面板结构工作文件的Hausman检验结果表明用随机效应,但固定效应模型的效果明显好于随机效应模型,我到底应该用固定效应还是随机效应呢?跪求各位大神指点!
Hausman检验结果:
Correlated Random Effects - Hausman Test
Equation: Untitled
Test cross-section random effects
Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob.
Cross-section random 11.272081 7 0.1272
固定效应结果:
| | | | |
| Coefficient | Std. Error | t-Statistic | Prob. |
| | | | |
| C | -0.089925 | 0.056005 | -1.605654 | 0.1098 |
| PENSION | 1.386228 | 0.222804 | 6.221744 | 0 |
| FIXASS | 0.008831 | 0.001341 | 6.585614 | 0 |
| GROWTH | -0.003768 | 0.001097 | -3.433634 | 0.0007 |
| LEVERAGE | -0.024431 | 0.005663 | -4.314027 | 0 |
| OWNER | -0.002944 | 0.001605 | -1.834491 | 0.068 |
| ROE | 0.004253 | 0.005614 | 0.757551 | 0.4495 |
| SIZE | 0.00674 | 0.0024 | 2.808712 | 0.0054 |
|
| | |
|
| Effects Specification | | |
|
| | | |
|
Cross-section fixed (dummy variables)
| | | |
|
| | | |
|
| Weighted Statistics | | |
|
| | | |
|
R-squared
| 0.98712 | Mean dependent var | | 0.151894 |
| Adjusted R-squared | 0.980112 | S.D. dependent var | | 0.227274 |
| S.E. of regression | 0.017414 | Sum squared resid | | 0.065201 |
| F-statistic | 140.8389 | Durbin-Watson stat | | 3.034818 |
| Prob(F-statistic) | 0 | | | |
|
| | |
|
| Unweighted Statistics | | |
|
| | | |
|
R-squared
| 0.711131 | Mean dependent var | | 0.056736 |
| Sum squared resid | 0.071397 | Durbin-Watson stat | | 3.059369 |
| | | | |
随机效应结果:
| Coefficient | Std. Error | t-Statistic | Prob. |
| | | | |
| C | 0.131581 | 0.04697 | 2.801406 | 0.0054 |
| PENSION | 0.76114 | 0.940187 | 0.809562 | 0.4188 |
| FIXASS | 0.011608 | 0.006607 | 1.757029 | 0.0799 |
| GROWTH | -0.002282 | 0.004355 | -0.523993 | 0.6006 |
| LEVERAGE | 0.011652 | 0.014096 | 0.826617 | 0.4091 |
| OWNER | -0.004143 | 0.005233 | -0.791741 | 0.4291 |
| ROE | 0.011359 | 0.014476 | 0.784649 | 0.4332 |
| SIZE | -0.00344 | 0.002127 | -1.61746 | 0.1067 |
|
| | |
|
| Effects Specification | | |
|
| | | S.D. | Rho |
| | | | |
| Cross-section random | | | 0.020118 | 0.5503 |
| Idiosyncratic random | | | 0.018187 | 0.4497 |
|
| | |
|
| Weighted Statistics | | |
|
| | | |
|
R-squared
| 0.028852 | Mean dependent var | | 0.026252 |
| Adjusted R-squared | 0.007935 | S.D. dependent var | | 0.018379 |
| S.E. of regression | 0.018306 | Sum squared resid | | 0.108912 |
| F-statistic | 1.379353 | Durbin-Watson stat | | 2.041862 |
Prob(F-statistic) | 0.213169 |
|
|
|
| | | |
|
| Unweighted Statistics | | |
|
| | | |
|
R-squared
| 0.025564 | Mean dependent var |
| 0.056736
|
| Sum squared resid | 0.240841 | Durbin-Watson stat | | 0.923362 |
谢谢各位大神啦!