Factor-based prediction of industry-wide bank stress

Abstract

This article investigates the use of factor-based methods for predicting industry-wide bank stress. Specifically, using the variables detailed in the Federal Reserve Board of Governors' bank stress scenarios, the authors construct a small collection of distinct factors. We then investigate the predictive content of these factors for net charge-offs and net interest margins at the bank industry level. The authors find that the factors do have significant predictive content, both in and out of sample, for net interest margins but significantly less predictive content for net charge-offs. Overall, it seems reasonable to conclude that the variables used in the Fed's bank stress tests are useful for identifying stress at the industry-wide level. The final section offers a simple factor-based analysis of the counterfactual bank stress testing scenarios. © 2014, The Federal Reserve Bank of St. Louis.