[Carlo Lazzaro]

Luca:

1)

 
 ----------------+----------------------------------------------------------------         sigma_u |  3.2101967         sigma_e |  5.0243405             rho |  .28988884   (fraction of variance due to u_i) --------------------------------------------------------------------------------- F test that all u_i=0: F(366, 18149) = 5.14                  Prob > F = 0.0000

you do have panel-wise effect here (the fact that sigma_u<sigma_e, per se, does not explain the whole complexity of the issue).

2) the fact that you SHOULD (you're kindly requested not to shout) do what you should, is your own business, not mine.

3) the fact that you should use different standard errors is good as it seems at its face value.

All that said, I'm going to bow out of this thread.

[Luca Toni]

Luca:

1)

----------------+---------------------------------------------------------------- sigma_u | 3.2101967 sigma_e | 5.0243405 rho | .28988884 (fraction of variance due to u_i) --------------------------------------------------------------------------------- F test that all u_i=0: F(366, 18149) = 5.14 Prob > F = 0.0000

you do have panel-wise effect here (the fact that sigma_u<sigma_e, per se, does not explain the whole complexity of the issue).

2) the fact that you SHOULD (you're kindly requested not to shout) do what you should, is your own business, not mine.

3) the fact that you should use different standard errors is good as it seems at its face value.

All that said, I'm going to bow out of this thread.

So a few days ago you wrote that it seems I don't have panel-wise effect becuase of sigma e>sigma u, and now when I showed the F test, you write that I indeed have panel-wise effect because F test is significant. How do I fix the sigma e>sigma u problem then? Could I still use FE and rely on the results?
I didn't mean to shout, by the way.