Chapter Four: Bank Debt and Market Debt

4.1. Introduction

Evidence
 Firm Reputation building (Diamond 1991)
 Capital availability (Holmstrom and Tirole 1997)
 Riskiness (Bolton and Freixas 2000)
4.2. A simple Model of Credit Market with Moral Hazard
1
 Assumed a firm can promise to repay some fixed amount R (its nominal debt) only
in case of success. The firm has no other source of cash, so the repayment is zero if
the investment fails.
 In the absence of monitoring, the firm will choose the good technology if and only if
this gives it a higher expected profit:

, Since 𝑃𝐺 > 𝑃𝐵 , this equation is equivalent to

𝑅 < 𝑅𝐶 = 2

 Where 𝑅𝐶 denotes the critical value of nominal debt above which the firm chooses
the bad technology.
 From the lender’s viewpoint, the probability P of repayment therefore depends on R:
𝑃𝐺 if R < 𝑅𝐶
𝑃𝑅 = 3
𝑃𝐵 if R> 𝑅𝐶
 In the absence of monitoring, a competitive equilibrium of the credit market
is obtained for R such that 𝑃𝑅 R=1

 This implies R < 𝑅𝐶 , and thus 𝑃𝐺 𝑅𝐶 ≥ 1, which is only satisfied when moral
hazard is not too important.

 If 𝑃𝐺 𝑅𝐶 < 1, the equilibrium involves no trade, and the credit market collapses
because good projects cannot be financed, and bad projects have a negative net
present value (NPV).
 Now a monitoring technology is introduced. At a cost C, specialized firms
interpreted as banks can prevent borrowers from using the bad technology.
 Assuming perfect competition between banks, the nominal value of bank
loans at equilibrium (denoted 𝑅𝑚 , where m stands for monitor) is determined
by the break-even condition .

4
 For bank lending to appear at equilibrium, two conditions are needed:
• The nominal repayment 𝑅𝑚 on bank loans at equilibrium has to be less than the
return G of successful firms. Given above equation , which determines 𝑅𝑚 , this is
equivalent to

𝑃𝐺 G - 1 > C 5

 i.e. the monitoring cost has to be less than the NPV of the good project. Unless
monitoring would be inefficient. Direct lending, which is less costly, has to be
impossible:
𝑃𝐺 𝑅𝐶 - 1 < 1 6
  Therefore, bank lending appears at equilibrium for intermediate values of the probability
1+𝐶 1
𝑃𝐺 𝑃𝐺 ɛ [ , ] 7
𝐺 𝑅𝐶

 provided this interval is not empty. Thus we have established the following
result assuming that the monitoring cost C is small enough so that
1 1+𝐶
>
𝑅𝐶 𝐺 8
 There are three possible regimes of the credit market at equilibrium:
1
1. If 𝑃𝐺 > (high probability of success), firms issue direct debt at a rate
𝑅𝐶
1
𝑅𝑑 = 9
𝑃𝐺
1+𝐶 1
2. If 𝑃𝐺 ɛ[ , ]
𝐺 𝑅𝐶
(intermediate probability of success), firms borrow from banks at a rate
1+𝐶
𝑅𝑏 = 10
𝑃𝐺
1+𝐶
3. if 𝑃𝐺 <
𝐺
(small probability of success), the credit market collapses (no trade equilibrium).
 4.3. Monitoring and Reputation Building Model-Diamond 1991

 This section is adapted from Diamond (1991). Its objective is to show, in a dynamic
extension of the previous model (with two dates, t =0,1), that successful firms can build a
reputation that allows them to issue direct debt instead of using bank loans, which are more
expensive. In order to capture this notion of reputation, assume that firms are
heterogeneous; only some fraction f of them has the choice between the two technologies.
The rest have access only to the bad one, and bank monitoring has no effect on them.
 Under some conditions of the parameters, the equilibrium of the credit market will be such
that
o at t = 0, all firms borrow from banks;
o at t =1, the firms that have been successful at t= 0 issue direct debt while the rest still borrow from
banks;
o banks monitor all the firms who borrow from them.
 This example starts with the case of successful firms. They will be able to issue direct debt
if and only if
 1 11
𝑃𝑠 > ,
𝑅𝐶

 where 𝑃𝑠 is the probability of repayment at date 2, conditionally on success

at date 0 (and given that all firms have been monitored at t = 0).
 Firm live and borrow for two periods
 A fraction f of risk-neutral firms are endowed with the G
technologies, while a fraction 1 - f is endowed with only bad projects
 Bayes’ formula gives the following:

12
 If (11) is satisfied, successful firms will be able to issue direct debt at a rate
𝑅𝑠 =1/ 𝑃𝑠 . On the other hand,
𝑅 < 𝑅𝐶 =
 The notion of reputation building comes from the fact that 𝑃𝑈 < 𝑃0 < 𝑃𝑆 , that is, the
probability of repayment of its debt by the firm is initially 𝑃0 but increases if the firm is
successful 𝑃𝑆 and decreases in the other case 𝑃𝑈 .
 Because of that, the critical level of debt, 𝑅𝐶 0 (above which moral hazard appears)
at t =0 is higher than in the static case.
 Indeed, firms know that if they are successful at t =0, they will obtain cheaper finance (
𝑅𝑆 instead of 𝑅𝑈 ) at date 1. If 𝛿 < 1 denotes the discount factor, the critical level of debt
above which strategic firms choose the bad project at t =0 (denoted by𝑅𝐶 0 ) is now
defined by

