Sufficient statistics for unobserved heterogeneity in structural dynamic logit models

We study the identification and estimation of structural parameters in dynamic panel data logit models where decisions are forward-looking joint distribution unobserved heterogeneity observable state variables is nonparametric, i.e., fixed-effects model. consider with two endogenous variables: lagged decision variable, time duration last choice. This class includes as particular cases important economic applications such market entry–exit, occupational choice, machine replacement, inventory investment decisions, or demand differentiated products. prove using a conditional likelihood approach. The structure model implies that there sufficient statistic function on this no longer depends – neither through current utility nor continuation value problem but still parameters. apply estimator to replacement