This assignment will examine whether credit rating agencies (CRAs) can be regarded as useful. In order to do so, two given academic papers will be analysed and discussed, supplemented by further academic literature. The first of the two is a working paper by Efraim Benmelech (2017) called “Credit Ratings: Qualitative Versus Quantitative Information”. The second one is called “What’s (Still) Wrong with Credit Ratings” and written by Frank Partnoy (2017).
Before looking at these papers in greater detail, it is important to provide a general definition of CRAs. Gavras (2012, p. 34) defines CRAs as “private companies [which] assess credit risk for companies and governments seeking to take out loans and issue fixed-income securities”. Arnold (2012, p. 190) states that this risk assessment comes in the form of a rating, which represents the CRAs’ opinion of the rated entity’s creditworthiness, i.e. its “ability to repay its debt”. Both Gavras (2012) and Arnold (2012) remark that the rating ultimately affects the price and interest rate paid by the borrower on the debt instrument.
Having generally defined CRAs, the authors of the two given papers also provide their own views. Both Benmelech (2017) and Partnoy (2017) agree with Gavras (2012) regarding the importance of CRAs in today’s credit markets and the strong influence these agencies have. They also agree that CRAs played a major part in the 2007/08 financial crisis. In his paper, Benmelech (2017, p. 24) evaluates rating by S&P Global Ratings Inc. (S&P) and concludes that they are vastly quantitative and, thus, can easily be predicted and substituted “by an algorithm that uses just ten financial variables”. On a similar note, Partnoy (2017) believes that rating methodologies are highly uninformative, that numerous reforms after the financial crisis have failed, and that the unchanged overreliance of investors and regulators on these ratings should be reduced.
2. Initial Summary
Benmelech (2017) states that, in contrast to the key role of CRAs, their method of determining ratings is still unclear. Therefore, the author examines the extent to which credit ratings are quantitative, to assess whether they deliver new information or simply mirror prevailing public data, the latter of which would render CRAs useless. To answer this query, Benmelech (2017, p. 1) analyses “all corporate credit actions issued by S&P Global Ratings Inc. between 1985 and 2015” and evaluates their quantitative extent and evolution during this time period. Given the fact that he uses global data and covers a wide time span of 30 years, the sample size of his study can be regarded as very large. This increases reliability as errors arising from atypical samples are marginalised.
The results of Benmelech’s (2017) calculations show that the proportion of quantitative data incorporated in S&P’s rating has varied widely throughout the examined time period. S&P’s ratings went from being highly quantitative in the 1980s and 1990s to quantitative proportions of approximately 45% in the aftermath of the global financial crisis, representing an enhanced focus on qualitative information. In recent years, however, Benmelech found a significant increase in S&P’s reliance on quantitative data, up to a proportion of 66.3% in 2015. In his opinion, this surge can be regarded a result of the public criticism of CRAs after the crisis, when they were accused of letting factors irrelative to the companies’ creditworthiness affect the ratings. Due to their highly quantitative character, the author shows that today’s S&P ratings can easily be predicted and replaced by a simple algorithm. Thus, he proves that, contrary to their prominence, CRAs are not useful as they barely generate any new information.
This argument is supported by another finding by Benmelech (2017, p. 19), which is the fact that credit “rating standards have become increasingly conservative over time”. This development of conservative ratings during times of credit growth is offset only by more lenient ratings during times of crisis, which indicates that S&P artificially held up their ratings during the financial crisis to avoid downgrades. Since investors and regulators heavily rely on ratings, this artificiality confirms the uselessness of CRAs as they would be better off relying on Benmelech’s calculations.
Like Benmelech (2017) and Hull (2009), Partnoy (2017) argues that CRAs played a major role in the financial crisis because of their misconduct in giving out high ratings for risky financial instruments, such as collateralised debt obligations. Partnoy (2017, p. 1407) further states that CRAs generate no or very little new information, which is due to the “primitive methodologies” they utilise to calculate ratings. In his opinion, neither investors nor regulators should put their faith in such flawed, subjective and unsophisticated methodologies, which is why he does not find CRAs useful. The author also discovers that all reforms implemented since the crisis have failed, which is why the pre-crisis overreliance of investors and regulators still remains unchanged and substantial today. He focuses his research on reforms in the US, in particular the 2010 Dodd-Frank Act. Analysing three key areas, he shows problems in regulatory reliance, oversight measures and accountability measures, proving the reforms’ uselessness.
