[Lutfi Ozturker]

Thank you and one last question:

In my set-up, how can I employ country and time fixed effects simultaneously by using the "xtreg" command which you deem necessary?

Is it simply the following code?

xtreg Y X CVs i.country i.year,r cluster (company) 

PS: The following code

xtreg Y X CVs i.country i.year,fe vce (cluster company) 

produces "country omitted because of collinearity" error message.

[Carlo Lazzaro]

Lutfi:
1) your first code includes an -re- specification, that makes estimating coefficienst of time-invariant variables feasible;
2) your second code includes an -fe- specification, that makes estimating coefficienst of within-panel time-invariant variables (-i.country) unfeasible. Therefore, there's nothing you can do to estimate -o.country- coefficient here.