School choice: an experimental study

We present an experimental study of three school choice mechanisms. The Boston mechanism is influential in practice, while the two alternative mechanisms, the Gale-Shapley and Top Trading Cycles mechanisms, have superior theoretical properties in terms of incentives and efficiency. Consistent with theory, this study indicates a high preference manipulation rate under the Boston mechanism. As a result, efficiency under Boston is significantly lower than that of the two competing mechanisms in the designed environment. However, contrary to theory, Gale-Shapley outperforms the Top Trading Cycles mechanism and generates the highest efficiency. Our results suggest that replacing the Boston mechanism with the Gale-Shapley mechanism in practice might significantly improve efficiency. ∗Chen: School of Information, University of Michigan, 1075 Beal Avenue, Ann Arbor, MI 48109-2112. Email: [email protected]. Sönmez: Department of Economics, Koç University, Sarıyer, 34450, İstanbul, Turkey. Email: [email protected]. We thank Atila Abdulkadiroğlu, Tim Cason, Vincent Crawford, Valerie Edwards, Carleton Jones, Peter Katuscak, Lutz Kilian, Aradhna Krishna, Parag Pathak, Alvin Roth and seminar participants at 2003 Economic Science Association Meetings (Pittsburgh, PA), 2004 AEA (San Diego, CA), ESSET 2004 (Gerzensee), Games 2004 (Marseille), Stanford SITE 2004, Michigan, Ohio State, and Pittsburgh for helpful discussions, and Kan Takeuchi for excellent research assistance. Chen acknowledges the research support of the National Science Foundation through Grant No. SES-0079001 and Sönmez acknowledges the research support of KoçBank via the KoçBank scholar program and Turkish Academy of Sciences in the framework of the Young Scientist Award Program via grant TS/TÜBA-GEBİP/2002-1-19. Any remaining errors are our own.