林世銘臺灣大學：會計學研究所黃美祝Huang, Mei-ZuhMei-ZuhHuang2007-11-282018-06-292007-11-282018-06-292006http://ntur.lib.ntu.edu.tw//handle/246246/61753本研究旨在利用民國89年至91年度台灣營利事業所得稅之申報與查核資料，應用聯立方程模型(simultaneous equations model, SEM)將稅捐機關與納稅者(營利事業)同時視為租稅逃漏賽局內之策略性參賽者，亦即在控制選案查核率與逃漏行為之內生性關係之下，探討我國營利事業所得稅之租稅逃漏情形。在此一聯立方程模型中，本文分別利用Tobit及Probit模型來分析營利事業所得稅租稅逃漏及稅捐機關於選案查核時之各項影響因素。 在選案查核模型上，實證結果顯示台灣稅捐機關於進行選案查核時，對各申報案件之選案查核機率深受各申報案件申報行為(如申報所得高低及逃漏傾向)、案件特性(如是否享有租稅減免及盈虧情形等)及查核資源(如各案件之平均稽徵預算)等因素所影響。由此可見，台灣的稅捐機關在有限的查核預算及查核人力下，會有系統性的採用特定之查核法則來進行其選案查核程序。因此研究者於進行租稅逃漏之研究時，實不應將稅捐機關之查核行為視為固定不變，或將之視為租稅逃漏模型之外生變數。 本文發現在其他條件不變下，採用會計師稅務簽證之申報案件、採用股份有限公司組織型式之申報案件、流動性較高及享有較多租稅抵減優惠等申報案件，其營利事業所得稅之逃漏相對而言較少；至於應付之課稅所得愈多及負債比率較高、財務相對較困難之營利事業，其所得稅逃漏情形則比其他營利事業為高。而逾期申報之營利事業相對於如期申報之營利事業而言逃漏情形亦較嚴重。另外，本文亦發現先前暫繳或扣繳不足於申報時仍須補稅之營利事業，比暫繳或扣繳過多可退稅之營利事業相對而言逃漏亦較多。綜上所述，本文之實證結果顯示營利事業所得稅之逃漏行為顯著受到各該營利事業本身之特性所影響。 然而，本文聯立方程之實證結果及各項敏感性分析皆未發現稅捐機關之選案查核率對各申報案件當年度及以後年度之租稅逃漏有遏阻之作用。此一結果與Alm et al. (2004)之結果不同，或許意味著台灣營利事業似乎未能清楚意識到或察覺到(perceived)他們自身之被查核機率的高低。此外，本文敏感性分析之結果亦指出稅務人員之人工選案查核似乎比電腦選案查核更能發現簽證會計師幫營利事業進行之不當的"租稅規劃"，或更能發現營利事業藉由租稅抵減優惠所進行之不當"租稅安排"。顯示出稅務人員多年累積之查帳經驗及其專業判斷，對於現今查核資源相對不足之稅捐機關而言，實為重要的一項資產。因此如何有效累積與傳承稅務稽查人員之查帳經驗、減少稅務人員之流動率，並提升其專業能力，乃稅捐機關必須持續努力的首要之務。This study uses Taiwan’s business income tax returns data to simultaneously investigate the determinants of audit selection and business income tax evasion behavior. The simultaneous equations model (SEM) that this study employs has two equations; the first equation is a Tobit model for predicting tax evasion, while the second equation is a probit model for predicting the probability that a return will be audited. The data set consists of business income tax returns filed in Taiwan during the period between 2000 and 2002. The simultaneous equations model has the advantage of treating the tax authorities and the taxpayers (firms) as strategic players in the compliance game. It also provides theoretically more robust statistical results on the relationships between audits and tax compliance. The estimation results provide significant evidence that Taiwan’s tax authorities employ a systematic, even if informal, audit selection process. That is, they use information reported by the taxpayers to select returns for audit. Firms are more likely to be selected for audit as they have larger underreported tax liabilities, as they have tax credits, and as they have net loss. Therefore, it should not be assumed that the behavior of the tax agency is given and exogenous to the tax compliance models. Corroborating some important findings in main foreign literature, our estimation results also imply that other characteristic variables, such as whether a CPA-attested tax return is used, the legal form, liquidity level, and tax credits status of a firm, have significant negative impacts on the firm’s behavior of tax compliance. On the other hand, the results imply that the income level, debt ratio, whether or not miss filing deadlines, and the withholding status of a firm at filing time have significant positive impacts on the firm’s behavior of tax compliance. However, in contrast to prior prediction, our results consistently suggest that audit selection rate of tax authorities has a positive relationship with firms’ contemporaneous and subsequent tax evasion. Therefore, this study fails to find that audit has a deterrent effect on tax noncompliance in Taiwan. This result seems to imply that in Taiwan most firms cannot discern the audit process, and consequently they do not modify their tax evasion strategy to reduce their probability of audit. In addition, our results also find that tax auditors can discover more underreporting tax from specific tax returns with their work experience and professional justification. Therefore, the experience and professionalism of tax auditors are important assets for tax authorities in this age of lacking for audit resources. Thus, how to retain outstanding tax auditors and enhance effectiveness and efficiency of audit selection are important matters nowadays.中文摘要.................................................I 英文摘要.................................................II 目錄.....................................................III 表次.......................................................V 第一章 緒論...........................................1 第二章 文獻探討.......................................5 第一節 選案查核率對租稅逃漏之影響....................5 第二節 影響租稅逃漏之各項決定因素....................8 第三節 探討國內租稅逃漏情形之相關文獻...............14 第三章 理論模型架構..................................17 第一節 租稅逃漏模型之資訊不對稱假設.................17 第二節 模型均衡下之必要條件.........................19 第三節 模型之比較靜態分析...........................21 第四章 研究方法......................................23 第一節 樣本篩選過程.................................23 第二節 實證研究模型.................................24 第三節 變數定義.....................................27 第五章 實證結果與分析................................34 第一節 樣本單變量統計分析...........................34 第二節 敘述性統計及相關性分析.......................44 第三節 聯立方程模型之分析結果.......................46 第六章 敏感性分析................................ ..54 第一節 使用租稅逃漏比率作為逃漏方程式之應變數........54 第二節 選案查核率之另一種定義........................57 第三節 只用有被查核過之樣本分析租稅逃漏..............61 第四節 當期選案查核率對下期租稅逃漏之影響............63 第五節 選案查核與租稅逃漏之年度效果..................67 第七章 結論與建議.....................................71 第一節 結論..........................................71 第二節 政策意涵與建議................................74 第三節 研究限制......................................76 參考文獻................................................78486431 bytesapplication/pdfen-US租稅逃漏選案查核率聯立方程模型tax evasionaudit selection ratesimultaneous equations modelotherhttp://ntur.lib.ntu.edu.tw/bitstream/246246/61753/1/ntu-95-D89722004-1.pdf