A reporter examined the bounds of authority of the Occupational Safety and Health Administration (OSHA), the agency charged with monitoring and regulating workplaces, to try to understand how many workers in the animal agriculture industry have oversight and protection.

To calculate the percentage of animal-ag operations that fall outside the federal agency’s jurisdiction, Civil Eats contacted the U.S. Department of Agriculture’s National Agricultural Statistics Service (NASS). According to the NASS, 1,075,130 U.S. farms specialize in livestock production. Of those farms, 234,991 hire workers—and 224,592 hire fewer than 10. (Because the OSHA appropriations rider exempts farms with 10 or fewer workers, rather than fewer than 10, the category NASS uses differs slightly.) Given that 224,592 is 95.57 percent of 234,991, about 96 percent of the livestock and poultry operations that hire workers hire in the U.S. employ fewer than 10 and are therefore not subject to OSHA oversight.

In further assessing OSHA’s limitations, reporters contacted state-level OSHA agencies to research which state OSHA programs allowed inspectors to investigate fatalities and injuries on small farms. Reporters found that 13 of the 22 states and territories with State Plan OSHAs that cover agricultural operations do not abide by the small-farm exemption and can inspect farms with 10 or fewer non-family employees.

Reporters also examined fatality and severe injury data from the federal OSHA. They searched what OSHA calls the Integrated Management Information System (IMIS) using the North American Industry Classification System codes, or NAICS codes, for the industries in which animal agriculture operations are concentrated. They then compared OSHA’s recorded fatalities with U.S. Bureau of Labor Statistics (BLS) fatality numbers for the same industries, using the same NAICS codes. These datasets are interrelated. In compiling its own Census of Fatal Occupational Injuries (CFOI), BLS includes incidents reported to OSHA, but gathers the data independently in the 50 states and at the federal level, relying on reports to state and federal OSHA agencies rather than on OSHA’s IMIS data. BLS also gathers data from numerous other sources, including reports to other state and federal agencies, death certificates, autopsy and coroner reports, hospital reports, police reports, and death certificates.

While federal OSHA records showed 149 animal-agriculture worker fatalities in the decade between 2011 and 2020, BLS data showed 1,006 deaths related to animal agriculture in those years. This gap accounts for the series conclusion that 85 percent of animal agriculture deaths were not reported to federal OSHA during that time.

There are significant differences in the methodologies BLS and OSHA use to collect agriculture-related deaths. Because BLS gathers data for statistical rather than regulatory purposes and is not subject to the small-farm exemption that limits OSHA’s response to farm fatalities, BLS’s data defines both “worker” and work-related incidents more broadly. The BLS fatality data thus includes deaths of non-employees like family members, contract workers, and volunteers, for example.

In the course of this research, reporters also queried OSHA about the contents and comprehensiveness of the data in its IMIS database to identify any other issues that could account for the gap in animal agriculture deaths recorded. IMIS is described as comprehensive in a guiding Note to Users, but OSHA staff and other sources noted possible data gaps in reports from the 22 states and territories under State Plan OSHA control prior to 2017 and possible lags in the timeliness of data uploads owing to a lack of staff to clean and load some data. OSHA’s own lack of clarity about the comprehensiveness of its data, combined with the complexity of the OSHA bureaucracy, made more detailed questions about IMIS data unanswerable, even after many months in conversation with the agency. For these reasons, and because of the complex way the state and federal OSHAs communicate and handle data, it was impossible to conclude how many worker deaths federal OSHA’s limited authority obscures, a fact noted in the text.

In further examining these acute injuries and fatalities, as well as long-term illness among animal-agriculture workers, reporters relied on information alerts from the National Institute for Occupational Safety and Health (NIOSH) and expert research in addition to the numerous interviews and references cited in the text. While examining the long-term health of workers, they also referenced an extensive web of science studying respiratory health, chemical exposure, exposure to microorganisms, fatal injuries associated with manure pits, and health trends in communities surrounding concentrated animal feeding operations, or CAFOs. These—and additional market reports and industry research examining consolidation trends, financial risks and policies, and the changing size and location of dairy farms—were too numerous to cite. In examining biogas incentives, reporters also researched media reports of farm deaths related to manure pits.

This methodology was updated on December 21, 2022, to better explain the relationship between the two datasets used in this analysis and to further articulate possible gaps within the OSHA data.