Hey everybody. I know I shouldn't exclude the terms of an interaction variable in my regression; however, If I include both terms I have the following problem:

1. Multicollinearity which I cannot address since the interaction variable is the product of two categorical variables

2. I have to exclude a set of fixed effects, since it actually controls for one of the term of the interaction variable. This, however, significantly reduces my R2.

Can I just include my fixed effects and exclude one of the 2 terms and in particular the one that is controlled by the fixed effects?

Thank you very much in advance.