前期預測較精準分析師之盈餘預測出現偏離時 – 精準分析師過度自信與錯價指標作為股票報酬之預測變數

Abstract

本研究探討分析師在歷經短期預測成功後，是否導致過度自信，進而使其預測產生系統性的偏離（deviation)；進一步分析此類偏離預測在不同資產錯誤定價情境下，是否實為偏誤（bias），並對風險調整後報酬產生影響。過去文獻指出，分析師預測行為可能受行為面影響，尤其在近期表現優異時，較易提出偏離市場共識的預測，其可信度或有不足。然而，現有研究對此類偏離預測是否具備資訊價值，仍缺乏系統性檢驗。 本研究針對 2010 至 2024 年間之美國股票市場分析師盈餘預測，篩選出連續四季預測優於同儕之「準確分析師」，並建構衡量其當期預測偏離全體分析師共識之指標。進一步採用事件研究法與迴歸模型，分析預測偏離、資產錯誤定價指標與異常報酬之間的關係。 實證結果顯示，準確分析師若提出高度樂觀的預測，其後續風險調整後報酬顯著為負；高度悲觀預測則對應正向報酬。此外，當偏離預測方向與資產錯誤定價方向一致時，更可能導致顯著的負向報酬，顯示分析師行為偏誤與市場定價錯誤具有交互放大效果。 本研究闡明，前幾期績效良好的分析師，可能因過度自信，進而影響預測品質，並對依此建構之投資組合的市場反應產生影響。此一發現有助於理解資訊中介在資本市場中的行為動態與所發布資訊意涵。

This study investigates whether and to what extent security analysts exhibit systematic forecast bias perhaps due to overconfidence resulting from their prior-period forecasting outperformance and further examines how such forecast deviations help explain risk-adjusted returns under varying degrees of mispricing. Prior research suggests that analysts may become self-confident after recent success, leading them to issue forecasts that deviate from market consensus. However, there is exists limited empirical evidence regarding whether such deviations carry informational value or how the market should respond to them. Using analyst earnings forecast data from the U.S. market between 2010 and 2024, this study identifies “accurate analysts” as those who outperform peers for four consecutive quarters and constructs measures to capture their current forecast deviation. Event study methodology and regression models are employed to assess the relationship among forecast deviation, asset mispricing, and abnormal returns. The empirical findings reveal that optimistic forecasts issued by previously accurate analysts are associated with significantly negative risk-adjusted returns, whereas pessimistic forecasts tend to yield positive abnormal returns. Moreover, when the direction of optimistic forecast deviation aligns with existing mispricing, the subsequent returns are even more negative, suggesting that behavioral bias and market mispricing may reinforce each other. This study highlights that even analysts with strong prior performance may exhibit overconfidence, perhaps driven by short-term success, adversely affecting their forecasting quality and market influence. The findings offer non-trivial insights into the behavioral dynamics and the informativeness of EPS forecasts issued by financial intermediaries in capital markets