基于动态因子Copula模型的收益率波动相依性分析

Analysis on Rate of Return Fluctuation Dependence Based on Dynamic Factor Copula Model

金融市场作为风险研究的对象,需要构建联合分布来分析其变量间的相依性变化。针对高维变量联合建模带来的“维数灾难”,动态因子Copula模型能有效处理金融数据的不对称性、厚尾性和动态相依性,解决参数估计问题。以2008年1月至2020年12月22个金融机构股票收益率为研究对象进行实证分析,研究结果表明:银行业与其他金融机构的相依性普遍更强,兴业银行受公共因子的影响最大,因子载荷均值为1.7973;保险业与其他金融机构相依性偏低,中国太保受公共因子影响最小,因子载荷均值为1.2013;证券业居中。

As the key object of risk research,financial market needs to construct joint distribution to ana-lyze the change of dependence among variables.In view of the"dimension disaster"caused by the joint modeling of high-dimensional variables,the dynamic factor copula model can effectively deal with the asymmetry,thick tail and dynamic dependence of financial data,and solve the problem of parameter esti-mation.Taking the stock returns of 22 financial institutions from January 2008 to December 2020 as the re-search object for empirical analysis,the research results show that:the dependence between the banking in-dustry and other financial institutions is generally stronger,Industrial Bank Co.,the most affected by pub-lic factors and the average factor load is 1.7973;the dependence between the insurance industry and other financial institutions has the lowest dependence,CPIC is least affected by public factors and the average factor load is 1.2013;the securities industry is in the middle.