The Autopsy / Case File · Violence & Homicide
The Register, the Cab, the Badge
Workplace homicide isn't spread evenly across the American labor force. Federal data traces it to three specific settings, each with its own documented mechanism: a retail counter someone can rob, a vehicle someone can get into, and a uniform someone can target. Two of the three have decades of granular government data behind them. The third — the taxi and rideshare industry — is tracked so inconsistently across six separate federal databases that no single agency can say with confidence how many of its workers are actually being killed.
BLS's Census of Fatal Occupational Injuries recorded 470 workplace homicides in 2024, up from 458 the year before, with shootings accounting for 379 of them and stabbings 43. A BLS study of the retail sector found 1,434 workplace homicides in that industry alone across just six years, 2003 through 2008 — 58 percent tied directly to robbery, concentrated overwhelmingly in convenience stores and gas stations, where three-quarters or more of all fatal injuries on the job are homicides. Taxi driving has long carried the single highest occupational homicide rate BLS has measured, at roughly 36 times the risk facing the average worker in 1998. And a February 2024 Government Accountability Office report found that taxi and rideshare driver deaths remain so inconsistently classified across BLS, CDC, and Justice Department databases that federal agencies cannot produce a reliable national count of how many are being killed today.
Join the discussion →
Homicide is not the leading cause of workplace death in America — transportation incidents hold that position by a wide margin. But it is the second leading cause, and unlike a falling scaffold or an unguarded machine, it is not evenly distributed across the kinds of harm a job can expose someone to. A construction site's hazards are mostly physics: gravity, weight, moving parts. A workplace homicide's hazards are mostly geography and role: whether the job puts a worker alone with cash after dark, alone in a vehicle with a stranger, or in a uniform whose entire function is to be the person who intervenes. BLS's Census of Fatal Occupational Injuries — the same annual accounting used throughout this series — gives the national total. Getting to where that total actually concentrates requires three separate, differently documented stories.
What the National Count Looks Like Now
BLS's most recent Census of Fatal Occupational Injuries, covering 2024, recorded 470 workplace homicides, up from 458 the year before and accounting for 64.1 percent of all fatalities the agency classifies as resulting from violent acts. Shootings caused 379 of those deaths; stabbings caused 43; the remainder involved other means. Women, who make up roughly 8.1 percent of all workplace fatalities across every cause, accounted for a disproportionate 15.3 percent of homicide victims specifically — a gap that starts to explain itself once the type of homicide gets broken down by relationship rather than treated as a single undifferentiated category. Protective service occupations — a BLS classification covering police officers, security guards, and similar roles — recorded 281 total fatalities in 2024, of which 97, or 34.5 percent, were homicides, the highest homicide share of any major occupational category BLS tracks.
The Register: Robbery With a Predictable Address
Retail trade's homicide problem is, more than any other industry's, a story about which specific storefronts carry the risk. A BLS analysis covering 2003 through 2008 counted 1,434 workplace homicides across the retail sector in those six years alone — an average of roughly 239 a year in that period, in a single industry. Fifty-eight percent were classified as robbery-related, 23 percent as non-robbery, and 19 percent of unknown motive. Break the non-robbery cases down further and a second pattern emerges: of those with an identifiable circumstance, roughly a third involved a customer or client, another third a personal relationship, and about a quarter a coworker — meaning even the homicides retail data doesn't attribute to robbery mostly aren't random either; they trace back to a specific, identifiable relationship gone fatal. But the robbery-driven majority is where the geography sharpens. Convenience stores carried a homicide rate roughly seven times higher than the average across all industries. Gasoline stations and beer, wine, and liquor stores were worse still by a different measure: homicides accounted for roughly three-quarters of all fatal occupational injuries at gas stations, and more than four-fifths at liquor stores — meaning that in those two specific retail subcategories, if a worker died on the job at all, the overwhelming likelihood is that someone killed them. Nearly two-thirds of all retail robbery-homicides concentrated in just two subsectors: food and beverage stores, and gas stations. Roughly three-quarters of the assailants in these cases were found to be robbers.
