Location Choice, Commuting, and School Choice

Abstract

We explore the impact of public school assignment reforms by building a households’ school choice model with two key features—(1) endogenous residential location choice and (2) opt-out to outside schooling options. Households decide where to live taking into account that locations determine access to schools—admissions probabilities and commuting distances to schools. Households are heterogeneous both in observed and unobserved characteristics. We estimate the model using administrative data from New York City’s middle school choice system. Variation from a boundary discontinuity design separately identifies access-to-school preferences from other location amenities. Residential sorting based on access-to-school preference explains 30% of the gap in test scores of schools attended by minority students versus their peers. If households’ residential locations were fixed, a reform that introduces purely lottery-based admissions to schools in lower- and mid-Manhattan would reduce the cross-racial gap by 7%. However, households’ endogenous location choices dampen the effect by half.

Keywords: Centralized School Choice, Neighborhood Sorting, School Segregation, Commuting to School

JEL Classification: D12, I21, I28, J15, R23