Dear Statalist,

I am estimating the effects of variable x on outcome y. Since x might be endogenous, I use z as the instrument. I estimate two models (FE and FEIV) as I have panel data. The results below show a big increase in the size of coefficient on x variable. Although FEIV is preferred to address endogeneity, I am not sure whether there is a technical explanation for the increase. Thank you.


Array Array