Do OSHA Inspections Reduce Injuries? A Panel Analysis
Wayne B. Gray, John T. Scholz
NBER Working Paper No. 3774
Using data on injuries and OSHA inspections for 6,842 large manufacturing plants between 1979 and 1985, we find evidence that OSHA inspections significantly reduce injuries. This effect comes exclusively from inspections that impose penalties, inspections which do not impose penalties appear to have no effect on injuries. Plants which are inspected (and penalized) in a given year experience a 22 percent decline in their injuries during the following few years. In our sample, total OSHA enforcement is predicted to have reduced injuries by about 2 percent. We take advantage of the panel nature of our data to test for a number of potential biases: autocorrelated injuries, plant-specific fixed-effects which are correlated with both inspections and injuries, and endogeneity of inspections (injuries causing inspections). These biases lead us to use the percentage change in injuries, rather than injury levels, as the dependent variable for our estimation. Our analysis shows that the estimated effect of inspections on the percentage change in injuries is not significantly affected by these biases, and thus seems to reflect a 'deterrence' effect of OSHA inspections on injuries.
Document Object Identifier (DOI): 10.3386/w3774
Published: published as "Does Regulatory Enforcement Work? A Panel Analysis of OSHA Enforcement", Law and Society Review, vol. 27, no. 1, (June 1993) P. 177-213