Abstract

The purpose of this chapter is to survey risk validation issues for equity portfolio models. Because risk is measured in terms of volatility, an unobservable time-varying variable, there are a number of difficulties that need to be addressed. This is in contrast to a credit risk model, where default or downgrading is observable. In the past, equity risk models have been validated rather informally, but the advent of Basel II has brought about a need for a more formal structure. Furthermore, a number of past high-profile court cases have considered the worthiness of the risk models used by fund managers as part of the broader question of whether the manager has been competently managing risk. To be able to demonstrate that the model user has been managing risk properly, it will be necessary, going forward, to demonstrate that back-testing and other forms of model validation have been practised; and that the systems in place have been designed for the efficient execution of such practise. There have been a number of insightful articles on portfolio risk written by the Faculty and Institute of Actuaries, see Gardner et al. (2000) and Brooks et al. (2002) among others. These papers address in some detail many of the problems concerning portfolio risk and make a number of recommendations. Any equity portfolio risk manager should read them. However, they do not say a great deal about risk model validation. Indeed, there are only occasional sections. For example, in Gardner