Challenges to Identification of Treatment Effect Parameters

• Grant number: 25/04857-0
• Support Opportunities: Regular Research Grants
• Start date: September 01, 2025
• End date: August 31, 2028
• Field of knowledge: Applied Social Sciences - Economics - Quantitative Methods Applied to Economics
• Principal Investigator: Vítor Augusto Possebom
• Grantee: Vítor Augusto Possebom
• Host Institution: Escola de Economia de São Paulo (EESP). Fundação Getúlio Vargas (FGV). São Paulo , SP, Brazil
• City of the host institution: São Paulo

Abstract

Many economists face identification challenges when conducting empirical research. These challenges are frequently associated with non-classical measurement error, self-selection into treatment, and heterogeneous treatment effects. This research proposal consists of six projects that tackle these three types of identification challenges in different contexts. The first project analyzes non-classical measurement error in proxies for agricultural productivity. The second project focuses on self-selection into treatment and is associated with the mediation analysis literature, decomposing the marginal treatment effect into a direct effect and an indirect effect. The third, fourth and sixth projects focus on heterogenous effects. The third project is a fully applied project and analyzes how the effect of judges' gender on judicial cases' outcomes varies by type of crime and type of evidence. The fourth project adopts a Bayesian perspective for judge fixed effect models. It discusses how to estimate the effect of judicial decisions on recidivism without the need to estimate judge's leniency rates in an initial stage. The fifth project identifies nonlinear treatment effects in shift-share designs via a control function approach. Lastly, the sixth project analyzes contamination bias in OLS regressions with multi-valued discrete treatment variables. (AU)