本研究以會計師事務所抑制裁決性應計數的能力作為審計品質之衡量指標,在控制自我選擇偏誤的情況下,探討安隆案後,臺灣四大會計師事務所的審計品質是否明顯優於非四大會計師事務所。實證證據顯示,樣本資料的確有自我選擇偏誤的現象,以內生性二元處理模型(endogenous binary treatment model)並輔以多種估計方法及固定效果迴歸模型,分離自我選擇效果後,四大會計師事務所相較於非四大會計師事務所更能抑制管理階層盈餘管理的空間;亦即從抑制公司盈餘管理的角度,四大會計師事務所的審計品質明顯優於非四大會計師事務所。此外,臺灣四大會計師事務所間儘管規模差異不小,但其間之審計品質並沒有明顯的差距。最後,證據也顯示,如像過去文獻一樣,若未控制自我選擇偏誤的問題,則檢測結果相當分歧,無法作出穩健的結論,本研究突顯相關議題之探討,對自我選擇偏誤控制的重要性。
Using abnormal accruals as proxies for audit quality, this study examines whether the audit quality of Big 4 firms is superior to that of non-Big 4 firms in Taiwan after Enron's accounting scandals. Empirical evidence suggests that client characteristics can confound inferences of the Big4 effect. Therefore, we employ an endogenous binary treatment model and a fixed effects regression model to control different client characteristics between the two auditor groups, and estimate the audit quality effects. The empirical results of these models show that Big 4 firms deliver higher audit quality than non-Big4 firms. In addition, the audit quality between Big 4 firms is homogenous even though the size of Big 4 firms is quite different in Taiwan. However, these findings become unclear under traditional OLS regression. The results indicate that control of self-selection bias is crucial in such audit empirical research.