计量经济软件Eviews上机指导及演示示例
第一阶段：LS CS C G CS(-1) 估计消费的简化式方程 GENR ECS=CS-RESID 计算消费的估计值
LS Y C G CS(-1) 估计收入的简化式方程
GENR EY=Y-RESID 计算收入的估计值
第二阶段：LS CS C EY CS(-1) 估计替代后的消费结构式方程 LS I C EY 估计替代后的投资结构式方程
Dependent Variable: CS Method: Least Squares Sample(adjusted): 1979 2003
Included observations: 25 after adjusting endpoints
Variable CoefficStd. t-Prob. ient Error Statistic C 273.399188.4341.450900.1609 0 5 9 EY 0.26518
0.18542
1.43013
0.166
4 6 1 7 CS(-1)
0.43372
0.45600
0.95115
0.351
9
4
1
9 R-squared 0.99805 Mean dependent 6526.8 var 600 Adjusted R-0.99788 S.D. dependent
4023.squared 1 var 411 S.E. of 185.186 Akaike info
13.39regression 9 criterion 277 Sum squared 754472. Schwarz
13.53resid
1 criterion 904 Log likelihood - F-statistic
5653.164.4097
343 Durbin-Watson 1.45643 Prob(F-0.000stat
9 statistic)
000
Dependent Variable: I Method: Least Squares Sample(adjusted): 1979 2003
Included observations: 25 after adjusting endpoints
Variable CoefficStd. t-Prob. ient
Error Statistic C -219.834-0.25126
计量经济软件Eviews上机指导及演示示例
258.6251
EY
0.40176
5
2
0.01338
9
1.176455 30.0070
4
4 0.000
0 5279.634 3703.951 15.69877 15.79628 900.4222 0.000000
R-squared Adjusted R-squared S.E. of regression Sum squared resid
Log likelihood Durbin-Watson stat
0.97509 Mean dependent 3 var 0.97401 S.D. dependent
0 var 597.132 Akaike info
5 criterion 8201047 Schwarz
. criterion -194.2347
F-statistic
0.75140 Prob(F-7 statistic)
方法二：
实际上在Eviews软件中，可以利用命令直接进行二阶段最小二乘估计，命令格式为：
TSLS Yi C 解释变量名 @ C 先决变量名
其中符号@前面是该结构式方程的所有解释变量名，包括内生变量和先决变量；符号@后面是联立方程模型中的所有前定变量。
因此本例可用TSLS命令直接写成： TSLS CS C Y CS(-1)@ C G CS(-1) TSLS I C Y @ C G CS(-1)
Dependent Variable: CS
Method: Two-Stage Least Squares Sample(adjusted): 1979 2003
Included observations: 25 after adjusting endpoints Instrument list: C G CS(-1)
Variable Coefficient 273.3999 0.26518
4 0.43372
9
Std. t-Error Statistic 177.3501 0.17451
9 0.42918
1
1.541583 1.51951
0 1.01059
5
Prob. C Y CS(-1)
0.1374 0.142
9 0.323
2 6526.600
27
R-squared 0.99828 Mean dependent 0 var
计量经济软件Eviews上机指导及演示示例
Adjusted R-0.99812 S.D. dependent
4023.squared 3 var 411 S.E. of 174.294 Sum squared
66832regression
0 resid 4.6 F-statistic 6382.06 Durbin-Watson
1.0033 stat 405
Prob(F-0.00000
statistic)
0
Dependent Variable: I
Method: Two-Stage Least Squares Sample(adjusted): 1979 2003
Included observations: 25 after adjusting endpoints Instrument list: C G CS(-1)
Variable CoefficStd. t-Prob. ient Error Statistic C -131.656-0.061258.6251 4 1.964394 7 Y
0.40176
0.00801
50.1044
0.000
5
9
4
