Parallel markets in school choice

Abstract

When applying to schools, students often submit applications to distinct school systems that operate independently, which leads to waste and distortions of stability due to miscoordination. To alleviate this issue, Manjunath and Turhan (2016) introduce the Iterative Deferred Acceptance mechanism (IDA); however, this mechanism is not strategy-proof. We design an experiment to compare the performance of this mechanism under parallel markets (DecDA2) to the classic Deferred Acceptance mechanism with both divided (DecDA) and unified markets (DA). Consistent with the theory, we find that both stability and efficiency are highest under DA, intermediate under DecDA2, and lowest under DecDA. We observe that some subjects use strategic reporting when predicted, leading to improved efficiency for all participants of the market. Our findings cast doubt on whether strategy-proofness should be perceived as a universal constraint to market mechanisms.