Capital Regulation

Financial Intermediation

Sergio Vicente
0. Why do banks want high leverage?
Tax bene…ts

I Tax savings because interest on deposits are generally

tax-deductible whereas dividends to shareholders are not
Managerial preference for ROE

I Executive compensation based on ROE

I Leverage increases ROE, albeit at the expense of reducing

the market equity value because of increased risk

I Deposits/debt are cheaper (deposit insurance not

well-priced/implicit or explicit bail-out guarantees) and
therefore leverage increases the return on equity
D
rE = rL + (rE rD ) (1 τC ) > rL
E
Government safety nets

I Safety nets such as deposit insurance (or implicit bailouts)

subsidize deposits (or debt, in general)

I Safety nets prevents the appropriate pricing of risk entailed

by increasing leverage:

I If deposits were not insured, depositors would require a

higher return to compensate them for the risk of losing
deposits

I If there were not implicit guarantees, debt investors (those

that lend to banks) would require higher rates of return to
compensate them for the risk of not repayment

I Proper pricing would reduce the incentive to increase

leverage
Remark: Which value increases?

I Tax bene…ts increase the value of the bank

I The bank pay less taxes, which would otherwise go to the
Government

I Government safety nets increase the value of the bank

I The bank gets an implicit subsidy in the form of cheaper
deposits and debt that in the absence of guarantees
I The Government bears the cost by supporting the guarantees

I Leverage increases the value of bank’s equity, not the value

of the bank
I Leverage increases the value of equity at the expense of the
depositors, who get rewarded through a low rate (rate on
deposits) even though the rate of return on loans is higher
1. The case for capital regulation
1.1 Reducing risk: Skin in the game and risk-shifting
Two projects

I Consider the following two projects:

I Safe: Invest $100 and obtain $140 for certain

I Risky: Invest $100 and obtain $300 with probability 0.25 and
$0 otherwise
Expected value of projects

I What is the expected value of each project?

I Safe: 140 100 = $40

I Risky: 0.25 300 100 = $25

Pro…t of projects when …nanced with deposits

I Suppose that the bank is funded with deposits in its entirety,

rewarded at rD

I What is the pro…t from each project?

I Safe:

140 100 (1 + rD )
= 40 100 rD

I Risky:

0.25 [300 100 (1 + rD )]

= 50 25 rD
Observations

I Risky is more pro…table (75% of the times the downside is

paid by depositors/deposit insurance)

I The higher the interest rate on deposits, the lower the bank’s
pro…t (self-explanatory!)

I The e¤ect of interest rates on pro…ts is more accentuated for

the safe project (interest rates are not paid 75% of the times in
the case of the risky project)
Pro…t of projects with a funding mix of deposits and
capital
I Suppose that the bank is funded with deposits rewarded at rD
and (external) capital, rewarded at rE rD

I What is the pro…t from each project?

I Safe:
140 (100 k ) (1 + rD ) k (1 + rE )
= 40 100 rD k (rE rD )

I Risky:
0.25 [300 (100 k) (1 + rD )] k (1 + rE )
2 3
6 7
= 50 25 rD k 4 (rE rD ) + 0.75 (1 + rD )5
| {z } | {z }
Costlier capital Downside
Observation I

I Capital erodes pro…ts (because 75% of the times a share k

the downside is paid by equityholders–also, rE > rD leads to
an additional loss, but observe that even if rE = rD , capital
always erodes pro…ts as long as the failure event has positive
probability)

I Banks would be willing to set capital to zero, that is k = 0

Observation II

I Capital favors the safe project: The factor

0.75 (1 + rD )–the downside e¤ect– is only paid in the risky
project

I Safe project is preferable for

13.33 + 100rD
k k
1 + rD

I Remark: The higher rD , the higher the capital requirement to

induce safety (the reward on deposits is most often paid for
the safe project, making the risky project relatively more
interesting
Which capital requirement?
I There is a social cost of capital requirements (imposing
capital on banks reduces the extent of the main banking
function: the (qualitative asset) transformation of illiquid
loans into liquid deposits)

I Each additional unit of capital entails a (shadow, social) cost ρ

I Clearly, the optimal capital requirement is either k = 0 or

13.33 +100rD
k=k 1 +r D

I A capital requirement larger than k is wasteful because capital

is costly and the bank chooses the safe project for any such
capital requirements

I A capital requirement in between 0 and k is wasteful because

capital is costly and the bank chooses the risky project for any
such capital requirements
Social welfare function

I What is the social welfare from each project?

I Safe:
40
|{z} k ρ
| {z }
NPV Social cost of capital

I The rest of the elements in the bank’s pro…t function is

transfers among agents (as long as the Social Planner regards
equityholders and depositors equally, as we usually assume)

I Risky:
25 k ρ
|{z} | {z }
NPV Social cost of capital
Social welfare function
I What is the social welfare from each capital requirement?

I
13.33 +100rD
k= 1 +r D (induces the safe project, at a social cost):

13.33 + 100rD
40 ρ
1 + rD

I k = 0 (induces the risky project):

25

I As long as
65 (1 + rD )
ρ
13.33 + 100rD
a positive capital requirement improves welfare
Social cost of capital?

I In practice, hard to tell what the social cost of capital is

I Under which circumstances are deposits–and, more generally,

the banking function–valuable?

I Under which circumstances capital may be socially more

costly? (capital scarcity in recessions implies that capital is
more costly than in expansions?)

