Understanding and Predicting Ultimate Loss-Given-Default on Corporate Debt

Abstract

Loss given default (LGD) is a critical parameter in various facets of credit risk modeling. This study empirically investigates the determinants of LGD and builds alternative predictive econometric models for LGD on bonds and loans using an extensive sample of most major U.S. defaults in the period 1985–2008. We build a simultaneous equation model in the beta-link generalized linear model (BLGLM) class, identifying several that perform well in terms of the quality of estimated parameters as well as overall model performance metrics. This extends prior work by modeling LGD both at the firm and the instrument levels. In a departure from the extant literature, we find the economic and statistical significance of firm-specific, debt, and equity-market variables. In particular, we find that information from either the equity or the debt markets at around the time of default (measures of either distress debt prices or cumulative equity returns, respectively) have predictive power with respect to the ultimate LGD, which is in line with recent prior recovery and asset pricing research. We also document a new finding, that larger firms have significantly lower LGDs while larger loans have higher LGDs.

Keywords: Recoveries, Default, Loss Given Default, Financial Distress, Bankruptcy, Restructuring, Credit Risk, Entropic Methods, Bootstrap Methods, Forecasting

JEL Classification: G33, G34, C25, C15, C52