Credit Booms, Financial Crises and MacroprudentialPolicy
Mark Gertler, Nobuhiro Kiyotaki, Andrea Prestipino
December 2018
What We Do
I We develop a model of banking panics in which:
1. Banking crises are usually preceded by credit booms
2. Credit booms often do not result in crises, i.e. good booms
I We study Macroprudential regulation in this model:
I How does Macroprudential policy weigh the benefits of preventing acrisis against the costs of stopping a good boom?
I What are the effects of macroprudential policy and the features ofoptimal regulation?
I Unintended consequences of regulation; Countercyclical buffers
Banking Crises in the Data (Schularick and Taylor)
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Credit Growth at time t-2 (" from Mean)
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Cre
dit G
row
th a
t tim
e t-
1 ( "
from
Mea
n)
No Crisis at time tCrisis at time t
Framework
I Endowment economy version of GKP (2018)
I Focus on how beliefs driven fluctuations can reproduce key empiricalproperties of banking crises in the data:
I Boom bust cycles in credit
I Unpredictability of crises
I Macroprudential regulation
Model Overview
I Capital is fixed Kt = K = 1 (normalized to unity)
I(Kbt
)intermediated by banks;
(Kht
)directly held by households :
1 = Kht + Kbt
I Households direct finance entails a quadratic deadweight loss
α
2
(Kht
)2
I Resource constraint is:
Yt = Zt −α
2
(Kht
)2= Ct
where Zt is an exogenous productivity shock
Marginal Rates of Return on Capital
Qt ≡ price of capital
I Intermediated capital
Rbt+1 =Zt+1+Qt+1
Qt
I Directly held
Rht+1 =1
1+αKhtQt
Rbt+1
i.e. increasing marginal cost of direct finance
Household and Bank Intermediation
���������
������ ��� ��� ����� � � ��
Qt Kt
Nt
Dt
HOUSEHOLDSCAPITAL
����������� �������������! �"����#�$��%
NO BANK RUN EQUILIBRIUM
Qt Kt
K
BANK RUN EQUILIBRIUM
HOUSEHOLDSQ*t K
CAPITAL
K
h
b
Bankers
I Objective
Vt = EtΛt,t+1[(1− σ)nt+1 + σVt+1]
I Net worth nt accumulated via retained earnings - no new equityissues
nt+1 = Rbt+1Qtk
bt − Rtdt if no run
= 0 if run
I Balance sheetQtk
bt = dt + nt
Deposit Contract
Rt ≡ deposit rate; Rt+1 ≡ return on depositspt ≡ run probability; xt+1 < 1 ≡ recovery rate
I Deposit contract: (One period)
Rt+1 =
{Rt with prob. 1− ptxt+1Rt with prob. pt
Limits to Bank Arbitrage
I Moral Hazard Problem:
I After banker borrows funds at t, it may divert fraction θ of assets forpersonal use.
I If bank does not honor its debt, creditors can
I recover the residual funds andI shut the bank down.
I ⇒ Incentive constraint (IC)
θQtkbt ≤ Vt
Solution
I Can show Vt = ψtnt with ψt ≥ 1 and independent of nt
I Combine with IC → endogenous capital requirement :
κt ≡nt
Qtkbt≥ θψt
I Note:
I ψt countercyclical→ market capital requirements relaxed in bad timesI nt ≤ 0⇒ bank cannot operate (key for run equilbria)
Bank Runs
I Self-fulfilling ”bank run” equilibrium (i.e. rollover crisis) possible if:
I A depositor believes that if other households do not roll over theirdeposits, the depositor will lose money by rolling over.
