Mardi | 2012-04-03
B103

Christophe BOUCHER – Jon DANIELSSON – Bertrand MAILLET – Patrick S. KOUONTCHOU

The recent experience from the global financial crisis has raised serious doubts about the accuracy of standard risk measures as a tool to quantify extreme downward risks. Risk measures are hence subject to a “model risk” due, e.g., to the specification and estimation uncertainty. Therefore, regulators have proposed that financial institutions assess the “model risk” but, as yet, there is no accepted approach for computing such a risk. We propose a general framework to compute risk measures robust to the model risk, while focusing on the Value-at-Risk (VaR). The proposed procedure aims empirically adjusting the imperfect quantile estimate based on a backtesting framework, assessing the good quality of VaR models such as the frequency, the independence and the magnitude of violations. We also provide a fair comparison between the main risk models using the same metric that corresponds to model risk required corrections.