Financial Regulation in a Quantitative Model of the Modern Banking System Juliane Begenau 1 Tim Landvoigt 2 1 Stanford & NBER 2 Wharton & NBER Indiana University November 7, 2018 Begenau & Landvoigt Financial Regulation 1 / 33
Financial Regulation in a Quantitative Model of theModern Banking System
Juliane Begenau1 Tim Landvoigt2
1Stanford & NBER
2Wharton & NBER
Indiana UniversityNovember 7, 2018
Begenau & Landvoigt Financial Regulation 1 / 33
Motivation
I Financial System:regulated (commercial) banks & unregulated (shadow) banks
I provide access to “intermediated” assets, e.g. long term creditI balance sheet: risky & illiquid assets funded with money-like liabilities
I Effects of financial regulation on a subset of banks?I Does tighter regulation cause shift to shadow banks?I Does it make financial system more risky?
I Requires quantitative general equilibrium analysis
I Study effect of capital requirements
Begenau & Landvoigt Financial Regulation 2 / 33
This PaperI Model
I comm. banks and shadow banks provide liquidity servicesI both have limited liability & costly bankruptciesI comm. banks: deposit insurance, subject to capital regulationI shadow banks: risky debt, no regulationI focus on (1) risk taking through bank leverage
(2) liquidity provision by banks
I Calibration matchesI aggregate liquidity premium of safe debtI size of shadow banking sectorI default risk of both types of banksI greater fragility of shadow banks (runs)
I Tighter capital requirementI causes shift to shadow sectorI only small increase in risk taking (leverage) by shadow banksI trade-off between financial fragility and liquidity provision
Begenau & Landvoigt Financial Regulation 3 / 33
Model Overview
Y Asset
Own Funds
Equity
Equity
Commercial Banks
Shadow banks
Intermediaries
Capital
Capital
Deposits
Debt
Capital
produced by
banks
C. Equity
C. Deposits
S. Equity
S. Debt
Households
Y Asset
(not intermediated)
Deposit Insurance
𝜚 deposits
withdrawn
early &
bailout
probability
Begenau & Landvoigt Financial Regulation 4 / 33
Overview of Talk
I Static ModelI What pins down size and leverage of shadow banks?I Effect of tighter capital requirementI Efficient allocation vs. equilibrium
I Dynamic quantitative modelI Differences to simple modelI Calibration highlightsI Quantitative results
Begenau & Landvoigt Financial Regulation 5 / 33
Setup
I Dates t = 0 and t = 1I Unit mass of households endowed with 1 unit of capital at t = 0I Unit mass of C-banks and S-bank purchase capital
and issue equity and debt to households
I Capital produces 1 unit of consumption at t = 1 if held by banksI Capital much less productive if held by households
I Household preferences: bank deposits provide liquidity services
U = C0 + β(C1 + ψH(AS ,AC ))
with Aj , j = S ,C , are deposits of banks held by households
Begenau & Landvoigt Financial Regulation 6 / 33
Setup
I Dates t = 0 and t = 1I Unit mass of households endowed with 1 unit of capital at t = 0I Unit mass of C-banks and S-bank purchase capital
and issue equity and debt to householdsI Capital produces 1 unit of consumption at t = 1 if held by banksI Capital much less productive if held by households
I Household preferences: bank deposits provide liquidity services
U = C0 + β(C1 + ψH(AS ,AC ))
with Aj , j = S ,C , are deposits of banks held by households
Begenau & Landvoigt Financial Regulation 6 / 33
Setup
I Dates t = 0 and t = 1I Unit mass of households endowed with 1 unit of capital at t = 0I Unit mass of C-banks and S-bank purchase capital
and issue equity and debt to householdsI Capital produces 1 unit of consumption at t = 1 if held by banksI Capital much less productive if held by households
I Household preferences: bank deposits provide liquidity services
U = C0 + β(C1 + ψH(AS ,AC ))
with Aj , j = S ,C , are deposits of banks held by households
Begenau & Landvoigt Financial Regulation 6 / 33
S-banks
I Each bank solves
maxKS≥0,BS≥0
qS(BS ,KS)BS − pKS︸ ︷︷ ︸equity raised at t = 0
+βmax ρSKS − BS , 0︸ ︷︷ ︸dividend paid at t = 1
I Bank issues risky debt at price qS(BS ,KS)I Creditors price default riskI Bank internalizes effect of choice (BS ,KS) on debt price
