台股分析師的從眾行為

Abstract

本文探討台灣股市分析師的從眾現象，觀察分析師是否有從眾或反從眾行為。研究期間為2006年至2011年，總共包含27,156筆分析師報告，以 Jegadeesh and Kim (2009) 的模型為基礎，並使用改良後的模型進行實證分析。研究結果發現台灣股市分析師有反從眾行為，當分析師發佈降等報告時，其反從眾行為將更為強烈。其中規模較大的券商或明星券商的從眾行為較為明顯。當分析師更新報告頻率較高，或評比的公司為電子產業時，也較容易產生從眾行為。最後發現，短天期下，市場對於分析師的反從眾行為幾乎完全反映，但長天期下，市場對分析師的反從眾行為會反應不全。

The purpose of this paper is to examine the herding behavior of the analysts for Taiwan Stock. We determine whether herding behavior or anti-herding behavior exists for analysts. Our data include 27,156 observations from analysts' report during the period from 2006 to 2011.The methodology that we employ in this study mainly revises from the model proposed by Jegadeesh and Kim (2009). The empirical results find that analysts for Taiwan Stock have anti-herding behavior, and this phenomenon is even pronounced when analysts report downgrade recommendation. In addition, analysts from star broker or big size broker have higher possibilities to herd. Furthermore, analysts with higher frequency of revision or who mainly rate for the electronic industry tend to herd. Finally, we find that the tendency of anti-herding behavior for the analysts has been reflected on the market at the revised day; however, in the long run, there is an underestimation for the anti-herding behavior of the analysts.