Article | Science | May 18, 2012

Randomized Government Safety Inspections Reduce Worker Injuries with No Detectable Job Loss

by David I Levine, Michael W. Toffel and Matthew S. Johnson


Controversy surrounds occupational health and safety regulators, with some observers claiming that workplace regulations damage firms' competitiveness and destroy jobs and others arguing that they make workplaces safer at little cost to employers and employees. We analyzed a natural field experiment to examine how workplace safety inspections affected injury rates and other outcomes. We compared 409 randomly inspected establishments in California with 409 matched-control establishments that were eligible, but not chosen, for inspection. Compared with controls, randomly inspected employers experienced a 9.4% decline in injury rates (95% confidence interval = -0.177 to -0.021) and a 26% reduction in injury cost (95% confidence interval = -0.513 to -0.083). We find no evidence that these improvements came at the expense of employment, sales, credit ratings, or firm survival.

Keywords: regulation; safety; occupational safety; evaluation; regression; matching; difference in differences; Safety; Health; Working Conditions; Governing Rules, Regulations, and Reforms; Competitive Advantage; Performance; Manufacturing Industry; California;


Levine, David I., Michael W. Toffel, and Matthew S. Johnson. "Randomized Government Safety Inspections Reduce Worker Injuries with No Detectable Job Loss." Science 336, no. 6083 (May 18, 2012): 907–911. (Featured by the head of US OSHA's blog and article, and in The Washington Post, U.S. News & World Report and many other news outlets. Basis of U.S. Congressional testimony on promoting safe workplaces.)