| 摘要: | 本研究旨在探討台灣地區本國銀行「盈餘管理」之現象與行為。
首先，採用Burgstahler & Dichev(1997)及Degeorge、Patel & Zeckhauser(1999)所發展之綜合橫斷面次數分配研究法(Pooled cross-sectional Distribution Approach)，依據直方分配圖與Z值及t值統計檢定方法，對「盈餘門檻」現象進行檢測，以瞭解本國銀行業是否存在「避免盈餘損失」及「避免盈餘下降」之「盈餘管理」行為。其次，更進一步分析該「盈餘管理」行為是否受景氣榮枯、資產規模、資本適足性、資產品質及股票信用交易考量等影響而有所不同。最後，再參考過去文獻與相關法令規定，萃取出本國銀行業「盈餘管理」權衡項目，利用複迴歸分析檢定各項「盈餘管理」權衡因子與稅後淨利之間的關係。
本研究以台灣經濟新報社資料庫(TEJ)中本國上市上櫃銀行，自民國84年至93年之年度財務報告資料，經整理篩選後計得樣本330筆，進行實證結果發現：
(1)本國銀行業會為了「避免盈餘損失」及「避免盈餘下降」而進行「盈餘管理」，且前者現象較為顯著。
(2)小額盈餘獲利及小額盈餘增加之本國銀行較小額盈餘損失及小額盈餘下降者更能看出「盈餘管理」現象。
(3)本國銀行業為了「避免盈餘損失」而進行「盈餘管理」的行為，受到資產規模與股票信用交易考量之影響而有所不同。
(4)本國銀行業為了「避免盈餘下降」而進行「盈餘管理」的行為，受到景氣榮枯、資產規模、資本適足性、資產品質及股票信用交易考量等影響而有所不同。
(5)放款總額、應收利息、逾期放款、備抵呆帳、各項提存及損失準備等五項「盈餘管理」權衡因子，皆與稅後淨利呈現顯著相關。其中放款總額、應收利息及備抵呆帳與稅後淨利呈現正相關；而逾期放款、各項提存及損失準備對稅後淨利呈負相關。五項權衡因子用來解釋盈餘的能力達50.89%。
This purpose of this research investigates the displayed phenomenon, patterns, and behavior due to the Earnings Management practices implemented by the native-owned banks in Taiwan.
In this research, first, we apply the Pooled cross-sectional Distribution Approach (developed by Burgstahler & Dichev(1997) and Degeorg, Patel & Zeckhauser(1999) as well as others) to the data of earnings and changes in earnings by constructing histograms as graphical tools and performing statistical inference tests such as z- and t-tests to examine the “earnings-losses avoidance” and “earnings-decreases avoidance” activities in order to detect the existence of Earnings Management practices exercised by the native-owned banks in Taiwan. Secondly, we further conduct a study to see if the detected Earnings Management practices were influenced by the economical factors such as business cycles, asset sizes, capital adequacy, and on-credit trading. Lastly, we shall perform a multiple regression analysis to investigate the relationship between the after-tax profits and various discretionary factors such as the total amount of loans, interest receivables, loans in arrear, allowance for bad debt, and preserves. The five discretionary factors are carefully selected into the regression model through a fairly comprehensive review of research literature and government regulations.
This research is based on a sample of size 330 by including thirty-four Taiwan’s publicly-traded, citizen-owned banks whose financial statement data are available from TEJ (Taiwan Economics Journal) database for the 10-year period from 1995 to 2004. It is noted here that the sample was finalized after deletions of few incomplete data.
The findings of our research are summarized as follows:
1.The native-owned banks in Taiwan did practice Earnings Management activities for the avoidance of “earnings losses” and “earnings decreases” for the period from 1995 to 2004. It was also observed that the activities for the former type of avoidance were more rampant than those for the latter;
2.The banks with small positive earnings or small earnings increases are more likely to exercise Earnings Management than those with small negative earnings or small earnings decreases;
3.The degree and types of Earnings Management activities exercised by the banks under study are significantly influenced by the asset size and on-credit trading;
4.The exercise of Earnings Management for the “earnings-decreases avoidance” by the banks under study is shown significantly affected by the economical factors such as business cycles, asset sizes, capital adequacy, and on-credit trading;
5.We find that the five Earnings Management discretionary factors: total amount of loans, interest receivables, loans in arrear, allowance for bad debt, and preserves are significantly correlated with after-tax profits. More specifically, it has been shown that the after-tax profits are positively correlated with the three factors: total amount of loans, interest receivables, and allowance for bad debt and are negatively correlated with the other two factors: the loans in arrear and preserves. The obtained regression model indicates that it has a 50.89% explanatory power of earnings using the five specified discretionary factors. |
