Journal of Risk Model Validation

Steve Satchell

University of Cambridge

This issue looks at two important contemporary matters: the first of these is the implementation issues for Basel II and the second is to do with the Summer of 2007. It has been recognized by anybody who has tried to validate an internal ratings system for a bank that standard mathematical methods require some sort of adaptation. There are many reasons for this. The first paper by Blochwitz outlines a comprehensive approach of validation for the internal ratings-based (IRB) approach in Basel II. This will be an essential read for anybody involved in validation in this area. The second paper by van der Burgt addresses a smaller problem within Basel II but nevertheless is one of considerable importance. Many banks and lending institutions simply do not have sufficient default data to build sensible measures of default probabilities based on the experiences in their mortgage book. This paper suggests procedures that one might follow in this situation.

The Summer of 2007 looks likely to produce an enormous amount of literature in due course and The Journal of Risk Model Validation is delighted to publish such relevant papers in a timely fashion. The first of these by Nyström focuses on the issues of model complexity in rating structured financial instruments. This necessarily brings into the discussion the role of the rating agencies that have already been criticized in the popular press. This paper does not inflame the discussion but instead rather focuses on methods that create universal, standardized, transparent and mathematically formalized approaches to the rating of structured finance products. I strongly recommend this paper to all our readers.

The second paper on the Summer of 2007 looks more at Basel II issues; in particular, value-at-risk (VaR). If we characterize the Summer as being a regime shift, not necessarily once and for all, but possibly reverting back at some stochastic date, then it would be interesting to model VaR under these assumptions. The authors, de la Pena and Rivera, go further than this and provide a dynamic backtesting approach that allows one to detect a regime shift at or near the point of occurrence. This has the further benefit that it can be used for stress testing for conventional models.

Are there any links between these two themes? One possibility is that Basel II became such an all absorbing issue for the regulators that they took their eye off the subprime ball. Whilst it would be difficult to prove this, the coincidence of these two major events is probably not entirely random.

We have completed the first year of the journal and I am extremely pleased by its progress. Subscriptions have exceeded expectations and we have been able to attract articles that are both scholarly and of interest to practitioners. We have an active and committed editorial board and in terms of timing, starting a risk model validation journal just when risk models decided to seize up over the summer makes one feel that there is such a thing as luck.