Upended by downgrades

The structured credit markets have been suffering as a result of problems arising from the US subprime mortgage crisis. Dealers expected worried investors to start questioning the viability of related collateralised commodity obligations, but they were not prepared for a surprise rating criteria change leading to major downgrades. By John Ferry

Barclays Capital knew it was on to something hot when it brought the first collateralised commodity obligation (CCO) to the market in 2004. Yield-hungry, diversification-seeking institutional investors lined up to buy the notes, which combined the structural features of a collateralised debt obligation (CDO) - subordination, tranching and a rating - with exposure to the commodity markets. Other dealers, most notably Credit Suisse, stepped in to offer their own versions of CCOs, and soon what had started off as a niche concept had turned into a multi-billion-dollar new segment of the structuring business, with Barclays Capital issuing almost $2 billion of notes globally.

Then the problems in the US subprime mortgage market stepped up a gear, causing mass downgrades of residential mortgage-backed securities and CDOs of asset-backed securities, uncertainty over valuations and a drying up of liquidity in the secondary market. Suddenly, the viability of anything that sniffed of being a CDO was called into question. With so much uncertainty surrounding the structuring and pricing of CDOs, should the risks and returns of CCOs also be called into question? Could the ratings applied to CCOs be taken as a true reflection of risk when so much doubt is being cast on rating agency analysis in the broader structured credit arena? And should the credit rating agencies have been assigning ratings that reflect the probability of large jumps in commodity prices - in other words, market risk - in the first place?

Just as investors and CCO issuers were starting to grapple with these questions, the market suffered a large upheaval in September, when rating agency Fitch suddenly changed it ratings criteria for CCOs - a move that led to the downgrade of a number of tranches. Key to the methodology change was the introduction of a scenario stress test overlay into its analysis system, a decision Fitch says reflects the possibility of a structural break in commodity price behaviour since 2003 that might not be captured in the historical data. Specifically, the test is designed to protect investors against a sharp reversal of the price gains experienced by some commodities over the past few years. The test introduces a minimum level of credit enhancement, so that a CCO would be able to withstand a large price decline of several of the reference commodities in the portfolio over a short period.

Almost all Barclays Capital's products were rated by Standard & Poor's (S&P) - which did not change its rating methodology - so none of its public products were affected. Credit Suisse, however, did suffer. The Swiss bank started issuing its Cargo series of CCO notes at the start of this year. The notes reference 17 different commodities through a portfolio of 100 long and 100 short commodity trigger swaps - the underlying derivatives products that form the basis of CCOs. Fitch downgraded one $70 million tranche of notes from AAA to A-, a six-notch move. Another $70 million tranche went from AAA to AA-, a EUR25 million series of notes was downgraded from AAA to A+, and a separate EUR6.8 million series also went from AAA to A+.

Collateralised commodity obligations are structurally similar to synthetic CDOs in that they employ tranching technology, credit enhancement and subordination, and often have ratings assigned to the notes sold to investors. However, CCOs reference a portfolio of trigger swaps linked to specific commodities, rather than credit default swaps. Commodity trigger swaps are essentially out-of-the-money European-style options. A 'default' occurs if the price of the underlying commodity has fallen below a pre-defined trigger level at the option's maturity - each loss event leads to a capital loss on the structure.

Barclays Capital was first to market in 2004 when it launched its Apollo CCO. The bank subsequently launched the first managed CCO, Everest 1, in conjunction with California-based asset manager Trust Company of the West last year. Earlier this year, it also launched the first super-senior CCO. Credit Suisse, meanwhile, completed its inaugural Cargo transaction comprising $190 million of notes across three- and five-year tranches, in April.

The September downgrades came as a blow to Credit Suisse, which had specifically structured its products to gain the highest credit rating. In justifying its decision immediately after the downgrade, Fitch said the price of several base metals had declined since the transaction had launched. In particular, the price of nickel had dropped to approximately 55-60% of its level at the time the transaction closed, increasing the likelihood of breaching at least some of the 15 long triggers linked to that particular metal. Aluminium, zinc, corn, silver and natural gas prices had also dropped in value by about 10-20%. Fitch explicitly stated that the downgrades were on the back of the amendment to its ratings criteria on CCOs.

