I regularly teach this course as a field course in the PhD Program in Economics (PIE), which is jointly offered by the Universities of Innsbruck and Linz. The course schedule can be found in the syllabus.
Objectives
This course covers the basic econometric concepts and methods used in modern applied microeconometrics to estimate causal effects in the presence of potentially unobserved confounding factors. Both theoretical and applied work will be discussed. The emphasis is, however, on application rather than on method. Methods will be illustrated by papers drawn from the most recent literature. Students will also use Stata to estimate models with actual data. By the end of the course, students should have a firm grasp of the types of research designs that allow a
convincing estimation of causal effects and be ready to apply methods to their own research questions and data.
Topics
1. Causality, Rubin Causal Model
2. Fixed Effects Estimation
3. Difference-in-Differences Estimator
4. Instrumental Variables Estimation
5. Regression Discontinuity Design
6. (Matching Estimators; if requested)
Prerequisites
This course is for PhD students with some prior training in econometrics, who are ideally working on an empirical dissertation. The methods discussed are also useful in an experimental setting. Students should be at least familiar with the basics of linear regression and have experience with Stata. Students without prior knowledge in Stata should contact me at the beginning of the semester, I can provide them with supporting materials and resources for self-study.
Readings
The primary textbook for this course is:
Angrist, J. and Pischke, J.-S. (2008). Mostly Harmless Econometrics: An Empiricist’s Companion. Princeton University Press, Princton, NJ
Additional reading material is listed in the reading list (reading list). Required readings are indicated by *.
Meetings
The course consists of 13 meetings (90 minutes each); broken down by three types:
Theory-meetings: In the theory-meetings, I will introduce the respective estimation method. The focus will be (as much as possible) on the intuition behind the method rather than on the algebra. Before class, students are expected to read assigned readings from the respective theory-section from the reading list.
Lab-sessions:In the Lab-sessions, we will actively practice the respective estimation methods by going through a set of prepared examples.
Application-meetings: In the application-meetings, we will discuss three papers drawn from the most recent literature (or somewhat older but seminal contributions). We will jointly select three papers from the respective application-section in the reading list. Each student is expected to read all three selected papers. Each paper will be presented by an assigned student (15 minutes) followed by a group discussion (15 minutes). The focus of the discussion should be on the identification strategy. The thorough discussion of actual applications should help to reinforce the theory-meetings.
Material
Rubin Causal Model: slides.
Fixed Effects Estimation & Difference-in-Differences: slides, state-package* (log-file, as pdf).
Instrumental Variables Estimation: slides, stata-package* (log-file).
Regression Discontinuity Design: slides, stata-package* (log-file).
Matching Estimators: slides, stata-package* (log-file).
applications 1, 2
* The full stata-package (including the data) will be provided in class.
Course Grading
Students are expected to read assigned readings and attend all classes. Grades for the course will be based on:
physical and intellectual attendance (10%);
a classroom presentation (30%);
and a final exam (60%).
Further details will be announced later.
Interesting links
- Equitable Growth in Conversation: An interview with David Card and Alan Krueger (link, April 2016)
- Podcast with Joshua Angrist on Econometrics and Causation (link, December 2014)