基于Copula函数和极值理论的金融传染度量--测度美国次贷危机对重要经济体的传染效应  

Financial Contagion Measurement Based on Copula Function & Extreme Value Theory——Contagion Effects of Sub-Prime Loan Crisis on the World's Most Powerful Economies

作　　者：郭立甫[1,2] 高铁梅[1] 姚坚[1] 

机构地区：[1]东北财经大学经济计量分析与预测研究中心,辽宁大连116025 [2]河北经贸大学信息技术学院,河北石家庄050061

出　　处：《数学的实践与认识》2013年第3期42-55,共14页Mathematics in Practice and Theory

基　　金：国家社会科学基金重大项目(10zd&010);国家自然科学基金(71173029);教育部社科规划基金(10YJA790021);霍英东教育基金会(131086);教育部新世纪优秀人才支持计划项目(NCET-11-1009)

摘　　要：基于Copula函数和极值理论研究美国次贷危机对重要经济体的传染效应,首先根据信息准则来选取Copula函数,然后用Cvm和Ks统计量来检验Copula函数的拟合程度,确保选取合适的Copula函数,并在此基础上计算一般相关系数和尾部相关系数;实证发现使用尾部相关系数度量金融传染并不可靠,因此基于Copula函数和极值理论的POT模型,构造了尾部附近相关系数并通过实证分析了其用于金融传染的有效性.结果表明发达国家所受传染较重,中国所受传染较轻.The paper studies the contagion effects of Sub-prime loan crisis on the world＇s most powerful economies based on Copula function ＆ extreme value, in which Copula function is selected according to information criterion, then goodness of fit is tested by Cvm ＆ Ks statistics, for proper Copula functions. Based on the Copulas, general correlation coefficients and tail dependence coefficients are calculated, which are discovered unreliable for estimating financial contagion. Therefore, the authors construct coefficients of tail near dependence based on Copula function and extreme value theory, then analyze theirs effectiveness empirically. The results show developed countries are infected seriously and china lightly.

关 键 词：金融传染 COPULA函数 极值理论 

分 类 号：F830[经济管理—金融学] O211.4[理学—概率论与数理统计]