I These are open questions, but the bottom line is that capital
requirements favors banks’safety but it entails a social
cost

I Capital regulation should deal with this tradeo¤

"appropriately"
Risk-shifting generally

I More generally, banks choose their risk level facing a

risk-return tradeo¤

I The bank would typically choose an ine¢ ciently high level

of risk, as compared to the social optimum

I This is due to the fact that the bank does not internalize
the losses associated with a failure: limited liability bounds
the losses to 0
Role for regulation

I Capital requirements improve social welfare because of an

ine¢ ciency: the bank does not internalize the cost
in‡icted on society when failing

I Limited liability and (inappropriately priced, because of

deposit insurance) debt-type funding induce this ine¢ ciency

I By reducing the wedge between the bank’s private

bene…t and society’s goals –hence an alignment of private
and social preferences-, capital requirements improve upon
social welfare
1.2 Preventing bank failures: Capital as a bu¤er
Capital as a bu¤er
I Consider a bank with:

I Stochastic returns with distribution:

R F Rlow , Rhigh

I Debt with face value D

I The bank becomes insolvent when R < D, that is, with

probability F (D )

I If the bank buys-back debt with equity, so that new debt

becomes D 0 = D K , now the bank becomes insolvent when
R < D 0 , that is, with probability F (D 0 ) < F (D )

I By reducing debt to D 0 , the bank can now (additionally!)

survive the events in which its returns R fall in the interval
[D 0 , D ], hence increasing the probability of bank’s survival
Capital as a bu¤er: An example
I Consider a bank with:

R Unif [0, 300]

I Bank raises k and 100 k in capital and in debt/deposits,

respectively

I Bank defaults if R < (100 k) (1 + rD ) D̂ (k )

I The probability of insolvency is given by:

D̂ (k ) (100 k ) (1 + rD )
PI (k ) Pr R < D̂ (k ) = =
300 300

I The probability of insolvency is decreasing in k (for rD = 0.05:

PI (0) = 0.35 , PI (20) = 0.28, PI (100) = 0)
The bu¤er case for capital requirements

I Capital reduces the probability of insolvency, but simply

by changing the nature of funding to the bank (from debt
to equity)

I Essentially, there is no risk reduction, but simply a change in

the type of investor that su¤er the losses (from debt-holders
–i.e. depositors- to equity-holders –i.e. the bank’s owners)

I Which is the gain from simply changing who su¤ers the loss?

I By reducing the likelihood of insolvency, there is a gain if

insolvency per se is socially costly because of panics,
contagion, side e¤ects,. . .

I Remind that capital is costly!

Why are defaults socially costly?
I Disruption costs (bail-ins)

I Investors may run on their investments (contagion)

I Investors may not invest at all (credit crunch, freeze,. . . )

I Tax-payers money (bail-outs)

I Shadow cost of public funds (less public investment, less

redistribution, distortionary taxes,. . . )

I Redistribution of bank losses to society at large

I Remark: A bail-out is a transfer which, as such, does not carry
any social cost per se for a utilitarian social welfare maximizer;
however, we may think that it is costly to transfer resources
from the average citizen to bank’s equityholders on equality
grounds
Are equityholders ready to absorb losses?

I Equityholders are investors whose return depends on the

bank earnings: If earnings are high, their return is high; if
pro…ts are low, their returns are low

I Bank insolvency is a state that equityholders (must!)

consider when investing, in which case their equity value gets
wiped out

I Unlike debtholders (especially depositors), who expect to get

their promised return back, equityholders are ready to
asume this loss
1.3 Reducing debt-overhang
Debt overhang

I Debt overhang: A …rm may pass on a positive NPV

investment instead of issuing equity to fund it because by
funding it may reduce the value of equity (since a large
share of the value created accrues to debtholders)

I This is typically the case when the …rm has a large share of
debt

I Food for thought: How leveraged are …nancial institutions?

An example of debt-overhang
The problem

I $70 owed to bondholders

I Payo¤ distribution: $100 w.p 0.5 and $0 otherwise

I The expected value of this gamble is $50

I Opportunity invest $20 to turn the project payo¤ into $75 w.p
1

I Should the …rm do it?

To invest or not to invest?

I A positive NPV investment:

$75 $50 $20 = $5

|{z} |{z} |{z}
New project value Old project (expected) value Cost

I Equity value from not investing:

0.5 (100 70) = $15

I Equity value from investing:

75 70 20 = $15
Why is a positive NPV opportunity missed?

I Most of the value of this (positive NPV) investment accrues

to the debtholder so that the …rm does not have an incentive
to invest
Aside: renegotiating debt

I Should debtholders accept a (50%) haircut to reduce their

debt claim from $70 to $35?

I Value of debt from no renegotiation:

0.5 70 = $35
Aside: renegotiating debt

I Value of debt from renegotiation:

I If the …rm does not invest, it would now get:

0.5 (100 35) = $32.5

I If the …rm invests, it would now get:

75 20 35 = $20

I Hence, the value of debt from renegotiation is smaller:

0.5 35 = $17.5 < $35

What about renegotiation with conditional aid to invest?
I Consider the following renegotiation package: Debtholders
provide $20 to invest and reduce their debt claim from $70 to
$55 (if you want, think of $55.0001)
I Notice that debtholders would then collect $55 for certain,
therefore getting–after subtracting the investmentof $20–a
return of $35

I Equity value from not investing:

0.5 (100 70) = $15

I Equity value from investing now:

75 55 = $20 > $15

I By letting the …rm cash the NPV of the project, the

investment can be undertaken