eviews软件对于我国城镇居民消费性支出和可支配收入的分析

我国城镇居民消费性支出和可支配收入的分析我国城镇居民消费性支出和可支配收入的分析一：研究目的及要求居民消费支出是指城乡居民个人和家庭用于生活消费以及集体用于个人消费的全部支出。居民可支配收入是居民家庭在调查期获得并且可以用来自由支配的收入。随着市场经济的稳定繁荣和改革开放的深入发展，我国人均生活水平有了大幅度提高，其主要表现在人均可支配收入的增长。为研究我国城镇居民消费支出与收入的相关性，探讨城镇居民可支配收入与消费性支出之间数量关系的基本规律，揭示可支配收入在居民消费性支出中的作用，对于宏观经济运行提出合理化建议，根据1994——2008年全国城镇居民消费性支出与可支配收入的基本数据，利用EVIEWS软件对计量模型进行了参数估计和检验，对城镇居民消费性支出与可支配收入之间数量关系进行分析从而证明增加居民收入来刺激消费，增加消费性支出的必要性。二、模型设定及其估计食品支出，居民住房，医疗保健以及衣着对于居民日常生活来说是必不可少的支出，因此我考虑的影响因素主要有食品支出X2，居住支出X3,医疗保健X4，衣着方面X5，建立了下述的一般模型：+et其中tY——居民的可支配收入2X——食品支出3X——居住支出4X——医疗保健5X——衣着支出et——随即扰动项。从1995---2009年的中国统计年鉴中收集到以下数据：年份Y收入（元）X2(食品支出)（元）X3(居住)（元）X4(医疗保健)（元）X5(衣着支出)（元）19943496.241422.49193.1682.89390.3819954282.951766.02250.18110.11479.2019964838.901904.71300.85143.28527.9519975160.321942.59358.64179.68520.9119985425.051926.89172.96257.15311.0119995854.021932.10453.99245.59482.3720006279.981958.31500.49318.07500.4620016859.582014.02547.96343.28533.6620027702.802271.84624.36430.08590.8820038472.202416.92699.39475.98637.7320049421.612709.60733.53528.15686.79200510493.032914.39808.66600.85800.51200611759.453111.92904.19620.54901.78200713785.813628.03982.28699.091042.00200815780.764259.811145.41786.201165.91利用Eviews软件，输入Y、X2、X3、X4、X5、X6等数据，采用这些数据对模型进行OLS回归，结果如表1：DependentVariable:YMethod:LeastSquaresDate:12/16/10Time:11:19Sample:19942008Includedobservations:15VariableCoefficientStd.Errort-StatisticProb.X22.1940210.5859043.7446790.0038X30.2143991.5857390.1352040.8951X46.3047982.0006123.1514340.0103X52.0980001.9080981.0995240.2973C-1227.160365.0907-3.3612460.0072R-squared0.997209Meandependentvar7974.180AdjustedR-squared0.996092S.D.dependentvar3628.636S.E.ofregression226.8423Akaikeinfocriterion13.94759Sumsquaredresid514574.4Schwarzcriterion14.18361Loglikelihood-99.60692F-statistic893.0849Durbin-Watsonstat1.471612Prob(F-statistic)0.000000表1表2残差图-400-20002004000400080001200016000949596979899000102030405060708ResidualActualFitted表2由表2可以看出，残差的变动有系统模式，连续为正和连续为负，表明残差项存在一阶正自相关，模型中t统计量和F统计量的结论不可信，需采取补救措施。根据表1可以看出，模型估计的结果为：ˉ1227.160+2.1940212X+0.2143993X+6.3047984X+2.0980005X(365.0907)(0.585904)(1.585739)(2.000612)(1.908098)t=(-3.361246)(3.744679)(0.135204)(3.151434)(1.099524)一、统计检验（1）拟合优度：由表1中数据可以得到：R2=0.9972，修正的可决系数为0.9961，可决系数很高，这说明模型对样本的拟合很好。（2）F检验：针对0234:0H，给定显著性水平0.05，在F分布表中查出自由度为k-1=3和n-k=11的Fα=893.0849〉F0.05(3,11)=3.59,明显显著，应拒绝原假设0234:0H，表明模型的线性关系在95%的置信水平下显著成立。（3）t检验：在5%的显著性水平下，自由度n-k-1=10的t统计量的临界值为t0.025（10）=2.23，则可以得出X3、X5前参数的估计值未能通过t检验。