Abstract
The purpose of this study is to evaluate the estimating ability of GAS models in the computation of the value-at-risk by applying the extreme-value theory. Our approach is the limiting result of an infinity shift of location. In this work, we use the generalized pareto distribution since it plays a central role in modelling heavy tail phenomena in many applications. A simulation study is performed to assess the estimated value-at-risk. Moreover, we examine the performance of the proposed method with daily returns of three stock market indices. The results prove that the presented approach gives good predictions for all indices.