I Limited liability with costly bankruptcy: if default, equity is wiped outand all assets lost (no recovery for creditors)
I Bank-idiosyncratic payoff shock ρS ∼ Uniform[0, 1]
Begenau & Landvoigt Financial Regulation 7 / 33
S-banks
I Each bank solves
maxKS≥0,BS≥0
qS(BS ,KS)BS − pKS︸ ︷︷ ︸equity raised at t = 0
+βmax ρSKS − BS , 0︸ ︷︷ ︸dividend paid at t = 1
I Bank issues risky debt at price qS(BS ,KS)I Creditors price default riskI Bank internalizes effect of choice (BS ,KS) on debt price
I Limited liability with costly bankruptcy: if default, equity is wiped outand all assets lost (no recovery for creditors)
I Bank-idiosyncratic payoff shock ρS ∼ Uniform[0, 1]
Begenau & Landvoigt Financial Regulation 7 / 33
S-banks
I Each bank solves
maxKS≥0,BS≥0
qS(BS ,KS)BS − pKS︸ ︷︷ ︸equity raised at t = 0
+βmax ρSKS − BS , 0︸ ︷︷ ︸dividend paid at t = 1
I Bank issues risky debt at price qS(BS ,KS)I Creditors price default riskI Bank internalizes effect of choice (BS ,KS) on debt price
I Limited liability with costly bankruptcy: if default, equity is wiped outand all assets lost (no recovery for creditors)
I Bank-idiosyncratic payoff shock ρS ∼ Uniform[0, 1]
Begenau & Landvoigt Financial Regulation 7 / 33
S-banks
I Each bank solves
maxKS≥0,BS≥0
qS(BS ,KS)BS − pKS︸ ︷︷ ︸equity raised at t = 0
+βmax ρSKS − BS , 0︸ ︷︷ ︸dividend paid at t = 1
I Bank issues risky debt at price qS(BS ,KS)I Creditors price default riskI Bank internalizes effect of choice (BS ,KS) on debt price
I Limited liability with costly bankruptcy: if default, equity is wiped outand all assets lost (no recovery for creditors)
I Bank-idiosyncratic payoff shock ρS ∼ Uniform[0, 1]
Begenau & Landvoigt Financial Regulation 7 / 33
S-Bank Problem
I Write time-1 dividend as
max ρSKS − BS , 0 = KS (1− FS(LS)) (E(ρS |ρS > LS)− LS)
with bank leverage LS = BS/KS and FS() is c.d.f. of ρS
I Using ρS ∼ Uniform[0, 1], FS(LS) = LS is default probability, andcontinuation value further simplifies
1
2KS (1− LS)2
I Using scale independence
vS = maxLS∈[0,1]
qS(LS)LS − p + β1
2(1− LS)2
maxKS≥0
KSvS
Begenau & Landvoigt Financial Regulation 8 / 33
S-Bank Problem
I Write time-1 dividend as
max ρSKS − BS , 0 = KS (1− FS(LS)) (E(ρS |ρS > LS)− LS)
with bank leverage LS = BS/KS and FS() is c.d.f. of ρS
I Using ρS ∼ Uniform[0, 1], FS(LS) = LS is default probability, andcontinuation value further simplifies
1
2KS (1− LS)2
I Using scale independence
vS = maxLS∈[0,1]
qS(LS)LS − p + β1
2(1− LS)2
maxKS≥0
KSvS
Begenau & Landvoigt Financial Regulation 8 / 33
S-Bank Problem
I Write time-1 dividend as
max ρSKS − BS , 0 = KS (1− FS(LS)) (E(ρS |ρS > LS)− LS)
with bank leverage LS = BS/KS and FS() is c.d.f. of ρS
I Using ρS ∼ Uniform[0, 1], FS(LS) = LS is default probability, andcontinuation value further simplifies
1
2KS (1− LS)2
I Using scale independence
vS = maxLS∈[0,1]
qS(LS)LS − p + β1
2(1− LS)2
maxKS≥0
KSvS
Begenau & Landvoigt Financial Regulation 8 / 33
C-banks
I Each bank solves
maxKC≥0,BC≥0
qCBC − pKC︸ ︷︷ ︸equity raised at t = 0
+βmax ρCKC − BC , 0︸ ︷︷ ︸dividend paid at t = 1
subject toBC ≤ (1− θ)E(ρC )KC
I Differences to S-bank problemI Government-insured debt is riskfree to creditorsI Regulatory capital requirement θ
Begenau & Landvoigt Financial Regulation 9 / 33
C-banks
I Each bank solves
maxKC≥0,BC≥0
qCBC − pKC︸ ︷︷ ︸equity raised at t = 0
+βmax ρCKC − BC , 0︸ ︷︷ ︸dividend paid at t = 1
subject toBC ≤ (1− θ)E(ρC )KC
I Differences to S-bank problemI Government-insured debt is riskfree to creditorsI Regulatory capital requirement θ
Begenau & Landvoigt Financial Regulation 9 / 33
Households and Government
I HH choose purchases of debt and equity of each bank to max utility
maxAj ,Sjj=S,C
C0 + β(C1 + ψH(AS ,AC ))
s.t. C0 = p︸︷︷︸sell cap.
−qSAS − qCAC − pSSS − pCSC︸ ︷︷ ︸buy securities
C1 = (1− LS)AS + AC − T
+ SS1
2KS (1− LS)2︸ ︷︷ ︸
div. from S-bank
+SC1
2KC (1− LC )2︸ ︷︷ ︸
div. from C-bank
whereT = LCBC
are lump-sum taxes required to bail out liabilities of failing C-banks
Begenau & Landvoigt Financial Regulation 10 / 33
Equilibrium
I Market clearing
SS = 1
SC = 1
AC = BC
AS = BS
KS + KC = 1.
