中文摘要：

本文对EGARCH模型进行了推广,得到了EPGARCH模型。对该模型的参数估计采用了带约束条件的非线性规划方法。利用直方图、时序图和Q统计量检验等方法对沪深指数收益序列进行了特征分析,得出收益序列具有高峰厚尾和波动聚集性。通过对沪深指数的VaR计算,得到在金融风险度量中基于稳定分布的EPGARCH模型比基于正态分布的EPGARCH模型更加有效。

英文摘要：

We proposed EPGARCH model, which is a generalization of EGARCH model. Nonlinear constrained programming method is adopted for parameter estimation of the model. Characteristics of return series of Shanghai and Shenzhen stock indices were analyzed with histogram, time series plot and Q statistic test. It shows that return series have some properties such as leptokurtosis, fat tail and volatility clustering. Calculated VaR for Shanghai and Shenzhen stock indices shows that EPGARCH with stable distribution is more efficient than EPGARCH with normal distribution in financial risk valuation.