计量经济学论文范文eviews

《我国财政收入影响因素分析》
班级：09财政1班
姓名：***
学号：************
指导教师：***
完成时间：2011年12月4日
摘要：对我国财政收入影响因素进行了定量分析，建立了数学模型，并提出了提高我国财政收入质量的政策建议。

关键词：财政收入实证分析影响因素
一、引言
财政收入对于国民经济的运行及社会发展具有重要影响。

首先，它是一个国家各项收入得以实现的物质保证。

一个国家财政收入规模大小往往是衡量其经济实力的重要标志。

其次，财政收入是国家对经济实行宏观调控的重要经济杠杆。

宏观调控的首要问题是社会总需求与总供给的平衡问题，实现社会总需求与总供给的平衡，包括总量上的平衡和结构上的平衡两个层次的内容。

财政收入的杠杆既可通过增收和减收来发挥总量调控作用，也可通过对不同财政资金缴纳者的财政负担大小的调整，来发挥结构调整的作用。

此外，财政收入分配也是调整国民收入初次分配格局，实现社会财富公平合理分配的主要工具。

在我国，财政收入的主体是税收收入。

因此，在税收体制及政策不变的情况下，财政收入会随着经济繁荣而增加，随着经济衰退而下降。

我国的财政收入主要包括税收、国有经济收入、债务收入以及其他收入四种形式，因此，财政收入会受到不同因素的影响。

从国民经济部门结构看，财政收入又表现为来自各经济部门的收入。

财政收入的部门构成就是在财政收入中，由来自国民经济各部门的收入所占的不同比例来表现财政收入来源的结构，它体现国民经济各部门与财政
收入的关系。

我国财政收入主要来自于工业、农业、商业、交通运输和服务业等部门。

因此，本文认为财政收入主要受到总税收收入、国内生产总值、其他收入和就业人口总数的影响。

二、预设模型
令财政收入Y（亿元）为被解释变量，总税收收入X1（亿元）、国内生产总值X2（亿元）、其他收入X3（亿元）、就业人口总数为X4（万人）为解释变量，据此建立回归模型。

二、数据收集
从《2010中国统计年鉴》得到1990--2009年每年的财政收入、总税收收入、国内生产总值工、其他收入和就业人口总数的统计数据如下：
obs 财政收入Y 总税收收入X1 国内生产总值X2 其他收入X3 就业人口总数X4 1990 2937.1 2821.86 18667.8 299.53 64749 1991 3149.48 2990.17 21781.5 240.1 65491 1992 3483.37 3296.91 26923.5 265.15 66152 1993 4348.95 4255.3 35333.9 191.04 66808 1994 5218.1 5126.88 48197.9 280.18 67455 1995 6242.2 6038.04 60793.7 396.19 68065 1996 7407.99 6909.82 71176.6 724.66 68950 1997 8651.14 8234.04 78973 682.3 69820 1998 9875.95 9262.8 84402.3 833.3 70637 1999 11444.08 10682.58 89677.1 925.43 71394 2000 13395.23 12581.51 99214.6 944.98 72085 2001 16386.04 15301.38 109655.2 1218.1 73025 2002 18903.64 17636.45 120332.7 1328.74 73740 2003 21715.25 20017.31 135822.8 1691.93 74432 2004 26396.47 24165.68 159878.3 2148.32 75200 2005 31649.29 28778.54 184937.4 2707.83 75825 2006 38760.2 34804.35 216314.4 3683.85 76400 2007 51321.78 45621.97 265810.3 4457.96 76990 2008 61330.35 54223.79 314045.4 5552.46 77480
2009 68518.3 59521.59 340506.9 7215.72 77995
三、模型建立
1、散点图分析
2、单因素或多变量间关系分析
Y X1 X2 X3 X4
Y 1 0.9989134611
47853
0.9934790452
90804
0.8770144886
79564
0.9836027198
41508
X1 0.9989134611
47853 1
0.9937402677
18469
0.8556377347
44782
0.9849352965
93492
X2 0.9934790452
90804
0.9937402677
18469 1
0.8561835802
28471
0.9862411656
80459
X3 0.8770144886
79564
0.8556377347
44782
0.8561835802
28471 1
0.8109403346
50381
X4 0.9836027198
41508
0.9849352965
93492
0.9862411656
80459
0.8109403346
50381 1
由散点图分析和变量间关系分析可以看出被解释变量财政收入Y与解释变量总税收收入X1、国内生产总值X2、其他收入X3、就业人口总数X4呈线性关系，因此该回归模型设为：
μβββββ+++++=443322110X X X X Y
3、 模型预模拟
由eviews 做ols 回归得到结果：
Dependent Variable: Y Method: Least Squares Date: 11/14/11 Time: 17:51 Sample: 1990 2009 Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob. C 7299.523 1691.814 4.314614 0.0006 X1 1.062802 0.021108 50.34972 0.0000 X2 0.001770 0.004528 0.391007 0.7013 X3 0.873369 0.119806 7.289852 0.0000 X4
-0.115975
0.026580
-4.363160
0.0006
R-squared 0.999978 Mean dependent var 20556.75 Adjusted R-squared 0.999972 S.D. dependent var 19987.03 S.E. of regression 106.6264 Akaike info criterion 12.38886 Sum squared resid 170537.9 Schwarz criterion 12.63779 Log likelihood -118.8886 F-statistic 166897.9 Durbin-Watson stat
1.496517 Prob(F-statistic)
0.000000
4
321115975.0873369.0001770.0062802.1523.7299X X X X Y -+++=
(4.314614) ( 50.34972 ) ( 0.391007) ( 7.289852) ( -4.363160)
999978.02=R 999972.02
=R 9.166897=F 496517.1.=W D
四、 模型检验 1.计量经济学意义检验 ⑴多重共线性检验与解决
求相关系数矩阵，得到：
Correlation Matrix
Y X1 X2 X3 X4 1 0.998913461147853 0.993479045290804
0.877014488679564
0.9836027198
41508
0.9989134611
1
0.99374026770.85563773470.9849352965
47853 18469 44782 93492
0.9934790452
90804 0.9937402677
18469 1
0.8561835802
28471
0.9862411656
80459
0.8770144886
79564 0.8556377347
44782
0.8561835802
28471 1
0.8109403346
50381
0.9836027198
41508 0.9849352965
93492
0.9862411656
80459
0.8109403346
50381 1
发现模型存在多重共线性。