I Resource constraints: C0 = 0 and
C1 =1
2
(1− KCL
2C − KSL
2S
)I Time-1 consumption clarifies fundamental trade-off
I Bank leverage causes bankruptcies and deadweight lossesI But some leverage necessary to produce liquidity services
Begenau & Landvoigt Financial Regulation 11 / 33
Planner Problem (1/2)
I Planner solves
maxKS ,LS ,LC
1
2
(1− (1− KS)L2
C − KSL2S
)+ ψH
LSKS︸ ︷︷ ︸=AS
, LC (1− KS)︸ ︷︷ ︸=AC
I Liquidity preference function
H(AS ,AC ) = (αAεS + (1− α)AεC )1/ε
I Marginal liquidity benefit (= “convenience yield”)
∂H(AS ,AC )
∂Aj= Hj(RS), where RS = AS/AC
I Solution isKS
KC=
(α
1− α
) 11−ε
, and LS = LC
with LS = ψHS(RS) and LC = ψHC (RS)
Begenau & Landvoigt Financial Regulation 12 / 33
Planner Problem (1/2)
I Planner solves
maxKS ,LS ,LC
1
2
(1− (1− KS)L2
C − KSL2S
)+ ψH
LSKS︸ ︷︷ ︸=AS
, LC (1− KS)︸ ︷︷ ︸=AC
I Liquidity preference function
H(AS ,AC ) = (αAεS + (1− α)AεC )1/ε
I Marginal liquidity benefit (= “convenience yield”)
∂H(AS ,AC )
∂Aj= Hj(RS), where RS = AS/AC
I Solution isKS
KC=
(α
1− α
) 11−ε
, and LS = LC
with LS = ψHS(RS) and LC = ψHC (RS)
Begenau & Landvoigt Financial Regulation 12 / 33
Planner Problem (1/2)
I Planner solves
maxKS ,LS ,LC
1
2
(1− (1− KS)L2
C − KSL2S
)+ ψH
LSKS︸ ︷︷ ︸=AS
, LC (1− KS)︸ ︷︷ ︸=AC
I Liquidity preference function
H(AS ,AC ) = (αAεS + (1− α)AεC )1/ε
I Marginal liquidity benefit (= “convenience yield”)
∂H(AS ,AC )
∂Aj= Hj(RS), where RS = AS/AC
I Solution isKS
KC=
(α
1− α
) 11−ε
, and LS = LC
with LS = ψHS(RS) and LC = ψHC (RS)
Begenau & Landvoigt Financial Regulation 12 / 33
Planner Problem (2/2)
I How does planner set leverage of each bank?
Lj = ψHj(AS/AC )
I RHS is marginal benefit to household from liquidityI LHS is marginal cost from failing banks (consumption losses)I LS = LC means that same marginal benefit from both types
I How does planner allocate capital?
KS
KC=
(α
1− α
) 11−ε
I Both banks have same production technology for consumption goodI So capital allocation only based on liquidity production technologyI Greater elasticity 1/(1− ε) tilts allocation towards bank with greater
weight in CES aggregator
Begenau & Landvoigt Financial Regulation 13 / 33
Planner Problem (2/2)
I How does planner set leverage of each bank?
Lj = ψHj(AS/AC )
I RHS is marginal benefit to household from liquidityI LHS is marginal cost from failing banks (consumption losses)I LS = LC means that same marginal benefit from both types
I How does planner allocate capital?
KS
KC=
(α
1− α
) 11−ε
I Both banks have same production technology for consumption goodI So capital allocation only based on liquidity production technologyI Greater elasticity 1/(1− ε) tilts allocation towards bank with greater
weight in CES aggregator
Begenau & Landvoigt Financial Regulation 13 / 33
Equilibrium Characterization: S-Bank Problem
I S-bank FOC for leverage
q′S(LS)LS + q(LS) = β(1− LS)
I Household FOC for S-bank debt
q(LS) = β(1− LS︸ ︷︷ ︸payoff
+ ψHS(RS)︸ ︷︷ ︸liq. premium
)
I Combining givesLS = ψHS(RS)
⇒ S-bank leverage choice is same as planner solution!
