How are Covered Bond Ratings Determined?

Remember covered bonds? They're ever so safe, because not only are they overcollateralized, they're also guaranteed by the issuing bank. As a result, Hank Paulson has been dropping hints that he'd like to see the mortgage industry move in the covered bond direction in future.

But "safe" doesn't mean "immune to downgrades", of course -- and Fitch has already started downgrading covered bonds issued by Washington Mutual. One problem is that it's very hard for a ratings agency to assign a rating to a both-and probability, which is what a covered bond is.

Let's say you need a 0.5% probability of default to be triple-A, and a 5% probability of default to be single-A. And, for the sake of argument, let's say that we're dealing with a bank which is only dabbling in mortgages, and has no chance of being brought down by its mortgage operations: it might fail, but if it fails it will be for reasons unrelated to mortgages.

So the bank structures a conventional mortgage-backed security with a whole bunch of tranches running from triple-A down to single-A and equity. Finding that it can't easily sell the single-A tranche in the open market, it covers (guarantees) the bond itself. And let's say that the bank itself has a single-A credit rating. What rating should the covered bond have?

It's tough. Mathematically speaking, if the chances of the collateralization failing are genuinely independent of the chances of the bank failing, then the chances of them both failing (which is what you need for a covered bond to default) are just 5% of 5%, which is 0.25%: easily low enough to be triple-A. But there's something intuitively a little dubious about two single-As making a triple-A quite so easily.

So one thing I'd like to see, if and when covered bonds become more popular, is the ratings agencies showing their work, as it were. We know what the rating of the bank is, and we know what the rating of the covered bond is, too. But in the interests of transparency, also tell us what the rating of the covered bond would be, if it didn't have the bank's guarantee. That would definitely help investors get an idea of what the chances are of one leg or other being kicked out from under the bond.