Subject 320: Advanced Econometrics: Microeconometrics Module

Course Outline

The Microeconometrics module introduces a number of econometric techniques that are widely used when modelling observed phenomenon at the micro-level. The type of models considered naturally arise from the analysis of individual behaviour. Many of the techniques employed have their roots in the statistics and biometrics literature.

The focus of this module is an examination of techniques that are used in current research. The topics cover a broad spectrum from the curse of dimensionality that characterises many qualitative response and limited dependent variable models, to models of duration which originated in bioassay. We also introduce a number of computationally intensive techniques such as simulation-based inference, Bayesian approaches to model uncertainty and the bootstrap.

Following recent developments, we also examine procedures for circumventing the often maintained assumption of iid observations in cross-section data. This will involve examining models of cross-sectional dependence and social interaction.

Finally we examine ways in which economists combine revealed and stated preference data to estimate models of willingness-to-pay.
An application to valuing attributes on mobile phones will be presented.

If time permits we will also consider an alternative to classical simulation, indirect inference, and in addition entropy and cross-entropy solutions to ill-determined problems in econometrics.

Lecture Notes

Topic 1: Models of Unobserved Latent Variables