Economics serving society

O. Tercieux – Are allocation algorithms worth saving? (June 2018)

JPEG - 416.2 kb

Julien Combe (UCL), Victor Hiller (University of Paris II), Olivier Tercieux (PSE, CNRS) and Camille Terrier (MIT) have just published a new Terra Nova paper, “Are the allocation algorithms worth saving? Affelnet, teaching deployments and Parcoursup” (6 June 2018).

Summary

In recent years, the use of algorithms as decision-making tools has become increasingly widespread in public administration. This is especially true in the national education system, where decisions relating to the management of human resources and the allocation of students to schools are largely made using algorithms.
These allocation algorithms are currently at the centre of a public debate. Sometimes accused of being at the origin of “inhumane” decisions, or out of touch with prevailing political preferences, they are the object of some virulent criticism. In this Note, we ask about the origin of several shortcomings observed in the past few years, and about the real or supposed role played by algorithms in the problems. Are allocation and assignment algorithms worth saving and if yes, how?
To answer these questions, we describe in detail and, we hope, in instructive fashion, how some of the allocation algorithms work. Our aim is to familiarise readers with the procedures that have such important consequences for our lives and in so doing, to reveal their normative content. We seek to explain that, far from being a simple technical tool, allocation algorithms can be a powerful public policy instrument. That is why algorithms must be chosen very carefully, and in the light of clear and precise evaluation of their consequences.

Press reviews