School Assignment By Match Quality
Research Seminars: Virtual Market Design SeminarParental school choice is a central education reform tool. Yet various policy objectives cannot be addressed in the standard model that guides the design of centralized admissions in school choice programs. The paper presented in this Virtual Market Design Seminar introduces student-school specific match quality and formalizes such policy objectives as maximizing aggregate match quality subject to stability constraints. The authors characterize subsets of stable assignments through admissible schools, and maximize match quality within these subsets with a minimum-cost flow solution. In comparison to the widely used Deferred Acceptance with a random tie-breaking algorithm, match quality optimization in New York City public school assignment reduces average traveling distance for high school students by about 1 mile, increases estimated Math and English standardized test scores for middle school applicants by about 3-5% of a standard deviation, and assigns around 7 percentage point more applicants to one of their two most preferred schools.