding to an annual return
of 6%. The corresponding price is therefore

623.1669 × (1.03)10 = $837.4842

Assume the current share price is $25 and the conversion price is $40. Thus the conversion ratio is

623.1669
= 15.5792
40
Exchangeable Bonds | 411

At the end of five years the choices available to the holder are the following: surrender it for $837.4842
or convert it into 15.5792 shares. For the conversion option to be attractive, the share price at the end
of five years must be greater than

837.4842
= $53.7566
15.5792

Because the current share price is $25, the required compounded average growth rate is

53.7566 1/10
[( ) − 1] × 2 = 15.9132%
25

Exchangeable Bonds
Exchangeable bonds are convertible bonds with a difference. In the case of a convert-
ible bond, if the bonds have been issued by a company ABC, then the bond holders
get shares of ABC if they choose to convert. However, if the bonds are exchangeable
in nature, then the holders get the shares of another company XYZ. One possibility is
that XYZ is a subsidiary of ABC. The other possibility is that ABC has taken a stake in
XYZ at some stage and, instead of offloading the shares in the stock market at a future
point in time, has chosen to link them with a bond issue in the form of an exchange-
able bond. The provision of the conversion option allows the bonds to be issued with
a lower coupon. Also, if the issuer has a significant stake in the other company, a stake
sale at a subsequent date may lead to a depression in the share price, unless the mar-
ket is buoyant enough to absorb the sale. The reduction in the stake by inducing the
bond holders to convert can achieve the same objective without the specter of a signif-
icant share price decline. It must be noted that if a convertible bond is converted to a
share of stock, there is a dilution of holding for the shareholders of the company that
issued the bonds. However, if an exchangeable bond is converted, there is no dilution
for shareholders of the firm that has chosen to issue these bonds.

Valuing a Convertible Bond with Built-in Call and Put Options

Consider a three-period convertible bond with a face value of $1,000 and a coupon of
7% per annum. The bond can be called back by the issuer at the end of one period,
or after two periods by paying six month’s coupon as a call premium. We assume that
each period is equivalent to six months. The bond also has a put option with a rolling
premium. If put back after one period, the issuer pays $20 as premium, whereas if it is
put after two periods, it pays $40 as premium. The conversion ratio is 25. The current
stock price is $40, and every period the stock may increase by 20% or decline by 20%.
The riskless rate is 5% per period.
412 | Chapter 13: Bonds with Embedded Options

From our discussion on the binomial model of option pricing in Chapter 10, we
know that the probability of an up move is3

(r − d) (1.05 − 0.80)
= = 0.625
(u − d) (1.20 − 0.80)

Thus the probability of a down move is 0.375.

The stock price tree is shown in Figure 13.4.

Figure 13.4: Stock price tree for the three-period case.

Let’s examine the bond node-wise, starting with the point of maturity.
Consider node-1 at time T.
If the bond is converted, the shares are worth 25 × 69.12 = $1,728. The cash flow
including the coupon is $1,763. The face value plus the coupon is $1,035. Thus the bond
will be converted, and the value will be $1,763.
Now consider node-2 at time T.
If the bond is converted, the shares are worth 25 × 46.08 = $1,152. The cash flow
including the coupon is $1,187. The face value plus the coupon is $1,035. Thus the bond
will be converted, and the value will be $1,187.

3 u represents the magnitude of an up move, and d is the magnitude of a down move. r is one plus the
periodic riskless rate.
Valuing a Convertible Bond with Built-in Call and Put Options | 413

Now consider node-3 at time T.

If the bond is converted, the shares are worth 25 × 30.72 = $768. The cash flow
including the coupon is $803. The face value plus the coupon is $1,035. Thus the value
of the bond is $1,035.
Finally consider node-4 at time T.
If the bond is converted, the shares are worth 25 × 20.48 = $512. The cash flow
including the coupon is $547. The face value plus the coupon is $1,035. Thus the value
of the bond is $1,035.
Now consider node-1 at time T − 1.
The model value is

1,763 1,187
0.625 × + 0.375 × = $1,473.33
1.05 1.05

If we add the coupon, the value is $1,508.33. If the bond is converted, the shares are
worth 25 × 57.60 = $1,440. The cash flow including the coupon is $1,475. The amount
payable if called is $1,070. The amount payable if put is $1,075. Because the model
value is more than the amount payable if called, the issuer will call. In response, the
investors will convert since they get more by doing so. The put option has no value in
this scenario. Thus the value of the bond is $1,475.
Now consider node-2 at time T − 1.
The model value is
1,187 1,035
0.625 × + 0.375 × = $1,076.1904
1.05 1.05

If we add the coupon, the value is $1,111.1904. If the bond is converted, the shares are
worth 25 × 38.40 = $960. The cash flow including the coupon is $995. The amount
payable if called is $1,070. The amount payable if put is $1,075. Because the model
value is more than the call price, the issuer will exercise the call option. However, as
the amount payable if called is more than what can be obtained by conversion, the
bonds will not be converted; therefore, if the call option is invoked, the issuer has
to pay $1,070 to the bond holder. However, from the bond holder’s perspective, the
put option if exercised leads to a higher cash inflow than both the conversion and the
call values. Thus the put option is valuable in this scenario, for the alternative is to
surrender it for $1,070, and the bond will therefore be put back. Thus the value of the
bond is $1,075.
Now consider node-3 at time T − 1.
The model value is
1,035 1,035
0.625 × + 0.375 × = $985.7143
1.05 1.05

If we add the coupon, the value is $1,020.7143. If the bond is converted, the shares are
worth 25 × 25.60 = $640. The cash flow including the coupon is $675. The amount
414 | Chapter 13: Bonds with Embedded Options

payable if called is $1,070. The amount payable if put is $1,075. Because the model
value is less than the amount payable if called, the issuer will not exercise the call
option. And, as the model value is greater than the conversion value, the bond holders
will not convert. However, the put option will be exercised in this scenario, and thus
the bond will be put back because the amount receivable is $1,075 which is greater
than the model value. Thus the value of the bond is $1,075.
Now consider node-1 at time T − 2. The model value is
1,475 1,075
0.625 × + 0.375 × = $1,261.9047
1.05 1.05
If we add the coupon, the value is $1,296.9047. If the bond is converted, the shares
are worth 25 × 48 = $1,200. The cash flow including the coupon is $1,235. Because
the model value is more than the amount payable if called, which is $1,070, the issuer
will exercise the call option. From the investor’s perspective, the value of the shares
is higher than the return from surrendering the bond. Thus the investors will convert.
The put option has no value in this scenario, for the investors stand to receive only
$1,055. Thus the value of the bond is $1,235.
Let’s now consider node-2 at time T − 2. The model value is
1,075 1,075
0.625 × + 0.375 × = $1,023.8095
1.05 1.05
If we add the coupon, the value is $1,058.8095. If the bond is converted, the shares
are worth 25 × 32 = $800. The cash flow including the coupon is $835. The amount
payable if called is $1,070, which is more than the model value. Thus, the issuer will
not exercise the call option. The put has no value in this case because it results in an
inflow of only $1,055 if exercised, which is less than the model value of $1,058.8095.
Thus the value of the bond is $1,058.8095.
Finally let’s move to time T − 3. The model value is

1,235 1,058.8095
0.625 × + 0.375 × = $1,113.2652
1.05 1.05
Thus this convertible bond, with built-in call and put options, has a price of $1,113.2652
at the outset based on the assumptions made.

Chapter Summary
In this chapter, we studied bonds with built-in options, such as callable bonds,
putable bonds, and convertible bonds. We examined the yield to call and its re-
lationship with the yield to maturity, for callable bonds. We used the short rates
obtained by the Ho-Lee model, as well as the BDT model, to value these bonds. In
this context we studied the yield differences between plain vanilla bonds and bonds
with call or put options, using the concepts of the yield spread, the static spread,
Chapter Summary | 415

and the option-adjusted spread (OAS). Although the static spread is equal to the yield
spread, the option-adjusted spread of a bond is zero if the bond is fairly priced. How-
ever, the presence of a non-zero spread need not imply that the bond is mispriced.
For model misspecification could lead to a similar conclusion. The chapter concludes
with a detailed study of convertible bonds. We examined the valuation of a convertible
bond using the binomial model for the evolution of the underlying stock price and
incorporated call and put options to make the analysis more realistic.
Chapter 14
Interest Rate Swaps and Credit Default Swaps
Swaps are a popular form of OTC derivative instruments. What is a swap? As the name
connotes, it is a contract to exchange or swap two cash flows. It is an exchange between
two parties of two payment streams that are different from each other. In the case of an
interest rate swap (IRS), the contract requires the specification of a principal amount
termed the notional principal, for it is not meant to be exchanged, but has been speci-
fied purely to facilitate the computation of interest. Each of the two counterparties use
a different benchmark for computing the interest on the specified notional principal.
For instance, one party may use a fixed rate of interest, whereas another may use a
variable rate such as the three-month LIBOR. This is referred to as a coupon swap. The
alternative is a contract, where both the parties use variable rates to compute their
respective obligations. This is referred to as a basis swap. After the obligations have
been determined, the party owing the higher amount pays the difference between the
two computed amounts to the counterparty. This is termed as netting. Netting is feasi-
ble because both cash flow streams are denominated in the same currency. Had they
been in two different currencies, this kind of netting would not be feasible. In this
chapter, we do not consider currency swaps, which entail the exchange of cash flows
in different currencies.
Although coupon and basis swaps are possible, a contract which requires both the
parties to pay a fixed rate is not feasible. This is because the party that is required to
pay the higher rate would be paying constantly, and the counterparty would be receiv-
ing constantly. This is clearly a manifestation of arbitrage, and consequently the party
that is required to pay a higher rate will never accept such an arrangement. In an in-
terest rate swap, whether it is a coupon swap or a basis swap, we do not know a priori,
which of the two counterparties will have to make a payment, and correspondingly
the identity of the receiver is unknown at the outset.

Interest Rate Swaps

We will now define briefly the various terms that need to be incorporated into a swap
contract.

Contract Terms

The following terms must be explicitly stated while designing an interest rate swap
contract:
– The identities of the two counterparties.

https://doi.org/10.1515/9781547400669-014
Interest Rate Swaps | 417

– The maturity of the swap. This is the date on which the last exchange of cash flows
takes place between the two parties.
– The interest rate used by the first party to calculate its payments. It may be fixed
or floating.
– The interest rate used by the second party to calculate its payments. It must be
floating if the first party is making payments based on a fixed rate of interest.
– The day-count convention for the computation of interest.
– The frequency of payment
– The notional principal

Example 14.1. Consider a four-year fixed-floating swap between Scotia Bank and Dominion Bank. Sco-
tia will make payments based on an annual rate of 5%. Dominion will compute its liability based on
the six-month LIBOR. We assume that every month consists of 30 days and that the year as a whole
consists of 360 days. The implication is that every semiannual period consists of 180 days, and conse-
quently the interest rate per semiannual period is one half of the annual rate. Other day-count conven-
tions such as Actual/Actual or Actual/360 are possible. In such cases, the amount payable by the fixed
rate payer varies from period to period, even though the interest rate remains constant, because the
number of days per semiannual period varies. The exchange of payments takes place on a semiannual
basis. The notional principal amount is $8 million.
Assume that the six-month LIBOR, as observed at semiannual intervals over the next four years,
is as shown in Table 14.1.

Table 14.1: Observed values of LIBOR.

Time LIBOR

0 5.25%
After 6 Months 5.40%
After 12 Months 5.10%
After 18 Months 4.75%
After 24 Months 4.30%
After 30 Months 4.00%
After 36 Months 4.80%
After 42 Months 5.20%

Using this data, let’s compute the payments to be made by the two parties every six months. Let’s
analyze the cash flows to be exchanged after six months. Scotia has to pay $200,000 every six months.
This may be calculated as follows:

0.5 × 0.05 × 8,000,000 = $200,000

This amount remains constant every period. If, however, we assume an Actual/360 day-count conven-
tion and that the actual number of days in the period is 184 days, the amount payable becomes

184
× 0.05 × 8,000,000 = $204,444.44
360
418 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

In the following period, if the number of days is assumed to be 181, the amount payable is

181
× 0.05 × 8,000,000 = $201,111.11
360

At the end of six months, Dominion has to pay

0.5 × 0.0525 × 8,000,000 = $210,000

Because Scotia owes $200,000 and Dominion owes $210,000, Dominion pays the difference of
$10,000 to Scotia. In the following period, Scotia once again owes $200,000. But the amount owed
by Dominion is

0.5 × 0.0540 × 8,000,000 = $216,000

Consequently Dominion has to pay $16,000 to Scotia.

The cash flows made by the two counterparties and the net cash flow are shown in Table 14.2.

Table 14.2: Cash flows in an interest rate swap.

