Dois ensaios sobre a melhoria da qualidade do ensino primário no Brasil: o efeito sobre o desempenho escolar a partir da projeção da infraestrutura escolar e o impacto do Programa Mais Educação

Abstract

The quality of Brazilian public education has been around the same level since the 2000s. This stagnation means that the educational gap between Brazil and the developed countries, and even other Latin American countries, is increasing. In addition, the country coexists with large regional socioeconomic disparities and provision of educational resources, which in turn may be associated with the difference in educational quality. Thus, the Brazilian government has been seeking different ways to improve student performance, with emphasis, in recent years, in policies of full time education. Given this scenario, this thesis is divided in two parts that aims at Brazilian public primary education. The first is to assess the level of efficiency, using a dynamic model, taking into account the differences in the socioeconomic levels of its students. This model includes an infrastructure index as a variable that links each period of time. We used three model’s specifications that differ in the usage of the socioeconomic variable. The results showed that the efficiency distributions of these models are not statistically equal, but the general conclusions are convergent. There was almost no evolution in school efficiency between 2007 and 2015, but our model indicate a possible way to improvement the efficiency by investing in school infrastructure. The second part of the thesis evaluates the Programa Mais Educação (PME) created by the Brazilian Federal Government with the aim of improving education at the fundamental level. This policy seeks to implement the model of school full-time, with at least seven hours a day, through the implementation of extra socio-educational activities in the public school second-shift. Our objective is to estimate the impact of the program on the average academic performance in Portuguese, mathematics and approval rate, in addition to the new violence index. For this, we propose the use of the augmented weighted propensity score method, which seeks to establish an adequate control group to compare with the group of treated schools, since the program is aimed mainly at the most vulnerable schools. Then, we estimated the impact of the program’s performance between 2007 and 2017 through the differences-in-differences method. The evidence obtained points to negative effects between 2009 and 2011, no statistical significance in 2013 and positive effects in 2015 and 2017. The cost-effectiveness of the program is on average 0.2 points in Portuguese and Mathematics and 0.4 % - 0.8 % in the approval rate for every ten thousand reais invested in the school. Finally, the program has a positive effect on school efficiency in all models.