Module description
What is the module about?
This course examines key methodological debates in modern econometrics, and explores how economic models can guide the interpretation of empirical evidence.
A central theme is the role of economic theory in empirical analysis, and the inherent trade-off between internal and external validity. An estimate has internally validity when it accurately captures the causal effect of X on Y in the specific context under study – for example, the true impact on wages of 1,000 immigrants with a particular skill mix arriving in a particular city at a particular time. But what happens if we double the number of immigrants? Or if the immigrants bring a different skill mix? Or if they arrive in a different city or country? An estimate has external validity if it is informative about causal effects in different contexts or under different policies. Such extrapolation requires a structural model – assumptions, grounded in economic theory, about how a market or system works. Stronger assumptions allow broader predictions, but those predictions are only as reliable as the assumptions themselves. Since the 1990s, empirical research has placed much greater emphasis on internal validity, in a trend often described as the “credibility revolution”. This has made econometric methodology much more rigorous, but some practitioners worry about a growing willingness to compromise on external validity.
We will explore how these questions inform the implementation of instrumental variable and panel data methods. And we will also consider how they shape debates on contemporary issues such as immigration and earnings inequality.
Who should do this module?
This module is available to students in the BSc Economics and Management, BSc Accounting and Finance and BSc Economics programs who took Introduction to Econometrics (5SSMN932) in their second year. It will complement material covered by Applied Econometrics (6SSMN961), but students are not required to have taken this course. Many problem sets will involve practical data exercises, and students will gain experience with Stata. The course will help students develop the skills to conduct independent empirical research – and is especially recommended for students who intend to go on to further study.
Outline of topics
- Role of economic models in econometrics
- Potential outcomes framework
- Structural models
- IV and simultaneous equation models
- Panel data and fixed effect models
Assessment details
50% In Person Examination
50% Project
Teaching pattern
Weekly Lecture
Weekly Tutorial
Suggested reading list
Key text or background reading
There is no course textbook: we will mostly rely on journal articles. As reference material, I recommend Mostly Harmless Econometrics (2008) by Angrist and Pischke.