This dissertation broadly focuses on how to improve equity in education. The first chapter focuses on education at the primary level and analyzes whether progressive vouchers in education can serve as a tool to decrease socioeconomic stratification at the school level and increase educational outcomes for low-income students. I use the Chilean setting, where a universal voucher system has been in place for over three decades, and analyze the impact of a major reform were voucher amounts were increased by 50 percent for students in the lowest 40 percent of the income distribution. Progressive vouchers were implemented in Chile to help low-income students benefit from school choice; increasing the revenues that schools receive for serving low-income students and lowering the relative prices of private voucher schools for eligible parents. I use a national dataset to implement a regression discontinuity design exploiting that eligibility is a discontinuous function of a socioeconomic ranking. Results reject that eligible students chose schools with higher test scores or average SES, and that they are doing better than non-eligible students in math and language test scores. Findings, I argue, are partly a consequence of the multiple barriers that low-income students face when choosing a school, including lack of information, the complexity associated with evaluating a substantial number of options, and issues of social belonging that prevent them from attending better performing schools.
The second chapter focuses on education at the tertiary level and analyses whether loans for higher education can help to increase tertiary education for low-income low-performing students. I use data from Chile and exploit the fact that access to loans for universities and technical institutions is a discontinuous function of students’ academic performance. The latter allows me to implement a regression discontinuity design to look at the causal impact of different types of loans on higher education access, persistence and graduation. Results show that loans for universities induce low-performing students away from technical institutions and towards higher quality university alternatives, where they have little chances of succeeding. This increases the total amount of time and money that students spend without substantially increasing, or even decreasing, their graduation rates and expected incomes. Loans for technical institutions are better in that they keep students away from alternatives that are too expensive or academically demanding. Results point to the unintended costs of offering university loans to low-performing students, steaming from a potential mismatch between low-performing students and higher quality university alternatives.
The third chapter, joint with Juan Matta, analyzes the role of social interaction in higher education choices. In particular, we analyze spillovers from older to younger siblings in the choice of college and major. We use data from Chile and exploit discontinuous admission rules generated by Chile’s centralized system of admission to postsecondary education. Our findings reveal strong sibling spillovers in the choice of major/institutions. Having an older sibling enrolling in a given major within an institution, as opposed to just applying, increases by 87% the likelihood of enrolling in that same major/institution combination, and it increases by 51% the probability of enrolling in any major within that same institution. An analysis of potential mechanisms suggests that spillovers are present even when siblings are far apart in age and are unlikely to attend college together, and even in cases where they are likely to be well informed about the program. Results provide an explanation as to why low-income students may be underrepresented in some high quality educational alternatives.
| Identifer | oai:union.ndltd.org:columbia.edu/oai:academiccommons.columbia.edu:10.7916/d8-km0d-2949 |
| Date | January 2019 |
| Creators | Aguirre, Josefa |
| Source Sets | Columbia University |
| Language | English |
| Detected Language | English |
| Type | Theses |
Page generated in 0.0021 seconds