NHTSA Says Enforcement Saved 2,614 Lives. The Fleet Math Says Otherwise.

NHTSA dropped its 2025 traffic fatality estimate last month: 36,640 dead, down 6.7% from 2024's 39,254. Transportation Secretary Sean Duffy credited law enforcement partnerships and cracking down on "dangerous foreign truck drivers." NHTSA Administrator Jonathan Morrison praised "safety strategies that reduce risky driving behaviors." Nobody mentioned the Chevy Impala finally dying.^[1]

I ran the FARS per-model fatality rates against fleet composition trends. It's blunt. Twelve vehicle models with death rates above 3.0 per 100 million VMT collectively killed 22,160 people between 2014 and 2023.^[2] The Impala (rate: 5.0), the Cobalt (5.1), the Chevy S-10 (4.83), all cheap, all old, all driven hard by people who could not afford something newer. Those vehicles are now rusting in junkyards, and the crossovers replacing them have fatality rates so low they barely register: the Mazda CX-5 at 0.12 per 100M VMT, the Toyota RAV4 at 0.19, the Honda HR-V at 0.13. Over 90 models with rates below 0.5 generated 20,345 deaths across roughly 15 times more total miles driven.

Read that again: fewer deaths on fifteen times the mileage.

Model year data sharpens the point even further. Ford's Explorer killed 503 people in its 2002 model year cohort alone. 2022 cohort? Eight. Tahoe: 311 deaths (2003 MY) down to 14, and the Accord cratered from 444 (2005 MY) to just 17. These aren't enforcement victories, and Secretary Duffy knows it, because nobody gets pulled over for driving a 2005 Cobalt. These are the inevitable result of vehicles with electronic stability control, automatic emergency braking, and crash structures designed after the Clinton administration replacing ones that weren't.^[3]

NHTSA's own quarterly data confirms the timeline problem with the administration's narrative. The decline started in Q3 2022, twelve consecutive quarterly drops before anyone in the current cabinet touched a briefing book. Q1 2025 posted a fatality rate of 1.05 per 100M VMT, the lowest since Q1 2019.^[4] Meanwhile, VMT went up 29.8 billion miles. Americans drove more and died less, and the most parsimonious explanation is that they did it in better cars.

Fair counterargument, and one worth hearing at full strength: DUI enforcement does reduce deaths. Roughly 30% of traffic fatalities involve impairment, and "Drive Sober or Get Pulled Over" campaigns have peer-reviewed evidence behind them.^[5] If the "Freedom Means Affordable Cars" initiative actually reduces new-car prices, that would accelerate fleet turnover, which is the mechanism that saves lives. So the administration might be directionally right about one of its three claimed levers. Just not the one that makes the press release.

One more number to sit with, and it stings. About that 2025 rate of 1.10 per 100M VMT, the "second-lowest in recorded history." That sounds historic until you check the 2014 rate: 1.08.^[6] We lost an entire decade, 2015 through 2024, to distraction, opioids, and a pandemic that turned roads into drag strips. Fleet turnover math is clawing us back to where we were eleven years ago. Progress? Sure. Credit-worthy? Only if you also take credit for the sunrise.

If you are still driving a pre-2012 sedan: you are in the highest-risk fleet segment by a wide margin. ESC became mandatory for model year 2012, and the per-model data shows it working exactly as advertised. A 2022 CX-5 or RAV4 has a fatality rate 25 to 40 times lower than a 2005 Impala or Cobalt. Your VIN is at nhtsa.gov/recalls, and checking it takes thirty seconds before you check the trade-in value.

Source: NHTSA FARS 2014–2023 per-model data; NHTSA 2025 early estimate and 2024 final FARS. Fatality rates are per 100M VMT using estimated fleet size and average annual mileage; uncertainty is ±15% for low-volume models. Fleet turnover is one contributing factor among several; we cannot isolate it from road design, speed limit changes, or ride-sharing effects without controlled studies. See methodology for caveats.