. xtabond TobinQ_A_1 INDEPENDENCE_1 BOARD_NUM_LN_1 DUAL_1 ControlProportion_1 SHAREHOLDERS_LN_1 WOMEN_CEO EMPLOYEE_LN_1 LEVE
RAGE_1 AGE_LN_1,lag(1) maxldep(2) endogenous(_WOMEN_1,lag(0,2)) twostep vce(r)
Arellano-Bond dynamic panel-data estimation Number of obs = 6,453
Group variable: id Number of groups = 1,326
Time variable: Year
Obs per group:
min = 1
avg = 4.866516
max = 12
Number of instruments = 56 Wald chi2(11) = 56.45
Prob > chi2 = 0.0000
Two-step results
(Std. Err. adjusted for clustering on id)
-------------------------------------------------------------------------------------
| WC-Robust
TobinQ_A_1 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------------+----------------------------------------------------------------
TobinQ_A_1 |
L1. | -.1837876 .0419806 -4.38 0.000 -.2660682 -.1015071
|
_WOMEN_1 | -12.12356 2.929896 -4.14 0.000 -17.86605 -6.381067
INDEPENDENCE_1 | -.4389985 1.443444 -0.30 0.761 -3.268096 2.390099
BOARD_NUM_LN_1 | -1.541146 .5921494 -2.60 0.009 -2.701738 -.3805548
DUAL_1 | .0830733 .1532713 0.54 0.588 -.217333 .3834796
ControlProportion_1 | -.0154079 .0065744 -2.34 0.019 -.0282936 -.0025223
SHAREHOLDERS_LN_1 | .4202823 .0984962 4.27 0.000 .2272334 .6133313
WOMEN_CEO | .3107486 .3367722 0.92 0.356 -.3493127 .97081
EMPLOYEE_LN_1 | -.3543356 .1707017 -2.08 0.038 -.6889049 -.0197664
LEVERAGE_1 | -.6177492 .5930614 -1.04 0.298 -1.780128 .5446298
AGE_LN_1 | .2057639 .1075345 1.91 0.056 -.0049999 .4165277
_cons | 6.76975 2.354841 2.87 0.004 2.154348 11.38515
-------------------------------------------------------------------------------------
Instruments for differenced equation
GMM-type: L(2/3).TobinQ_A_1 L(2/3)._WOMEN_1
Standard: D.INDEPENDENCE_1 D.BOARD_NUM_LN_1 D.DUAL_1 D.ControlProportion_1
D.SHAREHOLDERS_LN_1 D.WOMEN_CEO D.EMPLOYEE_LN_1 D.LEVERAGE_1
D.AGE_LN_1
Instruments for level equation
Standard: _cons
.
end of do-file
. do "E:\STATA14\STD04000000.tmp"
. estimates store DiffGMM6
.
end of do-file
. do "E:\STATA14\STD04000000.tmp"
. estat abond
Arellano-Bond test for zero autocorrelation in first-differenced errors
+-----------------------+
|Order | z Prob > z|
|------+----------------|
| 1 |-5.6077 0.0000 |
| 2 |-5.2578 0.0000 |
+-----------------------+
H0: no autocorrelation
我参考陈强老师高级计量经济及Stata应用的差分GMM的做法,发现二阶项存在自相关,是不是意味不能使用差分GMM