9,979 People Died in Crashes Involving a Drinking Driver — and the State You Die In Changes the Odds Threefold

NHTSA published the 2024 Fatality Analysis Reporting System file on November 20, 2025. It records 36,297 fatal crashes and 39,254 deaths — 1,771 fewer deaths than 2023, a decline of 4.3%.

In 8,990 of those crashes at least one driver was recorded as having been drinking, and 9,979 people died in them. That is 25.4% of all road deaths, down from 26.4% and 10,828 deaths the year before — 849 fewer.

One clarification before the state numbers, because it matters for anyone comparing this to an official figure. We count the FARS field DR_DRINK: a driver whom police recorded, or a test showed, to have been drinking. NHTSA's own headline measure — "alcohol-impaired driving fatalities", meaning a BAC of 0.08 or above — is higher, because it statistically imputes blood-alcohol levels for the many drivers never tested. That imputation is not something an outside analyst can reproduce, so we do not publish a number under its name. Ours is the conservative count.

Among states with at least 200 road deaths, the share of deaths involving a drinking driver runs:

• Montana 44.7% — very nearly one death in two.
• Nebraska 35.9%, Kansas 35.4%, Idaho 35.3%, Connecticut 35.2%.
• At the other end: New York 14.3%, Florida 16.6%, Georgia 17.0%, New Mexico 17.4%, Mississippi 18.6%.

A road death in Montana is more than three times as likely to involve a drinking driver as one in New York. Some of that is what people picture — rural roads, long distances, no alternative to driving home. Some of it is measurement: these are police-reported codes, and a state whose officers record the fact more often will report a higher share.

The tempting story is that lenient states have more of it. We have the first-offense license suspension for all 51 jurisdictions from the statutes, so we checked rather than assumed.

Across the 32 states with at least 200 road deaths, the correlation between the alcohol share and the length of the first-offense suspension is r = −0.26. Directionally what you would expect — longer suspension, slightly lower share — but weak enough that it accounts for roughly 7% of the variation between states. It is not a finding to act on.

The individual pairs make the point better than the coefficient. Georgia and New Mexico impose the longest first-offense suspensions in the country at 365 days, and both sit near the bottom of the alcohol share (17.0% and 17.4%). Kansas suspends for 30 days, the shortest, and sits near the top at 35.4%. But Connecticut suspends for 45 days and Montana for 180 — four times as long — and both are in the top five. Whatever drives the spread, penalty length is not most of it.

The national decline of 849 alcohol-involved deaths was not spread evenly. The largest falls in absolute terms were California (−179, from 1,267 to 1,088), Florida (−125), New York (−91, from 248 to 157), Georgia (−73) and Washington (−64).

Four states went the other way: Utah (+33), Kansas (+32), Minnesota (+20) and Wisconsin (+17).

One caveat on year-over-year comparisons. FARS annual files are revised after first release: the current 2022 file contains 42,721 deaths, while NHTSA's June 2024 research note reported 42,514 for the same year — about 200 cases of difference from vintage alone. At the scale of a 1,771-death national decline that does not change the direction, but at the scale of a single small state it can.