Reserve Design: Unintended Consequences and the Demise of Boston's Walk Zones

Admissions policies often use reserves to grant certain applicants higher priority for some (but not all) available seats. Boston’s school choice system, for example, reserved half of each school’s seats for local neighborhood applicants while leaving the other half for open competition. This paper shows that in the presence of reserves, the effect of the precedence order (i.e., the order in which different types of seats are filled) on distributional objectives is comparable to the effect of adjusting reserve sizes. Either lowering the precedence order positions of reserve seats at a school or increasing the number of reserve seats weakly increases reserve-group assignment at that school. Using data from Boston, we show that reserve and precedence adjustments have similar quantitative effects. Our results illustrate that policies about precedence, heretofore underexplored, are inseparable from other aspects of admissions policy. Moreover, our findings explain the puzzling empirical fact that despite careful attention to the importance of neighborhood priority, Boston’s implementation of its 50-50 reserve–open seat split was nearly identical to the outcome of a counterfactual system without any reserves. Transparency about these issues—in particular, how precedence unintentionally undermined the intended admissions policy—led to the elimination of Boston’s walk zones.