Publications by PSE researchers

We study a large market model of dynamic matching with no monetary transfers and a continuum of agents who have to be assigned items at each date. When the social planner can only elicit ordinal agents' preferences, we prove that under a mild regularity assumption, incentive compatible and ordinally efficient allocation rules coincide with spot mechanisms. The latter specify “virtual prices” for items at each date and, for each agent, randomly select a budget of virtual money at the beginning of time. When the social planner can elicit cardinal preferences, we prove that under a similar regularity assumption, incentive compatible and Pareto efficient mechanisms coincide with spot menu of random budgets mechanisms. These are similar to spot mechanisms except that, at the beginning of time, each agent chooses within a menu, a distribution over budget of virtual money.

Almost one third of kidney donation candidates are incompatible (HLA and/or ABO) with their directed recipient. Kidney paired donation allows potential donors to be exchanged and gives access to a compatible kidney transplant. The Bioethics Law of 2011 authorised kidney paired donation in France with reciprocity between 2 incompatible “donor-recipient” pairs. A limited number of transplants have been performed due to a too restricted authorization compared to other European practices. This study presents the perspectives of the new Bioethics Law, enacted in 2021, which increases the authorised practices for kidney paired donation in France. The two simulated evolutions are the increase of the number of pairs involved in a kidney paired donation to 6 (against 2 currently) and the use of a deceased donor as a substitution to one of living donor. Different scenarios are simulated using data from the Agence de la Biomedecine; incompatible pairs registered in the kidney paired donation programme in France between December 2013 and February 2018 (78 incompatible pairs), incompatible transplants performed during the same period (476 incompatible pairs) and characteristics of deceased donors as well as proposals made over this period. Increasing the number of pairs has a limited effect on the number of transplants, which increases from 18 (23% of recipients) in the current system to 25 (32% of recipients) when 6 pairs can be involved. The use of a deceased donor significantly increases the number of transplants to 41 (52% of recipients). This study makes it possible to evaluate the increase in possibilities of kidney transplants by kidney paired donation following the new bioethics law. A working group and an information campaign for professionals and patients will be necessary for its implementation.

To assign teachers to schools, a modified version of the well-known deferred acceptance mechanism has been proposed in the literature and is used in practice. We show that this mechanism fails to be fair and efficient for both teachers and schools. We identify a class of strategy-proof mechanisms that cannot be improved upon in terms of both efficiency and fairness. Using a rich dataset on teachers’ applications in France, we estimate teachers preferences and perform a counterfactual analysis. The results show that these mechanisms perform much better than the modified version of deferred acceptance. For instance, the number of teachers moving from their positions more than triples under our mechanism.

Top trading cycles (TTC) is Pareto efficient and strategy-proof in priority-based matching, but so are other mechanisms including serial dictatorship. We show that TTC minimizes justified envy among all Pareto efficient and strategy-proof mechanisms in one-to-one matching. In many-to-one matching, TTC admits less justified envy than serial dictatorship in an average sense. Empirical evidence from New Orleans OneApp and Boston Public Schools shows that TTC has significantly less justified envy than serial dictatorship.

We propose a new matching algorithm — Unpaired kidney exchange — to tackle the problem of double coincidence of wants without using money. The fundamental idea is that "memory" can serve as a medium of exchange. In a dynamic matching model with heterogeneous agents, we prove that average waiting time under the Unpaired algorithm is close to optimal, substantially less than the standard pairwise and chain exchange algorithms. We evaluate this algorithm using a rich dataset of kidney patients in France. Counterfactual simulations show that the Unpaired algorithm can match 57% of the patients, with an average waiting time of 440 days (state-of-the-art algorithms match about 34% with an average waiting time of 695 days). The optimal algorithm, which is practically infeasible, performs only slightly better: it matches 58% of the patients and leads to an average waiting time of 426 days. The Unpaired algorithm confronts two incentive-related practical challenges. We address those challenges via a modified version of the Unpaired algorithm that employs kidneys from the deceased donors waiting list. It can match 86% of the patients, while reducing the average waiting time to about 155 days.

We analyze the robustness of equilibria in sequential games when there is almost common certainty of payoffs. We show that a generic extensive-form game may have no robust equilibrium behavior, but has at least one robust equilibrium outcome, which is induced by a proper equilibrium in its normal-form representation. Therefore, backward induction leads to a unique robust outcome in a generic perfect-information game. We also discuss close relation between robustness to incomplete information and strategic stability. Finally, we present the implications of our results for the robustness of subgame-perfect implementation.

