Through returning examination of three different distributions(Normal,Student-t,GED),we can see GED distrition can delineate excess kurtosis and fat tail better,and then it can predict the value-at-risk of stock markets of Shanghai more accurately.

• 通过对三种不同分布(N orm al；Student-t；ged)进行返回检验；可看出ged分布能更好地刻画上证指数的尖峰厚尾特征；从而也能更准确地预测沪市的风险值.