TOPICS IN ECONOMETRICS I

G6427

Department of Economics

Autumn 2004

Columbia University

Rajeev H. Dehejia

 

 

Contact information:

807B IAB

420 W. 118th Street

Telephone:  854-4659

E-mail: dehejia@columbia.edu

 

Course information:

Lectures: Monday 6.10 – 8.00 pm, 501 IAB

Office Hours: MW 1 - 2 p.m.

 

Who should be interested:

This course will present selected topics in econometrics aimed at students interested in working with micro-level data.  The techniques discussed are relevant for students working in fields such as labor economics, public economics, international, finance, development, industrial organization… You get the idea. Second year students will learn a range of useful techniques. Students in the third year or higher will be encouraged to present their work in progress.

 

Course description:

The course has two themes: causal inference and Bayesian methods.

 

Causal inference refers to the body of theory that helps us to understand when a causal claim can be made. A huge proportion of applied papers in economics purport to make causal claims: education increases earnings, training programs increase employment, globalization increases income volatility, rules of corporate governance affect bankruptcy rates, access to credit affects child labor, etc. The course will present a framework within which to evaluate the legitimacy of these types of claims. The course will also revisit the issue of sample selection bias, presenting some additional techniques.

 

The second, overlapping theme is Bayesian econometrics. There has been a traditional divide in econometrics between the Bayesians and frequentists. This distinction is increasingly becoming irrelevant. We will present and understand Bayesian methods through the lens of simulation-based likelihood techniques. This is a powerful class of tools that is finding wide application in a range of fields including labor economics, finance, and industrial organization. Bayesian applications we will consider range from traditional tools like regression and IV, to causal inference, to extended topics such as portfolio choice.

 

Students are welcome to attend only one of the two modules of the course, but those planning to do so should attend the first class for organizational reasons.

 

Prerequisites: This course is intended for second- or higher-year Ph.D. students in economics. Full knowledge of the first-year Ph.D. sequence will be presupposed.

 

Course evaluation:

There will be no exams. There are two grading options. Students can present a paper that they are working on, or they can present a paper that they will select with my guidance. Auditors are welcome.

 

Readings

A course packet will be made available with copies of many articles. Several chapters from Andrew Gelman, John Carlin, Hal Stern, and Donald Rubin, Bayesian Data Analysis (London: Chapman and Hall) will be used (but will not be part of the packet).

 

 

COURSE OUTLINE

 

Lecture 1: A definition of causality; potential outcomes versus predictability.

Cox, D.R. (1992), “Causality: Some Statistical Aspects,” Journal of the Royal Statistical Society, Series A, part 2, 291-301.

 

*Holland, P. (1986), “Statistics and Causal Inference” (with discussion), Journal of the American Statistical Association, 81, 945-970.

 

Neyman, J. (1923), “On the Applicability of Probability Theory to Agricultural Experiments. Essay on Principles. Section 9,” translated in Statistical Science (with discussion), Vol. 5 (1990), No. 4, 465-80.

 

Rubin, D. (1974), “Estimating Causal Effects of Treatments in Randomized and Non-Randomized Studies,” Journal of Educational Psychology, 66, 688-701.

 

Pratt, J.W., and R. Schlaifer (1988), “On the Interpretation and Observation of Laws,” Journal of Econometrics, Vol. 39, No.1/2, 23-53.

 

Sims, C. (1972), “Money, Income, and Causality,” American Economic Review, 540-552.

 

Lecture 2: Randomized Experiments, Randomization Inference, and Blocking

Cox, D.R. (1958), Planning of Experiments, New York, Wiley, Chapters 1-3.

 

*Fisher, R.A. (1935), The Design of Experiments, Chapter 2, “The Principles of Experimentation, Illustrated by a Psycho-Physical Experiment”.

 

Neyman, J, “Statistical Problems in Agricultural Experimentation” (with discussion), Journal of the Royal Statistical Society, Supplement, Vol. II, No. 2.

 

Lecture 3: Application:  Natural Experiments:

*Meyer, B., W.K. Viscusi, and D. Durbin (1995), “Worker’ Compensation and Injury Duration: Evidence from a Natural Experiment,” American Economic Review, Vol. 85, 322-40.