Time Payment by LIBOR Payment by Net

Scotia Dominion Payment

Zero – 5.25% – –
6 Month $200,000 5.40% $210,000 $(10,000)
12 Month $200,000 5.10% $216,000 $(16,000)
18 Month $200,000 4.75% $204,000 $(4,000)
24 Month $200,000 4.30% $190,000 $10,000
30 Month $200,000 4.00% $172,000 $28,000
36 Month $200,000 4.80% $160,000 $40,000
42 Month $200,000 5.20% $192,000 $8,000
48 Month $200,000 $208,000 $(8,000)
TOTAL $1,600,000 $1,552,000 $48,000

The payments that are made by Dominion are based on the LIBOR that is observed at
the start of the period concerned. However, the interest itself is payable at the end of
the period. For instance, the payment made by Dominion at the end of the first period
is based on a interest rate of 5.25%, which corresponds to the LIBOR at the start of the
period. This system of computation is known as determined in advance and paid in
arrears and is the most commonly used system in practice.
The last column of Table 14.2 shows the net payment. Positive entries connote
that Scotia has to pay the counterparty, whereas negative amounts indicate that the
counterparty has to pay Scotia. The advantage of netting is that it reduces delivery
risk. For instance, if Scotia pays the gross amount it owes to Dominion at the end of
the first six months, it is exposed to the risk that the payment of $210,000 owed by
Dominion may never arrive. On the contrary, because of netting, the risk for Scotia in
Interest Rate Swaps | 419

this case is only the sum of $10,000 that would not be received if Dominion were to
renege.
In a swap contract, certain terms and conditions need to be specified at the very
outset to avoid ambiguities and potential future conflicts. Every swap contract must
clearly spell out the identities of the two counterparties to the deal. In our example,
the two counterparties are Scotia Bank and Dominion Bank.
Second, the tenor of the swap must be clearly stated. The tenor or maturity of
the swap refers to the length of time at the end of which the last exchange of cash
flows between the two parties takes place. In our example, the tenor is four years.
Unlike exchange-traded products like futures contracts, where the exchange specifies
a maximum maturity for contracts on an asset, swaps are OTC products that can have
any maturity that is agreed upon by bilateral discussions.
Third, the interest rates on the basis of which the two parties have to make pay-
ments should be clearly spelled out. To avoid ambiguities, the basis on which the cash
inflow and cash outflow are arrived at for both the counterparties should be explicitly
stated. In our example, Scotia Bank is a fixed rate payer with the rate of interest be-
ing fixed at 5.00% per annum, whereas Dominion Bank is a floating rate payer with
the amount payable being based on the six-month LIBOR prevalent at the start of the
interest computation period.
Fourth, the frequency with which the cash flows are to be exchanged has to be
clearly defined. In our example, we have assumed that cash flow exchanges will take
place at six-monthly intervals. In the market such swaps are referred to as semi-semi
swaps. Other contracts may entail payments on a quarterly basis or on an annual basis.
The benchmark that is chosen for the floating rate payment is usually based on the
frequency of the exchange. For instance a swap entailing the exchange of cash flows
at six-monthly intervals will specify six-month LIBOR as the benchmark, whereas a
swap entailing the exchange of payments at three-monthly intervals will specify the
three-month LIBOR as the benchmark. The most popular benchmark in the market is
the six-month LIBOR.
Fifth, the day-count convention that is used to compute the interest must be
explicitly stated. In our illustration we have assumed that every six-month period
amounts to exactly one-half of a year. The underlying convention is referred to as
30/360. That is every month is assumed to consist of 30 days while the year as a whole
is assumed to consist of 360 days.
Finally, the principal amounts on the basis on which each party has to figure out
the payment to the counterparty have to be clearly stated. In the case of interest rate
swaps, there is obviously only one currency that is involved. However, the magnitude
of the principal has to be specified to facilitate the computation of interest. In our
example, the principal is $8 million.
In a coupon swap, the party that agrees to make payments based on a fixed rate is
referred to as the payer. The counterparty, which is committed to making payments on
a floating rate basis, is referred to as the receiver. Quite obviously these terms cannot
420 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

be used in the case of basis swaps because both the cash flow streams are determined
based on floating rates. Consequently, in order to be explicit and avoid ambiguities, it
is a good practice to describe for each of the two parties, the rates at which they are
scheduled to make and receive payments. Thus in the preceding example, we would
state that Scotia is scheduled to pay a fixed rate of 5% and receive the six-month LIBOR,
whereas Dominion is scheduled to pay the six-month LIBOR and receive a payment
based on an interest rate of 5% in return. In the case of coupon swaps, some markets
refer to the fixed rate payer as the buyer and the fixed rate receiver as the receiver.
The payments in a swap may be settled either on a money-market basis or on a
bond-market basis, the difference being that the first convention is based on a 360-day
year and an Actual/360 day-count convention, and the second is based on a 365-day
year and an Actual/365 day-count convention. In practice, the fixed rates payable are
quoted on a bond basis, and floating rates are quoted on a money-market basis. To
convert from a bond basis to a money-market basis, we have to multiply the quote by
360/365; whereas to do the reverse, we multiply by 365/360.

Key Dates in a Swap Contract

There are four important dates that have to be specified in a swap contract. Consider
the four-year swap between Scotia Bank and Dominion Bank. Assume that the swap
was negotiated on 15 June 20XX with a specification that the first payments are for a
six-month period commencing on 1 July 20XX. 15 June is referred to as the transaction
date. The date from which the interest counter payments start to accrue is termed the
effective date. In our example, the effective date is 1 July.
Our swap, by assumption, has a tenor of four years and consequently the last ex-
change of payments take place on 30 June 20XX+4. This date consequently is referred
to as the maturity date of the swap. We assume that the eight cash-flow exchanges will
occur on 31 December and 30 June of every year. The first seven dates, on which the
floating rate will be reset for the next six-monthly period, are referred to as reset or
re-fixing dates.

The Swap Rate

The fixed rate of interest that has been agreed upon in a coupon swap is referred to as
the swap rate. If the swap rate is quoted as a percentage, it is referred to as an all-in
price. However, in certain interbank markets, the fixed rate is not quoted as a percent-
age. Instead, what is quoted is the difference, in basis points, between the agreed upon
fixed rate and a benchmark interest rate. The benchmark chosen to compute this dif-
ferential is usually the government security whose remaining term to maturity is clos-
est to the life of the swap in question. For instance, in the case of the Scotia-Dominion
Interest Rate Swaps | 421

swap, the fixed rate is 5% per annum. If the swap rate were quoted as an all-in price,
it would obviously be reported as such. However, in the second convention, the rate
would be quoted as follows. Assume that a four-year T-note has a yield to maturity of
4.80%. The swap price is then quoted as 5% minus 4.80% or as 20 basis points.

Risk

Whether it is a coupon swap or a basis swap, an interest rate swap exposes both the
parties to interest rate risk. In the case of Scotia, which is the fixed rate payer in this
case, the risk is that the LIBOR may decline during the life of the swap. If so, the pay-
ments due to Scotia may stand reduced, whereas the payments to be made by it are
invariant to interest rate changes. On the other hand, the risk for Dominion, the float-
ing rate payer, is that the LIBOR may increase over the life of the swap. If so, Domin-
ion’s payment amounts will increase, whereas the payments it is due to receive will
be invariant to rate changes. The same is true in the case of a basis swap. Assume one
party pays Rate-1 and the other pays Rate-2, where both the rates are variable. The risk
scenarios for the party paying Rate-1 are that the increase in Rate-1 is more than the
increase in Rate-2, the decline in Rate-1 less than the decline in Rate-2, or Rate-1 in-
creases and Rate-2 declines. For the counterparty, the situation is just the opposite.
That is, the increase in Rate-2 is more than the increase in Rate-1, the decrease in Rate-
2 is less than the decrease in Rate-1, or Rate-2 increases and Rate-1 declines. Also, in
every swap, both the counterparties are exposed to default risk.

Quoted Swap Rates

A swap dealer quotes two rates for a coupon swap, a bid and an ask. The bid is the
fixed rate at which the dealer is willing to do a swap that requires it to pay the fixed
rate, and the ask represents the rate at which the dealer will do a swap that requires
it to receive the fixed rate. The bid will be lower than the ask.
Consider the hypothetical quotes for US dollar-denominated interest rate swaps
on a given day, shown in Table 14.3. Assume that the corresponding floating rate is the
six-month LIBOR.
We have assumed a spread of 5 bp for all tenors, to make matters simple. Let’s
consider the rates for a 1-year swap. The bid is 3.70%, and the ask is 3.75%. Thus if the
dealer does a swap where it has to make a fixed rate payment in return for a cash flow
based on a floating rate, it agrees to pay 3.70% per annum. However, if the dealer is
asked to do a swap wherein it receives the fixed rate in exchange for a floating rate, it
asks for a rate of 3.75% per annum. Equivalently, if the dealer pays the fixed rate in a
coupon swap, it pays 20 bp over the prevailing rate on a one-year Treasury security,
422 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

Table 14.3: All-in prices and spreads for swaps.

Tenor All-in Prices Spread over Treasury

Bid Ask Bid Ask

1-Year 3.70% 3.75% 20 bp 25 bp

2-Year 3.90% 3.95% 15 bp 20 bp
3-Year 3.25% 3.30% 25 bp 30 bp
5-Year 2.90% 2.95% 15 bp 20 bp
10-Year 3.45% 3.50% 20 bp 25 bp

whereas if it receives the fixed rate, it demands a spread of 25 bp over the yield on a
comparable Treasury security.

Comparative Advantage and Credit Arbitrage

There could be situations where a party has a disadvantage from the standpoint of
borrowing rates with respect to another party in the markets for fixed rate as well as
for variable rate loans. However, despite this, the former may have a comparative ad-
vantage in one of the two markets, as the following example illustrates.

Example 14.2. A company called Mount Holly, can borrow at a fixed rate of 5.25% and a variable rate
of LIBOR + 1% in the US debt market. On the other hand, another company, Checkmate, can borrow at
4.90% in the fixed rate market and LIBOR + 25 bp in the variable rate market. Thus Mount Holly has to
pay 35 bp more than Checkmate does if it borrows in the fixed rate market. However, it has to pay 75
basis points more if it borrows on a floating rate basis. Because Checkmate can get funds at a lower
rate in both the markets, we say that it enjoys an absolute advantage in both markets compared to
Mount Holly. However, because the spread for Mount Holly is lower in the fixed rate market, compared
to the floating rate market, we say that Mount Holly enjoys a comparative advantage in the fixed rate
market.
Assume that Mount Holly wants to borrow at a floating rate, whereas Checkmate would like to
borrow at a fixed rate. However, if Mount Holly borrows in the fixed rate market, an arena where it has
a comparative advantage, and then swaps the interest payments with Checkmate, it could be a win-
win situation for both the entities. In other words, as a consequence of the swap, both the parties can
borrow at a reduced rate of interest compared to what they would have had to pay in the absence of it.
Let’s assume that Mount Holly borrows $10 million at a fixed rate of 5.25% per annum, and Check-
mate borrows the same amount at LIBOR + 25 bp. The two parties can then enter into a swap wherein
the former agrees to pay interest on a notional principal of $10 million at the rate of LIBOR + 40 bp
per annum in exchange for a fixed rate payment based on a rate of 4.90% from the latter. The effective
interest rate for the two parties may be computed as follows:
Mount Holly: 5.25% + LIBOR + 40 bp − 4.90% = LIBOR + 75 bp
Checkmate: LIBOR + 25 bp + 4.90% − LIBOR − 40 bp = 4.75%
Mount Holly has a saving of 25 basis points on the floating rate debt, and Checkmate has a saving
of 15 bp on the fixed rate debt. Checkmate has an advantage of 75 bp in the market for floating rate
debt and 35 bp in the market for fixed rate debt. The difference of 40 bp manifests itself as the savings
for both parties considered together. In our illustration, Mount Holly has saved 25 bp, and the counter-
The Role of Banks in the Swap Market | 423

party has saved 15 bp, which adds up to 40 basis points. In practice, we can have any pair of numbers,
as long as the sum total is 40 bp.
In practice, a swap dealer such as a commercial bank plays a role in the transaction. Assume that
Mount Holly borrows at 5.25% per annum and enters into a swap with Scotia Bank wherein it has to
pay LIBOR + 55 bp in return for a fixed-rate payment based on a rate of 5.00%. Checkmate on the other
hand borrows at LIBOR + 25 bp and enters into a swap with the same bank wherein in receives LIBOR
+ 20 bp in return for payment of 4.80%.
The net result of the transaction may be summarized as follows:
Mount Holly: Effective interest paid = 5.25% + LIBOR + 55 bp − 5.00 = LIBOR + 80 bp
Checkmate: Effective interest paid = LIBOR + 25 bp + 4.80% − LIBOR − 20 bp = 4.85%
Scotia Bank: Profit from the transaction = LIBOR + 55 bp − 5.00 − LIBOR − 20 bp + 4.80 = 15 bp
The difference in this case is that the comparative advantage of 40 basis points has been split
three ways. Mount Holly saves 20 basis points; Checkmate saves 5 basis points; and the bank makes
a profit of 15 basis points.

