本研究旨在從審計品質所扮演的資訊與公司治理角色,及從改善盈 餘品質之監督效果執行分析,希冀瞭解產業專家會計師是否可直接 或間接抑制受查客戶財稅差異的幅度? 研究結果顯示, 經由Heckman 兩階段估計法控制財稅差異樣本自我選擇偏誤,及工具變數之 兩階段最小平方法控制產業專家變數內生性後,產業專精程度與財 稅差異幅度呈顯著負相關。進一步透過兩階段估計法觀察產業專家 會計師是否經由提升盈餘品質而降低財稅差異之間接效果時,亦發 現產業專家會計師所查核的客戶其裁決性應計數較小,且因其具備 較佳的盈餘品質得以降低財稅差異的幅度。上述結論即使改以傾向 分數配對法控制產業專家變數內生性時,審計品質的效果仍然維持 相當的穩定。

This study aims to examine whether the auditor's industry expertise reduces book-tax differences (hereafter BTD) directly through its information and governance roles and indirectly through its effect in enhancing earnings quality. To alleviate the concern regarding the self-selection bias caused by the non-random sample ofBTD and the potential endogeneity issue related to auditor choice, we adopt the Heckman two-stage procedure and the twostage least squares with instrumental variables approach to deal with these issues. Empirical results show that smaller absolute BTD are associated with higher auditor industry expertise. Moreover, using two stage estimation methods to investigate the indirect effect of auditor industry expertise on BTD through earnings quality, we find that discretionary accruals is smaller for clients of auditor industry specialists than for clients of auditor industry non-specialists and that lower discretionary accruals in turn reduces the magnitude ofBTD. Even when we substitute the propensity score matching method for the two stage least squares approach to correct for endogeneity bias, the results remain qualitatively similar.