论文部分内容阅读
本文提出一个新的时变最优Copula模型,可以准确识别二元时间序列任意时点最优的相依结构。该模型构造了半旋转copula以刻画非对称的反向相依关系,并引入独立性的无分布检验证实相依关系的存在性。同时,我们对能源商品市场(原油、天然气)、外汇市场间动态相依关系进行了实证分析,实证结果表明跨市场相依结构类型确实是时变的,突发事件往往是相依结构突变的主因。另外,时变最优Copula模型的主要优势在于不仅能够捕捉相依方向和相依强度的动态性,还能有效捕捉相依结构类型的动态性。
In this paper, a new optimal time-varying Copula model is proposed, which can accurately identify the best dependent structure at any time in binary time series. The model constructs a semi-rotating copula to characterize the asymmetric inverse dependence and introduces the independence-free distribution test to confirm the existence of the dependence. At the same time, we empirically analyze the dynamic dependence between the energy commodity markets (crude oil, natural gas) and foreign exchange market. The empirical results show that the cross-market dependent structure types are indeed time-varying, and emergencies are often the main cause of the sudden changes in the dependent structures. In addition, the main advantage of the time-varying optimal Copula model is the ability to capture not only the dynamics of dependent directions and dependent intensities but also the dynamics of the dependent structure types.