paper Efeito formação docente

Abstract (full paper in Revista Brasileira de Economia: http://bibliotecadigital.fgv.br/ojs/index.php/rbe/article/view/76649)

This paper applies a value-added model to estimate the effects of 4^th and 5^th grades teachers’ education level (with and without higher education) and teachers’ major (Pedagogy, Math, Portuguese, and others) on the proficiency of students from Brazilian public schools. The identification strategy relies on between-class within-school variation of teachers’ majors, addressing teacher’s education level endogeneity, which could render standard Ordinary Least Squares (OLS) estimators inconsistent. The paper uses a unique longitudinal database drawn from multiple data sources that includes variables such as: (i) 2015 5^th grade students’ proficiency in Prova Brasil; (ii) 2015 5^th grade students’ proficiency in ANA 2013 (when they were in 3^rd grade); (iii) a set of students’ socioeconomic characteristics, (iv) education inputs (e.g. class size and school infrastructure) used for the 2015 5^th graders throughout the 4^th grade and 5^th grade (2014 and 2015 academic years, respectively). Our findings show that the improvement on proficiency of students whose teachers have tertiary education is not different from that obtained by pupils from teachers with no tertiary education in Math and Reading. However results are heterogeneous among different majors in Math. The estimated improvement on Math proficiency of students whose 4^th and 5^th grades teachers have a major in Math or Portuguese Language revealed to be 7% of σ or 2.6% of σ, respectively, higher than that obtained by pupils from teachers with no tertiary education. On the other hand, according to our results, teachers with major in Pedagogy do not impact their students' proficiency differently from teachers with no tertiary education.