The Minimum Wage Puzzle: Spillover Bias in Contiguous County Studies

Abstract

This thesis examines whether disagreement in the minimum wage literature over employment effects can be largely explained by bias in the contiguous county difference-in-differences design popularized by Card and Krueger. Focusing on New Jersey’s post-2014 minimum wage increases and its border with Pennsylvania, the analysis tests whether employment effects spill over into neighboring counties outside the treated state with spatially limited penetration. To do so, it compares three county-group specifications: a contiguous county replication, a within-Pennsylvania comparison used to detect spillovers, and a cross-state buffer-zone comparison designed to avoid spillover contamination. The results provide strong evidence of a spatially decaying spillover effect that exposes near-border Pennsylvania counties to the employment effects of New Jersey’s policy. By demonstrating that the control counties in the contiguous county framework are contaminated by this spillover, the findings suggest that such designs are biased toward insignificant estimates, helping to explain their subversive results and reconcile part of the disagreement in the literature. At the same time, the buffer-zone specification yields a positive and statistically significant employment effect for this case, suggesting that spillover bias explains part of the broader debate in the literature without resolving every disagreement over the employment effects of minimum wage increases.