訴訟之效果與效率向來為訴訟當事人所關切之議題。王全三、廖珮真與林敬偉(2010)曾分析所得稅訴訟案件中,專業訴訟代理人對案件之影響,但本文考慮到自選擇偏誤與樣本選擇偏誤存在於此議題之中,故採Heckman兩階段模型進行統計分析。 統計結果與先前研究有所不同,在訴訟效果方面,部分結果顯示會計師代理有助於提昇當事人之勝訴機率,至於律師代理對於當事人勝訴機率之影響則未有所發現;在訴訟效率方面,律師代理顯著延長訴訟天期,導致訴訟不經濟,至於會計師代理雖亦延長訴訟天期,但並未顯著。此結果足供實務界參考,有助於檢討2007年行政訴訟法修法之妥適性。
In the past, some essays analyzed the impact of agents’ professional background in income-tax administrative litigation, but most of them did not concern about that self-selection bias and sample selection bias may influence the outcomes in this issue. This article uses Heckman two-stage model to identify and correct those biases. The empirical evidence shows that accountant-agents could significantly give their clients higher winning rates, but there is no evidence to support that lawyer-agents could do so. Furthermore, the litigation with lawyer-agents takes more time than one without lawyer-agents. On the contrary, it is not significant that accountant-agents would increase litigation time period. The results could be references for the future amendment of the Administrative Litigation Act.