The Role of Banks in the Swap Market

When the swap market was at its nascent stage the normal practice was for investment
banks to play the role of an intermediary. These banks would arrange the transaction
by bringing together two counterparties and in return would be paid an arrangement
fee. Over a period of time, the role of an intermediary evolved from that of an agent that
facilitated a swap to that of a principal. One of the main reasons for this was that par-
ties to swaps did not want their identities to be revealed to the counterparty. Second,
as we have seen, swaps expose both counterparties to default risk. For this reason,
parties to a swap were more comfortable dealing with a bank, whose creditworthi-
ness was easier to appraise. As the market has evolved, such arrangement fees have
become extremely rare except perhaps for contracts that are very exotic or unusual.
In the days when the market was at its infancy, banks would primarily do rever-
sals. A reversal entails the offsetting of a swap with a counter agreement with another
client. For instance, they would do a fixed-floating deal with a party only if they were
hopeful of immediately concluding a floating-fixed deal for the same tenor with a third
party. Parties that carry equal and offsetting swaps in their books are said to be run-
ning a matched book. These days banks are less finicky about maintaining such a
matched position, and in most cases are willing to take on the inherent exposure un-
til they eventually locate a party for an offsetting transaction. It must be understood
that a dealer that maintains a matched book is exposed to default risk from both the
parties with which it has entered into swaps.

Valuing an Interest Rate Swap

In our illustration of a swap between Scotia Bank and Dominion Bank, we arbitrarily
assumed that the fixed rate or swap rate was 5% per annum. We now demonstrate how
this rate is determined in practice.
424 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

Consider a two-year swap between Bank Alpha and Bank Beta. Bank Alpha has to
pay a fixed rate of k% per annum on a semiannual basis, whereas the counterparty has
to pay the six-month LIBOR every six months. As an alternative, assume that instead of
entering into a swap, Bank Alpha has issued a two-year fixed rate note with a principal
of $8 million on which it has to make semiannual interest payments at the rate of k%
per annum. This money has been used to acquire a two-year floating rate note with the
same principal, and which pays coupons semiannually based on the LIBOR observed
at the start of the six-monthly period.
If we look at the cash flows of this alternate arrangement, the result is equivalent
to that on a two-year fixed-floating swap. At the outset there is an inflow of $8 million
for Bank Alpha when the fixed rate note is issued. But this amount is just adequate to
purchase the floating-rate loan. Thus the net cash flow is zero. Similarly, at the point of
termination, that is after two years, Bank Alpha receives $8 million when the floating
rate note matures, and this amount also is just adequate to retire the fixed rate note.
The net result is that there is no exchange of principal, either at the outset or at the
end, which is consistent with what we have seen for interest rate swaps.
Every six months, the floating rate note pays a coupon based on the LIBOR at the
start of the period. Bank Alpha receives this amount and is required to pay interest at
the rate of k% per annum to service the fixed rate note that it has issued. Consequently
the cash flows every six months are identical to that of the swap. Thus a long position
in a floating rate note coupled with a short position in a fixed rate note is equivalent to a
swap that requires fixed-rate payments in return for payments based on a floating rate.
Now let’s demonstrate as to how the fixed rate of a coupon swap can be determined.
Let’s use the same vector of spot rates that we used to calibrate the Ho-Lee and
BDT models in Chapter 11.

Table 14.4: Vector of spot rates.

Period Spot Rate

0 7.50%
1 7.00%
2 6.25%
3 6.75%

Because, by assumption, the current point in time is the start of the next six-monthly
period, the price of the two-year floating rate note is equal to its face value of $8 mil-
lion. For, on a coupon reset date, the price of a default risk-free floating rate bond
reverts to its face value. The question to be answered is what the coupon rate should
be for the fixed rate note so that it too has a current price of $8 million. Let’s first de-
termine the discount factors corresponding to the observed LIBOR rates.
Valuing an Interest Rate Swap | 425

The discount factor for a given maturity is the present value of a dollar to be re-
ceived at the end of the stated period. The convention in the LIBOR market is that if
the number of days for which the rate is quoted is N, then the corresponding discount
1
factor is given by N . For instance, the discount factor for an investment of 18
(1+i× 360 )
1
months is where i is obviously the quoted 18-month LIBOR. Table 14.5 shows
(1+i× 540
360
)
the vector of discount factors for our example.

Table 14.5: Discount factors.

Time to Maturity Discount Factor

6M 0.9639
12M 0.9346
18M 0.9143
24M 0.8811

If we denote the semiannual coupon by C2 , it must be that

C C C C
× 0.9639 + × 0.9346 + × 0.9143 + [ + 8,000,000] × 0.8811
2 2 2 2
C C
= 8,000,000 ⇒ 3.6939 × = 951,200 ⇒ = 257,505.61 ⇒ C = 515,011.22
2 2

Thus the swap rate is:

515,011.22
× 100 = 6.4376%
8,000,000

Valuing a Swap at an Intermediate Stage

Assume that three months have elapsed since the preceding swap was initiated. Con-
sider the term structure in Table 14.6.

Table 14.6: Spot rates and discount factors after three months.

Time to Maturity Rate Discount Factor

3M 7.75% 0.9810
9M 7.25% 0.9484
15M 6.50% 0.9249
21M 6.80% 0.8937
426 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

The value of the fixed rate note is computed as follows:

C
× [0.9810 + 0.9484 + 0.9249 + 0.8937] + 8,000,000 × 0.8937
2
= 257,505.61 × 3.748 + 8,000,000 × 0.8937 = 965,131.10 + 7,149,600
= $8,114,731

The value of the floating rate bond is computed as follows. Three months hence it
pays a coupon based on the original six-month rate which is 7.50%. The amount of
this coupon is

0.5 × 0.0750 × 8,000,000 = $300,000

When this coupon is paid, the value of the bond reverts to its face value of $8 million.
Consequently, its value today is

8,300,000 × 0.9810
= $8,142,300

From the standpoint of the fixed-rate payer, the swap is tantamount to a long position
in a floating rate note that is combined with a short position in a fixed rate note. Thus
the value of the swap is

8,142,300 − 8,114,731 = $27,568.95

For the counterparty, the value is obviously −$27,568.95, for a swap is also a zero sum
game. A negative value indicates that the position holder has to pay to assign the swap
to someone else, whereas a positive value indicates that the holder receives the value
if it chooses to assign the swap to another party.

Terminating a Swap

Let’s suppose that after three months have elapsed, the fixed-rate payer in the preced-
ing swap decides that it no longer wants to be a party to the swap. It can get out of the
current situation in a variety of ways. One way is by way of a reversal. That is, it can
enter into a swap with 21 months to maturity, wherein it is required to pay floating and
receive fixed. The index for the floating-rate payments must obviously be the same. In
this case two swaps exist. So the party is exposed to credit risk in both swaps. The sec-
ond way to exit the swap is by selling it to a new party. In this case, Scotia Bank has to
be paid $27,568.95 by the acquirer because the swap has a positive value. In this case,
the original counterparty to the swap, that is Dominion Bank, has to agree to the deal.
Finally, Dominion itself may buy out the swap from Scotia by paying the value. This is
known as buy-back or close-out.
Motives for the Swap | 427

Motives for the Swap

A party to a swap may enter into the contract with a speculative motive, or else with
an incentive to hedge. In addition such transactions may also be used to undertake
credit arbitrage arising due to the comparative advantage enjoyed by the participating
institutions, as we learned earlier. Let’s analyze each of these potential uses.

Speculation

Scotia Bank and Dominion Bank are both players in the Canadian financial market.
However they have very different perspectives about the direction in which interest
rates are headed. Scotia believes the domestic interest rates are likely to increase
steadily over the next decade. On the contrary, Dominion is of the opinion that do-
mestic interest rates will decline steadily over the next 10 years. A coupon swap with
10 years to maturity is a suitable speculative tool for both parties. Scotia, which is
bullish about interest rates, can enter the contract as the fixed-rate payer, and Domin-
ion, which is bearish about interest rates, can be the counterparty as the fixed-rate
receiver. Obviously both the speculators cannot earn a profit, for one of them will be
proved wrong subsequently. If interest rates rise, Scotia, the fixed-rate payer, stands
to benefit. On the contrary, if yields decline, Dominion Bank, the floating-rate payer,
gets positive cash inflows.

Hedging a Liability

Swaps can be used as a hedge against anticipated interest rate movements. Parties
may choose to hedge an asset or a liability, depending on their prior position. Carpen-
ters Inc., a company based in Kansas City, has taken a floating-rate loan on which it
has to pay an interest rate equal to the six-month LIBOR + 1%. Its apprehension is that
rates may increase, and consequently it would have to pay more. It can use an interest
rate swap to convert its existing liability into an effective fixed-rate loan. One alterna-
tive way to do so in practice is to renegotiate the loan and have it converted to a loan
carrying a fixed rate of interest. This may not be easy in real life, for there are a lot
of administrative and legal issues and related costs. However, it is relatively easier to
enter into a swap with a bank, wherein the company has to pay a fixed rate in return
for a LIBOR-based payment.
Assume that Prudential Bank agrees to enter into a swap with Carpenters wherein
it pays LIBOR in return for a fixed interest stream based on a rate of 4.25% per annum.
One possibility is that the bank is bearish about interest rates and wants to speculate.
The other possibility is that the bank has an asset on which it is earning a floating
rate of interest. Being bearish, it seeks to hedge by entering into a coupon swap as a
fixed-rate receiver.
428 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

The net result from the standpoint of Carpenters may be analyzed as follows:
– Outflow-1(interest on the original loan): LIBOR + 1%
– Inflow-1(receipt from Prudential Bank): LIBOR
– Outflow-2 (payment to Prudential Bank): 4.25%
– Net Outflow: 4.25% + LIBOR + 1% − LIBOR = 5.25% per annum

Thus the company has converted its variable rate liability, to an effective fixed-rate
liability carrying interest at the rate of 5.25% per annum.
Yet another reason for using swaps is to change the mix of fixed-rate and floating-
rate debt on a company’s balance sheet. Riviera Corporation has a liability profile con-
sisting of $50 million in fixed-rate debt and another $50 million in floating-rate debt.
It is now taking over Rodeo Corporation, and after the amalgamation it will have $100
million of fixed-rate debt and $70 million of floating-rate debt. The firm is eager to
maintain the 50:50 ratio between fixed and floating debt after the merger. The total
debt post-merger will be $170 million, and a 50:50 ratio implies a fixed-rate debt of
$85 million and a floating-rate debt of an equivalent amount. One way of restructur-
ing is to borrow another $15 million at a floating rate and repay $15 million worth of
fixed-rate debt. Another alternative is the use of an interest rate swap.
Riviera can enter into a swap wherein it receives interest at a fixed rate on a no-
tional principal of $15 million and pays interest on a floating rate basis on the same
amount. The portfolio of the existing liabilities plus the swap is effectively a fixed-rate
liability of $85 million and a floating-rate liability of $85 million.
Yet another reason for using a swap is the inability of a party to obtain loans at
a fixed rate of interest. In practice a small company, a new company, or even an ex-
isting company that is large but has a weak credit rating, may be unable to borrow at
a fixed rate of interest. Such a company has to raise capital in the bond market, and
the bond markets are very particular about the credit rating of the potential borrower.
One possibility is to issue junk bonds carrying high coupons. An alternative is to bor-
row at a floating rate of interest and then do a coupon swap, wherein the entity pays a
fixed rate and receives a floating rate. The net result is the conversion of a floating-rate
liability to a fixed-rate liability.

Hedging an Asset

A swap can be used by a borrower to convert its liability from a fixed-rate loan to a
floating-rate loan or vice versa. Such contracts, however, may also be used by entities
that seek to transform the income from their assets from a fixed rate cash inflow to a
floating rate cash inflow or vice versa. For instance, a company that has invested in
fixed rate bonds may use an interest rate swap to convert it into a synthetic floating-
rate asset.
Assume that a company has bought 100,000 bonds with a face value of $1,000 and
a coupon of 6% per annum paid semiannually. Let’s assume that we are on the issue
Motives for the Swap | 429

date and that the value of the bonds is equal to the par value. On the issue date, or any
coupon date, the accrued interest will be zero. A conventional coupon swap, wherein
the company pays fixed and receives LIBOR, is appropriate under these circumstances.
However, consider a situation where we are three months into the coupon period.
The bond has 21 months to maturity. Table 14.7 shows the prevailing LIBOR values.

Table 14.7: Spot rates and discount factors after three months.

