Open Collections

Abstract

In this thesis, I investigate how limited liability and aggregate uncertainty influence banks’ chosen portfolio liquidity. I use a three-period theoretical model inspired by Diamond and Dybvig (1983) and Allen and Gale (2004). The model features correlated withdrawals and interest rate shocks which cause asset values to fall and depositors to withdraw simultaneously. These features are inspired by the conditions that lead to the failures of Silicon Valley Bank and Silvergate Bank, which were characterized by unusually high interest rate and concentration risk related to cryptocurrency and venture-capital investments. Using this model, I examine outcomes under limited and unlimited liability and evaluate whether liquidity regulations similar to those introduced under Basel III can mitigate the likelihood and severity of runs and improve welfare. I find that limited liability generates a moral hazard by encouraging banks to under-insure against and tolerate a higher probability of runs. Liquidity requirements can prevent runs ex-post, but they do not improve welfare ex-ante as they fail to address the incentives and moral hazard created by the liability structure. When liability is unlimited, banks internalize the cost of runs and choose more liquid portfolios.