谁能帮我看看分别用xtabond2 和xtdpd 哪个结果好些吗?怎么分析呢?
1、xtabond2命令的结果:
xtabond2 lev l.lev size pro tang ndts liqu pca inop baoliu yingyun,gmm(lev size pro tang ndts liqu pca inop baoliu yingyun,lag(2 2)) nolevel smal l
Favoring space over speed. To switch, type or click on mata: mata set matafavor speed, perm.
Dynamic panel-data estimation, one-step difference GMM
------------------------------------------------------------------------------
Group variable: code_dum Number of obs = 255
Time variable : year Number of groups = 51
Number of instruments = 50 Obs per group: min = 5
F(10, 245) = 19.67 avg = 5.00
Prob > F = 0.000 max = 5
------------------------------------------------------------------------------
lev | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lev |
L1. | .0397423 .0789673 0.50 0.615 -.1157991 .1952836
size | .1075706 .021561 4.99 0.000 .0651019 .1500392
pro | -.0159616 .0170661 -0.94 0.351 -.0495766 .0176533
tang | -.1319296 .0572375 -2.30 0.022 -.2446699 -.0191894
ndts | .3441457 1.034333 0.33 0.740 -1.693174 2.381466
liqu | -.0973343 .0179064 -5.44 0.000 -.1326045 -.0620641
pca | -4.587896 1.060056 -4.33 0.000 -6.675883 -2.499909
inop | .0007705 .0005055 1.52 0.129 -.0002252 .0017663
baoliu | -.7476521 .1512282 -4.94 0.000 -1.045525 -.4497788
yingyun | -.0018036 .0006419 -2.81 0.005 -.003068 -.0005391
------------------------------------------------------------------------------
Instruments for first differences equation
GMM-type (missing=0, separate instruments for each period unless collapsed)
L2.(lev size pro tang ndts liqu pca inop baoliu yingyun)
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z = -3.38 Pr > z = 0.001
Arellano-Bond test for AR(2) in first differences: z = 0.20 Pr > z = 0.838
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(40) = 45.90 Prob > chi2 = 0.241
(Not robust, but not weakened by many instruments.)
2、xtdpd命令的结果
xtdpd lev l.lev size pro tang ndts liqu pca inop baoliu yingyun, twostep dgmmiv(lev size pro tang ndts liqu pca inop baoliu yingyun, lagrange(2))
> lgmmiv(lev size pro tang ndts liqu pca inop baoliu yingyun, lag(2)) artests(2)
Dynamic panel-data estimation Number of obs = 306
Group variable: code_dum Number of groups = 51
Time variable: year
Obs per group: min = 6
avg = 6
max = 6
Number of instruments = 191 Wald chi2(10) = 156381.22
Prob > chi2 = 0.0000
Two-step results
------------------------------------------------------------------------------
lev | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lev |
L1. | .3898956 .014208 27.44 0.000 .3620485 .4177428
size | .0638932 .0028806 22.18 0.000 .0582474 .069539
pro | -.0162189 .0025059 -6.47 0.000 -.0211304 -.0113073
tang | .0841234 .0082696 10.17 0.000 .0679153 .1003315
ndts | -2.659479 .13495 -19.71 0.000 -2.923976 -2.394982
liqu | -.063842 .0033097 -19.29 0.000 -.0703288 -.0573552
pca | -.4799973 .1392972 -3.45 0.001 -.7530148 -.2069798
inop | .0015522 .0000797 19.48 0.000 .001396 .0017083
baoliu | -.2815143 .0191556 -14.70 0.000 -.3190586 -.24397
yingyun | -.0005303 .0000886 -5.99 0.000 -.0007039 -.0003567
_cons | -.897393 .0527354 -17.02 0.000 -1.000753 -.7940334
------------------------------------------------------------------------------
Warning: gmm two-step standard errors are biased; robust standard
errors are recommended.
Instruments for differenced equation
GMM-type: L(2/.).lev L(2/.).size L(2/.).pro L(2/.).tang L(2/.).ndts L(2/.).liqu L(2/.).pca L(2/.).inop L(2/.).baoliu L(2/.).yingyun
Instruments for level equation
GMM-type: L2D.lev L2D.size L2D.pro L2D.tang L2D.ndts L2D.liqu L2D.pca L2D.inop L2D.baoliu L2D.yingyun
Standard: _cons
. estat sargan
Sargan test of overidentifying restrictions
H0: overidentifying restrictions are valid
chi2(180) = 40.73323
Prob > chi2 = 1.0000
. estat abond
Arellano-Bond test for zero autocorrelation in first-differenced errors
+-----------------------+
|Order | z Prob > z|
|------+----------------|
| 1 |-4.0164 0.0001 |
| 2 | 1.2848 0.1989 |
+-----------------------+
H0: no autocorrelation