Dummy variable in panel data regression

[Jodie Lewicki]

I have a dummy variable in my panel data regression that controls for country of origin. 0 for Germany and 1 for US.

I want to run the fixed-effect panel data regression, but it does not work only with random effects so: xtreg dv iv, re

Is that right or is there a way to also use fe when I have a dummy variable?

[Clyde Schechter]

Since country of origin is an attribute of a person (or firm, or whatever your units of analyses are) that, by definition, never changes over time, it is not possible to estimate the effect of country of origin on an outcome in a fixed effects model. That variable is necessarily colinear with the fixed effects themselves, so its effect is not estimable.

That said, what is your purpose for wanting to include this variable? If your purpose is to estimate its effect, then, yes, you must go to random effects, or perhaps a between-effects model. But if you are only including it to adjust for its effects ("as a control variable") then you can just ignore it altogether and use a fixed-effects model. In a fixed effects model the effects of all time-invariant attributes are automatically adjusted for by the fixed effects themselves.

[Jodie Lewicki]

My purpose is to estimate its effect as I am looking at how females on board influence corporate sustainability in Germany and the US and compare those countries.

Maybe you can also help me with this question: I ran the random effects model now and wonder how I can interpret the results now. So how can I see if the country of origin has an effect? Germany has the value 1 and the US 0 and the dummy variable shows this -0,59 under the coefficient. What does that mean?

[Clyde Schechter]

I'd be happy to help with the interpretation, but to do that I would want to see the exact command you used, and the complete output from the regression (as well as anything that appears between them in the Results window).