In gas stations and liquor stores, if a worker died on the job at all, the overwhelming likelihood is that someone killed them.
The Cab: The Occupation With the Ceiling
No occupation in BLS's historical data has carried a higher homicide rate than driving a taxi. In 1998, taxi drivers and chauffeurs faced 17.9 homicides per 100,000 workers — a rate BLS at the time calculated at roughly 36 times the risk facing the average employed person. That figure predates the rideshare era by more than a decade, and it is precisely the modern version of that same risk — spread now across traditional taxis and app-based ridesourcing drivers alike — that current federal data struggles hardest to quantify. A February 2024 Government Accountability Office report, produced in response to a congressional mandate known as Sami's Law, examined six separate federal databases that each collect some information relevant to driver safety: BLS's Census of Fatal Occupational Injuries and Survey of Occupational Injuries and Illnesses, the National Center for Health Statistics' National Vital Statistics System, NIOSH's emergency-department injury surveillance system, OSHA's enforcement database, and the Bureau of Justice Statistics' National Crime Victimization Survey. GAO's finding was blunt: none of the six was built to answer the specific question of how often taxi and rideshare drivers are assaulted or killed, and the ones that come closest can't be compared to each other or aggregated into a single reliable number.
A Category No Database Was Designed to Count
The specific ways the data breaks down are instructive on their own. CFOI and the National Vital Statistics System both use occupation codes that group taxi drivers, ridesourcing drivers, and even chauffeurs into a single category, meaning BLS cannot tell a taxi homicide from an Uber or Lyft homicide in its own numbers. GAO's own reporting of the only recent year with a publishable CFOI figure — 2019 — put fatal intentional-injury deaths across the combined taxi-and-ridesourcing industry at just 19 nationally; 2020's figure didn't meet BLS's publication threshold for reliability at all. The Survey of Occupational Injuries and Illnesses, meanwhile, excludes self-employed workers entirely — and BLS's own Employment Projections program estimated that roughly 89 percent of taxi drivers were self-employed as of 2021, meaning the survey structurally cannot capture the large majority of the workforce it would need to count. The Bureau of Justice Statistics' National Crime Victimization Survey, working from a different methodology, estimated 47,597 non-fatal violent victimizations of taxi drivers across 2016 through 2020 combined — a five-year aggregate, because the annual sample sizes were too small to produce a reliable yearly figure. Layered against all of that, the ridesourcing companies themselves publish their own numbers: Uber's safety reports counted 9 fatal physical assaults tied to trips arranged through its app in 2019 and 11 in 2020, against roughly 2.1 billion total trips; Lyft counted 4 in 2019 and had not yet published a 2020 figure at the time of GAO's report. None of these figures — the CFOI count, the NCVS estimate, the individual companies' self-reported numbers — use compatible definitions, and GAO explicitly found they cannot be combined into a single trustworthy total.
The Badge: A Risk Built Into the Job Description
Protective service occupations complete the pattern from the opposite direction: instead of an industry made dangerous by what happens to the people working in it, it is an occupational category built specifically around intervening in danger. That structural difference shows up directly in the 2024 CFOI numbers — 97 of 281 total protective-service fatalities, or 34.5 percent, were homicides, the highest share of any major occupational group BLS tracks, well above the 64.1 percent that violent acts represent among all fatalities from that specific cause nationally once every occupation is combined. Security guards, a related but distinct classification from sworn law enforcement, carried a homicide rate of roughly 14.3 per 100,000 workers in BLS's 2003–2008 data — lower than the taxi driver ceiling but still dramatically above the national average across all occupations, and high enough to place private security work among the small handful of jobs where a homicide, rather than an accident, is the leading way workers in that role actually die on duty.
