| 摘要: | 當衡量機構投資人從眾行為時, 如何定義買賣方相當重要。 因為這將會影響到從眾水準的衡量是否準確?Chu (2013)即指出使用持股比例與持股數衡量投資人買賣方會導致不同結果, 使用持股數衡量是較為正確的方式, 因持股比例會受到公司事件的影響, 導致從眾水準的衡量不準確。 本篇研究依據 Chu (2013)所提出之方法, 並利用Lakonishok et al.(1992)衡量機構投資人從眾之研究方法, 研究期間為1980年至2010年, 觀察並比較持股比例與持股數所衡量出之從眾結果。 為找出機構投資人從眾動機, 本篇研究首先利用兩種衡量指標, 透過迴歸分析檢測兩種衡量指標與後期股票報酬的關聯性, 以判斷機構投資人從眾行為是否基於資訊因素? 以持股數為衡量指標之實證結果顯示, 機構投資人從眾行為並非基於資訊因素。 為進一步找出從眾原因, 本篇研究參考Sias (2004)所提出三種非資訊從眾因素, 對此三種類別從眾因素一一做檢測, 結果顯示, 機構投資人從眾原因均可能受到股票特性、 市場潮流與機構投資人本身聲譽的影響。;When measuring institutional herding, different definitions of buyer and seller are important and will lead to different results. Chu (2013) points out that using ownership percentage and shares held cause different results of herding measure and using shares held is a more accurate method because ownership percentage may be affected by company events. Refer to Chu (2013) and take quarterly data during the 1980-2010 period, I use the herding measure proposed by Lakonishok et al. (1992) to distinguish the effect of using different methods, ownership percentage and shares held, to define buyer and seller and use a more accurate method, shares held, to find out what reasons drive institutional investors to herd. To test whether institutional investors’ trade are related to information or not, I first examine return reversals. The empirical results show that institutional herding is not related to information and the reasons of institutional herding may be driven by firm characteristic, fad, and reputation. |
