1、序列相关性检验（一）一元线性回归结果：Dependent Variable: YMethod: Least SquaresDate: 06/01/12 Time: 14:16Sample: 1981 2007Included observations: 27Variable Coefficient Std. Error t-Statistic Prob. C 4276.362 1079.786 3.960380 0.0005X 0.871668 0.029448 29.60012 0.0000R-squared 0.972258 Mean dependent var 24869.44Adjus
2、ted R-squared 0.971149 S.D. dependent var 25261.92S.E. of regression 4290.920 Akaike info criterion 19.63758Sum squared resid 4.60E+08 Schwarz criterion 19.73356Log likelihood -263.1073 F-statistic 876.1668Durbin-Watson stat 0.174669 Prob(F-statistic) 0.000000（二）拉格朗日乘数检验：含二阶残差项的回归结果：Breusch-Godfrey
3、Serial Correlation LM Test:F-statistic 120.8648 Probability 0.000000Obs*R-squared 24.65421 Probability 0.000004Test Equation:Dependent Variable: RESIDMethod: Least SquaresDate: 06/01/12 Time: 13:50Variable Coefficient Std. Error t-Statistic Prob. C 361.5102 372.6461 0.970117 0.3421X -0.025697 0.0132
4、22 -1.943398 0.0643RESID(-1) 1.477525 0.193620 7.631049 0.0000RESID(-2) -0.485298 0.229297 -2.116459 0.0453R-squared 0.913119 Mean dependent var -2.29E-12Adjusted R-squared 0.901787 S.D. dependent var 4207.593S.E. of regression 1318.618 Akaike info criterion 17.34251Sum squared resid 39991346 Schwar
5、z criterion 17.53449Log likelihood -230.1239 F-statistic 80.57655Durbin-Watson stat 1.772240 Prob(F-statistic) 0.000000含三阶残差项的回归结果：Breusch-Godfrey Serial Correlation LM Test:F-statistic 77.16026 Probability 0.000000Obs*R-squared 24.65663 Probability 0.000018Test Equation:Dependent Variable: RESIDMet
6、hod: Least SquaresDate: 06/01/12 Time: 13:53Variable Coefficient Std. Error t-Statistic Prob. C 340.4064 405.7832 0.838887 0.4106X -0.024688 0.015080 -1.637160 0.1158RESID(-1) 1.464982 0.214682 6.823974 0.0000RESID(-2) -0.441789 0.371964 -1.187721 0.2476RESID(-3) -0.039199 0.260256 -0.150618 0.8816R
7、-squared 0.913208 Mean dependent var -2.29E-12Adjusted R-squared 0.897428 S.D. dependent var 4207.593S.E. of regression 1347.559 Akaike info criterion 17.41555Sum squared resid 39950151 Schwarz criterion 17.65552Log likelihood -230.1100 F-statistic 57.87019Durbin-Watson stat 1.751706 Prob(F-statisti
8、c) 0.000000序列相关性消除（一）二阶迭代法回归结果：Dependent Variable: YMethod: Least SquaresDate: 06/01/12 Time: 15:22Sample(adjusted): 1983 2007Included observations: 25 after adjusting endpointsConvergence not achieved after 100 iterationsVariable Coefficient Std. Error t-Statistic Prob. C 921803.7 58583291 0.015735
9、 0.9876X 0.603615 0.087945 6.863519 0.0000AR(1) 1.519561 0.189668 8.011678 0.0000AR(2) -0.520079 0.203612 -2.554264 0.0185R-squared 0.998698 Mean dependent var 26697.20Adjusted R-squared 0.998512 S.D. dependent var 25384.37S.E. of regression 979.2577 Akaike info criterion 16.75711Sum squared resid 2
10、0137857 Schwarz criterion 16.95213Log likelihood -205.4639 F-statistic 5368.622Durbin-Watson stat 1.759566 Prob(F-statistic) 0.000000Inverted AR Roots 1.00.52再用拉格朗日乘数检验自相关是否已消除含二阶残差项回归结果：Breusch-Godfrey Serial Correlation LM Test:F-statistic 0.831638 Probability 0.450575Obs*R-squared 2.011998 Probab
11、ility 0.365679Test Equation:Dependent Variable: RESIDMethod: Least SquaresDate: 06/01/12 Time: 15:27Variable Coefficient Std. Error t-Statistic Prob. C 33620447 89317489 0.376415 0.7108X -0.038309 0.101577 -0.377143 0.7102AR(1) -0.009663 0.725955 -0.013310 0.9895AR(2) 0.028647 0.757768 0.037804 0.97
12、02RESID(-1) 0.157362 0.732418 0.214852 0.8322RESID(-2) -0.265974 0.409538 -0.649448 0.5238R-squared 0.080480 Mean dependent var 3.544879Adjusted R-squared -0.161499 S.D. dependent var 916.0045S.E. of regression 987.2043 Akaike info criterion 16.83319Sum squared resid 18516875 Schwarz criterion 17.12572Log likelihood -204.4149 F-statistic 0.332591Durbin-Watson stat 1.948202 Prob(F-statistic) 0.886935