计算各解释变量的相关系数，选择X2、X3、X4、X5数据，得相关系数矩阵表3X2X3X4X5X21.0000000.7417770.9465770.971292X30.7417771.0000000.6710580.950386X40.9465770.6710581.0000000.607836X50.9712920.9503860.6078361.000000表3由相关系数矩阵可以看出：各解释变量相互之间的相关系数较高，证实确实存在严重多重共线性。三、消除多重共线性采用逐步回归的办法，去检验和解决多重共线性问题。分别作Y对X2、X3、X4、X5的一元回归，结果如下：DependentVariable:YMethod:LeastSquaresDate:12/20/10Time:10:53Sample:19942008Includedobservations:15VariableCoefficientStd.Errort-StatisticProb.X24.6251330.18175625.447000.0000C-3181.529459.1190-6.9296380.0000R-squared0.980319Meandependentvar7974.180AdjustedR-squared0.978806S.D.dependentvar3628.636S.E.ofregression528.2685Akaikeinfocriterion15.50065Sumsquaredresid3627879.Schwarzcriterion15.59506Loglikelihood-114.2549F-statistic647.5497Durbin-Watsonstat0.708690Prob(F-statistic)0.000000DependentVariable:YMethod:LeastSquaresDate:12/20/10Time:10:55Sample:19942008Includedobservations:15VariableCoefficientStd.Errort-StatisticProb.X311.770550.81246014.487550.0000C1166.053525.32822.2196670.0448R-squared0.941675Meandependentvar7974.180AdjustedR-squared0.937188S.D.dependentvar3628.636S.E.ofregression909.4170Akaikeinfocriterion16.58705Sumsquaredresid10751511Schwarzcriterion16.68146Loglikelihood-122.4029F-statistic209.8891Durbin-Watsonstat1.493135Prob(F-statistic)0.000000DependentVariable:YMethod:LeastSquaresDate:12/20/10Time:10:56Sample:19942008Includedobservations:15VariableCoefficientStd.Errort-StatisticProb.X415.842050.94895616.694190.0000C1826.470421.59404.3322970.0008R-squared0.955433Meandependentvar7974.180AdjustedR-squared0.952005S.D.dependentvar3628.636S.E.ofregression794.9544Akaikeinfocriterion16.31801Sumsquaredresid8215383.Schwarzcriterion16.41242Loglikelihood-120.3851F-statistic278.6960Durbin-Watsonstat0.517894Prob(F-statistic)0.000000DependentVariable:YMethod:LeastSquaresDate:12/20/10Time:10:57Sample:19942008Includedobservations:15VariableCoefficientStd.Errort-StatisticProb.X514.516981.08233813.412610.0000C-1289.203735.3244-1.7532440.1031R-squared0.932607Meandependentvar7974.180AdjustedR-squared0.927423S.D.dependentvar3628.636S.E.ofregression977.5601Akaikeinfocriterion16.73156Sumsquaredresid12423110Schwarzcriterion16.82597Loglikelihood-123.4867F-statistic179.8981Durbin-Watsonstat1.301498Prob(F-statistic)0.000000统计如下为：变量X2X3X4X5参数估计值4.62513311.7705515.8420514.51698t统计量25.4470014.4875516.6941913.412612R0.97880.9371880.9520050.927423按2R的大小排序为：X2X4X3X5以X2为基础，顺次加入其他变量逐步回归。1)首先加入X5回归结果为：-3085.316+4.3021132X+1.07.02015Xt=(5.449761)(0.421283)R2=0.977374当取05.0时，tt0.05,X5参数的t检验不显著，予以剔除.2)再加入X4回归结果为：-1488.048+2.9138412X+6.2724784Xt=（10.87632）(6.748146)R2=