This thesis addresses issues that do not have a clear cut consensus in the economics of education literature. We begin by trying to identify regions of returns to schooling in the UK using an approach developed in Manski (1997) and Manski and Pepper (2000). Non-parametric bound analysis has the advantage to rely on relatively weak and somewhat testable assumptions. Applying this approach on two different large data sets we establish an identification region for the return to education and compare the results with some traditional parametric approaches commonly employed in the literature. The estimates show that the returns to education computed through weaker assumptions are smaller than (and in some cases well below) some of the point estimates usually reported in the literature. In the second part of this study, we investigate the effects of a funding education reform implemented in Brazil in 1998 (FUNDEF/FUNDEB), which largely increased educational expenditures across the country. The identification strategy comes from the fact that the exposure and intensity of the reform varied across municipalities and years. First, we analyse whether the redistributive effect of the reform reduce inequality in terms of schools resources between poor and rich regions. Second, we verify whether an increase in the availability of resources to schools brought by the reform translated into higher students’ performance. The results suggest a decrease in the inequality of school resources within and to some extent across regions. However, there is no evidence of effects on students’ test scores. We also assess the effects of the funding reform on somewhat long term student´s outcomes. The results suggest that the reform leads to an increase in education attainment for individuals who went to schools in the most affected regions.
|1 Apr 2015
|Unpublished - 2015