← The Crash Report
By The Numbers

The U.S. Just Posted a 0.99 Traffic Fatality Rate. The All-Time Record Is 0.98.

Aerial view of a nearly empty interstate highway at dawn, a few cars with headlights on under a pale gold and blue sky

I split NHTSA's first-quarter 2026 death estimates into exposure and rate, expecting the usual story. Instead the rate fell to 0.99 deaths per 100 million miles, one hundredth above the all-time quarterly record of 0.98.[1][4] Americans drove 11.1 billion more miles than the same quarter last year. Fewer of them died anyway: 7,770, down 4.3 percent.[1]

468
Lives saved in Q1 2026 by the fatality rate falling faster than driving exposure grew (author's decomposition of NHTSA early estimates)

Here is the part that made me rerun the spreadsheet. Last year's first-quarter rate works out to about 1.05 deaths per 100 million miles. Apply that rate to this year's driving and you expect roughly 8,238 deaths. The actual count was 7,770, which leaves a gap of about 468 lives saved in a single quarter by the rate falling faster than exposure grew. I ran it twice, and the gap stayed put.

This was the 16th consecutive quarter of year-over-year decline, a streak running back to mid-2022.[3] That was when the roads finally started digesting the pandemic surge, which had produced seven straight quarters of increases.[3] For perspective, 2021 killed 43,230 people, the most since 2005; 2025 is estimated at 36,640, with a rate of 1.10, the second-lowest annual figure ever recorded.[2][4] The pandemic bulge has been more than unwound.

Now the cold water, served at full strength. These are early estimates, and NHTSA revises them when the final FARS file closes. The first quarter is also seasonally the safest, so annualizing 468 into some triumphant yearly figure would be statistical malpractice, which is why I am not doing it. Some of the streak is just the 2021 anomaly draining out of the tub. And the national number hides real divergence: 30 states and Puerto Rico improved, which means the rest of the country did not.[1] Cyclists remain the exception that indicts the trend, with 1,148 pedalcyclist deaths in 2025, up 4 percent.[2]

If your car is more than a decade old, the data has a blunt suggestion for what to do with all of this: NHTSA's 2025 subcategory estimates show deaths in vehicles under 10 years old fell 12 percent, nearly double the overall decline.[5] Newer metal with automatic emergency braking and good IIHS scores is the cheapest life insurance in the dataset. Whatever you drive, check your VIN at nhtsa.gov/recalls; the repair is free, and the deadliest recall is the one you never heard about. And if you ride a bike, ride like the trend is trying to kill you, because it is.

Sources & References

  1. NHTSA, Traffic Crash Deaths: Early Estimates Jan–March 2026, July 2026. nhtsa.gov
  2. David Shepardson, Reuters, “US traffic deaths fell sharply in early 2026,” July 8, 2026. reuters.com
  3. CollisionWeek, “NHTSA Estimates 4.3% Drop in First-Quarter Traffic Deaths,” July 27, 2026. collisionweek.com
  4. Carscoops, “US Road Deaths Are Nearing A Record Low,” July 2026. carscoops.com
  5. NHTSA, Traffic Safety Facts Crash Stats: Early Estimates by Sub-Categories in 2025 (ROSA DOT 92422). rosap.ntl.bts.gov
  6. NHTSA, Fatality Analysis Reporting System (FARS). nhtsa.gov

Source: NHTSA early estimates for Q1 2026 and 2025 subcategory data. Early estimates are revised when final FARS files close; VMT figures are FHWA estimates, not odometer readings. The 468-lives decomposition is the author's calculation: Q1 2026 VMT (7,770 / 0.99) minus the reported 11.1B-mile increase gives Q1 2025 VMT; the implied Q1 2025 rate applied to Q1 2026 VMT yields ~8,238 expected deaths vs 7,770 actual; plausible VMT estimation error moves the figure by a few dozen either way; see methodology for further caveats.