悬赏 30 个论坛币 未解决
求助一下:动态面板系统gmm里面被解释变量的滞后阶数如何确定呢?如果采用解释变量的滞后项做工具变量如何确定之后滞后多少阶最好呢?
xtdpdsys lny $z ,lags(1) maxldep(1) endogenous(lnx,lag(0,1)) endogenous(lnx4,lag(0,1)) twostep
estat abond
estat sargan
----------------------------------------------------------------------------
lny | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lny |
L1. | .8130577 .0322903 25.18 0.000 .7497699 .8763456
|
lnx | -.0127754 .0102464 -1.25 0.212 -.0328579 .0073071
lnx4 | -.0085005 .0061721 -1.38 0.168 -.0205975 .0035965
lngdp | .0597195 .0275641 2.17 0.030 .0056948 .1137441
lnx91 | .0636574 .0219003 2.91 0.004 .0207335 .1065813
lnrenkou | .0537148 .020967 2.56 0.010 .0126202 .0948093
_cons | .060456 .0665772 0.91 0.364 -.0700328 .1909448
Warning: gmm two-step standard errors are biased; robust standard
errors are recommended.
Instruments for differenced equation
GMM-type: L(2/2).lny L(2/2).lnx L(2/2).lnx4
Standard: D.lngdp D.lnx91 D.lnrenkou
Instruments for level equation
GMM-type: LD.lny LD.lnx LD.lnx4
Standard: _cons
. esti store m1
. estat abond
Arellano-Bond test for zero autocorrelation in first-differenced errors
+-----------------------+
|Order | z Prob > z|
|------+----------------|
| 1 |-2.2929 0.0219 |
| 2 |-.37139 0.7103 |
+-----------------------+
H0: no autocorrelation
. estat sargan
Sargan test of overidentifying restrictions
H0: overidentifying restrictions are valid
chi2(47) = 23.53452
Prob > chi2 = 0.9983
------------------------------------------------------------------------------
lny | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lny |
L1. | .7845753 .0353794 22.18 0.000 .7152329 .8539177
|
lnx | .0098585 .0120646 0.82 0.414 -.0137876 .0335047
lnx4 | -.0146369 .0039246 -3.73 0.000 -.022329 -.0069449
lngdp | .023568 .023688 0.99 0.320 -.0228597 .0699957
lnx91 | .0962838 .0210245 4.58 0.000 .0550765 .1374911
lnrenkou | .0465209 .0289496 1.61 0.108 -.0102193 .103261
_cons | .1920225 .1223851 1.57 0.117 -.0478478 .4318928
------------------------------------------------------------------------------
Warning: gmm two-step standard errors are biased; robust standard
errors are recommended.
Instruments for differenced equation
GMM-type: L(2/3).lny L(2/2).lnx L(2/2).lnx4
Standard: D.lngdp D.lnx91 D.lnrenkou
Instruments for level equation
GMM-type: LD.lny LD.lnx LD.lnx4
Standard: _cons
. esti store m1
. estat abond
Arellano-Bond test for zero autocorrelation in first-differenced errors
+-----------------------+
|Order | z Prob > z|
|------+----------------|
| 1 |-2.2695 0.0232 |
| 2 |-.37249 0.7095 |
+-----------------------+
H0: no autocorrelation
. estat sargan
Sargan test of overidentifying restrictions
H0: overidentifying restrictions are valid
chi2(54) = 22.57473
Prob > chi2 = 0.9999
求助如何比较确定y的滞后阶数?以及x选择的滞后阶数?