標題: 從分析師預估EPS偏誤行為來看目標公司 累積異常報酬走勢─以台灣市場為例
Biases in the Earnings Forecasts of Analysts and the Cumulative Returns of Evaluated Firms: Evidence from Stock Market in Taiwan
作者: 黎佩香
Li, Pei-Hsiang
謝文良
葉銀華
Hsieh, Wen-Liang
Yeh, Yin-Hua
財務金融研究所
關鍵字: 分析師盈餘預測行為;過度反應;反應不足;過度樂觀;預測誤差;異常報酬率;累積異常報酬率;Analyst forecast behavior;Overreaction;Underreaction;Forecast error;Over-optimistic;Abnormal return;Cumulative abnormal return
公開日期: 2012
摘要: 國內外許多研究都指出分析師在盈餘預測時會有過度樂觀的現象產生,但在探討分析師對資訊是反應不足還是反應過度並沒有統一的結果,且分析師會受到不同資訊的影響,產生不同的偏誤結果。許多學者利用台灣市場資料來觀察分析師預估盈餘的偏誤類型,另一方面,也有學者將分析師報告的發佈當成事件分析探討是否會產生異常報酬,根據Easterwood & Nutt(1999)文章,本研究主要探討當分析師面臨目標公司前期EPS改變為正或負的訊息時,其預測誤差行為會呈現何種情況。最後,本研究將得出的結果進行累積異常報酬分析。 本研究利用台灣上市櫃2006年至2011年的分析師報告進行迴歸分析,結果支持分析師利用先前EPS資訊預測當年EPS時,對資訊會過度樂觀且對前期實際EPS改變資訊有過度反應的現象。利用過度反應的分析師報告資料進行累積異常報酬的觀察,發現當前期實際EPS改變為正(好消息),則累積異常報酬為負,反之,當前兩期實際EPS改變為負(壞消息),則累積異常報酬為正。
Many studies reach the same conclusion that analysts are over-optimistic when they forecast earnings; however, there are inconsistent results with regards to whether analysts overreact or underreact to prior information. In addition, the forecasts of analysts are biased by different sources of information. Some researchers observe the biased types of earnings forecasts in Taiwan market. On the other hand, regarding the announcement of analysts’ reports as an event, some researchers investigate whether there are abnormal returns. Therefore, based on Easterwood &Nutt(1999), this study examines the biased types of earnings forecasts for analysts in Taiwan. Specifically, this study first determines the biased types of earnings forecasts for analysts who face the positive or negative change in prior EPSs of the target firm. We then calculate the cumulative abnormal return after the release of analysts’ reports.. The empirical results indicate that analysts are overoptimistic and overreact to the information of change in prior EPSs when forecasting annual earnings. Moreover, we observe the cumulative abnormal returns (CARs) from the reports of analysts who overreact to the prior information. The results show that when the change in prior EPSs is positive (good news), the CAR will be negative. Conversely, the CAR will be positive when the change in prior EPSs is negative (bad news).
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT070053948
http://hdl.handle.net/11536/72266
顯示於類別:畢業論文