Andrew South, senior director and head of European structured finance research at S&P, argues that rating agencies can offer investors a valuable diversity of opinion

Investors in structured finance securities remain highly focused on triple-A ratings - and, consequently, so do issuers. But with banking system and sovereign risks recently increasing, triple-A ratings are becoming more difficult to achieve. In our view, market participants also tend to pay too little attention to what a triple-A rating for a given security actually means, based on the analysis behind it.

In particular, a security that one rating agency labels triple-A may not achieve the highest rating from another agency. This is because definitions and methodologies can vary significantly from one rating agency to the next.

This raises several questions: is the focus on triple-A ratings in the best interests of investors and issuers? Could a larger market develop for securities with a broader range of credit risk profiles? Should investors always view triple-A ratings from different rating agencies as comparable, or might they benefit from greater diversity of rating opinions?

There are no simple answers. However, in S&P's opinion, the ongoing financial crisis has shown that commonly held views and widespread use of similar methodologies can gradually heighten systemic risks in global markets. Therefore, we believe that wider acceptance of lower-rated securities and greater acknowledgement of different rating methodologies could be positive for the structured finance market.

The financial crisis triggered changes in credit analysis
The 2008-2009 recession and ongoing financial crisis have caused many market participants to reassess how they measure certain risks. Among rating agencies, this has led to a greater diversity in methodologies.

We at S&P have recalibrated some of our rating methodologies, aiming to enhance transparency and ensure comparability of our ratings across sectors. In some areas, we have tightened our ratings criteria. For example, we may look for more credit support and other risk-mitigating features before we consider rating securities triple-A.

Consequently, our assessment of the creditworthiness of some existing issuers and securities is now lower than it was before. In these cases, we have lowered our ratings.

In structured finance, this has led some transaction sponsors to request that we withdraw our ratings on their outstanding transactions, especially in cases where we no longer rated the senior securities triple-A. It may also be harder for transaction sponsors to receive triple-A ratings from us on new issuance.

Some transaction sponsors have cited this as a reason for not engaging us to provide ratings on their new transactions. We observe that investors have generally not stood in their way, remaining more focused on whether a security's outstanding ratings are triple-A and less focused on the methodologies behind the ratings and which agencies provide them. We believe, however, that both issuers and investors should welcome diverse credit opinions based on different methods of analysis.

Tighter rating criteria may not increase issuers' funding costs
If transaction sponsors perceive that one rating agency's criteria will lead it to assign lower ratings than another agency would, they may simply not engage the first agency to provide a rating opinion. This is known as 'ratings shopping', and we believe it is now happening in some areas of structured finance, effectively denying investors a potential diversity of opinion.

But is ratings shopping even in the issuer's interests? Not necessarily. One of an issuer's principal concerns is to minimise each new transaction's funding costs and it may perceive that engaging a rating agency with tighter criteria could increase such costs.

Minimising funding costs has historically been synonymous with maximising the portion of a transaction's capital structure that can be rated triple-A - that is, providing the minimum credit enhancement for the senior notes that would still allow them to have a triple-A rating. However, we believe that issuers may not always need to shop for the least stringent rating agency criteria in order to minimise costs - especially now that credit opinions and rating scales are becoming increasingly diverse.

In structured finance, engaging two rating agencies with different rating standards or methodologies may simply result in less of each issuance being rated triple-A by both agencies, with some other tranches bearing 'split ratings' of triple-A from one agency and double-A from another, for example. This may not necessarily raise a transaction sponsor's funding costs.

True, investors may demand higher spreads for tranches with split ratings. But they may also accept a lower spread on the senior-most tranche, which would have higher credit enhancement to support triple-A ratings from both rating agencies. The lower spread on the senior tranche may more than compensate the issuer for higher spreads on the other tranches, mitigating any effect on funding costs.

Other transaction costs can be an important consideration too. For example, in transactions that include hedging arrangements - such as interest rate or currency swaps - it may be relatively costly to obtain triple-A ratings.

This is because derivative agreements that mitigate counterparty risk sufficiently to support triple-A ratings on the notes tend to be expensive. An issuer that aims to structure, say, only double-A rated notes - even if the notes pay a higher coupon - might reduce its costs overall, while increasing the availability of lower-rated, higher-yielding structured finance securities and potentially appealing to new investors.

Investors may value different approaches to risk assessment
A more important consideration, in our view, is that both investors and regulators may not currently take into account important differences between rating agencies - and between their triple-A ratings - in ratings-linked investment guidelines and regulations. Any guidelines and regulations based on generic rating categories may rely heavily on metrics that are only loosely comparable among rating agencies, given that those agencies may not use common rating scales or standards. We believe that, by contrast, market participants should understand and value differences in ratings definitions for the unique perspectives that they provide.

In our experience, investors commonly assume that the rating scales of the major credit rating agencies are equivalent. This is perhaps not surprising, given that most agencies historically have used rating scales with about the same number of gradations (around 20) and, in some cases, even use the same letter designations.

The assumption that one agency's ratings are the same as another's has effectively become entrenched in certain regulations - for example, elements of the Basel 2 regulatory capital accord - and also in many buy-side participants' investment guidelines. However, we would warn against this view. Indeed, we have argued in favour of removing ratings references from legislation and regulatory guidelines precisely because this practice may lead to over-reliance on ratings and a false sense of equivalence among different agencies' ratings.

Consider, for example, the fundamentals of how different rating agencies assess creditworthiness. S&P's structured finance criteria generally consider - among other factors - an issuer's capacity to make full and timely payment of all its obligations on a particular security. This is not true of all credit rating agencies.

Another valid approach to assessing creditworthiness, in our view, may be to consider the size of the expected loss on the security. Differences such as this may hamper comparability between our ratings and those of other agencies.

Various definitions and methodologies among rating agencies may also lead to significantly different ratings behaviour over time. For example, we looked at a sample of French and German mortgage covered bond programmes that we rate triple-A and that Moody's rates Aaa. According to Moody's, the minimum overcollateralisation necessary to maintain the current Aaa rating is about 7% on average across these programs.

By contrast, S&P would lower its ratings below triple-A if overcollateralisation on average fell below 23%. For Fitch, we understand the equivalent figure is about 15%, according to data from Deutsche Bank.

Were it not for the focus on triple-A that ratings-based regulations and investment guidelines have introduced, we believe investors might value more highly the diversity in approaches and opinions among rating agencies. In our view, investors would benefit more from having access to varied opinions about a security's creditworthiness than from having fewer, less diverse opinions that all result in the same rating. Obviously, if each rating agency used the same methodology, a second rating on the same security would effectively add no further insight.

As long as triple-A ratings remain the de facto standard for many structured finance market participants, the scope for rating agencies to air divergent credit views will be limited. Today, a newly issued security is still more likely to have two triple-A ratings from agencies with similar approaches than it is to have triple-A and double-A split ratings from two agencies with different approaches, or with differing views of certain aspects of the risks to investors.

We can't know, of course, how perceptions and historical norms might shape investor and issuer demand for structured finance ratings in the future. But we believe that by sticking with conviction to their own rating definitions and methodologies, rating agencies can offer investors a valuable diversity of opinion.