The current subprime crisis has prompted us to look again into the nature of risk at the tail of the distribution. In particular, we investigate the risk contribution of an asset, which has infrequent but huge losses, to a portfolio using two risk measures, namely Value-at-Risk (VaR) and Expected Shortfall (ES). While ES is found to measure the tail risk contribution effectively, VaR is consistent with intuition only if the underlying return distribution is well behaved. To facilitate the use of ES, we present a power function formula that can calculate accurately the critical values of the ES test statistic. This in turn enables us to derive a size-based multiplication factor for risk capital requirement.
Value-at-Risk expected shortfall tail risk contribution saddlepoint technique risk capital