Leaving to one side profound uncertainty about investors risk appetite for CDO product in the medium term, there is a strong possibility that the triple-A rating will become a rarity in structured credit.
First, Fitch raised the markets ire by proposing to shift its methodology on synthetic corporate risk CDOs to a dramatically more conservative stance. The agency predicted, following the implementation of the new model, five-notch downgrades on average for deals rated under previous assumptions. Moodys proposes doing away altogether with its well-known alpha numeric ratings scale for complex securities (in other words CDOs) and replacing it with one based purely on numbers or perhaps a special signifier to denote the asset class. Standard & Poors has also made some modest changes to its RMBS-backed CDO methodology.
Both Fitchs and Moodys proposals are a reaction to the fact that many ABS CDOs rated by all the rating agencies have not lived up to the rankings initially bestowed on them. What was called a triple A was not in fact a triple A, and in many cases not even of investment grade. Fitchs approach is brave, although there is the possibility that it will lose a lot of business as a result of trying to match CDO ratings with standalone credit ratings. Moodys risks confusing everyone with a scale that means nothing to anyone.
"Our main aim was to produce a corporate CDO rating methodology that results in ratings that are consistent with other asset classes," says Ken Gill, managing director, structured credit, at Fitch Ratings. "Effectively there are two questions you need to answer in order to do that. The first is: what is the appropriate quantitative approach? Secondly, when should the models results be challenged?"
Credit researchers from BNP Paribas welcomed the move, saying that as corporate defaults are likely to increase over their long-term average of 4.5% over the coming years, it is sensible that Fitch is planning to use a peak rate of 9.3% instead.
"Concentration risk is something that we believed needed to be addressed," says Gill. "But in the first instance we looked to try the model out on a fundamental credit view. For example, single-A corporates should not default even in periods of peak defaults. Similarly, corporate CDOs rated single A and above also should not default even in periods of peak defaults. Its easy to produce a model that results in that being true for large diverse portfolios but we wanted the framework to also test the risks posed by concentration and adverse selection," says Gill.
Fitch is acknowledging that there is a fundamental difference between structured credit and standalone credits. An A-rated credit tranche cannot go out and raise equity in the same way a bank or corporate can.
"A triple-A rated tranche in a static structure is at the mercy of the underlying default and recovery rates, a fact not yet being explicitly acknowledged by the rating agencies. Technically speaking, being a AAA-rated tranche is equivalent to selling out of the money default and recovery rate options (extreme tail risk), a concept few AAA investors realize even after the sub-prime debacle," says BNP Paribas researcher Mehernosh Engineer.
Most bankers complained about the dramatic scale of downgrades that will take place if Fitch does follow through with its planned changes. CDOs in corporate credit have performed fine, leaving aside any market value-type structures such as CPDOs that Fitch did not rate anyway. While the indices and CDS spreads are at record wides, defaults remain at record lows. They were also wary of qualitative interference that Fitch might make in the ratings process.
Fitch looks to peak default rates
Ten-year rolling cohort triple-B default rate
"Maybe its right in principle but you want rating agency methodology to be transparent so you can create deals. Otherwise it is an inefficient process that is it too labour intensive," says one banker.
"What weve learnt is that concentration has crept in, as have other forms of adverse selection. It is those things we are trying to address by not only having a quantitative approach," says Gill.
CDO structurers are far from happy about the introduction of qualitative methodologies. But Fitch says that it will only be using such measures when the pool or structural characteristics warrant it. But the nature of banking, and structured credit specifically, is to seek out arbitrage opportunities. Fitchs proposals threaten to cut the opportunities for ratings arbitrage which is to take credits that are underperforming given their ratings and repackage them into CDOs. Fitch has also looked at changes in corporate governance and decided that recovery rates will probably be lower.
Rating agencies methodologies differ in default assumptions, recovery rate expectations and so on, so it is not unusual for there to be a meaningful gap in the actual ratings bestowed on standalone entities and for structured finance ratings. This also leads investment bankers to seek out the less conservative agencies for their clients. Given that there are only three or, at a push, four serious players in the international rating agency business, the implications are clear. The agency that is too out of step with its competitors will rate few transactions. Consequently Fitchs move appears to be extremely foolish or is it?
According to BNP Paribas, some $220 billion of synthetic CDOs will be affected. Fitch states that if this new methodology were applied to existing ratings of investment-grade synthetic CDOs, it would expect that on average there would be a five-notch downgrade. If this were applied in a uniform manner, triple As would be downgraded to perform in a manner consistent with single A. That would not please investors. There were a lot of rumours surrounding CDOs unwinding whether these are pre-emptive moves ahead of Fitchs possible shift is not known. But forced unwinds are a strong possibility if all goes ahead as planned.
In a mood
Bankers might grudgingly accept that Fitchs motives are sound but they were less flattering about Moodys proposals. Moodys issued a special comment that contains a wide range of options regarding the agencys ratings scale.
"From time to time we have heard market participants say: Structured finance ratings have performed differently from corporate and governmental ratings, and the underlying analysis is very different, so perhaps structured finance ratings should be presented in a different manner than ratings on corporates and governments," says Richard Cantor, managing director, structured finance at Moodys.
Moodys proposals include changing the scale from Aaa, Aa1, Aa2 and so on, all the way down to C, to a purely numerical scale ranging from 1 to 21. It is also considering using a scale that would simply include a suffix such as Aaa.v1, Aa1.v2.
Independent credit research company CreditSights points out that there are differences in the nature of default losses between financial and non-financials, or between lesser developed and developed sovereign countries. But rating agencies manage to bridge these differences to produce ratings that are generally comparable across sector divides. CreditSights was not alone in questioning whether what Moodys is proposing will help to shield it from claims that its CDO ratings were overly generous in the first place.
But Moodys officials deny any such intention. The aim is simply to help market participants know quickly what type of structure is being referred to, they say.
"We remain committed to our efforts to improve our rating methodologies and strengthen market confidence in structured finance transactions, but it also makes sense to see whether there is broad or narrow support for the anecdotal interest we have heard expressed for modifying the way structured finance ratings are presented," says Cantor.
It is true that many credit committees do not distinguish between different types of asset classes. This is one reason explaining why some investors have voiced an interest in such a change.
The structured finance sector is in deep trouble. A lack of trust in the rating agencies is widely cited as one of the reasons why there is little appetite for structured finance transactions. Fitch has admitted, implicitly at least, that its previous ratings were over-generous and that triple-A CDO tranches should in fact have been rated single-A, for instance.
No one could reasonably argue that there isnt a difference between securitized and non-securitized bonds but whether it is impossible to have similar scales for predicting expected loss between these two sectors is moot, to say the least.