Credit Rating Gaps in Japan: Differences between Solicited and Unsolicited Ratings, and "Rating Splits"

Credit ratings have become an indispensable part of the fundamental information infrastructure of credit markets. Credit ratings cover a wide range of issuers including governments, governmental organizations, municipalities, nonfinancial companies and financial institutions, and also cover securitized products. For users, credit ratings are readily available tools to grasp the credit quality of securities (or issuers), as they rank, in the simple form of letter symbols, the ability of issuers to repay creditors in a timely manner in accordance with contractual obligations. It is necessary to fully understand the credit rating criteria, policies, and characteristics of rating agencies when users refer to credit ratings in making investment decisions or for other purposes. This paper attempts to clarify the current status and facts behind the two types of "rating gaps," which must be taken note of by users of credit ratings. These are (1) differences between solicited and unsolicited ratings, and (2) differences in ratings assigned to the same securities or issuers by different rating agencies, i.e., the "rating splits." We focus on the credit ratings of Japanese corporations; i.e., B6nonfinancial companies and nonbank financial companies, excluding banks and other types of financial institutions. While these differences have often been qualitatively discussed, this paper places emphasis on quantitative and objective analyses. In the analyses, although we use data of specific rating agencies, our focus is purely on the "differences," and we do not intend to rank the appropriateness of individual credit ratings. With regard to differences between solicited and unsolicited ratings, our results show that unsolicited ratings tend to be lower than solicited ratings, concurring with general views, and that such differences have been narrowing and are on average less than one notch recently. We calculate the differences, based on several assumptions, using credit ratings assigned by Standard and Poor's Ratings Services and Rating and Investment Information, as they disclose the distinctions of solicited and unsolicited ratings. The backdrop of such rating gaps may include disparity in the level of information available to rating agencies as well as cherry picking actions by the issuers. Some point out that solicited-unsolicited gaps are narrowing due to improvements in corporate disclosure among other factors, and the results of our analyses are consistent with such view. In addition, looking at the practices of credit rating usages by investors, they do not seem to distinguish solicited and unsolicited ratings in many cases. However, strong and deep-rooted concerns over the reliability of unsolicited ratings remain especially among issuers. In this respect, the issue of differences between solicited and unsolicited ratings deserves further analyses from various perspectives. In regard to rating splits, the findings are that on average a three-notch difference exists for a certain issuer between the highest rating assigned by one rating agency and the lowest rating by another rating agency. The existence of rating splits may indicate that the same credit rating symbols may signify dissimilar credit qualities by each rating agency, in other words rating scales may differ. In many cases, investors, in using credit ratings, seem to make necessary adjustments considering rating splits. We believe comparison against default rates is the ultimate test of rating splits, and assessments using data currently available suggest that differences stay within a certain range. Further analyses are also required in this respect, however, because data constraints, such as available amount of data for default rates and lengths of the sample period, are still large in Japan.