Panel Data with fixed effect

[quynh_nguyennz]

Hi,
Sorry to bother you again. I already have the panel data and my supervisor asked me to choose the fixed effect . however when I choose fixed effect, it said near singular matrix. The attached below is my file
Thank you so much.

[EViews Glenn]

Usually it means that one of your variables is time invariant so that it is perfectly collinear with the effects.

A quick check of your workfile shows that DUM3 is time invariant.

[quynh_nguyennz]

Could you please specify it? Because I want to test whether the 5 dummy effect my resull
How can I fixed the problem
Thank you so much

[quynh_nguyennz]

Dum3 is for the countries in EU region

[EViews Glenn]

The only think you can do is to leave DUM3 out of the equation. The coefficient on DUM3 is not identified in the fixed effects specification.

[quynh_nguyennz]

Thank You Glenn, so the others dummy can use for fixed effect, except the dummy3
regards

[EViews Glenn]

I didn't check them. Possibly. Try it.

[quynh_nguyennz]

Hi Glenn,
I try to create another dummy variable today
Dum1=1 if the countries are greece, ireland, portugal, spain, france, germany otherwise denote eu countries
dum3=1 if the countries are uk,us, denmark, sweden, australia, canada. denote non eu countries
and dum2=1 if countries are greece, ireland, portugal,spain denote for debt crisis countries
when I estimate equation,
rsquared c dum1 dum2 dum3, it say nonsingular matrix. i think the problem is between dum1 and dum3. if I put dum1 and dum3 together i can not run equation
how can I fixed it
I attached my file below.
Thank you very much

[EViews Glenn]

You need to leave out either DUM1 or DUM3. Do a Google (or even a forum search) on the "dummy variable trap."