Econometric Analysis of Sequential Discrete Choice Models

This paper specifies an estimable dynamic model of sequential discrete choices in a controlled jump-process framework. We study sufficient conditions under which the agent's optimal policy is stationary. We show that the observable event histories at the micro-level are sample semi-Markovian. We provide, for the first time, sufficient and necessary conditions under which the destination-specific hazard functions belong to the proportional hazard family. We propose a computing algorithm for statistical inference of the structural parameters from longitudinal survey data.