The challenge of CDO squared

New angles

Richard Gambel, an analyst in the CDO group at Fitch Ratings in London, believes there should be considerable innovation this year. In addition to the inclusion of short positions, Gambel says the use of equity default swap (EDS) underlyings will grow. “These innovations will happen within vanilla single-layer structures, as well as within CDO squared structures,” he adds.

Fitch estimates that there was a 400% year-on-year increase in CDO squared issuance last year. For investors, the perceived attraction is simple: additional yield at a time when the structural arbitrage inherent in a standard synthetic CDO is less profitable. In the first half of 2004, for example, single-A tranches of CDO squared paid around 5–10 basis points more than a similarly rated tranche from a straight synthetic CDO. Further down the capital structure, the difference was even more marked, with a triple-B tranche potentially offering around 75bp more.

The pitch to investors looks attractive. But not all are sold on the idea. Philip Obazee, head of derivatives at Delaware Investments in Philadelphia, is one of the unconverted. “We haven’t got involved in CDOs of CDOs. They are complicated beasts and an understanding of the risks of these new products is still evolving,” he says. Delaware is the asset management arm of Lincoln Financial Group, which has around $100 billion worth of assets under management.

Kassy Kebede, principal and founder of Panton Capital in New York, is more optimistic. “Compared with a standard synthetic, the underlying names in a CDO squared are one step removed – hedging is more complex,” says Kebede. “There may be value to be found, though,” he adds. Kebede’s Panton Alternative Partners Fund was launched 12 months ago and follows a relative value credit arbitrage strategy.

Arturo Cifuentes, head of CDO research at Wachovia Securities in New York, is another sceptic. He sees only minimal arbitrage opportunities. “Many CDO of CDO deals are chasing similar assets,” he says. “And risk managing these deals is very problematic – you need a fair amount of modelling ability to deal with the complexity,” he adds.

Janet Tavakoli – a former head of financial engineering at Westdeutsche Landesbank in London – advises prospective CDO squared investors to be wary: “If you think you will be able to analyse each underlying CDO, then look at the overall structure and get a clean answer about risk, you are mistaken. It’s way too complex,” she says. Tavakoli now runs a consultancy – Tavakoli Structured Finance – in Chicago.

However, there are new developments in the modelling of CDO squared risk that might help. Semi-analytical models for such products where the interior single tranches do not contain the same credits do exist. They are not in widespread circulation though – there are only a handful of dealers that have a robust model at their fingertips.

Reoch Consulting, founded by Robert Reoch, a former head of credit derivatives at Bank of America, is one of the few non-dealer firms to have created such a model. James Wood, head of quantitative credit at the London-based credit derivatives consultancy, says modelling the non-overlap case is somewhat less onerous than modelling for structures with overlap, but that it’s still not straightforward. “It’s also plausible that one could create a semi-analytic model that could handle a limited amount of overlaps,” says Wood. “But the high level of overlap, and thus path dependency, in current market CDO squared trades begs the question of whether an efficient semi-analytic model is possible,” he adds. The path dependency that Wood refers to arises because in CDO squared the order in which underlying credits default matters, as only a subset of names are present in more than one of the inner CDO tranches.

Certainly, communications between dealers and the buy side about the risks of structured credit deals can be lacklustre – witness the steady trickle of legal disputes between dealers and disgruntled CDO investors over the past 18 months. And even before they have to grapple with the ongoing risk management of a CDO squared, there can be problems. Inexperienced CDO investors sometimes heavily rely on rating agencies’ assessments when forming opinions about the attractiveness of potential CDO squared deals – something Tavakoli says can be dangerous.

While the approaches of the major rating agencies to CDO squared share many common features, there is a fundamental difference: correlation assumptions. “Opinions [between rating agencies] diverge significantly when it comes to correlation,” says Matthias Neugebauer, an analyst at Fitch Ratings in London who helped develop the rating agency’s approach to CDO square analysis. “Different correlation assumptions can produce very different answers,” he adds. Caveat emptor.