I Condition met iff banks’ net worth nt goes to zero during a run
I nt = 0 → banks cannot operate
Conditions for Bank Run Equilibrium (BRE)
I Run equilibrium exists at t + 1 if
(Q∗t+1 + Zt+1
)Kbt < Dt R̄t (1)
where Q∗t+1 ≡ is the liquidation price:
Q∗t = Et{Λt,t+1(Zt+1 + Qt+1} − αKht
evaluated at Kht = 1
I ιt+1 ≡ sunpot variable; if ιt+1 = 1 depositors panic when runpossible
I Run occurs if (i) equation (1) is satisfied and (ii) ιt+1 = 1
Run Probability pt
I Assume sunspot occurs with probability κ.I → The time t probability of a run at t + 1 is
pt = Pr t{Zt+1 < ZRt+1} · κ
I ZRt+1 is the threshold value below which a run is possible
Q∗t+1(ZRt+1
)+ ZRt+1 =
Dt R̄t
Kbt
→ Higher leverage ratios Dt R̄tKbt
increase run probability
Run Equilibrium
0 1
No Run-Equilibrium Possible
Run-Equilibrium Possible
Negative Productivity Shock
B
A
Dt R̄t
Kb
Qt∗+1
(ZRt+1
)+ ZRt+1
Run Equilibrium
0 1
No Run-Equilibrium Possible
Run-Equilibrium Possible
Higher Leverage Ratio
BA
Dt R̄t
Kb
Qt∗+1
(ZRt+1
)+ ZRt+1
Run After a Negative 2 std Shock
0 10 20 30 40 50 60-2.5
-2
-1.5
-1
-0.5
0
% "
from
SS
Productivity
0 10 20 30 40 50 600
1
2
3
4
5
6
Leve
l (pc
t)
Run Probability
0 10 20 30 40 50 60-100
-80
-60
-40
-20
0
% "
from
SS
Bank Net Worth
0 10 20 30 40 50 60
Quarters
-100
-80
-60
-40
-20
0
% "
from
SS
Bank Intermadiation
0 10 20 30 40 50 60
Quarters
200
220
240
260
280
300
Leve
l Ann
ual B
asis
Poi
nts
Excess Return: ER b-Rfree (10 years)
0 10 20 30 40 50 60
Quarters
-8
-6
-4
-2
0
% "
from
SS
Output
Sunspot No Sunspot
Run after a sequence of bad shocks
Boom leading to the bust: news driven optimism
I Productivity:
Zt+1 = ρZt + �t+1
I Normally, E{�t+1} = 0
I Occasionally, bankers receive news about future productivity
I If news at t, bankers learn that unusually large realization �tB of size
B > 0 will happen at tB ∈ {t + 1, ..., t + T} with prob. PB0 < 1
I Pr t{tB = t + i} is a truncated Normal (discrete approx.)
I Agents update Pr t+i and PBt+i by observing �t+i
I Prob. at t + i of shock at t + i + 1 is Pr t{tB = t + i + 1} · PBt+i
Beliefs Driven Credit Boom
t=1
t=5.5
t=11
0.15
0.3
Prior cond. prob. of shock happening at time t
t=1
t=5.5
t=11
0
0.2
0.4
0.6
0.8
1Beliefs Evolution
PBt
Pr tftb = t + 1g
t=1
t=5.5
t=11
0
0.5
1
1.5
2
% "
from
SS
Productivity
Zt
EtZ
t+4
t=1
t=5.5
t=11
-0.5
0
0.5
1
% "
from
SS
Output
t=1
t=5.5
t=11
0
5
10
15
20
% "
from
SS
Credit: Q " Kb
t=1
t=5.5
t=11
0.1
0.2
0.3
Leve
l
Prob. of being in crisis zone: Pr { Zt+1
Boom Leading to a bust
0 10 20 30 40 50 60-3
-2
-1
0
1
2
% "
from
SS
Expected Productivity
0 10 20 30 40 50 60-3
-2
-1
0
1
2
% "
from
SS
Realized Productivity
Zt
ZRt+1
0 10 20 30 40 50 60-100
-50
0
50
% "
from
SS
Bank Intermediation
0 10 20 30 40 50 600
5
10
Leve
l (pc
t)
Run Probability (if no boom)
0 10 20 30 40 50 60-100
-50
0
50 %
" fr
om S
S
Bank Net Worth
0 10 20 30 40 50 60-10
-5
0
5
% "
from
SS
Output
0 10 20 30 40 50 60
Quarters
-100
-50
0
50
% "
from
SS
Capital Ratio: 5
0 10 20 30 40 50 60
Quarters
-15
-10
-5
0
5
% "
from
SS
Asset Price
0 10 20 30 40 50 60
Quarters
150
200
250
300
Leve
l Ann
ual B
asis
Poi
nts Excess Return: ER
b-Rfree (10 yrs)
Sunspot observed No Sunspot observed
Run After Credit Boom
False Alarms
0 10 20 30 40 50 60-3
-2
-1
0
1
2
% "
from
SS
Expected Productivity
0 10 20 30 40 50 60-3
-2
-1
0
1
2
% "
from
SS
Realized Productivity
Zt
ZRt+1
0 10 20 30 40 50 600
5
10
15
20
% "
from
SS
Bank Intermediation
0 10 20 30 40 50 600
5
10
Leve
l (pc
t)
Run Probability (if no boom)
0 10 20 30 40 50 60-10
0
10
20
30 %
" fr
om S
S
Bank Net Worth
0 10 20 30 40 50 600
0.5
1
% "
from
SS
Output
0 10 20 30 40 50 60
