Statistical properties and stability of ratings in a subset of US firms

Abstract

Standard explanatory variables that determine credit ratings do not achieve significant effects in a sample of 100 US non-financial firms in an ordered probit panel estimation. Sample size and selection as well as the distribution of explanatory variables across rating classes may be the cause this problem. Furthermore, we find evidence to suggest that variable coefficients vary over rating classes when analysed with an unordered loogit model. The sample reproduces well-established macroeconomic effects of credit ratings found by Blume et al. (1998) and highlights the influence of the rating agencies’ through-the-cycle approach on rating transitions.