I Constant returns also imply zero expected profits
vS = 0⇔ p − qS(LS)LS = β1
2(1− LS)2
Begenau & Landvoigt Financial Regulation 14 / 33
Equilibrium Characterization: C-Bank ProblemI Simplify problem as for S-bank
vC = maxLC∈[0,1]
qCLC − p + β1
2(1− LC )2
subject to
LC ≤1
2(1− θ) ,
and maxKC≥0
KCvC
I Household FOC for C-bank debt
qC = β( 1︸︷︷︸payoff
+ ψHC (RS)︸ ︷︷ ︸liq. premium
)
I C-bank constraint always binds if ψ > 0: LC = 12 (1− θ)
⇒ Moral hazard due to limited liability and deposit insurance
I Constant returns also imply zero expected profits
vC = 0⇔ p − qCLC = β1
2(1− LC )2
Begenau & Landvoigt Financial Regulation 15 / 33
Equilibrium Characterization: C-Bank ProblemI Simplify problem as for S-bank
vC = maxLC∈[0,1]
qCLC − p + β1
2(1− LC )2
subject to
LC ≤1
2(1− θ) ,
and maxKC≥0
KCvC
I Household FOC for C-bank debt
qC = β( 1︸︷︷︸payoff
+ ψHC (RS)︸ ︷︷ ︸liq. premium
)
I C-bank constraint always binds if ψ > 0: LC = 12 (1− θ)
⇒ Moral hazard due to limited liability and deposit insurance
I Constant returns also imply zero expected profits
vC = 0⇔ p − qCLC = β1
2(1− LC )2
Begenau & Landvoigt Financial Regulation 15 / 33
Equilibrium Characterization: C-Bank ProblemI Simplify problem as for S-bank
vC = maxLC∈[0,1]
qCLC − p + β1
2(1− LC )2
subject to
LC ≤1
2(1− θ) ,
and maxKC≥0
KCvC
I Household FOC for C-bank debt
qC = β( 1︸︷︷︸payoff
+ ψHC (RS)︸ ︷︷ ︸liq. premium
)
I C-bank constraint always binds if ψ > 0: LC = 12 (1− θ)
⇒ Moral hazard due to limited liability and deposit insurance
I Constant returns also imply zero expected profits
vC = 0⇔ p − qCLC = β1
2(1− LC )2
Begenau & Landvoigt Financial Regulation 15 / 33
Equilibrium Characterization: Relative Size of SectorsI Rewrite zero-profit conditions
S-bank: p = β1
2(1 + L2
S)
C-bank: p = β1
2(1 + L2
C ) + βLCψHC (RS)
I Combining both conditions to get indifference condition
LS = (L2C + LCψHC (RS))1/2 (L1)
I Condition implies that S-bank leverage is always higherthan C-bank leverage
I C-bank has key competitive advantage: deposit insuranceI To deliver same profit to equity owners, S-banks need to have higher
leverage (holding constant S-bank default risk)
I Also recall optimal S-bank leverage
LS = ψHS(RS) (L2)
Begenau & Landvoigt Financial Regulation 16 / 33
Equilibrium Characterization: Relative Size of SectorsI Rewrite zero-profit conditions
S-bank: p = β1
2(1 + L2
S)
C-bank: p = β1
2(1 + L2
C ) + βLCψHC (RS)
I Combining both conditions to get indifference condition
LS = (L2C + LCψHC (RS))1/2 (L1)
I Condition implies that S-bank leverage is always higherthan C-bank leverage
I C-bank has key competitive advantage: deposit insuranceI To deliver same profit to equity owners, S-banks need to have higher
leverage (holding constant S-bank default risk)
I Also recall optimal S-bank leverage
LS = ψHS(RS) (L2)
Begenau & Landvoigt Financial Regulation 16 / 33
Equilibrium Characterization: Relative Size of SectorsI Rewrite zero-profit conditions
S-bank: p = β1
2(1 + L2
S)
C-bank: p = β1
2(1 + L2
C ) + βLCψHC (RS)
I Combining both conditions to get indifference condition
LS = (L2C + LCψHC (RS))1/2 (L1)
I Condition implies that S-bank leverage is always higherthan C-bank leverage
I C-bank has key competitive advantage: deposit insuranceI To deliver same profit to equity owners, S-banks need to have higher
leverage (holding constant S-bank default risk)
I Also recall optimal S-bank leverage
LS = ψHS(RS) (L2)
Begenau & Landvoigt Financial Regulation 16 / 33
Equilibrium Characterization: Relative Size of Sectors
1
𝐿𝑆
𝑅𝑆
𝐿𝑆0
𝑅𝑆0
I Red line: indifference conditionLS = (1/4(1− θ)2 + 1/2(1− θ)ψHC (RS))1/2
I Blue line: leverage condition LS = ψHS(RS)
I Key property: decreasing returns, i.e. H ′S(RS) < 0 and H ′C (RS) > 0
Begenau & Landvoigt Financial Regulation 17 / 33
Effect of Higher θ
1
𝐿𝑆
𝑅𝑆𝑅𝑆0
𝐿𝑆0
I Indifference condition LS = (1/4(1− θ)2 + 1/2(1− θ)ψHC (RS))1/2 ↓I Higher θ makes C-banks less profitable and S-banks relatively more
profitable ⇒ S-bank sector expands: RS ↑I But decreasing returns ⇒ lower S-bank liquidity premium ⇒ LS ↓
Begenau & Landvoigt Financial Regulation 18 / 33
How Should Regulator Set θ?
Proposition
Index competitive equilibria by the factor m > −1, such that
LC = (1 + m)ψHC (RS),
and the function θ = f (m) that determines the value of θ implementingequilibrium m.
(i) There is no θ ∈ [0, 1] that implements the planner allocation.
(ii) In any equilibrium with m ≥ 0, an increase in the capital requirementθ is welfare-improving.