Time to Maturity Rate Discount Factor

3M 7.75% 0.9810
9M 7.25% 0.9484
15M 6.50% 0.9249
21M 6.80% 0.8937

In this case there is accrued interest for three months. In practice, investors prefer an
arrangement where the assets do not have any accrued interest, and where the swap
rate is equal to the coupon rate on the bonds. If we assume that the bonds are quoting
at 97-08, the clean price of the bond is

8
97 + 32
1,000 × = $972.50
100

The accrued interest, assuming a 30/360 day-count convention, is

30
1,000 × = $15
2

and consequently the dirty price is $987.50. In practice, investors entering into asset
swaps prefer contracts where the notional principal is equal to the par value of the
bonds. In our case the difference between the par value and the dirty price per bond
is $12.50, and for 100,000 bonds, it is $1,250,000. This amount is payable by the bond
holder to the counterparty. Thus the cash flows over the next 21 months are those in
Table 14.8.

Table 14.8: Cash flows from the asset holder’s perspective.

Time Amount

0 1,250,000
3M 3,000,000
9M 3,000,000
15M 3,000,000
21M 3,000,000
430 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

Each of the subsequent cash flows are $3 million, because the semiannual coupon
payment is $30 per bond.
The value of these cash flows given the LIBOR rates and using the market method
for computing the discount factors is

3,000,000 3,000,000
1,250,000 + 0.0775
+ 0.0725×3
(1 + 4
) (1 + 4
)
3,000,000 3,000,000
0.0650×5
+ 0.0680×7
(1 + 4
) (1 + 4
)
= $12,493,800

The corresponding fixed rate is given by

k
12,493,800 = × 0.9810 × 100,000,000
4
k
+ [0.9484 + 0.9249 + 0.8937] × 100,000,000
2
⇒ k = 7.6709%

Had the price of the bond been equal to par, and the counterparty had paid LIBOR,
the value of the cash flows would have been
100,000,000 × 0.0775 × 0.25 100,000,000 × 0.0725 × 0.50
0.0775
+ 0.0725×3
(1 + 4
) (1 + 4
)
100,000,000 × 0.065 × 0.5 100,000,000 × 0.068 × 0.5
0.0650×5
+ 0.0680×7
(1 + 4
) (1 + 4
)
= $11,382,938

The corresponding fixed rate is given by

k
11,382,938 = × 0.9810 × 100,000,000
4
k
+ [0.9484 + 0.9249 + 0.8937] × 100,000,000
2
⇒ k = 6.9889%

Thus the fixed rate is higher by 68.20 basis points because the bond is trading at a
discount. Consequently the counterparty pays LIBOR + 68.20 basis points.
Had the bond been at a premium, the analysis would be as follows. Assuming that
the bonds are quoting at 100-24, the clean price of the bond is

24
100 + 32
1,000 × = $1,007.50
100
Equivalence with FRAs | 431

The accrued interest is $15, and consequently the dirty price is $1,022.50. Thus the
difference between the par value and the dirty price per bond is $22.50, and for 100,000
bonds, it is $2,250,000. This amount is payable by the counterparty to the bond holder.
Each of the subsequent cash flows are $3 million, because the semiannual coupon
payment is $30 per bond. The value of these cash flows given the LIBOR rates and
using the market method for computing the discount factors is
3,000,000 3,000,000
− 2,250,000 + 0.0775
+ 0.0725×3
(1 + 4
) (1 + 4
)
3,000,000 3,000,000
0.0650×5
+ 0.0680×7
(1 + 4
) (1 + 4
)
= $8,993,800

The corresponding fixed rate is given by

k
8,993,800 = × 0.9810 × 100,000,000
4
k
+ [0.9484 + 0.9249 + 0.8937] × 100,000,000
2
⇒ k = 5.5220%

Had the bond been at trading at par, the corresponding fixed rate would have been
6.9889%. Thus the fixed rate is lower by 1.4669%, and the counterparty pays LIBOR −
1.4669%.

Equivalence with FRAs

We now show that a position in an interest rate swap is equivalent to a series of for-
ward rate agreements (FRAs). Consider a portfolio of one-period FRAs paying after
one, two, three, and four periods. Consider the case of an investor who has taken a
long position in the portfolio with a notional principal of $8 million. Let’s assume
that the FRAs are delayed settlement agreements that pay six months after the LIBOR
is determined. This is to be consistent with the payment convention of an interest rate
swap. Figures 14.1 and 14.2 show the interest rate, and state price trees, for the Ho-Lee
model that we derived in Chapter 11.

Determining the Fixed Rate

Let’s denote the unknown fixed rate by k. The value of the cash flow at time 1 is

8,000,000 × 0.5 × (0.075 − k)

0.075
(1 + 2
)
432 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

Figure 14.1: No-arbitrage interest rate tree.

t0 t1 t2 t3 t4

0.053160

0.112320

0.232247 0.215741

0.481928 0.340258

1.0 0.466755 0.328336

0.481928 0.343589

0.234508 0.222092

0.115651

0.056336

Figure 14.2: State price tree.

Equivalence with FRAs | 433

The value of the payoff from the FRA at time 2 is

8,000,000 × 0.5 × (0.075063 − k)

0.481928 × 0.075063
(1 + 2
)
8,000,000 × 0.5 × (0.055063 − k)
+ 0.481928 × 0.055063
(1 + 2
)

The value of the payoff from the FRA at time 3 is

8,000,000 × 0.5 × (0.067725 − k)

0.232247 × 0.067725
(1 + 2
)
8,000,000 × 0.5 × (0.047725 − k)
+ 0.466755 × 0.047725
(1 + 2
)
8,000,000 × 0.5 × (0.027725 − k)
+ 0.234508 × 0.027725
(1 + 2
)

Finally, the value of the payoff at time 4 is

8,000,000 × 0.5 × (0.112863 − k)

0.112320 × 0.112863
(1 + 2
)
8,000,000 × 0.5 × (0.092863 − k)
+ 0.340258 × 0.092863
(1 + 2
)
8,000,000 × 0.5 × (0.072863 − k)
+ 0.343589 × 0.072863
(1 + 2
)
8,000,000 × 0.5 × (0.052863 − k)
+ 0.115651 × 0.052863
(1 + 2
)

The unknown fixed rate k should be set such that the value of the portfolio is zero.
Using SOLVER in Excel, we get a value of 6.7485% per annum.
An identical solution can be obtained by computing the coupon rate correspond-
ing to a par bond:

1,000 × 0.5 × k 1,000 × 0.5 × k

1,000 = 0.075
+
(1 + ) 0.07 2
2 (1 + 2
)
1,000 × 0.5 × k [1,000 + 1,000 × 0.5 × k]
+ +
0.0625 3 0.0675 4
(1 + 2
) (1 + 2
)
k = 6.7485%

The equivalence of swaps and FRAs may be demonstrated as follows. The first pay-
ment due to Scotia Bank is known at the very outset because the applicable LIBOR is
determined at the start of the period. Hence the first transaction may be viewed as a
434 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

spot transaction in which Scotia Bank receives $300,600 from Dominion six months
hence. This may be computed as

8,000,000 × 0.5 × (0.075 − 0.067485) = $300,600

The remaining three transactions are FRAs in which Scotia Bank agrees to make pay-
ments based on a rate of 6.7485% per annum and receive payments based on the LIBOR
that prevails at the start of the corresponding period.
The par bond can also be valued as the present value of a series of cash flows
based on the implied one-period forward rates. These are

(1.035)2
= 1.032506 ≡ 3.2506%
1.0375
(1.03125)3
= 1.023791 ≡ 2.3791 %
(1.035)2
(1.03375)4
= 1.041286 ≡ 4.1286%
(1.03125)3

The forward rate for the first period is equal to the spot rate, which is equal to 7.50% per
annum. If we assume a face value of 1,000, the cash flows are $37.50 after six months,
$32.506 after 12 months, $23.791 after 18 months, and $41.286 after 24 months. If we
denote the unknown fixed rate by k

1,000 × 0.5 × k 1,000 × 0.5 × k

0.075
+
(1 + ) 0.07 2
2 (1 + 2
)
1,000 × 0.5 × k [1,000 + 1,000 × 0.5 × k]
+ +
0.0625 3 0.0675 4
(1 + 2
) (1 + 2
)
37.50 32.506
= 0.075
+
(1 + ) 0.07 2
2 (1 + 2
)
23.791 [1,000 + 41.286]
+ + = 1,000
0.0625 3 0.0675 4
(1 + 2
) (1 + 2
)
⇒ k = 6.7485%

Forward-Start Swaps
A plain vanilla interest rate swap starts at time zero. That is, the first cash flows arise
one period after inception. However, we can design contracts that are scheduled to
start at a later date. Such contracts are termed forward-start swaps. The contract rate
for such a swap will obviously be different from that for a corresponding plain vanilla
swap. However, it can be determined using the same vector of spot rates.
Forward-Start Swaps | 435

For instance, let’s consider a three-period (18 months) swap that is scheduled to
come into existence six months from now. The cash flows based on the one-period
forward rates are $32.506 after 12 months, $23.791 after 18 months, and $41.286 after
24 months. If we denote the unknown contract rate by k
1,000 × 0.5 × k 1,000 × 0.5 × k [1,000 + 1,000 × 0.5 × k]
+ +
0.07 2 0.0625 3 0.0675 4
(1 + 2
) (1 + 2
) (1 + 2
)
32.506 23.791 [1,000 + 41.286]
= + + ⇒ k = 6.4822%
0.07 2 0.0625 3 0.0675 4
(1 + 2
) (1 + 2
) (1 + 2
)

The same answer can be derived by computing the coupon rate corresponding to
a par bond. For this we need to compute the one-period, two-period, and three-period
forward rates one period from now.
The one-period forward rate, per annum, for a loan after one period is

3.2506 × 2 = 6.5012%

The two-period forward rate, per annum, for a loan after one period is
1/2
(1.03125)3
{[ ] − 1} × 2
(1.0375)
= 5.6278%

The three-period forward rate, per annum, for a loan after one period is
1/3
(1.03375)4
{[ ] − 1} × 2
(1.0375)
= 6.5006%

The coupon rate of the par bond may be calculated as

1,000 × 0.5 × c 1,000 × 0.5 × c [1,000 + 1,000 × 0.5 × c]
1,000 = + +
(1.032506) (1.028139)2 (1.032503)3
⇒ c = 6.4823%

Amortizing Swaps

In an amortizing swap, the notional principal declines steadily over the life of the
swap. Let’s reconsider the swap between Scotia Bank and Dominion Bank. Assume
that the initial notional principal is $8 million, and that it declines by $2 million at
the end of every six-monthly period. The fixed rate may be determined as follows
[250 + 1,000 × 0.5 × k] [250 + 750 × 0.5 × k]
1,000 = 0.075
+
(1 + ) 0.07 2
2 (1 + 2
)
436 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

[250 + 500 × 0.5 × k] [250 + 250 × 0.5 × k]

+ +
0.0625 3 0.0675 4
(1 + 2
) (1 + 2
)
k = 6.7375%

In-Arrears Swaps

In the case of an in-arrears swap, as soon as the LIBOR is determined at the end of
every period, the interest is paid out immediately. We can determine the fixed rate for
such a swap by viewing it as a portfolio of in-arrears FRAs.
Let’s denote the unknown fixed rate by k. The cash flow at time 1 is

0.481928 × [8,000,000 × 0.5 × (0.075063 − k)]

+ 0.481928 × [8,000,000 × 0.5 × (0.055063 − k)]

The payoff from the FRA at time 2 is

0.232247 × [8,000,000 × 0.5 × (0.067725 − k)]

+ 0.466755 × [8,000,000 × 0.5 × (0.047725 − k)]
+ 0.234508 × [8,000,000 × 0.5 × (0.027725 − k)]

Finally, the payoff at time 3 is

0.112320 × [8,000,000 × 0.5 × (0.112863 − k)]

+ 0.340258 × [8,000,000 × 0.5 × (0.092863 − k)]
+ 0.343589 × [8,000,000 × 0.5 × (0.072863 − k)]
+ 0.115651 × [8,000,000 × 0.5 × (0.052863 − k)]

The unknown fixed rate k should be set such that the value of the portfolio is zero.
Using SOLVER in Excel, we get a value of 6.5016% per annum.

