Multicollinearity in simple diference-in-differences model

[Laxman Sivaraman]

Hi All,
I have a simple 2-period panel data and I am trying to run a difference-in-differences model using the xtreg command.

generate treatxpost = treatment*post
xtreg sales post treatment treatxpost, fe

treatment and post are the dummy variables treatment group and post period.

I have balanced data for households and each household has sales in the pre and post period.

I am not getting an estimate for treatment as Stata is telling me: treatment omitted because of collinearity.

How do I get an estimate for both the treatment and post along with the interaction term treatxpost?

Appreciate your pointers on this.

Thanks,
Laxman.

[Joseph Coveney]

Search the list for Clyde Schechter's previous replies to this same question. It's been raised numerous times and he's always polite and patient—you'll benefit from his answers.

[Laxman Sivaraman]

Okay, I will do that. Thank you!