I Planner wants LC = ψHC (RS), so could choose θ = f (0)
I But always have LS > LC in equilibrium due to deposit insurance andcompetition ⇒ need additional policy tool to regulate S-banks
I Still welfare-improving to raise θ in world with m > 0
Begenau & Landvoigt Financial Regulation 19 / 33
Effect With More General Preferences
1
𝐿𝑆
𝑅𝑆𝑅𝑆0
𝐿𝑆0
𝐻 𝐴𝑆, 𝐴𝐶 =𝛼𝐴𝑆
𝜀 + (1 − 𝛼)𝐴𝐶𝜀 1/𝜀 1−𝛾𝐻
1 − 𝛾𝐻
I Ambiguous net effect of higher θ
1. makes C-banks less profitable and shadow bank equity more attractive
2. shadow bank share expands, liqu. premium on S-bank debt declinesrelative to C-bank debt
3. with γH > 0, marginal benefit of total liquidity goes up as H(AS ,AC )falls and leverage curve shifts up
Begenau & Landvoigt Financial Regulation 20 / 33
Overview of Talk
I Static ModelI What pins down size and leverage of shadow banks?I Effect of tighter capital requirementI Efficient allocation vs. equilibrium
I Dynamic Quantitative ModelI Differences to simple modelI Calibration highlightsI Quantitative results
Begenau & Landvoigt Financial Regulation 21 / 33
Dynamic Model: Key Differences
1. Infinite horizon model with bank-independent sector (endowment)and bank-dependent sector (production)
I Banks have access to standard investment technologyI Convex capital adjustment costs
2. Riskier S-banks: runs and implicit bail-out guaranteesI S-banks subject to stochastic deposit redemption shocks %t More Details
I Introduces additional losses through fire-saleI Government bails out S-bank liabilities with probability πB
3. Risk averse households with preferences
U(Ct ,H
(ASt ,A
Ct
))=
C 1−γt
1− γ+ ψ
([α(AS
t )ε + (1− α)(ACt )ε] 1ε
)1−γH
1− γH
I Portfolio choice of equity and debt of both types of banksI Inelastic labor supply
Begenau & Landvoigt Financial Regulation 22 / 33
State Variables and Solution Method
I Exogenous states
log(Yt+1) = (1− ρY )log(µY ) + ρY log(Yt) + εYt+1
Zt = φZYt exp(εZt )
and %t follows a two-state Markov-process
I Endogenous states
1. Capital stock2., 3. C-bank and S-bank debt
4. S-bank capital share
I Solve using non-linear projection methodsI Probability of default bounded in [0, 1]I Nonlinear dynamics because of bankruptcy option
I Report results for simulated model
Begenau & Landvoigt Financial Regulation 23 / 33
Calibration: Consolidated View of Shadow Banks
Assets
Comm.
Paper
Equity
MMMF
Shares
Comm.
Paper
AssetsDebt
Equity
Shadow banks
Finance Company Money Market Mutual Funds
Consolidated
Begenau & Landvoigt Financial Regulation 24 / 33
Key Parameters: Quarterly data 1999− 2015Values Target Data Model
Bank leverage and defaultδS 0.300 Quarterly corp. bond default rate 0.36% 0.31%δC 0.175 Quarterly net loan charge-offs 0.25% 0.26%ξC 0.515 Recovery rate Moody’s 63% 62%ξS 0.415 Recovery rate Moody’s 63% 63%πB 0.905 Shadow bank leverage 93% 93%
Liquidity preferencesβ 0.993 C-bank debt rate 0.39% 0.39%α 0.330 Shadow banking share (Gallin 2013) 35% 34%ψ 0.0103 Liquidity premium C-banks; KV2012 0.18% 0.17%γH 1.700 Corr(GDP, C-bank liquid. premium) -0.28 -0.39ε 0.420 S-bank liquidity elasticity 0.17% 0.16%
RunsδK 4 ×δK Max. haircut (GM 2009) 20% 19%Z 26% × Z Forecl. discount (Campbell et al 2011)% [0, 0.3] Fraction run
Prob%
[0.97 0.030.33 0.67
]Uncond. run prob. (Covitz et al 2013) 3% 3%
Begenau & Landvoigt Financial Regulation 25 / 33
Liquidity Preference Parameters (1/2)
I How are key liquidity preference parameters disciplined by data?
I ψ: level of liquidity premiumI Krishnamurthy & Vissing-Jorgenson 2012 estimate annual premium of
75 bpsI ψ directly scales marginal liquidity benefit in model
I α: market share of S-banksI Higher α raises S-bank relative to C-bank premiumI Lowers funding cost, increases demand for capital of S-banks
I γH : comovement of premium with GDPI Countercyclical in data: liquidity is abundant in good times,
so premium is lowI Model matches countercyclical premium with γH = 1.7 ⇒
downward-sloping demand curve for liquidity
Begenau & Landvoigt Financial Regulation 26 / 33
Liquidity Preference Parameters (1/2)
I How are key liquidity preference parameters disciplined by data?
I ψ: level of liquidity premiumI Krishnamurthy & Vissing-Jorgenson 2012 estimate annual premium of
75 bpsI ψ directly scales marginal liquidity benefit in model
I α: market share of S-banksI Higher α raises S-bank relative to C-bank premiumI Lowers funding cost, increases demand for capital of S-banks
I γH : comovement of premium with GDPI Countercyclical in data: liquidity is abundant in good times,
so premium is lowI Model matches countercyclical premium with γH = 1.7 ⇒
downward-sloping demand curve for liquidity
Begenau & Landvoigt Financial Regulation 26 / 33
Liquidity Preference Parameters (1/2)
I How are key liquidity preference parameters disciplined by data?
