Reg without constant term

[Daniel Gomez]

Hi,

I know that when I estimate a regression with fixed effects the constant term should not be included.
However, when I run the regression os Stata, it estimates the constant term.

How can I run my regression without a constant term?
or what is the interpretation of the constant term in the presence of the fixed effects?

Thanks!

[Joseph Coveney]

I'm not sure why you wouldn't usually want a constant term in a fixed-effects regression. Are you asking about something like the following?

sysuse bplong

quietly regress bp i.when ibn.patient, noconstant
margins when

xtreg bp i.when, i(patient) fe
margins when

[Daniel Gomez]

Yes, thanks Joseph Coveney.

I understand that no constant term is included in the presence of the fixed effects.
The equations with fixed effects usually do not have the constant term.

[Maarten Buis]

I understand that no constant term is included in the presence of the fixed effects.

There are multiple ways to identify a fixed effects model, some include the constant some don't, but in the end they are all exactly equivalent.

[Daniel Gomez]

Thanks Maarten Buis !

[Carlo Lazzaro]

Daniel:
you might be interested in taking a look at: https://www.stata.com/support/faqs/s...effects-model/

[Daniel Gomez]

Daniel:
you might be interested in taking a look at: https://www.stata.com/support/faqs/s...effects-model/

This is great!!!! Carlo Lazzaro