Summary & Background

At the request of Congress, the General Accounting Office (GAO) completed a study, Insurance Ratings: Comparison of Private Agency Ratings for Life/Health Insurers, covering the period between August 1989 to June 1992.

The GAO established a standard five-band rating scale and favorably compared the performance of Weiss ratings for life and health companies to that of the four rating agencies --A.M. Best, Duff & Phelps, Moody's, and Standard & Poor's. The GAO concluded that:

• Among all five rating agencies, Weiss was the first rater to assign "vulnerable" ratings in five of the six large insurance company failures.

• Weiss' ratings were first to reflect financial vulnerability three times more often than Best for insurance companies that became financially impaired or insolvent.

• Weiss' ratings warned of financial vulnerability an average eight months earlier than Best. In a majority of the cases studied, Weiss' warnings were issued 15 months ahead of Best.

• Of the 30 insurers that became insolvent or financially impaired and that were rated by both Weiss and Best during the study period, Weiss was first to assign "vulnerable" ratings 23 times. Best was first seven times. In four of the 30 cases, Best did not rate an insurer "vulnerable" until after the state officials took action.

This report updates and expands upon the GAO's conclusions, providing 29 months of additional data plus comparisons of failure rates.

Part I. Update of GAO Findings

In Part I, the GAO's primary conclusions are reviewed and updated for the period since the GAO study --- July 1992 through November 1994. During that period:

• Weiss was first to issue "vulnerable" ratings to five out of six companies that later became financially impaired or insolvent, while Best gave prior warning on one.

• In comparison to S&P Qualified Solvency Ratings ("Qualified Ratings"), Weiss was first in 13 cases, as compared to three in which S & P was first.

• Focusing on the five largest failures which have occurred since the GAO study, only the Weiss ratings and the S&P Qualified Ratings gave advance warnings. The Weiss ratings were first, with an average of 475 days advance warning; and S&P Qualified Ratings were second, with an average of 314 days.

• Weiss was also the quickest to reflect improvements in financial health. For example, after the large company failures in 1991, many large insurance companies strengthened their balance sheets. During this period, the other rating agencies announced a preponderance of downgrades. Weiss was the only rating agency that issued more upgrades than downgrades, reflecting improvements in a timely manner.

Part II. Comparison of Ratings Distributions

Critics of the GAO study argue that it considered only insurance companies that ran into trouble, but did not examine whether a sound company got an appropriate rating. These critics further claim that Weiss rated "too many" companies as "vulnerable" and therefore had a statistical advantage on the GAO's tests.

However, as evidenced by widespread downgrades by the other agencies in 1991-93, the industry as a whole was weaker in 1990 than the traditional rating systems had recognized.

At the time, the data in this study demonstrate that the performance of the Weiss ratings cannot be explained by differences in the ratings distribution.

Subsequent to the GAO study period, Best modified its scale and definitions in a way that resulted in a percentage of "vulnerable" companies very similar to that of Weiss. Best requested that the GAO apply this new retroactively to the 1989 period, but the GAO rejected this request, because the GAO believed "a consumer would not understand a ‘good' insurer to be ‘vulnerable.'"

The GAO adds that "if we had placed Best's ‘B' and ‘B-' ratings in the ‘vulnerable' category, Weiss would still have been first overall," by a margin of two to one. Thus, even with a very similar percentage of companies rated "vulnerable," Weiss was first most often in assigning a "vulnerable" rating.

Also subsequent to the beginning of the GAO study period, S&P introduced its Qualified Ratings which had a higher percentage of "vulnerable" ratings than Weiss. Nevertheless, these were second to Weiss in identifying "vulnerable" companies that later failed or became financially impaired.

Part III. Insurance Failure Rates Compared to Bond Default Rates

The other rating agencies also criticized the GAO study, proposing that the GAO should have presented the percentage of life/health insurers in each category that later became financially impaired. In Part III, we respond to this proposal by comparing the failure rates associated with the ratings of these agencies -- S&P, Best, and Weiss -- to corporate bond default rates. The results show that only the Weiss ratings are in line with widely accepted risk standards.

At S&P, the failure rates for insurance companies rated "AAA" and "AA" was roughly 40 times greater than the corresponding default rates for similarly rated corporate bonds. Also, the failure rate for insurance companies rated "A" and "BBB" by S&P is approximately equivalent to the default rate for junk bonds.

A.M. Best ratings exhibited a similar, but smaller, discrepancy. The failure rates form their "A+" (Superior) rated companies was roughly equivalent to that of BBB bonds, one grade level above junk bonds.

These risk levels are considered unacceptable for high rated companies.

Among all the rating agencies for which there is sufficient data, the failure rates associated with Weiss' insurance ratings show the closest correspondence to the default rates of equivalent bond ratings.

Part IV. Conclusions

With the GAO study and the update, one can now make a clearer distinction between two types of systems -- voluntary ratings and involuntary ratings.

Voluntary ratings include S&P's Claims Paying Ability Ratings, the ratings of A.M. Best and Duff & Phelps, and most of the ratings issued by Moody's. These ratings are issued at the request of the companies, in exchange for fees. In addition, these ratings play a very important role in the business of selling insurance.

We believe that, in these systems, the need to obtain company cooperation -- plus the agencies' dependence on the companies for revenues -- introduces an unavoidable conflict of interest into the rating process. Insurance companies are empowered to shop for the most liberal ratings and are usually allowed to withdraw from the process when they feel that the public disclosure of their problems would hurt business.

These conflicts of interest -- combined with the critical role that the ratings play in insurance sales and advertising -- can skew rating scales and bias the ratings. We believe this is the primary reason why these ratings systems have consistently failed to provide adequate advance warnings of vulnerability.

In contrast, the Weiss ratings and S&P Qualified Ratings are involuntary -- issued without company approval and without collecting fees. These systems have greater freedom to issue "vulnerable" ratings and have no reason to delay downgrades. In our opinion, this largely explains why the involuntary ratings are the only ones that have consistently provided advance warning of vulnerability.

The situation at S&P is especially illuminating: Their involuntary Qualified Ratings have outperformed their own voluntary Claims Paying Ability Ratings, despite the fact that S&P expends far less resources per company on the Qualified Ratings than it does on the Claims Paying Ratings.

Thus, we believe the presence or absence of bias in the rating process is the overriding factor determining rating accuracy and timeliness.

Unfortunately, consumers are not generally aware of the conflict-of-interest questions. They are not even informed about the relative position of each rating in the scales of most rating agencies.

The GAO study, by establishing a standard for comparing different rating scales, has made a significant step toward clearing up this confusion. But this information is not made available to consumers. And there has been little information available regarding the actual failure rates associated with each rating category. Consequently, it has been difficult -- even for sophisticated insurance consumers -- to adequately evaluate risk levels of insurers.

The GAO study plus this report can help reduce the information gap about rating systems and provide all insurance consumers with a firmer basis for making informed decisions.

We propose that the GAO's five-band scale, plus the associated failure rates for each rating, should be clearly disclosed to consumers in all insurance advertising and sales.