Hi all,

the title quite speaks for itself. I am trying to recover how the bias intervenes in the estimates of a fixed effect poisson model with endogeneity. Specifically, how the correlation between endogenous regressor, say y2, and error impact on such bias. Please, notice that I am not looking for a comparison between estimates without controlling for endogeneity vs estimates obtained after control function appraoch. This indeed would be rather simple but would not be informative on the reasons why for instance the estimates controlling for endogeneity are higher (this is actually my case, where the correlation between error term and endogenous is positive and the instrument has a negative and significant sign when inserted in the regression explaining the endogenous,i.e. first stage regression) than the ones without controlling for endogeneity.

Thank you all,

Federico