面板数据为38家银行11年的数据,我在进行回归时发现豪斯曼检验结果为零,显示我用随机效应和固定效应回归的系数一样,但是我单独回归时是不一样的,用了连玉君老师说的修正豪斯曼检验结果还是这样,这是什么原因呢?代码结果如下:
xtreg RWARATIO FINTECH LOGGDP LOGM2 ROA DPR CIR CAR LOGASSET,re
Random-effects GLS regression Number of obs = 418
Group variable: banks Number of groups = 38
R-sq: Obs per group:
within = 0.3606 min = 11
between = 0.3170 avg = 11.0
overall = 0.3357 max = 11
Wald chi2(8) = 225.62
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000
------------------------------------------------------------------------------
RWARATIO | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
FINTECH | 4.631615 1.044965 4.43 0.000 2.583521 6.679709
LOGGDP | 22.69639 26.53388 0.86 0.392 -29.30906 74.70183
LOGM2 | -10.86699 23.39898 -0.46 0.642 -56.72815 34.99417
ROA | 6.008272 1.599615 3.76 0.000 2.873085 9.143459
DPR | .2632869 .0302446 8.71 0.000 .2040086 .3225651
CIR | -.0222755 .0810162 -0.27 0.783 -.1810644 .1365134
CAR | -.4948777 .1222759 -4.05 0.000 -.734534 -.2552214
LOGASSET | -2.609663 .8681066 -3.01 0.003 -4.31112 -.908205
_cons | -2.539543 25.25365 -0.10 0.920 -52.03579 46.95671
-------------+----------------------------------------------------------------
sigma_u | 3.7115497
sigma_e | 4.2682227
rho | .43057733 (fraction of variance due to u_i)
------------------------------------------------------------------------------
. esti store re
. xtreg RWARATIO FINTECH LOGGDP LOGM2 ROA DPR CIR CAR LOGASSET,fe
Fixed-effects (within) regression Number of obs = 418
Group variable: banks Number of groups = 38
R-sq: Obs per group:
within = 0.3711 min = 11
between = 0.0332 avg = 11.0
overall = 0.0868 max = 11
F(8,372) = 27.43
corr(u_i, Xb) = -0.6999 Prob > F = 0.0000
------------------------------------------------------------------------------
RWARATIO | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
FINTECH | 4.33474 1.036846 4.18 0.000 2.295927 6.373553
LOGGDP | 32.07893 26.26679 1.22 0.223 -19.57107 83.72893
LOGM2 | -8.9939 23.52418 -0.38 0.702 -55.25095 37.26315
ROA | 4.206165 1.692855 2.48 0.013 .8773994 7.534931
DPR | .2252608 .033857 6.65 0.000 .1586857 .2918358
CIR | -.16203 .0904353 -1.79 0.074 -.3398584 .0157985
CAR | -.6533295 .1316505 -4.96 0.000 -.912202 -.3944571
LOGASSET | -10.73029 3.636778 -2.95 0.003 -17.88151 -3.579067
_cons | 39.03962 29.52025 1.32 0.187 -19.00785 97.0871
-------------+----------------------------------------------------------------
sigma_u | 7.8473731
sigma_e | 4.2682227
rho | .77170482 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(37, 372) = 10.01 Prob > F = 0.0000
. hausman fe
Note: the rank of the differenced variance matrix (0) does not equal the number of coefficients being tested (8); be sure this is what you expect, or there may be problems computing the test. Examine
the output of your estimators for anything unexpected and possibly consider scaling your variables so that the coefficients are on a similar scale.
---- Coefficients ----
| (b) (B) (b-B) sqrt(diag(V_b-V_B))
| fe . Difference S.E.
-------------+----------------------------------------------------------------
FINTECH | 4.33474 4.33474 0 0
LOGGDP | 32.07893 32.07893 0 0
LOGM2 | -8.9939 -8.9939 0 0
ROA | 4.206165 4.206165 0 0
DPR | .2252608 .2252608 0 0
CIR | -.16203 -.16203 0 0
CAR | -.6533295 -.6533295 0 0
LOGASSET | -10.73029 -10.73029 0 0
------------------------------------------------------------------------------
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(0) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 0.00
Prob>chi2 = .
(V_b-V_B is not positive definite)