接下来运用逐步回归法对模型进行修正:
①将各个解释变量分别加入模型,进行一元回归:
作Y与X1的回归，结果如下:
Dependent Variable: Y
Method: Least Squares
Date: 11/22/11 Time: 23:02
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C -755.6610 145.2330 -5.203094 0.0001
X1 1.144994 0.005760 198.7931 0.0000
R-squared 0.999545 Mean dependent var 20556.75
Adjusted R-squared 0.999519 S.D. dependent var 19987.03
S.E. of regression 438.1521 Akaike info criterion 15.09765
Sum squared resid 3455590. Schwarz criterion 15.19722
Log likelihood -148.9765 F-statistic 39518.70
Durbin-Watson stat 0.475046 Prob(F-statistic) 0.000000
作Y与X2的回归，结果如下:
Dependent Variable: Y
Method: Least Squares
Date: 11/22/11 Time: 23:06
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C -5222.077 861.2067 -6.063674 0.0000
X2 0.207689 0.005548 37.43267 0.0000
R-squared 0.987317 Mean dependent var 20556.75 Adjusted R-squared 0.986612 S.D. dependent var 19987.03 S.E. of regression 2312.610 Akaike info criterion 18.42478 Sum squared resid 96267005 Schwarz criterion 18.52435 Log likelihood -182.2478 F-statistic 1401.205 Durbin-Watson stat 0.188013 Prob(F-statistic) 0.000000
作Y与X3的回归，结果如下:
Dependent Variable: Y
Method: Least Squares
Date: 11/22/11 Time: 23:08
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C 2607.879 773.9988 3.369358 0.0034
X3 10.03073 0.294311 34.08209 0.0000
R-squared 0.984740 Mean dependent var 20556.75 Adjusted R-squared 0.983893 S.D. dependent var 19987.03 S.E. of regression 2536.645 Akaike info criterion 18.60971 Sum squared resid 1.16E+08 Schwarz criterion 18.70929 Log likelihood -184.0971 F-statistic 1161.589 Durbin-Watson stat 1.194389 Prob(F-statistic) 0.000000
作Y与X4的回归，结果如下:
Dependent Variable: Y
Method: Least Squares
Date: 11/22/11 Time: 23:08
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C -272959.3 37203.65 -7.336894 0.0000
X4 4.097403 0.518467 7.902918 0.0000
R-squared 0.776276 Mean dependent var 20556.75 Adjusted R-squared 0.763846 S.D. dependent var 19987.03 S.E. of regression 9712.824 Akaike info criterion 21.29492 Sum squared resid 1.70E+09 Schwarz criterion 21.39449 Log likelihood -210.9492 F-statistic 62.45611 Durbin-Watson stat 0.157356 Prob(F-statistic) 0.000000
②依据可决系数最大的原则选取X1作为进入回归模型的第一个解释变量,再依次将其余变量分别代入回归得:
作Y与X1、X2的回归，结果如下
Dependent Variable: Y
Method: Least Squares
Date: 11/22/11 Time: 23:09
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C -188.4285 239.0743 -0.788159 0.4415
X1 1.281594 0.049472 25.90568 0.0000
X2 -0.025055 0.009029 -2.774908 0.0130
R-squared 0.999687 Mean dependent var 20556.75
Adjusted R-squared 0.999650 S.D. dependent var 19987.03
S.E. of regression 374.0345 Akaike info criterion 14.82405
Sum squared resid 2378330. Schwarz criterion 14.97341
Log likelihood -145.2405 F-statistic 27118.20
Durbin-Watson stat 0.683510 Prob(F-statistic) 0.000000 作Y与X1、X3的回归，结果如下
Dependent Variable: Y
Method: Least Squares
Date: 11/22/11 Time: 23:10
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C -351.1054 83.15053 -4.222527 0.0006
X1 0.992813 0.018707 53.07196 0.0000
