allowing for individual coefficients on panel data

[Frdmn]

Hello,

I was wondering if there was a way of estimating using OLS on panel data with coefficient specific to each panel individual (instead of computing by hand an OLS for all panel individuals, just using the panel nature of data to speed things up).


Thanks,

Frdmn

[EViews Gareth]

Two options:
1) Convert your panel to a pool and use pool estimation, which does allow for individual coefficients.
2) Use @expand to generate the cross-section interaction terms yourself: y c @expand(@crossid)*x

[Frdmn]

Amazing! thanks Gareth, I like the 2nd way of doing, could I use another thing than @crossid ? for example if ids (Brazil, Peru..) are countries and I have another panel time serie where I have the region ('Latin America" for instance) for each country, can I have a coefficient per region?

[EViews Gareth]

Sure