There Are Many Ways To Be Precise

There appears to be a great deal of consensus that the rating system, as it is applied to single obligor securities, is fine. Problems arise, however, when the same scale is used for multi-obligor securities.

Some feel that we need to go to a numerical scale when evaluating a multi-obligor bond. Others present a letter-grading scheme that is more precise than the conventions now being used. Whether we choose letters or some numerical convention it really doesn’t matter.

Whether a high risk bond is rated “60” or a “Ca”, a “CC” or a “CCC”, is irrelevant. We simply need to ensure everyone understands the risk and can decide whether the profile is appropriate for their objectives.

What is far more critical than the scoring convention is to ensure credit rating agency analysts have the information, the time, the tools, and the models he or she needs to precisely ascertain risk.

So if a new rating system is devised, it needs a convention that can be attached at the individual obligor level. Each mortgage in a bundle, for example, needs to be scored on its own merits. These scores can then be aggregated and an accurate risk assessment made.

Once a bond is bundled, it’s too late.