The rating agencies provide an initial rating for a debt issue and monitor that issue during its life. Rating agencies base their decisions, in part, on publicly available data about the issue, the company, the industry and the economy. In addition, most agencies visit the business, interview key managers and obtain private information about performance, budgets, plans and forecasts. Some rating agencies employ statistical classification models, although ultimately, the rating is a judgment based on quantitative and qualitative data.
Most agencies give the organization advance notice of a proposed rating or proposed rating change and allow management an opportunity to respond. The ratings and rating changes are carried by financial newswires such as Reuters and PR Newswire, displayed on trading screens (such as Bloomberg and Telerate), reported in regular publications of the agencies and reviewed in the financial press.
Originally, rating services derived their revenue primarily from fees charged to subscribers to the rating bulletins. Now the revenues of US agencies come almost entirely from fees charged to the issuer of the security, although subscription fees are still important in some other markets outside the US. Typical issuer fees include an initial fee based on the size and complexity of the issue, together with monitoring fees. Both S&P and Moody’s have a policy of publishing ratings for all large corporations with significant outstanding debt, even if the issuer does not solicit the rating. Presumably, these unsolicited ratings enhance the comparability of solicited ratings.
Some controversies
The 1970s were difficult for the rating agencies. Criticism came from politicians, issuers and investors and the rating agencies were threatened with government supervision. The agencies are always tempting political targets for politicians from cities and districts that face higher borrowing costs as a result of an ‘unjustified’ (in the politicians’ view) bond rating downgrade. For example, the New York City financial crisis in the 1970s placed the agencies in a predicament. At one stage, S&P suspended the ratings on city debt and was vilified for causing, or at least exacerbating, the
subsequent financial chaos. On the other hand, Moody’s retained its A rating (perhaps attempting to counteract its ‘me too’ image). The bonds eventually defaulted and investors criticized the relatively safe rating assigned by Moody’s. In another example, the Australian government forbade any contact with Moody’s after losing its AAA rating in 1986.
Other spectacular financial collapses led to accusations that the agencies respond to bad news too slowly, perhaps out of concerns for the issuer who pays the rating fees. The most often cited examples of the alleged tardiness involve two instances where prominent companies (WT Grant and Penn Central) declared bankruptcy when their debt was highly rated and that of the real estate investment trusts and insurance companies in the 1970s. In the latter case their ratings were not changed until they experienced severe financial difficulties.
The corporate restructuring wave of the 1980s brought home a limitation of bond ratings that, although freely acknowledged by the agencies, had been ignored by many investors. Bond ratings do not reflect the vulnerability of the bond to what is known as ‘event risk’ — extraordinary changes in the financial or operating characteristics of a business. For example, some bonds are poorly protected by covenants. The default risk of those bonds changes dramatically if a company takes on massive quantities of additional debt or if it is acquired by a much riskier business. In several highly publicized instances, holders of highly rated debt experienced large losses when these events occurred. Despite disclaimers by the rating agencies, many critics attributed these losses to inadequate investigation by the agencies.
Similarly, the financial engineering innovations of the 1990s created securities with complex option-like features (such as mortgage-backed securities and their components) whose value could fluctuate dramatically even if their default risk was minimal. Currently, the antitrust division of the Justice Department is investigating the ratings industry for evidence of anti-competitive practices, apparently focusing on the issuing of unsolicited ratings.
In response to some of these criticisms, the rating agencies have greatly expanded their staffs, adopted new technology, worked at improving the timeliness of their ratings and stepped up their public relations efforts. In particular, in 1981 S&P introduced Credit Watch, which provides an early warning of a rating revision. Other major rating agencies quickly followed with similar services. All of the large agencies now have electronic services that notify subscribers of potential and actual ratings and revisions. The rating agencies have expanded not only the range of claims they rate but also, in response to criticisms discussed above, the types of ratings they issue. For example, S&P now issues supplementary ratings about event risk, designated E1-E5. These ratings are based on an analysis of the protection afforded by covenants, collateral and other contractual features of the issue (such as a put feature). They reflect judgment about an issue’s susceptibility to loss if a deleterious event occurs but not the likelihood of such an event happening.
