Charters Without Lotteries: Testing Takeovers in New Orleans and Boston

School Reform, December 2014

Lottery esti­mates sug­gest over­sub­scribed urban char­ter schools boost stu­dent achieve­ment marked­ly. But these esti­mates needn’t cap­ture treat­ment effects for stu­dents who haven’t applied to char­ter schools or for stu­dents attend­ing char­ters for which demand is weak. This paper reports esti­mates of the effect of char­ter school atten­dance on mid­dle-school­ers in char­ter takeovers in New Orleans and Boston. Takeovers are tra­di­tion­al pub­lic schools that close and then re-open as char­ter schools. Students enrolled in the schools des­ig­nat­ed for clo­sure are eli­gi­ble for “grand­fa­ther­ing” into the new schools; that is, they are guar­an­teed seats. We use this fact to con­struct instru­men­tal vari­ables esti­mates of the effects of pas­sive char­ter atten­dance: the grand­fa­ther­ing instru­ment com­pares stu­dents at schools des­ig­nat­ed for takeover with stu­dents who appear sim­i­lar at base­line and who were attend­ing sim­i­lar schools not yet closed, while adjust­ing for pos­si­ble vio­la­tions of the exclu­sion restric­tion in such com­par­isons. Estimates for a large sam­ple of takeover schools in the New Orleans Recovery School District show sub­stan­tial gains from takeover enroll­ment. In Boston, where we can com­pare grand­fa­ther­ing and lot­tery esti­mates for a mid­dle school, grand­fa­thered stu­dents see achieve­ment gains at least as large as the gains for stu­dents assigned seats in lot­ter­ies. Larger read­ing gains for grand­fa­ther­ing com­pli­ers are explained by a worse non-char­ter fallback.