自相关问检验的Eviews的操作方法.doc

计
量
经
济
学
实
验
报
告

实验目的：掌握自相关问题的检验以及相关的Eviews的操作方法。实验内容：消费总量的多少主要有GDP决定。为了考察GDP对消费总额的影响，可使用如下模型：Y=；其中，X表示GDP，Y表示消费总量。下表列出了中国1990-2000的GDP的X与消费总额Y的统计数据。
年份
GDP（X）
消费总额（Y）
年份
GDP（X）
消费总额（Y）
1990
18319.5
11365.2
1998
79003.3
46405.9
1991
21280.4
13145.9
1999
82673.2
49722.8
1992
25863.7
15952.1
2000
89112.5
54617.2
1993
34500.7
20182.1
2001
98592.9
58927.4
1994
46690.7
26796
2002
107897.6
62798.5
1995
58510.5
33635
2003
121730.3
67493.5
1996
68330.4
40003.9
2004
142394.2
75439.7
1997
74894.2
43579.4

一、估计回归方程

OLS法的估计结果如下：
Y=2329.401+0.546950X
（1.954322）（36.71110）
R=0.990446，=0.989711，SE=2091.475，D.W.=0.478071。

二、进行序列相关性检验
（1）图示检验法

（2）回归检验法
一阶回归检验

二阶回归检验

=1.144406e－0.343796e+ε
3）拉格朗日乘数（LM）检验法
Breusch-Godfrey Serial Correlation LM Test:
F-statistic
29.41781
Probability
0.000038
Obs*R-squared
12.63731
Probability
0.001802

Test Equation:
Dependent Variable: RESID
Method: Least Squares
Date: 12/17/12 Time: 21:51
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
37.31393
644.3315
0.057911
0.9549
X
-0.002008
0.009377
-0.214144
0.8344
RESID(-1)
1.744086
0.234326
7.442998
0.0000
RESID(-2)
-1.088243
0.315853
-3.445408
0.0055
R-squared
0.842487
Mean dependent var
4.37E-12
Adjusted R-squared
0.799529
S.D. dependent var
2015.396
S.E. of regression
902.3726
Akaike info criterion
16.67111
Sum squared resid
8957040.
Schwarz criterion
16.85992
Log likelihood
-121.0333
F-statistic
19.61188
Durbin-Watson stat
2.360720
Prob(F-statistic)
0.000101
C=37.31393 x=-0.002008 RESID(-1)=1.744086 RESID(-2)= -1.088243
三、序列相关的补救
Dependent Variable: DY
Method: Least Squares
Date: 12/17/12 Time: 22:07
Sample(adjusted): 1991 2004
Included observations: 14 after adjusting endpoints
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
2369.885
789.9844
2.999914
0.0111
DX
0.465880
0.029328
15.88520
0.0000
R-squared
0.954604
Mean dependent var
13875.68
Adjusted R-squared
0.950821
S.D. dependent var
5320.847
S.E. of regression
1179.971
Akaike info criterion
17.11593
Sum squared resid
16707973
Schwarz criterion
17.20722
Log likelihood
-117.8115
F-statistic
252.3397
Durbin-Watson stat
0.521473
Prob(F-statistic)
0.000000

（2）科克伦-奥科特法估计模型
Dependent Variable: Y
Method: Least Squares
Date: 12/17/12 Time: 22:09
Sample(adjusted): 1991 2004
Included observations: 14 after adjusting endpoints
Convergence achieved after 16 iterations
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
55169.41
54542.80
1.011488
0.3335
X
0.345292
0.057754
5.978675
0.0001
AR(1)
0.961253
0.042004
22.88491
0.0000
R-squared
0.998047
Mean dependent var
43478.53
Adjusted R-squared
0.997691
S.D. dependent var
19591.16
S.E. of regression
941.3171
Akaike info criterion
16.71985
Sum squared resid
9746856.
Schwarz criterion
16.85679
Log likelihood
-114.0389
F-statistic
2810.040
Durbin-Watson stat
0.941831
Prob(F-statistic)
0.000000
Inverted AR Roots
.96