We have updated our Terms of Use.
Please read our new Privacy Statement before continuing.

Lifting the CDO veil

16 January 2008 By Antony Currie

Asset-backed CDOs have become the bane of the financial markets. Few investment banks with pretensions of being a player in structured finance have escaped unscathed, and the writedowns firms have taken are in some cases eye-popping. Because these complex securities are illiquid and opaque, how their owners arrive at the valuations remains largely a mystery. But banks could give their shareholders a better road map than they have so far.

Collateralised debt obligations, to give them their full name, repackage all manner of debt, whether leveraged loans, corporate bonds or securitisations. They can work well, if constructed properly. But managers of asset-backed CDOs piled into subprime mortgages with abandon. As that market tanked, so did the expectation of losses on the low-rated mortgage bonds snaffled up by many of these CDOs. Bank losses from subprime-related writedowns now surpass $130bn – most of it from CDOs. Citigroup, Merrill Lynch and UBS alone account for almost half the amount.

How do the banks figure out how much money they ve lost? They don t rely on actual transactions. The banks say there are no valid market prices for such thinly traded, and often uniquely constructed, paper.

Instead, the calculations are based on the banks’ best estimates of the current value of the obligations. Using black-box modelling to decide just how much to write off is hardly ideal. But if voodoo valuations are the best available, they should at least come with enough information to let investors make more informed judgments about just how toxic this waste is. For the most part they don t.

No firm has taken the simplest step announcing the names of the CDOs they hold. With just that one piece of information, shareholders would be able to do their own research, tapping all manner of sources from ratings agency reports to third-party analysis of underlying loan performance. That works for cash deals, at least. CDOs constructed using mortgage bond derivatives are often private, and it s much harder for anyone other than those directly involved in the transaction to get information. How much of their holdings are in these so-called bespoke CDOs is another data point the banks ought to reveal.

Other disclosures so far vary considerably. The apparent winners in this market haven’t said much.

  • Barclays hasn’t explained its positions beyond its overall exposure. But it can’t spring any unpleasant surprises, since it has written the value of its holdings down to zero.
  • Credit Suisse has revealed little more than that its subprime exposure is de minimis.
  • Goldman Sachs and Lehman Brothers let it be known they have been net short CDOs for a year or more, but beyond that have said almost nothing.
  • Deutsche Bank’s chief executive, Josef Ackermann, has been out lecturing others about the need for transparency, but has been sparing with it where his own bank’s CDOs are concerned. We do, at least, know that Deutsche’s traders were among the first to slap on short positions that became so lucrative last year.

The losers, including Citigroup, UBS and Merrill Lynch, have, unsurprisingly, made more effort to explain, often providing helpful presentation slides to show a better breakdown of their subprime losses. Even these, though, have not supplied nearly enough to give a clear picture.

Here’s a short list of what investors want to know, and how much they do know. 

  • Providing the face value of all CDO exposure is the obvious, and necessary starting point. Worming this out of the banks hasn’t been easy – Citi made no mention of its $55bn exposure in its third-quarter earnings announcement, but then it took centre stage a few weeks later. One could be a tad more forgiving back in those early days, as few had revealed these figures before. Most banks now have, but often in an unclear way. 
  • Outlining the split between high-grade and mezzanine investments – in other words, bad paper and really bad paper – is a no-brainer. Most have done this, but the laggards should make it as plain as day. 
  • Very few banks have given much if any information on another important factor: vintage. Older CDOs should be safer. They often have more non-subprime collateral. Also, most mortgage bonds can pay back up to two-thirds of principal after three years, if defaults have been low. Citi announced last week that 58% of its remaining $29.3bn CDO exposure is from 2005 or earlier. Bank of America threw in some similar information, too. But the rest have been shy, allowing shareholders to assume the worst. 
  • The location of the underlying assets is another helpful data point that is sorely lacking. Although the US housing market is the most correlated it has been for years, there is still value to investors in knowing how much of a CDO is backed by mortgages from, say, really stressed markets such as Michigan or California, as opposed to as yet stronger markets such as New York. 
  • In fact, more information in general on the subprime bonds and the loans behind them making up the CDOs would go down well. Which lenders bonds were herded into the CDOs? How many of them were sold directly to borrowers, or bought from independent mortgage brokers?
  • Revealing the proportion of CDOs inside CDOs would really show a bank’s willingness to come clean to its investors. High-grade CDOs, especially last year, allotted up to a third of their portfolio to more risky mezzanine CDOs. That was the limit – any more and they would have to be classified as even riskier CDO-squareds. But regardless of the name, the losses still mount as the proportion of CDOs within CDOs rises. 
  • Finally, investors who really want to dig into the worth or otherwise of a bank’s CDO holdings would appreciate a better sense of the assumptions behind the secretive models. Estimates of house price falls and cumulative losses are critical. Citi and Merrill made one or the other public unprompted.

Societe Generale leads the pack in providing both, as well as breaking out vintage and stating that it has written down fully all CDO tranches held in its $4.8bn of CDOs. Without these additional slugs of data, shareholders have been left largely in the dark, which just encourages wild speculation. Disclosure might go some way to assuaging the persistent unrest. Even after a couple of rounds of write-downs, CDO exposure is still in the tens of billions of dollars. SG’s combined 27% writedown might look punchy to some – although it has been overshadowed by its rogue trading losses – but the French bank has at least, finally, sketched a good outline of how it got there. Its rivals should follow suit.


Email a friend

Please complete the form below.

Required fields *


(Separate multiple email addresses with commas)