基于分整特性视角的我国股市长记忆性识别

Parametric and Semiparametric Test on the Long Memory Property of Return and Its Volalitity in Stock Markets

金融时间序列的长记忆性检验常采用标度分析法,但结果往往不令人满意。从分整特性的新视角,利用KPSS检验和LW检验对我国股市收益及其波动的记忆性特征进行了深入研究。研究结果表明,我国股市的波动序列中存在显著的长记忆性,而收益序列本身无明显的长记忆性。这与成熟股票市场有关长记忆性的研究结论基本一致,与新兴股票市场的研究结论有所不同。此项结论对股市的长期投资者具有重要的决策意义。

Scaling unsatisfied results. methods are always applied to detect long memory property of financial time series, which produce Standing on the new viewpoint of fractional integration, KPSS test and LW test can find long memory of return and its volatility in Chinese stock market. The resuhs show that the volatility series has an obvious long memory property, while the return series has no such property. This conclusion is similar with mature stock markets, but different from developing stock markets, which has an important significance to the long-term investors.