𝑃𝐵 [(G-𝑅𝐶 0 ) + δ𝑃𝐺 (G- 𝑅𝑆 )] + (1 - 𝑃𝐵 ) δ 𝑃𝐺 (G- 𝑅𝑈 ) =

𝑃𝐺 [(G-𝑅𝐶 0 ) + δ𝑃𝐺 (G- 𝑅𝑆 )] + (1 - 𝑃𝐺 ) δ 𝑃𝐺 (G- 𝑅𝑈 )
 The left-hand side of this equality is the discounted sum of expected profits of a firm
that chooses the bad technology at t = 0.
 Note that at t =1 it will choose the good technology, either because it can issue direct
debt at rate 𝑅𝑆 after a success or because it borrows from a bank at rate 𝑅𝑈 after
being unsuccessful.
 The right-hand side represents the discounted sum of expected profits for a firm
that chooses the good technology at t =0 (and thus also at t =1). Solving for R, we
obtain
p G−𝑃 B
o 𝑅𝐶 0 = G𝑃 − 𝑃𝐵 +δ𝑃𝐺 (G- 𝑅𝑆 ) - δ 𝑃𝐺 (G- 𝑅𝑈 )
𝐺 𝐵

 which gives, after simplification:

o 𝑅𝐶 0 = 𝑅𝐶 + δ 𝑃𝐺 (𝑅𝑈 - 𝑅𝑆 )
Result : Under the following assumptions,
1 1 1 1+𝐶
o 𝑃0 ≤ 𝑅 0 , 𝑃𝑆 > 𝑅 , 𝑅 > 𝑃𝑈 > ,
𝐶 𝐶 𝐶 𝐺
 The equilibrium of the two-period version of Diamond’s model is characterized as
follows:

1. At t = 0, all firms borrow from banks at a rate

1+𝐶
𝑅0 =
𝑃0
2. At t = 1, successful firms issue direct debt at rate
1
𝑅𝑠 =𝑃
𝑠
Where as the rest borrow from banks at rate
1
𝑅𝑢 =𝑃 , which is higher than 𝑅0
𝑢

 See proof from Freixas-Rochet p-39

 Although this model is very simple, it captures several important features

of credit markets:
o Firms with a good reputation can issue direct debt.
o Unsuccessful firms pay a higher rate than new firms 𝑅𝑈 > 𝑅0 .
o Moral hazard is partially alleviated by reputation effects 𝑅𝑐 0 > 𝑅𝑐
4.3. Capital Availability Model- Holmstrom and Tirole 1997
 If the firm does not have enough capital for issuing direct debt, it can try to borrow
𝐼𝑚 from a bank (in exchange for a return 𝑅𝑚 in case of success), together with a direct
borrowing of 𝐼𝑢 from uninformed investors (in exchange for a return 𝑅𝑢 in case
of success).
See Freixas-Rochet
 4.3. Firm riskiness and dilution costs Model-Bolton and Freixas 2000

 Bolton and Freixas (2000) explore the coexistence of financial markets and financial
intermediaries in a world where borrowers differ in their credit risks.
 Beyond providing a natural framework for analyzing financial intermediation, their
goal is also to understand why equity issuing and bond financing are found
predominantly in mature and relatively safe firms, whereas bank finance (or other
forms of intermediated finance) is the only source of funding for start-up firms and
risky ventures (see Petersen and Rajan (1994) and (1995)).
 Firms choose among three different financial instruments (we assume that firms
cannot combine them).Since the bad firms will mimic the good ones, we only have
to are about the choices of the good firms:
 Bond financing implies a repayment R at date t = 1 (in case of success) and
nothing at t = 2. In case of default at time t =1, the firm is declared bankrupt and
is liquidated.
 Equity issue: a share a ɛ[0 1 ]of the cash flows generated by the firm is sold to the
investors.
 Bank debt implies a repayment Ŕ at t =1, and nothing at t =2. If the firm defaults
at t =1, there is renegotiation and the bank is able to extract the entire surplus at
t =2 because it can observe the probability of success at date 2.
 Each of the three financial instruments has its pros and cons. Equity financing
eliminates inefficient liquidations but generates high dilution costs for good
firms.
 On the contrary, bond financing has lower dilution costs but entails inefficient
bankruptcy costs for good firms.
 Finally, bank loans have the benefits of both because there is efficient
renegotiation in case of default and limited dilution costs, but there is an
intermediation cost.
 For each instrument, it is easy to compute the profits of good firms,
considering that investors require a non-negative return. Bad firms
systematically mimic the choices of good firms to avoid being identified.
 A fraction a of the firm’s capital is sold to outside investors. Because of adverse
selection about the probability of success at t =2, there is a dilution cost.
 Outside shareholders only anticipate an expected cash flow vy at t=2
 By comparing these expected profits for different values of p and v, we derive the
optimal financing choice of the firms (above fig ).
 Equity financing dominates when dilution costs are low, whereas bond financing
dominates when credit risk is low or dilution costs are high.