America Matched Its Safest Year Ever. 3,896 More People Still Died.
NHTSA's early 2025 estimate pegs the fatality rate at 1.10 deaths per 100 million vehicle miles traveled, within two hundredths of 2014's all-time record low of 1.08.[1] A triumph worth exactly one press release cycle, because 36,640 people died in 2025 compared to 32,744 in 2014.[2] Nearly four thousand additional corpses at essentially the same per-mile risk, because Americans collectively drove about 305 billion more miles in 2025 than they did when the rate last bottomed out.
I call this the VMT tax. It works like compound interest on a loan you never agreed to. Apply 2025's rate of 1.10 to 2014's VMT of roughly 3,026 billion miles and you get approximately 33,286 deaths. Apply it to 2025's estimated 3,331 billion miles and you get 36,640. That delta of 3,354 people is not a failure of vehicle engineering or driver education or trauma care. It is the arithmetic cost of putting ten percent more miles on the national odometer while holding everything else constant.[3]
To hold the 2025 body count at 2014's level, the rate would need to drop to 0.98 deaths per 100 million VMT. America has never achieved a rate below 1.00. Not once, in the entire history of FARS record-keeping.
Rate Improvement Is Real. It Just Isn't Enough.
Credit where it's earned: if 2025 had carried 2021's post-COVID spike rate of 1.37, the death toll would have been roughly 45,635, which means rate improvements between 2021 and 2025 saved approximately 8,995 lives in a single year.[2] That is genuine, measurable progress driven by ESC adoption, better crash structures, advanced airbag systems, and automatic emergency braking becoming standard across most new vehicles. But VMT growth clawed back about 3,354 of those saved lives, returning them to the ledger as a surcharge for the privilege of sprawl, longer commutes, and an infrastructure model that treats additional driving as economic output rather than systemic risk exposure.
Every one percent increase in national VMT at current rates produces approximately 367 additional fatalities per year. VMT has grown about one percent annually since 2014. So each year, rate improvement has to shave off roughly 367 deaths just to stay even, before it can start actually reducing the count. That's a treadmill, and the speed keeps increasing.
Counterargument, Stated Honestly
A transportation economist would tell you the rate is what matters, and they'd have a point. A driver in 2025 is individually safer per mile than at essentially any point since 2014. VMT growth reflects economic activity, population movement, and personal freedom. Aggregate deaths rising because the denominator grew is a population-level accounting identity, not a per-person risk increase. Framing additional miles as "killing people" conflates systemic exposure with individual hazard. That argument is structurally sound, but it is also a cold comfort to the 3,354 families who bury someone this year who would be alive if Americans drove the same number of miles they drove twelve years ago.
Limitations
NHTSA's 2025 figures are early estimates, historically within plus-or-minus three percent of final FARS counts but occasionally off by more during periods of rapid change. VMT figures rely on state-reported traffic counts extrapolated to national totals, not actual odometer readings. This analysis does not decompose VMT growth by vehicle type, road class, or urban-rural split, any of which could shift the marginal fatality rate of additional miles. Population growth accounts for some VMT increase, so per-capita comparisons would yield a slightly different picture. International comparisons using the CDC's 2019 figure of 11.1 deaths per 100,000 population (versus 4.8 for peer nations) suggest the problem extends beyond VMT, but that analysis exceeds the scope of this piece.[4]
What You Can Do
You cannot personally fix American land-use policy, but you can reduce your household's VMT contribution, because every mile not driven is a mile that cannot produce a fatality. Consolidate trips. If your employer offers remote work, take it; each round-trip commute eliminated removes roughly 30 miles of daily exposure. If you're choosing where to live, proximity to work, school, and groceries is a safety feature that doesn't show up on any window sticker. Check your current vehicle's safety record at iihs.org/ratings and run your VIN at nhtsa.gov/recalls. And when someone tells you the roads have never been safer, ask them: safer per mile, or safer per person, because those are two very different answers.
Sources & References
- NHTSA, Traffic Deaths 2025 Early Estimates & 2024 Annual, June 2026. nhtsa.gov
- IIHS, Yearly Snapshot of Fatality Statistics. iihs.org
- FHWA, Traffic Volume Trends, historical VMT data. fhwa.dot.gov
- CDC MMWR, International Comparison of Traffic Deaths, 2019 data. cdc.gov/mmwr