1
Lecture 4
Last week:
1) Graham (’00)
2) Mackie-Mason (’90)
3) Rajan & Zingales (’95)
4) Baker & Wurgler (’02)
Today: Gale & Hellwig (1985), Berger & Ofek (1995),
Lamont (1998)
Gale-Hellwig (’85): complete contracts. Costly State
Verification (see Townsend (’79))
Idea: when investors give money to firms /
entrepreneurs, they need to ensure they get their
money back. Firms generally know more about the
available cash flows. What contracts are optimal
[Mechanism Design / Security Design]
Results: debt contracts are optimal under some
assumptions (see also Innes ’92 for an ‘effort’ model)
Main problems: not time-consistent, not renegotiation
proof (the ‘two-period problem’ never got published)
Main contribution: economic intuition
PhD course in Corporate Finance Lecture 4 Jan Mahrt-Smith
Rotman School, U. of Toronto Capital Structure Winter 2007
2
Set-Up:
• A firm needs I to finance a project
• Uncertain Payoffs are X with probability
density h(X)
• The firm observes X
• The investor can observe X at a cost k
• Everyone is risk neutral, there are two dates
(one period), there is no discounting
• There is no renegotiation, and ‘verification’ is
not allowed to be random
• Capital Markets are competitive (investors
break even on average)
Timing and definitions:
• Firm borrows I
• Firm observes X ≥ 0
• Firm reports some ‘message’ about X, call it
m(X)
• Investors decide whether or not to verify, call
verification B(m) = 1, non-verification B(m)
=0
• ‘unaudited’ Payment R(m|B(m)=0) = R(m)
• ‘audited’ Payment R(m,X |B(m)=1) =
R(m,X)
PhD course in Corporate Finance Lecture 4 Jan Mahrt-Smith
Rotman School, U. of Toronto Capital Structure Winter 2007
3
Insight 1: because financial markets are competitive,
investors will just break even on average. Hence, the
firm’s optimization problem becomes a problem of
minimizing expected verification costs:
∞
min k × ∫ B ( X ) h ( X ) dX
0
Insight 2: REVELATION PRINCIPLE (RP)
The outcome of every mechanism can be replicated
with a truthful, direct mechanism
Direct: the entrepreneur’s message is the cash flow
Truthful: the entrepreneur chooses to report the truth
Note: the RP does NOT say that all optimal
mechanisms are truthful and direct. It simply states
that we will find ALL OPTIMAL OUTCOMES if we
restrict our search to truthful, direct mechanisms
(otherwise we would have to search among ALL
FEASIBLE mechanisms).
Note: the RP extends to multi-agent settings.
PhD course in Corporate Finance Lecture 4 Jan Mahrt-Smith
Rotman School, U. of Toronto Capital Structure Winter 2007
4
The Problem:
∞
min k × ∫ B ( X ) h ( X ) dX
{B ( i ); R ( X ); R ( X , X )} 0
s.t.
∞ ∞
IRI : ∫ B ( X ) ⎡⎣ R ( X , X ) − k ⎤⎦ h ( X ) dX + ∫ ⎡⎣1 − B ( X ) ⎤⎦ R ( X ) h ( X ) dX ≥ I
0 0
ICF ,1 : ∀m ≠ X s.t. ⎡⎣ B ( X ) = 0, B ( m ) = 0 ⎤⎦ => R ( X ) ≤ R ( m )
ICF ,1 : ∀m ≠ X s.t. ⎡⎣ B ( X ) = 1, B ( m ) = 0 ⎤⎦ => R ( X , X ) ≤ R ( m )
ICF ,2 : ∀m ≠ X s.t. ⎡⎣ B ( X ) = 0, B ( m ) = 1⎤⎦ => R ( X ) ≤ R ( m, X )
ICF ,1 : ∀m ≠ X s.t. ⎡⎣ B ( X ) = 1, B ( m ) = 1⎤⎦ => R ( X , X ) ≤ R ( m, X )
LLF ,1 : if ⎡⎣ B ( X ) = 0 ⎤⎦ => R ( X ) ≤ X
LLF ,2 : if ⎡⎣ B ( X ) = 1⎤⎦ => R ( X , X ) ≤ X
The Solution: Standard Debt Contracts are Optimal
K* ∞
- ∫ ⎡⎣ R ( X ) − k ⎦⎤h ( X ) dX + K * ∫ * h ( X ) dX = I
0 K
- for all X < K*, B*(X) = 1 and R*(X,X) = X
- for all X ≥ K*, B*(X) = 0 and R(X) = K*
Note: There might be other contracts that lead to
equivalent outcomes, but there will be no contract that
dominates the debt contract.
