Marcelo Barbosa Ferreira

Princeton Economics Ph.D. Candidate

Causal Inference · Data Science · Quantitative Modeling

Princeton Economics Ph.D. candidate, expected May 2027, and Brazilian Mathematical Olympiad Gold Medalist. My work combines large datasets and causal inference to understand how individuals respond to income and policy changes, economics and mathematical modeling to evaluate strategic decisions.

I am interested in bringing this expertise to the private sector, where rigorous analysis can inform consequential real-world decisions.

Resume (PDF)

marcelof@princeton.edu

Methods & Tools

Causal Inference

Event Studies, Differences in Differences, Panel Data, Instrumental Variables, Regression Discontinuity, Time Series, Linear Regression

Data Science

Big Data, Record Linkage, Data cleaning,

STATA, MATLAB, Python, SQL

Quantitative Modeling

Economic Modeling, Welfare Analysis, Public Finance, Macroeconomics, Game Theory,

Selected Research

Labor supply Responses to a Very Large Income Shock: Evidence from Daughters of Brazilian Military Officers

How does a large, permanent income shock affect employment? I link military-pension records with Brazil’s longitudinal employer–employee data and use an event-study design to estimate labor-supply responses. Formal employment falls approximately 7%, and after seven years each $1 of pension income reduces labor earnings by $0.08. I then incorporate these estimates into a model that quantifies the program’s welfare trade-offs.

The Welfare Effects of Tenure Requirements for Unemployment Insurance

How should Unemployment Insurance eligibility depend on job tenure? Using linked employer–employee records, I reconstruct unemployment insurance spells and exploit a 2015 Brazilian reform to estimate how eligibility requirements affect layoffs and unemployment duration. I then incorporate these estimates into a welfare model. The estimated efficiency cost of providing unemployment-insurance benefits is $0.11 per dollar transferred.