作　　者： ; ; ;

机构地区： 华南理工大学工商管理学院

出　　处： 《统计与信息论坛》 2011年第3期56-62,共7页

摘　　要： 以往大量文献研究表明基金经常发生风格漂移现象,但资本市场呈分形特征违背了有效市场假说,对传统的实证研究方法提出了巨大挑战。以2005年上半年成立的5只开放式基金为研究样本,利用3个信息准则确定ARFIMA(1,d1,1)-HYGARCH(1,d2,0)为最优模型,引入SKT分布结合该模型来探索研究基金投资风格漂移所带来的收益及波动过程的分形特征。实证结果表明:该模型能够较好刻画投资风格漂移日收益序列的双长记忆性分形特征,其波动过程均存在显著的长记忆性,但收益过程存在长、短记忆性不统一现象,这进一步说明基金投资风格经常发生无序漂移现象,也折射出其背后存在着巨大的漂移风险隐患。最后,Person吻合度检验证实了SKT分布能较好地拟合基金投资风格漂移日收益序列的分布。 Previous studies on fund investment style have shown that： Fund offen occurs style drift,but the fractal characteristics of the capital market was contrary to the efficient market hypothesis,put forward a huge challenge to the traditional empirical methods.This paper used five open-end funds for the research sample,which were founded in 2005.Used the three information criteria to determine ARFIMA（1,d1,1）-HYGARCH（1,d2,0） is the optimal model,Introducted the skt distribution combined with this model to firstly study on fund investment style drift return and volatility fractal characters.Empirical results show that： this model can depict the double long memory fractal characters of investment style drift return series,the volatility process was significant long memory,but the return process was the long or short memory that is not uncertain phenomenon,further research shows that fund often occurs disorder investment style drift,but also reflects there is a huge potential drift risks in the underlying.Finally,the Pearson goodness of fit test also confirmed the correctness of skew distribution.

关 键 词： 投资风格漂移 双长记忆性 模型 分形 开放式基金

领　　域： [经济管理]