IRVAPP ADVANCED SCHOOL 2020 - Advanced Methods for Impact Evaluation (Online)

The school is designed to be particularly beneficial to graduate students and researchers with intermediate proficiency in applied policy evaluation and quantitative micro-econometrics.It will run over six consecutive days, from December 14 to December 19, 2020, 2.00 pm (CET) to 5.30 pm (CET) (30mins break at 3.30 pm). It will focus on the estimation of policy effects with endogenous participation. Different solutions to the identification problem will be discussed, resulting in three thematic sessions.

Due to the ongoing pandemic, the school will be held on-line.

Targeted audience
This is an intermediate/advanced course on quantitative empirical methods for policy evaluation. Knowledge of methods for treatment effects and quantitative micro-econometrics equivalent to Ph.D. or Doctorate level coursework is required.

Programme

Causal inference, machine learning, interference (Fabrizia Mealli)

This part will provide an introduction to some exciting new areas in causal inference and
program evaluation. After an introduction to the basics of causal inference for experimental
and observational studies, machine learning methods for causal inference will be
introduced with a focus on issues of estimation of heterogeneous effects, assessment of
overlap, and sensitivity analysis. Next, it will review literature on observational studies with
interference, and present methods to estimate direct and spillover effects with cluster and
network data. Examples and case studies will be discussed.

Discontinuities (Enrico Rettore)

This part will address identification of policy effects when the probability of participation changes discontinuously or has kinks along observable dimensions. Starting with the basic methodology and implementation, the module will discuss extensions ranging from multiple cut-offs, to extrapolation and external validity, and randomization-inference methods.

Instrumental Variables (Erich Battistin)

This part will present solutions when (quasi-)experimental sources of exogenous variability for treatment participation become available. Identification of “local average” and “marginal” treatment effects will be discussed, offering a broad overview of ex-ante and ex-post evaluation methods when effects are heterogeneous across the statistical units.

The course will be taught in English.

Office hours:
The instructors will be available for individual meetings with participants willing to discuss on their research project. Given the limited amount of time available, requests of meeting will be dealt with on a first-in first-served basis.

Materials:
All the slides used by instructors will be sent to participants prior to the beginning of the school.

Tuition Fees
The registration fee includes workshop materials. We are not in the position to offer any scholarship.

• Students: EUR 200
• Others: EUR 300

All bank charges are for the registrant’s account.

Lectures
The links for the lectures link will be sent some days prior to the lecture.