Oslo Summer School in Comparative Social Science Studies 2006

Qualitative Choice Modelling for Applied Economics

Lecturer: Post.doc. Erik Ø. Sørensen,
Department of Economics,
Norwegian School of Economics and Business Administration, Norway

Main discipline: Economics

Dates: 31 July - 4 August 2006
Course Credits: 10 pts (ECTS)
Limitation: 30 participants

We will review the basic model for qualitative choice, and examine how this model has been applied in both static and dynamic settings in applied labour economics and industrial organization. The basic objective of the course is to provide a solid grounding in the theory of qualitative choice and how empirical work can respect the structure of standard theory.

The first part of the course will focus on standard theory that is well known by all practitioners, and I will base the lectures on the text-book literature. The more advanced topics will be based on journal articles with applications. Some of these are very technical in nature, and I will only cover the basic structure of these papers in class.

Main textbook and support literature

The main textbook for this PhD course will be

This book presents the basics of qualitative choice, and it covers much of the necessary computational detail. This book will be supplemented by other readings. All participants are expected to have taken an introductory econometrics course at the graduete level. Those who feel they need the support of a top-level microeconometrics text should definetely get Cameron, A. C. and P. K. Trivedi (2005). Microeconometrics: Methods and Applications. Cambridge University Press. This is a wonderful book. A basic understanding of micro economics will also be assumed.

Texts in the outline below marked with * will be printed in a booklet and sent to all participants in advance of the course.

Lecture outline:

Monday sessions: Introduction and overview. The logit-model
The first day we will go through the standard random utility model and get familiar with the logit-models that follow from assuming extreme-value distributed random terms.

Tuesday sessions: Static logits
In this lecture I will focus on extensions of the basic models to choicesituations that are a bit more complicated, including multiple choices, orderings and repeated choices. We will also discuss the choice-theoretic content of “independence of irrelevant alternatives”. If there is time we will start on the topics for wednesday.

Wednesday sessions: working around the IIA
The “Independence of irrelevant alternatives” that follows from assuming extreme-value distributed random elements is not always appropriate. We will look at two standard approaches to working around it: (1) GEV/Nested logits and (2) mixed logits. We will cover one very basic applied paper of each type.

Thursday sessions: Dynamic choices
Most choices in life are sequential. We will cover two classic papers, one on machine replacement and one on contraceptive choice. While slightly dated, these papers are model papers for classical maximum likelihood estimation (Rust ’87, Pakes ’86) and for the literature on identification of dynamic models (Hotz& Miller ’93).

Friday sessions: aggregating choices into market outcomes
If individuals are making discrete choices over buying goods, the random utility model provides a lot of structure on market choices.

The Lecturer

Dr. oecon Erik Ø. Sørensen is a researcher at Norwegian School of Economics and Business Administration in Bergen, and his research interests are in labour economics, experimental economics and applied econometrics.

Main page - Current Oslo Summer School Program