The stability of the equilibrium outcomes in the admission games induced by stable matching rules (Q2482680): Difference between revisions

The paper deals with a cooperative model known as the school choice model. Under this model and similarly to the College admission model, a set of students and a set of graduate centres are considered. Each student can be enrolled in one centre at most, while all centres have quotas. The students have ordered lists of preferences over the centres, and a given procedure is used by the centres to evaluate the students. The students are ranked by each centre according to the result of its evaluation procedure, and this ranking defines the preference list of the centre, over the students. For the model described above, the author considers a strategic game, called admission game, induced by a revelation mechanism defined by a stable matching rule. Due to the restriction on the preferences of the centres imposed by the market, these game participants behave straightforwardly, while the students' strategies are given by their preferences over the graduate centres. Given these, the paper deals with the following question: if at equilibrium, when students behave strategically, the mechanism continues to yield stable matchings with respect to the true preferences, just as it does when the students reveal their true preferences? Starting from the aforementioned question, the paper shows that the centre-optimal stable matching rule implements the set of stable matchings via the Nash equilibrium concept. For any other stable matching rule, however, the stability of the equilibrium outcomes may be lost. In fact, it is shown that for the student-optimal stable matching rule, the strategic behaviour of the students may lead to outcomes that are not stable under the true preferences. Then, uncertainty is introduced about the matching selected and it is proved that the natural solution concept is that of Nash equilibrium in the strong sense. Another general result shows that any stable matching rule, as well as the random stable matching rule, implements the set of stable matchings via the concept of Nash equilibrium in the strong sense and the student-optimal stable matching under the true preferences is the equilibrium outcome most preferred by the students. Finally, for the case where the students play, in equilibrium, truncations of the true preferences, additionally obtained results imply that the only Nash equilibrium outcome in the strong sense is the student-optimal stable matching rule under the true preferences. Furthermore, a profile of truncations of the true preferences is a strong equilibrium point for every stable matching rule, if and only if it determines a singleton set of stable matchings, whose only element is the student-optimal stable matching under the true preferences. As a consequence, if the students only play truncations of the true preference lists, then any stable matching rule implements the student-optimal stable matching under the true preferences in strong equilibrium in the strong sense and in Nash equilibrium in the strong sense.