Research

In Progress:

• "Family Culture and School Achievement in Brazil" (with Daniel Lopes and Leonardo Monasterio, IPEA).

The abstract is below:

The importance of family background in explaining educational outcomes inequality is well documented. According to the literature (see Björklund and Salvanes, 2011), more than 50% of the variation in years of schooling can be attributed to factors shared by siblings. On the other hand, the explanatory power of parental education in explaining educational outcomes is much lower, generally below 20%, what leaves a large margin for explaining factors other than parental education, such as family culture. This paper estimates the impact of family culture on educational outcomes of a cohort of Brazilian students. Based on a longitudinal data with student identification, we apply an algorithm of surname classification that assigns the student, based on the surnames of his/her parents and grandparents, to one of the following ancestry groups: Iberian, Japanese, Italian, German, Eastern European and Syrian-Lebanese. Our identification strategy relies on the epidemiological approach combined with an education production function setup. We show that, despite no significant differences in terms of preschool attendance rate, students with non-Iberian European or Japanese ancestry obtain statistically and substantively higher promotion rates from the 1^st to the 5^th grades and scores on 3^rd and 5^th grade nationwide standardized tests, even with a large set of individual, family and municipal controls and with a FE model that controls for students' place of birth and trajectory. We also control for the persistence of local institutions, established during the era of mass immigration to Brazil in the 19^th and 20^th centuries, and we show that the mechanisms of family transmission of culture remain robust.

Citation in midia: https://www1.folha.uol.com.br/educacao/2017/12/1945729-descendente-de-japones-esta-um-ano-na-frente-em-matematica-na-escola-publica.shtml

• "Does higher salary improve teacher quality and teaching profession attractiveness? A triple-differences analysis exploring an exogenous source of salary hike in Brazil " (with Cristine Pinto, EESP-FGV).

The abstract is below:

The main objective of this paper is to assess the impact of salary hikes on students' proficiency in the short run, and on the quality of teachers and on potential future teachers. In Brazil, not only are teachers’ salaries lower compared to alternative occupations, but also poor performing students seem to be more likely to pursue a teaching career. However, the minimum salary for teachers introduced in 2009 triggered a noticeable exogenous relative increase in municipal teachers’ salaries compared to other occupations. These changes were more noticeable in municipalities affected by the minimum salary law, particularly among those municipalities that complied with the law. Due to the lack of secondary data for base teacher salary, we had to carry out a survey with the municipal departments of education to gather information about teacher career structure and base salaries from 2008 to 2013. Triple-differences estimates reveal mild effects of teachers’ salary hikes on municipal teachers’ quality, proxied by their scores in a higher education completion exam (ENADE), and on the attractiveness of courses related to the teaching profession . The latter may induce improvements in teachers’ quality in the long-run. We apply a Triple-Differences model aiming at controlling for two types of potential confounding trends: (i) changes in the performance of municipal teachers (potential future teachers) across municipalities unrelated to the policy; and (ii) changes in the performance of teachers of all school systems working (students living) in municipalities affected by the new policy.

• "Quality of education and human capital decisions: experimental evidence from Brazil" (with Lucas Finamor - Yale University; Bruno Ferman and André Portela Souza, EESP-FGV).

The abstract is below:

This study makes use of the "Jovem de Futuro" program, an experimental intervention in Brazilian public high schools, to estimate the impact of quality of education on college and/or labor market decision. Our results indicate that high school students respond with more and better college enrollment after an increase in the quality of high school education. A better high school environment boosts the probability of students (i) to attend college; (ii) to be accepted into public colleges and into high quality and selective majors. These effects seem to be intermediated with admission through affirmative action policies (quotas). A reduction in working and only working (not studying) status is also detected. The effects are heterogeneous by baseline proficiency - as better is the inicial condition, greater are the effects. These results can be interpreted as suggestive evidence of the existence of credit constraint or heterogeneity in the returns to schooling function.

• "Qual o papel do Ensino Médio na geração de desigualdades educacionais entre indivíduos com diferentes níveis socioeconômicos?" (with Joana Costa, IPEA)

Estudos mostram que uma parte importante da desigualdade de renda é gerada na fase de formação das habilidades, antes mesmo do ingresso dos indivíduos no mercado de trabalho. O objetivo deste artigo é investigar a evolução das desigualdades educacionais, investigando o papel do Ensino Médio (EM) na geração de desigualdades educacionais entre estudantes de diferentes níveis socioeconômicos. Foi construída uma base de dados única com a junção de diversas bases administradas pelo Inep, Censos da Educação Básica e da Educação Superior, Prova Brasil e Enem, por meio do identificador único dos alunos pertencentes à coorte de matriculados no 9º ano do EF regular de escolas públicas brasileiras em 2011. Empreenderam-se análises multivariadas sobre a probabilidade de se alcançar o 3º ano do EM sem repetência e abandono, a probabilidade de realizar o Enem, o desempenho no Enem e a probabilidade de ingressar no ensino superior, bem como a qualidade desse curso. Realizou-se decomposição Oaxaca-Blinder para verificar quanto do diferencial dos indicadores educacionais por nível de renda e por qualidade da escola é explicado por características determinadas antes do ingresso no EM e quanto se deve ao EM. Os resultados corroboram outros estudos que apontam que parte relevante da desigualdade observada em determinada etapa de ensino se explica por características determinadas antes dessa etapa. Com o objetivo de verificar se a qualidade do EM pode afetar a desigualdade de resultados educacionais entre alunos de diferentes níveis socioeconômicos, acrescenta-se variável proxy da qualidade da escola de EM, baseada em medida de valor adicionado de proficiência. Tendo em vista o baixo poder explicativo da variável de qualidade da escola de EM, por fim, realiza-se análise especificamente para o grupo de estudantes de baixo nível socioeconômico com o objetivo de verificar se diferentes desempenhos entre estudantes do mesmo nível socioeconômico podem estar associados a características da escola, em particular a sua qualidade.