Research on Extreme Risk of Stock Market based on POT Model

摘要

To remedy the shortcomings of VaR model under the assumption that the sequence is normally distributed,this paper employs POT model to empirically estimate extreme risk of Shanghai stock market.The results show that POT model is effective to study sequences with fat tail.At the same time,the raising limit restrains the heterogeneity of extreme values,which results in a phenomenon that VaR model overestimate the extreme risk compared to POT model.In such circumstances,POT model is more effective than VaR model under normal distribution,then CVaRPOT is an effective indicator to reflect the nature of extreme risk instead of VaRPOT.