Hello,

I have some doubts about panel regression in fixed effects:

1) When I run fixed effects should I always test for heteroskedasticity, , autocorrelation and serial correlation independence? Or can I draw my conclusions through the interpretation of fixed effects?

2) In the case the three tests are positive (applied to fixed effects model), how should I solve when N is greater than T and T is greater than N?

Thanks,
Mary