Can dummy variable interact with year to overcome the collinear in Fixed effect?

[Yong tong ling]

I would like to ask if dummy variable found collinearity in Fixed effect, can dummy variable interact with year?
I measure ownership 1 is foreign ownership, otherwise 0.
the following is my model, but i am not sure whether Ownership * Year is right.


Y = B0 + B1 * Ownership + B2 * (Ownership * Year ) + Controls + GDP

eg:

COMPANY A	2023	1	2023
COMPANY A	2024	1	2024
COMPANY B	2023	0	0
COMPANY B	2024	0	0
COMPANY C	2023	0	0
COMPANY C	2024	0	0

[Carlo Lazzaro]

Yong:
welcome to this forum.
If your predictor is time-invariant, it is wiped out by the -fe- estimator.
From your (limited) example, it would seem that owbership is time-invariant.
Eventually, I fail to see the rationale behind your proposed interaction: what is the reaearch question it will answer to?

[Yong tong ling]

Thank you for reply.
The research question is the how ownership can impact company performance

[Carlo Lazzaro]

Yong:
given your research question, I would not interact Ownership with Year and code:

xtreg <depvar> i.Ownership i.year Controls GDP, fe

Depending on the number of your panels (30, at least) I would consider (or not) -vce(cluster panelid)- standard errors.