Advanced microeconometrics of discrete choices

Objectives

This main objective of the course is to present recent advances in discrete choice modeling.

After some reminders about the structure of discrete choice econometric models (random utility models) and the reference model used by econometricians (the Multinomial Logit model), the course aims at presenting two essential extensions of the reference discrete choice models: (a) discrete choice models with random parameters and (b) discrete choice models with agglomeration and/or congestion effects. These models are used in all fields of applied microeconometrics: choice of transport modes and routes, choice of business or residential location, modeling of market shares under imperfect competition, choice of recreational sites, …

The course will address issues related to the specification and estimation of these models, both in theory and through detailed examples.

Course outline

1. Reminders
a. Random utility models
b. The Multinomial Logit Model
c. The case of large choice sets

2. Models with random parameters
a. Interest and specification of random parameter models
b. Continuous mixing law: estimation by the MV
c. Discrete mixing law: estimation by the MV via the EM algorithm
d. Some practical aspects

3. Location models with congestion and/or agglomeration effects
a. Congestion and agglomeration effects and equilibria
b. Specification and estimation of discrete choice models with congestion or agglomeration effects

4. Supplement. Market share models

Prerequisites

​Econometrics, applied econometrics​