Measure of Stock Market Risk Based on Extreme Value Theory

Taking data of domestic listed company for an example and considering reality market backgrounds and economic significances this paper measures stock risk by POT model of extreme value theory and conditional Value-at-Risk (CVaR). That is to calculate VaR and CVaR by using EVT. It is found that VaR model based on generalized Pareto distribution describes stock risk and CVaR can effectively cover risk in these days that VaR estimation does not function. The results of model diagnosis imply that EVT is an effective method that measures risk loss under extreme market conditions. It can accurately portray quantile of tail and is proper to be used in studying financial asset returns with fat tail.