The Economonitor is off to the Antarctic for a little while, and his replacement – Stefan Geens, whose byline you should be seeing round these parts starting today – knows little and cares less about CPDOs. Which means that soon this blog might start becoming readable again. But I'm still here today, which means that I can blog Mark Gilbert, of Bloomberg, who's got a nifty column on my favourite credit derivative product:

"Dear Sir/Madam, your details were provided by someone who is assuring me of your honesty and integrity for an Urgent Business Proposal in Confidence of the Strictest Nature. I am the sales director of the Democratic Republic of Derivatives. Unbeknownst to my colleagues, I have discovered millions of dollars hidden in an unexplored corner of the republic in the form of Constant Proportion Debt Obligations, or CPDOs. I seek your assistance in unlocking this value for the benefit of both of us."

Pretty funny – and, if you continue reading, there's even substance there as well. Here's the bit which jumped out at me:

"At higher volatility, the structure will have a sharp increase in knock-out-scenarios," where losses become too big to recoup by making bigger bets, Citigroup said. "Knockouts will be more likely earlier in the life of the transaction." That would prompt an unwinding of the security, with investors forfeiting as much as 90 percent of their stake.

These things are starting to make a bit more sense – in fact, the CPDO critics and the credit rating agencies might both be right.

Simply put, to get the value of a bond, you take the value of its future cashflows if it doesn't default and add to that the value of its future cashflows if it does default – the so-called recovery value. So the ratings agencies might well be right that the bond is very unlikely to default. Here's Citi again:

"Jokingly, we have started calling the product a 'hydra;' almost every time we tried to kill it by subjecting it to severe stress, it seemed somehow to be able to recover par by maturity."

In the extreme situation where the bond does default, however, the recovery value on a CPDO will be much, much lower than the recovery value on a plain-vanilla triple-A bond. So that helps to explain why CPDOs can be riskier than other AAA credit, even if they have the same rating.

Gilbert also finds an example of a AAA-rated security which got downgraded to just one notch above junk in a very short amount of time:

CPDOs have been able to secure AAA ratings for both interest and principal payments. Ratings for derivatives-based securities, though, can change rapidly. In April, credit-linked notes issued by Barclays Plc and worth 70 million euros ($90 million) were cut by a staggering nine levels by Fitch Ratings, sliding to BBB-from AAA. The security, called Xelo III, took a beating after two U.S. auto-parts makers, Delphi Corp. and Dana Corp., filed for bankruptcy protection from creditors.
A truism of financial markets is that everyone agrees there's no such thing as a free lunch -- until they think they have spotted a buffet of risk-free basis points. CPDOs look too good to be true, so guess what? They probably are.

I can't wait to get my teeth further into this subject when I come back in three weeks.