All about priorities? Less school choice with bad schools

Parents worry a lot about which school might be best for their children. But can they really choose where to send their children to school? Can they only choose if they are willing to pay for a private school or by moving into a better neighborhood where public schools are better? The importance of allowing for school choice has been greatly emphasized in the literature and in the policy debate. In the past, children were systematically assigned to their neighborhood school. A large fraction of OECD countries has expanded choice in various ways in the last two decades.

As described in the Friedman Foundation for Educational Choice’s website: 

“School Choice is a commonsense idea that gives parents the power and freedom to choose their child’s education […] School Choice is a public policy that allows parent/guardian to choose a school regardless of residence and location.”

Many people prefer the same schools for their children. The main problem when introducing school choice is how to deal with this issue and defining the alternatives for applicants rejected from overdemanded schools. In most cases local authorities define priority rules that determine the applicants that shall be prioritized in case of overdemand. Priority is often given to families that have a sibling in the school, that live close to the school, or that have particular socioeconomic circumstances. The market design literature on school choice has extensively dealt with the mechanisms defining how applicants rejected from their first choice are reallocated.

In their Barcelona Graduate School of Economics working paper (May 2012), Caterina Calsamiglia and Antonio Miralles study the most common mechanisms, Gale Shapley Deferred Acceptance (DA) and the Boston mechanism (BOS), with priority rules defined to break ties for overdemanded schools. The authors show that the resulting allocation may be close to that defined by the tie breaking rules rather than families’ preferences, provided there exist bad schools. In most cities where school choice is implemented there are some so-called failing schools, schools that are explicitly considered as bad by local authorities. In the US the requirement of the federal No Child Left Behind Public Choice Program requires that local school districts allow students in academically unacceptable schools (F-rated schools) to transfer to higher performing, non-failing schools in the district (if there is capacity available).

Under DA, the presence of bad schools that all families rank below or equal to their neighborhood school, leads to most children being allocated to the school in their neighborhood, as if there was no school choice. In the case of BOS, the authors show that, under some conditions there exists a quality level for the bad school for which all families apply and get accepted to the school where they have highest priority, that is, the neighborhood school. In other words, the authors show that under the presence of bad schools in the system, school choice with priorities to resolve overdemand does not guarantee that the final allocation is shaped by families’ preferences. Instead, priority rules, which in principle should only help break ties, mostly determine where children go to school. In particular, priorities fully determine the allocation when the number of students in the neighborhood of good schools is larger or equal than the capacities of those schools. The authors also demonstrate that choice is amplified only when priorities are such that they explicitly discriminate those families who live in the bad school’s proximity, by giving them lowest priority to all good schools. But that clearly implies assigning those living in the bad school neighborhood to the bad school.

There is recent empirical support of this result regarding BOS: Calsamiglia and Güell (2013) exploit a change in neighborhood design in Barcelona, where the Boston mechanism is used, to show that a lot of families do apply for their neighborhood school because it is where the children are most likely to be accepted.

The authors emphasize that the two most debated and used mechanisms in the literature and in the policy debate may both be very limited in their capacity to allocate children according to their preferences whenever there are priority rules to break ties. Under certain conditions preferences play no role in determining the final assignment. Therefore, school choice does not automatically provide choice. In the end, the only way to choose might be by choosing to live in a different neighborhood.


Abdulkadiroglu, A. and Sönmez, T. (2003) “School Choice: A Mechanism Design ApproachAmerican Economic Review, 93 3.

Calsamiglia, C. and Güell, M. (2013) “The Illusion of School Choice: Evidence from Barcelona” mimeo.

1The seminal work of Abdulkadiroglu and Sönmez (2003) introduces and describes these two mechanisms in detail. This work constitutes one of the main applications of the recently awarded Nobel Prize in 2012 to Alvin Roth and Lloyd Shapley for “Stable Allocations and the Practice of Market Design.”