为了研究金融市场间非线性的动态相关结构,提出了一类具有变结构特性的分阶段Copula模型以及相应的二元正态Copula模型变结构点的诊断程序.构建了分阶段二元正态Copula-GARCH模型并用于上海股市各板块之间动态相关结构的研究.结果表明,在刻画金融收益序列之间动态相关结构的能力上,变结构二元正态Copula模型优于时变相关二元正态Copula模型.
In order to catch dynamic non-linear dependence between financial markets, a type of staged copula model with structural change is provided. At the same time, a change-points detection program of bivariate normal copula model is given. A staged bivariate normal Copula-GARCH model is constructed to study dynamic dependence structure of Shanghai market. The empirical results show that the bivariate normal copula model with structural change is superior to time-varying bivariate copula model.