Credit Suisse would not comment directly on the Fitch downgrades. However, the bank is said to be surprised that the downgrades took place when the underlying commodity markets had not changed immediately pre and post the downgrade.

Some market participants say Fitch's move was a deliberate attempt to plug a hole in its model. "S&P's models had greater credibility to begin with," says one fixed-income manager. "I've discussed the Fitch models with people in the industry who felt there were some flaws with them from the start. These views were expressed to me before the liquidity crunch that began in August."

The criticism of the rating agency's methodology only adds to the mounting disapproval of the way the agencies in general deal with complex products. Questions first started to be asked last year when highly leveraged constant proportion debt obligations were assigned AAA ratings - some analysts claimed this was, for some products at least, a result of ratings arbitrage (Risk May 2007, pages 64-66).¹ Then, in July, Moody's Investors Service and S&P downgraded hundreds of securities linked to the subprime mortgage market. While the agencies argue that their ratings are based on statistical analysis of historical loss data, and are not intended to reflect mark-to-market volatility or liquidity, regulators in the US and Europe have nonetheless begun looking closely at the role of rating agencies (Risk September 2007, pages 25-28).²

Derivative Fitch, the part of Fitch that rates structured credit and its related products, insists it is not playing catch-up, and that its models are some of the most advanced around (see box). A spokesman for the company says that in rating CCO transactions, Derivative Fitch's analysts employ an approach they feel "fits the performance profile of commodity assets better than alternative models". The spokesman continues: "The timing of the rating actions was driven by the performance of the underlying assets. Fitch's performance analytics team observed price trends for commodities over a period of time before deciding to take action. The simultaneous methodology change introducing an additional stress test amplified some of the downgrades, but did not trigger the rating actions."

So what will the rating volatility and the uncertainty surrounding market risk structures mean for the development of the CCO market in the months and years to come? Opinion is divided. "Now, any kind of structured product makes investors a bit leery, which means CCOs will get a lot more detailed scrutiny. But that means it will just take investors longer to commit to a CCO. Through their increased scrutiny, they will see that the source of return is, in fact, from something other than credit markets," says Bob Greer, Newport Beach, California-based executive vice-president and real return product manager at Pimco, a fixed-income manager.

Greer says he expects to see new issuance of CCOs at some point in the future. However, there is scepticism from other quarters. "There could be significant model risk," says a structuring expert at a leading US credit house. "It is difficult for CCO investors to gauge whether the coupon they receive is appropriate to the risk they are accepting."

Ratings volatility obviously makes investors nervous, but despite this Credit Suisse is still upbeat about the prospects for the products. "We've been trading them throughout the recent market volatility, so unlike some of the other CDOs, where pricing transparency disappeared for a period of time, our CCOs have shown consistently transparent, liquid prices, and the performance has been relatively good throughout the turbulence," says Adam Knight, London-based managing director in fixed income at Credit Suisse, and head of the business partnership the bank has built with Switzerland-based commodities company Glencore International.

Knight argues that it is easy to observe a specific commodity price at a specific point in time, unlike some CDOs, where credit deterioration could take time to show up in the capital structure - and this gives CCO investors a degree of comfort that is lacking in structured credit. The market is therefore not likely to come to a complete and prolonged standstill, he says - although the bank accepts it will have to do more work to educate prospective investors about the benefits of CCOs.

"People are still interested in gaining exposure to commodities in a variety of ways, and this is one way they can gain that exposure," adds London-based Mark Harvey, managing director and head of structuring and marketing for Credit Suisse's metals tie-up with Glencore.

Barclays Capital refuses to comment on the performance of its products or the current state of the CCO market. However, a source close to the bank says it is confident that its clients are seeing positive mark-to-market values on their CCO investments, and that constant transparency and liquidity has been provided.

