我国股市收益率长记忆性的区制转移分析  

Regime-Switching Analysis of Long Memory of Stock Market Returns in China

作　　者：王亮 Wang Liang(Economics and Management College,Dalian Minzu University, Liaoning, 116600)

机构地区：[1]大连民族大学经济管理学院,辽宁大连116600

出　　处：《技术经济与管理研究》2018年第10期68-72,共5页Journal of Technical Economics & Management

基　　金：国家社科基金一般项目(15BMZ090);教育部人文社会科学研究规划基金项目(17YJA790072)

摘　　要：本文使用MS-ARFIMA模型对沪深两市收益率轨迹的动态特征进行了新的计量检验。结果显示:不仅均值和波动性存在着明显的结构非线性,而且沪深两市收益率的长记忆性也表现出显著的区制转换动态。沪市收益率在"熊市"区制下表现为具有均值回复能力的长记忆过程(d1=0.7732),在"牛市"区制下表现为短记忆过程(d2=-0.1854)。而深市收益率在"熊市"区制下表现为短记忆过程(d1=-0.1575),在"牛市"区制下表现为协方差平稳的长记忆过程(d2=0.2625)。这一新的研究结论意味着我国股市对信息冲击的反应模式和市场有效性也存在相应的区制转移变化。这要求有关部门在制定股市政策时,既要考虑均值和波动性的区制变化,又要兼顾长记忆性和市场有效性的区制变化。This paper performs new econometric tests on dynamics of Shanghai and Shenzhen stock market returns path by adopting MS-ARFIMA model.Results show that,not only mean and volatility show obvious structural non-linear,but also long memory of Shanghai and Shenzhen stock market returns show obvious regime-switching dynamics.Under"bear market"regimes,the returns of Shanghai stock market show long memory process with mean reversion(d1=0.7732),and show short memory process(d2=-0.1854)under"bull market".While the returns of Shenzhen stock market show short memory process(d1=-0.1575)under"bear market",and show long memory process(d2=0.2625)with covariance stationary under"bull market".These new results imply the reaction pattern and market efficiency of China's stock market to news also exist relative regime-switching changes,which requests the departments concerned to not only take regime changes of mean and volatility into consideration,but also pay enough attention to regime changes of long memory and market efficiency when stipulating stock regulations.

关 键 词：股市收益率 MS-ARFIMA模型 区制转移 

分 类 号：F064.1[经济管理—政治经济学]