统计学论文

1计量经济学论文各地区的国内生产总值的影响因素班级：09级国贸5班姓名：郝红杰学号：200919111200时间：2012年6月2摘要:各地区国内生产总值受多种因素影响，根据全国31个省市的相关经济变量数据，对其进行计量分析。应用计量经济学所学知识对根据经济理论选取的影响各地区国内生产总值的各因素进行分析、检验，并对其影响程度的大小进行定量分析，进一步明确和完善相关的经济学知识。本文选取2010年的相关数据关键词：生产总值消费性支出固定资产投资政府支出出口额3一、前言：我国各地区居民的消费支出、固定资产的投资、政府支出及出口额在很大程度上决定了当地的生产总值。现利用计量经济学中的知识对此进行分析，研究各影响因素的影响程度。二、模型的选择与建立本文选择了四个解释变量对我国2010年各地区的国内生产总值进行分析，并利用计量经济学方法对所建立模型进行定量分析。模型的变量选择如下：Y---国内生产总值（亿元）X1----居民消费性支出（元）X2----固定资产投资（亿元）X3----政府支出（亿元）X4----各地出口额（万美元）模型的变量数据如下：地区生产总值消费性支出固定资产投资政府支出出口额北京14113.5819934.484916.5262717.3175543621天津9224.4616561.775896.5191376.843748483河北20394.2610318.3212922.662820.2442255644山西9200.869792.655526.6031931.364470281.5内蒙古1167213994.628687.9862273.505333442.6辽宁18457.2713280.0415106.333195.8164309871吉林8667.5811679.047395.2321787.248447584.9黑龙江10368.610683.926292.6742253.2691628079上海17165.9823200.44630.4723302.88618071398江苏41425.4814357.4917416.474914.0627053869浙江27722.3117858.28438.0753207.88318046478安徽12359.3311512.5510281.292587.6141241289福建14737.1214750.017385.7811695.0917149313江西9451.2610618.697856.9411923.2631341606山东39169.9213118.2418844.414145.03210422560河南23092.3610838.4913934.823416.1431052937湖北15967.6111450.979405.632501.4031444180湖南16037.9611825.338617.9842702.475795598.9广东46013.0618489.5312599.265421.543453191164广西9569.8511490.086383.2562007.591960307.3海南2064.510926.711257.5581.3379232033重庆7925.5813335.026170.611709.035748893.7四川17185.4812105.0911061.384257.9811884063贵州4602.1610058.292609.3591631.479192018云南7224.1811074.085052.6132285.723760576.7西藏507.469685.54404.9806551.036277102.9陕西10123.4811821.887569.9012218.828620821.5甘肃4120.759895.352808.5511468.581163778.5青海1350.439613.79840.0123743.403346619.7宁夏1689.6511334.431292.802557.5285117000.2新疆5437.4710197.093065.1331698.9131296865作出各解释变量与被解释变量的散点图：（1）x1与Y的散点图：01000020000300004000050000500010000150002000025000X1Y（2）X2与Y的散点图：50100002000030000400005000005000100001500020000X2Y（3）X3与Y的散点图：010000200003000040000500000100020003000400050006000X3Y（4）X4与Y的散点图：60100002000030000400005000001000000020000000300000004000000050000000X4Y由散点图知，Y与X1、X2、x3、X4呈线性关系，所以设模型为：Y=β0+β1x1+β2x2+β3x3+β4x4+μ三、模型参数估计与回归结果分析Y与X1、X2、X3、X4的回归分析结果如下：DependentVariable:YMethod:LeastSquaresDate:06/02/12Time:08:13Sample:131Includedobservations:31VariableCoefficientStd.Errort-StatisticProb.C-1304.3632237.124-0.5830530.5649X10.0085080.1724650.0493340.9610X21.2629460.1821266.9344580.0000X31.1678230.9277471.2587730.2193X40.0005807.49E-057.7350380.0000R-squared0.964615Meandependentvar14098.13AdjustedR-squared0.959171S.D.dependentvar11401.35S.E.ofregression2303.768Akaikeinfocriterion18.46917Sumsquaredresid1.38E+08Schwarzcriterion18.70046Loglikelihood-281.2721F-statistic177.19457Durbin-Watsonstat1.991912Prob(F-statistic)0.000000Y=-1304.36+0.0085X1+1.26X2+1.17X3+0.00058X4(-0.58)(0.049)(6.93)(1.26)(7.74)R2=0.9646F=177.19D.W.