The Long Shadow of the Paddle? Evidence from State Corporal Punishment Bans

Abstract

We investigate whether state-level bans on corporal punishment in public schools affected long-term educational and economic outcomes. Using the staggered adoption of bans across 33 U.S. states between 1971 and 2023, we implement a difference-in-differences design with 3.2 million observations from the American Community Survey (2017–2022). Our event-study analysis reveals clear violations of the parallel trends assumption: pre-treatment coefficients show systematic differences between early-ban (predominantly Northeastern) and never-ban (predominantly Southern) states that predate any policy change. The counterintuitive finding that bans increase disability rates is a strong indicator of residual confounding from divergent regional trends rather than a causal effect. We conclude that the stark socioeconomic and cultural differences between states that adopted bans early versus never preclude credible causal identification with standard two-way fixed effects methods. Our results underscore the importance of pre-trends diagnostics and careful counterfactual construction in staggered adoption designs, and motivate the use of synthetic control methods when parallel trends cannot be maintained. \bigskip Keywords: corporal punishment, education policy, difference-in-differences, parallel trends, identification, school discipline \bigskip JEL Codes: I21, I28, C21, J24