The Limits of Marginal Aid: A Regression Discontinuity Estimate of Place-Based Policy in Appalachia

Abstract

The Appalachian Regional Commission has spent over \$3.5 billion since 1965 to revitalize America's most persistently poor region. I exploit a sharp threshold in ARC's county classification system—counties in the worst national decile are labeled "Distressed" and receive 80% federal match rates instead of 70%—to estimate whether this marginal designation improves local economic conditions. Using a regression discontinuity design on 3,317 county-year observations from 369 Appalachian counties over 2007–2017, I find the Distressed label has no effect on unemployment, per capita income, or poverty; I rule out even a 4% income improvement or a 0.6 percentage-point reduction in unemployment. The null survives bandwidth variation, polynomial order changes, donut-hole specifications, and placebo thresholds. Three independently measured BEA outcomes confirm the finding. Marginal increases in federal match rates—the architecture of most U.S.\ place-based programs—cannot bend the economic trajectory of chronically poor communities.