Regulation by Shaming: Direct and Spillover Effects of Publicizing Violations of Workplace Safety and Health Laws

Ratings, scores and other forms of information provision are increasingly seen as a tool to incentivize firms to improve their quality or attributes, but less is understood about policies which exclusively publicize firms discovered to have the lowest quality attributes–i.e. “shaming.” Shaming may affect the decisions of publicized firms, and perhaps of greater policy relevance is its deterrent effect on not-yet publicized firms. After a 2009 policy change, the Occupational Safety and Health Administration (OSHA) began issuing press releases about workplaces found to be in violation of safety and health regulations if the penalties levied for those violations exceeded a cutoff. Using quasi-random variation induced by this cutoff, I estimate the effect of publicizing violations by one workplace on its own subsequent compliance with OSHA regulations, and also the spillover effect onto the compliance of “peer” workplaces most likely exposed to the publicity. Workplaces with a peer whose violations were publicized significantly improve compliance with safety and health regulations. These spillover effects decrease with geographic and industry “distance” to the publicized workplace, and the response appears to be driven by employers acting to avoid their own future publicity, rather than by updating beliefs over the probability of future regulatory enforcement. Finally, the magnitude of spillover effects associated with a press release diminish in the years after the policy’s introduction, suggesting the information content relevant to employers was strongest for the earliest press releases. JEL: J28, J81, L51 ∗Email [email protected] I wish to thank Kevin Lang, David I. Levine, and Johannes Schmieder for invaluable guidance, as well as Kavan Kucko, Michael Lipsitz, John Mendeloff, Alison Morantz, Benjamin Ogden, Claudia Olivetti, Mike Toffel, David Weil, Russell Weinstein, and seminar participants at Boston University and the 2014 Alliance for Research on Corporate Sustainability (ARCS) Annual Meeting for helpful discussion. I would also like to thank the staff at OSHA’s headquarters and each of OSHA’s regional Office of Public Affairs for providing me with details on OSHA’s press release policies and patiently answering my unending questions. Keith Ma provided invaluable guidance with geocoding in ArcGIS. Jing Li provided excellent research assistance. All errors are mine and mine alone.