Cullen, J. B., Jacob, B. A. & Levitt, S. D. (2000)

‘The impact of school choice on student outcomes:
an analysis of the Chicago Public Schools.’

NBER working paper 7888

• Explores the impact of a limited school choice program (through ‘open enrolment’) within the Chicago Public Schools (CPS) system.
• Under the system, students are guaranteed a place at a local school, but can ‘opt out’ and apply to any other high school in the CPS (over 60 schools). Roughly half of all parents do so. Note, however, that money does not follow the pupil in this system, and schools are not free to expand or contract.
• Uses cross-section data to compare the actual allocation of students to that which would have occurred if all pupils had gone to their assigned school.
  

Key results:

• Controlling for a wide range of observables (including 8th grade test scores) opting out is found to be associated with close to an 8% increase in completion of 10th grade, 11th grade and graduation on time. All estimates are highly statistically significant.
• Sorting of pupils based on ability is found to be much greater with open enrolment than without. The very highest performing students are segregated in schools with average test scores more than a grade-equivalent higher than would have been expected without school choice.
• The authors find no evidence of ‘negative spillovers’ from school choice (i.e. departing students imposing negative externalities on those left behind).
• However, the improved graduation rates of students who ‘opt out’ appear to derive from these students being superior in unobserved attributes such as motivation.