Competition Among the Big Three: How Multiple Credit Ratings Pay Off for Investors

Asymmetric information is an important characteristic of the securitization market, where products exhibit complex architecture and information about the underlying credit portfolio is highly opaque. In order to overcome these information asymmetries, issuers use rating agencies who act as agents and provide a credit rating at tranche-level for each issued security. In fact, based on our data, most of these tranches were rated by more than one rating agency and few of the tranches had been downgraded by the beginning of 2008. Ultimately, investors have to bear the costs of multiple ratings in the form of lower interest rates on the tranche. But what is the benefit of paying a second or third rating agency, when all of them failed to predict adverse events such as the financial crisis? In a recent study, we provide empirical evidence for why it makes sense for investors to support their investment decision with ratings from more than one credit rating agency.

A rather prominent example in this context is the market for U.S. residential mortgage-backed securities (RMBS). These securitization transactions have been at the forefront of the financial crisis as their underlying portfolios consist of U.S. mortgages, which lost substantial value as a result of the bursting of the U.S. real estate price bubble. This development, alongside misaligned incentive schemes and flawed regulation, triggered a very high portion of defaults. Investors heavily depend on rating agencies to better understand the complex transaction structures, which can involve several layers of securitization. In an efficient market environment, however, one rating agency should suffice to fulfill the monitoring function on behalf of the investor base. We know little about why the vast majority of issuers, and ultimately investors, are willing to pay for more than one credit rating.

Some scholars argue that more ratings reduce uncertainty about the underlying credit quality of a security. As investors are adverse to uncertainty, issuers may apply for additional ratings due to the demand for increased information production. Investors are interested in additional information, since it allows them to better assess the credit quality of the underlying debt instrument. Rating agencies may also apply different models or specialize in evaluating particular drivers of default and might thereby develop comparative advantages to justify their existence. Thus, the advantageous effect of rating agencies’ different perspectives is expected to provide additional information on the uncertainty associated with credit quality and default probabilities. But does this argument really pay off for investors? Do additional ratings really lead to more and better information?

Based on 28 years of rating migration data on all U.S. RMBS tranches ever rated by Standard & Poor’s, Moody’s, and Fitch, our findings confirm that multiple ratings are indeed of some avail to investors: We find empirical proof that rating agencies demonstrate more effort with respect to their monitoring activities when tranches have assigned more than one credit rating. Rating agencies publish more reports and comments and we find that it is more likely that they become active in the light of deteriorating market conditions than in the case of single-rated tranches. For tranches which eventually entered into default, this difference in probabilities increases particularly between 2002 and 2006. Thus, additional outstanding ratings induce credit rating agencies to work harder for their money. As a result, investors get on average more information from each rating agency, or in other words, competition matters.

The underlying default probabilities of their assets are of primary concern to investors. The next question therefore is whether higher levels of revision effort on behalf of rating agencies also relate to improved predictive power of credit ratings. More precisely, one should be able to observe an improvement in rating agencies’ ability to correctly discriminate between different levels of credit risk. Indeed we show that rating agencies not only publish more, but also more accurate information in case of multiple ratings, as average default prediction accuracy is significantly higher compared to single-rated tranches.

In supplemental tests, we find that disagreement between rating agencies increases over a tranche’s lifetime. On average the predicted pairwise disagreement widens by about 2-3 rating categories during the first three years after issuance. Based on our observation of the divergence over time, Moody’s has a tendency to publish the most pessimistic credit assessments. The systematic and time-persistent differences in rating levels provide a strong motivation for issuer clients to engage in rating shopping activities. Issuers might be tempted to exploit the inconsistent credit assessments of different rating agencies in order to maximize the rating of their securities. We argue that by restricting its focus only on the point of issuance, existing empirical research underestimates the incentives for rating shopping, which are likely to be also driven by issuer’s expectations about relative future rating migration.

Finally, we report a rather devastating result with regard to the overall performance of RMBS transactions. Out of the 154,608 tranches in our sample, about one third was rated in or close to default during the recent financial crisis, peaking at a default rate of three quarters for tranches issued in 2007.

The results of our analysis shed light on the performance of rating agencies and offer insights for investors and regulators for future investment decisions. From a regulatory perspective, we provide empirical evidence that a multiplicity of ratings reduces information asymmetries and lowers overall industry opaqueness. Additional ratings therefore increase market transparency and regulators should support initiatives to foster competition between rating agencies. The economic implications suggest that more disciplined behavior on behalf of rating agencies among single-rated tranches could have contributed to a more timely and accurate prediction of about 110 billion USD of potential losses to investors.

The preceding post comes to us from Stefan Morkoetter, Professor at the University of St. Gallen, Roman Stebler, PhD Student at the University of St. Gallen, and Simone Westerfeld, Professor at the University of Applied Sciences Northwestern Switzerland. It is based on their recent article entitled “Rating Agencies and Information Efficiency: Do Multiple Credit Ratings Pay Off?”, which is available here.