What's in a change in an industrial bond rating?

Abstract

Bond quality rating changes (BQRC) for industrial bonds are analyzed using both univariate statistical methods and discriminant analysis to find significant variables and their relationship with the changes. The single most important explanatory variable is found to be the rate of return on assets (ROA), followed by the trend in the return on assets (ROATREND). The univariate analysis found six of the seven proposed explanatory variables significant beyond the 0.01 level. The two-group discriminant analysis model achieved a correct classification rate of over 77%. The paper shows how the results of the two-group discriminant analysis can be used for a three-way prediction (upgrade, downgrade, or no change of bond ratings). The results of this study show that models based on financial statement data can predict rating changes with good accuracy and therefore may be a useful tool for rating agencies, at least as an initial screening device.