Testing the differences between the determinants of moody's and standard & poor's ratings an application of smooth simulated maximum likelihood estimation

Abstract

This paper extends previous studies on bond ratings by modeling as a system of equations the determinants of a municipality's decision to obtain a bond rating and the determinants of the municipality's rating for the two major rating agencies. Our model provides a framework to examine formally the differences between the two agencies in the determinants of the ratings. We estimate the four-equation system by smooth simulated maximum likelihood estimation and then construct minimum x 2 tests on cross-equation restrictions based on optimal minimum distance estimation. Self-selection is found to be important in Moody's ratings while not in those of S&P. Split ratings appear to reflect differences in both the weight attached to specific determinants of the ratings and differences in the way the bonds are classified. 52C.-G. MOON AND J. G . STOTSKY ratings. These studies have primarily employed cross-section analysis with several different statistical techniques. Although early research used ordinary least squares and multiple discriminant analysis techniques to investigate the determinants of bond ratings, some recent research has used univariate ordered probit and logit analysis. These techniques are appropriate in analyzing this issue because, unlike ordinary least squares, the dependent variable-the rating-is not constrained to lie along a linear scale but only along an ordinal scale, and the spacing of the categories is determined by the data.Several studies have used ordered categorical variable techniques to examine the determinants of municipal bond ratings. Farnham (1984, 1985) consider the determinants of Moody's and Standard & Poor's (S&P's) ratings using univariate ordered probit estimation and Wescott (1984) analyzes the determinants of Moody's ratings using cluster and probit analysis. Several studies have also applied ordered categorical variable techniques to the analysis of corporate debt ratings.The first econometric extension in this analysis is to account for self-selection. A selectivity problem may arise in this analysis from the practice of considering only observations in a sampIe for which a bond rating has been obtained rather than from considering all observations with outstanding debt in the sample. The selectivity problem arises in analyzing these ratings in that the rating is observed only for those municipalities that have chosen to get rated. Considering the sample of municipalities that issue general obligation debt, it is inappropriate to estimate the determinants of the ratings with only the municipalities that have a rating if there is some systematic reason that the municipalities that do not have a rating have chosen not to obtain one. In this case, the expectation of the error term of the regression equation explaining the determinants of the rating, conditional on that municipality having a bond rating, is not equal to zero. [See Heckman (1979).] The bias that results from neglecting the self-selection problem is indeterminate, as it depends on the nature of the relationship b...

An analysis of the determinants of sovereign ratings

“…However, this is not the case since a rating of AAA/Aaa conveys different information as compared to a rating of AA+/Aa1. Hence, the use of OLS method is argued to be inappropriate for some multinomial choice variables, which are inherently, ordered such as ratings (see for example, Mckelvey & Zaviona, 1975;Moon & Stotsky, 1993).…”

The value of private sector business credit information sharing: The US case

“…Newer approaches to predicting bankruptcies have employed more sophisticated techniques. Examples are Theodossiou (1993), who uses a dynamic approach to testing for shifts in the mean of a time series to predict business failures, Cheng and Titterington (1994) who survey neural net approaches, and Moon and Stotsky (1993), who use smooth simulated maximum likelihood estimation. These studies are almost all based upon financial statement data, typically various financial ratios.…”

Determinants of sovereign ratings: A comparison of case-based reasoning and ordered probit approaches

“…However, this is not the case since a rating of AAA/Aaa conveys different information as compared to a rating of AA+/Aa1. Hence the use of OLS method is argued to be an inappropriate for some multinomial choice variables, which are inherently, ordered such as ratings (see for example, Mckelvey and Zaviona (1975); Moon and Stotsky (1993), BissoondoyalBheenick (2005)). …”