最近我正在做一个动态面板的估计问题,我发现使用xtabond2和xtdpdsys对同一个方程做系统GMM估计时,估计的结果有很大的差异,以下是我的命令和估计结果:
xtdpdsys h y educ pse,lags(1) endogenous(fert) twostep
System dynamic panel-data estimation Number of obs = 186
Group variable: id Number of groups = 31
Time variable: year
Obs per group: min = 6
avg = 6
max = 6
Number of instruments = 44 Wald chi2(5) = 26668.66
Prob > chi2 = 0.0000
Two-step results
------------------------------------------------------------------------------
h | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
h |
L1. | .8713852 .0246559 35.34 0.000 .8230605 .9197099
|
fert | -3.934089 .6614846 -5.95 0.000 -5.230575 -2.637603
y | .1372989 .0183964 7.46 0.000 .1012426 .1733551
educ | -.005879 .0045482 -1.29 0.196 -.0147932 .0030353
pse | .0170601 .0081368 2.10 0.036 .0011122 .033008
_cons | -.3237235 .0875093 -3.70 0.000 -.4952387 -.1522084
------------------------------------------------------------------------------
Warning: gmm two-step standard errors are biased; robust standard
errors are recommended.
Instruments for differenced equation
GMM-type: L(2/.).h L(2/.).fert
Standard: D.y D.educ D.pse
Instruments for level equation
GMM-type: LD.h LD.fert
Standard: _cons
xtabond2 h l.h y educ fert pse, gmm(h fert,lag(2 .) eq(diff)) gmm(l(h fert),lag(0 0) eq(level)) iv(y educ pse,eq(diff)) nomata twostep
Dynamic panel-data estimation, two-step system GMM
Group variable: id Number of obs = 186
Time variable : year Number of groups = 31
Number of instruments = 44 Obs per group: min = 6
Wald chi2(4) = 497967.84 avg = 6.00
Prob > chi2 = 0.000 max = 6
h Coef. Std. Err. z P>z [95% Conf. Interval]
h
L1. .8436513 .0244426 34.52 0.000 .7957447 .8915579
y .1094211 .0189805 5.76 0.000 .07222 .1466221
educ .0419908 .0042817 9.81 0.000 .0335988 .0503828
fert 1.084262 .7909442 1.37 0.170 -.4659604 2.634484
pse -.0269193 .0087809 -3.07 0.002 -.0441295 -.0097091
_cons -.1194238 .1270722 -0.94 0.347 -.3684808 .1296332
Warning: Uncorrected two-step standard errors are unreliable.
Instruments for first differences equation
Standard
D.(y educ pse)
GMM-type (missing=0, separate instruments for each period unless collapsed)
L(2/.).(h fert)
Instruments for levels equation
Standard
_cons
y educ pse
GMM-type (missing=0, separate instruments for each period unless collapsed)
DL.(h fert)
Arellano-Bond test for AR(1) in first differences: z = -2.59 Pr > z = 0.010
Arellano-Bond test for AR(2) in first differences: z = 1.68 Pr > z = 0.093
Sargan test of overid. restrictions: chi2(38) = 90.83 Prob > chi2 = 0.000
(Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(38) = 28.94 Prob > chi2 = 0.855
(Robust, but weakened by many instruments.)
请问这是什么原因?