Andrew Shephard, KU Leuven
Dynamic Discrete Choice Copula Models
Abstract
A copula-based specification is introduced to model serial dependence in utility shocks while preserving logit tractability in dynamic discrete choice models. With Type-I extreme value marginals and a Bernstein-polynomial copula, persistence is summarized by an endogenous finite-state latent rank Markov process: conditional choice probabilities and inclusive values are closed-form finite mixtures of logit expressions, and the likelihood integrates out the latent state by standard hidden Markov model filtering.
Andrew Shephard holds a PhD in Economics from University College London. Before joining the KU Leuven faculty, he was Associate Professor of Economics at the University of Pennsylvania. His research interests include labor economics, public economics, microeconometrics, and applied microeconomics.
You can rad more about Andrew Shepard here
CEBI contact: Thomas Høgholm Jørgensen