先前的文獻多探討盈餘宣告時點對股價是否有反應，衡量其資訊內涵程度，且大多數國內研究結果均未發現盈餘宣告時點具有資訊內涵。本研究以分析師預期與公司經理人宣告盈餘之間的差距（即未預期盈餘），探討我國一般產業上市公司利用盈餘宣告時點，以裁決性應計項目之操縱來掩飾其本身經營之實際情況，試著從財務報告間接證實盈餘宣告時點具有資訊價值。實證結果發現：不論是好消息或壞消息之情況下，延後宣告盈餘者與其他宣告時點相比，其管理盈餘幅度均最大。延後宣告樣本與異常應計數相關性之測試，亦呈顯著正相關，表示延後天數越多，其盈餘管理幅度越大。而在探討延後宣告盈餘者之公司屬性方面。全部延後宣告樣本之檢定，除負債比率變數達顯著水準外，其餘均不顯著。好消息延後樣本中，負債比率變數與公司規模代理變數均呈顯著相關。壞消息延後樣本中，僅負債比率變數呈顯著相關。

Previous literature defined the information content of earnings announcement by testing their response to the stock price. Most internal empirical results have not found the information content of earnings announcement timing. According to the disparity of earnings announcement timing between the analyst and manager announcement, this study discusses the public company in Taiwan's earnings management problem controlled earnings announcement timing. This research tries try to verify earnings management via discretionary accrual conditioned earnings announcement timing. Empirical results found that late announcement of the good news sample group and the bad news sample group manipulates range more than the others. This research found significant positive correlation exists between the deferred announcement and abnormal accruals. Finally, in the characteristic of late announcement company, this research also found significant positive correlation exists between the days of deferred announcement and debt rate.