Abstract:
This paper studies identification in binary response dynamic panel data models with switching state dependence. Departing from the standard approach of modelling binary response dynamic panel data models, where last period’s choice enters as an additional regressor, this paper allows for switching dependence where current period’s decision depends on whether this period’s choice differs from last period’s choice. This form of correlation causes inertia in individual choices and is suitable for modeling cases where individuals face some form of high \switching costs”. This contemporaneous effect in choices, where the choice an individual makes in the current period directly affects current period’s latent utility, results in the model being logically inconsistent, making the model both incomplete and incoherent, which might result in lack of point-identification.