中国A股行业溢出效应的实证分析——基于小波分解和多元GARCH-VAR模型  

An Empirical Study on Industry Spillovers in China Stock Market——Based on Wavelet Analysis and Multivariate GARCH-VAR Model

作　　者：苏民[1] SU Min(School of Economics and Management, Taiyuan University of Technology, Taiyuan 030024 ,Chin)

机构地区：[1]太原理工大学经济管理学院,山西太原030024

出　　处：《系统工程》2017年第12期15-24,共10页Systems Engineering

基　　金：山西省软科学研究一般项目(2017041017-1);太原理工大学校基金资助项目(2016RS13)

摘　　要：本文首先对A股市场进行聚类分析,找出比较典型的3个代表性行业,之后进行小波分解,提取出这些行业数据的低频部分(趋势)和高频部分(周期和不规则)。然后,对行业的低频数据建立VAR模型,进行Granger因果关系分析和方差分解,以检验行业间的均值溢出效应。研究结果表明A股市场不同行业之间的均值溢出效应很显著。最后,对行业的高频数据部分,进行DCC-GARCH和BEKK-GARCH建模,分析不同行业间的时变关系和波动溢出效应。研究结果表明A股市场不同行业间的相关性是时变的,行业间的双向波动溢出效应明显,但单向溢出效应中有个别行业并不显著。Using a cluster analysis,this paper firstly finds three representative industries and applies a wavelet analysis to extract the low-frequency part（trends）and high-frequency part（cycle and irregular）.Then,this paper establishes a VAR model based on the low-frequency data to do Granger causality analysis and variance decomposition for testing inter-industry mean spillovers.According to the results,the mean spillover effect between different industries in the A-share market and is significant.Finally,based on the high-frequency data,this paper establishes the DCC-GARCH and BEKK-GAECH model for analyzing time-varying relationships and volatility spillovers effect between different industries.The results show that the correlation between different Industries in A-share market is time-varying and a bi-directional volatility spillover effect is obvious.But,unidirectional spillover effect is not significant in some Industries.

关 键 词：行业效应 小波分解 多元GARCH 波动溢出 

分 类 号：F831[经济管理—金融学]