X3 1.356936 0.165109 8.218410 0.0000
R-squared 0.999908 Mean dependent var 20556.75
Adjusted R-squared 0.999898 S.D. dependent var 19987.03
S.E. of regression 202.1735 Akaike info criterion 13.59361
Sum squared resid 694859.9 Schwarz criterion 13.74297
Log likelihood -132.9361 F-statistic 92839.33
Durbin-Watson stat 1.177765 Prob(F-statistic) 0.000000
作Y与X1、X4的回归，结果如下
Dependent Variable: Y
Method: Least Squares
Date: 11/22/11 Time: 23:10
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C 11853.46 1824.522 6.496748 0.0000
X1 1.185886 0.006645 178.4608 0.0000
X4 -0.186645 0.026984 -6.917003 0.0000
R-squared 0.999881 Mean dependent var 20556.75
Adjusted R-squared 0.999867 S.D. dependent var 19987.03
S.E. of regression 230.8464 Akaike info criterion 13.85886
Sum squared resid 905931.0 Schwarz criterion 14.00822
Log likelihood -135.5886 F-statistic 71206.90
Durbin-Watson stat 1.459938 Prob(F-statistic) 0.000000
③在满足经济意义和可决系数的条件下选取X3作为进入模型的第二个解释变量,再次进行回归则:
作Y与X1、X3、X2的回归，结果如下
Dependent Variable: Y
Method: Least Squares
Date: 11/22/11 Time: 23:13
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C -76.04458 100.1724 -0.759137 0.4588
X1 1.085924 0.029801 36.43881 0.0000
X3 1.210853 0.133444 9.073877 0.0000
X2 -0.014073 0.003944 -3.567901 0.0026
R-squared 0.999949 Mean dependent var 20556.75
Adjusted R-squared 0.999939 S.D. dependent var 19987.03
S.E. of regression 155.5183 Akaike info criterion 13.10826
Sum squared resid 386975.0 Schwarz criterion 13.30741
Log likelihood -127.0826 F-statistic 104602.9
Durbin-Watson stat 1.196933 Prob(F-statistic) 0.000000
作Y与X1、X3、X4的回归，结果如下
Dependent Variable: Y
Method: Least Squares
Date: 11/22/11 Time: 23:13
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C 6781.764 1024.745 6.618003 0.0000
X1 1.068642 0.014514 73.62764 0.0000
X3 0.891069 0.107949 8.254551 0.0000
X4 -0.107639 0.015451 -6.966675 0.0000
R-squared 0.999977 Mean dependent var 20556.75
Adjusted R-squared 0.999973 S.D. dependent var 19987.03
S.E. of regression 103.7654 Akaike info criterion 12.29900
Sum squared resid 172276.1 Schwarz criterion 12.49814
Log likelihood -118.9900 F-statistic 234970.9
Durbin-Watson stat 1.451447 Prob(F-statistic) 0.000000
④可见加入其余任何一个变量都会导致系数符号与经济意义不符,故
最终修正后的回归模型为:
Dependent Variable: Y
Method: Least Squares
Date: 11/30/11 Time: 12:18
Sample: 1990 2009
Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob.
C -351.1054 83.15053 -4.222527 0.0006
X1 0.992813 0.018707 53.07196 0.0000
X3 1.356936 0.165109 8.218410 0.0000
R-squared 0.999908 Mean dependent var 20556.75
Adjusted R-squared 0.999898 S.D. dependent var 19987.03
S.E. of regression 202.1735 Akaike info criterion 13.59361
Sum squared resid 694859.9 Schwarz criterion 13.74297
Log likelihood -132.9361 F-statistic 92839.33
Durbin-Watson stat
1.177765 Prob(F-statistic)
0.000000
31356936.1992813.01054.351X X Y ++-=
(-4.222527) ( 53.07196) ( 8.218410)
999908.02=R 999898.02
=R 33.92839=F 177765.1.=W D
⑵异方差检验与修正
① 图示法
ee 与X1的散点图如下：
说明ee 与X1存在单调递增型异方差性。

