Application of Extreme Value Theory to Estimate Tail Thickness for Asset Return Distributions

Abstract

The main topics covered in this chapter are:assessment of value-at-risk (VaR) and average value-at-risk (AVaR) through extreme value theory (EVT) using the peaks-over-threshold method;how to combine the standard GARCH model which takes into account the clustering of volatility with EVT for risk estimation;extensive in-sample and out-of-sample analysis of the upper and the lower tail thickness of the return distributions of market indices;AVaR-based and standard VaR-based statistical tests to assess the out-of-sample behavior of the risk model.