Abstract
We study school switches in which students who have already been assigned seats request reassignment. Unlike in regular school choice, stability need not be legally binding in such programs, since students’ priority-based claims have already been satisfied upstream. Efficiency then naturally becomes the primary design objective. To quantify the potential legal efficiency gains, we compare Deferred Acceptance (DA) with Top Trading Cycles (TTC) and an efficiency-adjusted DA (EADAM) using Swedish data. TTC and EADAM triple switching rates compared to DA. This difference is driven by the absence of slack in effective capacity and shrinks rapidly when introducing additional seats.
•Priority-based stability need not be legally binding in school-switch programs.•We compare DA, TTC, and EADAM using data from six Swedish school districts.•TTC and EADAM triple switching rates relative to DA.•Tight effective capacity drives the efficiency gap.•The gap shrinks rapidly when additional school seats are introduced.