The Cyclical Behavior of Default and Recovery Rates

Abstract

Using a regime switching framework we investigate the determinants of default clustering. We find that a common credit cycle, modeled as a two-state Markov chain, can account for a large portion of default correlations, with the residual clustering being captured by a factor structure. During credit crunches default rates increase, and so does the conditional residual correlation. Using data for the period 2000-2006 we find that this common cycle is robust in explaining a large part of the default behavior of individual industries and recovery rates. We also find that structured products that depend on the distribution tail are the first to respond to a deterioration of the credit environment.

Keywords: copula, credit Rating, default correlation, credit cycle, regime switching, recovery rates