Journal of Risk Model Validation

This issue of The Journal of Risk Model Validation presents, as usual, four papers, two of which discuss that hardy perennial, value-at-risk (VaR). It is extraordinary, after all the research that has been done, that one can find new and interesting things to say about VaR. In my view, the authors here have succeeded in doing so. The other two papers are both welcome contributions in examining different features of debt modeling and validation.

In the first paper, titled "Value-at-risk estimation with the Carr-Geman-Madan-Yor process: an empirical study on foreign exchange rates", Sun-Yong Choi investigates whether the Carr-Madan-Geman-Yor (CGMY) distribution is appropriate for describing the fat tails of the distribution of foreign exchange rate return. VaR is estimated for this model and for a number of others, and attractive properties are found for the distribution.

In "Research on equity release mortgage risk diversification with financial innovation: reinsurance usage", the second paper in this issue, Kuo-Shing Chen looks at equity release mortgages and provides fair value pricing based on Brownian motion assumptions and recognition of the options structure. Risk validation is carried out via some numerical calculations. This is one of the very few papers in this area that we have published and it is of considerable general interest.

Our third paper, "Testing value-at-risk models in emerging markets during crises: a case study on South Eastern European countries" by Nikola Radivojevic, Nikola V. Curcic, Dragana Milojkovic and Vuk S. Miletic, contains further discussion on VaR. The authors carry out a case study of South Eastern European countries to see if VaR can be applied to emerging markets. They conclude that the most popular VaR models do not perform all that well in this context and suggest refinements to address this.

The issue's fourth paper, by Mark Rubtsov and Alexander Petrov, is called "A point-in-time-through-the-cycle approach to rating assignment and probability of default calibration". Here, the authors propose a methodology for constructing through-the-cycle rating grades and assessing the resulting degree of point-in-time-ness. The models are calibrated and the results are used to understand validation tests. The authors give an example using a portfolio of corporate customers; other applications are possible.

Looking at the overall health of the journal, I am pleased to reveal that we are now receiving a regular supply of good submissions, and there are other indicators of good publishing health. Risk model validation continues to be of significant importance and is likely to become more, rather than less, important going forward.

Steve Satchell
Trinity College, University of Cambridge