我国上证股市波动率关联传导网络分析  

An Analysis of the Correlation Transmission Network of Volatility in China's Shanghai Stock Market

作　　者：罗玉波 何星佐 李梦尧 LUO Yu-bo;HE Xing-zuo;LI Meng-yao(School of Mathematics and Statistics,Beijing Technology and Business University,Beijing 100048,China)

机构地区：[1]北京工商大学数学与统计学院,北京100048

出　　处：《数学的实践与认识》2022年第10期75-85,共11页Mathematics in Practice and Theory

基　　金：北京工商大学数据科学研究院平台建设项目(2022)

摘　　要：将增广因子模型和广义动态因子模型相结合,提出了增广广义动态因子模型,在此基础上用长期方差分解网络模型挖掘上证股市波动率关联传导网络.模拟实验结果表明,所提出的方法不仅能够较好处理维度灾难和股票数据的尖峰厚尾性问题,还能利用额外可观测信息提高模型参数估计的表现.利用本文的方法,根据2002年至2021年我国上证180指数所有成分股的日度收益率数据,构建了六个时期的上证股市波动率关联传导网络.实证发现:1)我国上证股市波动率关联传导网络的动态演化趋势明显,各时期处于网络中心位置的行业不尽相同,且各行业风险溢出水平也有动态调整;2)在大多数时期,金融和能源两个行业都处于网络的核心位置,并在风险传导路径中扮演着至关重要的角色,当前阶段信息技术和电信行业的重要性也明显提升。This paper combines the augmented factor model with the generalized dynamic factor model,and proposes the augmented generalized dynamic factor model.On this basis,the long-term variance decomposition network model is used to mine the volatility correlation conduction network of Shanghai stock market.Simulation results show that the proposed method can not only better solve the dimension disaster and the problem of peak and thick tail of stock data,but also improve the performance of model parameter estimation by using additional observable information.Using the method of this paper,according to the daily return data of all constituent stocks of Shanghai 180 index from 2002 to 2021,this paper constructs the volatility related transmission network of Shanghai stock market in six periods.The empirical results show that:1)The dynamic evolution trend of volatility related transmission network in Shanghai stock market is obvious,the industries in the center of the network are different in each period,and the Risk Spillover level of each industry also has dynamic adjustment;2)In most periods,finance and energy industries are always at the core of the network and play a vital role in the risk transmission path,at the current stage,the importance of information technology and telecommunications industry has also increased significantly.

关 键 词：广义动态因子模型 增广因子模型 长期方差分解网络 

分 类 号：F832.51[经济管理—金融学]