=1.99（一）经济意义检验从经济意义上说，各地区的生产总值Y与各地消费性支出x1、各地固定资产投资X2、各地政府支出X3、各地出口额X4成正相关。由分析结果知:变量x1、x2、x3、x4的系数均为正数，故符合实际经济意义，保留四个解释变量。（二）统计检验由回归结果表明，R2的值接近于1，表明模型的拟合优度较好。在α=0.05的显著性水平下，x2、x4的t统计量的伴随概率均为零，所以x2、x4在95%的水平下影响显著，通过了变量显著性检验；x1、x3的t统计量的伴随概率均大于0.05，所以x1、x3在95%的水平下影响不显著，未通过变量显著性检验。F统计量的伴随概率为零，所以模型的线性关系在95%的置信水平下显著成立。（三）计量经济学检验1、多重共线性检验：（1）做Y与X1的回归，结果如下：DependentVariable:YMethod:LeastSquaresDate:06/02/12Time:08:38Sample:131Includedobservations:318VariableCoefficientStd.Errort-StatisticProb.C-6782.7217430.367-0.9128380.3689X11.6354290.5639602.8999040.0070R-squared0.224795Meandependentvar14098.13AdjustedR-squared0.198063S.D.dependentvar11401.35S.E.ofregression10210.01Akaikeinfocriterion21.36247Sumsquaredresid3.02E+09Schwarzcriterion21.45498Loglikelihood-329.1182F-statistic8.409445Durbin-Watsonstat1.667767Prob(F-statistic)0.007048Y=-6782.72+1.64X1（-0.91）（2.90）R2=0.2248F=8.41D.W.=1.67（2）作Y与x2的回归，结果如下:DependentVariable:YMethod:LeastSquaresDate:06/02/12Time:08:45Sample:131Includedobservations:31VariableCoefficientStd.Errort-StatisticProb.C-1200.1532129.272-0.5636450.5773X22.0208940.2392498.4468120.0000R-squared0.711008Meandependentvar14098.13AdjustedR-squared0.701042S.D.dependentvar11401.35S.E.ofregression6233.917Akaikeinfocriterion20.37574Sumsquaredresid1.13E+09Schwarzcriterion20.46825Loglikelihood-313.8239F-statistic71.34863Durbin-Watsonstat1.790849Prob(F-statistic)0.0000009Y=-1200.15+2.02x2(-0.56)(8.45)R2=0.7010F=71.34D.W.=1.79（3）作Y与x3的回归，结果如下:DependentVariable:YMethod:LeastSquaresDate:06/02/12Time:08:46Sample:131Includedobservations:31VariableCoefficientStd.Errort-StatisticProb.C-6698.2131832.491-3.6552510.0010X38.7256080.68869812.669720.0000R-squared0.846983Meandependentvar14098.13AdjustedR-squared0.841707S.D.dependentvar11401.35S.E.ofregression4536.146Akaikeinfocriterion19.73988Sumsquaredresid5.97E+08Schwarzcriterion19.83240Loglikelihood-303.9682F-statistic160.5219Durbin-Watsonstat2.017763Prob(F-statistic)0.000000Y=-6698.21+8.73x3(-3.66)(12.67)R2=0.8470F=160.52D.W.=2.02（4）作Y与x4的回归，结果如下:DependentVariable:YMethod:LeastSquaresDate:06/02/12Time:08:46Sample:13110Includedobservations:31VariableCoefficientStd.Errort-StatisticProb.C9276.0211366.6946.7871