會計穏健性原則下前後多期盈餘之不對稱性

Abstract

本研究旨在探討我國前後多期會計盈餘在會計穏健性原則影響下認列之不對稱性。本研究首先釐清我國會計盈餘落後股票價格反映當期經濟事項之時效性，意即檢視前後多期盈餘與當期股票報酬率之關聯性；其後，再參酌Basu（1997）模型，進一步分析前後多期會計盈餘對當期好、壞消息認列之不對稱性，並據以判斷我國會計盈餘認列之穏健程度。本研究以民國82年至民國96年之台灣上市（櫃）公司作為研究對象，利用複迴歸模型分析落後至第7年之前後年度盈餘與當期股票報酬率間之關係。實證結果顯示： 一、當期股票報酬率與當期及落後第1年之盈餘為正相關；當期股票 報酬率與落後第2年之盈餘正、負向關係並不一定。當期股票報酬率與落後第3年與第4年之盈餘為負相關。 二、當期股票報酬率與落後至第2年內之會計盈餘最具關聯性；當盈餘落後至第3年後，當期股票報酬率與未來盈餘之關聯性逐漸轉為不明顯；而當盈餘落後第5年至第7年時，當期股票報酬率與未來盈餘不具關聯性。 三、壞消息傾向當期認列，而好消息則傾向分期認列。就整體而言，相較於壞消息，好消息對前後多期盈餘之影響較為持續；惟當盈餘落後第2年至落後至第4年時，壞消息仍持續對未來盈餘發生影響，顯示我國會計盈餘認列尚非明顯具穏健性。 四、相隔同樣期數下，同單位盈餘對好消息之反應係數較高，而對壞消息之反應係數較低。就整體而言，相較於好消息，盈餘反映壞消息之及時性較佳；惟實證結果顯示我國壞消息並未如預期於當期全數認列損失，當盈餘落後第2年至第4年時，壞消息仍持續影響未來盈餘。

This study is to investigate the the asymmetric effect of accounting conservatism on the multi-period earnings. Fisrt, we examine the extent that the recognition of economic events in accounting earning lags that of the market in Taiwan. In other words, we assess the association between current annual stock returns and both current and future periods’ accounting earnings. Further, similar to Basu’s model (1997), we use postive and negative annual returns to proxy for “good news” and “bad news”, respectively, to test the asymmetric effect of accounting conservatism on the multi-period earnings. The samples used for the study consist of all Taiwan listed and OTC firms-year observations from the period 1993 to 2007. We ues the multiple regression analysis to test the association between current annual stock returns and both current and the next 7 periods’ annual accounting earnings. The empirical results are as follows: 1. The correlations between current annual stock returns and both current and 1 year lagged earnings are positive. The correlations between current annual stock returns and 2 years lagged earnings are not clear.The correlations between current annual stock returns and both 3 as well as 4 years lagged earnings are negative. 2. The association between current annual stock returns and current as well as the next 2 years’ earnings are relatively stronger. After the next 3 years, the association between current annual stock returns and future earnings are less significant. Further, there is no association between current annual stock returns and the next 5 to 7 years’ earnings. 3. The study shows that bad news tends to reflect mostly in the current period’s earnings and that good news tends to reflect in the current and future multi-periods’ earnings. In general, good news earnings is more persistet than bad news earnings. However, this study reveals that bad news also reflects in 2 to 4 years lagged earnings. These results suggest that the degree of conservatism in accounting in Taiwan is not much significant. 4. In the same intervals, the study shows that the coefficient on earnings is larger for positive than negative return. These results suggest that bad news earnings is more timely than good news earnings. However, the study also shows that not all losses for bad news are recognized immediately in the current period as predictions. In contrast, bad news reflects in 2 to 4 years lagged earnings.