America Had Its Safest Six Months in a Decade. Then Summer Happened.

January through June 2025, the United States posted a traffic fatality rate of 1.06 deaths per 100 million vehicle miles traveled, a number NHTSA hadn't seen at midyear since 2014.^[1] Fifteen consecutive quarters of decline. Roads safer than any point since Obama's second term. The kind of number that gets press releases and podium speeches and a line in the Transportation Secretary's talking points about how American roads are finally healing.

Then July arrived. And the math collapsed.

NHTSA’s full-year 2025 estimate landed at 36,640 fatalities, a rate of 1.10 per 100 million VMT, second-lowest in recorded history, behind only 2014’s 1.08.^[2] That sounds like continued progress, and it is. But buried inside the annual number is a brutal structural fact: if H1 produced roughly 17,140 deaths at 1.06, then H2 produced approximately 19,500 at an implied rate around 1.14. The second half of the year was 14% deadlier per mile than the first half, a gap wide enough to bury the best six months in a decade under a pile of summer coffins.

I ran the counterfactual because nobody else seemed interested. Had the July–December period maintained January–June’s 1.06 rate, the full-year death toll would have been approximately 34,280 instead of 36,640. The difference is about 2,360 human beings who would be alive if Americans drove in August the way they drive in February. That’s not a rounding error. It is more than the entire annual traffic death toll of thirteen states.^[3]

What Kills in Summer That Doesn’t Kill in Winter

The seasonal pattern is not noise. NHTSA’s own subcategory data shows that alcohol-involved fatal crashes spike sharply from Memorial Day through Labor Day, with July 4th alone producing 438 fatalities across a 78-hour measurement window in recent years. Motorcycle fatalities, which account for roughly 14% of all traffic deaths despite motorcycles representing 3% of registered vehicles, concentrate overwhelmingly between May and September when ridership peaks. And recreational VMT surges in summer months, putting unfamiliar drivers on unfamiliar roads at higher speeds, a combination that does not end well and does not discriminate by brand or model year.

But the seasonal death tax is not just about more people on the road, because the per-mile rate itself climbs in every measurable dimension once the thermometer crosses seventy degrees. Americans don’t merely drive more in summer; they drive worse, propelled by longer daylight hours that reward speed, by warm-weather alcohol consumption that rises steadily from May through September, and by teen drivers who flood the fleet between June and August carrying the least experience and the most bravado. NHTSA’s H1 2025 data showed fatality decreases in 38 states, but Hawaii posted a staggering 47.8% increase, Kansas rose 21.2%, and Idaho climbed 18.6%, all states where summer recreational driving dominates the VMT mix.^[1]

The Floor Nobody Reaches

The H1 2025 rate of 1.06 matters because it proves something uncomfortable: America already knows how to achieve sub-1.08 driving, the standard set as the all-time annual record in 2014, but we have never once sustained it past June in any year for which records exist. Every year, the first half builds a safety dividend, and the second half spends it, plus interest, on barbecues and road trips and one more beer at the lake house before the drive home.

If the 1.06 rate were the annual baseline, it would save roughly 2,360 lives per year at current VMT. Compounded over a decade, that is 23,600 people, roughly the population of a mid-sized college town. The infrastructure to achieve this rate already exists for six months. The question is whether it can exist for twelve, and the answer, for every year on record, has been no.

The Strongest Case Against This Math

VMT is not evenly distributed across halves. FHWA preliminary data typically shows H2 carrying 51–52% of annual vehicle miles traveled, driven by summer vacation and holiday season traffic. If H2 has 2–3% more VMT, a proportional share of its excess deaths is simply exposure, not behavioral deterioration. Adjusting for a 52/48 VMT split reduces the seasonal death tax from 2,360 to roughly 1,800, still substantial, still more than Wyoming, Vermont, or Rhode Island’s entire annual toll, but meaningfully smaller than the headline number.

There is also a weather argument: H1 2025’s record-low rate may have been aided by severe winter storms in January and February that suppressed discretionary travel in the Midwest and Northeast. A particularly harsh winter would artificially compress H1 deaths by removing marginal trips from the exposure pool, making the seasonal gap look wider than it structurally is. I cannot fully rule this out without month-by-month VMT granularity that FHWA has not yet released for 2025.

What You Can Do

If you drive between Memorial Day and Labor Day, you are statistically operating in America’s deadliest driving window at America’s deadliest driving rates. The per-mile risk does not merely increase because more cars are on the road; it increases because the cars on the road are driven by people who are more impaired, more distracted, more hurried, and less familiar with where they are going.

Check your VIN for open recalls before any summer road trip at nhtsa.gov/recalls. Avoid driving between midnight and 4 AM on holiday weekends, the window with the highest per-hour alcohol-involvement rate in FARS data. And if your vehicle predates 2012 and lacks electronic stability control as standard equipment, understand that you are operating without the single technology IIHS credits with the largest per-vehicle fatality reduction of the last two decades.^[4]

Source: NHTSA FARS 2014–2023 and NHTSA early estimates for 2024–2025. H2 death counts are derived by subtracting NHTSA’s H1 estimate from the full-year estimate; individual H2 quarterly breakdowns were not available at publication. VMT split assumptions are based on historical FHWA seasonal patterns. See methodology for caveats.