我国居民消费水平的变量因素分析2010级工程管理赵莹201000271120改革开放以来,我国居民收入与消费水平不断提高,居民消费结构升级和消费需求扩张成为我国经济高速增长的主要动力,特别是进入20世纪90年代以来,居民消费需求对国民经济发展的影响不断增大,对国民经济产生了拉动作用。
我国经济逐步由短缺经济走向过剩经济、由卖方市场转向买方市场,社会消费需求不足,居民消费问题显得更加突出。
特别市对于如何启动内需,扩大居民消费变得越来越重要。
因此,及时把握国民经济发展格局中居民消费需求变动趋势,制定符合我国现阶段情况的国民消费政策,对于提高我国经济增长速度和质量都有重要意义。
我选取了全国1990年-2009年居民消费水平及其影响因素的统计资料,详情如下表所示。
1、建立回归模型并进行参数估计导入数据后得到下表:Dependent Variable Y FzlethGid Least Squares Date 12-13/12 Time: 22:19Sample 1990 2009 Included observations 20Variable Coefficient Std Error 卜 Statist 祀 ProbX1 0 4029000 046073z 7437590..0000X2 -0 023108 0 016025 -1 442040 0 1G8G X3 0-0044740.0055810 001593 0-4345 C-78 54985 50 62100-1 5547960 1396R-squared0.999564Mean dependent \ar 3923.300Adjusted F?-squared 0 999463 S D dependent var 2406 042 S.E. of regression 54.71930 Akaike info criterion11.01317 Sum squared resid47907 22Sch.varz Gritericin11 21831 Log likelihood■106 1917F-statistic12239.64 Du 市in-Watson stat0 92174-9Pro biF-stati stic;0 000000表2由表2可知,模型估计的结果为:Y? 0.403X 10.023X 2 O.OO4X 3 78.550(0.046) (0.016)(0.006)(50.521) t= (8.743) (-1.442) (0.802)(-1.555)R 20.999564 R 20.999483F=12239.64 n=20 D.W.=0.9217、异方差性的检验用怀特检验进行异方差性的检验,得出下表:^Equation: IHITITLEBTo rkfile: 越莹\Hntitled冋区V Proc ObjectF-statistic 1.458301Probability 0.231690Obs*R-squaidd 11 35292Probability0 252291Test Equation;Dependent Variable RESIDE M&thad Least SquaresDate 12/13/12 lime 22 39Sample 1900 2009Included observations 20Variable Coetficieni Std. Error t-Statistic Prob.C-76190.093B52T.01-1 9775760 07 62X1113 432653,16658 2 1336320 C687X1A2-0 0360410,013153-27402030世帕xrx2-0 0179260000845-2.0267330 0702xrx30 007400l UJ2-J13Z5717200 0273X220 733419 989820 2 0754540 0647X2*2-0 0Q326D0 001745-1 3681730 0913X2*X30 0022440 001161 1 9321200 0Q21X3-10.56911 4 S2&719-2 1901630 0533X3^2■0 0004240000187-2 2677390 C467R-squaied0.567646Mean dependent .-ar2395 361Adjusted ht-squared O.170SZB S D dependent 価4157 563S.E. of regression3795.492Akaike 血critenorr19.62787Sum scuared rssid 1.44E+0B Schwarz criterion20.12573Log likehhond-186 2787F-statistic 1 458801Durbin stat2 777616ProbfF-statistic}0 281690表32由表3可知,nR 11.35292,由怀特检验,在a =0.05的情况下,查可知20.05(9)16.92>nR211.35292,表明模型不存在异方差性。
三、序列相关性的检验由表2中结果可知D.W.=0.9217,D.W检验结果表明,在5%的显著性水平下,n=20, k=2,查表得d L 1.20,d u 1.41,由于0<D.W.=0.9217<d L 1.20,故存在正自相关。
得出残差图如下:表4由残差图可知,残差的变动有系统模式,连续为正和连续为负, 存在一阶正自相关。
