For the First Time in 30 Years, Most Passengers Dying in Crashes Were Wearing Their Seatbelts

IIHS published its 2024 yearly snapshot in June. Buried in the seat belt table is a line that took 30 years to arrive: 46% of fatally injured passengers were belted, versus 42% unbelted. First time the belted column has ever been the bigger number for passengers.^[1]

This is not a story about belts failing. Flip the numbers around. Unbelted occupants make up roughly 8.8% of the front-seat population, according to NHTSA's 2024 observational survey, which pegs national belt use at 91.2%.^[2] That 8.8% still produces about 48% of known-status passenger fatalities. Per capita, being unbelted makes you approximately 5.5 times more likely to end up in a morgue.

So why did the crossover take until 2024? Passengers approached the line in 2019: 45% belted, 45% unbelted. A dead tie. Then the pandemic blew the curve backward. Belt compliance dropped, reckless driving surged, and by 2020 the unbelted share climbed back to 47%. It took four years to recover what COVID erased in one.

Drivers, by contrast, crossed over a decade ago. In 2013, belted drivers first outnumbered unbelted among fatalities. The gap between driver and passenger crossovers is a real engineering and behavioral chasm. Driver-side belt reminders became standard in the 1970s. Passenger-side reminders followed slowly, sporadically, and with weaker chimes. Rear seats still often have no audible warning at all. NHTSA proposed a rear-seat belt reminder rule that its own research estimated could save between 50 and 1,600 lives per year.^[3] The rule has not been finalized.

Here is the actionable part. At 91% belt use, every marginal percentage point you add converts about 3.3 million more front-seat occupants from unbelted to belted. Given the 5.5x mortality gap, that single point could prevent several hundred deaths per year. No redesign. No recall. No new regulation. Just getting the person in seat 2B to click.

The strongest counterargument: maybe crash severity is rising, and belts that once saved occupants are now overwhelmed by heavier, faster striking vehicles. It is a fair point. The average new vehicle sold in the U.S. weighs over 4,200 pounds, and SUVs and pickups now account for more than 80% of new sales. Side-impact energy scales with striker mass. But if belts were broadly failing, you would expect the unbelted-to-belted mortality gap to narrow. It has not. The 5.5x overrepresentation of unbelted occupants in 2024 is within the historical range. Belts are not getting weaker. The population of unbelted holdouts is getting smaller.

Limitations worth stating: the 91.2% belt-use figure comes from a daytime observational survey of front-seat occupants, which systematically misses nighttime driving, rear passengers, and rural roads where belt use runs lower and fatality rates run higher. FARS codes restraint use from crash reports, which rely on officer judgment and sometimes vehicle inspections. About 12% of 2024 fatally injured occupants had unknown belt status, and if those unknowns skew unbelted, the true crossover year may still be ahead.

The 2024 crossover is a milestone the same way breaking 4:00 in the mile was a milestone. It tells you the trajectory is right. But 7,535 unbelted drivers and 2,023 unbelted passengers still died last year.^[1] That is 9,558 people whose single cheapest, most effective safety intervention was already installed and within arm's reach.