A leading regulator gets it. Finally. Maybe.

Sheila Bair, chairman of the Federal Deposit Insurance Corporation, is trying to wake her fellow supervisors up to the issues in the internal models approaches to capital in Basel 2. She is reported to have said:

Current financial market turmoil has shown up the weaknesses of the models used in the advanced approaches to assessing credit risk under the international Basel II bank safety rules...

The first lesson of the crisis in terms of the Basel II capital adequacy rules is “beware models!”

Bair is right of course. The models are necessarily inaccurate, since calibration is problematic, procyclical, and tend to support asset price bubbles.

The weaknesses “startlingly revealed” in the Basel II models by the turmoil “are a very bright, flashing yellow light warning us to drive very carefully,” ...

The FDIC chairman has previously expressed doubts about what she has described as the “untried” risk models underlying the advanced Basel II approaches that banks can use to assess their credit and operational risks and determine the minimum capital they need absorb shock losses.

Remember just because it backtests for some period doesn't mean it is right. A stopped clock backtests well in the few seconds around the time it is stopped at.

Update. The Telegraph (yes, I know, but I couldn't find a reference from a reputable paper in a hurry) carries a report on a slightly alarmist but interesting speech from Peter Spencer of the Ernst and Young Item Club. Spencer says that the Basel 2 rules

..are the root cause of the crunch and were serving to worsen the City's plight.

Dismissing the assumption that banks are not lending to each other on the money markets because they lack confidence in each others' potential solvency, he argued that they were, in practice, prevented from lending the cash at all because it could leave their balance sheets falling foul of the Basel regulations.

Presumably the idea is that now lines of credit below 364 days are not free in capital terms, banks are rationing capital and hence not lending. This seems unlikely, especially as the capital usage of short term interbank lending is low. I wonder what evidence Spencer has.