The central question of this paper is "how well can corporate bond ratings be explained by available financial and operating statistics?" This question may be reversed to ask, "how important to the determination of bond ratings are the 'intangibles of judgment' and the data that are presumably available to the professional rater, but not available in widely published data on the firm?" To provide evidence on this question, we have estimated the parameters of several regression equations designed to explain variations in the bond ratings. Before addressing this question, the authors present some comments on the importance and history of bond ratings (Section It). Prior studies of bond ratings are reviewed in Section III. Section IV discusses statistical methods and data. Variables that are believed to Be relevant determinants of bond ratings are defined, and the samples of industrial, railroad, and public utility bonds used in this study are described. This study is limited to the first four bond quality ratings — namely, Aaa, Aa, A, and Baa for each of these three traditional industrial classifications. Five explanatory variables are used: earnings coverage, long-term debt to capitalization, profitability, earnings instability, and asset size. Average values of these explanatory variables for each quality rating and industrial classification for the period 1961 through 1966 are also presented in this section. When presenting the empirical results of the study (Section V), we discuss the relative importance of the five variables in the statistical explanation of bond ratings, as well as their overall ability to explain bond ratings. [ABSTRACT FROM AUTHOR]