Gaming the Boston School Choice Mechanism in Beijing

Résumé

The Boston mechanism has been criticized for its poor incentive and welfare performance compared with the deferred-acceptance mechanism (DA). Using school choice data from Beijing where the Boston mechanism without school priorities is adopted, I investigate parents' behavior and allow for possible mistakes. Evidence shows that parents are overcautious because they play ``safe'' strategies too often. There is no evidence that wealthier/more-educated parents are more adept at strategizing. If others behave as indicated in the data, an average naive parent who always reports her true preferences experiences a utility loss in switching from the Boston to the DA mechanism (equivalent to random serial dictatorship in this setting), corresponding to an 8% increase in the distance from home to school or substituting a 13% chance at the best school with an equal chance at the second-best school. She has a 27% (55%) chance of being better (worse) off. If parents are instead either sophisticated (they always play a best response against others) or naive, the results are mixed: under DA, naive parents enjoy a utility gain on average when less than 80% of the population is naive, while still about 42% are worse off and only 39% are better off. Sophisticated parents always lose more.

Mots-clés

Boston Immediate-Acceptance Mechanism; Gale-Shapley Deferred-Acceptance Mechanism; School Choice; Bayesian Nash Equilibrium; Strategy-Proofness; Moment Inequalities; Maximin Preferences;