基于自正则的K.S方法对2005年第1季度至2011年第4季度QFII羊群行为的均值变点进行检验,发现QFII在中国A股市场上羊群行为度序列确实存在均值变点,QFII羊群行为发生变化的时点分别为2007年第3季度与2009年第4季度.与传统的Kolmogorovo Smirnov检验以及“滑窗”检验方法相比,基于自正则的K—S检验可以避免长程方差的相合估计与窗宽参数的选取。在检验小样本容量数据时,比传统的K—S检验更加适用。
A self-normalization based Kolmogorov-Smirnov test is employed to test the change points of QFII's herding from the first quarter of 2005 to the fourth quarter of 2011, it suggests that there really are change points in QFII's herding behavior. And the QFII herding behavior changed at the time of the third quarter of 2007 and the fourth quarter of 2009. Compared with the traditional Kolmogorov- Smirnov test and the lag-window test, the new method doesn't require a long-run variance estimator, avoiding the annoying choice of bandwidth parameter. It is more applicable compared with the traditional Kolmogorov-Smirnov test when testing data in small sample.