As public debt markets developed and expanded outside the US, S&P has grown internationally with an aggressive acquisition and affiliation strategy and both S&P and Moody’s have opened local offices throughout the world. Foreign rating agencies have not penetrated the US market to any great extent, perhaps reflecting the barriers created by the NRSRO certification process.
Fig.1 Industrial yields. | m | ||||||
July 1988 - December 1994 | |||||||
AAA | AA | A | BBB | BB | B | T-bond | |
Interest coverage | 13.90 | 9.40 | 4.50 | 2.70 | 1.60 | 0.80 | |
Average yield (%) | 8.49 | 8.81 | 9.26 | 9.74 | 10.96 | 12.86 | 7.96 |
Std dev yield (%) | 0.88 | 0.87 | 0.84 | 0.94 | 1.21 | 2.34 | 0.94 |
Average spread (bp) | 53.40 | 84.90 | 130.40 | 178.30 | 300.50 | 489.60 | |
Std dev spread (bp) | 18.20 | 18.40 | 20.40 | 35.40 | 67.00 | 194.40 | |
January 1983 - December 1994 | |||||||
AAA | AA | A | BBB | BB | B | T-bond | |
Interest coverage | 13.80 | 9.20 | 5.10 | 3.00 | 2.00 | 1.20 | |
Average yield (%) | 9.43 | 9.79 | 10.14 | 10.70 | |||
Std dev yield (%) | 1.53 | 1.53 | 1.47 | 1.58 | |||
Average spread (bp) | 49.00 | 85.00 | 120.50 | 176.80 | |||
Std dev spread (bp) | 25.40 | 28.00 | 29.80 | 35.40 |
Table captionSource: Standard and Poor's CreditWeek
Evaluating the performance of rating agencies requires an appreciation of their economic purpose. In principle, the agencies have at least two roles: information processors and certifiers. The first role is primarily market driven; the second is primarily driven by regulation.
If there are economies of scale and special skills or training associated with information gathering and processing, rating agencies can reduce search costs by providing investors with information about the default risk of bonds, just as restaurant guides can provide value to discerning diners. Moreover, published debt ratings reduce the information costs ’of others who contract with an organization (such as suppliers, customers and employees) and are concerned about the company’s financial health. In particular, ratings can reduce the costs associated with evaluating counterparty risk in over-the-counter derivatives transactions.
Certification demand is driven by laws and regulations that restrict the investment behavior of financial institutions, such as those described in Table 2. Certification demand is partly market driven because it reduces the cost of writing contracts by providing a convenient summary statistic (the bond rating, akin to a restaurant’s rating in the Michelin guide.) For example, security ratings allow a corporation to restrict the behavior of someone in charge of short-term cash management to investing in securities with at least a given rating.
Academic research has emphasized the information content of bond ratings, the accuracy of bond ratings and the extent to which ratings can be predicted with publicly available information.
Timeliness of the information is more crucial with regard to information than for certification. Rating changes provide information to investors only if that information is not already incorporated in the price of the security. Research demonstrates that the announcement of a rating change for the debt of an organization affects the price of the
Evaluating the bond-rating agencies
Fig.2 Industrial yield spreads over
■m
:
_
Basis points 1200
183 184 185 186 187 188 189 190 191 192 193 194 195
Source: Standard and Poor's CreditWeek
debt and the company’s equity. The effects are stronger for downgrades than for upgrades, though the effect is relatively small (less than two per cent of the market value for the two-day period around the announcement). Furthermore, the rating changes follow much larger changes in the price of the security, suggesting that much of the information leading to the rating decision is incorporated in the security’s price by the time the change is announced.
Of course, if it were not for the looming presence of the rating agencies, some of that earlier information might not be released as quickly. Does this mean that the agencies act too slowly? Research has yet to answer that question because the optimum speed of action depends on unknowns such as the costs and benefits of gathering and disseminating information more rapidly. However, competition among rating agencies should provide incentives for the agencies to provide more timely information if it is cost effective. Some of the innovations such as Credit Watch and the increased use of electronic dissemination of rating changes are examples of agencies reacting to their perception of market demand. The accuracy of bond ratings has been addressed by studies of bond yields and returns (see Figures 1 and 2). Yields on bonds are highly correlated with bond ratings. Yields vary within rating classes, providing evidence that investors do not accept ratings mindlessly, though there is not enough evidence to address the question thoroughly.