Conflict of Interest in Investor-paid Credit Ratings

After the 2008 global financial crisis, most of the criticism about credit rating agencies (CRAs) has been revolving around the conflict of interest inherent to their issuer-paid business model: The very issuer that is being evaluated for its creditworthiness is also paying for the rating service. An alternative business model, often considered in debates, is an investor-paid model. The premise of the latter is that it does not present the same problem of conflicting interests because the CRA is being compensated by the investors rather than the issuers.

In this paper, I provide economic arguments as well as empirical support to show that the regulatory environment could still lead to a rating inflation problem even under the investor-paid model. That is, even if investors pay for the rating service, a conflict of interest might still arise, this time not with the investor but with the regulator. To make a case for my arguments, I focus on the credit ratings of corporate bonds market in the United States.

The proposed economic mechanism behind such conflict of interests has three main ingredients: First, investors’ propensity to reach for yield; second, the investors’ demand for relief from rating-contingent capital requirements; and third, the ability of the investor-paid CRA to supply such relief to investors.

Recent studies have found strong evidence that investors tend to reach for yield, ie, they have a propensity to take more risk by acquiring bonds with higher yield, which are often riskier, in each credit rating category. Behavior of asset managers and a low interest rate macroeconomic environment have been considered as the main drivers of such behavior. But this reaching for yield behavior is constrained by capital requirement regulations. Many large institutional investors are required to hold more equity capital if they take more risk. However, the evaluation of risk is still based on credit ratings of the assets that investors hold. Insurance companies in the United States, who are among the largest buyers in the corporate bonds market, are subject to such regulations. These regulations, set by NAIC, extensively use credit ratings of several CRAs to calculate capital charges.

The above constraint could induce investors to demand inflated ratings because it could provide them with regulatory relief. A higher rating assigned to a higher yield bond means lower capital charges, rendering them less expensive to hold and less likely to be subject to a fire sale. Given such demand for regulatory relief, the next question is whether an investor-paid CRA actually caters to investors. This question must be addressed empirically.

In the paper, I perform an empirical investigation of credit rating behavior of Egan-Jones Rating Company (EJRC). EJRC is an investor-paid CRA approved by the SEC and an official provider of credit ratings used by the National Association of Insurance Commissioners (NAIC) for risk-based capital regulation of insurance companies. The empirical tests indicate that, indeed, EJRC assigns more favorable ratings to those bonds with a higher yield. The association is stronger for bonds that are traditionally targeted by insurance companies.

Further, I investigate whether we can observe a systematic difference between the rating behavior of EJRC and an issuer-paid CRA (namely Moody's). The test reveals that for those bonds that have received ratings from both types of CRAs, bonds with a higher yield are assigned a better rating by EJRC. The results are stronger if a better rating could push the bond to a risk-category with lower capital requirements.

If one single conclusion can be drawn based on this research, it is that rating-contingent financial regulations could lead to rating inflation, regardless of compensation structure of the rating agencies.

Nima Fazeli is an assistant professor of finance at the Paris School of Business.