The fixed effects negative binomial model revisited

In this paper I show that the conditional fixed effects negative binomial model for count panel data does not control for individual fixed effects unless a very specific set of assumptions are met. I also propose a score test to verify whether these assumptions are met.