Do we must to control the individual fixed effects?

[Wang Jia-jia]

Hi. I have a quesstion about individual fixed effects.

We always add the individual fixed effects in our model to control for unobeserved plant-level characteristics which doesn't change over time.

But if my independent variable is 0-1(dummy variable such as did), could I just add industry-year fixed effects and province-year fixed?

Thanks for your advice.

Jiajia

[Jared Greathouse]

I'm unclear why your focal predictor being 0 or 1 is important here. You can usually add fixed effects in any panel dataset, but whether this makes sense is a separate question.

The better question for me is what are you investigating? What're you researching? how are you designing your paper?

[Carlo Lazzaro]

Wang:
as an aside to Jared's helpfule reply, what uis recommended in -xtreg,fe- is to add -i.timevar- in the right-hand side of your regerssion equation.
Oftentimes, interacting time-invariant predictors with time gives back coefficients that, if helpful in understanding the data generating process you're investigating, are difficult to explain.