我做的回归很简单,ind=a+rat+ind(-1)+b
xtabond2 ind ind0 rat,gmm(ind0,collapse)iv(rat) noleveleq small
结果如下
Dynamic panel-data estimation, one-step difference GMM
------------------------------------------------------------------------------
Group variable: industry Number of obs = 14022
Time variable : t Number of groups = 114
Number of instruments = 124 Obs per group: min = 123
F(2, 14020) = 4.28 avg = 123.00
Prob > F = 0.014 max = 123
------------------------------------------------------------------------------
ind | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
ind0 | .0001847 .0009783 0.19 0.850 -.0017329 .0021024
rat | .0030668 .0010548 2.91 0.004 .0009992 .0051344
------------------------------------------------------------------------------
Instruments for first differences equation
Standard
D.rat
GMM-type (missing=0, separate instruments for each period unless collapsed)
L(1/123).ind0 collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z = -51.51 Pr > z = 0.000
Arellano-Bond test for AR(2) in first differences: z = 4.95 Pr > z = 0.000
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(122) = 146.38 Prob > chi2 = 0.066
(Not robust, but not weakened by many instruments.)
Difference-in-Sargan tests of exogeneity of instrument subsets:
iv(rat)
Sargan test excluding group: chi2(121) = 111.81 Prob > chi2 = 0.713
Difference (null H = exogenous): chi2(1) = 34.57 Prob > chi2 = 0.000