Extendable and Cancelable Swaps

An extendable swap gives one of the counterparties, usually the fixed-rate payer, the
option to extend the maturity date beyond the scheduled date. From the perspective of
a fixed-rate payer, the option to extend is likely to be invoked in an economic environ-
ment in which interest rates are increasing, in which the cash inflows due to floating-
rate payments are likely to be higher than the counter payments based on the fixed
rate. The extension option needs to be priced and has a higher fixed rate compared to
a plain vanilla interest rate swap.
Swaptions | 437

A cancelable swap gives one of the counterparties the option to terminate the swap
prior to the scheduled maturity date. From the perspective of a fixed-rate payer, the
option to cancel is likely to be invoked in an economic environment in which interest
rates are declining, in which the cash inflows due to floating-rate payments are likely
to be lower than the counter payments based on the fixed rate. A swap with this option
has a higher fixed rate compared to a plain vanilla swap. A swap that is cancelable by
the fixed-rate payer swap is also termed a callable swap.
Another type of a cancelable swap is a putable swap. This gives the floating-rate
payer the right to terminate the swap prior to the original maturity date. Obviously a
payer chooses to do so when interest rates are rising. Consequently the fixed rate for
a swap with a put option is lower than that of a plain vanilla swap.

Swaptions
A swaption is an option on a swap. Such an option requires the buyer to pay an up-front
premium, for the right to enter into a coupon swap. There are two types of swaptions.
The holder of the option may enter into a coupon swap as a fixed-rate payer or as a
fixed-rate receiver. In the case of a payer swaption, the option holder can exercise it to
enter into the swap as a fixed-rate payer. On the other hand, a receiver swaption gives
the holder the right to enter into a swap as a fixed-rate receiver.
The exercise price specified in the swaption is an interest rate. The underlying as-
set is a swap with a specified term to maturity. A payer swaption is exercised only if the
prevailing rate for a swap with the specified maturity is higher than the exercise price
of the swaption. Quite obviously, a receiver swaption is exercised only if the prevailing
swap rate is lower than the exercise price. Swaptions may be European or American
in nature.
Swaptions may be useful for entities that are not sure whether they will face in-
terest rate exposure in the future. Consider the case of a corporation that decides to
borrow in the future. It obviously worries about the specter of a rising interest rate. It
can protect itself by buying a payer swaption. If rates go up, it borrows at the market
rate and exercises the option, which results in receipt of a series of positive cash flows
because the market rates are higher than the swap rate. However, if rates decline, it
refrains from exercising the option and borrows at the market rate, which by assump-
tion has declined and consequently is favorable to it.
An asset holder, such as a bank that holds a portfolio of mortgages, will be
perturbed about falling interest rates, which are likely to lead to substantial pre-
payments. Such an entity can protect itself by buying a receiver option. If rates go up,
the protection is not required, and the option is allowed to expire. However, if rates
decline, the bank exercises the swaption and receives a series of positive cash flows
that help mitigate the effect of prepayments from the mortgage holders.
438 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

Issuers of callable bonds can also use swaptions. Assume that a company has is-
sued 10-year deferred callable bonds with a 5-year call protection period. Let’s assume
that the coupon rate is 7.50% per annum. At the end of two years, rates decline substan-
tially, to say 4% per annum. Given a choice, the issuer wants to call back the bonds.
However, it is constrained by the call protection feature. In such circumstances, it can
sell a receiver option with a strike price of 7.50%, for which it obviously receives a
premium. Three years hence if interest rates are more than 7.50%, the swaption is not
exercised, and the bonds are not recalled. However, if the rate falls to say 5%, the coun-
terparty does exercise the option. The issuer of callables receives LIBOR in return for
a fixed-rate payment of 7.50%. It can then call back the bonds and issue fresh floating
rate bonds carrying an interest of LIBOR. The funds required to pay the coupon come
from the counterparty to the swaption. Thus the consequence of the transaction is that
the company continues to pay 7.50% for the remaining life of the bond. At the outset,
however, it receives a premium due to the sale of the option. It is as if it has sold the
call option inherent in the callable bond, although the inherent option per se is not a
tradable asset. Consequently, this strategy is termed call monetization.
It is not necessary that the company sell a swaption with an exercise price equal to
the coupon of 7.50%. It can choose to sell a swaption with a lower or a higher exercise
price. If it chooses a lower exercise price of, say, 6.00%, and the rate after two years is
4%, the counterparty will exercise. The effective rate on its debt is 6%. The premium
received for the receiver swaption is lower in this case because the exercise price of
6% is lower than the exercise price of 7.50% that was assumed earlier. If the company
sells a swaption with a higher exercise price of, say, 9%, it has to pay a higher rate on
its debt, but receives a higher premium at the outset.
In practice, the swap rate and the bond rate need not be perfectly correlated. As-
sume that the company has sold a receiver swaption with an exercise price of 7.50%.
The LIBOR is assumed to be 4%. The holder pays the LIBOR in return for a fixed rate
of 7.50%. However, the credit quality of the bond issuer may have deteriorated, and
as a consequence, it has to pay LIBOR plus 1.50% if it issues floating rate bonds to
refinance its debt. The net result is that the rate for the issuer is 9% per annum.

Credit Default Swaps

A credit default swap (CDS) is a contract between two parties, known as the protection
buyer and the protection seller, respectively. At the outset, the buyer of protection has
to pay a premium to the counterparty. In return it gets protection against credit risk
or default risk pertaining to an issuing entity, or to a specific security that has been
issued by an entity. The security or entity for which protection has been sought, is
known as the reference asset. In practice it may be a standalone asset, or a basket or
portfolio of assets. The protection seller is required to make a payment if a specified
credit event occurs. If the event does not happen, the seller makes no payment. In
Credit Default Swaps | 439

the case of the credit event happening, the seller pays the loss due to the event to the
buyer. The contract then terminates. The purpose of a CDS is to transfer the credit risk
pertaining to an asset without transferring its ownership.
The specified credit event may be one of the following:
– Bankruptcy or insolvency
– Default on a payment
– Decline in the price of a specified asset by more than a certain prespecified amount
– Credit downgrade of a security or an issuing entity

If a credit event occurs, it leads to one of the following situations. The protection seller
determines the post-default value, or in other words, the percentage of the distressed
asset’s value that can be recovered. The par value minus this post-default value is
then paid to the protection buyer. There also is something called a digital swap. In this
case, the post-default price or recovery value is a prespecified amount. This mode of
settlement is termed cash settlement. Finally, the seller may pay the entire par value to
the buyer and take over the security in return. This is referred to as delivery settlement.
Here is an example of cash settlement.

Example 14.3. Arizona Bank has bought a CDS from Xylo Bank. The notional principal is $25 million.
Every year the buyer has to pay a premium of 80 basis points to the seller. Thus the premium in this
case is

25,000,000 × 0.008 = $200,000

If there is a credit event, and assuming that the recovery rate is 60% of par, the seller pays

(100 − 60)
25,000,000 × = $10,000,000
100

In the absence of the event, nothing is payable, and the seller retains the premiums.

In practice the premiums are paid quarterly or semiannually in arrears, and it is not
always necessary that the credit event conveniently occurs at the beginning of a pe-
riod. If a credit event occurs in the middle of a period, the accrued premium has to be
paid. A CDS is a tool that facilitates the trading of credit risk, just the way other tools
facilitate the trading of market risk. It should be noted that the maturity of the swap
contract does not need to match the maturity of the reference asset, and in practice, it
does not in most cases.
If the contract is on an issuing entity and not on a specific issue, the buyer may
have a choice of deliverable assets or what are termed deliverable obligations. In such
a situation, the buyer delivers the cheapest of the assets that are eligible for delivery.
In principle, there is no difference between cash settlement and delivery settlement
from the standpoint of protection against a credit event. However, in practice, deliv-
ery settlement offers a benefit to the seller. If the status of the issuer or the security
440 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

improves after the seller has taken possession of the underlying asset, it may be able
to derive a higher recovery value than what was anticipated at the time of default. In
the market, buyers prefer cash settlement. This is because, if they are not in prior pos-
session of the underlying asset, they have to acquire it in the market. In this case, if
there are liquidity issues, the price of acquisition of the asset may be higher than what
the recovery rate warrants. In practice, physical settlement is commonly used because
it facilitates the determination of the security’s market value. In practice, if there is a
credit event, a cooling period such as a fortnight is provided so that the market stabi-
lizes, before the post default value is determined. The contracts usually stipulate that
if the value cannot be ascertained after a credit event, the value of a similar asset in
terms of maturity and credit quality can be used.
In the case of a basket CDS, which is based on a portfolio of assets, one possibility
is a first to default swap. As the name suggests, the swap terminates the moment one of
the constituent assets suffers a credit event, unless of course the CDS itself terminates
before any of the component securities experience a credit event. In the case of such
swaps, if the buyer seeks protection for the remaining constituents of the portfolio, it
has to enter into a fresh contract, for the original contract becomes void the moment
a credit event occurs. An alternative is an all to default swap, which terminates when
all the assets in the basket have defaulted, unless of course the contract itself termi-
nates prior to that. In this case, the buyer receives compensation for all the defaulting
securities.

Valuation of a CDS

A combination of a long position in a risky bond and a long position in a CDS is equiv-
alent to a long position in a synthetic T-bill. Thus the premium for the CDS should be
approximately equal to the default risk premium inherent in the bond’s yield. Take the
case of a party that short sells a riskless bond with a coupon of r% and buys a risky
bond with a coupon of c%, where c = r + p, where p is the default risk premium. To
keep the issue simple, let’s assume that p is a constant. If we assume that both bonds
are trading at par, the proceeds from the short sale are adequate to buy the risky bond,
and consequently there is neither a net inflow nor a net outflow. Assume that the in-
vestor goes long in a CDS and has to pay a premium of s% every period. At the end of
every coupon period, there is an inflow of c = r + p from the risky bond and a outflow
of r. There also is an outflow on account of the swap premium. The net cash flow is
c − (r + s).
On the coupon date, if the risky bond does not default, the holder can sell it and
retire the short position in the riskless bond. However, if the risky bond defaults, the
holder can deliver it under the swap in return for the par value if it is delivery settled, or
sell it at the prevailing price, receive the deficit under the swap, and use the proceeds
to retire the short position. In either case the net cash flow is zero. Because the cash
Credit Default Swaps | 441

flows at the outset and at the end are zero, all intermediate cash flows should be zero
to rule out arbitrage. Thus

c − (r + s) = 0 ⇒ c = r + s = r + p ⇒ s = p

This confirms our claim that the swap premium should be equal to the default risk pre-
mium. There are some inherent assumptions in this argument. First, the credit spread
p is a constant. Second, the bonds are floating rate issues, which reset to par on every
coupon date. And as usual, we assume that there are no market imperfections such
as transaction costs, and if an asset is short-sold, the full proceeds are immediately
available for use.

Using Default Probabilities to Determine the Swap Rate

Consider an eight-year swap. Every period has a 2.50% probability of default, assum-
ing that there is no earlier default. We assume that the swap premium is paid in arrears
at the end of each year. If a default occurs, it happens at the middle of the period. The
recovery rate is assumed to be 60%. We denote the unknown swap premium by s per
dollar of notional principal. The survival probabilities are given in Table 14.9.

Table 14.9: Default and survival probabilities.

Time Default Survival

Probability Probability

1 0.025000 0.975000
2 0.024375 0.950625
3 0.023766 0.926859
4 0.023171 0.903688
5 0.022592 0.881096
6 0.022027 0.859069
7 0.021477 0.837592
8 0.020940 0.816652

The survival probability at the end of the first year is one minus the default prob-
ability, which is 1.00 − 0.025 = 0.975. The default probability for the second year
is 0.025 × 0.975000 = 0.024375. Thus the probability of survival after two years is
0.975000 − 0.024375 = 0.950625. In general the survival probability at the end of pe-
riod t is (1.000 − 0.025)t . The default probability for period t is the survival probability
at the end of the previous period multiplied by the default intensity of 2.50%.
442 | Chapter 14: Interest Rate Swaps and Credit Default Swaps

Assume that the discount rate is 6% per annum. Thus the discount factor for a
cash flow occurring at time t is

1
(1.06)t

At the end of every period, if there has been no prior default, a premium of s is payable
per dollar. Hence, the expected payoff at the end of the year is the corresponding sur-
vival probability multiplied by s.

Table 14.10: Expected premium payments and their present values.

Time Survival Expected Discount PV of Expected

Probability Premium Pmt. Factor Premium Pmt.

1 0.975000 0.975000s 0.943396 0.919811s

2 0.950625 0.950625s 0.889996 0.846052s
3 0.926859 0.926859s 0.839619 0.778208s
4 0.903688 0.903688s 0.792093 0.715805s
5 0.881096 0.881096s 0.747258 0.658406s
6 0.859069 0.859069s 0.704960 0.605609s
7 0.837592 0.837592s 0.665057 0.557046s
8 0.816652 0.816652s 0.627412 0.512377s
Total 5.593314s
Pmt. stands for Payment.

Table 14.11: Expected accrued premium payments and expected payoffs and their present values.

Time Default Expected Discount PV of Expected PV of Exp.

