Syllabus: Frontier Topics in Empirical Economics
Introduction
The goal of this course is to review some basic methods widely used in empirical economic research and introduce some new techniques developed by statisticians and econometricians. To take this course, students must have finished their first-year econometrics class and understand some basic ideas of causal inference. It would be even better if they have learned some structural model-based methods (but not required). The content of this course is methodologically intensive, but the main target is not on the details of math. The most important part is how to use these methods correctly in the empirical research. We will focus more on the design-based approach and give only a brief introduction to the structural-based approach. There will be many homework papers for students to read. They are the applications of the methods introduced in class. It is extremely important for students to read them carefully and write a report for each one.
Office Hour:
Textbook:
-
J.D.Angrist and J.Pischke (2009) Mostly Harmless Econometrics (Main)
-
B.Neal (2020) Introduction to Causal Inference https://www.bradyneal.com/causal-inference-course
-
J.Pearl (2009) Causality: Models, Reasoning, and Inference
-
K.Train (2009) Discrete Choice Methods with Simulation https://eml.berkeley.edu/books/choice2.html
-
B.Hansen (2021) Econometrics
-
J. Wooldridge (2010) Econometric Analysis of Cross Section and Panel Data
-
James, Witten, Hastie and Tibshirani (2017) An Introduction to Statistical Learning
-
James, Witten, Hastie and Tibshirani (2008) The Elements of Statistical Learning, 2nd Edition
Content
Week 1: Outline of Causal Inference Slides Handout
Structural vs Design-based approach, Potential Outcome Framework, RCT, control variable and Simpson Paradox, matching and regression
Reading:
J.D.Angrist (2009) Chapter 2, 3.1, 3.2, 3.3
B.Neal (2020) Chapter 1,2
Paper Report: Bertrand and Mullainathan (2004) Are Emily and Greg More Employable than Lakisha and Jamal: A Field Experiment on Labor Market Discrimination
Week 2: Non-parametric Method Slides Handout
Parametric vs non-parametric vs semi-parametric method, NW and kernel estimator, local polynomial regression, partially linear model, bootstrap
Reading:
Hansen (2021) Chapter 10.6, 10.7, 19.1, 19.2, 19.3, 19.4, 19.5
Carol’s Notes Class 1.2, 2.3, 3.1.1, 3.1.4, 5.2
Paper Report: Dube, Jacobs, Naidu, and Suri (2020) Monopsony in Online Labor Markets
Week 3: Machine Learning and Model Selection Slides Handout
Introduction to machine learning, model selection, penalized regression, tree-based method, random forests, causal forests, neural networks
Reading:
James, Witten, Hastie and Tibshirani (2008) Chapter 2.1, 2.2, 3.1-3.4, 5.1, 6.2, 8.1, 8.2, 9.1, 9.2, 15.1-15.3 (JWHT 2008 is a simpler version of JWHT 2017. You can also refer to the corresponding chapters in JWHT 2017.)
James, Witten, Hastie and Tibshirani (2017) Chapter 11.3
Wager and Athey(2016) Estimation and Inference of Heterogeneous Treatment Effects Using Random Forests
Athey and Imbens (2019) Machine Learning Methods Economists Should Know About
Paper Report: Ro’ee Levy (2021) Social Media, News Consumption, and Polarization: Evidence from a Field Experiment (Please also read Online Appendix C.5)
Causal graph, DAG framework of causal inference, more on bad control and control choice
Reading:
B.Neal (2020) Chapter 3, 4, 5, 6
Imbens(2020) Potential Outcome and Directed Acyclic Graph Approaches to Causality: Relevance for Empirical Practice in Economics
Paper Report: Pinto (2015) Selection Bias in a Controlled Experiment: The Case of Moving to Opportunity
Week 5: Instrumental Variable Method (I) Slides Handout
IV, LATE, GMM, Oster bound
Reading:
J.D.Angrist (2009) Chapter 4.1, 4.2, 4.4, 4.6.1
Hansen (2021) Chapter 13.1-13.9, 13.21
Oster (2019) Unobservable Selection and Coefficient Stability: Theory and Evidence
Paper Report: Clark et al (2021) Compensating for Academic Loss: Online Learning and Student Performance during the COVID-19 Pandemic
Week 6: Instrumental Variable Method (II) Slides Handout
Choice model and causal inference, Non-parametric interpretation of IV beyond LATE and monotonicity, MTE
