基于动态极值VaR股指分析的中国行业风险研究  

Study of Chinese industry risk based on stock index analysis of dynamic EVT-VaR

作　　者：刘庆斌[1] 姜薇[2] 

机构地区：[1]天津工业大学管理学院,天津300160 [2]军事交通学院基础部,天津300161

出　　处：《东南大学学报（自然科学版）》2009年第6期1246-1251,共6页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(70702012)

摘　　要：应用动态极值VaR模型分析中国行业股票价格指数,对不同市场条件下中国行业风险进行实证研究.在McNeil等提出的GARCH模型与极值理论基础上,运用Huisman等提出的VaR法修正上述风险值估计方法,构建了动态极值VaR模型.在该模型的构建中考虑了风险因子的时变特性,采用EVT方法对风险因子的厚尾特性进行建模,简化了风险值的估计过程,并提高了估计的准确性.动态极值VaR模型能够较好地刻画风险尾部分布特性,同时能够比较准确地度量所研究时间段内的市场风险和行业风险.研究结果表明,不同市场条件下各行业的风险特性具有显著的变化.最后利用Kruskal-Wallis一致性检验对上述结论进行了验证.据此,在资产组合管理过程中,有必要依据市场环境和行业风险特性对资产配置比例进行调整.Chinese industry stock price index is analyzed and empirical studies on China＇s industrial risk under different market conditions are made with GARCH-EVT（generalized autoregressive conditional heteroskedasticity-extreme value theory） VaR（value at risk） model.On the basis of the GARCH model and extreme value theory presented by McNei and other scholars,this paper modifies the above method of risk evaluation and constructs a GARCH-EVT VaR model by applying VaR method put forward by Huisman and others.Taking time-varying property of the risk factor into consideration while constructing the model,this paper models the fat-tail of the risk factor using EVT,thus simplifying the evaluation process of VaR and improving the accuracy of evaluation as a result.The GARCH-EVT VaR model can better describe the tail of the distribution,and more accurately measure the market risk and industrial risk within studied period as well.The results indicate that risks of different industries vary in different market conditions.The above conclusions are validated by using Kruskal-Wallis Consistency tests.It is essential to adjust the distributions of assets in accordance with market conditions and features of industrial risks in the process of access combination management.

关 键 词：行业股票价格指数 GARCH-EVT POT模型 风险值 行业风险 

分 类 号：F830.9[经济管理—金融学]