黃河泉 发表于 2018-12-27 10:45 
1. 请看看 xtabond2 中之 collapse 选项。2. AR(2) 没通过,不太好处理!
老师您好,我使用39个国家15年的数据,运用系统GMM方法做出的结果如下,. xtdpdsys y y1 k l fdi ed1,lag(2) maxldep(3) endogenous(y1,lag(0,1)) twostep
note: y1 dropped because of collinearity
System dynamic panel-data estimation Number of obs = 507
Group variable: country Number of groups = 39
Time variable: year
Obs per group: min = 13
avg = 13
max = 13
Number of instruments = 79 Wald chi2(7) = 29330.25
Prob > chi2 = 0.0000
Two-step results
------------------------------------------------------------------------------
y | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
y |
L1. | .2242291 .0056216 39.89 0.000 .213211 .2352472
L2. | -.3355919 .0052522 -63.90 0.000 -.3458861 -.3252977
|
y1 | .0000347 2.62e-06 13.27 0.000 .0000296 .0000399
k | .4162729 .013596 30.62 0.000 .3896253 .4429206
l | -.9014696 .1689993 -5.33 0.000 -1.232702 -.5702372
fdi | .0131003 .0005908 22.17 0.000 .0119424 .0142582
ed1 | 5.213435 1.100529 4.74 0.000 3.056439 7.370432
_cons | -8.653763 .2007631 -43.10 0.000 -9.047252 -8.260275
------------------------------------------------------------------------------
Warning: gmm two-step standard errors are biased; robust standard
errors are recommended.
Instruments for differenced equation
GMM-type: L(2/4).y L(2/2).y1
Standard: D.y1 D.k D.l D.fdi D.ed1
Instruments for level equation
GMM-type: LD.y LD.y1
Standard: _cons
. estat abond
Arellano-Bond test for zero autocorrelation in first-differenced errors
+-----------------------+
|Order | z Prob > z|
|------+----------------|
| 1 |-4.1058 0.0000 |
| 2 |-.46717 0.6404 |
+-----------------------+
H0: no autocorrelation
. estat sargan
Sargan test of overidentifying restrictions
H0: overidentifying restrictions are valid
chi2(71) = 36.88563
Prob > chi2 = 0.9997
请问老师,这样的结果是否可以用?