The Wall Street Journal ran an odd article yesterday about the unpredictability of sovereign credit ratings that are below the investment-grade cutoff (BB+ and lower). Check out the table from the IMF of S&P’s sovereign ratings.
The WSJ article seemed to air some highly paid bond-fund managers’ whining that ratings were not a useful signal for when they should buy and sell bonds of specific countries. The complaint was also that ratings don’t include certain data sets that are important, such as fund flows in banks, and that they don’t have the agility of credit default swaps.
The guys quoted in the WSJ might be right on the data sets that raters use and they are certainly right about CDS being more agile than ratings. But ratings are not intended to mirror market sentiment like CDS does. They are supposed to stand above market panics and routs and give a 30,000 foot view of an issuer’s credit condition.
Raters should incorporate every type of information available to them. Markets want hot, accurate information. Raters create reputations for themselves for good analysis, and reputations rise and fall on good predictive ability. Information is money, and markets want to make money.
It’s a well known phenomenon in fixed-income markets that once a bond falls into the junk or speculative category that its rating becomes murkier for predicting default. So maybe credit ratings are more useful for investment-grade bonds and CDS are more useful for junk-grade securities? If there are any academic studies on the subject that you know of, please add them in the comments.