Dois ensaios sobre os determinantes da desigualdade educacional brasileira a partir de dados longitudinais

Abstract

The aim of this thesis is to investigate the most important determinants of the Brazilian education inequality, due to the relationship of educational attainment and individual social and economic success. The data set used comes from the GERES Project – Educational Generation 2005 – in which a sample of children was tracked during the first four years of elementary school. Following the Theory of the Life Cycle of Skills Formation (Cunha and Heckman, 2007) this study analyzes the role of school and preschool in the children’s proficiency achievement. This study also investigates the relationship among social origin and children’s proficiency achievement, based on the Theory of Equality of Opportunity (Roemer, 1998). This thesis is composed by two independent essays which are organized as chapters, besides an introductory chapter and a chapter of conclusion. The first essay investigates the schools’ role in the generation/ reduction of inequalities of students’ skills. The central assumption is that the initial level of skills inequality limits the effects of schools’ intervention in skills formation. Specifically, this study intends to (i) measure the evolution of inequality and the effect of schools on the opportunities in skills formation and (ii) find empirical evidence for Brazil' dynamic process of learning. The first goal is performed by comparing the evolution of learning for groups with different socioeconomic characteristics and by the decomposition of inequality of proficiency using the Theil-T index-, while the second point applies techniques of dynamic panel, based on the model of technology of skill formation (Cunha and Heckman, 2007). The results indicate a positive relationship between skill level and socioeconomic background. They also show evidence of the dynamic effects of learning, highlighting the role of previous skill level. These findings shed some light on the limited role of the regular schooling reducing inequalities of skills and the importance of investments in early childhood education. The second essay estimates the preschool effect on the child’s proficiency level of students from the first four years of elementary school. Following Cunha & Heckman (2007) and Todd & Wolpin (2003), the preschool effect is estimated through an educational production function that incorporates a dynamic process of learning. The empirical analysis shows evidence of cumulative effect of inputs and the presence of individual heterogeneity. In the case of Mathematics the direct effect of preschool on child’s proficiency was positive and significant, while in Portuguese, there is no direct effect. Furthermore, the indirect effect shown by the coefficient of initial proficiency is positive and increases over the years for Math. On the other hand, the students with higher skills in language at the beginning of elementary school expand their differences in relation to less skilled students, but this effect inverts over the years, indicating that the inequality of skills in Portuguese can be reduced during the scholar period. Finally, the overall results point that the educational inequality is partially explained by the difference in the school resources. However, the initial level of children’s skills limits the role of schools in reducing the inequality of abilities. Therefore, investing in the young children - like an increasing in preschool attendance - can contribute to reduce the educational inequality.