问题1、我用的是非均衡的面板数据,通过Hausman检验的结果显示:
chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 17.38
Prob>chi2 = 0.0264
(V_b-V_B is not positive definite),说明应该选用固定效应模型还是随机效应模型?
问题2、固定效应模型如下:命令:xtreg y x1 x2 x3 x4 x5 x6 x7 x8,fe
结果:
Fixed-effects (within) regression Number of obs = 106
Group variable: company Number of groups = 16
R-sq: within = 0.5369 Obs per group: min = 3
between = 0.0766 avg = 6.6
overall = 0.2635 max = 10
F(8,82) = 11.88
corr(u_i, Xb) = -0.5800 Prob > F = 0.0000
------------------------------------------------------------------------------
y | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
x1| .0001679 .0002239 0.75 0.456 -.0002776 .0006134
x2 | .003266 .0044786 0.73 0.468 -.0056434 .0121755
x3 | .0001283 .0000663 1.94 0.056 -3.59e-06 .0002601
x4 | .0021561 .0004661 4.63 0.000 .0012288 .0030833
x5 | -.0059285 .0048169 -1.23 0.222 -.0155109 .0036539
x6 | .0229708 .007865 2.92 0.005 .0073249 .0386167
x7 | .0016107 .0028482 0.57 0.573 -.0040552 .0072766
x8 | -.0012728 .0014106 -0.90 0.370 -.0040789 .0015333
_cons | -.0159413 .0059565 -2.68 0.009 -.0277907 -.0040919
-------------+----------------------------------------------------------------
sigma_u | .00365279
sigma_e | .00183084
rho | .79922075 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(15, 82) = 5.35 Prob > F = 0.0000
结果显示的
我在做随机效应模型时也遇到了同样的问题,求前辈们指点!!



雷达卡





京公网安备 11010802022788号







