发帖
雷达卡
做个了panel模型，可是做出来r2非常小，请教下是不是模型设定有问题？
xtreg y lnlabint lntechint lncapint lnlink lneg lnexp lngj lngat lnws lnmy, fe
Fixed-effects (within) regression Number of obs = 2279
Group variable: code Number of groups = 269
R-sq: within = 0.0099 Obs per group: min = 1
between = 0.0002 avg = 8.5
overall = 0.0008 max = 10
F(10,2000) = 2.00
corr(u_i, Xb) = -0.4701 Prob > F = 0.0295
y Coef. Std. Err. t P>t [95% Conf. Interval]
lnlabint -5.738045 3.770261 -1.52 0.128 -13.1321 1.656005
lntechint 5.756236 2.354177 2.45 0.015 1.13934 10.37313
lncapint 4.539947 2.816781 1.61 0.107 -.9841861 10.06408
lnlink -13.34634 17.81885 -0.75 0.454 -48.2918 21.59912
lneg -.8905182 .8737519 -1.02 0.308 -2.604077 .823041
lnexp -.0863584 1.630519 -0.05 0.958 -3.284051 3.111335
lngj -.7480789 .6542707 -1.14 0.253 -2.031202 .5350446
lngat -2.303603 1.50543 -1.53 0.126 -5.255978 .6487724
lnws .6136329 1.634061 0.38 0.707 -2.591008 3.818274
lnmy -5.680633 1.861163 -3.05 0.002 -9.330654 -2.030612
_cons -14.59944 23.74623 -0.61 0.539 -61.16937 31.97049
sigma_u 12.399334
sigma_e 21.805498
rho .24433855 (fraction of variance due to u_i)
F test that all u_i=0: F(268, 2000) = 1.36 Prob > F = 0.0003
.
. est store fe
.
. xtreg y lnlabint lntechint lncapint lnlink lneg lnexp lngj lngat lnws lnmy, re
Random-effects GLS regression Number of obs = 2279
Group variable: code Number of groups = 269
R-sq: within = 0.0019 Obs per group: min = 1
between = 0.0890 avg = 8.5
overall = 0.0167 max = 10
Random effects u_i ~ Gaussian Wald chi2(10) = 38.55
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000
y Coef. Std. Err. z P>z [95% Conf. Interval]
lnlabint 1.670329 1.467299 1.14 0.255 -1.205524 4.546181
lntechint 4.056968 1.190392 3.41 0.001 1.723843 6.390094
lncapint -1.030132 1.418221 -0.73 0.468 -3.809793 1.749529
lnlink 5.174985 9.010687 0.57 0.566 -12.48564 22.83561
lneg -1.337603 .5225849 -2.56 0.010 -2.36185 -.3133554
lnexp 1.586588 .8494239 1.87 0.062 -.0782525 3.251428
lngj .1880504 .428719 0.44 0.661 -.6522233 1.028324
lngat -.1844872 .7652223 -0.24 0.809 -1.684295 1.315321
lnws 1.370395 .9377967 1.46 0.144 -.4676529 3.208443
lnmy .2334082 1.067996 0.22 0.827 -1.859825 2.326641
_cons -32.82805 11.83598 -2.77 0.006 -56.02613 -9.62996
sigma_u 0
sigma_e 21.805498
rho 0 (fraction of variance due to u_i)
.
. est store re
.
. hausman fe
---- Coefficients ----
(b) (B) (b-B) sqrt(diag(V_b-V_B))
fe re Difference S.E.
lnlabint -5.738045 1.670329 -7.408374 3.473025
lntechint 5.756236 4.056968 1.699268 2.031039
lncapint 4.539947 -1.030132 5.570079 2.433703
lnlink -13.34634 5.174985 -18.52133 15.37267
lneg -.8905182 -1.337603 .4470847 .7002481
lnexp -.0863584 1.586588 -1.672946 1.391787
lngj -.7480789 .1880504 -.9361294 .494237
lngat -2.303603 -.1844872 -2.119116 1.296439
lnws .6136329 1.370395 -.7567619 1.338168
lnmy -5.680633 .2334082 -5.914041 1.524241
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(10) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 34.41
Prob>chi2 = 0.0002
京公网安备 11010802022788号
