Does the Nationally Recognized Statistical Rating Organization certification matter for Japanese credit rating agencies?

Highlights

• The designations of NRSROs increase the market share of new NRSROs in Japan.

• Newly certified NRSROs’ bond ratings are more informative for predicting yields.

• NRSRO designations enhance the reputation of Japanese rating agencies.

Abstract

The SEC’s NRSRO designation for Japanese credit rating agencies is suitable for examining certification and monitoring effects, as Japanese domestic bond markets are not subject to SEC regulations. We find that the certification increased the market share of new NRSROs compared to that of incumbent NRSROs. Moreover, bond issues rated by newly certified NRSROs obtained lower yields and their ratings became relatively more informative for predicting the yield spread after the NRSRO designation. Our findings suggest that the NRSRO designation enhances the reputation of newly certified NRSROs and the value of their ratings information.

Introduction

We use the Nationally Recognized Statistical Rating Organization (NRSRO) designations of the U.S. Securities and Exchange Commission (SEC) on Japanese credit rating agencies (CRAs) to identify the certification and monitoring effects of the NRSRO designation on Japanese CRAs. SEC has been highly selective with its NRSRO designation and granted it only to Moody’s, S&P, and Fitch until recently.¹ After Enron’s collapse, however, the U.S. Congress passed the Credit Rating Agency Reform Act of 2006 (CRARA) to promote competition among NRSROs. In May 2007, out of political pressure under the CRARA, the SEC issued NRSRO certifications for seven CRAs including two Japanese CRAs: Rating & Investment Information (R&I) and Japan Credit Rating Agency (JCR). Moreover, the law specifically prohibits the SEC from regulating an NRSRO’s rating methodology.

Given that Japanese CRAs assign higher ratings than U.S. CRAs (Shimoda and Kawai, 2007) and Japanese CRAs’ rating fees are lower than U.S. CRAs’ (Han et al., 2019), issuers may find advantages in hiring newly certified NRSROs. Consequently, some issuers were able to switch to new NRSROs with less uncertainty about their prospective ratings, increasing the market share of new NRSROs. Furthermore, S&P and Moody’s reputations have been deteriorating following failures of their ratings to detect a series of corporate collapses, such as Enron and WorldCom. The advantage of the new NRSRO certification depends on the CRA’s effort to reap subsequent opportunities (Merton, 1973) and the incumbent NRSROs’ efforts to maintain their competitive standing (Bruno et al., 2016).

Yet, the ratings of Japanese yen bonds are not subject to the regulatory rules of the SEC. More importantly, the Financial Service Agency (FSA, the Japanese equivalent of the SEC) has abolished regulations that specify rating-based “eligibility standards” since 1996 (Japan Bond Market Guide, 2012, p. 14). Thus, the NRSRO designation is unlikely to show regulatory effects but will reflect the market participants’ recognition of the certification and monitoring effects. Moreover, unlike U.S. issuers who almost uniformly choose both Moody’s and S&P with variation only in their choice of additional ratings (Bongaerts et al., 2012), Japanese issuers show much greater variation in their CRA choices. Thus, the Japanese market provides a clean setting for investigating the certification/monitoring effects of the NRSRO designation on the information value of ratings and the competitive standing of CRAs.

If the market perceives NRSRO certification as quality assurance with continuous monitoring by the SEC, more issuers would switch to newly certified CRAs, intensifying competition among CRAs. With more competition, CRAs will be more concerned about maintaining their reputation, which will in turn improve the ratings quality. This was the intention of the CRARA. However, a major criticism of the NRSRO designation is the anti-competitive nature of the regulatory barriers. Moreover, it remains controversial if more competition in the ratings industry improves the quality of ratings (Becker and Milbourn, 2011; Bae et al., 2015). Our study contributes to resolving these issues by investigating the effects of NRSRO certification on competition among CRAs and the information content of their ratings. Our study also contributes to understanding the role of government certification for the behavior of information intermediaries, which has important implications for regulators and investors.

We use a sample of new Japanese yen-denominated corporate bonds between 2001 and 2011 to test for changes in market share and the information value of ratings by newly certified NRSROs. In particular, we use the NRSRO designations of R&I and JCR in 2007 as quasi-natural experiments in the difference-in-difference (DiD) approach. In Japan, R&I has been the largest CRA in terms of market share; this is a unique situation from a global perspective since S&P and Moody’s usually hold a stronger position than local CRAs. Arikawa et al. (2020) suggest that split ratings between Japanese and U.S. agencies result from systematic differences in the assessment of credit risk determinants, such as business risk and company-specific risk, but not from upward rating bias by the Japanese rating agencies.

