基于Rolling Regression及VAR-DCC-GARCH模型的股市时变协动研究——发达市场对中国大陆股市存在金融传染吗?

Research on the Time-Varying Interdependence among Stock Markets based on Rolling Regression and VAR-DCC-GARCH Model——Is There Financial Contagion Effect from Developed Market to China Mainland?

本文以2002年5月至2013年12月为研究区间,利用协整模型、滚动回归模型及VARDCC-GARCH模型对中国大陆股市与日本、香港、新加坡股市间的时变协动性进行了研究。结果表明:长期内,中国大陆股票市场与香港股市存在时变协整关系(截距与斜率均具有时变性),与新加坡股市存在着弱协整关系,但与日本股票市场不存在协整关系;中国大陆股市与香港股市、日本股市间的时变相关系数具有长记忆特征,而与新加坡股市时变相关系数并不具有持续性;亚洲发达经济体对中国大陆股市存在显著的金融传染效应。

The paper investigates the time -varying comovements between China Mainland stock market and developed markets in Asia such as Hong Kong , Japan and Singapore using cointegration , rolling regression and VAR -DCC -GARCH models respectively based on daily data ranging from May 2002 to December 2013 .The results show that： the time-varying （ slope and intercept ） cointegration between China Mainland and Hong Kong is significant and strong and the cointegration between China mainland and Singapore is weak , while there is no cointegration relationship between China Mainland and Japan; The DCCs between China Mainland and Hong Kong or Japan has long memory of persistence, while there is no persistence with Singapore;Asia′s developed economies has significant financial contagion effect on China Mainland stock market .