S&P is also broadly positive about the CCO market. Kimon Gkomozias, a London-based quantitative analyst at the firm, confirms that S&P has not performed any CCO rating changes. He says the deals the agency has rated are currently stable partly because they reference European-style options, where the exercise of commodity trigger swaps could be five years away, if triggers take place at all. "There has to be some serious structural change to commodity markets before there would be a requirement for downgrades or upgrades," he says. "Before we see any kind of breakdown in fundamentals in these markets, my guess is there isn't going to be any sort of ground for rating action."

Market participants say no issuance is taking place at the moment, but this is partly down to the fact that spread compression has eaten away at deal value, in the same way that ever-tighter spreads drastically reduced the profit opportunity from basic CDO structures before the credit crisis kicked in. Indeed, if it had not been for the credit crisis, the next logical development in the CCO market might have been the introduction of more highly leveraged structures that would provide investors with increased returns, dealers say.

Value

Ironically, with credit spreads widening again, attention could fall away from CCOs, not because they are tainted with the negatives associated with structured credit, but because more value can be achieved in structured credit itself - assuming, of course, that investors return to the structured credit market. Whether we see further CCO issuance in future will mainly come down to the economics of the deals in comparison to other transactions, says Pimco's Greer: "The inherent value of CCOs will have to move in a direction that will allow the risk to provide enough return that it is competitive with some of these other products."

In other words, whether or not Barclays Capital, Credit Suisse and others issue more CCOs could be a function of the movements of forward curves on commodities and option prices as much as uncertainty surrounding structured credit.

Modelling CCO risk

Fitch's CCO rating model is called Vector CCO. It uses Monte Carlo simulations of correlated commodity returns, then compares them with the applicable triggers of the CCO and their respective loss severity to come up with a joint portfolio loss distribution.

Fitch says that in developing its framework, it considered a range of options and found that a multivariate approach gave the statistically best fit among alternatives to historically observed commodity return distributions. It therefore jointly models the universe of commodities rather than each commodity individually.

"The commodities universe is modelled using principal components analysis to derive a correlated factor structure for the universe, combined with an asymmetric generalised autoregressive conditional heteroskedasticity (Garch) process with jumps and a constant mean return that is used to model each individual factor," says Fitch.

Principal components analysis identifies common factors contributing to each commodity's return process and also quantifies the percentage of risk in the overall portfolio contributed by each factor. Overall, Fitch says its model takes account of volatility across commodities and of observed patterns in empirical data, such as fat tails, price jumps, volatility clustering and correlation.

In stating the case for its scenario stress test overlay, Fitch says there may have been a structural break in the price performance of some commodities, especially base metals, since 2003, when prices increased significantly over a period of two to three years. "While it is still premature to draw conclusions as to whether the current levels of prices are likely to persist in a more challenging economic environment, there is a risk that the price increases could be at least partially reversed and that simulation using a model calibrated to historical data may not be able to fully capture the price dynamics since 2003," says Fitch.

The scenario stress test overlay was included as part of Fitch's CCO analysis to account for the possibility that the boom in commodity prices could quickly reverse. The agency says the aim of the overlay is to ensure that any of the products it rates can withstand a decline in the price of a number of reference commodities over a short period of time from their current price towards their long-term mean price, which is calculated using historical price data from 1989 to 2003. The size of the price shock that each commodity in the reference portfolio is expected to withstand depends on the remaining tenor of the deal and the target rating of the note. In coming up with its final CCO rating, Fitch considers the higher of the rating loss simulated by its Vector CCO model and the loss rate determined by the scenario test overlay.

S&P says it uses a mean-reverting jump diffusion model to analyse CCO risk. Kimon Gkomozias, a London-based quantitative analyst at S&P, says his firm's model implicitly takes account of the possibility of a structural reversal in commodity price behaviour. "We do a lot of analysis of economic factors and fundamentals of the commodity markets. These are incorporated by construction into our models," he says.

1 www.risknews.net/public/showPage.html?page=445525

2 www.risknews.net/public/showPage.html?page=465516