Rating Comparisons in the Literature
Among the first papers aimed to compare the ratings of different agencies was the one by Beattie and Searle (1992). Long-term credit ratings were gathered from 12 international credit rating agencies (CRA) that used similar scales. The sample of differences between the pairs of ratings for the same issuer was found. Around 20 % of the pairs in that sample involved differences in excess of two gradations. That may be explained by differing opinions about the financial stability of the issuers, as well as by different methodologies used by the rating agencies. But the average difference between ratings of the main international agencies S&P and Moody’s was insignificant.
Cantor and Packer (1994) compared Moody’s ratings of the international banks with the ratings of nine other rating agencies. It was found that the differences were greater on average than those discussed earlier. The average rating difference among the biggest international and three Japanese rating agencies was nearly three gradations.
The CRAs sometimes explain this effect in terms of a conservative approach when dealing with an unrequested rating because they do not have as much information about a company with which they have a rating contract, as they would with a company that has entered into a rating agreement. Poon (2003) empirically concluded that unrequested ratings were lower on average than the requested ratings, and found that the effect could be explained as self-selection.
The questions connected with the desire of issuers to use rating shopping to obtain the best ratings were developed to overcome the difficulties to apply ratings for regulatory aims (Cantor and Packer 1994; Karminsky and Peresetsky 2009).
A lot of studies have analyzed the reasons for differences in ratings from different agencies rather than constructing a mapping between the different scales. Liss and Fons (2006) compared the national rating scales supported by Moody’s with its global rating scale.
Ratings have also been compared in Russia by some authors (Hainsworth et al. 2012), according to Russian bank ratings connected both national and international agencies. Matovnikov (2008) looked at the relationship between the gradations of rating scales and the total assets and capital of banks. Hainsworth used an iterative application of linear regressions to find mappings between the rating scales of all the credit rating agencies.
A wide array of literature on rating modeling uses econometric models; for example, for bank ratings (Caporale et al. 2010; Iannotta 2006; Peresetsky and Karminsky 2011). Typical explanatory variables from publicly available sources have been defined for models of ordered choice. Examining changes in rating gradation over time for a limited sample of international CRAs was fulfilled.
The selection of the explanatory variables is an important step for the elaboration of such models. Firstly, quantitative indicators that are employed by the rating agencies may be examined (see, for example, Moody’s 2007), as well as nonconfidential indicators that have previously been employed by other researchers. Typical informative indicators are connected with the CAMELS classification and include the size of the company, its profitability, stability, liquidity, and structure of the business, as expressed through companies’ balance-sheet figures. In recent years, the use of such factors as state support for banks or companies, and support from the parent company or group of companies has also become more frequent.
Secondly, the use of macroeconomic indicators has become popular recently (Carling et al. 2007; Peresetsky and Karminsky 2011). Among the most common indicators there are inflation index, real GDP growth, industrial production growth and oil prices, and changes in the foreign exchange cross-rates of currencies for export-oriented countries. Because of the correlation between the majority of macroeconomic indicators they may be used mostly separately. Thirdly, the potential efficiency of market indicator exploration (Curry et al. 2008) for public companies should be mentioned. It should also be noted that alternate indicators may be informative for developing and developed markets.
At the Higher School of Economics and the New Economic School in Moscow there has been research on modeling the ratings of international credit rating agencies in Russia (Peresetsky et al. 2004; Karminsky et al. 2005; Peresetsky and Karminsky 2011). These studies have focused on finding economic and financial explanatory factors, that affect ratings, and on comparing the ratings of international agencies.