The Microeconometrics and Policy Evaluation programme presents recent developments in the microeconomic analysis of impact evaluation, with courses taught by experts in their fields.
- “Methods of policy evaluation” (David Margolis) introduces the main methods currently used for program evaluation
- “Machine learning for policy evaluation” (Philipp Ketz) presents recent advances in machine learning techniques for policy analysis
Students will attend practice class everyday. According their level of STATA, students will attend one of the following groups:
- Policy Evaluation in Stata (intermediate level)
- Advanced Topics in Stata (advanced level)
The courses are designed for graduate students wishing to gain additional skills in applied microeconometrics for policy evaluation, as well as for researchers and professionals in public and private institutions who work on impact evaluation.
Methods of Policy Evaluation - David Margolis
The econometric techniques for evaluating the impact of policies using micro-level data have evolved substantially over time, and understanding which technique is most appropriate in any particular situation is crucial to providing credible estimates of the effects of policy. This course will explore a wide variety of techniques, focusing not only on the estimators themselves and why they work, but also on the types of data they require and the assumptions that must hold for the estimates to be valid.
1. Introduction: setting and real world experiments
2. Comparing similar individuals
a. Regression models
b. Matching models
c. Regression discontinuity
3. Simulating unobserved outcomes
a. Instrumental variables
b. Selection models
4. Intertemporal comparisons
c. Synthetic controls
5. Summarizing methods
- W.H. Greene, Econometric Analysis, Prentice Hall.
- A.C. Cameron & P.K. Trivedi, Microeconometrics: Methods and Applications, Cambridge University Press.
Other higher-level texts
- J.M. Wooldridge, Econometric Analysis of Cross Section and Panel Data, MIT Press.
- R. Davidson & J.G. MacKinnon, Estimation and Inference in Econometrics, Oxford University Press.
Machine Learning for program evaluation - Philipp Ketz
The objective of this course is to understand how machine learning methods can be useful for program evaluation and what the difficulties and challenges are. We start with a short introduction to machine learning, highlighting its usefulness for prediction. Then, we discuss two recent adaptations of machine learning in economics, where the focus lies with causal inference. They concern the consistent estimation of the average treatment effect in the presence of many control variables and the analysis of heterogeneous treatment effects in the context of randomized controlled trials (RCTs), respectively.
1. Introduction to machine learning
a. Bias-variance tradeoff
c. Overview of popular methods
2. Estimating average treatment effects in the presence of many controls
a. Problem with “naive” estimators
b. “Double” machine learning as solution
3. Estimating heterogeneous treatment effects in RCTs
a. Tree-based methods
b. Generic methods
- Belloni, A., V. Chernozhukov, and C. Hansen (2014a). High-dimensional methods and inference on structural and treatment effects. Journal of Economic Perspectives 28(2), 29–50.
- Davis, J. and S. B. Heller (2017). Using Causal Forests to Predict Treatment Heterogeneity: An Application to Summer Jobs. American Economic Review 107 (5), 546–50.
- James, G., D. Witten, T. Hastie, and R. Tibshirani (2013). An introduction to statistical learning. Springer.
- Mullainathan, S. and J. Spiess (2017). Machine learning: an applied econometric approach. Journal of Economic Perspectives 31(2), 87–106.
- Varian, H. R. (2014). Big data: New tricks for econometrics. Journal of Economic Perspectives 28(2), 3–28.
- Athey, S. and G. Imbens (2016). Recursive partitioning for heterogeneous causal effects. Proceedings of the National Academy of Sciences 113 (27), 7353–7360.
- Belloni, A., V. Chernozhukov, and C. Hansen (2014b). Inference on treatment effects after selection among high-dimensional controls. The Review of Economic Studies 81(2), 608–650.
- Chernozhukov, V., D. Chetverikov, M. Demirer, E. Duflo, C. Hansen, W. Newey, and J. Robins (2018). Double/debiased machine learning for treatment and structural parameters. The Econometrics Journal 21(1), C1–C68.
- Chernozhukov, V., M. Demirer, E. Duflo, and I. Fernandez-Val (2018). Generic machine learning inference on heterogenous treatment effects in randomized experiments. Working paper.
- Hastie, T., R. Tibshirani, and J. Friedman (2009). The elements of statistical learning. Springer.
- Wager, S. and S. Athey (2018). Estimation and Inference of Heterogeneous Treatment Effects using Random Forests. Journal of the American Statistical Association 113 (523), 1228–1242.
Practice classes - Margherita Comola and Liam Wren-Lewis
All students are expected to have basic knowledge of the STATA software. Students will attend 2 hours a day of practice classes, according to their level:
- The intermediate group will attend the course Policy Evaluation in Stata (professor Liam Wren-Lewis) which illustrates the estimation methods presented in the theoretical class taught by David Margolis using observational data from various sources (e.g. cross sectional databases collecting information on labor market and health-related individual outcomes, as well as firm-level data).
- The advanced group will attend the course Advanced Topics in Stata (professor Margherita Comola). This course is intended for students who don’t need guidance to use built-in Stata commands, and it provides an introduction to Stata programming tools. Topics discussed in class include: matrix-based Mata sintax, log-likelihood functions, coverage probabilities, matching algorithms.