The information bound of a dynamic panel logit model with fixed effects

In this paper, I calculate the semiparametric information bound in two dyna mic panel data logit models with individual specific effects. In such a mod el without any other regressors, it is well known that the conditional maxi mum likelihood estimator yields a rootn-consistent estimator. In the case w here the model includes strictly exogenous continuous regressors, Honore an d Kyriazidou (2000, Econometrica 68, 839-874) suggest a consistent estimato r whose rate of convergence is slower than rootn. Information bounds calcul ated in this paper suggest that the conditional maximum likelihood estimato r is not efficient for models without any other regressor and that rootn-co nsistent estimation is infeasible in more general models.