Three Mechanisms, One Category
Put the register, the cab, and the badge side by side and the shape of workplace homicide in America stops looking random. It is a story about cash and isolation in retail, about a stranger and a closed vehicle in transportation, and about a role whose job description is to be the target in protective services — three distinct, well-documented mechanisms that happen to get summed into one BLS total every year. The one place that total gets genuinely uncertain is exactly the place where the underlying industry has changed fastest: taxi driving didn't stop being the most dangerous job in America when ridesourcing apps arrived, but the federal data infrastructure built to count that danger was designed for an industry that, in important structural ways, no longer exists in the form the databases assume.
And Then You're Dead
A 470-person national total sounds like a single phenomenon. It isn't. It's a convenience-store clerk working the register alone after dark in a subsector where homicide is the leading cause of on-the-job death outright; it's a driver alone with a stranger in a vehicle, in an industry the federal government's own data systems still can't fully see; and it's someone in a uniform whose entire job, on paper, is to be the one who steps toward the danger instead of away from it. Three different reasons to die at work, filed under one number.
That's it. That's the whole thing.
Sources
- 470 workplace homicides recorded nationally in 2024 (up from 458 in 2023), 64.1 percent of all violent-act fatalities, 379 by shooting and 43 by stabbing, women accounting for 8.1 percent of all workplace fatalities but 15.3 percent of homicide victims, and protective-service occupations recording 281 total fatalities including 97 homicides (34.5 percent) — U.S. Bureau of Labor Statistics, National Census of Fatal Occupational Injuries in 2024, released February 19, 2026 (USDL-26-0230)
- 1,434 workplace homicides in the retail trade sector from 2003 to 2008, 58 percent robbery-related, 23 percent non-robbery (roughly a third customer/client, a third personal relationship, a quarter coworker among identifiable circumstances), convenience stores carrying roughly seven times the average industry homicide rate, homicides accounting for roughly three-quarters of fatal occupational injuries at gas stations and more than four-fifths at liquor stores, nearly two-thirds of retail robbery-homicides concentrated in food/beverage stores and gas stations, and roughly three-quarters of assailants identified as robbers — U.S. Bureau of Labor Statistics, "Assaults and Violent Acts in the Private Retail Trade Sector, 2003–2008," Monthly Labor Review/Compensation and Working Conditions; BLS, "Homicides in Retail Trade, 2003–2008," The Economics Daily; peer-reviewed replication in "Non-Robbery-Related Occupational Homicides in the Retail Industry, 2003–2008"
- Taxi drivers and chauffeurs facing 17.9 homicides per 100,000 workers in 1998, roughly 36 times the risk facing the average employed person, and security guards carrying a homicide rate of roughly 14.3 per 100,000 workers in 2003–2008 data — U.S. Bureau of Labor Statistics historical occupational fatality-rate data
- A finding that six federal databases (BLS's CFOI and SOII, NCHS's National Vital Statistics System, NIOSH's NEISS-Work, OSHA's Information System, and BJS's National Crime Victimization Survey) each collect incompatible, incomplete data on taxi and rideshare driver assaults and fatalities, including a 2019 CFOI count of 19 fatal intentional-injury deaths across the combined taxi/ridesourcing industry with no publishable 2020 figure, SOII's structural exclusion of the roughly 89 percent of taxi drivers who are self-employed, and NCVS's 47,597 non-fatal violent victimizations of taxi drivers estimated only as a 2016–2020 five-year aggregate — U.S. Government Accountability Office, "Ridesharing and Taxi Safety: Information on Assaults against Drivers and Passengers," GAO-24-106742, February 2024
- Uber's and Lyft's self-reported fatal physical assault counts tied to app-arranged trips (Uber: 9 in 2019, 11 in 2020, against roughly 2.1 billion total trips; Lyft: 4 in 2019, 2020 not yet published) — cited within GAO-24-106742, sourced to Uber and Lyft public safety reports
- Photo: "Night view of 7-Eleven store at 2546 Niagara Falls Boulevard, Tonawanda, New York," Andre Carrotflower, CC BY-SA 4.0, via Wikimedia Commons
Comments
Loading…