ee与X3的散点图如下：
说明ee与X3存在单调递增型异方差性。

②G-Q检验
对20组数据剔除掉中间四组剩下的进行分组后，第一组（1990-1997）数据的回归结果：
Dependent Variable: Y
Method: Least Squares
Date: 11/30/11 Time: 12:54
Sample: 1990 1997
Included observations: 8
Variable Coefficient Std. Error t-Statistic Prob.
X1 0.984123 0.016255 60.54320 0.0000
X3 0.851518 0.156688 5.434472 0.0029
C -28.34275 45.36993 -0.624703 0.5596
R-squared 0.999686 Mean dependent var 5179.791 Adjusted R-squared 0.999560 S.D. dependent var 2099.840
S.E. of regression 44.05899 Akaike info criterion 10.68893 Sum squared resid 9705.972 Schwarz criterion 10.71872 Log likelihood -39.75573 F-statistic 7947.575 Durbin-Watson stat
1.663630 Prob(F-statistic)
0.000000
残差平方和RSS1=9705.972 第二组（2002-2009）数据的回归结果：
Dependent Variable: Y Method: Least Squares Date: 11/30/11 Time: 12:55 Sample: 2002 2009 Included observations: 8
Variable Coefficient Std. Error t-Statistic Prob. X1 1.066404 0.027747 38.43321 0.0000 X3 0.847228 0.215114 3.938503 0.0110 C
-1184.159
261.8258
-4.522698
0.0063
R-squared 0.999932 Mean dependent var 39824.41 Adjusted R-squared 0.999905 S.D. dependent var 18639.16 S.E. of regression 182.0047 Akaike info criterion 13.52594 Sum squared resid 165628.5 Schwarz criterion 13.55573 Log likelihood -51.10375 F-statistic 36705.08 Durbin-Watson stat
1.326122 Prob(F-statistic)
0.000000
残差平方和RSS2= 165628.5
所以F= RSS2/RSS1= 165628.5/9705.972=17.0646 在给定α=5%下查得临界值 39.6)4,4(05.0=F ，)4,4(05.0F F >
因此否定两组子样方差相同的假设，从而该总体随机项存在递增异方差性。

