CAUSAL INFERENCE FOR PUBLIC POLICY ANALYSIS

This course focuses on the study of modern empirical methods for causal policy evaluation based on counterfactuals. It will discuss the rationale behind the use of counterfactuals for public policy evaluation, and related key concepts including treatments and controls, different treatment effects (ATE, ATET, ATU, ITT, LATE), selection bias, heterogenous treatment effects, conditional independence, SUTVA. Next, it will present state-of-the-art experimental and quasi-experimental estimation techniques: Randomised Control Trials, Instrumental Variables, Regression Discontinuity Design, Difference-in-Differences, Event Studies, Synthetic Control Method. The course will discuss in depth key literature contributions adopting these methods, analysing advantages, disadvantages, underlying assumptions, and complementarities of each estimation strategy. The textbook used in the course is Cunningham, S. (2021) “Causal Inference. The Mixtape”, Yale University Press. The course will also run computer classes giving students the possibility to see the methods in practice through STATA exercises.

Course objectives / expected outcomes
The students are expected to learn modern empirical methods assessing the causal effect of public policies, developing the capability to recognise the most suitable setting for each method. They will be able to understand whether the fundamental conditions for the application of a method are met, and whether/when different methods can be combined to make empirical analyses more credible. They are also expected to be able to see the advantages of counterfactual approaches relative to more traditional estimation techniques, and understand their limitations. In addition, they will learn the most appropriate robustness tests to be implemented in each context. By the end of the course, they will also learn how to apply these methodologies using standard econometric software.

Knowledge of econometrics is required

Course topics:
- Introduction to Counterfactual Policy Evaluation
- Randomised Experiments (RCTs)
- Matching
- Instrumental Variables (IV)
- Regression Discontinuity Design (RDD)
- Difference-in-Differences (DiD)
- Staggered Difference-in-Differences and Event Studies
- Synthetic Control Method (SCM)
- Application of DiD, RDD and Synthetic Control Method

Cunningham, S. (2021). Causal Inference. The Mixtape, Yale University Press.

Long take-home essay (5000 words). Students will select a topic, dataset, methodology and produce a short paper using the techniques learnt in the course.

Lectures
Classes with practical applications of methods

scritto