 

Meyer, Bruce (1989), “A Quasi-Experimental Approach to the Effects of Unemployment Insurance,” NBER Working Paper No. 3159.

 

Meyer, Bruce (1994), “Natural and Quasi- Experiments in Economics,” NBER Technical Working Paper No. 170.

 

Application: Fisher’s Randomization Test for Diffs-in-Diffs Models

Bertrand, Marianne, Esther Duflo, and Sendhil Mullainathan, “How Much Should We Trust Differences-in-Differences Estimates?” NBER Working Paper No. 8841. Note: do not refer to the published version of this paper in the QJE.

 

Observational Studies with Ignorable Treatment Assignment

*Rubin, D.B. (1977), “Assignment to a Treatment Group on the Basis of a Covariate,” Journal of Educational Statistics, 2, 1-26.

 

Rubin, D.B. (1990a), “Comment: Neyman (1923) and Causal Inference in Experiments and Observational Studies,” Statistical Science, 5, 472-480.

 

Rubin, D.B. (1984), “William G. Cochran’s Contributions to the Design, Analysis and Evaluation of Observational Studies,” in Poduri and Rao (eds.), W.G. Cochran’s Impact on Statistics.

 

Rubin, D.B. (1991), “Practical Implications of Modes of Statistical Inference for Causal Effects and the Critical Role of the Assignment Mechanism,” Biometrics, Vol. 47, 1213-1234.

 

Lecture 4:  Application: Regression Discontinuity Design

Lee, David (2004), “Economic Impacts of New Unionization on Private Sector Employers: 1984-2001”, with John DiNardo, Forthcoming in Quarterly Journal of Economics Working paper version: Economic Impacts of Unionization on Private Sector Employers: 1984-2001, NBER Working Paper #10598, July 2004.

 

Lee, David, “Randomized Experiments from Non-random Selection in U.S. House Elections”, revised September 2003. Previous version: The Electoral Advantage to Incumbency and Voters' Valuation of Politicians' Experience: A Regression Discontinuity Analysis of Elections to the U.S. House, NBER Working Paper #8441, August 2001

 

Hahn, Jinyong, Petra Todd, and van der Klaauw, Wilbert (1999), “Identification and Estimation of Treatment Effects with a Regression-Discontinuity Design”.

 

Application: The Limitations of Regression and Econometric Adjustment

*Lalonde, Robert (1986), “Evaluating the Econometric Evaluation of Training Programs,” American Economic Review, Vol. 76, 604-20.

 

Lecture 5: The Role of the Propensity Score

Rosenbaum, P., and D. Rubin (1983), “The Central Role of the Propensity Score in Observational Studies for Causal Effects,” Biometrics, 70 (1), 41-55.

 

Rosenbaum, P., and D. Rubin (1984), “Reducing Bias in Observational Studies Using Subclassification on the Propensity Score,” Journal of the American Statistical Association, Vol. 79, 516-524.

 

Rosenbaum, P., and D. Rubin (1985), “Constructing a Control Group Using Multivariate Matched Sampling Methods that Incorporate the Propensity Score,” American Statistician, Vol. 39, 33-38.

 

*Dehejia, Rajeev H. and Sadek Wahba (1999),  “Causal Effects in Non-Experimental Studies:  Re-Evaluating the Evaluation of Training Programs,” Journal of the American Statistical Association, Volume 94, Number 448 (December 1999), pp. 1053-1062.


Hirano, Kei, Guido Imbens, and Geert Ridder (2003), 
"Efficient Estimation of Average Treatment Effects using the Estimated Propensity Score," (with G. Imbens and G. Ridder), 2003, Econometrica 71, 1161-1189. Earlier version appeared as NBER Technical Working Paper 251.


Abadie, Alberto, and Guido Imbens (2003). "Large Sample Properties of Matching Estimators for Average Treatment Effects." Manuscript.


Application: Semi-parametric Diffs-in-Diffs

Alberto Abadie, “Semiparametric Difference-in-Differences Estimators,” Harvard University Working Paper also in Review of Economic Studies.

 

Lecture 6: Instrumental Variables

*Imbens, Guido, and J. Angrist, “Identification and Estimation of Local Average Treatment Effects,” Econometrica, Vol. 62,  467-75.

