Advanced Applied Econometrics
Background and course objectives
The OECD economics department identified a training need in econometrics. There was demand for training at an advanced level, but focused on researchers doing applied work, illustrating tools with recent, well published applied studies in the literature. In terms of topics, there is strong interest in micro-econometric techniques, like difference-in-differences and (dynamic) panel data models. In terms of teaching methods, there is demand for a combination of lectures, interactive tutorials and some guidance to the use of statistical software, over a longer period of time (possibly even multiple courses).
This course was designed with these training needs in mind, and is loosely is modeled after an advanced Ph.D. level course in applied econometrics, which I designed, coordinated and co-taught at the graduate program in economics of the Universitat Pompeu Fabra. It is an advanced course, meaning that I assume that participants have already followed standard graduate or advanced undergraduate courses in econometrics, and have a working knowledge of statistical software. The course is taught at high speed, starting from the basics, but advancing quickly to the research frontier. At the same time, this is an applied course, geared towards researchers doing applied work on a regular basis. The course’s aim is not to give a survey of econometric theory. Rather, we discuss a range of techniques and empirical approaches, which are regularly and successfully used in applied work, and discuss applications, practical implementation, and potential problems and pitfalls.
Course content and teaching philosophy
The course includes six topics that I consider as essential for almost all applied researchers using micro-econometric methods. In this first part of the course, participants receive training in the modern way of thinking about applied micro-econometrics, the so called experimental approach, which borrows a lot of terminology and its way of thinking about identification from laboratory experiments. In addition, we discuss particular issues that arise in panel datasets, in particular the estimation of dynamic panel data models. Finally, we discuss cases, in which standard errors do not give an accurate reflection of the uncertainty associated with the point estimates, and ways to correct for these problems. If further, more specialized, needs arise or are revealed during the first part of the course, additional topics may be added as a second part.
To ensure effective learning, each topic consists of a lecture and an interactive workshop. The lectures start with an application of the technique under study. Using this application, typically either a classic or a recent contribution to the literature, we discuss what the methodological problem is, and why it is important in practice. Then, I explain how to use the method in practice, how to compare it to other methods (e.g. OLS), how to use statistical software to implement it, how to best present the results, and what problems to be aware of. If necessary, we discuss additional applications to illustrate particular strengths or weaknesses of the method. The workshops offer the opportunity to practice these skills using guided exercises (problem sets) or replicating existing studies. I am available to answer questions and offer suggestions related to the exercise.
In addition to lectures and workshops, I have individual meetings with the participants. These meetings are meant to offer OECD researchers the opportunity to consult with me on issues or problems related to their own work, whether related to the course material or not. Individual meetings serve a dual purpose. First, they offer direct feedback for the participants on their work for the OECD. Second, they give me a better insight into what methods and techniques are of interest to OECD researchers, allowing me to better focus the content of the lectures, particularly possible additional topics in part II.
As a statistical software package, Stata has become the standard for academic research in economics. Other packages, like eViews and SAS, are good alternatives, but since I am much less familiar with these, I can offer less support. If necessary for the workshops, I make code available in Stata format. Data are provided as text files, which are readable in all packages.
1. Regression Analysis and the Experimental Approach
· Introduction to the course
· Review of regression analysis
· Regression and causality
· Experimental approach
· The ‘basement’ of applied research
Joshua Angrist and Jörn-Steffen Pischke (2009), Mostly Harmless Econometrics: An Empiricist’s Companion, chapters 1-3
Joshua D. Angrist and Alan B. Krueger (1999). Empirical strategies in labor economics, in: O. Ashenfelter & D. Card (ed.), Handbook of Labor Economics, edition 1, volume 3, chapter 23, pages 1277-1366
Peter E. Kennedy (2002). Sinning in the Basement: What Are the Rules? The Ten Commandments of Applied Econometrics, Journal of Economic Surveys, 16(4)
Journal of Economic Perspectives, 24(2), Spring 2010
Joshua D. Angrist and Steve Pischke (2010). The Credibility Revolution in Empirical Economics: How Better Research Design is Taking the Con out of Econometrics.
Edward Leamer. Tantalus on the Road to Asymptopia.
Michael Keane. A Structural Perspective on the Experimentalist School.
Christopher Sims. But Economics Is Not an Experimental Science.
Aviv Nevo and Michael Whinston. Taking the Dogma out of Econometrics: Structural Modeling and Credible Inference.
Journal of Economic Literature, 48(2), June 2010
Angus Deaton. Instruments, Randomization, and Learning about Development.
Journal of Economic Perspectives, 25(3), Summer 2011 (Symposium: Field Experiments)
Jens Ludwig, Jeffrey R. Kling and Sendhil Mullainathan. Mechanism Experiments and Policy Evaluations.
David Card, Stefano DellaVigna and Ulrike Malmendier. The Role of Theory in Field Experiments.
Oriana Bandiera, Iwan Barankay and Imran Rasul. Field Experiments with Firms.
· Effects of the minimum wage on employment
· Difference-in-Differences (DD)
· DD and regression
· DD with multiple groups and fixed effects (FE)
· Review of FE regression
· Presenting results
· What is the source of identifying variation?
