计量经济学 实验报告 4.docx
计量经济学实验报告四开课实验室:财经科学实验室2014年5月25日班级:财政1112班 学号: 班11410060 姓名:蓝锦达实验项目名称:序歹ll相关 成绩:实验性质:口验证性口综合性口设计性指导教师签字:【实验目的】掌握用Eviews软件进行参数估计、检验的方法,并进行预测。【实验要求】熟悉基本操作步骤,读懂各项上机输出结果的含义并能进行分析,预测。【实验软件】Eviews, Excel 软件【实验方案与进度】1.学习软件操作,运用EViews软件建立新文件obsXY1960I157I143.00001961162.0000146.00001962169.0000153.00001963176.0000160.00001964188.0000169.00001965200.0000180.00001966211.0000190.00001967220.0000196.00001968230.0000207.00001969237.0000215.00001970247.0000220.00001971256.0000228.00001972268.0000242.00001973287.0000253.00001974285.0000251.00001975290.0000257.00001976301.0000271.00001977311.0000283.00001978326.0000295.00001979335.0000302.00001980337.0000301.00001981345.0000305.00001982348.0000308.00001983358.0000324.00001984384.0000341.00001985396.0000357.00001986409.0000371.00001987415.0000382.00001988432.0000397.00001989440.0000406.00001990448.0000413.00001991449.0000411.00001992461.0000422.00001993467.0000434.00001994478.0000447.00001995493.0000458.00003 .进行图形分析、回归分析105-0-5-QS 山江1002003004005004 .残差序列进行滞后一期的自回归5 .利用广义差分法进行,对变化后的模型进行回归【实验过程】下表给出了美国1960-1995年36年间个人实际可支配收入X和个人实际消费支出丫的 数据。美国个人实际可支配收入和个人实际消费支出单位:100亿美元年份个人实际可 支配收入X个人实际消费支出Y年份个人实际可 支配收入X个人实际 消费支出Y196015714319783262951961162146197933530219621691531980337301196317616019813453051964188169198234830819652001801983358324196621119019843843411967220196198539635719682302071986409371196923721519874153821970247220198843239719712562281989440406197226824219904484131973287253199144941119742852511992461422197529025719934674341976301271199447844719773112831995493458注:费料来源于Economic Report of the President,数据为1992年价格。要求:(1)用普通最小二乘法估计收入一消费模型;Yt=/3y+/32X2+ut建立个人实际消费支出对个人实际可支配收入的回归模型,利用Y和X的数据表,Dependent Variable: Y Method: Least SquaresDate: 05/14/14 Time: 10:35Sample: 1960 1995Included observations: 36VariableCoefficientStd. Error t-StatisticProb.C-9.4287452.504347 -3.7649510.0006X0.9358660.007467125.34110.0000R-squared0.997841Mean dependent var289.9444Adjusted R-squared0.997777SD dependent var95.82125S E. of regression4.517862Akaike info criterion5.907908Sum squared resid693.9767Schwarz criterion5.995881Log likelihood-104.3423F-statistic15710.39Durbin-Watson stat0.523428Prob(F-statistic)0.000000得到回归分析结果为:Y =-9.428745 +0.935866XSE=(2. 504347)(0. 007467)t=(-3. 764951)(125. 3411)r2 =0 997841, /?2 =0.997777, F = 15710.39, DW = 0.523428(2)检验收入一消费模型的自相关状况(5%显著水平);对残差序列进行滞后一期的自回归,得自回归方程6 = 0.728550e-由此可知,p = 0.728550所以 DW = 2(1 - 0.728550) = 0.5429在。=0.05下,查表可知% =1.402,% = 1.519,模型中 OW = 0.5428<4 =1.402,说 明在5%的显著水平下,存在正自相关。(3)用广义差分法的方法消除模型中存在的问题,并检验修正后的模型是否存在序列 相关。注:用弓二爪1+匕来估计力。由于样本容量较大,不必对第一个观测值进行变换。采用广义差分法,变化后模型的回归结果(表2)Dependent Variable: YP Method: Least Squares Date: 05/25/14 Time: 14:45 Sample: 1960 1995 Included observations: 36Variable Coefficient Std. Error t-Statistic Prob.C-3.7789691.842792-2.0506760.0481XP0.9482880.01857351.056530.0000R-squared0.987125Adjusted R-squared0.986746S E of regression3.022923Sum squared resid310.6942Log likelihood-89.87701Durbin-Watson stat2.097821Mean dependent var86.72292S.D. dependent var26.25780Akaike info criterion5.104279Schwarz criterion5.192252F-statistic2606.769Prob(F-statistic)0.000000et = 0.728550,.I/ -1g=-3.778969 + 0.948288X;Se = (1.842792)(0.018573)t = (-2.050676)(51.05653)R2 = 0.987125, F = 2606.769, DW = 2.097821查5%显著水平的DW统计表可知 = L402,d = 1.519 ,模型中OW = 2.097821 > d” ,说明广义差分模型中已无自相关。同时,可决系数R2、t统计量、F统计量均达到理想水平。由查分方程式有:£)=3.7789691-0.728550= 13.921418最终的消费模型为:»* -13.9214184-0.948288X;由此模型可知,个人实际消费支出对个人实际可支配收入的边际消费倾向为 0. 948288,即个人实际可支配收入增加1单位,个人实际支出将增加0. 948288个单位。注:应包括数据、图表、计算等。【实验结果及分析】运用EViews软件进行模型建立和检验,通过进行DW检验和广义差分法的模型变化,我们可以得到相关结论:1 .在a = 0O5下,查表可知& =1.402,4, =1.519,模型中QW = 0.5428<4 =1.402, 说明在5%的显著水平下,存在正自相关。2 .查5%显著水平的DW统计表可知4 =1.402,4, =1.519,模型中DW = 2.097821d,说明广义差分模型中已无自相关。同时,可决系数R?、t统计量、F统计量均达到理想水平。由查分方程式有:B° =3.7789691 0.728550= -13.921418最终的消费模型为:g* =-13.921418 + 0.948288%;由此模型可知,个人实际消费支出对个人实际可支配收入的边际消费倾向为0.948288,即个人实际可支配收入增加1单位,个人实际支出将增加0. 948288个单位。