Skip to main content

Randomized safety inspections and risk exposure on the job: Quasi-experimental estimates of the value of a statistical life (Lee & Taylor, 2019)

Absence of conflict of interest.

Citation

Lee, J. M., & Taylor, L. O. (2019). Randomized safety inspections and risk exposure on the job: Quasi-experimental estimates of the value of a statistical life. American Economic Journal: Economic Policy, 11(4), 350-374.

Highlights

  • The study's objective was to examine the impact of surprise plant inspections conducted by the federal Occupational Safety and Health Administration (OSHA) on plants’ fatality rates and average hourly wages. 

  • The study uses an instrumental variables design where OSHA inspections, which are random conditional on state, industry, and recent inspection history, are an instrument for plant-level workplace risks. The study uses OSHA inspection records, which include plant injury and fatality rates, and data on plant-level employment and wages from the Census of Manufactures and Annual Survey of Manufactures. The authors use a statistical model to compare outcomes at plants before and after receiving an OSHA inspection with those of different plants receiving their first inspection in different years. 

  • The study found that receiving a surprise OSHA inspection was associated with a lower fatality rate in the following years and with slightly lower average hourly wages per worker.  

  • This study receives a moderate evidence rating. This means we are somewhat confident that the estimated effects are attributable to surprise plant inspections conducted by OSHA, but other factors might also have contributed. 

Intervention Examined

Surprise safety inspections conducted by OSHA

Features of the Intervention

The federal Occupational Safety and Health Administration (OSHA) was created in 1970 to oversee workplace safety standards. As part of its mission, OSHA conducts workplace inspections both randomly, through unannounced or surprise visits known as “programmed” inspections, and in response to workplace events, such as employee complaints, serious accidents, or follow-ups from a previous inspection to ensure compliance. During the observation period, OSHA randomly selected plants for surprise inspections by state within high-risk manufacturing industries, exempting plants with fewer than 11 employees or that had recently been inspected.  

The study focuses on the 28 states (plus District of Columbia) that have opted to have OSHA administer their workplace safety programs rather than administering individual state-specific safety programs. The main analyses focus on seven high-risk manufacturing industries: food; paper; printing and publishing; chemicals; stone, glass, and clay products; primary metal industries; and fabricated metal products.  

Features of the Study

The study uses an instrumental variables design where OSHA inspections are an instrument for plant-level workplace risks. This instrumental variables design is combined with a difference-in-differences (DID) approach to compare fatality rates and the resulting effects on wages between inspected plants versus uninspected plants. The models control for time-varying plant characteristics like number of employees and turnover, and the preferred models contain fixed effects for year, plant, industry, and the three-way interaction of year, industry, and state.  

The study uses OSHA inspection records, which include plant injury and fatality rates, and data on plant-level employment and wages from the Census of Manufactures and Annual Survey of Manufactures. Matching the OSHA inspection data with the Census of Manufactures requires an iterative process, including manual matching to achieve the best matching rates. The authors conducted the necessary manual matching for seven industries with high fatality rates; these seven industries are the focus of the authors’ main analyses.  

The main analyses use data from 1987 to 1997 and focus on plants that receive their first OSHA inspection between 1988 and 1996, thus having at least one observation prior to and one observation after the first inspection. The analytic sample includes 28,071 plants and 112,865 observations across the time period. The treatment condition is manufacturing plants that received an OSHA inspection. The comparison condition is manufacturing plants that were eligible to be randomly selected for inspection but were not inspected at that time.  

Findings

Health and safety

  • The study found that receiving a surprise OSHA inspection is associated with significantly fewer fatalities, 3.5 per 10,000 workers annually, compared with not receiving an OSHA inspection. 

Earnings and wages

  • The study found that average wages are significantly lower ($0.21 lower per hour or $420 lower per year) post-inspection for plants that received a surprise OSHA inspection compared to those that did not. 

Considerations for Interpreting the Findings

The authors interpret the lower wages after a plant is randomly inspected by OSHA as evidence of “compensating wage differentials,” meaning that, because workers’ environment is safer after the plant remedies violations found in the inspection, the plant does not need to pay as much to induce workers to stay—safety compensates for wages. An alternate interpretation might be that plants spend funds addressing OSHA violations and thus chooses to pay their employees less to make up those costs. 

Causal Evidence Rating

The quality of causal evidence presented in this report is moderate because it was based on a well-implemented nonexperimental design. This means we are somewhat confident that the estimated effects are attributable to the surprise plant inspections conducted by OSHA, but other factors might also have contributed.  

Reviewed by CLEAR

April 2022