Abstract:
This paper revisits an important analysis of enterprise zones (EZs) by Ham, Swenson, Imrohoroğlu, and Song (2011), who report substantial poverty reductions from state and federal EZs, as well as improvements in other labor market outcomes. In our re-analysis, we find that a data error accounts for a large share of the estimated impact of state EZs in reducing poverty. More generally, we find that both state and federal EZs appear to be endogenously selected based on prior changes in poverty and other labor market outcomes. Once we account for this selection, much of the evidence that state and federal EZs reduce poverty largely evaporates, as does most of the evidence for other beneficial effects of enterprise zones, with the main exception of some limited evidence for federal Empowerment Zones. Thus, we confirm the more widely-prevailing view that EZs – and especially state EZs – have for the most part been ineffective at reducing urban poverty or improving labor market outcomes in the United States.