Our results show that new NRSRO certifications have significantly increased the market share of new NRSROs (R&I and JCR) relative to that of incumbent NRSROs (S&P and Moody’s). Interestingly, after its NRSRO certification, JCR rose from having the lowest market share in Japan to today being the second dominant player. Our results suggest that the market share of our new NRSROs grew by 61%–67% after their certifications. Such increases in market share are not due to rating shopping by issuers (Skreta and Veldkamp, 2009; Farhi et al., 2013) or catering by CRAs (Griffin et al., 2013). Moreover, yield spreads (bond yield minus maturity matched Japanese government bond yield) for bonds rated by newly certified NRSROs decreased by 23–25 bps relative to those of incumbent NRSROs, representing economically significant information effects as well. These findings suggest that the ratings by the new NRSROs become as informative as those by the incumbent NRSROs. Additionally, the results show that the yield spread after NRSRO designation is lower by 11–12 bps for bonds rated by R&I and JCR, whereas this change is insignificant for bonds rated by S&P and Moody’s. The theoretical literature suggests that a certification allows a less reputed CRA to free ride on the reputation of other certified firms (Polidoro, 2013), which implies that the effect of NRSRO certification on market share is greater for JCR than for R&I. Indeed, our tests show that JCR has gained about 26%–27% more market share than R&I.

Further, we show that issuers with multiple ratings have switched to the newly certified NRSROs (JCR and R&I) from the incumbent NRSROs (S&P and Moody’s). The probit regression estimates suggest that the probability of issuers selecting the pair of newly certified NRSROs has increased by 20%, while that of choosing the pair of incumbent NRSROs has decreased by 4%–6%. This evidence further suggests that issuers recognize new NRSRO certifications. The results remain intact even after excluding the 2007–2009 financial crisis period or unsolicited ratings. The results are also robust after the inclusion of Fitch’s ratings.

We also check for the potential regulatory effect of the NRSRO designation on Japanese CRAs. Bruno et al. (2016) suggest that due to the regulatory requirements, either issuers choose a CRA that is most likely to rate their issues with investment grade or the CRA is reluctant to issue a speculative grade across this threshold, resulting in significant kinks in the distribution of ratings at the investment grade threshold. We find no evidence of such kinks for Japanese corporate bond ratings over our sample period. Moreover, our results show no significant differences in the NRSRO effects between smaller and larger issuers and between issuers with more and less foreign ownership or sales.

We also investigate the effect of R&I’s withdrawal from the NRSRO certification in 2011. If the NRSRO status is important to attract U.S. investors, R&I’s withdrawal should materially reduce Japanese issuers’ demand for R&I’s ratings. We find that R&I’s withdrawal did not reduce its relative market share. We do find, however, that the bond yields for issues rated by R&I relative to issues rated by other NRSROs significantly increased after its withdrawal from the NRSRO certification. This finding is consistent with the argument that investors attach more informational value to issues rated by NRSROs because the SEC’s monitoring enhances their transparency and reduces the conflict of interest. We also check the effect of Japanese regulatory rules that may potentially confound with the NRSRO designation. The results show that the NRSRO designation of Japanese CRAs has unique effects.

Two important implications arise from our findings. On the one hand, the increased market share of newly certified CRAs suggests that NRSRO certifications make it easier for issuers to switch from incumbent NRSROs to the newly certified NRSROs by removing uncertainty about less-known CRAs. On the other hand, the lower yields associated with bonds rated by the newly certified NRSROs imply that the SEC’s monitoring enhances the information value of their ratings.

Our study makes several important contributions. This is the first study that examines the effects of NRSRO certification on non-Anglo-American CRAs. In particular, we examine whether the NRSRO designation has any bearing on the Japanese market. Our findings suggest that the SEC plays a significant role through its NRSRO certification program, even in a non-U.S. market.

Second, we improve the empirical identification of the effects of SEC certification and monitoring of CRAs. Previous studies provide mixed evidence on the effects of an NRSRO designation. In particular, whether the SEC enhances CRAs’ credibility through its NRSRO certification and monitoring or entrenches them with its regulatory requirements remains unresolved. This study contributes to resolving this issue by showing the effects of NRSRO certification/monitoring in isolation from regulatory effects.

Third, our study provides tests for the competitive implications of the NRSRO certification. Uncertainty over a CRA implies high switching costs (Schmalensee, 1982) for bond issuers. The NRSRO certification resolves this uncertainty. After the certification, issuers change their extrapolation of the new NRSRO’s service in such a way that competing NRSROs move closer in quality due to the anti-lemons effect (MacLeod and Urquiola, 2015). Consequently, new NRSROs gain more market share. Our findings are consistent with this argument.

Fourth, this study provides an alternative explanation to the findings of Han et al. (2012), who document that Japanese bonds rated by Moody’s and S&P (NRSROs) had lower yields compared to those rated by R&I and JCR (non-NRSROs), which disappeared after 2007. They speculate that the difference reflects the shocks to S&P’s and Moody’s reputations. Our findings suggest that the NRSRO certifications of R&I and JCR explains this documented disappearance of the yield differential between issues rated by Japanese and U.S. CRAs. In conclusion, the NRSRO certification affects the competitive environment in the credit rating industry and the information value of the ratings beyond the regulatory effects.

The rest of our study is outlined as follows. In Section 2, we review the relevant literature, and in Section 3, we describe our empirical methods. In Section 4, we provide details on our sample and univariate results. In Section 5, we report the multivariate results and in Section 6 we provide robustness checks. Finally, in Section 7, we provide conclusions.