The effects of field office closures on social assistance take-up and targeting

How does in-person support affect access to safety net programs? I study this question by examining how the closure of field offices has changed welfare and disability assistance caseloads. Using rich administrative data and a staggered difference-in-differences design, I estimate that closures, on average, reduced local caseloads by 11.5% for welfare and a statistically insignificant 1.6% for disability assistance. Declines in welfare caseloads (i) occurred across demographic, health, education, and eligibility groups, (ii) were somewhat larger among young and healthier individuals, and (iii) were suggestively larger among persons less familiar with the programs. On the whole, I find limited change in the relative targeting of benefits.