Bank Rescues and Bailout Expectations: The Erosion of Market Discipline During the Financial Crisis

Transcription

1 Bank Rescues and Bailout Expectations: The Erosion of Market Discipline During the Financial Crisis Florian Hett Goethe University Frankfurt Alexander Schmidt Deutsche Bundesbank & Goethe University Frankfurt Bailouts, bail-in, and financial stability Paris, November 28, 2014 Opinions and views expressed in this paper do not necessarily reflect the opinions of the authors institutions.

2 What is the aim of this paper? Develop a new methodology to test market discipline in the US financial sector Apply this method to study the effects of public interventions during the crisis (Bear Stearns, Lehman Brothers) Test whether bail-outs lower market discipline by increasing moral hazard

3 Main findings Prior Bear Stearns: existence of (some) market discipline Between Bear Stearns and Lehman: market discipline declined, in particular for large banks After Lehman Brothers: further strong decline in market discipline across the whole financial sector

4 Why is this question important? Any aid to a present bad bank is the surest mode of preventing the establishment of a future good bank. (Walter Bagehot, 1873) Costs of bank bail-out? Direct fiscal costs Loss in market discipline (Calomiris and Kahn, 1991) (Our definition:) Market discipline means a state in financial markets where claim holders put pressure on institutions so that they adhere to a proper risk behaviour.

5 Why care about market discipline and bailouts? Partial mis-alignment of incentives between debt and equity Debt holders want to maximize the repayment probability Equity holders want to maximize expected discounted residual profits Debt holders can withdraw their funds (Calomiris and Kahn, 1991) Banks are disciplined to implement appropriate risk behavior Market discipline is an important requirement of financial intermediation BUT: promise of a bail-out lowers incentives to monitor and control the bank

6 Bail-outs induce moral hazard and affect market discipline Promise of a bail-out ˆ= free put option Debt assumes a guaranteed repayment Moral Hazard: Incentives to monitor and control the bank vanish Drop in Market Discipline: bond holders do not control bank risks anymore as their claims are essentially risk-free Optimal reaction of banks: Risk-shifting more likely and more pronounced future financial crises Expectations of future bail-outs harm market discipline by stimulating moral hazard

7 Digression: The structural firm value approach Idea pioneered by Merton (1974): both debt D t and equity E t can be priced as derivatives on fundamental firm value V t Debt ˆ= risk-free investment + short put option on firm value Equity ˆ= call option on firm value Changes in equity and debt prices are linked via firm value Debt-to-equity sensitivity: β = D E E D = D V E E V D = HR ( 1 1) (1) Lev

8 Option pricing theory and debt-to-equity sensitivity Debt-to-equity sensitivity ( beta ): β = HR ( 1 Lev 1) = D V E V }{{} unobservable From general option pricing theory we know: ( 1 Lev 1) }{{} observable DV and E V depend on probability of ending in the money ( = default probability) with increasing downside risks, value of debt put increases and equity call declines hence for higher risks, DV (E V ) increases (decreases) Hedge ratio is higher for more risky banks

9 Hedge ratios, firm risks and bailouts From general option pricing theory we know: Hedge ratio increases in firm risks X We show that in the presence of a bailout probability PB: HR = (1 PB) HR (2) Test for changes in bailout expectations reduces to test for structural changes in hedge ratio

10 Empirical approach 1. Estimate fundamental CDS-equity semi-elasticity in control period (Jan Feb 2008) : CDS = 1 T β r E + ɛ = 1 T HR( 1 Lev 1) r E + ɛ = 1 T (α 0 + α j X j ) ( 1 Lev 1) r E + ɛ = α 0 r AE α j X j r AE + ɛ (3) with r AE = 1 T ( 1 Lev 1) r E 2. Test for structural changes in α j after major events (Bear Stearns, Lehman)

11 What do we expect? Hypothesis 1: The rescue of BSC lead to a decline in market discipline, absolute values of α j are lower than in control period Hypothesis 2: Possible effects of Lehman Brothers A: Default of LEH re-established pre-crisis level of market discipline, absolute values of α j increase to pre-bear Stearns level B: Default of LEH weakened market discipline permanently absolute values of α j are still lower than in control period Hypothesis 3 & 4: Effect heterogeneity with respect to size and firm type, Changes of α j are stronger for investment banks and large institutions Hypothesis 5: Effects should be less pronounced for non-financial companies

12 Our data set 31 US financial institutions: Commercial and Investment Banks, Insurance Companies and Other Financial Institutions (78 non-financial firms) Time span covered: Jan Sept 2010 Drop March and September 2008 due to potential excess correlation problem Due to better liquidity, we estimate CDS spread- equity (semi-)elasticity Total number of weekly observations: 8,760 Control period: 6,064 After BS: 548 After LEH: 2,148

13 Firm specific risk characteristics Implied Volatility Skewness weighted FCP spread Leverage

14 Baseline regression results ( , weekly data)

15 Results for largest financial firms ( , weekly data)

16 Results for investment banks ( , weekly data)

17 Results for non-financial firms ( , weekly data)

18 Conclusion New methodology to test for changes in market discipline Evidence for significant costs of public bail-outs due to moral hazard Market discipline declined already after Bear Stearns rescue, but effect is even more pronounced after Lehman default Effects are heterogeneous with respect to firm size and type Larger banks are more impacted than other financial companies Robustness analysis: non-financial firms are much less affected than financial companies

19 Results for unadjusted returns ( , weekly data)

20 Illustration of theoretical semi-elasticity and hedge ratio Credit spread Semi elasticity Hedge Ratio