crystal8832 发表于 2015-3-9 14:01 
面板数据一般不需要考虑多重共线问题
哦哦,谢谢了~能再问下Hausman检验的时候,怎么看是否存在内生性问题呀?我用如下命令:
quietly xtpoisson noentrant density frac_close frac_sameindus frac_syndicate excess_exit_rate coeff_monthly_nodeals pe_ratio gdp_growthrate lg_inflow lg_gdp,fe
estimates store fe
quietly xtpoisson noentrant density frac_close frac_sameindus frac_syndicate excess_exit_rate coeff_monthly_nodeals pe_ratio gdp_growthrate lg_inflow lg_gdp,re
estimates store re
hausman fe re
得到的结果如下:
---- Coefficients ----
| (b) (B) (b-B) sqrt(diag(V_b-V_B))
| fe re Difference S.E.
-------------+----------------------------------------------------------------
density | -3.531657 -6.008735 2.477077 .6990346
frac_close | -20.94529 -1.272254 -19.67304 3.896646
frac_samei~s | 3.084396 .0266254 3.057771 .5326415
frac_syndi~e | .4146716 -.3505008 .7651724 .1962509
excess_exi~e | .211647 .5838781 -.3722311 .174546
coeff_mont~s | -.2430681 -.3151531 .072085 .016465
pe_ratio | -.0021849 -.0021707 -.0000143 .0001507
gdp_growth~e | .0554463 .0157283 .039718 .0131107
lg_inflow | .0371133 .0998151 -.0627018 .
lg_gdp | .1709649 .1768322 -.0058674 .0062433
------------------------------------------------------------------------------
b = consistent under Ho and Ha; obtained from xtpoisson
B = inconsistent under Ha, efficient under Ho; obtained from xtpoisson
Test: Ho: difference in coefficients not systematic
chi2(10) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 13.96
Prob>chi2 = 0.1747
(V_b-V_B is not positive definite)
我只知道这个是接受原假设,所以要用re,但是怎么看内生性呢?