Wachovia Securities CDO analysts last month sought to find a new way to answer an old question: Is there a method to project future new-issue CDO performance using historical data? Their ultimate finding was yes, although with some caveats.
In a nutshell, the idea is to use historical ratings and recovery data, by CDO type and credit level, in order to derive a risk-adjusted return when applied to current spread levels. The risk adjusted return would provide a value for how well the securities would fare now, assuming they performed identical to historical averages. The CLO sector came out smelling rosy under the methodology - as it was the only group to display positive returns all the way down the capital structure. "I think this is a good way to benchmark relative value across sectors, but I wouldn't focus on one number," said Steven Todd, a senior analyst at Wachovia. "I would look at the general trends, and then examine how sensitive the relative value rankings are to rating transition and recovery assumptions."
Trying to judge future CDO behavior based on past deals can be tricky, given constant structural and collateral changes as time passes. For example, today's average ABS CDO would contain a much more homogenous pool of assets - namely RMBS - as opposed to older deals, which suffered from manufactured housing and aircraft exposure. Such a divergence makes judging future performance based on past rating and recovery data rather difficult. Some sectors within CDO land have also crawled into new structural shells. CDOs backed by commercial real estate, for example, have only recently begun to look more like ABS CDOs.
But in the latter of two reports issued last month by the researchers, the analysts argue that the methodology can be used, particularly in assessing relative value across CDO sectors, although changes in deal characteristics and variances in economic climate mean results should be taken with the proverbial grain of salt. Indeed, rating agencies and investors are paying more attention to ratings transitions and secondary pricing volatility, not only because of mark-to-market risk in investor portfolios, but because of the burgeoning CDO credit default swap sector, market players point out. Fitch Ratings earlier this year began incorporating ratings stability scores, which attempt to place a likelihood that a particular credit will maintain its rating for a given time period.
Certain sectors haunted by past mistakes
Risk adjusted returns were derived from Moody's Investors Service ratings transition data from 1990 to 2006 and sector specific data that covered shorter time periods; generalized CDO mean loss given default rates from 1993 to 2005 were used to calculate recoveries. The analysts focused on CLOs, synthetics, ABS mezzanine CDOs and high yield CBOs backed by corporate debt, both with generalized CDO and sector-specific data. Across the asset classes, all annualized returns ended up with a negative adjustment. That would mean downgrades have a higher average impact on performance than upgrades.
Combining the return adjustment to a given asset class' current spread would create a risk-adjusted return based on historical averages. For example, single-A rated CLOs had a sector-specific value of negative seven, meaning that if spread differentials did not change over the next year, and CLOs experience a rating transition that matches their historical average, investors would lose seven basis points of spread. That means that instead of earning 65 basis points over Libor, based on current spreads, the investment would yield 58 basis points over, according to Wachovia. Meanwhile, double-B rated high yield CBOs would stand to make out the worst, shaving 18.74% a year. It was noted, however, that this asset class took a blow due to past performance woes stemming from exposure to the telecom and dot com sectors between 2000 and 2002. Triple-B tranches of synthetic corporate CDOs fared the best out of the credit sector in the bunch, only losing 1.94%. On the triple-A side, leveraged loan CLOs actually came out even - they didn't lose anything. Synthetic corporate CDOs had a negative 0.31%, while ABS mezzanine CDOs had a negative .10% and high yield CBOs a negative 0.23%.
Triple-A CLOs a safe bet for buy-and-hold
Applying the risk adjustment figures to current spread levels, the analysts found a shining star - CLOs. The asset class showed positive risk-adjusted spreads for all points in the capital structure, whereas the other asset classes displayed mostly negative risk-adjusted returns below the triple-A level. "Based on this simple analysis, we may conclude that CLOs appear cheap relative to most other CDOs." Wachovia analysts wrote, but not without noting the historical performance blunders, like manufactured housing, that skewed certain results. Also, the method's rating transition methodology assumed that notes downgraded below single-B generate zero recoveries.
Triple-A CLO notes were found to be the safest bet for buy and hold accounts. They were also found to have the lowest mark-to-market risk - the analysts do not expect spreads to return to wide levels seen in December 1998 and January 1999. Increasingly leveraged corporate borrowers and changing rating criteria, however, could result in increased structural leverage, making double-B notes vulnerable to mark-to-market risk, they noted.
Even though triple-B rated ABS mezzanine CDOs offer the highest spreads among cash CDOs, performance will ultimately depend on a combination of the housing market and various structural nuances in both the deals and their collateral, Wachovia found. And mark-to-market risk is high in this area. "We would not be surprised if triple-B spreads widened to 400 to 500 basis points. Last year, we saw triple-B price talk as wide as 375 to 400 basis points," they wrote, adding that some inventory was rumored to have been sold at a discount consistent with a 500 basis point spread.
(c) 2006 Asset Securitization Report and SourceMedia, Inc. All Rights Reserved.