對企業做報導分析師數及其餘預測落差之資訊意涵

Abstract

本研究探討分析師數的資訊意涵，檢測了解分析師數多寡與市場投資人對盈餘預測落差反應間的關聯，以及分析師數變動與盈餘預測準確度的關聯性。文中採用美國2001至2010年間在紐約交易所交易的公司樣本，以分析師數分為五組作為衡量基準，再用迴歸與T檢定等方法檢驗累積超額報酬率與各變數間之關係。 研究貢獻是（1）提出分析師數與股市面對盈餘落差反應強度的估計模型；（2）檢測不同公司規模分群下對上述模型之影響；（3）探討股價報酬率與前期預測準確度及盈餘操控度的關聯性；（4）驗證分析師數與盈餘預測準確度之關係，進而描述分析師數變動與預測準確度排名變動的關聯性。 實證結果發現：分析師數多公司之盈餘落差似因具備較多的資訊意涵，因此較受市場投資人的重視，而低資產(較小規模企業)組別或因平日較少受媒體關注， 分析師數較少者之會計盈餘公佈便成為較重大之資訊事件，未預期盈餘重要性隨分析師數下降。同時，「前期預測準確度」與股價反應為正向關係；「盈餘操控度」與股價反應並無顯著關係。「分析師數」與「盈餘準確度」有正向關係，且「分析師數變化」與「盈餘準確度排名變化」亦呈正向關係。

This study investigates the information implications of the number of security analysts following individual company and examines whether such number of analysts will be associated with investors’ different extent of reaction to earnings surprise, namely, the difference between the consensus forecast and realized earnings. Furthermore, I clarify the correlation of changes in earnings forecast accuracy and increases in the number of analysts. This study uses U.S. companies listed on the New York Stock Exchange from 2001 to 2010 as the full sample and divides these companies into five groups by the number of analysts. The regression analysis and t-tests are adopted to investigate how variables regarding analyst characteristics help explain the variability of cumulative abnormal returns. The contributions of this study are that (1) it proposes the estimation model for stock market response to the unexpected earnings based on the number of analysts, (2) it examines the impact of company size on the aforementioned model, (3) it confers the association between security returns and prior-period forecast accuracy as well as the extent of earnings manipulation, (4) it verifies the relationship between number of analysts and forecast accuracy, with which it elaborates the relevance of changes in accuracy ranking in changes in analyst number. Empirical results are consistent with the notion that investors put more emphasis on earnings announcements of firms with more analysts coverage as the number of analysts has stronger information implication. Furthermore, the importance of analyst number declines in the lower assets group as these stocks may be less frequently covered by the news media, thus the smaller the number of analysts, the more important the earnings announcement event becomes. Consistent with our prior conjecture, prior-period forecast accuracy has a positive relationship with stock returns while discretionary accruals, the proxy for the extent of earnings manipulation, has an insignificantly relationship with stock returns. Moreover, the results show that analyst forecast accuracy increases with analyst number. Likewise, changes in analyst number appear to help explain the changes in accuracy ranking.