【摘要】 与现有研究方法不同,本文通过考察Akaike、Schwarz、Shibata、Hannan-Quinn四个信息准则,建立了描述深圳股票市场收益过程和波动过程双长记忆性特征的ARFIMA-FIGARCH模型。实证分析说明采用ARFIMA(0,m,1)-FIGARCH(1,d,0)模型拟合最好。研究结果表明:深圳成分指数日收益序列无长记忆,但波动序列具有较强的长记忆特征。更多还原

【Abstract】 This paper can simultaneously capture double long memory properties of return process and volatility process of China stock market by ARFIMA-FIGARCH model. The results of empirical study to the behavior of Shenzhen stock market show that there is no long memory for Shenzhen daily return, while there is strong long memory for volatility series of Shenzhen daily return. According to Akaike、Schwarz、Shibata and Hannan-Quinn criterions , ARFIMA(0,m,1)- FIGARCH(1,d,0) model is the most appropriate for simulating Shenzhen daily return.更多还原