Priorities vs. Precedence: Theory and Evidence from Boston

School Assignment, January 2013

Boston’s school choice plan has a 50/50 split of school seats into a walk half and open half. This fea­ture emerged as a com­pro­mise when the plan was for­mu­lat­ed in 1999. When a choice plan has a slot-spe­cif­ic pri­or­i­ties like this, the prece­dence order, i.e. the order in which seats are deplet­ed by appli­cants with spe­cif­ic claims, is a lever to achieve dis­tri­b­u­tion­al goals that has effects com­pa­ra­ble to pri­or­i­ties under the deferred accep­tance algo­rithm. While Boston gives pri­or­i­ty to neigh­bor­hood appli­cants in half of the seats at each school, the intend­ed effect of this pol­i­cy is almost ful­ly lost because of the prece­dence order of the seats; its out­come is near­ly equiv­a­lent to that of a mech­a­nism with­out any neigh­bor­hood pri­or­i­ty. This fact shows how the prece­dence order can under­mine the intend­ed role of priorities.