I'm interested in potentially using Poisson panel Generalized Method of Moments to hopefully gain consistent estimates for my parameters of interest. However, I have an endogenous binary variable (a time varying dummy variable I suspect is correlated with other time varying unobservables).

Following this working paper here: GMM for panel count data models by Frank Windmeijer, section 3.3, I notice that the Wooldridge moment condition is conditioned based on a 2 period lag of instruments of interest?

Is this correct? Could I use further lags as well? Any help would be definitely appreciated. Thank you. Wooldridge's transformation for endogeneity


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.