Every Age Group in America Is Dying Less on the Road. Except One.

NHTSA released its sub-category breakdown of 2025 traffic fatalities in July, and the numbers are genuinely encouraging for almost everyone. Overall deaths fell 6.7% to 36,640, with speeding deaths dropping 11%, unrestrained occupant deaths plummeting 13%, and children under 15 seeing a 12% decline. Young adults aged 25 to 34 experienced the steepest drop of any cohort at 13%, erasing most of their pandemic-era surge in a single year.^[1]

Then there are Americans over 65. Count barely twitched: 8,060 dead in 2024, 8,039 in 2025. A statistical rounding error, which NHTSA rounds to zero percent change, and no other age bracket came close to that stagnation. One tier below, the 55-to-64 cohort managed a 3% decline; everybody younger did better than that.

Not a single bad quarter buried in seasonal noise, either. Monthly proportions tell the story more clearly than the totals do. In December 2024, the 65-plus group accounted for 23% of all traffic deaths. By December 2025, that share had risen to 24%. November 2025 hit the same mark, the highest monthly proportion in the dataset. Across nine of twelve months, the elderly share of road deaths expanded compared to the same month a year earlier, even as absolute fatality counts fell for nearly every other demographic.^[1]

To contextualize the gap, consider the math. If the 65-plus group had improved at the national average of 6.7%, roughly 540 fewer people over 65 would have died on American roads last year, and if they had matched the 25-to-34 cohort’s 13% decline, the number spared would have been closer to 1,045. Instead, 21 fewer people died. Twenty-one.

An obvious counterargument writes itself, and it deserves full credit: the 65-plus population is growing rapidly. Census estimates put it north of 60 million in 2025, up from roughly 52 million in 2018, as the youngest Baby Boomers cross the threshold and older Gen Xers approach it. More people in the denominator means a flat death count could mask a real per-capita improvement. NHTSA’s sub-category report does not publish age-adjusted rates, so this analysis cannot rule out that possibility.^[2]

But even the most generous demographic adjustment cannot close the gap entirely. A 15% population increase paired with a 0% fatality decrease would imply roughly a 13% improvement in per-capita risk, plausible but still behind the 25-to-34 group’s 13% fatality decline from a stable or shrinking population. And it would still leave the 65-plus cohort as the worst-performing age bracket by a wide margin. Growth in the elderly population explains part of the plateau; it does not explain all of it.

Part of the answer is biology, and it is unforgiving. Human thoraxes become more brittle with age, which means a 75-year-old occupant experiencing the same crash forces as a 35-year-old is categorically more likely to die from it, regardless of the vehicle’s crash rating. IIHS and NHTSA test protocols use crash test dummies calibrated to a 50th-percentile male, roughly 171 pounds, mid-thirties physiology. Nobody tests for an 80-year-old sternum.^[3]

Part of the answer is vehicle choice, and the FARS data from 2014 to 2023 makes the case in brutal arithmetic. FARS reveals a striking lethality gradient correlated with vehicle age and class. Vehicles with the highest crash-to-death conversion rates — Saturn S-Series at 92.4%, Buick Park Avenue at 88.2%, Buick LeSabre at 82.3%, Lincoln Town Car at 80.5% — are overwhelmingly older full-size sedans associated with elderly ownership demographics. Contrast that with modern crossovers and trucks: Ram 2500 at 20.5%, Toyota RAV4 at 42.8%, Chevrolet Traverse at 39.3%. Nine out of ten. That is the lethality ratio for a Park Avenue in FARS-involved crashes, which means that when a fatal crash involves a Park Avenue, the person inside is almost always the one who does not survive, while fewer than half of RAV4 occupants in equivalent crashes share that outcome, because the vehicle itself acts as a multiplier on biological fragility, turning age-related brittleness into a death sentence that modern crossover steel and airbag geometry would have absorbed.^[4]

Modern safety technology compounds the mismatch in a way that produces precisely this kind of divergence. AEB, lane-keeping assist, blind-spot monitoring, and forward-collision warning are crash avoidance technologies that prevent the impact from happening entirely. For a 30-year-old driver in a 2024 RAV4, that means never entering FARS at all; the crash that would have killed them simply does not occur. For a 78-year-old in a 2012 LeSabre, the technology doesn’t exist on the vehicle. Crash happens. And when it does, the occupant’s age and the vehicle’s structure conspire to produce fatality rates that modern crossover buyers will never experience.

What would move the number, and how likely is any of it to actually happen?

First: Accelerating fleet turnover for elderly drivers. If the 65-plus population transitioned from older sedans to current-generation crossovers with standard AEB, the crash avoidance technology alone could prevent hundreds of fatal crashes annually. Economics is the practical barrier: fixed-income retirees are not buying $38,000 RAV4s.

Second: Age-calibrated crash testing, in which IIHS and NHTSA would develop and publish results using dummies that represent elderly physiology: lower bone density, reduced chest deflection tolerance. Europe’s Euro NCAP has explored this concept, but America has not implemented it.

Third: Medical fitness screening, which several states have entirely neglected when it comes to license renewals past age 65. Illinois requires a road test at every renewal after 75, but most states require nothing beyond the standard renewal cycle, and political appetite for mandatory elderly driving restrictions remains approximately zero.

What 2025 lays bare is a truth nobody in the safety establishment is eager to articulate publicly: America’s traffic safety improvements are disproportionately benefiting younger, wealthier drivers in newer vehicles equipped with technology that did not exist when the average 75-year-old bought their car. Seatbelt campaigns worked. ESC mandates worked. AEB is working. But none of them are working fast enough for the demographic that needs them most — the one that absorbs 22% of all traffic deaths while receiving 0% of the annual improvement, because the technology arrives on new vehicles that elderly Americans on fixed incomes do not buy, installed in body structures calibrated to protect bodies that do not resemble theirs.