I ψ: level of liquidity premiumI Krishnamurthy & Vissing-Jorgenson 2012 estimate annual premium of
75 bpsI ψ directly scales marginal liquidity benefit in model
I α: market share of S-banksI Higher α raises S-bank relative to C-bank premiumI Lowers funding cost, increases demand for capital of S-banks
I γH : comovement of premium with GDPI Countercyclical in data: liquidity is abundant in good times,
so premium is lowI Model matches countercyclical premium with γH = 1.7 ⇒
downward-sloping demand curve for liquidity
Begenau & Landvoigt Financial Regulation 26 / 33
Liquidity Preference Parameters (2/2)
qCt − qSt = β
(MRSC
qC−
MRSS
qS
)γCt+1 +
βF S
qSF St+1
+ β
((1− ε− γH)
(MRSC
qC−
MRSS
qS
)α
(AS
H
)ε+ (1− ε)
MRSS
qS
)ASt+1
+ β
((1− ε− γH)
(MRSC
qC−
MRSS
qS
)(1− α)
(AC
H
)ε− (1− ε)
MRSC
qC
)ACt+1
I ε: elasticity of substitution between S- and C-bank debtI Log-linear approximation of spread 1/qS − 1/qC (shadow rate −
deposit rate)I If ε = 1 (perfect substitutes) and γH = 0 (CRS in liquidity), quantities
of debt (AS ,AC ) do not matter for spreadI If ε < 1, would expect negative sign on AC and positive sign on AS
I Regression of CP − Tbill spread on Tbill supply, shadow debt supply(and controls) gives elasticity of 17 bp
I Matched in model with ε = 0.42 (net substitutes)
Begenau & Landvoigt Financial Regulation 27 / 33
Increasing Capital Requirement
Larger shadow banking share, C-banks “exit”, S-bank “enter”
Benchmark 13% 17% 20% 30%mean mean mean mean
Capital and Debt1. Capital 4.005 +0.2% +0.6% +0.9% +2.2%2. Debt share S 0.349 +3.4% +6.2% +8.1% +16.0%3. Capital share S 0.342 +1.1% +0.6% -0.1% -1.7%4. Leverage S 0.933 +0.1% +0.4% +0.6% +1.0%5. Leverage C 0.900 -3.3% -7.8% -11.1% -22.4%6. Early Liquidation (runs) 0.004 +0.3% +0.8% +1.2% +2.3%
Prices7. Deposit rate S 0.49% -0.7% -2.0% -3.0% -6.5%8. Deposit rate C 0.39% -4.1% -9.1% -13.0% -28.6%9. Convenience Yield S 0.23% +2.0% +5.3% +8.0% +17.8%10. Convenience Yield C 0.31% +5.2% +11.5% +16.4% +36.2%
Begenau & Landvoigt Financial Regulation 28 / 33
Increasing Capital Requirement
C-banks become safer, but S-banks riskier
Benchmark 13% 17% 20% 30%mean mean mean mean
Capital and Debt1. Capital 4.005 +0.2% +0.6% +0.9% +2.2%2. Debt share S 0.349 +3.4% +6.2% +8.1% +16.0%3. Capital share S 0.342 +1.1% +0.6% -0.1% -1.7%4. Leverage S 0.933 +0.1% +0.4% +0.6% +1.0%5. Leverage C 0.900 -3.3% -7.8% -11.1% -22.4%6. Early Liquidation (runs) 0.004 +0.3% +0.8% +1.2% +2.3%
Prices7. Deposit rate S 0.49% -0.7% -2.0% -3.0% -6.5%8. Deposit rate C 0.39% -4.1% -9.1% -13.0% -28.6%9. Convenience Yield S 0.23% +2.0% +5.3% +8.0% +17.8%10. Convenience Yield C 0.31% +5.2% +11.5% +16.4% +36.2%
Begenau & Landvoigt Financial Regulation 28 / 33
Increasing Capital Requirement
Interest rates fall as liquidity premia rise ⇒ more investment
Benchmark 13% 17% 20% 30%mean mean mean mean
Capital and Debt1. Capital 4.005 +0.2% +0.6% +0.9% +2.2%2. Debt share S 0.349 +3.4% +6.2% +8.1% +16.0%3. Capital share S 0.342 +1.1% +0.6% -0.1% -1.7%4. Leverage S 0.933 +0.1% +0.4% +0.6% +1.0%5. Leverage C 0.900 -3.3% -7.8% -11.1% -22.4%6. Early Liquidation (runs) 0.004 +0.3% +0.8% +1.2% +2.3%
Prices7. Deposit rate S 0.49% -0.7% -2.0% -3.0% -6.5%8. Deposit rate C 0.39% -4.1% -9.1% -13.0% -28.6%9. Convenience Yield S 0.23% +2.0% +5.3% +8.0% +17.8%10. Convenience Yield C 0.31% +5.2% +11.5% +16.4% +36.2%
Begenau & Landvoigt Financial Regulation 28 / 33
Increasing Capital Requirement
DWL from C-banks decline, from S-banks rise
BM 13% 17% 20% 30%mean mean mean mean
