The Most Effective Safety Program in America Is a Car Crusher

NHTSA announced 36,640 traffic deaths in 2025, down 6.7% from 2024's 39,254.^[1] Everyone scrambled to claim credit; the Trump administration pointed to enforcement while IIHS gestured vaguely at technology adoption. Nobody thanked the junkyard.

Cross-tabulate FARS fatality data by model year and a pattern emerges that nobody in Washington wants to discuss.^[2] Model years 2000 through 2009 account for 50.4% of all fatal-crash deaths over the past decade: 94,290 people killed while riding in a single ten-year production window. Peak lethality belongs to model year 2005, which contributed 11,363 deaths by itself. These cars rolled out before electronic stability control became mandatory, before automatic emergency braking existed as a concept, before the IIHS small-overlap test revealed how eagerly their passenger compartments folded inward under asymmetric load.^[3] Many shipped without side curtain airbags. Some left the factory floor already trying to kill their owners.

Now those vehicles are 16 to 25 years old, scrapping at 5–6% per year. Over the past five to six years, roughly a quarter of the MY2000–2009 fleet has been crushed, auctioned to salvage yards, or abandoned to oxidize quietly in fields across the lower 48. At their historical kill rate of approximately 9,429 deaths per year, removing 25% of the fleet eliminates around 2,357 annual fatalities.^[2] Observed 2025 decline: 2,614 deaths.^[1] Fleet turnover alone could account for most of the decline, possibly all of it at the higher end of estimated scrappage rates. Every press conference was a politician standing in front of a car crusher and calling it a policy victory.

Post-ESC-mandate vehicles, model year 2012 and newer, represent just 28.1% of fatal-crash deaths despite constituting a growing majority of registered cars on the road.^[2] The safety technology works. But it is not saving more lives this year than last year; the improvement comes from the most dangerous vehicles finally disappearing from the fleet, one tow truck at a time.

What to do with this

If you drive a 2000–2009 model year vehicle, this is the actuarial case for replacing it now rather than squeezing out another year. Check whether your car has electronic stability control at safercar.gov. IIHS estimates ESC prevents roughly one-third of all fatal crashes.^[3] If your vehicle predates the 2012 mandate and the original buyer did not option ESC, you are driving the statistical problem described above. No amount of defensive driving fully compensates for a structural deficit measured in federal body counts. If replacement is not in the budget right now, make it the priority for when it is: a 2012-or-newer vehicle with ESC is the single highest-impact safety upgrade available at any price point.

Strongest counterargument

Post-pandemic drunk driving may be unwinding on its own. Police traffic enforcement has rebounded from a 33% pandemic-era collapse in stops. Trauma medicine keeps pulling people back from injuries that would have killed them a decade ago. All real factors, and all capable of producing a fatality decline even if not a single old car left the road. But behavioral explanations cannot account for why 65.4% of deaths cluster in pre-2010 vehicles. If enforcement or medical improvement were the primary mechanisms driving the decline, fatality reductions would distribute more evenly across all model years. They do not.

Limitations

Full 2025 FARS microdata will not be available until approximately 2027. This analysis uses the 2014–2023 FARS window to characterize fleet composition in fatal crashes, then estimates the 2025 effect from published scrappage rates of 5–6% annual attrition for vehicles over 15 years old. Actual attrition varies by region, economic conditions, and used-car market demand. Total dataset: 187,058 deaths across all model years, covering fatal crashes only, not the roughly 6.7 million total crashes per year. Multiple causal factors contribute simultaneously, because the universe is not in the business of offering clean single-variable experiments on public roads. Fleet turnover need not be exclusive to be dominant.