计量经济学论文(eviews分析)影响我国人口预期寿命的可能因素分析》

影响我国人口预期寿命的可能因素分析摘要：笔者针对各个地区间的人均寿命存在差异比较迷惑，是什么因素造成了地区间的这种差异呢？在这篇文章中，我以2000年全国的各地区的人均寿命（每10年才统计一次）以及一些相关数据来进行分析。寻找造成各地人均寿命存在差异的原因。关键词：人均寿命人均GDP卫生机构数单位面积污染比教育支出（由于造成寿命的长短的因素很多，笔者只是找到几个关键因素进行分析）地区地区人均寿命y人均GDP（万元）x1卫生机构数x2单位面积污染比x3教育支出（万元）x4北京76.11.8266476176184402503068天津74.911.637722298314822.0339685499.8河北72.540.762505206634978.787881559084.1山西71.650.506099137364070.55215794624.1内蒙古69.870.5906457852387.642276580861.6辽宁73.341.115932125646205.263161463314.7吉林73.10.67904255441580.51282902941.6黑龙江72.370.8544798038884.6625771175133上海78.142.7734513689086.68732008864.6江苏73.911.171384128138484.112152905677.2浙江74.71.31339170346140.566042200449.9安徽71.850.48333467052720.689661129954福建72.551.14958198072262.41225753.9江西68.950.48278480481275.86207752776.6山东73.920.949371171187659.748432462775.5河南71.540.54149107644273.563221709181.4湖北71.080.717503110652909.743591706416.4湖南70.660.562615246781629.680371523036.8广东73.271.253695134994291.752583609720.5广西71.290.431608137071872.76423934717.9海南72.920.65713626891225.9887223093.3重庆71.730.51401793752234.19204698720.6四川71.20.466233351940.7480311619987.5贵州65.960.26451889922193.22034528486.6云南65.490.46099713356668.856448976174.5西藏64.370.45527112371.1811023681549.8陕西70.070.455796107371111.682241014584.4甘肃67.470.3845767191597.014925535287.2青海66.030.509845184780.9333333128177.5宁夏70.170.47936813612088.15029148649.4新疆67.410.7378916705116.407186702243.4表一（来源于国家统计局网站）一、建立模型并回归建立回归方程：Y=a0+a1*X1+a2*X2+a3*X3+a4*X4+u(式1)Y地区人均寿命，X1人均GDP（万元），X2卫生机构数，X3单位面积污染比，X4教育支出（万元运用OLS估计方法对式1中的参数进行估计，得回归分析结果：（表2）DependentVariable:YMethod:LeastSquaresDate:12/25/10Time:20:49Sample:131Includedobservations:31VariableCoefficientStd.Errort-StatisticProb.C65.976461.13033158.369160.0000X14.9821041.6594543.0022540.0059X27.02E-057.05E-050.9958130.3285X3-2.99E-054.42E-05-0.6769880.5044X45.41E-077.19E-070.7534090.4580R-squared0.652727Meandependentvar71.24387AdjustedR-squared0.599300S.D.dependentvar3.191195S.E.ofregression2.020053Akaikeinfocriterion4.390815Sumsquaredresid106.0960Schwarzcriterion4.622103Loglikelihood-63.05763F-statistic12.21726Durbin-Watsonstat1.285255Prob(F-statistic)0.000010从以上的报告单可得到回归方程为:Y=65.97646203+4.982103524*X1+7.01887465e-005*X2-2.990581087e-005*X3+5.414193515e-007*X4二、模型的经济意义检验Y=65.97646203+4.982103524*X1+7.01887465e-005*X2-2.990581087e-005*X3+5.414193515e-007*X4回归方程表明：⑴X1人均GDP（万元），X2卫生机构数，X3单位面积污染比，X4教育支出都=0时，地区人均寿命是65.97646203表明没有任何卫生机构没有污染和教育支出等因素的影响下人均寿命约是66岁。⑵当其他条件不变，每增加一单位的人均GDP，人均寿命增长五岁。⑶当其他条件不变时，每增加一单位的卫生机构，人均寿命增加7.01887465e-005个单位。