Abstract
This study re-examines Dube (2019), which finds large and statistically significant poverty-reducing effects of the minimum wage. We show that his estimated elasticities are fragile and sensitive to (1) time period under study, (2) choice of macroeconomic controls, (3) limiting counterfactuals to geographically proximate states (“close controls”), which poorly match treatment states' pre-treatment poverty trends, and (4) accounting for potential bias caused by heterogeneous and dynamic treatment effects. Using data spanning nearly four decades from the March Current Population Survey and a dynamic difference-in-differences (DiD) approach, we find that a 10 percent increase in the minimum wage is associated with a (statistically insignificant) 0.17 percent increase in the probability of longer-run poverty among all persons. With 95% confidence, we can rule out long-run poverty elasticities with respect to the minimum wage of less than -0.129. Our null results persist across a variety of DiD estimation strategies, including two-way fixed effects, stacked DiD, Callaway and Sant'Anna, and synthetic DiD. We conclude that, to date, the preponderance of evidence suggests that minimum wage increases are an ineffective policy strategy for alleviating poverty.