PhD course in Corporate Finance Lecture 4 Jan Mahrt-Smith
Rotman School, U. of Toronto Capital Structure Winter 2007
5
Admissible Mechanisms:
Define R ≡ inf { R ( X ) s.t. B ( X ) = 0}
X
a. Then for all X s.t. B(X)=0, we get R(X) = R
b. Also, for all X s.t. B(X)=1, we get R(X,X) ≤ R
Call mechanisms that satisfy a. and b. and that are
feasible (R(X) ≤ X and R(X,X) ≤ X) ‘admissible’
B(X)=1
B(X)=0
= B(X)
= R(X)
R X
Standard debt is the ‘best’ admissible contract. They
minimize the ‘verification region’.
PhD course in Corporate Finance Lecture 4 Jan Mahrt-Smith
Rotman School, U. of Toronto Capital Structure Winter 2007
6
Main Problems:
1) The multi-period model is very difficult
2) Renegotiation: once a firm reports small cash
flows, there is no need to ‘verify’ (everyone tells
the truth). In fact, both parties should agree to
split the ‘savings’ from not ‘verifying’
Somewhat Similar model: Diamond (’84) – the
foundation of banking
Hart (2001)
Incomplete Contracts: What Are They? How Can We
Use The Idea?
Hart (’95) – Good introductory survey to incomplete
contracts.
PhD course in Corporate Finance Lecture 4 Jan Mahrt-Smith
Rotman School, U. of Toronto Capital Structure Winter 2007
7
Diversified Firms: a Window into How Firms
Work
Q: Why might Diversified Firms be different?
A1: Risk Sharing – little conclusive evidence
A2: IO – reasons (competition) – not interested in
this from a finance perspective
A3: Internal Capital Markets might be good/bad
Theories: Gertner, Scharfstein, Stein (QJE ‘94), Stein
(JF’97)
Idea: in diversified firms, the providers of finance are
the ‘owners’ of the assets, while in stand-alone firms,
the providers of finance are NOT the owners. I.e.
managerial control (ownership) matters and has
interesting effects
Results: in diversified firms …
1) … investors ‘own’ the assets. Thus, they will
monitor more. That is because monitoring
generates information, and information is only
useful if you have the control rights to act on it.
PhD course in Corporate Finance Lecture 4 Jan Mahrt-Smith
Rotman School, U. of Toronto Capital Structure Winter 2007
8
More results: in diversified firms …
2) … managers are no longer the owners of the
assets, they will exert less ‘effort’. This is
because, when their effort is successful in
generating opportunities, the ‘owners’
(headquarters) will extract some of the rents by
exercising (or threatening to exercise) their
control. Whenever a non-effort exerting party can
share in some of the rents, the effort exerting
party will ‘undersupply’ effort. [Note: this is
related to the monitoring story of equity
ownership dispersion – Burkart et al. (QJE 98)
show, too much monitoring that comes with
concentrated equity will also destroy managerial
initiative].
3) … finance providers are also the ‘owners’. Thus,
assets can often re-deployed more easily (say to
other internal divisions). This is also related to
control and information (Stein ’97)
PhD course in Corporate Finance Lecture 4 Jan Mahrt-Smith
Rotman School, U. of Toronto Capital Structure Winter 2007
9
Scharfstein & Stein (’00) and Rajan, Servaes &
Zingales (’00)
Idea: internal capital is allocated in a ‘political’ way.
This can lead to ‘corporate socialism’ where capital
goes to ‘poor’ divisions. It also leads to inefficient
‘influencing activities’ that are wasteful and distortive.
Results: what types of divisions should be combined
in a conglomerate, what are their optimal sizes, …
PhD course in Corporate Finance Lecture 4 Jan Mahrt-Smith
Rotman School, U. of Toronto Capital Structure Winter 2007
