TY - JOUR T1 - Modeling Ultimate Loss Given Default on Corporate Debt JF - The Journal of Fixed Income SP - 6 LP - 20 DO - 10.3905/jfi.2011.21.1.006 VL - 21 IS - 1 AU - Michael Jacobs, Jr. AU - Ahmet K. Karagozoglu Y1 - 2011/06/30 UR - https://pm-research.com/content/21/1/6.abstract N2 - 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 1985–2008 period. The authors build simultaneous equation models 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, the authors find the economic and statistical significance of firm-specific debt and equity market variables. In particular, they find that information from either the equity or the debt markets at around the time of default (measures of either cumulative equity returns or distressed debt prices, respectively) have predictive power with respect to the ultimate LGD, which is in line with recent prior recovery and asset pricing research. They also document a new finding: Larger firms have significantly lower LGDs while larger loans have higher LGDs.TOPICS: VAR and use of alternative risk measures of trading risk, big data/machine learning, credit risk management ER -