国际市场恐慌情绪传染分析与风险预警

Panic Sentiment Infection Analysis and Risk Warning in International Market

市场情绪是影响市场走向的重要因素。为了考察多个市场中恐慌情绪的联动问题,本文采用半参数的时变藤Copula函数对多市场情绪的联动结构进行刻画,并利用支持向量机为其估计变量的边缘密度函数,构建不依赖模型与分布假设的SVM-Dynamic Vine Copula系统;以美国、韩国、香港三个市场的VIX恐慌指数为研究对象的实证发现,三个市场间的相依结构存在显明的时变效应,且当前市场相依结构与次贷危机前期非常相似,作为预警信息值得关注;此外,压力测试发现,市场反应存在不对称性,美国市场更容易受香港市场的影响。

Market sentiment is an important factor affecting the market trends. In order to evaluate the hnkage among the panic sentiment of multi markets, the paper proposes semiparametric dynamic vine copula to describe the linkage, and u- ses support vector machine to estimate the density function, then it gets SVM- Dynamic Vine Copula system which doesn＇t depend on assumptions of distributions and models; based on the VIX indexes of US, Korea and HK, the result shows that the dependencies of the 3 markets is time varying, and the current dependencies of them are similar to early subprime crisis which should be attended; stress test analysis gives the results that the reaction of markets is asymmetric and US market is more sensitive to HK market.