This paper provides evidence on whether sell-side security analysts herd around the consensus when they issue target price forecasts. Using a modified version of the methodology developed by Jegadeesh and Kim (2010), we find that, on average, analysts issue biased target forecasts toward the consensus and the market recognizes their herding tendency by reacting more to targets away from the consensus than those closer to the consensus. Analysts herd more when they issue unfavorable targets than favorable targets, possibly because being bold and negative at the same time is hazard to analysts' career. We find that analysts who previously made more precise target forecasts issue less herding target prices, an evidence consistent with the over-confidence hypothesis. Target herding behaviors can also be explained by the incentive structure faced by analysts and brokerages, whereas it is unrelated to the degree of information sufficiency.
本文首度提出實證證據顯示證券分析師發布的目標價呈現往共識（consensus）目標價偏移的「從眾」傾向。研究方法使用調整後的Jegadeesh and Kim（2010）模型，結果發現目標價傾向往發布當時的市場平均目標價靠攏，顯示目標價有從眾傾向；而市場也認知此傾向，對於偏離共識目標價的預測給予額外的報酬反應，以補充其刻意壓抑的私有資訊。悲觀的預測目標價呈現較強烈的從眾傾向，反映了分析師不願單獨出具明顯異於同儕觀點的負向評估報告。實證結果並發現目標價預測較準確的分析師後續所發布的目標價較不從眾，此現象符合過度自信假說。部分結果也顯示分析師從眾行為受其獎懲動機的影響，然而資訊充分性並不影響目標價的從眾程度。