基于混合Copula函数的金融市场非线性极端风险传染研究  

The Study on the Financial Market Nonlinear Risk Contagion Based on Mixed Copula Model

作　　者：淳伟德[1] 付君实 赵如波[1] 

机构地区：[1]成都理工大学商学院,四川成都610059

出　　处：《预测》2015年第4期53-58,共6页Forecasting

基　　金：国家社会科学基金资助项目(12BGL024);国家自然科学基金资助项目(71171025);四川省科技计划资助项目(2012ZR0045;2013ZR0067);成都理工大学优秀科研创新团队资助项目(KYTD201303)

摘　　要：本文在金融市场典型事实约束下,运用ARFIMA和FIAPARCH模型分别对金融收益率与波动率进行建模,以排除金融市场典型事实对风险传染效应的影响,进而运用极值理论(EVT)对标准收益的极端尾部建模,并运用由Clayton、Frank和Gumbel组成的混合Copula模型对金融风险传染进行实证研究。研究结果表明:次贷危机的爆发对于股市的长记忆性具有一定的影响;在次贷危机后,中国大陆股市与香港股市、日本股市以及新加坡股市发生了显著的极端风险传染,而香港股市与日本股市、新加坡股市以及日本股市与新加坡股市之间未发生显著的极端风险传染。The paper uses the ARFIMA and FIAPARCH model to model return and volatility based on some stylized facts respectively, which can eliminate the effect on the financial risk contagion caused by stylized facts, and uses extreme value theory to measure the extreme tail. At last, the financial risk contagion is researched by the mixed copula model, which is composed of Clayton, Frank and Gumbel. The empirical results show that the outbreak of the subprime crisis has a certain impact on the long memory of stock markets, the extreme risk contagion between Chinese mainland stock market, Hong Kong stock market, Japanese stock market, and Singapore stock market is obvious after the subprime crisis. However, there is no obvious extreme risk contagion existing between Hong Kong stock market, Singapore stock market and Japanese stock market, nor does between Japanese stock market and Singapore stock market.

关 键 词：极端风险传染 典型事实 极值理论 混合Copula 

分 类 号：F830.9[经济管理—金融学]