Detecting Estimation Errors in Value at Risk

For a variety of reasons, many bank and corporate finance executives are implementing enterprise risk management systems to better understand and manage the firm’s risks. Our objective is to provide tools that evaluate the accuracy of the output from these systems. Especially in markets where the risks are large and complex, such as energy firms, it is crucial to establish the accuracy of the risk system. Often, firms are so pleased to finally have an enterprise risk management system in place that they may forget to address whether the resulting daily risk assessment has any legitimacy. The ability to certify that a system is without significant bias is extremely useful for building confidence in the risk-management decision-making process. Also, there is a direct feedback loop between theoretical market models and valuation models used within these systems and certifying the system. It is possible to identify superior market models and superior valuation models through the effort of system certification. Although we focus in this paper on value at risk, the methodologies developed are directly applicable to many other risk measures.