Valuing Corporate Liabilities

We implement a structural bond pricing framework on a large panel of US industrial issues using an efficient maximum likelihood methodology. Although, like others before us, we underpredict yield spread levels when using only stock market data in the estimation, our errors are much less dispersed. In addition, we show that when our model underpredicts spreads, the errors are correlated with liquidity proxies, suggesting that an underestimation of total yield spreads may be economically plausible. When we include bond price information in our estimation, our errors become similar in magnitude to those found in recent implementations of reduced form models.