【摘要】 针对非参数方法研究国内股市长记忆性时结论参差不齐的现状,本文研究了更为稳健的半参数估计方法,即局部W h ittle(LW)估计和对数周期图(LP)回归。通过对不同频率高频数据的分析,证实了LW估计方法尽管需要数值最优化,但仍然要优于LP回归。进而将LW估计首次应用于中国股市,结果表明不同频率绝对收益序列的长记忆强度基本一致;同时发现,重大突发事件发生时的长记忆性表现得最为强烈,且事件后比事件前表现的要强烈,这说明股票市场的溢出效应在事件后增强,此项结论对我国证券市场有一定的借鉴意义。更多还原

【Abstract】 Aimed at the fault of using methods of non-parametric to stock market,the paper investigated the methods of semi-parametric estimation based on frequency domain,including local Whittle estimation and log-periodogram regression.Through different high-frequency data,LW estimation was exemplified to be more efficient than LP regression,even though it involves numerical methods.Then LW estimation was applied in China stock market.The results present that long memory be generally consistent across various temporally aggregated returns.And the long memory is found greater during the gravely outburst events,which is stronger in the post-event than in the pre-event.Thus,the spillover effects in stock market become strong in the post-event,which is important for China securities business.更多还原