Description: This article examines how parameter uncertainty leads to uncertainty in Value at Risk VaR estimates. Lacking a closed form solution, we use a first order approximation of VaR to map parameter changes into changes in the VaR value. When the asymptotic distribution of parameter estimates is normal, this implies a normal approximation of the confidence interval about VaR. We describe this approximation in detail and explore some extensions to risk factor modeling.

Description: Because actuaries fit models for a variety of situations, particularly with regard to data modifications, it is useful to have a single approach. Though not optimal or exact for a particular model or data structure, the method should be reasonable for most all settings. Such a method is proposed in this article. Actuarial Research Clearing House 2003, VOL. 1