Resumen Paper






Rocio Valdebenito1, Esteban López Ochoa2  y Patricio  Aroca3.

1. Universidad Adolfo   Ibañez,
2. Universidad Adolfo   Ibañez,
3. Universidad Adolfo   Ibañez,



Commuting effects on performance of workers and students has been greatly documented as producing in most cases negative effects. Most school choice literature has disregarded commuting effects given that in many countries school attendance boundaries are strictly enforced, and in cases of open school access, school generally provide suitable transportation as a way to alleviate the problem. In the case of Chile, given the large-scale school reform in the early 80s that introduced the voucher system, allowed families to choose any school within the city. The main assumption claimed by officials and program authorities, was that the new system would create a competitive market in which parents pay for quality and only the best schools will survive in the long-run. This set up was translated into a race of families for higher quality schools, creating incentives for long distance commutes that would assure high quality.

This paper aims to research to what extent the previous claims (usually held in the media) are supported by data, and if there is evidence of long distance commuting, we estimate the effect that it would have had in student’s performance. We rely on counterfactual techniques of quasi-experimental non-parametric design, to build two particular subsets of data (treated and control) referring to students with similar characteristics including the location of origin, but with the difference that control stu- dents attend to the neighborhood schools, while treated students do not.

We use a unique data set which pulls together students’ records from several points in time where performance was assessed. In particular, we take sample of students in the capital region of Santiago that took the standardized test during fourth grade in 2002, and then we find them later on in the last year of high school when they took the national standardized test for college admissions in 2010. Then we geocoded all students in this sample with valid records to obtain their residency location. Together with the location of the school they use to attend, we were able to construct Euclidean and road distances variables to test our hypothesis.

We believe this type of study is important for the current education policy agenda. In 2017, the voucher system was derogated instead of a new system where previously voucher schools have to choose to either become fully private or fully public. Additionally, there are discussions proposing   to ban selection at the school level to replace it with a new student selection system that will place students randomly based on some basic criteria. These current and upcoming changes have definitely implications in terms of the distances students must travel to their school along with the potential effects that this brings.


Participación en sesión: Economía regional y urbana.