Understanding Changes in Corporate Credit Spreads

New evidence is reported on the empirical success of structural models in explaining changes in corporate credit risk. A parsimonious set of common factors and company-level fundamentals, inspired by structural models, was found to explain more than 54 percent (67 percent) of the variation in credit-spread changes for medium-grade (low-grade) bonds. No dominant latent factor was present in the unexplained variation. Although this set of factors had lower explanatory power among high-grade bonds, it did capture most of the systematic variation in credit-spread changes in that category. It also subsumed the explanatory power of the Fama and French factors among all grade classes.We provide new evidence on the empirical success of structural models in explaining corporate credit-risk changes. The basic factors used in the model were a set of common factors—the equity market return, change in prospects for company growth (proxied by change in the price-to-book ratio, or P/B), change in aggregate idiosyncratic equity volatility, the slope of the term structure, and change in the spot rate—and a set of company-level characteristics—the stock return (as a proxy for leverage), stock momentum, change in idiosyncratic equity volatility, and change in the P/B. We also examined the three Fama–French factors (for beta, size, and value–growth) and whether Federal Reserve Board policy was expansionary or recessionary.Studying 2,375 U.S. corporate bonds that were diverse in credit quality (ranging from AAA to D) in the 1990–2003 period, we found that a structural model with our set of factors was impressively successful in explaining credit-spread changes in the medium- and low-grade bond segments. Specifically, we found the model explained more than 54 percent of individual-bond credit-spread changes in the medium-grade group and 67 percent of changes in the low-grade group. In the highest-grade group, the explanatory power was about 36 percent.Explanatory power differs among bond classes partly because of the different roles played by company-level fundamentals, such as volatility, leverage, and growth opportunities. These variables are particularly important in the low-grade segment but play little role in the high-grade segment. Moreover, the common factors captured twice as much variation in the low-grade group as in the high-grade group.No dominant latent factor was present in the unexplained variation. Principal-components analysis applied to bond portfolios revealed no latent factor in the residual variation of low- or medium-grade portfolios but a strong latent factor in high-grade portfolios. Analysis of individual regression residuals revealed no variation, however, common to all the credit-risk groups, which provides solid evidence that structural models capture essentially all systematic variation in individual credit-spread changes for all credit-risk classes.We provide new evidence that idiosyncratic volatility and the P/B are important in empirical corporate bond pricing. The P/B and idiosyncratic volatility are both driven by uncertainty about a company’s future profitability, which affects default probability in a structural model framework. We show that changes in idiosyncratic volatility and the P/B at both the aggregate and company level are economically and statistically significant in explaining the time-series variation in corporate credit-spread changes.Our parsimonious set of variables subsumes the explanatory power of the Fama–French factors, which have traditionally been used in equity pricing. We found the Fama–French factors on their own to be significant and to explain about 26 percent of the variation in credit-spread changes, but when added to our proposed set of structural model variables, the Fama–French factors lost significance and did not increase overall explanatory power. This finding suggests that structural model factors capture the systematic risk in credit-spread changes better than do the Fama–French factors.