Basel is fighting the last war and that rather badly
The Basel Committee has put out a set of proposals for revising the Basel II capital requirements.
One of the things that Basel is now correcting is a discrepancy between the risk level of 99% that was laid down during the market risk amendment of 1996 to Basel I and the level of 99.9% that was laid down in Basel II in 2004 for the banking book. The proposed Guidelines for computing capital for incremental risk in the trading book require risk in the trading book to be measured at the 99.9% level and at a one year horizon. The Committee admits that:
Owing to the high confidence standard and long capital horizon of the IRC, robust direct validation of the IRC model through standard backtesting methods at the 99.9%/one-year soundness standard will not be possible. Accordingly, validation of an IRC model necessarily must rely more heavily on indirect methods including but not limited to stress tests, sensitivity analyses and scenario analyses, to assess its qualitative and quantitative reasonableness, particularly with regard to the model’s treatment of concentrations. Given the nature of the IRC soundness standard such tests must not be limited to the range of events experienced historically. The validation of an IRC model represents an ongoing process in which supervisors and firms jointly determine the exact set of validation procedures to be employed.
I think this is a futile attempt to preserve the 1990s era risk management technology (value at risk, linear correlations and normal distributions) embodied in Basel II. The only way to get to the 99.9% level in any plausible way is to use fat tailed distributions (say student with four degrees of freedom) explicitly and also to move to non linear dependence models (copulas); when one is doing all this, one might as well give up the theoretically discredited value at risk measure and move to a “coherent risk measure” like expected shortfall. Suggesting the use of stress tests as a way to arrive at the 99.9% standard is akin to changing the subject when you do not know what to say.
What I found even more troubling is the following statement in the other consultation document “Revisions to the Basel II market risk framework”
In addition, a bank must calculate a ‘stressed value-at-risk’ based on the 10- day, 99th percentile, one-tailed confidence interval value-at-risk measure of the current portfolio, with value-at-risk model inputs calibrated to historical data from a period of significant financial stress relevant to the firm’s portfolio. For most portfolios, the Committee would consider a 12-month period relating to significant losses in 2007/08 to be a period of such stress, although other relevant periods could be considered by banks, subject to supervisory approval. This stressed value-at-risk should be calculated at least weekly.
This document has been in the making for a longer period and perhaps reflects the Committee’s thinking at an earlier point of time – it still talks of 99% instead of 99.9%. That apart, what puzzled me is the belief that 2007-08 represented the ultimate in terms of financial stress. Since they say 2007/08 and not 2007/2008, it clearly refers to the financial year 2007-08 and excludes the severe stress in the second half of calendar year 2008.
More importantly, the idea that even calendar year 2008 is the ultimate in stress is debatable. There have been no sovereign defaults (ignoring Ecuador) while the big risk for 2009 and 2010 is certainly the possible default of a G7 sovereign and the related possibility of the break-up of the Eurozone. Risk managers who think that the worst is over in the current crisis are not worth their salaries today. I am shocked that the Basel Committee is encouraging this kind of shoddy thinking.
Posted at 1:51 pm IST on Wed, 21 Jan 2009 permanent link
Comments