Monday, September 27, 2010

Credit Ratings Reversals

The debate continues over the usefulness of credit default swaps (CDS) spreads as alternatives to ratings.

Today, Moody’s Corporation announced that its Analytics division – separate from its ratings group – has improved the ability of its EDF (expected default frequency) model to estimate default probability as a result of the incorporation of CDS spreads to the platform.

Moody’s Analytics clearly agrees that CDS spreads provide useful predictive content. So did a fellow panelist of ours at a distressed debt conference on Friday.

Jerome Fons, EVP of Kroll Bond Rating Agency, included the following slide in his presentation (click here to download the presentation in its entirety).

Among other things, it shows CDS spreads to be better predictors of default probability (see the higher Accuracy Ratio).

The slide also displays the lower frequency with which credit ratings are reversed by rating analysts, versus the regularity with which CDS spreads can move from one bucket to another as per the market’s whims.

This feature, as displayed by Ratings Reversals and Rating Changes, reminds us of the human nature of rating agency analysts and in particular their psychological predisposition against reversing a prior rating action. The obvious upside is ratings stability – at the expense of volatility -- to the extent we care for it. Would we want our regulatory capital ratios to move on a daily or secondly basis, as a stock price may trade on the news, or on gossip?

For example, consider the case of Arlington CDO tranche A3. Moody’s and S&P both started off at Aa2/AA ratings, respectively, in the year 2000. In 2002, Moody’s downgraded it more aggressively than S&P, a situation which lasted until 2006, at which stage Moody’s upgraded the bond to A3, which was the then-current equivalent of S&P’s rating of A-. 2009 arrives and Moody’s drops to Caa3, before upgrading to B3 in early 2010 and then Ba3 last week. Moody’s is now just short of S&P’s current equivalent rating of BB+.

While certain market participants might benefit from more regular rating actions, others no doubt value ratings stability above all else. But either way, it seems entirely unlikely that rating stability and ratings accuracy go hand-in-hand.

We remain very interested in the topics of ratings alternatives and the comparison of ratings performance. Let us know if you have a similar interest in these topics.

For more on CDS spreads as alternatives to ratings, click here; to visit our submission to the Fed, OCC, OTS and FDIC on this topic, click here.

No comments: