影响税收的几个因素分析计量经济学论文.docx

1、影响税收的几个因素分析计量经济学论文金融计量学论文影响税收的几个因素分析班级：金融工程1003学号：020*姓名：李瑶成 绩1数据选取（20分）2模型建立与数据分析（40 分）3Eviews 应用（10 分）4结论陈述（10分）5整体行文（20分）6总 分摘要税收是我国财政收入的基本因素，影响着我国的经济发展。 本文通过查阅相关网站信息数据对影响我国税收的因素进行论 述。通过 Eviews 计量学软件对国内生产总值、财政支出、商品 零售价格指数等税收收入影响因素进行一定的证明与研究分析， 得出相关结论，并对我国的税收提一些建议。关键词：税收 Eviews 国内生产总值AbstractThe t

2、ax is the basic factors of Chinas fiscal revenue impact onChinas economic development. This article discusses the factors affect ing Chin as tax reve nue by access to the releva nt site information and data. Some formal research and analysis, the in flue ncing factors of the the Eviews metrology sof

3、tware, the tax revenues of the gross domestic product (GDP), fiscal spending, the retail price in dex and draw releva nt con clusi ons, and our tax some advice.Key words: tax Eviews GDP序言税收是国家为实现其职能， 凭借政治权力， 按照法律规定， 通过 税收工具强制地、 无偿地征收参与国民收入和社会产品的分配和再分 配取得财政收入的一种形式。税收主要用于国防和军队建设、国家 公务员工资发放、道路交通和城市基础设施

4、建设、科学研究、医 疗卫生防疫、文化教育、救灾赈济、环境保护等领域。而税收一 方面受经济发展的制约， 但同时又对经济宏观发展起到重要作用。 因此，我们需要对影响税收的重要因素加以分析。一 变量的选取 从整体来看，经济的增长是税收增长的主要源泉。因此，选 择国内生产总值作为解释变量 x1。税收是财政收入的一个主体， 社会经济的发展会对公共财产 产生需求。则财政支出可以代表，作为解释变量 x2。我国的税制结构以流转税为主，以现行价格计算的 GDP 等 指标和经营者的收入水平都与物价水平有关。所以选取商品零售 指数作为物价水平的代表作为变量解释 x3 。数据的选取以下是选取的样本数据，单位均为亿元年

5、份国内生产总值国家财政支出商品零售物价指数税收收入19804545.6241228.830106.0000571.700019814891.5611138.410102.4000629.890019825323.3511229.980101.9000700.020019835962.6521409.520101.5000775.590019847208.0521701.020102.8000947.350019859016.0372004.250108.80002040.790198610275.182204.910106.00002090.730198712058.622262.180107

6、.30002140.360198815042.822491.210118.50002390.470198916992.322823.780117.80002727.400199018667.823083.590102.10002821.860199121781.503386.620102.90002990.170199226923.483742.200105.40003296.910199335333.924642.300113.20004255.300199448197.865792.620121.70005126.880199560793.736823.720114.80006038.04

7、0199671176.597937.550106.10006909.820199778973.039233.560100.80008234.040199884402.2810798.1897.400009262.800199989677.0513187.6797.0000010682.58200099214.5515886.5098.5000012581.512001109655.218902.5899.2000015301.382002120332.722053.1598.7000017636.452003135822.824649.9599.9059020017.312004159878.

8、328486.89102.806225718.002005183867.933930.28100.777430866.002006210871.040422.73101.028237636.00（以上数据来源于中国统计年鉴及中宏数据库）Depe ndent Variable: YMethod: Least SquaresDate: 06/02/12 Time: 23:51Sample: 1980 2006In cluded observati ons: 27VariableCoefficie ntStd. Errort-StatisticProb.C-6357.3062589.143-2.45

9、53710.0221X1-0.0111910.014037-0.7972610.4335X20.9670820.07682112.588750.0000X357.1184124.003452.3795920.0260R-squared0.994954Mean depe ndent var8681.087Adjusted R-squared0.994296S.D.dependent var9909.343S.E. of regressi on748.4057Akaike info criteri on16.20972Sum squared resid12882553Schwarz criteri

10、 on16.40170Log likelihood-214.8312F-statistic1511.718Durb in -Watson stat0.691548Prob(F-statistic)0.000000由上表可以得出回归方程：Y=-6357.306-0.011191*x1+0.967082*x2+57.11841*x32589.1430.0140370.07682124.00345T=-2.455371-0.79726112.588750.0260R2 二 0.9949542R-0.994295f=1511.718四.模型检验1、经济意义检验在假定其他变量不变的情况下，每当国内生产总

11、值增加一亿时， 税收便减少0.011191% ;每当国家财政支出增加一亿时，税收增 加0.967082% ;每当商品零售物价指数增加一亿时，税收增加 57.11841%。其中我认为国民生产总值与物价零售指数有一定出入， 下文会有所校正。2、 统计检验拟合优度由表中得出的两个数据 （以下两个），可知模型对样本拟合的较好2R2 =0.994954 R =0.994295T检验中三个解释变量的t值分别是t0=-2.455371 ,t仁-0.797261,t2=12.58875,t3=2.379592. 在 5%显著 性水平下自由度为 n-k-仁27-3-1=23 的t的临界值t。.。25（23）=2

12、.069 其中截距的t值小于临界值说明截距与零没有显著性差异，三个偏斜 率有一个没有通过显著性检验，t2与t3通过了显著性检验3、 多重共线性的检验YX1X2X3Y1.0000000.9797460.996789-0.383615X10.9797461.0000000.984833-0.407265X20.9967890.9848331.000000-0.416781X3-0.383615-0.407265-0.4167811.000000由上图可知x1与x2之间的相关系数高达0.984833，两者高度正相关。将国内生产总值x1对国家财政支出x2进行回归分析Depe nde nt Variab

13、le: X1Method: Least SquaresDate: 06/03/12 Time: 19:23Sample: 1980 2006In cluded observati ons: 27VariableCoefficie ntStd. Error t-StatisticProb.X25.3643510.189013 28.380900.0000C7063.3492797.184 2.5251640.0183R-squared0.969897Mean depe ndent var60995.78Adjusted R-squared0.968693S.D.dependent var6027

14、7.90S.E. of regressi on10665.50Akaike info criteri on21.45860Sum squared resid2.84E+09Schwarz criteri on21.55459Log likelihood-287.6911F-statistic805.4757Durb in -Watson stat0.144634Prob(F-statistic)0.000000X1i=7063.349+5.3643512R2 二0.969897DW=0.144634F=805.4757因此，x1与x2之间存在显著地线性关系VIF=1/（ 1-R2）=33.21

15、92810 因此该模型具有多重共线性多重共线性修正结果分析运用OLS方法逐一求y对各个结束变量的回归2Y与 x1: y=-1143.176+0.161065x1 R=0.959902VariableCoefficie ntStd. Error t-StatisticProb.C-1143.176559.4057 -2.0435540.0517X10.1610650.006584 24.463690.0000R-squared0.959902Mean depe ndent var8681.087Adjusted R-squared0.958298S.D.dependent var9909.343

16、S.E. of regressi on2023.592Akaike info criteri on18.13432Sum squared resid1.02E+08Schwarz criteri on18.23031Log likelihood-242.8134F-statistic598.4724Durb in -Watson stat0.170737Prob(F-statistic)0.000000R2=0.9935892R =0.147161Y与 x2:y二-292.7317+0.892575x2In eluded observati ons: 27VariableCoeffieie n

17、tStd. Error t-StatistieProb.X20.8925750.014340 62.244310.0000C-292.7317212.2144 -1.3794150.1800R-squared0.993589Mean depe ndent var8681.087Adjusted R-squared0.993332S.D.dependent var9909.343S.E. of regressi on809.1614Akaike info eriteri on16.30106Sum squared resid16368556Schwarz eriteri on16.39705Lo

18、g likelihood-218.0643F-statistie3874.355Durb in -Watson stat0.501126Prob(F-statistie)0.000000= =Y与 x3:y= 68011.85+-564.9916x3Depe ndent Variable: YMethod: Least SquaresDate: 06/03/12 Time: 19:51Sample: 1980 2006In eluded observati ons: 27VariableCoeffieie ntStd. Error t-StatistieProb.C68011.8528622.

19、30 2.3761840.0255X3-564.9916272.0256 -2.0769790.0482R-squared0.147161Mean depe ndent var8681.087Adjusted R-squared0.113047S.D.dependent var9909.343S.E. of regressi on9332.439Akaike info eriteri on21.19157Sum squared resid2.18E+09Sehwarz eriteri on21.28756Log likelihood-284.0862F-statistie4.313843Dur

20、b in -Watson stat0.179687Prob(F-statistie)0.048232由上面的三个基本回归方程可知，x2是最重要的解释变量，所以选择第二个基本回归方程作为出事的回归模型逐步回归将其余变量逐一代入式y=-292.7317+0.892575x2得出如下几个模型Y x2 x3 : y=-6394.656+0.906950x2+56.73074x3R2 二 0.9943832R =0.994815 DW=0.652300 F= 2302.212Depe ndent Variable: YMethod: Least SquaresDate: 06/03/12Time: 20

21、:11Sample: 1980 2006In eluded observati ons: 27VariableCoefficie ntStd. Error t-StatisticProb.C-6394.6562568.992 -2.4891690.0201X20.9069500.014480 62.636270.0000X356.7307423.81565 2.3820780.0255R-squared0.994815Mean depe ndent var8681.087Adjusted R-squared0.994383S.D.dependent var9909.343S.E. of reg

22、ressi on742.7027Akaike info criteri on16.16291Sum squared resid13238574Schwarz criteri on16.30689Log likelihood-215.1993F-statistic2302.212Durb in -Watson stat0.652300Prob(F-statistic)0.000000 丫 x1 x2 x3:丫=-6357.306-0.011191*x1+0.967082*x2+57.11841*x32R2 =0.994954 R =0.994295 DW=0.652300 F=1511.718通

23、过以上分析，得出x1税收影响并不显著，故将其剔除。在删除 X1后模型的统计检验有较大改善，经过以上分析， 丫对X2、X3的回归模型较优。最终回归结果如下：y=-6394.656+0.906950x2+56.73074x3氏=0.994815R2 二 0.994383DW=0.652300 F= 2302.2124、异方差性y=-6394.656+0.906950x2+56.73074x3由G-Q检验，对样本按x2由大到小排序，去除中间的4个样本，剩 余22个样本，再分成两个样本容量为11的子样本，对两个子样本分 别用OLS法做回归子样本1:Depe ndent Variable: YMetho

24、d: Least SquaresDate: 06/03/12Time: 20:58Sample: 1980 1990In eluded observati ons: 11VariableCoeffieie ntStd. Error t-StatistieProb.C-20728.7112657.86 -1.6376150.1401X20.9368950.031075 30.149240.0000X3191.9785126.6051 1.5163570.1679R-squared0.991411Mean depe ndent var17713.26Adjusted R-squared0.9892

25、63S.D.dependent var9994.315S.E. of regressi on1035.583Akaike info eriteri on16.95032Sum squared resid8579459.Schwarz eriteri on17.05883Log likelihood-90.22675F-statistie461.6999Durb in -Watson stat0.713244Prob(F-statistie)0.000000= =Y=-20728.71+0.936895x2+191.9785x3Rf=O. 0.991411 RSS=8579459子样本2:Dep

26、e ndent Variable: YMethod: Least SquaresDate: 06/03/12 Time: 21:14Sample: 1996 2006In eluded observati ons: 11VariableCoeffieie ntStd. Errort-StatistieProb.C-2531.9821207.246-2.0973210.0692X21.2134650.11085010.946950.0000X316.5979012.252481.3546570.2125R-squared0.960384Mean depe ndent var1621.469Adj

27、usted R-squared0.950480S.D.dependent var895.8196S.E. of regressi on199.3467Akaike info eriteri on13.65497Sum squared resid317912.7Sehwarz eriteri on13.76349Log likelihood-72.10233F-statistie96.97020Durb in -Watson stat1.934652Prob(F-statistie)0.000002Y 二2531.982+1.213465x2+16.59790x3氏=0.960384 RS9=3

28、17912.7计算 F 统计量：F=RSSl/(11-2-1) - RSS7(11-2-1)=26.9868在5%的显著性水平下，自由度为(8, 8)的f的分布临界值为Fo.oo5(8,8)=3.44,于是拒绝了同方差的假设，表明元模型存在异方差。异方差性修正结果分析采用加权最小二乘法进行估计:以1/|ei|为权重进行加权最小二乘法，则有Depe ndent Variable: YMethod: Least SquaresDate: 06/03/12 Time: 21:37Sample(adjusted): 1996 2006Included observations: 11 after adjusting endpointsWeighti ng series: 1/ABS(E1)VariableCoefficie ntStd. Errort-StatisticProb.C-2251.989