The Recall System Has a Defect: 39% of Fixes Never Happen, and Some That Do Don't Work

Two announcements landed in the same week. NHTSA reopened its investigation into over a million Ram trucks whose rollaway fix may not have worked.^[1] And Kia issued its second recall for 462,869 Tellurides because the first fire remedy failed, producing seven more underseat fires and eleven melting incidents after dealers had already turned the wrench.^[2] Roughly 1.5 million vehicles whose recall "fix" either didn't happen or didn't hold. The recall system, it turns out, has its own defect. Nobody's issued a recall for it.

NHTSA's Report to Congress on recall completion rates puts the number without flinching: only 61% of vehicles recalled by major manufacturers between 2010 and 2022 were repaired by the fifth quarter after the recall launched.^[3] Fifteen months. The best year on record hit 69%. The worst scraped 52%. Four in ten recalled vehicles stay on the road, unrepaired, well after the system has had its shot.

Now apply the second failure. Of the vehicles that do reach a dealer, some fraction leave with a fix that doesn't fix. Kia's Telluride is the cleanest case study. The original defect: a power seat slide knob gets bumped, the motor churns continuously, heat builds under the seat. First recall solution, issued months ago: reinforce the switch cover, replace the switch. Result: fires kept happening. Kia's own investigation found "sporadic dealer workmanship issues" and suspected some technicians reinforced the cover but forgot to replace the switch. Their new fix installs an electronic fuse that trips if the motor draws too much current. It's an engineering solution to the fact that the first engineering solution required flawless human execution, which it didn't get.

Ram's rollaway problem follows the same arc. Truck shifts out of park. No foot on the brake. No key in the ignition. Previous recalls attempted a fix. Twenty complaints later, including injuries and crashes, NHTSA is back. A Ram 2500 weighs north of 7,000 pounds. FARS data from 2014 to 2023 shows that when Ram 2500s are involved in fatal crashes, only 20.5% of the fatalities are Ram occupants.^[4] The other 79.5%? People in the other vehicle, on the sidewalk, on a bike. When a truck that heavy rolls unmanned into traffic, the physics are not ambiguous. The truck survives. Someone else doesn't.

Do the arithmetic and it stings. Roughly 39% of recalled vehicles never see a wrench. For the Ram rollaway recalls covering over a million trucks, that's potentially 400,000 unfixed pickups still parking themselves wherever gravity wants. The ones that were "fixed"? Twenty complaints is enough to reopen a federal investigation. NHTSA can't quantify the remedy failure rate because it doesn't systematically track post-fix defect recurrence. They wait for owners to complain. Which means they're measuring the failures people bother to report, not the failures that exist.

Kia's Telluride sits at the opposite end of the FARS spectrum. With a fatality rate of 0.04 per 100 million miles traveled and only 31 deaths across the dataset, it is among the safest vehicles in America when it's being driven.^[4] The fire risk is a parked-car phenomenon, invisible to crash statistics. FARS can't see a seat motor slowly melting while a family shops at Target. That's the blind spot: the recall system's own failure mode occurs in a space the fatality database doesn't monitor.

What would an honest accounting look like? NHTSA should publish post-remedy complaint rates alongside completion rates. Right now, a recall with 69% completion and a 5% remedy failure rate would leave roughly 33% of affected vehicles either unrepaired or improperly repaired. Nobody calculates that number. Nobody reports it. The system tracks whether you showed up at the dealer. It does not track whether you left with a car that works.

Limitations

FARS captures only fatal crashes, not parked-vehicle fires or property-damage-only rollaways. The 61% completion rate is a fleet-wide average across all major manufacturers; individual recall campaigns vary widely (some exceed 95%, others crater below 20%). Post-remedy failure rates cannot be derived from NHTSA data because remedy effectiveness is tracked only through voluntary complaint submission, not systematic follow-up. The deaths-per-crash ratio for Ram 2500 (0.205) reflects occupant deaths relative to crash involvements, not a controlled survivability metric.

Strongest Counterargument

The recall system, for all its leakiness, does eventually fix millions of vehicles. NHTSA's 2022 completion rate of 69% represents tens of millions of repaired cars. And the Cadillac Vistiq recall this same week shows the system working as intended: GM self-audited after Hyundai's Palisade recall, found the same defect in its own vehicle, and issued a preemptive recall before anyone was hurt. The system's failures are real, but its successes don't generate headlines.

What You Should Do

Check your VIN at nhtsa.gov/recalls today. If a recall shows "remedy available" but you skipped it, schedule the appointment. If a recall shows "remedy complete" and the problem came back, file a complaint at nhtsa.gov/report-a-safety-problem. Post-fix complaints are how NHTSA detects remedy failures. Your report might be the one that triggers a reinvestigation. And if you own a Kia Telluride, park it away from structures until the new fuse fix is available in August.