The Design of Teacher Assignment: Theory and Evidence

Abstract

In several countries, teachers' assignment to schools is managed by a central administration. One of the objectives of the reassignment process is to make sure that teachers obtain an assignment which they weakly prefer to their current position. The Deferred-Acceptance mechanism (DA) proposed Gale and Shapley (1962) fails to satisfy this constraint. As a solution, a variation on this mechanism has been proposed in the literature and used in practice - as for the assignment of French teachers to schools. In this paper, we show that this mechanism yields assignments that can be improved in terms of both efficiency and "fairness". For each of the two efficiency notions considered in the literature (two-sided or one-sided), we identify the class of mechanisms which cannot be improved upon in terms of efficiency and fairness. Additionally, for two-sided efficiency, we show that a unique mechanism in the associated class is strategy-proof. For one-sided efficiency, no mechanism in the class is strategy-proof. Finally, using a rich dataset on teachers' applications to transfer, we empirically assess the extent of potential efficiency and fairness gains associated with the adoption of our mechanisms. These empirical results confirm the poor performance of (the variation on) the DA mechanism, and the significant improvement brought by our proposed mechanisms in terms of both efficiency and fairness. Joint with Julien Combe (PSE) and Camille Terrier (LSE)