Bonilla-Angel, Juan DiegoContracting out public schools to private institutions is an instrument for reforming public education as it may facilitate academic innovation and improve student academic performance through higher school accountability and autonomy. The degree of autonomy that different providers have may vary substantially depending on the contractual and institutional arrangements they are subject to. In principle, contractual differences should generate different sets of incentives for providers that may ultimately affect student academic performance. One can expect, for example, that programs with limited achievement accountability rules might invest sub-optimally in resources aiming at improving the academic performance of their students. In this dissertation, I evaluate short- and longer-run achievement effects of the \emph{Colegios en Concesi\'on} (CEC) program, a large-scale initiative implemented in 2000 in Bogota, Colombia, which contracted out the administration of some traditional public schools (TPS) to reputed, not-for-profit private schools and universities. This program allows participating schools to operate outside public schools' collective bargaining provisions in return for being accountable, among other things, for the academic performance of their students in the ICFES test, a high-stakes college entry national standardized test. The major empirical challenge in studies of alternative school models is selection bias. Students who attend CEC schools may differ in a number of ways from public school students. To overcome potential selection bias of CEC attendance, I exploit variation in distance from a student's residence to the closest CEC institution as an instrument for CEC attendance. While distance may in theory be correlated with unobservable characteristics of students, I demonstrate using a variety of empirical strategies that this instrument is conditionally exogenous of unobserved determinants of academic achievement. I first evaluate the effects of attending a CEC school on ICFES test scores. Instrumental variables results indicate that CEC students exhibit important and significant gains in test scores on the ICFES test. That is, the two-stage least squares estimates obtained indicate that CEC students score 0.6 and 0.2 standard deviations higher in math and verbal tests, respectively, relative to TPS students. I provide evidence that the positive test score results of CEC attendance are not driven by unintended strategic responses by CEC schools such as excluding low-performing students from the pool of test-takers or via test specialization in the curriculum, or by significant differences in education inputs such as teachers' education, student-teacher ratios, or expenditures per student. I also provide suggestive auxiliary evidence that the estimated results are a consequence of an institutional arrangement that makes CEC schools accountable for the academic performance of their students. I also evaluate whether attending a CEC school translates into longer-run gains in potentially more meaningful outcomes such as increasing the probability of investing in post-secondary schooling, attending a more selective tertiary institution, or being admitted in high-return academic programs. The results on college attendance indicate that CEC students exhibit a significantly higher probability of attending a higher education institution and to attend a vocational program relative to TPS students. Moreover, CEC students have a slightly higher probability of attending a selective public institution and are not more likely to drop out from college relative to TPS students. The overall results provide compelling evidence that the contractual arrangement that defines the operation of CEC schools are successful at improving the academic performance of their students relative to TPS.DissertationEducation policy