Restaurant Employment, Minimum Wages, and Border Discontinuities

Dube, Lester and Reich (2010), using cross-state border county pairs (BCPs), did not detect restaurant employment losses from higher minimum wages. Jha, Neumark and Rodriguez- Lopez (2024, JNR), using multi-state commuting zones (MSCZs), do find disemployment effects. However, JNR’s results are confounded by 1990s-era parallel trends violations amplified by the two-way-fixed-effects (TWFE) model, which is also unreliable for detecting such violations. With post-2000 data, JNR’s specification does not indicate substantial disemployment. BCPs are much less affected by such pre-trends than MSCZs. Importantly, modern event study designs (using either MSCZs or BCPs) show no local pre-tends, and find no meaningful employment decline.