In the area of regulatory reliance, the Dodd-Frank Wall Street Reform and Consumer Protection Act (2010) states that US federal agencies should amend legislation to remove references to credit ratings and substitute them with suitable replacements. The purpose of this was to decrease and revoke the reliance of regulators on CRAs. Partnoy (2017) notes that, despite this change, the Dodd-Frank Act is not far-reaching enough and a number of US federal laws still contain references to credit ratings. In the area of oversight, the Dodd-Frank Act (2010) established a number of new measures, most significantly the creation of the Office of Credit Ratings. Its purpose was to oversee CRAs by monitoring their practices as well as examining them annually and reporting on the findings of the examination. According to Partnoy (2017, p. 1417), these reports, however, do not identify violators by name, which is why CRAs do not “suffer reputational costs” and get away with negligible penalties. The author argues that this ineffectiveness continues with regards to accountability where the attempt to revoke CRAs’ privileges of exemptions from liability failed.
Altogether, Partnoy (2017) believes regulators should reverse their current course and thoroughly implement all measures of the Dodd-Frank Act. Additionally, investors should reduce their reliance on ratings and concentrate on fundamental measures with greater informational value.
3. Main Discussion
According to Rhee (2015), there are two contrasting views regarding the usefulness of CRAs. He argues that the supportive view believes CRAs are useful because they bridge the information gap between investors and issuers and because they reduce regulation costs. The opposing view states that CRAs are not useful because they do not create any or very little informational value. The latter view is shared by many commentators, including Benmelech (2017) and Partnoy (2017), who support their views with empirical evidence.
In order to prove the uselessness of CRAs, Benmelech (2017) estimates annual cross- sectional regressions of S&P’s rating determinants for all years of his study and then examines the development of these regressions’ R to discover the explanatory variables’ significance. For this, he chooses ten explanatory variables with regards to the company’s profitability, financial structure, asset composition and margins. Commenting on his study, it should be noted that he only uses operating companies’ ratings in his sample and that he excludes sovereigns, investment holding companies as well as financial institutions. Additionally, the study only covers ratings by S&P which means its findings may not necessarily be applied to all rating agencies in the industry. In my opinion, however, his study nonetheless qualifies as robust evidence, due to the very large sample size. As mentioned before, Benmelech’s calculations show that due to the high quantitative proportion of ratings, CRAs can easily be replaced by simplistic calculations and are, thus, not useful. Partnoy shares this first argument and also finds that the methodologies of establishing CRA ratings are seriously flawed and not only simplistic but also highly arbitrary. His study of rating approaches of the two largest CRAs, S&P and Moody’s, shows the uselessness of CRAs as their methods have not changed since the financial crisis and continue to be extremely opaque. His analysis examines the methodologies used to assess credit ratings in a detailed manner, basing his discussion on information from the CRAs’ US websites. Since he analyses the two largest CRAs, his results are likely to be applicable to other rating agencies as well, which makes his study relevant.
I agree with this view because, as Arnold (2012) mentions, ratings should consist of both quantitative and qualitative factors. I also agree with Gavras (2012) that the basic justification for the existence of CRAs was the fact that they offer a more elaborate approach to measure credit risk, different from simple calculations based on publicly available company data. If this is not the case, there is no need for CRAs. I do, however, also believe that, if CRAs were to change their methodologies and integrate more valid qualitative information in their ratings, they might become useful.
As CRAs are a frequently and vastly discussed subject, not all topical literature has the same view on their usefulness. Thus, it has to be noted that while many academics agree with Benmelech (2017) and Partnoy (2017), there is also a literature in support of the usefulness of CRAs. In this context, Rhee (2015) argues that CRAs have an extremely important function, which he calls sorting function. He explains that by providing a very cost-effective solution for this, CRAs systematise market information and enhance market information, which leads to greater market efficiency. Despite this, in his opinion, useful function, he is aware of the fact that information sorting is very different from information creation, essentially agreeing with Benmelech (2017) that CRAs create very little new information. Moreover, Athanassiou and Theodosopoulou (2015) state that, despite their drawbacks concerning rating quality and objectivity, CRAs have still managed to offer reasonably decent credit ratings over time.
A second important opposing argument discussed widely in the relevant literature is the regulatory reliance on CRAs, which, according to Partnoy (2017) started to increase in the 1970s. Athanassiou and Theodosopoulou (2015) argue that by giving some CRAs the NSRSO (Nationally Recognised Statistical Rating Organisation) status, the US government outsourced some parts of regulation, giving these NRSROs what Rhee (2015, p. 164) calls a “regulatory license”. He further states that, as a result, the government was able to save regulation costs as it did not have to establish its own investment analysis groundwork. In Partnoy’s (2017) opinion, this led to today’s strong overreliance on CRAs in many essential federal laws and, at the same time, helped NRSROs to become highly profitable, despite their low informational value. This argument is shared by Macey (2006, p. 23) who even labels this development of increased importance of and reliance on a small number of NRSROs as “cartelization” of the industry.
 See Gavras (2010); European Commission (2015); Cane, Shamir and Jodar (2012); Devine (2011)