摘要
运用股指高频数据,首先构建基于一般已实现波动率与双幂次变差的RV-regular vine copula模型,比较分析两个模型在刻画资本市场之间的波动跳跃结构的差异;然后,考虑股票价格存在跳跃的情形,分析离散跳跃的统计特征,研究资本市场之间跳跃相依结构,并通过模拟计算市场在险价值来探讨市场之间的跳跃相依风险。研究发现:由于跳跃的存在,采用双幂次变差所构建的RV-regular vine copula模型在参数估计结果方面表现更佳;跳跃序列存在尖峰厚尾的特征以及跳跃聚集的现象,而跳跃相依风险可以通过构建合适的资产组合来降低。
This paper is firstly to model the RV-regular vine copula framework based on realized volatility and bi-power variation by using high frequency data of stock indexes in capital markets,and compare the differences of capturing the volatility jump among the markets.Then it takes jumps of stock prices into consideration,analyzes the statistical features of the discrete jumps,and tests the dependence among the jumps of the markets.Further,it simulates Value-at-Risk for discussing the jumping dependence risks in the markets.The empirical results show that the built RV-regular vine copula model based on bi-power variation is optimal to capture the dependence of the capital markets.Meanwhile,leptokurtosis,fat tails and clustering are displayed in the serials of jumps.And the induced jumping dependence risks could be reduced by building proper portfolios.
出处
《统计与信息论坛》
CSSCI
北大核心
2017年第11期34-41,共8页
Journal of Statistics and Information
基金
教育部人文社会科学研究青年基金项目<带跳跃的碳排放交易市场状态相依结构:基于高频数据的研究>(16YJC790030)
安徽省自然科学基金项目<基于高频数据的资本市场跳跃相依与极值风险模型构建及实证研究>(1708085QG163)