Fixed Effects check up

[Manel Berjano]

Hey guys,

so I have a naive question, after a FE regression for Panel data what procedure should one follow to check if the model is ok?
For example how does one test if the strict exogeneity assumption is valid in the model?
Or should the errors be white noise ?

Thanks for the help in advance.

[Carlo Lazzaro]

Manel:
welcome to this forum.
1) strict exogeneity relates to the epsilon error only (see paragraph 8.9.2 in

https://www.stata.com/bookstore/microeconometrics-stata

). Under the -fe- framework, the panel-wise effect (if any) is correlated with the vector of regressors (weak endogeneity).
2) Model misspecification can be checked via a procedure similar to the one detailed under -linktest- entry, Stata .pdf manual.