An Historical Loss Approach to Community Bank Stress Testing

49 Pages Posted: 3 Jan 2019

See all articles by Cao Fang

Cao Fang

University of Arkansas - Sam M. Walton College of Business

Timothy J. Yeager

University of Arkansas - Sam M. Walton College of Business

Date Written: December 18, 2018

Abstract

We develop a top-down macro stress test that assesses a community bank’s ability to withstand a severe and prolonged period of high credit losses. The model groups banks by geography and subjects them to the 90th percentile chargeoff rates that banks experienced between 2008 and 2012. Our historical loss approach better reflects patterns of community bank stress than econometric approaches that estimate the relationship between macroeconomic conditions and bank performance. We put all U.S. community banks at year-end 2017 through the test and highlight two results. First, banks are much better prepared to withstand an adverse shock than they were on the verge of the financial crisis because capital ratio distributions are higher and construction loan concentrations are lower. Second, the Tax Cuts and Jobs Act of 2017 may reduce the ability of banks to survive an adverse shock because it weakens the automatic stabilizer effect from operating losses. We also show that the stress test accurately projects in-sample insolvency risk for regional banks as well as community banks.

Keywords: Community Banks, Stress Testing, Financial Crisis, Loan Diversification, Regional Banks, Insolvency Risk

JEL Classification: G01, G21

Suggested Citation

Fang, Cao and Yeager, Timothy J., An Historical Loss Approach to Community Bank Stress Testing (December 18, 2018). Available at SSRN: https://ssrn.com/abstract=3303422 or http://dx.doi.org/10.2139/ssrn.3303422

Cao Fang

University of Arkansas - Sam M. Walton College of Business ( email )

Fayetteville, AR 72701
United States

Timothy J. Yeager (Contact Author)

University of Arkansas - Sam M. Walton College of Business ( email )

Fayetteville, AR 72701
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
56
Abstract Views
440
Rank
664,767
PlumX Metrics