School Choice with Multiple Priorities

Minoru Kitahara,Yasunori Okumura

DOI: https://doi.org/10.48550/arXiv.2308.04780

2024-10-17

Abstract:This study considers a model where schools may have multiple priority orders on students, which may be inconsistent with each other. For example, in school choice systems, since the sibling priority and the walk zone priority coexist, the priority orders based on them would be conflicting. We introduce a weaker fairness notion called M-fairness to examine such markets. Further, we focus on a more specific situation where all schools have only two priority orders, and for a certain group of students, a priority order of each school is an improvement of the other priority order of the school. An illustrative example is the school choice matching market with a priority-based affirmative action policy. We introduce a mechanism that utilizes the efficiency adjusted deferred acceptance algorithm and show that the mechanism satisfies properties called responsiveness to improvements and improved-group optimally M-stability, which is stronger than student optimally M-stability.

Theoretical Economics

What problem does this paper attempt to address?

This paper attempts to address the fairness and efficiency issues in the school choice problem under multiple priority orders. Specifically, when there are multiple priority orders of schools for students and these priorities may conflict with each other, how to design a mechanism to achieve more effective matching while ensuring a certain level of fairness. The paper introduces a weak fairness concept called M - fairness and proposes a new mechanism using the Efficiency - Adjusted Deferred Acceptance algorithm (EADA algorithm), which can satisfy improved responsiveness and improved group - optimal M - stability.

### Main contributions of the paper:
1. **Introduction of M - fairness**: The paper defines a new fairness concept - M - fairness, which allows the violation of a school's priority order in some cases as long as this violation can be rationalized by another priority order of the school. This provides a flexible method for dealing with multi - priority systems.
2. **Design of a new mechanism**: The paper proposes a new mechanism. This mechanism first constructs a single priority order from multiple priority orders and then uses the EADA algorithm for matching. This mechanism can improve the efficiency of matching while maintaining a certain level of fairness.
3. **Improved responsiveness and group - optimal M - stability**: The paper proves that the proposed mechanism can not only achieve student - optimal M - stable matching but also achieve improved group - optimal M - stable matching. This means that in some specific groups, the mechanism can generate matching results that are not Pareto - dominated by any other M - stable matching.
4. **Application examples**: The paper shows the application of this mechanism in the actual school choice system through specific examples, especially in cases involving priorities based on affirmative action policies, such as providing more educational opportunities for minority students.

### Explanation of formulas and symbols:
- **Priority order**: Denote a priority order of school \(s\) for students by \(\succ_s\), where \((i, j)\in\succ_s\) means that student \(i\) has a higher priority than student \(j\).
- **M - fairness**: A matching \(\mu\) is called M - fair if for any students \(i\) and \(j\) and school \(s\), if \(s\succ_i\mu(i)\) and \((i, j)\in\succ_s\), then there exists \(\succ'_s\in\succ_s\) such that \((j, i)\in\succ'_s\).
- **EADA algorithm**: An improved deferred acceptance algorithm for finding stable matchings in multi - priority systems.

### Conclusion:
By introducing M - fairness and designing a new matching mechanism, the paper provides an effective method for solving the school choice problem under multiple priority orders. This method not only improves the efficiency of matching but also ensures fairness to a certain extent, especially for specific groups (such as minority students).