Dear Statalisters,
I am interested in estimating the effect of X on Y, where both X and Y are counts (non-negative integers). I have an instrument for X. I would like to use a Control Function approach. Can I estimate the first stage with Poisson, and then use the generalized residuals in another Poisson regression of Y on X? I am aware of the paper by Jeff Wooldridge (2015) about CF for a binary endogenous variable but I am not sure this can be applied to my case where the endogenous variable (X) is a count. I was planning on using ppmlhdfe for both stages.
Many thanks!
Iacopo
I am interested in estimating the effect of X on Y, where both X and Y are counts (non-negative integers). I have an instrument for X. I would like to use a Control Function approach. Can I estimate the first stage with Poisson, and then use the generalized residuals in another Poisson regression of Y on X? I am aware of the paper by Jeff Wooldridge (2015) about CF for a binary endogenous variable but I am not sure this can be applied to my case where the endogenous variable (X) is a count. I was planning on using ppmlhdfe for both stages.
Many thanks!
Iacopo
Comment