首次利用经典的LSV模型以及新发展起来的FHW模型来估计中国基金的羊群行为水平的上下界,得到基金羊群行为水平位于区间[4.9%,14.4%]之内,相对于国外发达的资本市场,中国基金表现出了较高的羊群行为水平,但相对于国内已有的研究结果,中国基金羊群行为水平有所降低。进一步利用FGLS对中国基金羊群行为的影响因素的研究发现,参与股票交易的基金数量越多,羊群行为越容易发生,而规模越大的股票越不容易产生羊群行为,另外,股票收益、流通A股比例对基金羊群行为在统计上没有显著的影响,基金在对不同行业的股票交易中表现出了差异化的羊群行为。最后提出可以从提高基金交易信息的透明度、降低基金的集中度、扩大股票流通规模等方面来降低基金羊群行为的水平。
This paper uses the classic LSV model and the newly developed FHW model to estimate the downward andupward bounds of Chinese Funds' herding level for the first time, and finds that it ranges in the interval [ 4.9%, 14.4% ]. Comparing to foreign developed capital markets, Chinese funds exhibit a high herding level, but relative to theexisted domestic research results, it has decreased. Using FGLS to further study the impact factors of Chinese funds'herding found that the more the number of funds involved in stock trading, the more prone to herd, while the larger thestock scale, the less prone to herd. In addition, stock return and proportion of tradable A shares have no significant impacton the fund herding statistically, and in the stock trading of different industries fund herding is different. Finally,improving the transparency of fund trading information, reducing the concentration of fund and expanding the scale ofstock liquidity are proposed to reduce the fund herding level.