Evaluation of the French Affelnet procedure
Victor Hiller et Olivier Tercieux
Every year, in every academy, education authorities have an impact on the lives of thousands of students in hundreds of lycées, juggling the preferences of the parents and the positions of priority allocated to the students (1). The complexity of the allocation problem renders it is impossible to resolve without the help of algorithms generated by computers. Consequently, the choice of algorithm is crucial: it determines the final allocation, which determines the essential elements for the public decision makers, such as meeting parents’ desires, and social and academic diversity in schools. The study of the properties of these algorithms has received particular attention in economics since the appearance of the matching theory developed by, among others, the Nobel prize winners Alvin Roth and Lloyd Shapley.
In this article, Victor Hiller and Olivier Tercieux use the tools of matching theory to analyse the strengths and weaknesses of the algorithm used in France, the Affelnet procedure. They show that this procedure is far from free of criticism. First, it does not satisfy the parents wishes and leads to some of them wanting to trade their children’s allocations. More problematically, the allocation achieved through the Affelnet algorithm does not respect the criteria of priorities laid down by the education authorities. In fact, using this procedure, it is possible that a student whose rank is enhanced because, for example, he or she has obtained good grades, or a scholarship, can, because of that gain, end up in what is considered a lesser lycée. More than 50 years ago, Gale and Shapley proposed an alternative algorithm. The allocation obtained through the use of their algorithm reduces the incidence of parents wishing to trade their children’s allocations and more importantly, renders impossible the disadvantaging of students because they have met a criterion that the education authorities have accorded high priority. This alternative algorithm is currently used with great success in many parts of the world and this article offers some arguments in favour of its adoption in France.
(1) The level of “priority” is determined by a combination of several criteria, including the proximity of the school to the place of residence, the pupil’s grades and the award of scholarships.
Original title of the article : “Choix d’écoles en France. Une évaluation de la procédure Affelnet”
Published in : Revue Economique, 65, pp 619-656 - mai 2014
Available at : http://hal-pjse.archives-ouvertes.fr/PJSE/halshs-00986984/fr/
© AntonioDiaz - Fotolia.com