39,345 Dead and NHTSA Threw a Party

Somewhere in Washington, a bureaucrat typed the words “first time since 2020 that the number of fatalities fell below 40,000” into a press release and hit send. Presumably there were doughnuts. NHTSA’s April announcement of the 2024 full-year traffic fatality estimate read like a quarterly earnings call where the company lost eight percent less money than last year and the CEO described it as “encouraging.”

The number is 39,345. That is how many Americans NHTSA estimates died on the road in 2024.^[1] Eleven consecutive quarters of decline. A fatality rate of 1.20 per 100 million vehicle miles traveled. Below the sacred 40,000 line for the first time in four years. And yet, buried in the agency’s own announcement, the tell: “Total road fatalities remain significantly higher than a decade ago,” the release conceded, and “America’s traffic fatality rate remains high relative to many peer nations.”

Significantly higher, and the arithmetic is not flattering. In 2014, the United States recorded 32,744 traffic fatalities.^[2] The 2024 figure exceeds that by 6,601 bodies, a gap of 20.2 percent spanning a decade of advanced driver-assistance systems, automatic emergency braking mandates, and billions in safe-streets grants that NHTSA dutifully cataloged in its press materials. In 2019, the last pre-pandemic year that anyone can use with a straight face as a benchmark, 36,355 people died, and the 2024 number is still 2,990 above that, an 8.2 percent overshoot from a year the agency never once called a success when it was happening.

But what NHTSA won’t compute for you is the running tab. If American roads had simply held at the 2019 level from 2020 onward, which was itself unremarkable, the nation would have been spared 23,429 deaths over five years, not a rounding error but a mid-sized arena full of people who did not need to die, decomposed into a data footnote because the comparison class shifted from “before the pandemic” to “the worst year of the pandemic.” The math, year by year: 2,652 excess deaths in 2020, then 6,875 in 2021 at the peak, then 6,366 and 4,546 and 2,990, the long slow bleed back toward normal that NHTSA has rebranded as “progress.”^[2]

The sleight of hand is structural. Every quarterly release compared the current number to the same quarter of the previous year, a chain of year-over-year calculations anchored to a 2021 peak that nobody in 2019 would have predicted. By the time you stack eleven of those declines end-to-end, you have regressed from catastrophe to merely terrible and called the trajectory a victory. The fatality rate of 1.20 per 100 million VMT is the lowest since 2019, NHTSA noted. It is also above the seven-year pre-COVID average of 1.13, a detail the agency included, presumably so no one could accuse it of omitting the context it was clearly hoping you would not pause to absorb.^[1]

For counterpoint: the decline is real, and ignoring it would be dishonest. Eleven straight quarters is not noise. The post-COVID behavior shift was genuinely dangerous, with higher speeds, more impaired driving, lower seatbelt use, and the fact that it did not calcify into a permanent new normal is worth acknowledging. But “we partially reversed a catastrophe we failed to prevent” is a different sentence than “we’re encouraged by the continued decline in traffic fatalities,” which is what NHTSA Deputy Administrator Sophie Shulman actually said. The first is accurate; the second is a press release.

So what should you actually do with this information? When you read a headline that says traffic deaths are falling, ask: falling from what? If the baseline is 2021, when America recorded 43,230 road deaths because everyone decided simultaneously that speed limits were optional and seatbelts were a suggestion, then yes, falling. If the baseline is any year before the national road-rage psychotic break, you are still waiting for the decline to arrive. Check your own state: NHTSA estimates fatalities decreased in 35 states in 2024 but increased in 14 states and the District of Columbia.^[1] Your state might be on the wrong list. And the next time a federal agency celebrates a body count that is merely 8.2 percent above its own pre-crisis benchmark, remember the 23,429 people who occupied the space between the trend line and the spike, the ones nobody put in a press release because they died above the old normal but below the new one.