hausman fe re
---- Coefficients ----
| (b) (B) (b-B) sqrt(diag(V_b-V_B))
| fe re Difference S.E.
-------------+----------------------------------------------------------------
lnlew | .013798 -.0532699 .0670679 .
lnmew | -.014493 -.0196906 .0051976 .
lnhew | .0038009 -.0281014 .0319023 .
lnland | .00183 .01711 -.01528 .0072026
lndcs | .0285488 -.0210769 .0496257 .0124952
lnfcs | -.0197679 -.0074297 -.0123383 .0049028
lnt13 | .025137 .0124746 .0126624 .
lnt14 | -.0060789 .0141996 -.0202785 .
lnt15 | -.0075596 -.0283399 .0207803 .
lnt16 | .0084235 -.0130725 .021496 .
lnt17 | .0001869 .0392335 -.0390466 .
lnt18 | .0051126 -.0179904 .0231031 .
lnt19 | -.0018246 .0033119 -.0051365 .
lnt20 | -.0007303 -.0074581 .0067279 .
lnt21 | .0048407 .0057937 -.000953 .
lnt22 | -.000359 -.0298752 .0295162 .
lnt23 | -.0007569 -.0031509 .002394 .
lnt24 | .0024903 -.0137755 .0162658 .
lnt25 | .0001254 .0103887 -.0102633 .
lnt26 | -.0144952 .0131174 -.0276126 .
lnt27 | -.0002999 -.0022209 .001921 .
lnt28 | -.00426 -.0050844 .0008243 .
lnt29 | .0041553 -.0050577 .009213 .
lnt30 | .0068889 .0387275 -.0318386 .
lnt31 | -.0074758 -.0172259 .0097501 .
lnt32 | -.0077801 -.0064641 -.0013159 .
lnt33 | -.0025043 -.0085045 .0060002 .
lnt34 | -.0024529 .0023026 -.0047555 .
lnt35 | .0138619 .0055335 .0083284 .
lnt36 | .0072613 .0241174 -.016856 .
lnt37 | .002092 .0040526 -.0019606 .
lnt39 | -.0025413 -.0004549 -.0020864 .
lnt40 | -.000218 .0156884 -.0159064 .
lnt41 | .0064464 .0060964 .0003499 .
iyear1 | .0386065 -.003337 .0419434 .0114416
iyear2 | .0382804 -.0031285 .0414089 .0090285
iyear3 | .0319524 -.0076588 .0396112 .0074283
iyear4 | .0210788 -.0137816 .0348604 .0057736
iyear5 | .0152002 -.0148647 .0300648 .0037779
iyear6 | .0133038 -.0109972 .0243009 .
iyear7 | .0104632 -.0084403 .0189035 .
iyear8 | .0071372 .0008829 .0062543 .
------------------------------------------------------------------------------
b = consistent under Ho and Ha; obtained from xtreg
B = inconsistent under Ha, efficient under Ho; obtained from xtreg
Test: Ho: difference in coefficients not systematic
chi2(42) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 140.58
Prob>chi2 = 0.0000
(V_b-V_B is not positive definite)
结果表明应该选择随机模型。
xttest0
Breusch and Pagan Lagrangian multiplier test for random effects:
outputshare[pvcode,t] = Xb + u[pvcode] + e[pvcode,t]
Estimated results:
| Var sd = sqrt(Var)
---------+-----------------------------
outputs~e | .0013585 .0368576
e | .0000645 .0080288
u | 0 0
Test: Var(u) = 0
chi2(1) = 66.45
Prob > chi2 = 0.0000
Hausman检验又应该选择FE模型。
到底应该怎么选?FE和RE都做了,感觉用RE做显著的变量多些。