以下是我跑的结果,初步判断无二阶自相关,工具变量有效。变量先对数化,然后差分。本人初学GMM,很多不懂,请各位大大看看的我这个还有什么问题?有什么好的建议?先谢谢各位啦~
. xtdpdsys dlgdp dlrc dluc dlgc dlci,lags(1) maxldep(2) twostep
System dynamic panel-data estimation Number of obs = 154
Group variable: city Number of groups = 14
Time variable: year
Obs per group: min = 11
avg = 11
max = 11
Number of instruments = 34 Wald chi2(5) = 1035.49
Prob > chi2 = 0.0000
Two-step results
------------------------------------------------------------------------------
dlgdp | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
dlgdp |
L1. | .1533626 .058521 2.62 0.009 .0386635 .2680616
|
dlrc | .0421651 .0056054 7.52 0.000 .0311786 .0531515
dluc | .0266207 .0052712 5.05 0.000 .0162893 .036952
dlgc | .0541615 .0181326 2.99 0.003 .0186222 .0897007
dlci | .0681643 .0156701 4.35 0.000 .0374514 .0988771
_cons | .0750033 .0065036 11.53 0.000 .0622564 .0877502
------------------------------------------------------------------------------
Warning: gmm two-step standard errors are biased; robust standard
errors are recommended.
Instruments for differenced equation
GMM-type: L(2/3).dlgdp
Standard: D.dlrc D.dluc D.dlgc D.dlci
Instruments for level equation
GMM-type: LD.dlgdp
Standard: _cons
. estat abond
Arellano-Bond test for zero autocorrelation in first-differenced errors
+-----------------------+
|Order | z Prob > z|
|------+----------------|
| 1 |-2.0754 0.0379 |
| 2 | .63449 0.5258 |
+-----------------------+
H0: no autocorrelation
. estat sargan
Sargan test of overidentifying restrictions
H0: overidentifying restrictions are valid
chi2(28) = 12.04143
Prob > chi2 = 0.9963