基于极值理论的沪综指尾部风险度量  

Fat-tails Measurement and Risk Estimation of Composite Index of Shanghai Stock Exchange on Extreme Value Theory

作　　者：高岳[1] 朱宪辰[1] 

机构地区：[1]南京理工大学经济管理学院,江苏南京210094

出　　处：《财贸研究》2009年第5期102-108,共7页Finance and Trade Research

基　　金：国家自然科学基金项目"异质性与共享资源自发治理的集体行动研究"(批准号:70573046);教育部高等学校博士学科点专项科研基金资助课题(20060288016);江苏省高校哲学社会科学基金重点项目(06SJB790001)

摘　　要：以沪综指为实证研究对象,利用AR-GARCH模型捕获沪综指对数收益序列中的自相关和异方差现象,获得近似独立同分布的残差序列,再利用极值理论中的POT方法对残差序列进行极值分析,并采用给出了极值分布的各参数极大似然估计,进一步计算置信水平下的VaR值,结果表明,根据AR-GARCH-POT方法所得的VaR对沪综指实际损失风险的刻画准确。而基于所得的VaR序列分析沪综指收益率的系统性风险历史波动趋势,结果发现系统性风险历史趋势存在先高后低,且自2006年后逐年升高的特征,目前阶段的总体性风险处于较高水平。This paper detects shanghai composite index from 1996 to 2008. An AR-GARCH Model is firstly build to fit the correlationship and heteroskedasticity of return series. POT method is employed to analyze innovations and estimate the interval of VaR. The parameters are solved by MLE. Result shows that the VaR series solved by AR-GARCH-POT model have a good coverage rate to factual lost. Finally, historical VaR changing trend is analyzed to study the system risk in the Shanghai Stock Exchange, which indicates that the system risk trend was high in the early years and was getting lower in following years, it increased from 2006, and it is in a comparatively high level now.

关 键 词：POT模型 VAR 厚尾 风险度量 

分 类 号：F830.9[经济管理—金融学] F224.0