Welfare11. DWL S 0.001 +5.2% +11.7% +16.7% +31.8%12. DWL C 0.003 -68.6% -94.6% -98.8% -100.0%13. Default Rate S 0.31% +3.7% +9.9% +15.1% +29.7%14. Default Rate C 0.26% -67.2% -94.1% -98.7% -100.0%
15. GDP 1.365 +0.0% +0.1% +0.1% +0.2%16. Liquidity Services 1.969 -2.3% -5.2% -7.3% -14.4%17. Consumption 1.261 +0.13% +0.19% +0.20% +0.24%18. Vol(Liquidity Services) 0.068 -2.5% -6.1% -8.9% -18.7%19. Vol(Consumption) 0.005 +0.4% +0.6% +0.7% +10.5%
20. HH Welfare -114.596 +0.106% +0.129% +0.116% +0.048%
Begenau & Landvoigt Financial Regulation 29 / 33
Increasing Capital Requirement
More consumption and lower liquidity provision
BM 13% 17% 20% 30%mean mean mean mean
Welfare11. DWL S 0.001 +5.2% +11.7% +16.7% +31.8%12. DWL C 0.003 -68.6% -94.6% -98.8% -100.0%13. Default Rate S 0.31% +3.7% +9.9% +15.1% +29.7%14. Default Rate C 0.26% -67.2% -94.1% -98.7% -100.0%
15. GDP 1.365 +0.0% +0.1% +0.1% +0.2%16. Liquidity Services 1.969 -2.3% -5.2% -7.3% -14.4%17. Consumption 1.261 +0.13% +0.19% +0.20% +0.24%18. Vol(Liquidity Services) 0.068 -2.5% -6.1% -8.9% -18.7%19. Vol(Consumption) 0.005 +0.4% +0.6% +0.7% +10.5%
20. HH Welfare -114.596 +0.106% +0.129% +0.116% +0.048%
Begenau & Landvoigt Financial Regulation 29 / 33
Welfare
10 15 20 25 30
%
0
0.05
0.1
0.15
Wel
fare
: con
sum
ptio
n eq
uiv.
uni
ts %
Begenau & Landvoigt Financial Regulation 30 / 33
Transition Dynamics
0 4 8 12 16 200
0.05
0.1
0.15Consumption
0 4 8 12 16 20-6
-5
-4
-3
-2
-1
0Liquidity
0 4 8 12 16 20-0.1
0
0.1
0.2
0.3
0.4Capital
0 4 8 12 16 200
0.1
0.2
0.3
0.4
0.5S Cap Share
0 4 8 12 16 200
0.1
0.2
0.3
0.4
0.5
0.6S Leverage
0 4 8 12 16 200
2
4
6
8S Debt Share
Begenau & Landvoigt Financial Regulation 31 / 33
Other Policies
I Set time-varying insurance fee such that fund breaks evenI Fair κ does not reduce C-bank leverage, but shifts activity to S-bankI ⇒ Less liquidity provision and higher deadweight losses
θ=17% fair κ Corr(θt ,Yt) Corr(θt ,Yt) Minn. plan< 0 > 0
Debt share S +6.18% +9.04% +5.95% +6.12% -63.88%Leverage S +0.39% -0.22% +0.42% +0.37% +0.02%Leverage C -7.78% -0.00% -7.77% -7.79% -14.45%
DWL S +11.70% +2.57% +11.83% +11.99% -70.70%DWL C -94.63% -5.49% -80.47% -83.73% -99.72%Liquidity Services -5.16% -1.43% -5.08% -5.16% -14.02%Consumption +0.19% -0.00% +0.16% +0.17% +0.28%Welfare +0.129% -0.011% +0.115% +0.118% +0.144%
Begenau & Landvoigt Financial Regulation 32 / 33
Other Policies
I Set cyclical capital req’s with mean 17%I similar effects as with static optimal θ
θ=17% fair κ Corr(θt ,Yt) Corr(θt ,Yt) Minn. plan< 0 > 0
Debt share S +6.18% +9.04% +5.95% +6.12% -63.88%Leverage S +0.39% -0.22% +0.42% +0.37% +0.02%Leverage C -7.78% -0.00% -7.77% -7.79% -14.45%
DWL S +11.70% +2.57% +11.83% +11.99% -70.70%DWL C -94.63% -5.49% -80.47% -83.73% -99.72%Liquidity Services -5.16% -1.43% -5.08% -5.16% -14.02%Consumption +0.19% -0.00% +0.16% +0.17% +0.28%Welfare +0.129% -0.011% +0.115% +0.118% +0.144%
Begenau & Landvoigt Financial Regulation 32 / 33
Other PoliciesI “Minneapolis plan”: θ =23% and tax on S-bank debt of 30 bps
I Shrinks S-banks while making C-bank saferI Large drop in liquidity production, but greatest overall welfare gainI Consistent with welfare analysis in simple model: both banks have too
high leverage in status quo
θ=17% fair κ Corr(θt ,Yt) Corr(θt ,Yt) Minn. plan< 0 > 0
Debt share S +6.18% +9.04% +5.95% +6.12% -63.88%Leverage S +0.39% -0.22% +0.42% +0.37% +0.02%Leverage C -7.78% -0.00% -7.77% -7.79% -14.45%
