This thesis examines how banks choose their optimal capital structure and cash reserves in the presence of regulatory measures. The first chapter, titled Bank Capital Structure and Tail Risk, presents a bank capital structure model in which bank assets are subject to both diffusion and tail risk. Of these two types of risk, tail risk causes uninsured deposits to be risky, as the bank's asset value can unexpectedly fall below the value of deposits in case of default. The model shows that tail risk, rather than diffusion risk, is the main driver of the risk on deposits when the bank is unregulated and of the endogenous deposit insurance premium when the bank is regulated. Keeping total volatility constant, the model shows that a high tail risk component leads to higher credit spreads, default risk, and magnitude of bank losses in default than a high diffusion risk component.The second chapter, titled Bank Regulation and Market Discipline in the Presence of Risk-Taking Incentives, presents a bank capital structure model in which equity holders can increase asset risk once debt is in place. I study the effects of capital requirements and subsidized deposit insurance on the bank's privately optimal funding and operational risk level. The model predicts that there are synergetic effects of regulation and market discipline. When the regulator sets the capital charge and deposit insurance premium payments sufficiently high for a risky portfolio, the bank commits to the low-risk asset portfolio by setting a lower leverage ratio and substituting market debt for deposits. This market discipline effect disappears when the regulatory costs become too high. In the third chapter that is titled Dividend Restrictions and Asymmetric Information, we develop a dynamic model of a bank whose management has superior information about the impact of a pending shock to the bank's cash holdings and can signal the bank's type through its dividend policy. Banks that will be adversely affected by the shock have incentives to pool with unaffected banks to increase their market value. To avoid being mimicked, the unaffected banks can credibly signal via a more aggressive payout strategy. Dividend payout restrictions have the potential to prevent a separating equilibrium from forming. This leads to the bad type adopting a more aggressive payout policy with a higher risk of default but mitigates the distortion of the good type's policy. We identify a number of scenarios where this trade-off presents an opportunity for regulatory intervention and some where it does not.
EPFL2021