 

*Angrist, J., G. Imbens, and D. Rubin, “Identification of Causal Effects Using Instrumental Variables” (with discussion), Journal of the American Statistical Association, 91, 444-72.

 

Angrist, J. (1990), “Lifetime Earnings and the Vietnam Era Draft Lottery:  Evidence from Social Security Administrative Records,” American Economic Review, 80, 313-36.

 

Application: Estimating a Demand Function

*Graddy, K. (1995), “Testing for Imperfect Competition at the Fulton Fish Market,” RAND Journal of Economics, 26, 75-92.

 

Application: Connection to Latent Variable Models

Heckman, J. (1979), “Sample Selection Bias as a Specification Error,” Econometrica, 47, 153-61.

 

*Barrow, Burt, Glen Cain, and Arthur Goldberger (1983?),  “Issues in the Analysis of Selectivity Bias,” Evaluation Studies, Vol. 5., 43-59.

 

 

 

 

Lecture 1: Statistical decision theory and the role of Bayes rules.

Readings: Lecture notes 1 and 2.

 

Lecture 2: The normal likelihood and classical regression. Posterior simulation. The Predictive Distribution.

Readings: Lecture note 3. Gelman, et al., chapters 2, 3, and 8.

 

Lecture 3: Markov chain Monte Carlo: Gibbs sampling and data augmentation.

Readings: Lecture note 4.

 

Applications: The Tobit model. The probit model.

Chib, Siddhartha (1992).  “Bayes Inference in the Tobit Censored Regression Model,” Journal of Econometrics, 51, 79-99.

 

Albert, J. and S. Chib (1993). “Bayesian Analysis of Binary and Polychotomous Response Data,” Journal of the American Statistical Association, 88, 669-679.

 

Lecture 4: Panel data, random effects, and hierarchical models.

Readings: Lecture note 5. Gelman et al., chapter 13.

 

Dehejia, Rajeev H. (2000), “Was There a Riverside Miracle? A Framework for Evaluating Multi-Site Programs,” National Bureau of Economic Research Working Paper No. 7844 (August 2000).

 

Lecture 5: Bayesian models of causal inference.

Imbens, Guido, and Donald Rubin (1997), “Estimating Outcome Distributions for Compliers in Instrumental Variable Models,” Review of Economic Studies, October 1997.

 

Imbens, Guido, and Donald Rubin (1997), “Bayesian Inference for Causal Effects in Randomized Experiments with Noncompliance,” Annals of Statistics, 25, 305-327.

 

Lecture 6: Decisions, Program Evaluation

*Dehejia, Rajeev H. (1997), “Program Evaluation as a Decision Problem,” National Bureau of Economic Research Working Paper No. 6954 (February 1999).

 

Lecture 7: Applications to Financial Economics

Kandel, S., and R. Stambaugh (1995), “Predicting Returns in the Stock and Bond Markets,” Journal of Finance, 17, 357-90.

 

*Barberis, N. (2000) “ Investing for the Long Run when Returns are Predictable,” Journal of Finance, 55, 225-64.

 

Stambaugh, R (1997). “Predictive Regressions,” Journal of Financial Economics, 54, 375-421.

 

Aguilar, Omar, and M. West (2000), “Bayesian Dynamic Factor Models and Portfolio Allocation,” Journal of Business and Economic Statistics, 18, 8-57.

 

Polson, N. and B. Tew (2000), “Bayesian Porfolio Selection: An Empirical Analysis of the S&P 500 Index, 1970-1996,” Journal of Business and Economic Statistics, 18, 164-73.

 

Vrontos, I., P. Dellaportas, and D. Politis (2000), “Full Bayesian Inference for GARCH and EGARCH Models,” Journal of Business and Economic Statistics, 18, 187-98.

 

Watanabe, Toshiaki (2000), “Bayesian Analysis of Dynamic Bivariate Mixture Models: Can They Explain the Behavior of Returns and Trading Volume,” Journal of Business and Economic Statistics, 18, 199-210.

 

Waggoner, Daniel F. and Tao Zha (1999), “Conditional Forecasts in Dynamic Multivariate Models,” Review of Economics and Statistics, 81, 639-51.