基于Markov区制转换模型的极值风险度量研究  

Study on Extreme Risk Measurement Based on Markov Regime Switching Model

作　　者：张保帅[1] 周孝华[1] 李强[1] 冯梦雨[2] 

机构地区：[1]重庆大学经济与工商管理学院,重庆400030 [2]中国人民银行延安市中心支行,陕西延安716000

出　　处：《统计与信息论坛》2012年第9期26-31,共6页Journal of Statistics and Information

基　　金：中央高校基本科研业务费资助项目<基于Copula理论的地方投融资平台风险研究>(CDJXS11021112)

摘　　要：将马尔科夫区制转换模型与极值理论相结合研究金融风险度量问题。首先用SWARCH-t模型捕捉收益率序列的剧烈波动和结构变换特征,然后将收益序列转化为标准残差序列,在此基础上通过SWARCH-t模型与极值理论相结合拟合标准残差的尾部分布,进而构建基于SWARCH-t-EVT的动态VaR模型,最后对模型的有效性进行检验。研究表明,SWARCH-t-EVT模型能够有效识别上证综指的波动区制特征,且能有效合理地测度上证综指收益风险,尤其在高的置信水平下表现更好。The main innovations of this paper is to study financial risk measurement by combining Markov switching model with extreme value theory. This paper captures the volatility and structure transformation characteristics of the return series by SWARCH--t model, then transforms the return series into the standard residuals series, on this basis, fits the tail distribution of the standard residuals series by combining Markov switching model with extreme value theory, and establishes a new financial risk measurement mode based on the SWARCH--t--EVT dynamic VaR model, finally tests the validity of the model. The results show that the SWARCH--t--EVT model can effectively identify SSEC Index volatility regime eharaeteristics, also effectively and reasonably measure the risk of SSEC Index, especially in the high confidence level.

关 键 词：SWARCH-t-EVT模型 极值风险 阀值 

分 类 号：F831.5[经济管理—金融学]