③White 方法检验
White Heteroskedasticity Test:
F-statistic 6.142010 Probability 0.003919 Obs*R-squared
12.41812 Probability
0.014498
Test Equation:
Dependent Variable: RESID^2 Method: Least Squares Date: 11/30/11 Time: 13:21 Sample: 1990 2009 Included observations: 20
Variable Coefficient Std. Error t-Statistic Prob. C 24856.50 19211.30 1.293848 0.2153 X1 -20.57327 7.549127 -2.725252 0.0156 X1^2 0.000212 8.04E-05 2.639982 0.0186 X3 237.1813 78.61323 3.017067 0.0087 X3^2
-0.024073
0.006568
-3.665230
0.0023
R-squared 0.620906 Mean dependent var 34743.00 Adjusted R-squared 0.519815 S.D. dependent var 49156.00 S.E. of regression 34062.86 Akaike info criterion 23.92212 Sum squared resid 1.74E+10 Schwarz criterion 24.17105 Log likelihood -234.2212 F-statistic 6.142010 Durbin-Watson stat
1.560937 Prob(F-statistic)
0.003919
12.418120.620906202=⨯=*R n
α=5%下，临界值488.9)4(05.02=χ拒绝同方差性
④
修正
Dependent Variable: Y Method: Least Squares Date: 11/30/11 Time: 14:29 Sample: 1990 2009 Included observations: 20 Weighting series: 1/E1
Variable Coefficient Std. Error t-Statistic Prob. C -314.2074 43.68550 -7.192486 0.0000 X1 0.979758 0.008622 113.6336 0.0000 X3 1.457291
0.065922
22.10629 0.0000
Weighted Statistics
R-squared 0.999999 Mean dependent var 27246.27 Adjusted R-squared 0.999999 S.D. dependent var 74471.17 S.E. of regression
73.91795 Akaike info criterion
11.58127
Sum squared resid 92885.67 Schwarz criterion 11.73063 Log likelihood -112.8127 F-statistic 3138195. Durbin-Watson stat
0.956075 Prob(F-statistic) 0.000000
Unweighted Statistics
R-squared 0.999902 Mean dependent var 20556.75 Adjusted R-squared 0.999891 S.D. dependent var 19987.03 S.E. of regression 209.0283 Sum squared resid 742778.2
Durbin-Watson stat
1.365483
31457291.1979758.02074.314X X Y ++-=
(-7.192486) ( 113.6336) ( 22.10629)
999999.02=R 999999.02
=R 3138195=F 365483.1.=W D
⑶序列相关性检验
①从残差项e2与e2(-1)及e 与时间t 的关系图（如下）看，随机项呈现正序列相关性。

600
400
200
-200
-400
-600
90929496980002040608
E2
②Q统计量检验
由图可以看出，存在一阶序列相关
③回归检验
残差e2与e2（-1）做回归得：
Dependent Variable: E
Method: Least Squares
Date: 12/04/11 Time: 15:21
Sample (adjusted): 1991 2009
Included observations: 19 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 16.81525 45.69611 0.367980 0.7174
E(-1) 0.303570 0.231114 1.313508 0.2065
R-squared 0.092138 Mean dependent var 25.28519 Adjusted R-squared 0.038734 S.D. dependent var 201.1252 S.E. of regression 197.1916 Akaike info criterion 13.50553 Sum squared resid 661036.6 Schwarz criterion 13.60494 Log likelihood -126.3025 F-statistic 1.725303 Durbin-Watson stat 1.776498 Prob(F-statistic) 0.206464
e与e(-1)、e(-2)做回归得：
Dependent Variable: E
Method: Least Squares
Date: 12/04/11 Time: 15:24
Sample (adjusted): 1992 2009
Included observations: 18 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 7.449760 46.20912 0.161218 0.8741
E(-1) 0.419564 0.244475 1.716187 0.1067
E(-2) -0.379894 0.278641 -1.363380 0.1929
R-squared 0.192570 Mean dependent var 16.45940 Adjusted R-squared 0.084912 S.D. dependent var 203.1349 S.E. of regression 194.3193 Akaike info criterion 13.52789 Sum squared resid 566399.7 Schwarz criterion 13.67629 Log likelihood -118.7510 F-statistic 1.788727 Durbin-Watson stat 2.055382 Prob(F-statistic) 0.201043
由上表明不存在序列相关性。

④D.W检验
由异方差检验修正后的结果：
31457291.1979758.02074.314X X Y ++-=
999999.02=R 999999.02
=R 3138195=F 365483.1.=W D
得D.W=1.365483
取α=5%，由于n =20，k =3(包含常数项)，查表得：
dl =1.10， du =1.54
由于dl<DW=1.365483< du ，故: 序列相关性不确定。

⑤拉格朗日检验
Dependent Variable: E Method: Least Squares Date: 12/04/11 Time: 15:05 Sample (adjusted): 1992 2009
Included observations: 18 after adjustments
Variable
Coefficient Std. Error t-Statistic Prob. Y 0.000984 0.002548 0.386217 0.7051 C -14.14792 73.42247 -0.192692 0.8500 E(-1) 0.392009 0.261633 1.498316 0.1563 E(-2)
-0.347730
0.298739
-1.163992
0.2639
R-squared 0.201082 Mean dependent var 16.45940 Adjusted R-squared 0.029885 S.D. dependent var 203.1349 S.E. of regression 200.0765 Akaike info criterion 13.62841 Sum squared resid 560428.6 Schwarz criterion 13.82627 Log likelihood -118.6557 F-statistic 1.174565 Durbin-Watson stat
2.010385 Prob(F-statistic)
0.354679
02164.4201082.020*2=⨯==R n LM
取α=5%，2χ分布的临界值815.7)3(05.02=χ LM < )3(05.02χ
故: 存在序列相关。

