Aims

This module is a further investigation into econometric methods and techniques with a particular focus on techniques commonly used with micro data. It is a particularly useful module for students intending to undertake empirical analysis in their dissertations, intending to take a Masters course or for those who are considering entering into employment involving data or policy analysis.

The module deals with questions around causal inference. OLS always identifies some form of partial correlation. However, if people can self-select or if not all relevant variables are observable, these partial correlations might not have a causal interpretation. To frame our thinking, we will consider two useful frameworks to think about causality - the first is a definition of causality based on the potential outcomes framework (also called the Rubin causal model) that enables us to define the effect of a treatment on an outcome as the difference between an observed outcome and a hypothetical “counterfactual” outcome that would have prevailed had a different treatment state been realised. We will also consider a more recent approach based on graphical methods pioneered by Judea Pearl. We will then talk about estimators and research designs that allow us to infer causal relationships from observational data.

Outline Of Syllabus

(0) A brief revision of OLS (this is not formally part of the course but might be a useful refresher for the material covered in stage 2)
(1) Intuition of causal inference – why is this hard?; potential outcome framework, directed acyclical graphs
(2) Regression as a causal inference tool
(3) Fixed effects
(4) Difference-in-differences
(5) Instrumental variables
(6) Regression discontinuity designs

Teaching Rationale And Relationship

The lectures will cover both theoretical material and applications. We will also have seminars for some topics. In these, we will usually discuss two applied papers.

Assessment Rationale And Relationship

The essay will tests students' ability to critically assess and explain the application of econometrics techniques in an applied setting.