Why Effort on Structural Models is Worthwhile: New Evidence from a Parasian-Option Style Model

Abstract

Since the seminal work of Merton(1974), structural models have frequently been implemented to price defaultable bonds. Empirical evidence shows significant skewness of results using Merton's model. The question is whether modifications can sufficiently improve the performance of structural models. This paper compares existing models with a relatively new variant which takes into account lengthy bankruptcy negotiations.

We investigate three structural models for pricing defaultable bonds as contingent claims: (i) the original Merton model augmented to allow for coupons; (ii) extension to include stochastic interest rates; (iii) a further extension involving a ParAsian option feature to include the Chapter 11 effect (Yu et al., 2004). The fact that creditors tend to allow a grace period during distress is largely recognized but rarely incorporated into bond-pricing models. This is the motivation for our new model.

To test the three models empirically, data from corporations and bond issues are collected. For each model, the numerical scheme is described briefly and computations are performed using parameters estimated from the data. Comparisons of numerical results are then presented and compared.

The primary conclusion is that the effect of lengthy bankruptcy is priced by the market and that by incorporating this factor our model helps explain the well-known spread between the Merton model and real data.

Keywords: bankruptcy resolution, ParAsian option, bond pricing, structural model, empirical test