Multi-District School Choice: Playing on Several Fields

Yannai A. Gonczarowski,Michael Yin,Shirley Zhang
2024-03-07
Abstract:We extend the seminal model of Pathak and Sönmez (2008) to a setting with multiple school districts, each running its own separate centralized match, and focus on the case of two districts. In our setting, in addition to each student being either sincere or sophisticated, she is also either constrained - able to apply only to schools within her own district of residence - or unconstrained - able to choose any single district within which to apply. We show that several key results from Pathak and Sönmez (2008) qualitatively flip: A sophisticated student may prefer for a sincere student to become sophisticated, and a sophisticated student may prefer for her own district to use Deferred Acceptance over the Boston Mechanism, irrespective of the mechanism used by the other district. We furthermore investigate the preferences of students over the constraint levels of other students. Many of these phenomena appear abundantly in large random markets.
Computer Science and Game Theory,Theoretical Economics
What problem does this paper attempt to address?
### What problems does this paper attempt to solve? This paper attempts to extend the seminal model of Pathak and Sönmez (2008) to cover the situation of multiple school districts (each district independently operates its own centralized matching mechanism), with a particular focus on the two - district setting. The authors study the impact of students' strategic choices and district choices on student - school matching results in a multi - district environment. Specifically, the paper explores the following aspects: 1. **Student types and behaviors**: - Students can be "sincere" (i.e., always truthfully report their preferences for schools) or "sophisticated" (i.e., be able to strategically submit preference lists). - Students can also be "restricted" (only able to apply to schools in their own district) or "unrestricted" (able to choose any district to apply to). 2. **Choice of matching mechanisms**: - The paper compares two common matching mechanisms: the Boston mechanism (BM) and the deferred acceptance mechanism (DA). BM is not strategy - proof, while DA is strategy - proof. - In the context of a single district, Pathak and Sönmez (2008) found that sophisticated students prefer BM to DA because BM gives them an advantage. However, in a multi - district environment, these conclusions may change. 3. **New phenomena in a multi - district environment**: - The research finds that in a multi - district environment, sophisticated students may prefer their district to use DA rather than BM, regardless of which mechanism the other district uses. - Sophisticated students may prefer some sincere students to become sophisticated, which is the opposite of the conclusion in a single - district environment. - The authors also study students' preferences for other students' constraint types, for example, whether an unrestricted student prefers another student to also become unrestricted, or whether a restricted student prefers students in another district to become restricted. 4. **Large - market analysis**: - By analyzing large - scale random markets, the authors show that the above phenomena are widespread in large markets, emphasizing the important impact of district choice on students' strategies and matching results. ### Summary The core problem of this paper is to explore how different types of students becoming sophisticated or changing their district choices in a multi - district environment affect the overall matching results, and how these changes challenge the previous conclusions drawn in a single - district environment. Through theoretical models and large - market analysis, the authors reveal new complexities and potential policy implications in a multi - district environment.