U.S. Traffic Fatality Data, 2019–2024
How dangerous is
your county?
Every fatal crash on a U.S. public road is recorded by NHTSA in the Fatality Analysis Reporting System — a census, not a sample. This is all of it, 2019 through 2024, ranked by county.
The deadliest hour of the week
Traffic deaths are not spread evenly across the week. Each cell is one hour of one weekday, every fatality in the United States from 2019 through 2024 counted into it.
By hour of day
By day of week
Who dies
Fatalities by road user and by age, 2019–2024. Pedestrian and motorcyclist deaths are the share of the problem that vehicle safety engineering does the least for.
Road users
Age of people killed
Six years, one spike
Deaths surged in 2020–2021 even as Americans drove less, then fell three years running. 2024 is the lowest total since 2019.
The leaderboards
Annual traffic deaths per 100,000 residents, pooled over six years. County lists are limited to populations of 100,000 or more — see the note on small counties below.
What it costs
NHTSA puts the economic cost of a single traffic fatality at $1.61 million and the comprehensive cost — economic loss plus the valuation of lost quality of life — at $11.26 million, both in 2019 dollars.
What the $1.61 million is made of
Share of total U.S. crash economic cost by component, 2019. Lost market and household productivity together are 31% — the part a forensic economist is retained to quantify.
Method & sources
Where the numbers come from
Fatality counts are computed directly from NHTSA's Fatality Analysis Reporting System (FARS)
national CSV files for 2019 through 2024 — the accident and person tables,
read in full, not sampled. FARS is a census of every crash on a U.S. public road that killed
someone within 30 days.
Validation. The totals computed here reproduce NHTSA's published national figures exactly: 36,355 (2019), 39,007 (2020), 43,230 (2021), 42,721 (2022), 41,025 (2023) and 39,254 (2024). Road-user and hour-of-day splits are derived from the same records.
Denominators. County and state population are Census Bureau Vintage 2024 population estimates. Rates use the mean of the 2021 and 2024 estimates as a mid-window population, and are expressed as deaths per 100,000 residents per year.
County boundaries. FARS still codes Connecticut to its eight legacy counties while the Census Bureau moved to nine planning regions in 2022; legacy-county populations from the Vintage 2020 series are used for Connecticut, scaled to the 2024 state total. Renamed and split areas (Oglala Lakota SD, Kusilvak AK, Prince of Wales-Hyder AK, and Valdez-Cordova AK) are merged so no deaths are dropped. 3,142 counties and county equivalents are covered; 12 deaths out of 241,592 could not be assigned to a county and appear only in national and state totals.
Definitions. "Night" means the crash was coded dark, whether or not the road was lit. "Rural" is the FARS land-use code. Road users are assigned from the FARS person-type field so that the four categories sum exactly to each state's total: motorcyclists are occupants of vehicles with a motorcycle body type; pedestrians are person type 5; cyclists and other non-occupants cover bicyclists, other cyclists, people on personal conveyances such as scooters, occupants of non-motor transport devices, and other or unknown non-motorists. Bicyclist counts reconcile to NHTSA's published pedalcyclist series; the "other non-occupant" residual is roughly 200 deaths a year.
Which vintage of FARS. These are the current annual files, which contain NHTSA's revised counts. A reader checking against an older NHTSA publication will see small differences — 2023, for example, was first published at 40,901 and is 41,025 in the current file. Pedestrian and motorcyclist totals move the same way.
Rankings are computed here. The county leaderboards are this analysis's own calculation from FARS and Census data, not a NHTSA-published ranking. NHTSA does not publish a national county league table.
Caution on rates. Resident population understates true exposure wherever through-traffic dominates. For small counties the population-band rank is more meaningful than the national rank, and both are shown.
Sources
NHTSA, Overview of Motor Vehicle Traffic Crashes in 2024, DOT HS 813 791
NHTSA, The Economic and Societal Impact of Motor Vehicle Crashes, 2019 (Revised), DOT HS 813 403
U.S. DOT, Departmental Guidance on Valuation of a Statistical Life
FHWA, Highway Statistics 2024, Table VM-2
U.S. Census Bureau, County Population Totals, Vintage 2024
Download the full county dataset: us-county-traffic-deaths-2019-2024.csv (3,142 counties)
Reuse
FARS and Census data are U.S. Government works in the public domain. This page is free to embed, screenshot, cite or link. Attribution to Forensic Economic Services appreciated but not required.
Turning a fatality into a number is what we do for a living.
NHTSA's $1.61 million is a national average. In an actual case the figure is built from that person's earnings history, worklife expectancy, fringe benefits, household services and personal consumption, discounted to present value. Forensic Economic Services has been retained by counsel nationwide to do exactly that since 1993.
Send us the party names. We clear conflicts immediately.
No engagement, no cost, no sales process. If we're conflicted or not the right expert, we'll say so and point you somewhere useful.