Making School Place Allocation More Efficient and Equitable
ResearchZEW Study Reveals: School Choice Better Left to Parents and Students Themselves
Algorithms are employed in Germany to allocate children and teenagers to schools. However, the mechanisms commonly used so far perform poorly in terms of efficiency and equality, as demonstrated by a study conducted by researchers from ZEW Mannheim and Queen’s University Belfast. An approach that prioritises the preferences of parents and students proves to be far more effective for allocating school places to students.
The most commonly used mechanisms for allocating school places are the deferred acceptance (DA) mechanism and the top-trading cycles (TTC) mechanism. These mechanisms match the preferences of children and parents with the allocation criteria of schools, and then assign school places according to predefined rules. Unfortunately, these mechanisms exhibit low efficiency and often result in unfair outcomes. According to the ZEW study, which has since been published in the renowned journal Games and Economic Behaviour, a rank-minimizing mechanism (RM) performs better: “The rank-minimizing mechanism ensures that students end up in schools they would prefer attending over those assigned to them by the two prevailing algorithms,” explains Thilo Klein, a researcher in ZEW’s “Market Design” Unit and professor at Pforzheim University. On average, students are assigned places at schools ranked third on their personal list of preferred school programmes, whereas with the other mechanisms, it would only be schools ranked ninth or twelfth.
Greater efficiency through RM mechanism based on children’s and parents’ preferences
This is shown by ZEW and Queen’s University using the example of Hungary, a country where the allocation of school places is currently done using the DA mechanism. For their empirical analysis, the authors utilise Hungarian data from 2015 on the admission of students to secondary schools. In total, they consider the preferences of 10,131 students and the allocation criteria of 244 school programmes in Budapest. The students indicate the order in which they would consider attending the schools. When the RM mechanism is applied to this data, the average student is assigned a school that aligns more closely with their preferences.
More equality among students
Furthermore, the RM mechanism promotes greater equality, as it assigns less than 2 per cent of students to a school ranked 10 or lower. In contrast, TTC and DA allocate a larger proportion, with 16 and 41 per cent respectively, to such schools. “It would, therefore, be significantly more efficient and fair to put the wishes of parents and children first, as the RM mechanism does, and to take the preferences of the schools into account as a secondary factor,” emphasises market design expert Thilo Klein. To achieve this in practice, schools could apply broader criteria such as the catchment area of the school or whole-number grades, instead of fine-grained factors like the distance to school or the average grade with decimal places. By incorporating these broader criteria among schools, there is more flexibility to accommodate the preferences of parents.
