Association for Public Policy Analysis & Management

Critics have long lamented the design of public housing projects, charging that the resulting concentrated poverty creates problematic micro-neighborhoods with limited opportunity and poor quality services that hinder the wellbeing of residents overall and children, in particular. There is, however, little rigorous evidence documenting negative causal effects of public housing due to the difficulty of disentangling the underlying characteristics of public housing residents from the impact of residence in the projects per se. One exception, Schwartz, McCabe, Ellen, and Chellman (2010) use cross-sectional data on New York City public school students and find students living in public housing have worse academic performance than classmates living elsewhere, even conditional on a set of sociodemographic characteristics. Weinhardt (2014) exploits the plausibly random precise timing of entry into social housing in England, finding little evidence of any negative (or positive) effects. We build upon this research, drawing on rich student-level longitudinal data to examine the trajectories of academic performance of students living in public housing during their school years (grades 3-8) and shed light on the effects of public housing on student outcomes. Using an event study framework, we focus on those entering public housing in grades 5-7, exploring trajectories in test scores, attendance, school mobility, and school quality before and after the residential move and using a variety of sociodemographic controls and student fixed effects. Before exploring heterogeneity and mediating factors, we perform a series of robustness checks to explore the case for a causal interpretation.

New York City’s Housing Authority (NYCHA), aims to “increase opportunities for low- and moderate-income New Yorkers by providing safe, affordable housing and facilitating access to social and community services.” Importantly, NYCHA is the nation’s largest public housing system, serving more than 400,000 residents in its 326 public housing developments. Our analyses focus on the 117,723 public school students, K-12, who lived in NYCHA housing at some point between academic years 2008-09 and 2016-17. Test scores analyses include 10,778 students taking standardized reading and math tests in grades 3-8, and event studies focus on the cohorts entering NYCHA public housing in grades 5-7. Key data include student residential address, zoned schools and actual schools attended, reading and math scores, attendance, and sociodemographic variables, such as age, gender, race, nativity, educational program participation (e.g. students with disabilities and English language learners), and eligibility for free or reduced-price lunch.

We estimate both event study and more parsimonious difference-in-difference style models to compare academic trajectories pre- and post-entry into public housing and shed light on school mobility, changes in school quality, among other potential mediators. Preliminary results suggest no immediate change in performance, followed by steady improvements over time but little evidence of significant pre-trends. School mobility jumps with entry, potentially explaining stalled academic performance. Future analyses will examine changes in attendance and multiple measures of school quality (including peer characteristics). Robustness checks will consider alternate specifications, subgroups, and samples. The study results will inform the policy debate about the impact of public housing on student outcomes.