Probability Payoff Factor Expected Accrual Accrual
Payment Payoff Payment Payment
0.5 0.025000 0.010000 0.971286 0.009713 0.012500s 0.012141s
1.5 0.024375 0.009750 0.916307 0.008934 0.012188s 0.011168s
2.5 0.023766 0.009506 0.864441 0.008217 0.011883s 0.010272s
3.5 0.023171 0.009268 0.815510 0.007558 0.011586s 0.009448s
4.5 0.022592 0.009037 0.769349 0.006953 0.011296s 0.008691s
5.5 0.022027 0.008811 0.725801 0.006395 0.011014s 0.007994s
6.5 0.021477 0.008591 0.684718 0.005882 0.010739s 0.007353s
7.5 0.020940 0.008376 0.645960 0.005411 0.010470s 0.006763s
Total 0.059063 0.073830s
Exp. stands for Expected.

The present value of the expected swap payment is 5.593314s+0.073830s = 5.667144s.

The break-even CDS spread is given by

5.667144s = 0.059063 ⇒ s = 0.010422 ≡ 104 basis points

Chapter Summary | 443

The expected accrual premium at time t is the default probability multiplied by half
the premium, as default is assumed to occur in the middle of the year. The expected
payoff from the swap is the default probability multiplied by 0.40 because the recovery
rate is assumed to be 60%.

Chapter Summary
This chapter looked at two important OTC derivatives, namely interest rate swaps and
credit default swaps. We demonstrated the valuation of an interest rate swap given a
vector of spot rates. We showed that it is equivalent to a combination of a fixed rate
bond and a floating rate bond. We also illustrated that an interest rate swap is equiv-
alent to a portfolio of FRAs and showed how to compute the swap rate given a model
for determining the short rates such as the Ho-Lee model. The alternatives to a plain
vanilla interest rate swap, such as a forward-start swap, an amortizing swap, and an
in-arrears swap, were also discussed. It was shown that swaps could be used for both
hedging and speculation, in conjunction with both assets and liabilities. While dis-
cussing the possible motives for a swap, the concept of credit arbitrage was discussed
in detail. The chapter concluded with a study of credit default swaps. We showed how
to determine the swap premium for such contracts by assuming survival probabilities
and a recovery rate.
With this final chapter, the book is concluded. This book seeks to provide a com-
prehensive treatment of bonds, bond valuation, and yield computation. Derivatives
based on interest rate products are also extensively covered. To facilitate the com-
prehension, two detailed chapters, on time value of money, and derivatives, are pre-
sented. At every appropriate step, the use of Excel for solving problems is demon-
strated in adequate detail.
Appendix A
Goal Seek
Goal Seek is a tool in Excel, which can be used to determine the value that needs to
be input in a cell so that the target value is achieved in a formula cell. It is called
Goal Seek because the objective is to achieve a computational goal, by using Excel to
determine the appropriate value of the input cell. We will use Goal Seek to determine
the target present value of a series of cash flows, by varying the discount rate. Consider
the following Excel sheet.

Table A.1: Goal Seek example.

Row Time Cash Flow Present Value Input Value

Number Column A Column B Column C Column D
1 1 2,000 = PV($D$1, A1, , −B1) 0.10
2 2 3,000 = PV($D$1, A2, , −B2)
3 3 4,000 = PV($D$1, A3, , −B3)
4 4 6,000 = PV($D$1, A4, , −B4)
5 Total Present Value = SUM(C1 : C4)

The objective is to determine the value in cell D1 that will make the sum of the present
values equal to $12,500. We have chosen to use an initial input of 10% per period,
which will obviously be entered as 0.10. Notice that to compute the present value of
a cash flow we need to always use the value in cell D1. Consequently we enter the
rate parameter in the PV function as $D$1. The formula for PV need not be input four
times. We can simply enter the function in cell C1, and drag and use the control D
option in Excel.
If we invoke Goal Seek it asks for the following:
– Set cell
– To value
– By changing cell

We give C5, 12,500, and D1 as the inputs. Then we need to click on OK. The answer is
6.4948%. The result is shown in the table below.

Table A.2: The result from Goal Seek.

Row Time Cash Flow Present Value Input Value

Number Column A Column B Column C Column D
1 1 2,000 1,878.03 0.064948
2 2 3,000 2,645.24
3 3 4,000 3,311.88
4 4 6,000 4,664.85
5 Total Present Value 12,500

https://doi.org/10.1515/9781547400669-015
Solver | 445

Solver
Solver also enables us to achieve a target value for a formula cell. However it allows us
to change the values of multiple cells to achieve our goal. In our illustration, however,
we need to change the value of only a single input cell.
If we use Solver to achieve a goal it asks for the following:
– Set objective:
– To: (there are three options: Max; Min; and Value Of)
– By changing variable cells:
– Subject to the constraints:
– Option to Make Unconstrained Variables Non Negative:
– Select a Solving Method:

We give C5, 12,500 (by choosing the value of option), and D1 as the inputs. We do not
input any constraints and do not choose the option to make unconstrained variables
non negative. The default solving method is GRG Nonlinear, and that is adequate for
most problems. We then need to click on Solve. The solution once again is 6.4948%.
Bibliography
[1] Annaert, J., Claes, A. G. P., de Ceuster, M. J. K., and Zhang H. Estimating the Yield Curve Using
the Nelson Siegel Model: A Ridge Regression Approach. International Review of Economics and
Finance, 2012.
[2] Babbel, D. F. and Merrill, C. B. Valuation of Interest-Sensitive Financial Instruments.
Schaumburg, IL: Society of Actuaries, 1996.
[3] Bierwag, G. O. Duration Analysis. Cambridge, MA: Ballinger Publishing, 1987.
[4] Blake, D. Financial Market Analysis. New York, NY: John Wiley, 1999.
[5] Bodie, Z., Kane, A., Marcus, J. A., and Mohanty P. Investments. Chennai, McGraw Hill Education,
2018.
[6] Brown, P. J. Bond Markets. Chicago: Glenlake Publishing, 1998.
[7] Chance, D. M. An Introduction to Options and Futures. New York: The Dryden Press, 1991.
[8] Chance, D. M. An Introduction to Derivatives and Risk Management. Mason, OH:
South-Western, 2004.
[9] Choudhry, M. An Introduction to Repo Markets. London: Securities Institute, 2001.
[10] Choudhry, M. An Introduction to Bond Markets. London: Securities Institute, 2001.
[11] Choudhry, M. Analyzing and Interpreting the Yield Curve. New York: John Wiley, 2004.
[12] Comotto, R. The Money Market. London: Euromoney Publications, 1998.
[13] Coyle, B. Corporate Bonds & Commercial Paper. Canterbury U.K.: Financial World Publishing,
2002.
[14] Coyle, B. Government Bonds. Canterbury U.K.: Financial World Publishing, 2002.
[15] Coyle, B. Interest-Rate Options. London: Financial World Publishing, 2001.
[16] Coyle, B. Interest-Rate Swaps. London: Financial World Publishing, 2001.
[17] Coyle, B. Money Markets. London: Financial World Publishing, 2001.
[18] Daigler, R. T. Financial Futures and Options Markets: Concepts and Strategies. Harper Collins,
1994.
[19] Davidson, A. S., and Herskovitz, M. D. The Mortgage-Backed Securities Workbook. New York:
Irwin McGraw Hill, 1996.
[20] Davidson, A., Sanders, A., Wolff, L., and Ching, A. Securitization. New York: John Wiley, 2003.
[21] Dodd, R. Primer on Repurchase Agreements Market. Online at www.financialpolicy.org.
[22] Douglas, L. G. Bond Risk Analysis. New York: New York Institute of Finance, 1990.
[23] Dubofsky, D. A. Options and Financial Futures: Valuation and Uses. New York: McGraw-Hill,
1992.
[24] Dym, S. I. The Complete Practitioner’s Guide to the Bond Market. New York: McGraw Hill, 2010.
[25] Edwards, F. R., and Ma, C. W. Futures and Options. New York: McGraw-Hill, 1992.
[26] Fabozzi, F. J. Bond Markets: Analysis and Strategies. Englewood Cliffs, NJ: Prentice Hall, 2015.
[27] Fabozzi, F. J. Fixed Income Mathematics. New York: McGraw Hill, 2006.
[28] Fabozzi, F. J. Handbook of Fixed Income Securities. New York: McGraw Hill, 2012.
[29] Fabozzi, F. J., Bhattacharya, A. K., and Berliner, W. S. Mortgage-Backed Securities. New York:
John Wiley, 2007.
[30] Fabozzi, F. J., and Mann, S. V. Introduction to Fixed Income Analytics. New Hope, PA: Frank J
Fabozzi Associates, 2001.
[31] Fabozzi, F. J., and Modigliani, F. Mortgage and Mortgage-Backed Securities Markets.
Cambridge, MA: Harvard Business School Press, 1992.
[32] Figlewski, S., Silber, W. L., and Subrahmanyam, M. G. Financial Options: From Theory to
Practice. New York: Irwin, 1990.
[33] Garbade, K. D. Fixed Income Analytics. Cambridge: MIT Press, 1996.
[34] Geroulanos P. An Introduction to Swaps. London: The Securities Institute, 1998.

https://doi.org/10.1515/9781547400669-016
Bibliography | 447

[35] Gray, S. Repo of Government Securities: Handbooks in Central Banking, No. 16. London: Bank
of England, 1998.
[36] Hegde, S. P., and Parameswaran, S. K. Futures Markets: 200 Questions and Answers. New York:
John Wiley, 2007.
[37] Hull, J. Fundamentals of Futures and Options Markets. Englewood Cliffs, NJ: Prentice Hall,
2004.
[38] Hull, J. Options, Futures, and Other Derivatives. Upper Saddle River, NJ: Prentice Hall, 2006.
[39] James, T. Interest Rate Derivatives. London: Risk Books, 2006.
[40] Kellison, S. G. The Theory of Interest. New York: Irwin McGraw-Hill, 1991.
[41] Kolb, R. W. Futures, Options, and Swaps. New York: Blackwell, 2003.
[42] Kolev, I. Mortgage-Backed Securities. Washington, DC: Financial Policy Forum, 2004.
[43] Liaw, K. T. Capital Markets. Mason, OH: Thomson/South-Western, 2004.
[44] Liaw, K. T. and Moy, R. L. Stocks, Bonds, Futures, and Options. New York: McGraw Hill, 2001.
[45] Livingston, M. Bonds and Bond Derivatives. New York: Blackwell Publishing, 2004.
[46] Logue, D. E. The WG&L Handbook of International Finance. Mason, OH: South-Western, 1995.
[47] McDonald, R. L. Derivatives Markets. New York: Pearson Education, 2006.
[48] Mclnish, T. H. Capital Markets: A Global Perspective. New York: Blackwell, 2000.
[49] Mobius, M. Bonds. Singapore: John Wiley, 2012.
[50] Needles, B. E. and Powers, M. Principles of Financial Accounting. Boston: Houghton Mifflin,
2008.
[51] Parameswaran, S. K. Fundamentals of Financial Instruments. Singapore: John Wiley, 2011.
[52] Parameswaran, S. K. Futures and Options: Concepts and Applications. Singapore: McGraw Hill,
2011.
[53] Rendleman, R. J. Applied Derivatives. Oxford, U.K.: Wiley-Blackwell, 2002.
[54] Saunders, A., and Cornett, M. M. Financial Institutions Management. New York: McGraw-Hill,
2003.
[55] Siegel, D. R., and Siegel, D. F. Futures Markets. New York: The Dryden Press, 1990.
[56] Smith, D. J. Bond Math. New Jersey: John Wiley, 2011.
[57] Steiner, B. Foreign Exchange and Money Markets. Oxford, UK: Butterworth-Heinemann, 2002.
[58] Steiner, R. Mastering Financial Calculations. London: Financial Times Prentice Hall, 1998.
[59] Stigum, M., and Robinson, F. L. Money Market and Bond Calculations. New York: Irwin, 1996.
[60] Stigum, M., and Crescenzi, A. Stigum’s Money Market. New York: McGraw-Hill, 2007.
[61] Strumeyer, G. Investing in Fixed Income Securities. New York: John Wiley, 2005.
[62] Sundaresan, S. Fixed Income Markets and Their Derivatives. San Diego: Academic Press, 2009.
[63] Taff, L. G. Investing in Mortgage Securities. St. Lucie Press, 2003.
[64] Taggart, R. A. Quantitative Analysis for Investment Management. Englewood Cliffs: Prentice
Hall, 2002.
[65] Veronesi, P. Fixed Income Securities. New Delhi: Wiley India, 2012.
[66] Windas, T., and Miller, T. Introduction to Option Adjusted Spread Analysis. Chennai: Viva
Books, 2009.
Index
30/360 European method 75, 169 Black-Scholes model for option valuation 301
30/360 method 73, 74, 169 Bond equivalent yield 81, 164, 165, 168, 182,
234, 237
Bond insurance 60
A Bootstrapping 100–102
Accounting for bonds Break-even for a convertible bond 409
– the effective interest method (EIM) 54, 56 Buy-down 203
– the straight line method (SLM) 54–56
ACCRINT function in Excel 67, 71
C
Accrual bonds 243, 247 C-Strips 52, 53
Accrued interest 64, 65, 68–73, 75, 76, 78, 96, Calibrating the BDT model 319, 424
123, 126, 127, 139, 170, 171, 173, 174, 176, Calibrating the Ho-Lee model 311, 326, 390
191, 364–366, 368–370, 378, 429, 431 Call monetization 438
Actual term to maturity (ATM) 34, 173 Call options 287, 288, 291, 303, 408
Actual/360 method 76, 77, 169, 343 Callable bonds 33, 385, 401, 414, 438
Actual/365 method 75–77, 169, 343 Callable swaps 437
Actual/actual method 66, 67, 72, 75, 169 Cancelable swaps 437
Additive adjustment 273, 274 Caps 184, 200, 331, 333, 340
Adjustable-rate mortgages (ARMs) 203, 210, 211 Captions 339, 340
Adjusted simple margin 190 Carryovers 213
Adjusted total margin 192 Cash and carry arbitrage 269–271, 349, 352
All to default swap 440 Cash flow yield 12, 14, 234, 237
American options 286, 290, 292, 294, 304 Changes in the conversion ratio 407, 408
Amortization with a balloon payment 23, 24 Changing the duration of a bond portfolio 381,
Amortized loans 29, 204 382
Amortizing bonds 45, 64 Cheapest to deliver (CTD) bond 367, 382
Amortizing swaps 435 Clearinghouse 161, 260, 261, 267
Approximate yield to call 388 Collars 200, 249, 331, 336, 340
Approximate yield to maturity 79, 81 Collateralized mortgage obligations 238, 254
Arrow-Debreu securities 312, 313, 340 Commercial paper 114, 177, 182
Average life 64, 94, 95, 118, 222, 223, 225, 229, Comparative advantage 422, 423, 427
234–236, 241, 243, 246, 250, 253, 254 Competitive and non-competitive bids 49, 50,
162
Conditional pre-payment rate (CPR) 227, 229
B Continuously callable bonds 386
Bankers’ acceptances 181, 182 Continuously compounded interest 7
Basis 1–4, 11, 35, 42, 50, 63, 71–74, 77, 80, 81, Contraction risk 242, 254
89, 90, 129, 140, 158, 160, 164, 166, 168, Conversion clause 203
169, 178, 182, 186, 281, 286, 333, 341, 342, Conversion factors 273, 364, 382
362, 416, 419, 421, 422, 428 Conversion factors for T-Note and T-Bond futures
Basis risk 351, 357 363
Bermudan options 386, 405 Conversion premium 406, 410
Bid to cover ratio 50 Conversion price 405, 406, 408, 409
Bills of exchange 180, 181 Conversion ratio 405–411
Binomial model for option valuation 295 Conversion value 406, 407, 410, 414
Black-Derman-Toy (BDT) model 305, 307, 326, Convertible bonds 33, 45, 385, 405, 406, 408,
340, 345, 348, 390, 397 409, 411, 414, 415

