<P>Along with the likelihood of default, recovery upon default is a key ingredient of creditrisk<BR>models. We document the empirical determinants of the recovery on defaulted securities<BR>in the United States over the period 1982–1999, using the prices of defaulted securities at<BR>the time of default and at the time of emergence from default or from bankruptcy. In addition<BR>to seniority and security of the defaulted securities, industry conditions at the time of<BR>default are found to be robust and important determinants of the recovery rates: Recovery<BR>in a distressed state of the industry (median annual stock return for the industry firms being<BR>less than −30%) is lower than the recovery in a healthy state of the industry by 10 to 20<BR>cents on a dollar; a better liquidity position of the peers of the defaulted firm also implies<BR>higher recovery at emergence; recovery is negatively correlated with asset-specificity of the<BR>industry when the industry is in distress, but not otherwise; finally, recovery is not a ected<BR>by macroeconomic and aggregate bond-market supply conditions, once the industry distress<BR>e ect is accounted for. Our results underscore the existence of substantial variability in<BR>recoveries, in the cross-section of securities as well as in the time-series, and suggest that in<BR>order to capture recovery risk, the next generation of credit risk models should include an<BR>industry factor in addition to the factor representing the firm value or the default process.</P>
<P>Keywords: Recovery, Loss given default, Credit risk, Default, Bankruptcy.</P><BR>