Investigating the Sources of Default Risk: Lessons from Empirically Evaluating Credit Risk Models

Abstract

From a credit risk perspective, little is known about the distress factors -- economy-wide or firm-specific -- that are important in explaining variations in defaultable coupon yields. This paper proposes and empirically tests a family of credit risk models. Empirically, we find that firm-specific distress factors play a role (beyond treasuries) in explaining defaultable coupon bond yields. Credit risk models that take into consideration leverage and book-to-market are found to reduce out-of-sample yield fitting errors (for the majority of firms). Moreover, the empirical evidence suggests that interest rate risk may be of first-order prominence for pricing and hedging. Measured by both out-of-sample pricing and hedging errors, the credit risk models perform relatively better for high grade bonds. Controlling for credit rating, the model performance is generally superior for longer maturity bonds compared to its shorter maturity counterparts. Using equity as an instrument reduces hedging errors. This paper provides an empirical investigation of credit risk models using observable economic factors.