DWL S +11.70% +2.57% +11.83% +11.99% -70.70%DWL C -94.63% -5.49% -80.47% -83.73% -99.72%Liquidity Services -5.16% -1.43% -5.08% -5.16% -14.02%Consumption +0.19% -0.00% +0.16% +0.17% +0.28%Welfare +0.129% -0.011% +0.115% +0.118% +0.144%
Begenau & Landvoigt Financial Regulation 32 / 33
Conclusion
I Tractable quantitative GE model with two types of banks
I Increasing capital requirement on commercial banks
I makes C-banks less, S-banks more profitableI leads to larger and riskier S-bank sectorI less liquidity provisionI no negative effects on production and investment in total
I Welfare trade-off: greater consumption (fewer bank failures)versus reduced liquidity provision
I Key Model Lessons
I GE effects (e.g. deposit rate response to higher cap reg)I Nature of competition between S-bank & C-bankI Slight increase in S-bank risk does not undermine intended benefits of
tighter capital regulation
Begenau & Landvoigt Financial Regulation 33 / 33
Fraction of Liquid Wealth in MMA at Household Level
0 20 40 60 80 100
Percentile of Wealth Distribution
0
0.1
0.2
0.3
0.4
0.5
0.6
Fra
ctio
n M
MA
+ M
MM
F
Source: 2013 SCF Back
Begenau & Landvoigt Financial Regulation 34 / 33
S-bank problem Back
vS(Zt) = maxbSt+1≥0,kS
t+1≥0kSt+1
(qS(bSt+1)bSt+1 − pt
)− φK
2
(kSt+1 − 1
)2
+ kSt+1Et
[Mt,t+1 ΠS
t+1ΩS(LSt+1)],
with
ΩS(LSt ) = (1− F Sρ,t)
(ρS,+t
(1− `St
(1− xSt
))− LSt + (1− `St )
vS(Zt)
ΠSt
)− F S
ρ,tδS
I Endogenous liquidation (fraction of assets)
`St =%St B
St
KSt ΠH
t
I Probability of default F Sρ,t = F S
ρ (ρSt ) with threshold
ρSt =LSt − (1− `St ) vS (Zt)
ΠSt− δS
1− `St(1− xSt
)increasing in leverage, liquidation fraction, and fire sale discount
Begenau & Landvoigt Financial Regulation 35 / 33
Bankruptcy & Deposit InsuranceI C-bank default
I Government bails out liabilities of failing C-banksI Recovers
rC (LCt ) = (1− ξC )ρC ,−t
LCtper bond issued by C-banks
I S-banks defaultI Benchmark: government does not bailout failing S-banks bails out
liabilities of failing S-bank with probability πBI Recovery value per bond
rS(LSt ) = (1− ξS)(1− `St (1− xt))ρS,−t
LSt
I Required taxes in addition to deposit insurance revenue
Tt = FCρ,t(1− rC (LCt ))BC
t − κBCt+1︸ ︷︷ ︸
Net payment to defaulting C-bank depositors
+ πBFSρ,t(1− rS(LSt ))BS
t︸ ︷︷ ︸Bailout for S-bank
Begenau & Landvoigt Financial Regulation 36 / 33
Bankruptcy & Deposit InsuranceI C-bank default
I Government bails out liabilities of failing C-banksI Recovers
rC (LCt ) = (1− ξC )ρC ,−t
LCtper bond issued by C-banks
I S-banks defaultI Benchmark: government does not bailout failing S-banks bails out
liabilities of failing S-bank with probability πBI Recovery value per bond
rS(LSt ) = (1− ξS)(1− `St (1− xt))ρS,−t
LSt
I Required taxes in addition to deposit insurance revenue
Tt = FCρ,t(1− rC (LCt ))BC
t − κBCt+1︸ ︷︷ ︸
Net payment to defaulting C-bank depositors
+ πBFSρ,t(1− rS(LSt ))BS
t︸ ︷︷ ︸Bailout for S-bank
Begenau & Landvoigt Financial Regulation 36 / 33
Bankruptcy & Deposit InsuranceI C-bank default
I Government bails out liabilities of failing C-banksI Recovers
rC (LCt ) = (1− ξC )ρC ,−t
LCtper bond issued by C-banks
I S-banks defaultI Benchmark: government does not bailout failing S-banks bails out
liabilities of failing S-bank with probability πBI Recovery value per bond
rS(LSt ) = (1− ξS)(1− `St (1− xt))ρS,−t
LSt
I Required taxes in addition to deposit insurance revenue
Tt = FCρ,t(1− rC (LCt ))BC
t − κBCt+1︸ ︷︷ ︸
Net payment to defaulting C-bank depositors
+ πBFSρ,t(1− rS(LSt ))BS
t︸ ︷︷ ︸Bailout for S-bank
Begenau & Landvoigt Financial Regulation 36 / 33