Hazard versus Probit in Predicting U.S. Bank Failures: A Regulatory Perspective over Two Crises
39 Pages Posted: 24 Aug 2009 Last revised: 17 Jul 2014
Date Written: July 16, 2014
In this paper, we examine two bank early warning models – one based on the time-varying hazard model developed by Shumway (2001) and the other based upon the static probit model used by Cole and Gunther (1998) – from a unique perspective. Distinguished from existing empirical studies, we compare the out-of-sample accuracy of both models in forecasting U.S. bank failures from both academic and regulatory perspectives. When we limit both models to financial data available at the time of prediction, we find that the probit model slightly outperforms the hazard model, indicating that the superior performance of hazard models documented in previous empirical research is attributable to use of more timely financial data rather than to incorporation of time-varying covariates. We also find that a parsimonious specification fit to data over 1985-1993 performs well in forecasting bank failures during 2009-2010 – evidence that the characteristics of “distressed banks” have experienced little change over the past two decades despite substantial changes in structure and regulation of the U.S. banking industry. Our findings support supervision focusing on banks’ traditional CAMELS risk ratios, particularly those related to capital adequacy, liquidity and asset quality.
Keywords: bank failure, bank regulation, failure prediction, hazard model, probit model, financial crisis
JEL Classification: G17, G21, G28
Suggested Citation: Suggested Citation