Unveiling the Link between Tail Risk Measures and Bank Failure: Empirical Evidence from Us Bank Holding Companies

Abstract

This study presents an empirical investigation into the impact of tail risk measures such as value-at-risk (VaR), Cornish-Fisher Value-at-Risk (VaRCF), and Expected Shortfall (ES), on the probability of bank failures by using a novel definition of bank failure. We develop failure prediction models using quarterly bank data between 1988 and 2019 for publicly traded bank holding companies (BHCs) in the United States. The dataset includes 15,693 bank-quarter observations of 202 bailout BHCs and 34,980 of 1,108 non-bailout BHCs. Correlated random-effects logistic analysis and robustness checks confirm a significant positive relationship between tail risk measures and bank failures.

Keywords: bank bailout, Tail risk, value-at-risk, expected shortfall