https://doi.org/10.1515/9781547400669-017
Index | 449

Convexity 117, 118, 133–141, 143, 144, 149, 189, Duration of a level annuity 121
200 Duration of a perpetuity 122, 124
Convexity of a floating rate bond 189 Duration of a perpetuity due 122, 124, 148
Convexity of a level annuity 140 Duration of a risky floater 189, 199, 201, 202
Convexity of a perpetuity 142, 149 Duration of an annuity due 128, 144, 146, 148
Convexity of a perpetuity due 142, 149 Dutch auctions 49
Convexity of a zero coupon bond 137, 139
Convexity of an annuity due 141
E
Cost of carry relationship 271
EFFECT function in Excel 6
COUPDAYBS function in Excel 77
Effective convexity 144
COUPDAYS function in Excel 77
Effective cost of a CD 177
COUPDAYSNC function in Excel 77
Effective duration 130, 144
COUPNUM function in Excel 77
Effective interest rates 7, 360, 422
Coupon effect 99, 117, 402
End-of-month option 374, 377
Coupon rolls 53
End/end rule 156
Coupon strips 51
Equal principal repayment loan 228
Coupon yield curves 102
Equity 31, 48, 203, 405, 406
Coupon-Linkers (C-Linkers) 193, 195, 200
Equivalence of a swap with FRAs 433
Credit arbitrage 422, 443
Equivalence of interest rates 7
Credit default swaps (CDS) 416, 438–440, 443
Equivalent life 95
Credit risk or default risk 204
Escrow accounts 219
Credit watch/rating watch 60
Eurobonds and foreign bonds 40, 64
Cum-dividend bond price 69, 70, 79
Eurodollar (ED) futures 341, 348, 382
Current yield 79, 80
European options 284, 286, 291, 292
CUSIP 50, 52
Ex-dividend bond price 68–70
Exchangeable bonds 385, 405, 411
D
Extendable swaps 436
DATE function in Excel 71
Extension risk 242, 243, 249
Day-count conventions 64, 66, 67, 73, 76, 96,
332, 417
Debentures 1, 30, 31 F
Deferred callable bonds 386 Federal funds 153, 161, 173, 360, 361, 363
Delayed settlement FRAs 347 Federal funds futures 382
Deleveraged floating rate bonds 186 Federal Home Loan Mortgage Corporation
Delta of an option 303 (FHLMC) 233
Descartes’ rule of signs 13 Federal National Mortgage Association (FNMA)
Digital CDS 439 233
DISC function in Excel 168 First to default swap 440
Discount margin 190, 192, 193, 196–198, 200 Fitch 64, 154
Discretely callable bonds 386 Floating rate tranches 247, 254
Dispersion 117, 136–139, 144 Floors 184, 200, 331, 333, 340
Dollar convexity 139 Floortions 339, 340
Dollar duration 129, 133, 134 Following business day convention 155, 156
Dual-indexed floating rate bonds 187 Foreign exchange risk 63
Duration 117–131, 133–139, 142–144, 147–149, Forward rate agreements (FRAs) 304, 382
189, 199–202, 211, 332, 380–383 Forward rates 104, 111, 117, 196, 197, 305, 324
DURATION function in Excel 129 Forward-start swaps 434
Duration of a floating rate bond 126, 139, 143, French auctions 49
148, 188, 189 FV function in Excel 12
450 | Index

G K
Government National Mortgage Association Key dates in a swap contract 420
(GNMA) 233, 234
Graduated payment mortgages 213, 229 L
Growing annuities 24, 25, 29 Letters of credit (LCs) 178
Growing annuity dues 26, 29 Level annuities 14
Growing equity mortgages 214 Level annuities due 17
Growing perpetuities 26 Leverage 31, 63
Growing perpetuity dues 28 LIBID 158, 182
Guaranteeing fee 234 LIBOR 157, 158, 182–184, 186–188, 191–193,
196, 208, 209, 212, 213, 247, 333, 337,
H 341–343, 348, 350, 351, 355, 357, 418, 421,
Haircuts 171 424, 428, 431, 434, 436, 438
Hedging 275–277, 280, 281, 304, 341, 342, 354, LIMEAN 158
358, 360, 378, 380, 382, 443 Liquid yield option notes (LYONs) 410
Hedging assets with swaps 428 Liquidity preference theory 110, 111, 113, 115
Hedging liabilities with swaps 427 Liquidity risk 61, 204
Hedging the CTD bond 380 Loan to value (LTV) ratio 221
High yield 126 Locking in a borrowing rate using ED futures
Ho-Lee model 305, 307, 309, 311, 321, 330, 340, 349
401, 405, 414, 431, 443 Locking in a lending rate using ED futures 351
Horizon/Holding period yield (HPY) 88
Hybrid ARMs 210 M
Maintenance margin 171, 264
I Market method for bond valuation 67
ICE LIBOR 158 Market segmentation hypothesis 114
Implied repo rate (IRR) 270, 271, 368, 372 Marking to market 171, 261–263, 277–280, 284
Implied reverse repo rate (IRRR) 371, 373 MDURATION function in Excel 129, 130
In-arrears FRAs 436 Modified duration 128–130, 134
In-arrears swaps 436 Modified following business day convention 155
Inflation indexed bonds 193 Modified simple margin 190, 191
Inflation risk 62, 63 Money market forward rates 159
Initial margin 262, 264, 265 Money substitute hypothesis 114
Interest equalization tax 41 Moneyness of an option 288
Interest rate caps 212 Moody’s 58, 60, 64, 154, 179, 180
Interest rate swaps 416, 419, 421, 424 Mortgage insurance 221
Interest-only (IO) ARMs 211 Mortgage servicing 218
Interest-only (IO) strips 249 Multiplicative adjustment 273, 367
Interest-rate risk 62
Internal rate of return (IRR) 12, 79 N
Intrinsic value 289–292, 294, 300, 301, 331 Negative accrued interest 70, 79
Inverse floating rate bonds 185 Negative amortization 207, 208, 211–214, 243
Inverse floating rate tranches 247 Negative carry/negative funding 80
IPMT function in Excel 22, 23 Negotiable certificates of deposit 173
ISIN 50 Nelson-Siegel model 106, 107
Nelson-Siegel-Svensson model 108
J NOMINAL function in Excel 6
Jensen’s inequality 112 Nominal interest rates 1, 6, 29
Jumbo mortgages 221 NORMSDIST function in Excel 301
Index | 451

Notional interest-only tranches 248, 254 Re-openings of securities 48

Novation 260 Realized compound yield (RCY) 64, 87, 88, 90,
96
O Realized compound yield with taxes 89
Option price/option premium 285, 289, 290, Receiver swaptions 437, 438
297 Relationship between YTC and YTM 387, 388
Original term to maturity (OTM) 34, 173, 174 Repurchase agreements (Repos) 169, 182
Reverse cash and carry arbitrage 271, 272, 349,
P 353, 371
P-Strips 52 Reverse repos 182
Par bond yield curves 102, 103 Risk in arbitrage strategies due to multiple
Parity value 406 deliverable grades 370
Payer swaptions 437 Rolling premium put option 408
Payment to income (PTI) ratio 221
Payment-in-kind (PIK) bonds 47, 64 S
Perpetuities 19, 139, 140, 142, 144, 146, 149 Serial bonds 93
Perpetuity dues 128 Servicing fee 219, 220, 234
Perturbation hedge ratios 380 Settlement price 264, 265, 374, 378
Planned amortization class (PAC) bonds 249 Short rates 305, 309, 324, 325, 345, 414, 443
PMT function in Excel 22, 23 Sight bills 181
Points 203, 290, 351 Simple (Japanese) yield to maturity 80
Portfolio convexity 138 Simple margin 190
Portfolio duration 131, 132 Single month mortality (SMM) 223–227, 229,
Portfolio yield 90, 91 235–238, 249, 250, 253–257
PPMT function in Excel 22, 23 Single period immunization 143, 144, 149
Pre-payment burnout 207 Sinking funds 93
Pre-payment risk 249 Special purpose vehicles (SPVs) 51
Preceding business day convention 155, 156 Speculation 117, 281–283, 304, 341, 382, 427
Preferred habitat theory 115 Speculation with swaps 443
PRICE function in Excel 67, 71 Spot rates 98–103, 105, 106, 108–111, 113, 115,
Principal-linkers (P-Linkers) 193, 200 117, 305, 326, 397, 402, 424, 434, 443
Principal-only (PO) strips 249 Spot-futures convergence 266, 280
Pull to par effect 37, 38, 64, 387 Stack hedges 359
Pure and mixed cash flows 13, 29 Standard and poor’s 58
Pure or unbiased expectations hypothesis 109 Step-down coupons 47, 64
Put options 130, 183, 286–288, 291–295, Step-up coupons 47
301–303, 385, 395, 405, 408, 409, 411, Straight value 406
414, 415 Strike price/exercise price 285
Put-call parity 292, 302 Strip hedges 358
Putable bonds 33, 385, 394, 395, 397, 409, 414 Sub-prime mortgages 221
Putable swaps 437 SUMPRODUCT function in Excel 120, 223
PV function in Excel 386 Support bonds 249, 254
Swap rates 421
Q Swaptions 437, 438
Quality option 378 Synthetic zero coupon bonds 42

R T
Range notes 187 T-bill bond equivalent yield 164
RATE function in Excel 45, 79, 81, 90 T-bill discount yield 163
452 | Index