⑷当其他条件不变时，每增加一单位的面积污染比，人均寿命减少-2.990581087e-005。⑸当其他条件不变时，每增加一单位教育支出，人均寿命增加5.414193515e-007个单位。经济意义检验表明卫生机构、面积污染比、教育支出与人均寿命的关系微乎其微。我们可以得出人均寿命只与人均GDP存在很明显的关系。现在单独对人均寿命和人均GDP进行回归。结果的报告单为DependentVariable:YMethod:LeastSquaresDate:12/25/10Time:21:18Sample:131Includedobservations:31VariableCoefficientStd.Errort-StatisticProb.C67.510130.70612395.606750.0000X14.5951260.7326516.2719170.0000R-squared0.575632Meandependentvar71.24387AdjustedR-squared0.560999S.D.dependentvar3.191195S.E.ofregression2.114395Akaikeinfocriterion4.397756Sumsquaredresid129.6494Schwarzcriterion4.490271Loglikelihood-66.16521F-statistic39.33694Durbin-Watsonstat1.169499Prob(F-statistic)0.000001报告单数据表明最后的回归方程为Y=67.51013261+4.595125638*X1经济意义是：⑴当人均GDP=0时人均寿命约是67.5⑵当人均GDP没增加一个单位，人均寿命4.6个单位。三、统计意义的检验报告单数据表明：１、F-statistic39.33694Prob(F-statistic)0.000001F检验的P值是0.0000010.05T检验的P值是00.05F检验和T检验说明人居ＧＤＰ在方程中显著２、R-squared0.575632AdjustedR-squared0.560999Ｒ＝0.575632＞０．３表明样本方程呗解释变量解释的程度是５７．５６３２％四、经济计量学检验⑴异方差检验①、散点图检验做出ｘ１与残差的散点图散点图表明可能不存在异方差。②进一步进行异方差检验，采用怀特检验。怀特检验的报告单为WhiteHeteroskedasticityTest:F-statistic2.303808Probability0.118507Obs*R-squared4.380453Probability0.111891TestEquation:DependentVariable:RESID^2Method:LeastSquaresDate:12/25/10Time:21:49Sample:131Includedobservations:31VariableCoefficientStd.Errort-StatisticProb.C11.565183.7171093.1113360.0043X1-13.661637.092647-1.9261680.0643X1^24.0023112.5403491.5754960.1264R-squared0.141305Meandependentvar4.182237AdjustedR-squared0.079970S.D.dependentvar6.093022S.E.ofregression5.844318Akaikeinfocriterion6.460583Sumsquaredresid956.3694Schwarzcriterion6.599355Loglikelihood-97.13903F-statistic2.303808Durbin-Watsonstat1.862642Prob(F-statistic)0.118507怀特检验结果：Obs*R-squared4.380453Probability0.111891怀特检验的统计量＝4.380453，Ｐ值为0.11189＞０．０５怀特检验表明不存在异方差。（二）自相关性检验①残差的图示检验残差的图示表明残差之间可能不存在自行相关性。②ＬＭ检验ＬＭ的检验报告表如下ARCHTest:F-statistic0.758037Probability0.478663Obs*R-squared1.597835Probability0.449816TestEquation:DependentVariable:RESID^2Method:LeastSquaresDate:12/25/10Time:22:11Sample(adjusted):331Includedobservations:29afteradjustmentsVariableCoefficientStd.Errort-StatisticProb.C3.4098191.5219552.2404200.0338RESID^2(-1)0.2281300.2030581.1234720.2715RESID^2(-2)0.0334500.2029260.1648400.8703R-squared0.055098Meandependentvar4.468796AdjustedR-squared-0.017587S.D.dependentvar6.201487S.E.ofregression6.255782Akaikeinfocriterion6.602587Sumsquaredresid1017.505Schwarzcriterion6.744031Loglikelihood-92.73750F-statistic0.758037Durbin-Watsonstat1.957170Prob(F-statistic)0.478663ＬＭ的检验表明Obs*R-squared1.59