0 R-squared 0.99106 Mean dependent 5279.7 var 634 Adjusted R-0.99067 S.D. dependent
3703.squared 8 var 951 S.E. of 357.616 Sum squared
29414regression
6 resid 61. F-statistic 2510.45 Durbin-Watson
0.8145 stat 465
Prob(F-0.00000
statistic)
0
方法三：
还可以在方程说明窗口中，选择估计方法为TLSL，并在工具变量兰（Instrument List）输入模型中的所有先决变量。
方法四：
借助于Eviews中的System命令，可以直接进行TSLS估计。
28
计量经济软件Eviews上机指导及演示示例
（1）创建系统：在主菜单上单击Objects → New Object，并在弹出的对象列表框中选择System；然后在打开的系统窗口输入结构式模型的随机方程
CS=C(1)+C(2)*Y+C(3)*CS(-1) I=C(4)+C(5)*Y INST G CS(-1)
（2）估计模型：在系统窗口单击Estimate，在弹出估计方法选择窗口中选择TSLS方法后，单击OK。
System: UNTITLED
Estimation Method: Two-Stage Least Squares Sample: 1979 2003
Included observations: 25
Total system (balanced) observations 50 Instruments: G CS(-1) C
CoefficStd. t-Prob. ient Error Statistic C(1) 273.399177.3501.541580.1309 1 3 2 C(2) 0.26518
0.17451
1.51951
0.135
4 9 0 6 C(3) 0.43372
0.42918
1.01059
0.317
9 1 5 6 C(4) -131.656
-0.055
258.6251 4 1.964394 7 C(5)
0.40176
0.00801
50.1044
0.000
5
9 4
0
Determinant residual 1.35E+0
covariance
9
Equation: CS=C(1)+C(2)*Y+C(3)*CS(-1) Observations: 25 R-squared 0.99828 Mean dependent
6526.0 var 600 Adjusted R-0.99812 S.D. dependent
4023.squared 3 var 411 S.E. of 174.294 Sum squared
66832regression 0 resid 4.6
Durbin-Watson 1.00340
stat
5
Equation: I=C(4)+C(5)*Y 29
计量经济软件Eviews上机指导及演示示例
Observations: 25 R-squared Adjusted R-squared S.E. of regression Durbin-Watson stat
0.99106 Mean dependent
7 var 0.99067 S.D. dependent
8 var 357.616 Sum squared
6 resid 0.81446
5
5279.634 3703.951 2941461.
第四部分 演示示例
EVIEWS在计量经济学教学过程中的演示示例（一）
练习一（习题集P17第8题）
一、按题意在EVIEWS中输入数据；
二、在主窗口QUICK菜单下选择estimate equation，在弹出对话框中输入Y C X，进行最
小二乘估计参数。
三、在equantion窗口viiew菜单下选择representations选项，可得回归方程。
Y = 49.82200092 + 0.7944335972*X
四、选择equantion窗口viiew菜单下estimation output或stats按钮，可得回归结果输出：
Dependent Variable: Y Method: Least Squares
Date: 10/13/05 Time: 19:47 Sample: 1 30
Included observations: 30 Variable
C X R-squared
Adjusted R-squared S.E. of regression Sum squared resid Log likelihood
Durbin-Watson stat CoefficienStd. Error t-Statistic
t 49.82200 0.794434 143.7013 0.025719 0.346705 30.88857 Prob. 0.7314 0.0000 0.971490 Mean dependent var 4342.317 0.970472 S.D. dependent var 1166.015 200.3663 Akaike info 13.50251
criterion
1124106. Schwarz criterion 13.59592 -200.5377 F-statistic 954.1039 1.607925 Prob(F-statistic) 0.000000 五、标准报告形式 Y = 49.82200092 + 0.7944335972*X
（0.346705） （30.88857）
R-squared＝0.971490 S.E. of regression＝200.3663 F-statistic＝954.103925503
六、点预测
Y0 = 49.82200092 + 0.7944335972*1000=844.222(元)
30