ESC Saved Truck Drivers. It Barely Dented the SUV Body Count.

I ran a simple cross-tab: every vehicle in FARS, grouped by class, split into two model-year eras. MY2005–2011 (before the ESC mandate) versus MY2012–2018 (after FMVSS 126 made electronic stability control standard).^[1] Then I watched the death counts diverge. Pickups dropped 43%. Vans dropped 44%. Sedans dropped 29%. SUVs dropped 5.5%. Sports cars got 8% worse.^[2]

That is not a rounding error. SUV deaths went from 11,355 to 10,734. A delta of 621 people over seven model years. In that same window, pickups shed 4,661 deaths. And sports cars? They went from 1,547 to 1,672 deaths. An 8.1% increase. The only class on the board that moved in the wrong direction.^[2]

IIHS data says ESC reduces fatal single-vehicle SUV crashes by 56%.^[3] Fifty-six percent should produce a cliff. Instead we got a ramp that barely tilted. Why? Because between 2012 and 2024, SUVs and crossovers went from roughly 33% of new-vehicle sales to over 55%.^[4] Each individual SUV got meaningfully safer. But the fleet multiplied faster than the per-vehicle risk fell. More SUVs, more miles, more exposure, more crashes.

A Jevons Paradox wearing a crossover badge.

Pickups held steady at 18–20% of sales. No fleet explosion to absorb the gains. ESC did its job and the death count fell. Vans shrank. Same story. In our FARS fleet data, SUVs represent 39.6% of registered vehicles but account for only 24.3% of deaths (an overrepresentation ratio of 0.61).^[2] Per vehicle, they are the safest class on the road. But "safest per vehicle" and "fewer total deaths" are different charts entirely.

NHTSA's preliminary 2025 data pegged total traffic deaths at 36,640, a 6.7% decline and the second-lowest fatality rate ever recorded (1.10 per 100M VMT, behind only 2014's 1.08).^[5] Good news, genuinely. But the class-level story says sedans aging out and pickups getting safer are carrying that decline. SUVs contribute a per-vehicle dividend that their own fleet growth immediately offsets.

What This Means for You

If you are buying one vehicle, SUVs remain the individually safer choice. A 2024 RAV4 beats a 2024 Camry on crash survival by almost every metric. But if you are an insurer pricing fleet risk or a regulator projecting total fatalities, do not assume that mandating a safety technology automatically reduces total deaths in the fastest-growing vehicle class. Fleet growth can eat the safety dividend whole. Verify ESC functionality at your next service appointment and check your VIN at nhtsa.gov/recalls for outstanding campaigns.

Methodology

FARS model-year deaths (crash years 2014–2023) grouped by NHTSA vehicle class. Pre-ESC era: MY2005–2011. Post-ESC era: MY2012–2018. Death reduction = (pre − post) / pre. Pickups: (10,810 − 6,149) / 10,810 = 43.1%. Vans: (2,767 − 1,538) / 2,767 = 44.4%. Sedans: (32,055 − 22,700) / 32,055 = 29.2%. SUVs: (11,355 − 10,734) / 11,355 = 5.5%. Sports cars: (1,547 − 1,672) / 1,547 = −8.1%. Fleet share ratios use estimated registrations from industry sales data scaled by NHTS average annual mileage. MY2019+ model years have fewer crash-years of exposure in the dataset and are excluded from the era comparison.

Limitations

This analysis compares aggregate deaths without normalizing by fleet-miles per model year. The 5.5% figure conflates declining per-vehicle risk with rising fleet size, which is exactly the point, but means the per-vehicle improvement for SUVs is likely comparable to other classes. ESC cannot be separated from concurrent safety tech (side curtain airbags, stronger roofs, better A-pillar geometry) arriving in the same model years. FARS covers only fatal crashes, not the approximately 6.7 million annual total. And our death-by-class counts do not distinguish SUV occupant deaths from deaths in the other vehicle.

Strongest Counterargument

The SUV stagnation IS the mechanism by which sedan deaths declined. Every buyer who traded a 2006 Camry for a 2016 RAV4 improved their personal crash odds. Those "new" RAV4 deaths appearing in the SUV column would have been worse had the driver stayed in a sedan. The system improved overall. The SUV column absorbed the people. This objection is valid. It does not change the aggregate math: ESC reduced per-vehicle SUV risk, but the body count barely moved.