Reading:
J.D.Angrist (2009) Chapter 4.5
Heckman and Vytlacil (2007) Econometric Evaluation of Social Programs, Part I
Heckman and Vytlacil (2007) Econometric Evaluation of Social Programs, Part II
Paper Report: Kline and Walters (2016) Evaluating Public Programs with Close Substitutes: The Case of Head Start
Week 7: Instrumental Variable Method (III) Slides Handout
Two approaches to understand Bartik IV
Reading:
Pinkham, Sorkin and Swift (2020) Bartik Instruments: What, When, Why, and How
Borusyak, Hull and Jaravel (2022) Quasi-Experimental Shift-Share Research Designs
Adao, Kolesar and Morales(2019) Shift-share Designs: Theory and Inference
Paper Report: Autor, Dorn and Hanson (2013) The China Syndrome: Local Labor Market Effects of Import Competition in the United States
Week 8: Causal Inference with Panel Data (I) Slides Handout
How to use FE, DID, and event study, what are their implications, what are their regression assumptions, parallel trend testing, synthetic control
Reading:
J.D.Angrist (2009) Chapter 5.1, 5.2
Roth (2021) Pre-test with Caution: Event-study Estimates after Testing for Parallel Trends
Abadie (2021) Using Synthetic Controls Feasibility, Data Requirements and Methodological Aspects
Paper Report: Abadie, Diamond and Hainmueller (2014) Comparative Politics and the Synthetic Control Method
Week 9: Causal Inference with Panel Data (II) Slides
Staggered DID, more general Two-Way Fixed Effect model and their issues
Reading:
Chaisemartin and D'Haultfoeuille (2020) Two-Way Fixed Effects Estimators with Heterogeneous Treatment Effects
Chaisemartin and D'Haultfoeuille (2022) Two-Way Fixed Effects and Differences-in-Differences Estimators with Several Treatments
Callaway, Bacon and Sant'Anna (2021) Difference-in-Differences with a Continuous Treatment
Paper Report: Roth et al (2022) What’s Trending in Difference-in-Differences: A Synthesis of the Recent Econometrics Literature
Week 10: Regression Discontinuity Slides
Identification and estimation of RDD and RKD
Reading:
J.D.Angrist (2009) Chapter 6
Lee (2008) Randomized Experiments from Non-random Selection in U.S. House Elections
Gelman and Imbens (2019) Why High-order Polynomials Should not be Used in Regression Discontinuity Designs
Hahn, Todd, and Van der Klaauw (2001) Identification and Estimation of Treatment Effects with a Regression-discontinuity Design
Calonico, Cattaneo and Titiunik (2014) Robust Nonparametric Confidence Intervals for Regression-Discontinuity Designs
Card et al (2015) Inference on Causal Effects in a Generalized Regression Kink Design
Card et al (2016) Regression Kink Design: Theory and Practice
Paper Report: He, Wang, and Zhang (2020) Watering Down Environmental Regulation in China
Week 11: Standard Errors’ Issues Slides
Why and how to cluster standard errors, design-based uncertainty
Reading:
J.D.Angrist (2009) Chapter 8.2
Abadie, Athey, Imbens and Wooldridge (2020) Sampling‐Based versus Design‐Based Uncertainty in Regression Analysis
Abadie, Athey, Imbens and Wooldridge (2023) When Should You Adjust Standard Errors for Clustering
Week 12: Discrete Choice Model (I) Slides
Static discrete choice model, Logit, Probit, nested Logit,
Reading:
K.Train (2009) Chapter 2.1-2.5, 3.1-3.3, 3.5-3.6, 4.1, 4.2
Week 13: Discrete Choice Model (II)
Endogeneity in non-linear model, IV-probit, control function approach, BLP, Heckman two-step
Reading:
K.Train (2009) Chapter 13.1, 13.2, 13.4
J. Wooldridge (2010) Chapter 15.7.2
Berry, Levinsohn and Pakes (1995) Automobile Prices in Market Equilibrium
Paper Report: Arcidiacono (2005) Affirmative Action in Higher Education: How Do Admission and Financial Aid Rules Affect Future Earnings
Acknowledgement:
I want to deliver my special thanks to Junjian Yi. The applied econometrics reading group organized by him inspires the design of this course. I also want to thank Ronni Pavan, Lisa Kahn, Nese Yildiz, John Singleton, Kegon Tan, Gregorio Caetano, and Carolina Caetano for all empirical techniques I learned when I was at the University of Rochester. I thank Dengke Chen for carefully going through all slides in the first draft and giving me many suggestions.