Econometrics

Econometrics uses economic theory, mathematics, and statistical inference to quantify economic phenomena, converting qualitative statements (such as "the relationship between two variables is positive") into quantitative ones (such as "an increase in income by Rs 100 increases consumption by Rs 90"). Essential to the social sciences, public policy evaluation, and business practice, econometrics is particularly concerned with untangling cause and effect: inferring that one variable (e.g., education) influences another (e.g., worker productivity), all else equal.

The first half of the course builds foundations in econometric theory, covering bivariate and multivariate regressions with both continuous and dummy variables, estimation using OLS, and the challenges and limitations of these methods. The second half turns to causal inference, introducing instrumental variables, difference-in-differences, and regression discontinuity. Classic research papers in economics that employ these methods are discussed throughout.

The first half of the course begins with introducing with what Econometrics is and its applications. We then move on to discussing estimation of simple and multiple linear regression models with OLS, hypothesis testing of regression coefficients, reading regression tables and interpreting regression results, and proving the Gauss Markov Theorem.  Typical problems that arise in regression analysis are discussed, including – omitted variable bias, measurement error, multicollinearity, heteroskedasticity, etc. Different functional forms and dummy variables in multiple regression models are discussed. Limited dependant variable models are also discussed.

The second half of the course will begin with introducing the notion of causal inference and briefly discussing the concept of random assignment – why it is useful for the purpose of sorting cause and effect. Then we discuss regression as a tool for causal inference.  The course then moves to exploring advanced techniques of causal inference, including difference-in-differences, instrumental variables, regression discontinuity, and propensity score methods – diving deep into each of these techniques one-by-one, with their multiple applications from research papers in economics to be discussed in the classroom lectures, student presentations, and homework assignments.

Students will be trained in Stata Software (We prefer Stata be used to run all regression models for the course)

By the end of this course, students should be able to:

1. Make sense of simple and multiple linear models and estimating them with OLS.
2. Understand hypothesis testing of regression coefficients, reading regression tables, and interpreting regression results.
3. Understand potential problems recurring in regression analysis, including omitted variable bias, measurement error, multicollinearity, heteroskedasticity, etc. and ways to address them.
4. Understand different functional forms and limited dependant variable models.
5. Make sense of the idea of causal inference and why randomized assignment is so useful for the purpose of sorting out cause and effect.
6. Make sense of how regression models can be used as a tool for causal inference.
7. Apply difference-in-differences techniques to study causality when experiments happen naturally in society.
8. Illustrate the use of instrumental variables in evaluating causality in complex real-world applications.
9. Explain how regression discontinuity is used to draw inferences about causal effects from rules constraining human behaviour.
10. Demonstrate expertise in the use of Stata Software to run Econometric models.

After completing this course, students should be able to:

1. Describe regression as technique to study data and make quantitative statements in Economics.
2. Explain the estimation of regression models with OLS, along with hypothesis testing of regression coefficients.
3. Demonstrate expertise in the use of Stata Software to run Econometric models; read, understand and interpret regression results.
4. Explain potential problems recurring in regression analysis, including omitted variable bias, measurement error, multicollinearity, heteroskedasticity, etc. and ways to address them.
5. Describe regressions with different functional forms and limited dependant variable models.
6. Explain of the idea of causal inference and why randomized assignment is useful for the purpose of sorting out cause and effect.
7. Explain how regression models can be used as a tool for causal inference.
8. Illustrate the use difference-in-differences techniques to study causality when experiments happen naturally in society.
9. Illustrate the use of instrumental variables in evaluating causality in complex real-world applications.
10. Explain how regression discontinuity is used to draw inferences about causal effects from rules constraining human behaviour.

• Angrist, J. D., Pischke, J. (2014). Mastering 'Metrics: The Path from Cause to Effect. Princeton University Press.
• Gujarati, Damodar, Porter, Dawn C, Gunasekar, Sangeetha (2009). Basic Econometrics, Sixth Edition, McGraw Hill
• Stock, J.H. and Watson, M.W. (2011) Introduction to Econometrics. 3rd Edition, Addison-Wesley, Boston.
• Wooldridge, J. M. (2022). Introductory Econometrics: A Modern Approach, 7E. Cengage Learning India