基于Kendall’sτ秩相关系数的优越性和定义,本文提出了新的具有明确经济意义的动态条件相关copula模型,将常用的Gaussian、Clayton和Gumbel函数统一根据该演化方程实现动态化,构造出三种Kendall’sτ动态条件相关copula模型,可用于刻画不同的相关模式。这些模型不仅参数少、容易估计,避免了现有动态条件相关copula模型构建方法各异导致的在实证中不利于比较的缺点,而且能够进行多步向前预测,有效地减少了进行样本外预测时的计算量,从而为刻画时变、非线性、非对称性和尾部相关等复杂的动态相关模式提供了新方法。%Based on the superiority and definition of Kendall＇s τ rank correlation coefficient,this paper puts forward a new Kendall＇s τ Dynamic Conditional Correlated Copula model which is of economic significance.And we construct three kinds of Kendall＇s τ-Dynamic Conditional Correlated Copula model to describe different correlations by applying this new model to Gaussian,Clayton and Gumbel copula.These models not only have fewer parameters and are less time-consuming in calculation,but also overcome the shortcoming of the existing Dynamic Conditional Correlation Copula models that it is not convenient to evaluate because of their different constructing methods.Moreover,this new model can carry out multi-step ahead prediction with less time-consuming in out-of-sample prediction,providing a new approach to characterize the time-varying,nonlinear,non-symmetry,tail dependence and other complicated correlation.
展开▼
