连老师:您好!我在运用系统GMM分析全国分省资本形成与金融发展动态关系时,运用一步法时的结果如下:
xtabond2 gdzb L.gdzb L(0/1).( deploa lndp ) , gmm(L.gdzb,lag(2 5)) iv(L(0/1).
( deploa lndp ) )
Favoring space over speed. To switch, type or click on mata: mata set matafavor s
> peed, perm.
Warning: Number of instruments may be large relative to number of observations.
Dynamic panel-data estimation, one-step system GMM
Group variable: code Number of obs = 900
Time variable : year Number of groups = 30
Number of instruments = 139 Obs per group: min = 30
Wald chi2(5) = 26540.87 avg = 30.00
Prob > chi2 = 0.000 max = 30
gdzb Coef. Std. Err. z P>z [95% Conf. Interval]
gdzb
L1. 1.043036 .0078442 132.97 0.000 1.027661 1.05841
deploa
--. .1061822 .0417231 2.54 0.011 .0244064 .1879581
L1. -.1191385 .0418687 -2.85 0.004 -.2011997 -.0370773
lndp
--. .115945 .0451242 2.57 0.010 .0275032 .2043869
L1. -.126873 .0443236 -2.86 0.004 -.2137457 -.0400002
_cons .0012758 .1066421 0.01 0.990 -.2077388 .2102905
Arellano-Bond test for AR(1) in first differences: z = -10.11 Pr > z = 0.000
Arellano-Bond test for AR(2) in first differences: z = -2.08 Pr > z = 0.037
Sargan test of overid. restrictions: chi2(133) = 434.70 Prob > chi2 = 0.000
(Not robust, but not weakened by many instruments.)
Difference-in-Sargan tests of exogeneity of instrument subsets:
GMM instruments for levels
Sargan test excluding group: chi2(105) = 369.84 Prob > chi2 = 0.000
Difference (null H = exogenous): chi2(28) = 64.86 Prob > chi2 = 0.000
ivstyle(L(0/1).( deploa lndp ))
Sargan test excluding group: chi2(129) = 419.95 Prob > chi2 = 0.000
Difference (null H = exogenous): chi2(4) = 14.75 Prob > chi2 = 0.005
.
以上结果中的AR(2)和Sargan 检验均不理想,但我在xtabond2语句后加 twostep,其结果较为理想。
xtabond2 gdzb L.gdzb L(0/1).( deploa lndp ) , gmm(L.gdzb,lag(2 5)) iv(L(0/1).
> ( deploa lndp ) ) twostep
Favoring space over speed. To switch, type or click on mata: mata set matafavor s
> peed, perm.
Warning: Number of instruments may be large relative to number of observations.
Warning: Two-step estimated covariance matrix of moments is singular.
Using a generalized inverse to calculate optimal weighting matrix for two-step
> estimation.
Difference-in-Sargan statistics may be negative.
Dynamic panel-data estimation, two-step system GMM
------------------------------------------------------------------------------
Group variable: code Number of obs = 900
Time variable : year Number of groups = 30
Number of instruments = 139 Obs per group: min = 30
Wald chi2(5) = 137124.14 avg = 30.00
Prob > chi2 = 0.000 max = 30
------------------------------------------------------------------------------
gdzb | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
gdzb |
L1. | 1.038108 .01015 102.28 0.000 1.018214 1.058001
deploa |
--. | .1009442 .0239268 4.22 0.000 .0540485 .1478399
L1. | -.1123348 .0164816 -6.82 0.000 -.1446382 -.0800314
lndp |
--. | .1112644 .0177955 6.25 0.000 .0763858 .1461429
L1. | -.1267078 .0195746 -6.47 0.000 -.1650733 -.0883423
_cons | .0324825 .1346606 0.24 0.809 -.2314475 .2964125
------------------------------------------------------------------------------
Warning: Uncorrected two-step standard errors are unreliable.
Arellano-Bond test for AR(1) in first differences: z = -4.02 Pr > z = 0.000
Arellano-Bond test for AR(2) in first differences: z = -1.58 Pr > z = 0.115
Sargan test of overid. restrictions: chi2(133) = 434.70 Prob > chi2 = 0.000
(Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(133) = 29.29 Prob > chi2 = 1.000
(Robust, but weakened by many instruments.)
Difference-in-Hansen tests of exogeneity of instrument subsets:
GMM instruments for levels
Hansen test excluding group: chi2(105) = 29.29 Prob > chi2 = 1.000
Difference (null H = exogenous): chi2(28) = 0.00 Prob > chi2 = 1.000
ivstyle(L(0/1).( deploa lndp ))
Hansen test excluding group: chi2(129) = 29.79 Prob > chi2 = 1.000
Difference (null H = exogenous): chi2(4) = -0.51 Prob > chi2 = 1.000
以上一步法结果和两步伐结果出现明显差别,我究竟以哪一个结果为准?理由如何解释?特请教连老师!!
多谢!!