David Card and Alan B. Krueger (1994). Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania, American Economic Review, 84(4), pp.772-93
David Card (1992). Using Regional Variation to Measure the Effect of the Federal Minimum Wage, Industrial and Labor Relations Review, 46, pp.22-37
Joshua Angrist and Jörn-Steffen Pischke (2009), Mostly Harmless Econometrics: An Empiricist’s Companion, section 5.1-5.2
Guido W. Imbens and Jeffrey M. Wooldridge (2009). Recent Developments in the Econometrics of Program Evaluation, Journal of Economic Literature, 47(1), pp.5-86
3. Experiments in Economics and Instrumental Variables
· A typical field experiment in development economics: cost sharing
· A field experiment in the US: labor market discrimination
· An experiment in macroeconomics: fiscal stimulus
· Natural experiments
· Review of instrumental variables (IV) regression
Jessica Cohen and Pascaline Dupas (2010). Free Distribution or Cost-Sharing? Evidence from a Randomized Malaria Prevention Experiment. Quarterly Journal of Economics, 125 (1), pp.1-45
Marianne Bertrand and Sendhil Mullainathan (2004). Are Emily and Greg More Employable than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination. American Economic Review, 94(4), pp.991-1013
Jonathan A. Parker, Nicholas S. Souleles, David S. Johnson, and Robert McClelland (2010). Consumer Spending and the Economic Stimulus Payments of 2008, working paper Northwestern University
Joshua Angrist and Jörn-Steffen Pischke (2009), Mostly Harmless Econometrics: An Empiricist’s Companion, chapter 4
4. Regression Discontinuity Design
· Effect class size on achievement
· Sharp RDD
· Choosing the bandwidth
· Choosing the control function
· Fuzzy RDD
· Checklist for implementation
· Creating variation by using the cross-section
Joshua Angrist and Jörn-Steffen Pischke (2009), Mostly Harmless Econometrics: An Empiricist’s Companion, chapter 6
David S. Lee and Thomas Lemieux (2010). Regression Discontinuity Designs in Economics, Journal of Economic Literature, 48, pp.281-355
Guido Imbens and Thomas Lemieux (2008). Regression Discontinuity Designs: A Guide to Practice, Journal of Econometrics, 142(2), pp.615-635
Joshua D. Angrist and Victor Lavy (1999). Using Maimonides’ Rule to Estimate the Effect of Class Size on Scholastic Achievement, Quarterly Journal of Economics, 114(2), pp.533-775
5. Dynamic Panel Data Models
· Examples: income and democracy, convergence, education and growth
· Dynamic models with fixed effects
· Lagged dependent variable
· Bounding the true coefficient
· Using lags as instruments
· GMM estimators (Arellano-Bond, Blundell-Bond system GMM)
· Implementation in Stata
· Endogenous explanatory variables
· Serially correlated errors
· Dynamics in the effect of interest
Joshua Angrist and Jörn-Steffen Pischke (2009), Mostly Harmless Econometrics: An Empiricist’s Companion, section 5.3-5.4
Daron Acemoglu, Simon Johnson, James A. Robinson and Pierre Yared (2008). Income and Democracy, American Economic Review, 98(3), pp.808-842
Francesco Caselli, Gerardo Esquivel, and Fernando Lefort (1996). Reopening the Convergence Debate: A New Look at Cross-Country Growth Empirics, Journal of Economic Growth, 1(3), pp.363-389
Stephen R. Bond, Anke Hoeffler, and Jonathan Temple (2001). GMM Estimation of Empirical Growth Models, CEPR Discussion Papers 3048
Coen Teulings and Thijs van Rens (2008). Education, Growth, and Income Inequality, Review of Economics and Statistics, 90(1), pp.89-104
6. Standard errors
· Inference problems
· Standard error of the OLS estimator
· Generalized Least Squares (GLS)
· Robust standard error estimation
· Cluster-robust standard errors
· How much does clustering matter?
· Clustering as double-counting
· Autocorrelation in panels
Joshua Angrist and Jörn-Steffen Pischke (2009), Mostly Harmless Econometrics: An Empiricist’s Companion, chapter 8
Kurt Schmidheiny (2010). Clustering in the Linear Model, Short Guides to Microeconometrics
Brent R. Moulton (1986). Random Group Effects and the Precision of Regression Estimates, Journal of Econometrics, 32(3), pp.385-397
Jeffrey M. Wooldridge (2003). Cluster-Sample Methods in Applied Econometrics, American Economic Review (P&P), 93(2), pp.133-138
Marianne Bertrand, Esther Duflo and Sendhil Mullainathan (2004). How Much Should We Trust Differences-in-Differences Estimates?, Quarterly Journal of Economics, 119(1), pp.249-275
Two-way clustering made easy (thanks to Jean-Marc Fournier for pointing out these references)
Christopher F Baum, Austin Nichols and Mark E Schaffer (2010). Evaluating one-way and two-way cluster-robust covariance matrix estimates, presentation at the BOS’10 Stata Conference, July 2010
Samuel B. Thompson (2009). Simple Formulas for Standard Errors that Cluster by Both Firm and Time, working paper.
· Matching and the propensity score
· Heterogeneity and Local Average Treatment Effect
· Weak instruments
· Discrete choice models (logit, probit, tobit)
· Quantile regression
· Non-parametric estimation
· Synthetic panel data (repeated cross-sections)