I am still trying to catch up with my backlog of blogging; work as been interfering rather drastically of late, and by the time I got home I seldom had the energy to blog. Hopefully this week will be better. One of the loose ends are comments on financial derivatives, which I referred to recently. I said that I was surprised that we had not seen (publicly) any major blow-ups of equity CDOs. A CDO is a Collateralized Debt Obligation, and comes in many forms, ranging from the harmless to the volatile. The riskier ones are backed by high-yield bonds (more commonly known as "junk bonds," but "high yield" sounds so much better). These are the bonds rated below investment grade, i.e. lower than BBB-. The default risk of these bonds is much higher than for safer bonds with higher ratings. But even high-grade bonds can default. WorldCom's debt was rated AA when it defaulted, for instance.

Let's take a junk bond CDO as an example. What you do is you borrow a chunk of money in the markets and use to buy a portfolio of junk bonds. You want it to be well-diversified so that any one default won't blow your portfolio out of the water while you still profit from the higher yields of these bonds. Now you start redistributing the risk of this portfolio. You create a synthetic AAA-rated bonds for instance, which pays a spread of government bonds of, say 200 basis points. In order to get that AAA rating, you need to overcollateralize it with junk bonds. Essentially the holders of the synthetic AAA bond get first dibs on any cash flows coming from the underlying junk portfolio. Then the buyers of the AA-rated synthetic bond get their share, etc. What's left at the end is the equity portion of the CDO. It concentrates the risk of the junk portfolio in a small part of the capital; the overall aggregate risk of the bits and pieces into which you've sliced the underlying junk remains the same (otherwise you could make risk evaporate), and if some parts have much lower risk than the underlying, some other parts must have much higher risk. And that's the equity portion of the CDOs.

Junk bonds have not done well in the last five years. The spreads over government bonds have widened dramatically, far more than nominal yields on government bonds have fallen. Coupled with increasing default rates, this has meant that junk bond portfolios have had negative returns for several years now. Credit spreads have started to tighten again in the fourth quarter of 2002, so the worst may be over now. But institutions who bought equity CDOs have seen those investments get marked to zero. Individual investors will not have been affected, since you typically need tens of millions of dollars to buy into such a structure. In some cases, the issuing investment bank will keep the equity CDO on its own books to show its commitment and confidence in the deal. In many cases, equity CDOs have been sold on to third parties. Big buyers in Europe have been primarily insurers and some banks. I've talked to some people who've been involved with CDOs over the last years, and they say that it has hurt the equity CDO buyers, but that they were generally well diversified and did not invest too much of their portfolios in these things. But it's hard to tell; these deals are very private, and nobody likes to talk about them in the first place, let alone when they go sour. So we're unlikely to see these deals reflected explicitly in the investment results unless some institution really blows up in a big way. The losses are there, but CDOs have not been big enough (yet?) to cause a blow-up. I sincerely hope it stays that way.