Ordinary Series for Panel Data

[LuqmanSaeed]

Greetings,

My inquiry is related to panel data. I have time-series cross section data on provinces of Pakistan.The dependent variable is count variable and i am regressing negative binomial with sets of cross section specific variables. However there is one variable which is common to all the cross section. How do i put that in the work file? The series follows a pattern where every cross section time series data is followed by next cross section, for example, data on education is like,

edu
1-81 25
1-82 30
till
2010
2-81 30
2-81 35
till 2010.
where 1 and 2 represents cross section followed by year. The variable which is same for every cross section, will its observation be repeated for every cross section or there is some other specific way to put the value for such variable in workfile.

[EViews Gareth]

Repeated.

[Aijan]

Hi there,

I have a panel data where some of the variables are the same for each cross-section (the variable is a country gdp, and firms in this country represent different cross-sections) . As I understood, I need to copy the data on the common variable for each cross section. So that it looks like this:

Year1 cross-section1 gdp1
Year2 cross-section1 gdp2
...
Year1 cross-section2 gdp1
Year2 cross-section2 gdp2
Etc...

But the problem is that when I run a regression, I get the following error "Near singular matrix".

Could you advise whether it is related to copying the same variable across all cross sections, or maybe something else.. And how can I avoid this problem.

Thanks!

[EViews Gareth]

You probably have perfect collinearity between your regressors (including any fixed effects, if you have them). Without knowing more details, it is impossible to advise further.

[Aijan]

Thank you) I will look into econometrics textbooks for the theory of FE