Wednesday, February 27, 2008

S&P on moolines: "no problem"

(Accrued Interest) S&P and Moody's have now both affirmed MBIA. S&P also more or less said they would affirm Ambac as well if the reported $3 billion capital infusion is completed. MBIA's infamous 14% surplus note is now trading comfortably above par ($101 bid, $104 offer last night). So are we out of the woods with the monolines?

Well let's start by looking at S&P's methodology. In short, S&P will bestow a AAA rating if they believe an insurer can survive their "stressed" scenario. Here are their assumptions for various types of mortgage-related securities.

First, the cumulative losses for various types of loans. Note these are losses, not default rates. If you assume, say 50% recovery on first liens, then the default rates assumed would be roughly double what you see here. For some context, during the 2001 recession subprime default rates got as high as 9%, so clearly these loss rates are a multiple of historic norms, even during a recession.
Table 2 shows S&P's loss assumptions on direct RMBS transactions. That's where the monoline has insured a pool of loans directly, as opposed to a tranche of a RMBS transaction. So here a AAA-rated RMBS tranche means that S&P estimated the whole pool was worth a AAA at the outset.

Table 3 is starting to get to the real problems. What they mean by a "tranched" RMBS is one where there is a subordination credit enhancement. In other words, losses hit the lower-rated tranches first, and higher rated tranches only if the lower-rated pieces are completely wiped out.

You can't get much worse than those 2006-2007 loss levels, even on AA-rated tranches. Note that these loss levels will dictate the loss levels on ABS CDOs, which is table 4.

I noted in a previous post that CDOs was the monolines biggest exposure, so this table is the real kicker. Remember that the monolines basically only insured senior tranches of CDOs, so these loss rates would seem to represent an assumption that all the subordination will turn out to be worthless. When they say "high-grade" CDOs, they mean a CDO made up of ABS paper rated A or better. That does mean that the transaction is more leveraged, i.e., there is less subordination under the AAA-rated tranche. "Mezzanine" CDOs were usuallly made up of paper rated A or BBB. Since a lot of these mezz ABS pieces are turning out to be worthless, its not surprising that the CDOs are going to take huge losses. I'd say the only thing preventing these losses from being 100% is that most CDOs had more than just RMBS in them, and not all that paper will turn out to be worthless.

So all in all, I'd say that's a pretty stressful scenario.

No comments: