第一章 緒論 本章分三節，說明研究動機、目的、方法、內容以及研究限制。 第二章 我國營利事業所得稅結算申報調查現況檢討 本章分二節，檢討現行調查制度的績效及調查所面臨的問題。 第三章 選案調查制度建立的必要性。 本章分六節，說明建立選案調查制度的理由。 第四章 選案方法之理論分析－區別分析的理論。 本章分三節，敘述本文所使用之選案方法的理論及建立模式的步驟。 第五章 選案模式建立的實證研究。 本章分五節，根據實際的抽樣資料，進行選案模式的建立及說明如何應用 。 第六章 結論與建議。 本章根據上述研究結果提出總結及建議。

本研究探討影響臺灣跨國企業進行移轉訂價規避納稅義務之主要因素，並進一步檢視臺灣現行選案條件之合宜性。此外，本研究驗證陳明進與蔡麗雯（2010）實證研究使用公司「移轉訂價偏離同業毛利率之差異數」作為衡量移轉訂價代理變數之可信度。本研究使用2004年至2013年度經稅捐稽徵機關以不合常規移轉訂價查核調整之公開發行公司案件計101筆資料進行分析。實證結果發現，在其他條件不變情況下，公司毛利率、營業淨利率及純益率低於同業，進行移轉訂價程度愈高；資產報酬率愈高者，進行移轉訂價規避稅賦程度愈高；研發支出愈多及資產規模愈大等發展能力愈強之公司，愈容易進行移轉訂價；海外子公司占所有子公司比率愈高之跨國企業公司，進行移轉訂價程度愈高。另現行移轉訂價選案條件有關損益鉅幅波動及獲利低於集團其他企業，與企業進行移轉訂價並無顯著關係，可能有進一步檢討該項選案條件之必要。 至於檢視以「移轉訂價偏離同業毛利率之差異數」衡量移轉訂價代理變數之合宜性，資產報酬率、營業成長率及企業規模與本研究實證結果一致，惟大股東及董監經理人持股比、四大會計師事務所、負債比例及電子業等其他變數之結果並不一致，顯見以移轉訂價偏離同業毛利率之差異數作為移轉訂價之代理變數，可能與實際查核狀況不儘相同。 / This study examines the major factors affecting multinational enterprises (MNEs) in Taiwan to avoid tax obligation on transfer pricing (TP), and analyzes if the current criteria of TP case selection is appropriate. In addition, this paper tries to verify the credibility of an indicator developed by Chen and Tsai (2010) to estimate abnormal TP transactions of companies. We measure transfer pricing noncompliance in terms of tax audit adjustments made by tax authorities. Base on a sample of 101 TP audits on public companies in Taiwan from 2004 to 2013, our empirical results show that companies which have lower gross profit margin, operation margin and net profit margin than the profit standard of the same business (D_PROFIT), higher return on assets (ROA), higher research and development (RD) expenditure, bigger firm size (SIZE), and higher proportion of foreign affiliates in all subsidiaries (MULTI) have more TP aggressiveness. Moreover, there are no significant relationships between the current criteria on TP case selection in Taiwan such as huge fluctuations in profits (D_CH_NI) and profits below the other companies in the MNE group (D_GROUP) and TP aggressiveness of MNEs. As regards the credibility of the indicator to estimate abnormal TP transactions of companies, we find that the results of ROA, growth rate (GROWTH), and firm size (SIZE) are consistent with this study, but the results of other factors such as shareholding ratios of the major shareholders, the directors, and the supervisors (SH), four major accounting firms (CPA4), ratio of debt (LEV) as well as electronics industry (ELECO) are not consistent with this paper. It means that the indicator to estimate abnormal TP transactions may not reflect the real TP audit adjustments.