基于非线性相依的市场间金融传染度量——测度2015年中国股灾对重要经济体的传染效应  

Financial contagion measurement between nonlinear inter-dependent markets:Detecting the contagion effects of Chinese stock market crash in 2015 on the world’s important economies

作　　者：苑莹[1] 王海英[1] 庄新田[1] YUAN Ying;WANG Haiying;ZHUANG Xintian(School of Business Administration,Northeastern University,Shenyang 110169,China)

机构地区：[1]东北大学工商管理学院,沈阳110169

出　　处：《系统工程理论与实践》2020年第3期545-558,共14页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(71271047);国家社会科学基金(18BJY238);教育部人文社会科学基金(17YJCZH235);中央高校基本科研业务费(N180614002)。

摘　　要：针对真实市场间所具有的非线性、尾部极值相依性及时变性等相依特征,本文以2015年6月中国爆发的股灾为背景,从传染效应的存在性、传染强度、传染方向三个视角深入研究中国股市对日本、美国、韩国三个重要经济体股市的金融传染效应.首先,将极值理论(extreme value theory,EVT)与时变Clayton Copula函数相结合构建时变Clayton Copula-EVT模型,估计下尾极值动态相依系数,并进行统计检验,发现中国股市对日本、美国股市存在风险传染效应,而对韩国股市不存在风险传染效应;进一步地,量化中国股市对日本、美国股市的风险传染强度,发现中国股市对日本、美国股市的传染强度都较大,且中国股市对美国股市的传染强度明显强于其对日本股市的传染强度;最后,利用基于时间延迟的去趋势交叉相关性分析方法研究中国股市与日本、美国股市之间的风险传染方向,发现股灾后市场间风险传导方向发生了改变,风险主要由中国股市传染至日本、美国股市.上述实证结果为深入研究市场间金融风险传染的非线性相依特征机理提供了有益的参考.Considering the complex interdependence characteristics such as nonlinearity,tail extreme dependence and time-varying pattern between real financial markets,this paper quantitatively studies the financial contagion effects of Chinese stock market crash in June 2015 on Japanese stock market,U.S.stock market and South Korean stock market from the perspectives of the contagious existence,contagious intensity and contagious direction.Firstly,combining the extreme value theory(EVT)with the timevarying Clayton Copula function,we construct a time-varying Clayton Copula-EVT model to examine lower tail extreme value dynamic dependence.And an additional statistical test helps us verify that there is financial contagion effect between Chinese stock market and Japanese stock market as well as between Chinese stock market and U.S.stock market,however,there is no contagion effect between Chinese stock market and South Korean stock market.Furthmore,the financial contagious intensities are considered between Chinese stock market and Japanese stock market,U.S.stock market respectively.It is found that the contagious intensities are stronger between Chinese stock market and Japanese stock market,U.S.stock market respectively.In addition,the contagious intensity between Chinese stock market and U.S.stock market is much stronger than that between Chinese stock market and Japanese stock market.Finally,we investigate the risk contagious directions between Chinese stock market and Japanese stock market,U.S.stock market respectively using the time-delayed detrended cross-correlation analysis.It is shown that the directions of risk transmission are changed after the breakout of Chinese stock market crash,and the directions of risk contagion are in general from Chinese stock market to Japanese stock market and U.S.stock market.These results can help us to get a better understanding of the nonlinear dependence mechanism of financial risk contagion.

关 键 词：金融传染 极值理论 时变Clayton COPULA 基于时间延迟的去趋势交叉相关性分析 

分 类 号：F830.99[经济管理—金融学]