⑥修正
为了更好的提高模型的精度，我们用广义差分法对模型进行修正。

首先用杜宾（durbin）两步法估计ρ。

Dependent Variable: Y
Method: Least Squares
Date: 12/04/11 Time: 16:18
Sample (adjusted): 1992 2009
Included observations: 18 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C -36.85790 81.18933 -0.453975 0.6606
Y(-1) 0.730610 0.345304 2.115847 0.0635
Y(-2) 0.358104 0.364519 0.982402 0.3516
X1 1.097355 0.030377 36.12488 0.0000
X1(-1) -0.872470 0.400852 -2.176541 0.0575
X1(-2) -0.355699 0.409249 -0.869149 0.4073
X3 0.755747 0.218272 3.462405 0.0071
X3(-1) -0.272101 0.460341 -0.591086 0.5690
X3(-2) -0.083096 0.402994 -0.206198 0.8412
R-squared 0.999986 Mean dependent var 22502.69
Adjusted R-squared 0.999973 S.D. dependent var 20158.96
S.E. of regression 104.6672 Akaike info criterion 12.44630
Sum squared resid 98597.03 Schwarz criterion 12.89149
Log likelihood -103.0167 F-statistic 78825.65
Durbin-Watson stat 2.219316 Prob(F-statistic) 0.000000
由上表可得回归方程ρ1=0.730610，ρ2=0.358104，对原模型进行广义差分，下表为广义差分结果。

Dependent Variable: Y1
Method: Least Squares
Date: 12/04/11 Time: 16:47
Sample (adjusted): 1992 2009
Included observations: 18 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C -402.0982 84.12065 -4.780018 0.0002
X11 1.041509 0.018988 54.85006 0.0000
X33 1.107351 0.185136 5.981271 0.0000
R-squared 0.999844 Mean dependent var 16547.52
Adjusted R-squared 0.999824 S.D. dependent var 14812.28
S.E. of regression 196.6902 Akaike info criterion 13.55215
Sum squared resid 580305.5 Schwarz criterion 13.70054
Log likelihood -118.9693 F-statistic 48198.10
Durbin-Watson stat 1.385664 Prob(F-statistic) 0.000000
其中Y1=Y-0.730610*Y(-1)+0.358104*Y(-2),X11=X1-0.730610*X1(-1)+0.358104*X1(-2),
X33=x3-0.730610*X3(-1)+0.358104*X3(-2)
D.W=1.385664<du，仍存在序列相关。

下面我们用采用科克伦-奥科特迭代法估计ρ
Dependent Variable: Y
Method: Least Squares
Date: 12/04/11 Time: 15:33
Sample (adjusted): 1991 2009
Included observations: 19 after adjustments
Convergence achieved after 107 iterations
Variable Coefficient Std. Error t-Statistic Prob.
C -21511.24 677371.7 -0.031757 0.9751
X1 1.086097 0.022027 49.30646 0.0000
X3 0.825966 0.128930 6.406292 0.0000
AR(1) 0.995597 0.142149 7.003896 0.0000
R-squared 0.999968 Mean dependent var 21484.10
Adjusted R-squared 0.999962 S.D. dependent var 20087.80
S.E. of regression 123.6723 Akaike info criterion 12.65781
Sum squared resid 229422.7 Schwarz criterion 12.85664
Log likelihood -116.2492 F-statistic 158291.4
Durbin-Watson stat 2.273071 Prob(F-statistic) 0.000000
Inverted AR Roots 1.00
取α=5% ，du=1.54<D.W=2.273071<4-du=2.46
表明:广义差分模型已不存在序列相关性。

同时可决系数，t，F 统计量也均达到理想水平。

五、模型的最终确定
310.8259661.086097-21511.24X X Y ++=
(-0.031757) (49.30646) (6.406292) 0.9999682=R 0.9999622=R 158291.4=F 2.273071.=W D。