Jefferson Leal

Jefferson Leal

he/him
PhD Candidate
University of Rochester
jefferson.leal (at) rochester.edu


Hi! I am a Ph.D. Candidate in Political Science at the University of Rochester. My research examines how candidate strategy shapes public policy provision in developing democracies. I study how ideological positioning, far-right extremism, populism, and political experience affect public policy outcomes, drawing primarily on Brazilian local politics while engaging with broader comparative evidence. Methodologically, I am interested in causal inference, text-as-data, and machine learning.

Before coming to Rochester, I received my M.A. in political science and my B.A. degrees in economics and international relations from the University of São Paulo, in Brazil. In my free time, I love to hike and learn tango.



Research

Publications

Anderson Frey & Jefferson Leal. “When Crisis Hits, Does Experience Matter? Evidence from COVID-19 Vaccination” PDF
Conditionally accepted at the British Journal of Political Science

Abstract Politicians often highlight prior office experience to appeal to voters, who seem to value this attribute. However, the role of political experience in determining actual policy outcomes remains overlooked by the literature, particularly during times of crisis when experience becomes uniquely salient to the electorate. This article examines how political experience affects policy implementation during a crisis using data on COVID-19 vaccinations across Brazilian municipalities in 2021. The findings show that vaccination rates were 12 p.p. higher in cities governed by incumbent mayors reelected in November 2020—who already had extensive experience managing the pandemic—compared to cities that elected newcomers. These effects cannot be attributed to bureaucratic turnover or differences in reelection incentives. They are also concentrated in municipalities severely affected by the pandemic earlier in 2020, supporting the argument that early crisis management experience, through learning-by-doing, significantly enhanced mayors’ performance during the vaccine rollout.


Book Chapters

Eduardo Lazzari & Jefferson Leal. “Brazilian Tax Policy under the Lens of Inequality” Publisher
In The Policies of Politics · in Portuguese

Abstract Tax policy is a central instrument for reducing income inequality, yet in Brazil it has drawn far less attention than social spending. This chapter examines the personal income tax from 1985 to 2017, the federal tax with the most direct distributive effect and therefore the clearest record of what successive governments intended. Drawing on data from the Federal Revenue Service and the National Treasury Secretariat, and on every statute that altered the tax over the period, we trace revenue composition, the treatment of income across brackets, and the tax's legislative history. Regressivity held steady from at least 1990 to 2015, and no government implemented significant changes aimed at reducing inequality through this tax. Most amendments cut progressivity through exemptions and deductions. Neither democratization, electoral competition, nor left-wing incumbency produced redistributive tax policy, which follows a political logic of its own.


Working Papers

Jefferson Leal. “Local Right-Wing Populism and Vaccination: Evidence from the COVID-19 Pandemic in Brazil” PDF

Abstract Despite a growing literature on the impacts of national leaders' populist stances and voters' populist attitudes, less is known about how local politicians' right-wing populism influences policy outcomes. The study focuses on the Brazilian case during the COVID-19 pandemic, where the right-populist President Jair Bolsonaro successfully employed an anti-science stance to spread skepticism towards vaccinations and social distancing measures. I use a Naïve Bayes classifier and leverage over 14,000 mayoral manifestos to present a novel measure of local-level right-wing populism. Using a close-elections regression discontinuity design, I document a significant decrease in COVID-19 vaccination rates in municipalities run by right-wing populist mayors. This effect is most likely attributable to these authorities sending mixed signals about the severity of the health crisis and exerting low effort to promote vaccination.


Jefferson Leal. “Which Issues Divide? Interpretable Ideal-Point Estimation from Political Text”

Abstract Political actors reveal ideology through both the positions they take and the issues they emphasize. Yet, most text-based ideal-point measures capture only one of these two channels. Existing approaches also impose a trade-off: unsupervised methods struggle where language is not sharply polarized, while supervised ones require human-labelled training data and ex ante knowledge of how relevant issues and policy positions map onto ideology. I develop an item-response model that jointly scales ideal points from issue salience and issue-specific stance. My approach discovers topics inductively and uses transfer learning with multilingual data from the Manifesto Project to classify stance, requiring no additional human labels. I apply the model to campaign platforms from 4,507 candidates in US House primaries and 16,830 candidates in Brazilian mayoral elections. The estimated positions align with benchmarks based on roll-call votes, campaign contributions, and expert surveys, even in the Brazilian corpus, where local campaign language is largely non-polarized. The model further recovers substantively distinct salience and stance divisions, providing an interpretable account of which issues structure ideological competition.




Teaching

ICPSR Summer Program in Quantitative Methods

Teaching Assistant

Graduate

University of Rochester

Teaching Assistant
Best Teaching Assistant in Political Science, 2024

Undergraduate

Universidade Paulista, Brazil

Instructor of Record

Undergraduate (2018–2021)

IPSA São Paulo Summer School in Methods

Teaching Assistant

Graduate

University of São Paulo

Teaching Assistant

Graduate



Policy Memos

Ian Prates, Rogério Barbosa & Jefferson Leal. “Black men and black women are the most vulnerable in the crisis; however, a ‘new vulnerable’ group emerges: white men and white women in non-essential services” PDF Folha Jornal da USP
Solidary Research Network · Technical Note 3, April 2020

Abstract This bulletin classifies workers in Brazil by their vulnerability to the Covid-19 economic crisis, combining how unstable a worker's job is with how badly the pandemic hit the sector in which they work. We also map how vulnerability varies across regions, education, gender, and race. The distribution of vulnerable groups is broadly similar across the country, with some nuances. Less developed regions, such as the North and Northeast, have larger shares of workers in essential sectors and in unstable jobs, while in the South, Southeast, and Midwest vulnerability comes mainly from the large presence of workers in non-essential sectors. Men and women differ mainly by sector, since men work more often in essential activities and women in non-essential ones. Black and white workers differ mainly by job stability, since white workers hold more secure positions. We also identify a "new vulnerable" group: white men and women, mostly college-educated and in relatively stable jobs, yet concentrated in essential sectors that were badly hit or in non-essential ones. Black men and women remain the "traditionally vulnerable," and Black women are the most vulnerable of all, because they combine the least stable jobs with non-essential sectors. The traditionally vulnerable are still worse off than the new vulnerable.


Ian Prates, Rogério Barbosa, Jefferson Leal et al. “Emergency benefit providing R$600 needs to continue and could be funded by emergency contribution on high incomes” PDF Jornal da USP
Solidary Research Network · Technical Note 8, May 2020

Abstract This bulletin examines labor market trends in Brazil during the first months of the Covid-19 pandemic, focusing on rising unemployment and the negative impact on household income following an executive order that allowed employers to cut hours and pay, with government-funded partial compensation. Prolonged social distancing disrupted the labor market and reduced both employment and household income, making the continuation of the Emergency Basic Income (EBI) of R$600 a month (about US$110 at the time) an unavoidable question for Congress and for Brazilian society. We estimate that these hours-and-pay agreements alone lowered per capita household income by almost R$200, and that formal workers who could not access the EBI were left without compensation for those losses. Cutting the benefit to R$200 (about US$37), as the Minister of Economy proposed at the time, would push another 20 million people into poverty. We show that keeping the R$600 benefit for three additional months could be fully funded by an emergency social contribution on high incomes.