接下来确定是否存在序列相关性,得出结果、] Fro :| Objec r 比11] 1 亠崗色;rr 吃|Fy 獣卅t Sr^ts 抵id?Breu^ch-Godfrey Serial Gortelation LM TestF-statiaticK 5937D Probflbilit^ 0 004B530b9*R-9<|Li8(ed8 414207P 门凸hiltj0 303TZ3Tast EquationDependent Vaiiabls RE5ID Lkthod LeastDale 1Z14J 2 Tin 出20 53Pres a HI pie mies ng a u@ lagtj^d rasidLials s&t tc zeroVariableCoeficie niSid Error t-StatiSlicProb.X1 心2皿20 043165 4 795265 OOiild X20 027565 0015114 1 B23B5S OOBB2其3 0 009310 00052391 051253 0 03J 9CJ2.J0340 41739971 016014 0 3267 RESIDE)0 793053 0 2417343 30C5T 30 00^19 R-sqlidieJ0 420710(/ledii depsndtfiil Ydi -5.21E-13adjusted R-squared0266233S.D. dependent varS.E of regre9fion43.01330 Aka ike info criterion 10.67321 turn squared resid 2f 752.1b yctivj 白性 cnle r icn10.02215Loe hk^lhcod ■ 100.7321 F-statistic 2.72344GDurbin-讥白站or stat1 727777Pf3biF-£tStlSt (C0.W5235Equation : UMHUED lorkfile:匚叵 表明残差项 如下表:由表 5 可知 nR 2 8.414207 ,查表可得2o.o5(1) 3.84 ,22nR 8.414207 >0.05( 1) 3.84 ,RESID(-1)未通过5%勺显著性检验,表明存在一阶序列相关性。
■ Equation: UNTITLED Vorkf lie: 赵莹\ITnt.-Print Name =reeie =orecast Stats ResideSreusch-Godfrey Serial Correlation LM Test'Test Equation:Dependent Variable RE SID Mei hod Least Squares Date: 12/14/12 Time 21 06Presainple missing \alue lagged residuals set io 7eraVariableCoefficientStd Errort-StatisticProbXI 0 053148 0 055420 -0 958934 0 3533 X2 -0 019331 0.019120 ■1 013679 0.327 X3 0.006315 0.006S09 1 000796 0.3339 C 27 31816 47 33530 0 577120 0 5730 RESID(-1> 0 8380G2 0 252001 3 325626 0.0059 RESID(-2)-0.219797 0305682-0 7190380.4839 R-squared0441341 Mean dependent war -6.21 E-13 Adjusted R-squar 已对 0 241821 £ D dep 已nclEnt T ;ar 50 21387 S E of regression 43 72297 Akaike info criterion 10.G3695 Sum squared res id 26763.78 Schwarz critericn10.93667 Log likelihood-100 3695 F -statistic2212006 Durbin-Watson stat1 &98410PrebfF-statistic}0 111103表6由表 6 可知,nR 2 8.826830 > 2o.°5(2) 5.99, RESID (-2 )未通过 5%的显著性检验,说明不存在二阶序列相关性。
F-statisticObs ft R-squared5 530016Probability 8 826330Probabilrty0 0169830 0'121U(U.x'v 邑: 刊址| Cbj 雲 t] F 「int]忖 a me ★海 | gg 価| F ggg £: 15馆也| 只“血 |Dependent Variable Y Method - Least Squares Date 12」W12 Tine: 21:20 Sample (adjusted; 1991 2009Included observations 19 after adjustments Con'.-ergence achiie/ed after 56 iterationsVariable Coefficient Std Error 卜StatisticProti X1 0 321038 0 082548 3 889123 0 00 tc X2 -0 038591 0 028807 -1 33S921 0 2029 X3 0.0122660 0096501.2710930 224C 36 62573163 73220.2235930 8262AR ⑴07066G0 0 2731742 5S6S510 0215 R-squared0 999686 Mean dw 卩endent var 4005 947 Adjusted R.-squsred 0999597 S D dependent \sr 2356 3 OS S.E of regression 47 32352 Akaike info criterion 10.77233 Sum squared resid 31353 22 Schwarz critenon 11 0213CLog likelihood-97 34155 F -statistic11152 S3D jrbin-'/.'atson stat1 454-231Prab(F-statistic)0 000000Inverted AR Roots.71表7由表7可知,一阶广义差分的估计结果为Y? 0.321X 1 0.039X 2 O.OI2X 3 36.626 0.707AR(1)t= (3.889) (-1.336) (1.271)R 20.999597 D.W.=1.454231由于d u 1.40<D.W.< 1- d u ,判断是不存在序列相关性。