台灣地區上市公司累計盈餘與股票報酬之關聯性研究

Abstract

在財務領域的學術研究中，研究股票報酬與公司盈餘關聯性的文章不勝枚舉，包含直接針對一段期間之盈餘與報酬的相關性做研究，亦有發展為以公司財務因素預測未來股價變動的模型等等。但綜觀國內外相關主題的實證研究，盈餘與報酬的計算期間從一年、一季至一週甚至幾天不等，以上述短期的盈餘與此段期間的報酬或股價作相關性分析，而統計迴歸分析下所求得之盈餘對報酬之解釋能力往往極低。 本研究在假設盈餘為報酬的解釋變數之前提下，將盈餘與報酬之研究期間（即計算期間）拉長，以消除因當期盈餘與當期股價的資訊內涵不同（即所謂盈餘誤差與不完美盈餘）所造成盈餘之解釋力變低的影響。故本研究將長時間的公司盈餘加總而得「累計盈餘」，以研究長期累計盈餘相較於短期盈餘對於報酬是否有較高的解釋能力。而實證結果支持了此項論點。

The relationships between stock returns and the accounting earnings have been studied intensively. The relationships between earnings and returns during a time period and the application of financial statement analysis in the prediction of stock returns are most popular ones among various approaches. Although such topics are so popular, the time period of calculated stock returns and accounting earnings are all short-term time interval (from one year, several months to one day). Even though the issues underlying the annual and shorter return interval studies differs, the results of both forms of empirical analysis have been similarly disappointed in the sense of unimpressive correlations between returns and earnings. There is always low R2 under the return-earning regression analysis. In other words, the empirical findings suggest that earnings have low explanatory power to returns. Under the assumption of earnings as a natural variable for explaining stock earnings, the paper expand the interval over which earnings and returns are determined to minimize the effects of the asymmetry information contents between returns and earnings (errors in earnings and imperfect earnings) which leads to low explanatory power of earnings. This paper aggregates the annual earning during different period to get the long term earnings. The relationships between long term aggregate earnings and stock returns are compared with that for the short term interval. The empirical findings support the hypothesis that long term interval eliminate the effects of the asymmetry information content.