I have the following, small question. I have panel data and have played around with it a bit. In my first estimation (1) I assumed that the coefficient of union is biased and therefore wanted to include time fixed effects (2) and individual fixed effects in the model (3). Would you say that including time fixed effects or individual fixed effects was necessary? I would say that it is definitely necessary to include fixed individual effects. Fixed time effects did not really change the coefficient, but it still makes sense to control for them. Would you agree with that?