论文部分内容阅读
加强个人所得税征管、促进纳税遵从一直是税务部门的重要工作之一。本文利用多分类有序反应变量Logistic回归模型,以江苏某市2010年度年所得达12万元以上个人所得税纳税申报中的数据为样本,剔除缺失值以后,以“应补税额”等级的五分类有序变量为反应变量,以纳税人的年龄、应纳税所得额、应纳税额、性别、职业大类、行业大类6个影响因素为自变量,进行多分类有序反应变量逐步Logistic回归分析,并进一步根据多分类Logistic模型的分析,构建基于遵从概率估计的风险预警模型,期望研究结果为税务系统加强年所得12万元以上高收入者个人所得税的申报与征管、促进纳税遵从提供新的研究思路。
It is always one of the important tasks of the tax department to strengthen the collection of personal income tax and promote tax compliance. In this paper, a multi-class order logistic regression model was used to take the data of personal income tax returns of more than 120,000 yuan in a certain city of Jiangsu in 2010 as a sample. After removing the missing value, Five categories of orderly variables as response variables, the taxpayers age, taxable income, tax liability, gender, occupational categories, industry categories six influencing factors as independent variables, multi-class orderly response variable stepwise Logistic Regression analysis and further based on the analysis of multi-classification Logistic model to build a risk early warning model based on compliance probability estimation. It is hoped that the research results will provide tax system to strengthen the declaration and collection of personal income tax of high-income earners with income over 120,000 yuan annually, and promote tax compliance New research ideas.