This paper sits at the intersection of three lines of research and policy making: valuation of liquidated assets, debt valuation and credit risk modeling. A variety of work in the past 15 years has found that liquidation values of corporate assets have a material bearing on the debt capacity of the firm prior to default. Though there are competing stories for this connection, the basic intuition has that differences inability to liquidate the firm translate back into differences in risk to lenders. As asset values after default decrease (increase), lenders are less (more) willing to extend credit.
Within the credit risk literature, this concept is well established as variation in loss-given-default (LGD). Research on liquidation costs has increased in recent years (see for example, Altmanetal. 2002,2005b, Acharyaetal. forthcoming ) and has sought, among other things, to explain the relationship between probabilities of default (PD) and LGD. Much of the empirical literature draws on the concept of a single underlying factor (see Frye 2000a, b, cforexamples) to explain the observed correlations.
A reading of the literature would suggest that a current open question is whether the correlation between PDandLGDshould imply differences in debt valuation that are not reflected in existing models. In particular, with a positive correlation of PD and LGD, one should see much larger changes in debt value (for relatively small changes in default probability) than occur incurrent models. Because existing structural debt valuation models make an assumption that PD and LGD enter independently, this valuation effect will be muted.
Resolving the debt valuation question requires linking the literatures. In particular, one needs a modifiation of the structural models of default used in the theoretical literature to account for the intuition derived in the credit risk fi eld. Key here is the presence of a single common factor that drives both PD and LGD. As will become apparent, for such models to reproduce the observed patterns, it is convenient to introduce an explicit dependence between PD/LGD and potentially time-varying asset market liquidity.
