Fixed effect in a pooled regression

[suyi mcmaster]

Hi Statalist,

I wonder whether it make sense to add the industry and year fixed effects in a pooled regression? The code will be reg y x i.industry i.year, r

Best regards,
Suyi

[Fei Wang]

If you have a pooled cross-sectional of firms over multiple years, then yes and you'd better include such fixed effects.

[Carlo Lazzaro]

Suyi:
welcome to this forum.
As an aside to Fei's helpful advice, please note that your non-default option for standard error should be -vce(cluster panelid)-, as, under -regress-, -robust- deal with heteroskedasticity only.

[suyi mcmaster]

Thanks for your reply!