This dissertation examines how policy interventions shape outcomes for vulnerable populations, focusing on homelessness and foster care in North America. The first chapter studies the effects of court-imposed limits on police enforcement of anti-homelessness laws. Leveraging variation from key federal and Supreme Court rulings as a natural experiment, it estimates the impact of enforcement constraints on visible homelessness across cities. The analysis draws on a novel multi-city dataset that integrates enforcement records, street-level homelessness counts, and shelter system data collected through extensive outreach to local agencies.
The second chapter examines how foster care placement decisions affect long-term outcomes for families, highlighting institutional tradeoffs between child safety and family preservation. It contributes to policy debates around child welfare by analyzing administrative data that follows families over time.
To support both projects, the dissertation develops tools for causal inference, including new methods for variance estimation in panel data. Together, these studies combine original data collection, applied econometrics, and large-scale administrative record linkage to generate new evidence on how systems of care and control influence marginalized groups. The work offers a scalable, data-driven approach to evaluating complex institutions and informing more effective and equitable social policy.
