How To Get To AAA

When a debt security carries a high credit rating, like AA or AAA, what makes it secure? Sometimes it's just the fact that the issuing entity has a lot of financial strength: think UPS, for instance. But the number of companies with that kind of financial strength is decreasing rapidly, as mangement succumbs to shareholder pressure for more leverage. Yves Smith of Naked Capitalism notes in an email to me that

in the stone ages of finance, there was a fair bit of AAA paper (most money center banks, many utilities, AT&T when it was Ma Bell, most sovereign credits) but now there are very few native AAA credits. So there has been an imbalance in that slice of the market.

That imbalance is driving demand for more artificial AAA bonds, from investors who can't find the native stuff any more.

Meanwhile, Tanta at Calculated Risk is quoting John Mauldin on how the ratings agencies rate mortgage-backed securities:

The rating process was not the same as the ratings that were used in the corporate world. But the problem is that the ratings used the same designations. Instead of creating a whole new type of rating standard (say, using numbers like "CDO rank 1-10"), they used the same designations that bond investors were used to.
I think it is disingenuous for a rating agency to explain the difference in paragraphs 457-503 in 7-point type and dense legalese in their disclosure document. Investors had (and should have) a certain level of expectation when the designation "AAA" is used. GE and Exxon types of expectations.

The main difference between a native AAA and an asset-backed AAA is in mark-to-market volatility. A credit rating does not mean that an investor is unlikely to lose money if he marks his investments to market. It means only that an investor is unlikely to lose money if he holds his security to maturity. The problem is that if you're buying native AAA bonds, you're getting safety on both counts, whereas if you're buying asset-backed AAA bonds, you're getting safety only on the latter count. But investors knew that perfectly well: it's not the ratings agencies' fault that they chose to ignore it.

How does a bunch of subprime nuclear waste become a AAA bond? That's the question I hinted at yesterday, when I said that diversification was key. In response, I got a number of comments, both on the blog and by email, telling me, quite rightly, that I'd ignored the other key way of boosting credit ratings: overcollateralization.

But in the CDO market, I'm not convinced that overcollateralization always works particularly well. Consider three different CDOs:

1. A $100 million CDO backed by $200 million in mezzanine tranches of subprime-backed mortgage-backed securities.
2. A $100 million CDO backed by $100 million in AA tranches of subprime-backed mortgage-backed securities.
3. A $100 million CDO backed by $100 million in a mixture of residential MBS, commercial MBS, and corporate syndicated loans.

All three of these CDOs can and would be tranched, and one of those tranches would end up carrying a AAA rating. But I think I would feel safer with the triple-A tranche of CDO3 than of CDO2, and I would feel safer with the triple-A tranche of CDO2 than of CDO1.

How can CDO1 be the least safe of the lot, when it's backed by twice as much collateral? Because in the event of a serious subprime meltdown, mezzanine tranches of subprime bonds could be wiped out. And twice zero is still zero. Once the mezzanine tranches are wiped out, the higher-rated tranches start taking losses as well. But at least they retain some cashflow, which means that the AAA parts of CDO2 – the parts which get paid first – would probably be fine.

Meanwhile, the owners of the AAA parts of CDO3 are really quite sanguine about subprime losses. Those losses are absorbed entirely by the equity portion of the CDO, and the AAA parts can be paid entirely with cashflow from commercial mortgages and corporate syndicated loans.

In other words, the most diversified CDO is the safest, while the most overcollateralized CDO is the weakest. Now of course this is an artificially-constructed thought experiment, and I'm sure there are lots of overcollateralized CDOs which are perfectly safe even if they don't have much diversification. But as a general rule I think a CDO buyer should be looking first at diversification, and then to other sorts of credit enhancement, such as guarantees or wraps from insurance companies who specialize in such things.