Quarters
-5
0
5
10
% "
from
SS
Capital Ratio: 5
0 10 20 30 40 50 60
Quarters
-1
0
1
2
3
% "
from
SS
Asset Price
0 10 20 30 40 50 60
Quarters
190
200
210
220
Leve
l Ann
ual B
asis
Poi
nts Excess Return: ER
b-Rfree (10 yrs)
Boom Happens No Sunspot is Observed
False Alarms
Unpredictability of Crises: Data and Model
-1.5 -1 -0.5 0 0.5 1 1.5
Log " of Credit from trend at t-2
-1.5
-1
-0.5
0
0.5
1
1.5
Log "
of C
redi
t fro
m tr
end
at t-
1
No Crisis at tCrisis at time t
-1.5 -1 -0.5 0 0.5 1 1.5
% " of Credit from mean at year t-2
-1.5
-1
-0.5
0
0.5
1
1.5
% "
of C
redi
t fro
m m
ean
at y
ear
t-1
No Crisis at tCrisis at time t
Regulation
I Macroprudential regulator sets time varying capital requirement κ̄t
I Equilibrium capital ratios are
κt = max {κ̄t , κmt }
where κmt =θψt
are the market imposed capital ratios
I We restrict policy to be deteremined by simple rule
κ̄t =
{κ̄ if Nt ≥ N̄0 if Nt < N̄
I We look for(κ̄, N̄
)that maximize welfare
Regulation
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
Bank Net Worth
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
Capital Ratio
Unregulated Equilbrium
Regulation
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
Bank Net Worth
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
Capital Ratio
Unregulated EquilbriumCapital Requirement
Regulation
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
Bank Net Worth
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
Capital Ratio
Unregulated EquilbriumCapital RequirementRegulated Equilibrium
Avoiding a Run with Regulation
0 10 20 30 40 50 60-3
-2
-1
0
1
2
% "
from
SS
Expected Productivity
0 10 20 30 40 50 60-3
-2
-1
0
1
2
% "
from
SS
Realized Productivity
Zt
ZRt+1
(Unregulated)
0 10 20 30 40 50 60-100
-80
-60
-40
-20
0
20
% "
from
SS
Capital Ratio: 5
0 10 20 30 40 50 60
Quarters
-100
-80
-60
-40
-20
0
20
% "
from
SS
Bank Intermediation
0 10 20 30 40 50 60
Quarters
0
2
4
6
8
Leve
l (pc
t)
Run Probability
0 10 20 30 40 50 60
Quarters
-8
-6
-4
-2
0
2
% "
from
SS
Output
Regulated Unregulated
Avoiding Runs with Macro Pru
Responding to False Alarms: No Sunspot Observed
0 10 20 30 40 50 60-3
-2
-1
0
1
2
% "
from
SS
Expected Productivity
0 10 20 30 40 50 60-3
-2
-1
0
1
2
% "
from
SS
Realized Productivity
Zt
ZRt+1
(Unregulated)
0 10 20 30 40 50 60-100
-80
-60
-40
-20
0
20
% "
from
SS
Capital Ratio: 5
0 10 20 30 40 50 60
Quarters
-100
-80
-60
-40
-20
0
20
% "
from
SS
Bank Intermediation
0 10 20 30 40 50 60
Quarters
0
2
4
6
8
Leve
l (pc
t)
Run Probability
0 10 20 30 40 50 60
Quarters
-8
-6
-4
-2
0
2
% "
from
SS
Output
Regulated Unregulated
Run After Credit Boom
Effect of Regulation
Unregulated Economy��� = 0; �N = 0
� Optimal Regulation��� = :13; �N = :85 �NDESS
� Fixed Capital Requirements��� = :13; �N = 0
�Run Frequency :8 pct :45 pct :3 pct
AVG Output Cond. No Run(� from Decentralized Economy)
0 �:4 pct �1:7 pct
AVG Output(� from Decentralized Economy)
0 :1 pct �:9 pct
Welfare Gain(� Permanent Consumption)
0 :16 pct �1:16 pct
1
Recovery From a Run
0 10 20 30 40 50 60
Quarters
-16
-14
-12
-10
-8
-6
-4
-2
0
% "
from
SS
Asset Price
0 10 20 30 40 50 60
Quarters
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
% "
from
SS
Net Worth
0 10 20 30 40 50 60
Quarters
-7
-6
-5
-4
-3
-2
-1
0
% "
from
SS
Output
Regulated Fixed Unregulated Regulated Countercyclical
Recovery from a run: Forgiveness VS No Forgiveness
Conclusion
I Develop model of banking panics that captures boom-bust cyclesand unpredictability of runs
I Study macroprudential policy
I Future work
I Ex-post intervention
I Regulated and Unregulated Banks
I Multiple assets