Abstract:
We develop a structural bond pricing approach and implement it on a large panel of US industrial bonds using an efficient maximum likelihood methodology. We evaluate the model's ability to predict yield spread levels and changes out-of-sample. Errors are smaller and distinctly less variable than those found in previous implementations of structural as well as reduced form models. Furthermore, our analysis provide evidence that bond yield spreads incorporate a substantial liquidity component on top of the default spread structural models are designed to capture.