Hi, Dr. Lian,
I have a quick question regarding my research really struggling me. I watch your vedio of panel data. It's really a good tutoring.

Thanks for any suggestion in advance.

I have a fixed effect model to examine the firm's performance in a panel data set.

The dependent variables is firm's Return on asset(ROA).

ROA= f(Firm level control, state level control)

Can I put the state level control on the right hand side? I cluster the standard error by firm ID, however, if I put the state level variables on the right hand side, Do I  need to do cluster the standard error by the state.

Or this is totally wrong by put two different level control in a single model?

Thanks
Best regard.