Clean Slate Laws and Aggregate Labor Market Outcomes: Evidence from Staggered State Adoption

Abstract

This paper documents patterns in aggregate labor market outcomes following the adoption of "Clean Slate" automatic criminal record expungement laws, while highlighting fundamental identification challenges that preclude causal interpretation. Between 2019 and 2024, seven U.S. states implemented automatic expungement programs. Using a staggered difference-in-differences design with the Sun-Abraham estimator, I find statistically significant associations with employment and labor force participation. However, event study analysis reveals severe pre-trends violations: 6 of 11 pre-treatment coefficients are statistically significant, indicating that Clean Slate adopting states were on systematically different employment trajectories than non-adopting states prior to implementation. The point estimates—0.15 percentage points for employment and 0.37 percentage points for labor force participation—cannot be interpreted causally given this selection. I also find a counterintuitive positive effect on unemployment that further undermines identification. These findings illustrate the difficulty of evaluating recent state policy innovations using aggregate data and the need for individual-level administrative data or alternative identification strategies.