| Tax noncompliance is as old as taxation itself. In order to prevent tax losing caused by tax noncompliance, many research works have been done through economics science, management science, law science and psychology science. Based on the empirical data and standing by the side of tax administration, this paper first analyzes the results of sampling survey of tax compliance in China. Second, we discuss the tax audit decision model manipulated by the practitioners. Then we try to discriminate whether the taxpayer's declaration is correct or not. Also we make further efforts to discriminate what kind of tax evasion means might be applied by evaders.The main ideas of this dissertation are as following:Chapter one focuses on the importance of the research and the relevant background.Chapter two describes the tax-collection and tax compliance abroad. The western countries' countermeasures are time-consuming and require systematic works. Based on tax administration practice in China, improving the technology of tax audit decision is relatively practical since it needs less cooperation and costs less.Chapter three analyzes tax compliance in China at the sight of the tax administration. Results suggest that about 50% of all the taxpayers (companies) are evading their tax returns. Also we discuss the key factors that influencing tax compliance. Research confirms that low penalty is the key issue which induces tax evasion. While uncertainty of laws, equity enforcement and limits of tax management make it impossible to raise it to reduce tax evasion.Chapter four reviews the relevant researches on technology of tax audit decision making, including peak-value analysis, discriminating analysis, neural network model and expert system.Chapter five describes the tax audit decision model manipulated by practitioners. Using AHP, we list out and rank the variables adopted by tax auditors. We associate the model with its efficiency (below 50%) and conclude that the key attributes used in this tax audit decision model have long been in concern of taxpayers. To a considerable extent, these data are disguised, which is supported by some earlier research. So we strongly suggest that other attributes such as stock-turnover ratio should be of more importance. Yet, variables which deal with intangible factors can be part of the discriminating attributes in computer-aided tax audit decision making models.Chapter six tries to discriminate the honest report type and dishonest report type (mainly on value-added tax). Using SPSS 10.0 software, we choose 19 variable trying to develop a discriminating model. First on combined data, then splits them into sub-industries to develop discriminating functions respectively. Results suggest that splitting efforts do help improving the discriminating efficiency and discriminating model can do even better as long as enough data are available.Chapter seven makes further efforts to discriminate what kind of tax evasionmeans may be applied by evaders. Based on empirical sampling, we conclude 13 means of value-added tax evasion and try discriminating. Results pints out that enlarging sample quantity and getting a more equiponderant sample may be of great use in future work.In the last chapter of this thesis, the research is summarized and some important conclusions follow the next. The author explicates some developing aspects for future work.The original points of this dissertation include the following:(1) We analyze the tax compliance in China at the sight of the tax administration. From the results of the questionnaire, we get some aspects the same as earlier research and also some different ones.(2) We describe the tax audit decision model manipulated by practitioners. Using AHP, we rank the intangible and tangible variables adopted in the model. So we can incorporate the intangible attributes in discriminating model in future research.(3) Using SPSS 10.0 software, we choose 19 variable trying to develop discriminating models based on sub-industries by stepwise me...
|
PDF Full Text Request