T-bill holding period yield 166 V

T-bill money market yield 163 Valuation of a callable bond 390
T-bill tail 166 Valuation of a CDS 440
T-bond futures 363, 382 Valuing a callable bond 390
T-note futures 341 Valuing a convertible bond 411
Tax shield 30–32 Valuing a putable bond 395
Taxable equivalent yield (TEY) 92, 96 Valuing a swap 425
TBILLEQ function in Excel 168 Variation margin 265
TBILLPRICE function in Excel 167, 168 Volatility 116, 293, 294, 302, 307, 319, 321
TBILLYIELD function in Excel 167
Term certificates of deposit 174 W
Term structure of interest rates 100, 182, 200, Weighted average coupon (WAC) 228
307, 319
Weighted average maturity (WAM) 228
Terminating a swap 426
When issued (WI) trading 53
The option-adjusted (OAS) spread 404, 415
Wild card option 374
The static spread 402, 415
The traditional yield spread 401
Y
Time or usance bills 181
Time value 289–291, 299 Yield curves 97, 100, 109
Timing option 373–378 YIELD function in Excel 79
Timing risk 63, 385, 395 Yield to average life 94–96
Treasury auctions 48 Yield to call (YTC) 387, 388
Treasury inflation-protected securities (TIPs) Yield to equivalent life 95
193 Yield to maturity 33, 34, 37, 44, 64, 119, 123,
Treasury method for bond valuation 67 124, 126, 385, 387, 389, 414
Yield to maturity with taxes 395
U Yield to worst 389
Using default probabilities to find the rate for a
CDS 441 Z
Using ED futures to create a fixed rate loan 357 Zero cost collars 337
Using short rates to determine the contract rate Zero coupon bonds 42, 45, 82, 98, 100, 122,
for a FRA 345 124, 137, 158, 340

MVU
100% (1)
MVU
72 pages
ETFs Are Eating The Bond Market
No ratings yet
ETFs Are Eating The Bond Market
31 pages
Barney, J. B & Hesterly, W. S. (2019) Strategic Management and Competitive Advantage. Concepts and Cases
0% (1)
Barney, J. B & Hesterly, W. S. (2019) Strategic Management and Competitive Advantage. Concepts and Cases
21 pages
Merrill Lynch Credit Derivatives Handbook 2006 Volume 1
No ratings yet
Merrill Lynch Credit Derivatives Handbook 2006 Volume 1
118 pages
Unofficiall Abbott 2nd Edition Solutions
No ratings yet
Unofficiall Abbott 2nd Edition Solutions
210 pages
PCA Using R
No ratings yet
PCA Using R
12 pages
Across - The - Curve - in - Rates - and - Structured - Products - and - Across - The - Grade - in - Credit - Products - December 19, 2006 - (Bear, - Stearns - & - Co. - Inc.) PDF
No ratings yet
Across - The - Curve - in - Rates - and - Structured - Products - and - Across - The - Grade - in - Credit - Products - December 19, 2006 - (Bear, - Stearns - & - Co. - Inc.) PDF
60 pages
Expected Loss of Credit Products and Credit Portfolios
100% (1)
Expected Loss of Credit Products and Credit Portfolios
4 pages
2020 08 JRP Inflation Expectations
No ratings yet
2020 08 JRP Inflation Expectations
542 pages
Financial Engineering Proposed Applications
No ratings yet
Financial Engineering Proposed Applications
19 pages
Lecture 1 - Structured Product
No ratings yet
Lecture 1 - Structured Product
39 pages
Integrated Bank Risk Management: Interest Rate Risk in The Banking Book
No ratings yet
Integrated Bank Risk Management: Interest Rate Risk in The Banking Book
10 pages
Understanding Analysis Solutions: Michael You
No ratings yet
Understanding Analysis Solutions: Michael You
155 pages
FINC001
0% (1)
FINC001
33 pages
Deviation Interest Rate Parity
No ratings yet
Deviation Interest Rate Parity
58 pages
CB Primer Gs
No ratings yet
CB Primer Gs
141 pages
GPF Interest Rates 1963 To 2024
No ratings yet
GPF Interest Rates 1963 To 2024
1 page
Lucius Riccio Presentation, 2014 MIT Sloan Sports Analytics Conference
100% (1)
Lucius Riccio Presentation, 2014 MIT Sloan Sports Analytics Conference
47 pages
Practical Guide To Principal Component N R
No ratings yet
Practical Guide To Principal Component N R
43 pages
Sherman, Rick - Business Intelligence Guidebook
50% (2)
Sherman, Rick - Business Intelligence Guidebook
1 page
Week 3 - 03
No ratings yet
Week 3 - 03
70 pages
Market Skew
No ratings yet
Market Skew
39 pages
Year Ahead 2022
No ratings yet
Year Ahead 2022
72 pages
Standards of Value: Theory and Applications
From Everand
Standards of Value: Theory and Applications
Jay E. Fishman
No ratings yet
Ef4331 Case Study 1
100% (1)
Ef4331 Case Study 1
4 pages
Audit Planning Memo - Sample
100% (1)
Audit Planning Memo - Sample
11 pages
RV YTM Model PDF
0% (1)
RV YTM Model PDF
47 pages
Market Consistency: Model Calibration in Imperfect Markets
From Everand
Market Consistency: Model Calibration in Imperfect Markets
Malcolm Kemp
No ratings yet
Management Lessons From Mahabharat
100% (1)
Management Lessons From Mahabharat
18 pages
Trading Plan Harian - Binary Options
No ratings yet
Trading Plan Harian - Binary Options
6 pages
Credit Risik Default Models Credit Derivatives
No ratings yet
Credit Risik Default Models Credit Derivatives
192 pages
High Speed Computations
No ratings yet
High Speed Computations
16 pages
1, DOA LEASE BG-SBLC 11+2% BPU (HSBC) (12) (1) (7) (2) (5) (4) (4) (5) (6) (2) (Enregistré Automatiquement) Corrige
No ratings yet
1, DOA LEASE BG-SBLC 11+2% BPU (HSBC) (12) (1) (7) (2) (5) (4) (4) (5) (6) (2) (Enregistré Automatiquement) Corrige
28 pages
Equity and Foreign Exchange Hybrid Models For Pricing Long-Maturity Financial Derivativesn
No ratings yet
Equity and Foreign Exchange Hybrid Models For Pricing Long-Maturity Financial Derivativesn
184 pages
Manufacturing and Managing Customer-Driven Derivatives
From Everand
Manufacturing and Managing Customer-Driven Derivatives
Dong Qu
No ratings yet
1.1.6 Topic-Architectural Features About Hotels HMO-341
No ratings yet
1.1.6 Topic-Architectural Features About Hotels HMO-341
18 pages
Development Assistant Cover Letter
100% (1)
Development Assistant Cover Letter
8 pages
Singular Value Decomposition Tutorial - Kirk Baker
No ratings yet
Singular Value Decomposition Tutorial - Kirk Baker
24 pages
Case Study of HY Diaries
100% (1)
Case Study of HY Diaries
4 pages
Human Resource Management in Organizations
100% (1)
Human Resource Management in Organizations
30 pages
Bloomberg Fair Value Market Curves: Michael Lee Fixed Income Specialist, Bloomberg LP
100% (2)
Bloomberg Fair Value Market Curves: Michael Lee Fixed Income Specialist, Bloomberg LP
20 pages
Ng-Phelps-2010-Barclays Capturing Credit Spread Premium
No ratings yet
Ng-Phelps-2010-Barclays Capturing Credit Spread Premium
24 pages
Immunization Strategies
No ratings yet
Immunization Strategies
75 pages
Council Tax Bill 2023/24
No ratings yet
Council Tax Bill 2023/24
2 pages
Financial Mathematics
No ratings yet
Financial Mathematics
14 pages
Green Finance
No ratings yet
Green Finance
4 pages
Mason C-RE HDB Test Report - Shear 2016
No ratings yet
Mason C-RE HDB Test Report - Shear 2016
5 pages
David A. Freedman - The Limits of Econometrics PDF
100% (1)
David A. Freedman - The Limits of Econometrics PDF
13 pages
Life Settlements and Longevity Structures: Pricing and Risk Management
From Everand
Life Settlements and Longevity Structures: Pricing and Risk Management
Geoff Chaplin
No ratings yet
Stochastic Volatiity Models 2005 PDF
No ratings yet
Stochastic Volatiity Models 2005 PDF
35 pages
L5 Castagna and Mercurio
No ratings yet
L5 Castagna and Mercurio
6 pages
Convexity Adjustments and Forward Libor Model
No ratings yet
Convexity Adjustments and Forward Libor Model
19 pages
A Simple Multi-Curve Model For Pricing SOFR Futures and Other Derivatives
No ratings yet
A Simple Multi-Curve Model For Pricing SOFR Futures and Other Derivatives
18 pages
Relative Value Models (Feb04)
100% (1)
Relative Value Models (Feb04)
18 pages
Next Generation Credit Curves
No ratings yet
Next Generation Credit Curves
2 pages
Laplace Transform
No ratings yet
Laplace Transform
35 pages
The Exponencial Yield Curve Model
No ratings yet
The Exponencial Yield Curve Model
6 pages
The Definitive Guide To CDOs
100% (1)
The Definitive Guide To CDOs
4 pages
Expected Returns On Stocks and Bonds
No ratings yet
Expected Returns On Stocks and Bonds
21 pages
Credit Models and the Crisis: A Journey into CDOs, Copulas, Correlations and Dynamic Models
From Everand
Credit Models and the Crisis: A Journey into CDOs, Copulas, Correlations and Dynamic Models
Damiano Brigo
No ratings yet
Jio-Bill Mar25
No ratings yet
Jio-Bill Mar25
1 page
Asset and Equity Volatilities
No ratings yet
Asset and Equity Volatilities
40 pages
(Andrew Davidson & Co) An Implied Prepayment Model For MBS
No ratings yet
(Andrew Davidson & Co) An Implied Prepayment Model For MBS
13 pages
Ipath S&P 500 Vix Short-Term Futures Etn (Nysearca: VXX) : Investment Thesis
100% (1)
Ipath S&P 500 Vix Short-Term Futures Etn (Nysearca: VXX) : Investment Thesis
4 pages
Derivatives Exotic Path Dependent Options
No ratings yet
Derivatives Exotic Path Dependent Options
21 pages
Percentage Sheet - 06
No ratings yet
Percentage Sheet - 06
3 pages
AD Prepay Model Meth
No ratings yet
AD Prepay Model Meth
21 pages
European Bond Futures 2006
No ratings yet
European Bond Futures 2006
10 pages
(Deutsche Bank) Credit Derivatives - Issues & Trends PDF
No ratings yet
(Deutsche Bank) Credit Derivatives - Issues & Trends PDF
4 pages
How to Implement Market Models Using VBA
From Everand
How to Implement Market Models Using VBA
Francois Goossens
No ratings yet
TTP 0611 Can You Time LongDuration
No ratings yet
TTP 0611 Can You Time LongDuration
3 pages
11 17 06 Modeling Sovereign Correlations
No ratings yet
11 17 06 Modeling Sovereign Correlations
21 pages
Equity Variance Swaps With Dividends OpenGamma
No ratings yet
Equity Variance Swaps With Dividends OpenGamma
13 pages
Atrlist
No ratings yet
Atrlist
42 pages
Economics of Digitization
No ratings yet
Economics of Digitization
12 pages
Modelling Risk in Central Counterparty Clearing Houses:: A Review
No ratings yet
Modelling Risk in Central Counterparty Clearing Houses:: A Review
13 pages
Longstaff Schwartz (95) Risky Debt (P)
No ratings yet
Longstaff Schwartz (95) Risky Debt (P)
18 pages
松下公司人事变动
No ratings yet
松下公司人事变动
33 pages
An Alternative Bond Relative Value Measure: Determining A Fair Value of The Swap Spread Using Libor and GC Repo Rates
No ratings yet
An Alternative Bond Relative Value Measure: Determining A Fair Value of The Swap Spread Using Libor and GC Repo Rates
6 pages
Agricutural Structure and Farmers
No ratings yet
Agricutural Structure and Farmers
16 pages
Deliver
No ratings yet
Deliver
2 pages
CDO Base Correlations
No ratings yet
CDO Base Correlations
7 pages
A Brief Comparison of Interpolation Methods in Yield Curve Construction
No ratings yet
A Brief Comparison of Interpolation Methods in Yield Curve Construction
3 pages
A - 7LLAW024W.2 International Project Finance Law 2024-UPDATED
No ratings yet
A - 7LLAW024W.2 International Project Finance Law 2024-UPDATED
3 pages
Convertible Warrant
No ratings yet
Convertible Warrant
3 pages
FAQs MBA Admissions22-23
No ratings yet
FAQs MBA Admissions22-23
3 pages
General Mamert Natividad, Nueva Ecija
No ratings yet
General Mamert Natividad, Nueva Ecija
1 page
Sunil Sir
No ratings yet
Sunil Sir
1 page
Sole Source Fog Data
No ratings yet
Sole Source Fog Data
3 pages
Test I. Matching Type. Write The Letter On The Space Provided. No Erasures!
No ratings yet
Test I. Matching Type. Write The Letter On The Space Provided. No Erasures!
1 page