Equitable Access in School Choice: Evidence from Washington, D.C.

• In Event: The Gap Between Promise and Practice: Three Logics of School Choice Under Scrutiny

Abstract

School choice holds the promise of expanding access to high-quality schools by decoupling assignment from residential location. However, a persistent concern is that choice may exacerbate segregation as families differ in resources and capacity to participate. In response, some districts have begun embedding equity goals directly into assignment mechanisms (i.e. controlled choice). This paper evaluates a recent reform to the Washington, D.C. unified school lottery designed to increase access to high-demand schools for economically disadvantaged students.

The reform, known as “Equitable Access,” introduced a priority for economically disadvantaged students within a student-proposing deferred acceptance (DA) algorithm. In theory, priority-based (“soft”) rules preserve desirable properties of DA, such as strategy-proofness, unlike quota-based (“hard”) rules. However, the effectiveness of priorities depends on applicant demand, raising an important empirical question: do families respond to these rules in ways that offset or amplify policy effects? While a large theoretical literature studies controlled choice, there is limited empirical evidence on how such policies operate in practice.

I address three questions: (1) how did the policy change match rates to high-demand schools for targeted and non-targeted students, (2) did the reform induce behavioral responses that meaningfully affected assignment outcomes, and (3) what were the resulting welfare consequences? Evaluating these questions is challenging because assignment outcomes are equilibrium objects without directly observable counterfactuals. To address this, I perfectly replicate DC’s assignment mechanism and simulate counterfactual assignments under alternative policy scenarios. I then combine these simulations with a Shapley decomposition to isolate the contributions of policy rules, applicant preferences, and school capacity.

The analysis uses administrative lottery records from 2018–2024, focusing on PK3 applicants, where policy exposure is most concentrated. Between 2020 (the last pre-pandemic lottery) and 2023 (the second year of the policy), match rates to targeted schools for low-income students increased from 22 to 38 percent. Over half of this increase was due to the rules alone (9 percentage points). This effect was more than twice as large as the change from the concurrent 20 percent expansion in seats (4 percentage points). I find no evidence that changes in applicant preferences affected match rates beyond normal year-to-year variation. Welfare effects are modest: while higher-income students experience small declines in first-choice matches, both groups are, on average, more likely to match to higher-ranked schools. Results are robust to alternative base years and implementation years.

This paper provides direct evidence that applicant preferences remain stable in response to rule changes in a DA system and offers a framework for evaluating similar reforms in other districts that are currently being tested, such as Denver and Chicago.

Author

• Michael Connolly Quinn, College of the Holy Cross
  Presenting Author