本研究的主要目的是為了探討共同基金的從眾行為及共同基金對外資與散戶投資行為的影響。本研究的研究期間由1995年至2000年8月,利用91家曾被選為摩根指數成份股的公司為研究樣本,觀察國內196檔共同基金的月持股資料,研究方法則利用Lakonishok et al. (1992) 提出了方法來衡量共同基金的從眾行為,研究發現國內共同基金的從眾行為程度約為5.22%,高於美國市場的3.4% (Wermers, 1999),而且國內的基金經理人較傾向於跟隨其他經理人買進相同的股票。進一步探討公司規模對從眾行為的影響,發現大型股的從眾程度高於小型股;另外在產業別方面,以電子產業的從眾程度最高,金融產業次之,而一般產業的從眾程度最低,符合產業風險愈高,從眾行為程度愈高的論點。 因為本研究的研究期間含概整個東南亞金融風暴的發生期間,故本研究進一步探討外在環境對從眾行為的影響,發現當外在環境極為惡劣時,共同基金的買進從眾行為的程度會下降而賣出從眾行為的程度則會增加,此現象以小公司、金融產業及電子產業最為明顯。雖然國內共同基金之間存在著從眾行為的現象,但本研究發現國內共同基金的從眾行為是因為共同反應新的訊息而形成的,並不是因為代理問題或是為了維持私人聲望而造成的,因此共同基金的存在有助於使證券市場更有效率;但在金融風暴發生期間,共同基金的從眾行為則較不理性,尤其是在投資金融產業方面。 根據Grinblatt et al. (1995) 指出美國市場的共同基金有77%會採行正向回饋交易策略,國內的共同基金對一般產業採行正向回饋交易策略,而且追漲的現象較殺跌明顯,而對金融產業則改以負向回饋交易策略為主,其程度以金融風暴間期較明顯,致於電子產業的投資策略則無一定的規則,只有在金融風暴期間有輕微追漲殺跌的現象。 雖然法人的投資比重正逐年增加,但國內證券市場至今為止仍以個別投資人居多,法人與個別投資人之間的互動關係亦值得我們深入研究,於是本研究利用Probit 模型來探討投信、外資與個別投資人之間的互動關係,結果發現投信與外資之間是正向互動關係,其程度以金融產業最為密切,其次為電子產業與一般產業;個別投資人的投資行為則與共同基金明顯相反,而與外資輕微的反向操作,其程度以電子產業最明顯,一般產業次之,但在金融產業方面,共同基金與散戶之間則沒有明顯的關係。根據從眾行為對股票報酬影響的結果可知,共同基金擁有優異的選股及擇時能力,特別是在投資電子產業方面,由此可知個別投資人總是資訊落後者,其會買在最高點而賣在最低點。
This study examines the trading activity of the mutual fund industry from 1995 through 2000 to determine whether funds “herd” when they trade stocks and to investigate the impact of herding on stock prices on the Taiwan Stock Exchange (TSE). We also use “Probit model” to examine interdependencies among mutual funds, foreign investors, and individual investors. The overall average level of herding is about 5.22 percent, which is higher than reported by Wermers (1999) for their sample of mutual funds in U.S. (3.4 percent). We find much higher levels in trades of large market capitalization stocks and high-tech stocks. We also find that herds form much more often on the buy-side (BHM) than on the sell-side (SHM) in our sample. Because our sample period includes financial crisis period from the second half of 1997 through 1998, we can investigate the herding behavior during the financial crisis period separately from all sample period. We find that the herding measures decrease in buy-side and increase in sell-side during the financial crisis. The effects magnify in trades of small stocks, banking stocks, and technology stocks. We find no evidence that trades by mutual find had a destabilizing effect on the Taiwan Stock Exchange over non-crisis period. Our results are consistent with mutual fund herding speeding the price-adjustment process in the stable investment environment. However, during financial crisis, we find some evidences of overreaction, especially in financial shares. Finally, we find that mutual funds have the contemporarily positive correlations with foreign investors, especially in trading financial shares. However, the investment behaviors between mutual funds and individual investors are significant negative correlations.