IIHS Saved 48,352 Lives. The Vehicles That Failed Its Tests Are Still Killing People.

On Tuesday, the Insurance Institute for Highway Safety threw itself a birthday party. Thirty years of crash testing. To celebrate, the institute bought a pristine 1996 Chevrolet Blazer, lined it up against a 2026 Blazer, and ran them into each other at 40 mph in a moderate overlap frontal collision.^[1]

The dummy in the new Blazer walked away. Bumps and bruises, maybe a sore ankle.

In the old one, the dummy lost its head. Literally. Impact crushed the cabin, shoved the steering column into the dummy's lap, and then the airbag caught it under the chin and snapped the neck backward with enough force to detach the skull from the spine. IIHS noted, with clinical understatement, that "this isn't likely to happen to a human driver" but "illustrates the extreme forces the dummy absorbed."

Alongside the carnage, IIHS released a study quantifying what three decades of testing actually accomplished.^[2] The number is 48,352 lives saved from 1999 to 2024, worth an estimated $538 billion to society. That moderate overlap test alone, the one they just used to behead the Blazer dummy, accounts for the majority of 28,697 lives saved across all three front-crash evaluations, while side-impact testing saved another 18,224. Insurance companies spent $600 million funding all of it, making it a 900-to-1 return that is, by any honest accounting, extraordinary.

But here is the part IIHS did not mention at the party.

The 1996 Blazer earned a "Poor" rating when it was first tested. That rating covers every Blazer built from 1995 through 2004, when Chevrolet discontinued the nameplate. And when we cross-referenced NHTSA's Fatality Analysis Reporting System data for the Blazer, we found that 89.9% of all Blazer deaths in the federal database involve the old model: 310 fatalities in model years 1992 through 2005, versus just 35 in the 2019-and-later relaunch that earns IIHS's top "Good" rating.^[3] The vehicle IIHS just destroyed on camera is the same vehicle still doing the destroying in the real world.

That is not a Blazer-specific problem; it is a fleet-wide structural failure. Across every make and model in FARS, 40.6% of all occupant deaths involve vehicles built before 2005, meaning they were manufactured before or during the first decade of IIHS testing, when most models still earned marginal or poor ratings and the side-impact and roof-strength evaluations did not yet exist, with another 34.8% involving 2005–2012 model-year vehicles. Three-quarters of everyone dying in cars on American roads is dying in vehicles that predate the safety era IIHS is celebrating.

Zoom out further and the pattern sharpens: all five of the deadliest model years in FARS fall between 2003 and 2007, each responsible for more than 10,000 deaths, with model-year 2005 leading at 11,363 fatalities. These were the vehicles rolling off assembly lines during the exact period when automakers were scrambling to redesign around IIHS tests. Redesigns worked, for the people who bought the redesigned cars. But the old ones did not evaporate. They got cheaper, moved downmarket, became someone else's daily driver.^[4]

IIHS knows this. Their own study methodology compares fatality rates for good-rated vehicles against everything else, which means the denominator in their 48,352-lives-saved calculation includes the ongoing body count from vehicles that never earned a good rating in any test. They measured the gap between the best and the rest, multiplied by exposure years, and called the difference "lives saved." Which it is, mathematically. But the lives on the other side of that equation are not hypothetical. They are people driving a 2003 Trailblazer to work because that is what they can afford.

What this means for you

If your vehicle was built before 2012, look up its IIHS rating at iihs.org/ratings. If it earned "Poor" or "Marginal" in any crashworthiness test, understand that you are driving a vehicle whose structural performance is categorically worse than anything built in the last decade. Check your specific VIN for open recalls at nhtsa.gov/recalls. A 2015 Honda CR-V or Toyota RAV4 with 100,000 miles offers dramatically better crash protection than a 2003 Blazer with 60,000, and on the used market they often cost the same.

Limitations

FARS captures only fatal crashes, a subset of the approximately 6.7 million annual crashes nationwide. A vehicle with a low fatality rate may still have a high injury rate. Our model-year analysis uses aggregate FARS data from 2016 to 2020; individual model-year death counts for low-volume vehicles carry wide uncertainty. IIHS derived the 48,352 figure using counterfactual modeling and the value of a statistical life, which represents willingness to pay for risk reduction, not the direct cost of a death.

Strongest counterargument

IIHS testing cannot retroactively fix vehicles that already left the factory. Its mandate is to influence new-vehicle design through market pressure, and by that standard, the program has been spectacularly successful: virtually every new vehicle earns a "Good" rating today versus a small minority 30 years ago. Blaming IIHS for the persistence of old vehicles on the road is like blaming a building code for the existence of buildings constructed before it was written. Blame for that belongs to the absence of scrappage programs, fleet-turnover incentives, or mandatory inspection regimes that most other developed nations maintain. It is a strong counterargument, and it is correct, which is what makes the body count so frustrating.