Cet article s’intéresse à l’algorithme utilisé en France pour affecter les enseignants du second degré dans les établissements scolaires. Partant du constat que la procédure actuelle tend à limiter la mobilité des enseignants, nous proposons une procédure d’affectation alternative qui permettrait d’augmenter de plus de 30 % le mouvement des enseignants titulaires tout en prenant en compte les spécificités des académies les moins attractives. Pour ces académies, nous fournissons un outil de pilotage qui permet de faire des simulations et de tester différentes stratégies de gestion des ressources humaines – augmentation, maintien ou diminution du mouvement dans ces académies.

We study Pareto efficient mechanisms in matching markets when the number of agents is large and individual preferences are randomly drawn from a class of distributions, allowing for both common and idiosyncratic shocks. We provide a broad set of circumstances under which, as the market grows large, all Pareto efficient mechanisms—including top trading cycles (with an arbitrary ownership structure), serial dictatorship (with an arbitrary serial order), and their randomized variants—produce a distribution of agent utilities that in the limit coincides with the utilitarian upper bound. This implies that Pareto efficient mechanisms are uniformly asymptotically payoff equivalent “up to the renaming of agents.” Hence, when the conditions of our model are met, policy makers need not discriminate among Pareto efficient mechanisms based on the aggregate payoff distribution of participants.

L'affectation des enseignants au sein des établissements recouvre une multitude d'enjeux importants : attractivité de la profession, inégalités géographiques, réussite des élèves, etc. L'arbitrage entre ces différentes dimensions s'avère délicat. Assurer une mobilité forte des enseignants peut se faire au prix d'une augmentation des inégalités entre les différentes académies en termes d'expérience des enseignants affectés, et, in fine, au détriment de la réussite des élèves dans les académies les moins attractives. Dès lors, la procédure informatique utilisée afin d'affecter les enseignants du second degré s'avère être un levier important pour arbitrer entre mobilité et égalité entre académies. Dans cet article, nous montrons dans un premier temps que de la procédure actuelle résulte un fort manque de mobilité des enseignants. Nous proposons une procédure d'affectation alternative et quantifions l'impact que pourrait avoir l'adoption de celle-ci par rapport au système actuel. L'une des procédures alternatives que nous proposons permet d'augmenter de plus de 30 % le mouvement des enseignants titulaires tout en prenant en compte les spécificités des académies les moins attractives. Pour ces académies, nous fournissons un outil de pilotage qui permet de faire des simulations et de tester différentes stratégies RH – augmentation, maintien ou diminution du mouvement dans ces académies. Ce travail souligne ainsi l'impact positif important que pourrait avoir une modification du système actuel d'affectation des enseignants en termes de mouvement.

We consider a model of evolution in games in which a revising agent observes the actions of a random number of randomly sampled opponents and then chooses a best response to the distribution of actions in the sample. We provide a condition on the distribution of sample sizes under which an iterated p-dominant equilibrium is almost globally asymptotically stable under these dynamics. We show under an additional condition on the sample size distribution that in supermodular games, an almost globally asymptotically stable state must be an iterated p-dominant equilibrium. Since our selection results are for deterministic dynamics, any selected equilibrium is reached quickly; the long waiting times associated with equilibrium selection in stochastic stability models are absent.

Dans cet article, nous analysons, à l'aune de la théorie de l'appariement, la procédure d'affectation des élèves dans les lycées utilisée en France : la procédure Affelnet. Nous montrons que cette dernière correspond à la procédure d'acceptation différée école-proposant, définie par Gale et Shapley [1962]. Nous en déduisons un certain nombre de propriétés de la procédure Affelnet. En particulier, elle est stable et difficile à manipuler. En revanche, elle possède deux défauts majeurs : d'une part, dans l'ensemble des procédures stables, c'est celle qui conduit à la moins bonne affectation du point de vue des élèves ; d'autre part, la procédure Affelnet ne respecte pas les priorités : si un élève obtient une plus haute priorité dans tous les lycées (grâce à de meilleurs résultats scolaires, par exemple), il peut s'en trouver pénalisé dans son affectation finale. Nous discutons deux composantes additionnelles de la procédure Affelnet : la limite sur le nombre de vœux que les parents peuvent reporter, ainsi que l'existence d'un bonus pour les écoles classées en premier dans la liste de préférences. Finalement, nous pointons l'existence d'une procédure alternative, la procédure d'acceptation différée élève-proposant, qui respecte les priorités et est la plus efficace parmi les procédures stables. Nous simulons l'impact du passage à cette procédure sur le bien-être des élèves, et ce pour différentes caractéristiques de leurs préférences.