Optimal Deposit Insurance Eduardo Dvila (NYU, Stern) Itay Goldstein (U-Penn, Wharton) Banks and the Threat of Runs $UXQRQ$PHULFDQ8QLRQ UXQRQ$PHULFDQ8QLRQRQRQ$PHULFDQ8QLRQ$UXQRQ$PHULFDQ8QLRQ P HULFDQRQ$PHULFDQ8QLRQ8 QRQ$PHULFDQ8QLRQLRQRQ$PHULFDQ8QLRQ 2 %DQRQ$PHULFDQ8QLRQN

Banks and the Threat of Runs Contd Banks provide maturity and liquidity transformation This can improve welfare, but It exposes banks to the risk of a run Many investors demand early withdrawal out of the self fulfilling belief that others will do so History of many bank failures around the world

3 A Possible Solution: Deposit Insurance Insurance of deposits may reduce the incentive of investors to run Deposit insurance was enacted in the US in 1933 and had a great success in stabilizing the banking system Many countries in the world have followed this experience enacting different forms of deposit insurance

Supported by theoretical literature, going back to Diamond and Dybvig (1983) 4 Optimal Amount of Coverage Key question in design of insurance: How much should be insured? In Diamond and Dybvig (1983): Unlimited insurance: insurance works to prevent failures altogether and so has no cost

In the real world: Insurance always limited; e.g., in US current maximum for insurance is $250,000, which was increased from $100,000 in 2008 What is different in the real world? Failures sometimes happen generating costs Insurance causes frictions How to set the optimal amount? 5

History of Deposit Insurance Amount in the US 6 Ingredients of A Model Three dates: 0, 1, 2 Banks: Take deposits at time 0 Allow withdrawal in time 1 or 2 Invest in asset that pays uncertain return in time 2 (higher in

expectation than liquidation value in time 1) Set deposit amounts in competition for deposits Investors withdraw early in time 1 (run) if Have early liquidity need or Think return will be higher in time 1 (due to a run or to fundamentals) Act based on information and coordination 7 Ingredients of A Model Contd

Government: Sets deposit insurance; determines amount covered in case of failure Maximizes expected utility of depositors taking into account response of bank and depositors to insurance Equilibrium: Government sets deposit insurance coverage Banks set deposit rate, given insurance, and taking into account the effect on runs Depositors choose whether to run, given insurance and deposit rates Runs happen if fundamentals are below threshold (panic and fundamental runs)

Higher insurance coverage reduces run probability 8 Sufficient Statistic Approach Usually, getting quantitative prescriptions from a model of this kind requires calibration and estimation of exogenous deep parameters of the model This is a difficult task The sufficient statistic approach targets

endogenous high level variables that are potentially observable Illustration in next slide is based on Chetty (2009) 9 Sufficient Statistic Approach Contd 10 Optimal Level of Deposit

Insurance Based on Sufficient Statistic 11 Intuition Benefit from deposit insurance: reducing the probability of a run and increasing consumption as a result Cost of deposit insurance: causes fiscal costs in case a failure does happen Note: moral hazard concerns associated with

banks behavior only enter the fiscal cost (which is not internalized by banks) Other implications of banks behavior are internalized (envelope theorem, competition) 12 Measurement The variables in the formula are either observable or could be inferred from the data In particular, B, C, and D are fairly easily observable or known from various other studies

A is more difficult, need more data to figure out historical sensitivity Theory tells us what we need to measure Ideally, regression of failures on deposit insurance amount 13 Numerical Example: Effect of Coverage on Welfare 14

Uses of Formula Use formula to find optimal amount Usually interim maximum Welfare decreases initially, then increases, then decreases again Too ambitious? Use cost vs. benefit to tell whether an increase or a decrease is desirable at current level of coverage Back out change in failure probability or sensitivity that would rationalize recent insurance coverage

increases 15 Deposit Insurance Premium Paper extended to allow the government to set deposit insurance premium Used to make banks internalize the effect of their deposit rates on fiscal costs Can also be used for funding the insurance in addition to setting incentives Deposit ceiling can also be used

Formula can be adjusted to tell optimal coverage given the pricing of premiums 16 Extensions Cross-section of depositors (heterogeneity) General portfolio and investment decisions (moral hazard) Global game (information structure and equilibrium selection)

Macro effects (spillovers) 17 Conclusion Optimal amount of deposit insurance is first-order question with little scientific guidance to date Paper provides characterization of optimal deposit insurance as a function of a few sufficient statistics For a wide range of environments Additional characterization of optimal ex-ante policies, such as insurance premium

Paper provides guidance for what we need to measure in the data 18