Literature on Unsolicited Ratings

Poon (2003) was the first author to empirically analyze unsolicited credit ratings, using S&P long-term credit ratings of 265 corporations in different industries across 15 countries from 1998 to 2000. Poon (2003) found significant self-selection bias in the rating decision. Unsolicited ratings were lower than solicited ratings, and profitability and sovereign credit risk were the two major factors in determining long-term corporate credit ratings. Byoun and Shin (2003) used the unsolicited and solicited ratings of non-US corporations between 1996 and 2002 to study the effects of solicited and unsolicited ratings on firm value. They found that stock prices fell after downgrades of unsolicited ratings and that there were positive market reactions to upgrade announcements. In contrast, for solicited ratings, they found only significant positive market reaction to upgrade announcements.

Butler and Rodgers (2003) examined a sample of 360 corporate bond ratings of 153 nonfinancial companies from Moody's and S&P during 1997. Their results suggest that when relationships existed in the assignment of solicited ratings compared to unsolicited ratings, rating agencies relied less on publicly available “hard” information, and were better able to assess “soft” (or private) information about bond issuers. Using a similar methodology to Butler and Rodgers (2003), Gan (2004) distinguished unsolicited bond issues from solicited bond issues using the rating fees reported in the registration statement. Using a sample of S&P and Moody's credit ratings assigned to 1,410 bond issues of 303 firms between 1994 and 1998, Gan (2004) examined whether Moody's and S&P used consistent standards in solicited and unsolicited issue ratings, and empirically tested two hypotheses—the punishment hypothesis and the private information hypothesis. The results of Gan's study showed that both agencies gave significantly lower ratings to unsolicited issues, but there was no significant difference between the performances of solicited and unsolicited groups. However, Gan (2004) believed that these rating agencies gave lower ratings to unsolicited issuers not as blackmail or punishment, but as a necessary adjustment for the difference in the true and unobserved quality, such as for private information.

Poon and Firth (2005) used an international sample of 1,060 bank ratings from 82 countries during 2002 to analyze shadow ratings,² which were based largely on public information, to shed light on the controversy surrounding unsolicited ratings. Their results indicated that shadow ratings were lower than nonshadow ratings and that banks that received shadow ratings were smaller and had weaker financial profiles than the other group. This suggests that bank size, profitability, asset quality, liquidity, and sovereign credit risk are important in determining bank ratings. While Poon and Firth (2005) applied Heckman's two-step estimation method (Heckman 1979) to their estimation model of bank ratings, Van Roy (2006) conducted a study of the credit ratings of Asian banks rated by Fitch, using an endogenous switching regression model. He found that unsolicited ratings appeared to be lower than the solicited ratings after controlling for differences in observed bank characteristics, mainly because they were based exclusively on public information. He suggested that unlike solicited ratings, which incorporated both public and private information, unsolicited ratings tended to be more conservative as a consequence.

Using a comprehensive sample of the issuer ratings of 460 commercial banks in 72 countries from S&P for the period 1998–2003, and adopting an endogenous regime-switching model, Poon, Lee, and Gup (2009) investigated whether solicitation mattered in bank credit ratings. They found that solicited ratings tended to be higher than unsolicited ratings. Banks with solicited ratings tended to be larger, have relatively lower ratios of nonperforming loans to gross loans, and have higher returns on equity than unsolicited banks. Their results indicated that the observed differences between solicited and unsolicited ratings could be explained by both the solicitation status and financial profiles of the banks.

Fairchild, Flaherty, and Shin (2009) employed Moody's solicited and unsolicited credit ratings, collected from a previous survey of Japanese firms, as their sample and came to similar conclusions to those of the authors of the previous studies using S&P ratings. That is, unsolicited ratings were still lower than solicited ratings even though firms with unsolicited ratings had provided Moody's with some degree of inside information. Comparing the unsolicited ratings given by S&P and Moody's, the authors did not find a significant difference between the ratings assigned by the two rating agencies. They believed that firms with unsolicited ratings provided incomplete private information to rating agencies and that as a result their ratings were lower.

Shimoda and Kawai (2007) conducted a thorough study of the recent credit rating gaps in Japan, using ratings from two Japan-based rating agencies (Japan Credit Rating Agency and Rating and Investment Information) and three global rating agencies (Moody's, S&P, and Fitch). The two types of rating gaps that they investigated were: (i) differences between solicited and unsolicited ratings, and (ii) rating splits among different rating agencies. Focusing on the credit ratings of nonfinancial corporations in Japan, the authors found that unsolicited ratings were lower than solicited ratings, which was consistent with the results of other studies using international samples, but these differences seemed to be smaller than before. They suggested that the disparity in the level of information available to rating agencies and cherry-picking actions by the issuers (similar to sample-selection bias) might have contributed to the differences in the ratings.

Recently, Behr and Güttler (2008) and Bannier, Behr, and Güttler (2008) conducted empirical studies on the informational content of unsolicited ratings and the explanations for rating differences between solicited and unsolicited ratings, respectively. Behr and Güttler (2008) examined whether the stock market reacted to initial unsolicited ratings and changes in unsolicited ratings for a sample of firms rated by S&P from January 1996 to December 2005. They found significant negative market reactions to these unsolicited rating actions. Hence, the authors believed that unsolicited ratings did convey new information to stock markets as S&P's unsolicited ratings were mainly based on public information. Bannier, Behr, and Güttler (2008) attempted to explain why unsolicited ratings of non-US firms were lower than solicited ratings with adverse selection and strategic rating (e.g., agency conservatism or blackmailing) arguments, using a sample of S&P ratings data from January 1996 to December 2006. The adverse selection hypothesis was not rejected for the full sample. Moreover, they found that the strategic rating explanation appeared to be important in the subsample of banks.

Download this Paper [ PDF 293.6KB| 26 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.