The Federal Car Seat Test Doesn't Include a Car

FMVSS 213 is the federal safety standard that governs every child restraint system sold in the United States. It determines whether the plastic shell cradling your infant can survive a crash. The test procedure bolts the seat to a flat bench, fires it down a sled track at 30 mph, and measures what happens to the dummy inside. The bench has no front seatback in front of it, no headrest, no reclining mechanism, nothing that resembles the interior of an actual vehicle.

That omission sat quietly in the federal code for decades until a 2026 SAE International paper pulled it into the light.^[1] Researchers tested what happens to rear-facing infant seats when a front seatback, occupied and loaded, yields rearward during a rear-end collision. In a real car, with a real adult sitting in front of a rear-facing infant seat, the seatback moves backward on impact, compressing the infant seat's plastic shell from above and behind. The shell, which was never designed to resist loading from that direction, collapses and deforms until head injury risk for the infant spikes.

Nobody mentions the flip side: a separate study found that in frontal crashes, placing a rear-facing seat closer to the front seatback actually reduces head injury by roughly 60 percent.^[2] Every millimeter of gap between the car seat and the front seatback up to 70 mm adds 5.3 points to the Head Injury Criterion score, which means touching or nearly touching is actually safest. So the positioning that protects an infant in a head-on collision is the same positioning that exposes them in a rear-end hit, and the federal test for car seats evaluates neither scenario because the test bench lacks a front seatback entirely.

Rear-impact collisions account for roughly 6 percent of fatal crashes, far behind frontal and side impacts, but they still kill about 2,500 people annually. That percentage is small enough that regulators can argue the risk doesn't justify restructuring a compliance regime that governs a $10 billion market. Car seat manufacturers can point out that adding structural reinforcement to resist front seatback intrusion would make seats heavier, pricier, and potentially harder to install correctly, which could increase misuse rates and create a different safety problem. Both arguments have some empirical support, and ignoring them would be dishonest.

But something should bother you about a testing protocol that evaluates child safety equipment on a flat bench without the vehicle interior components that define the crash environment. European standard ECE R129 includes a test with a representative vehicle seatback for forward-facing seats. The U.S. standard has no equivalent for any orientation. The SAE researchers demonstrated that infant seats designed to withstand realistic seatback intrusion conditions can improve outcomes, which means redesigned seats exist but the test to validate them does not.

If your infant rides rear-facing behind an occupied front seat, you are relying on a plastic shell that passed a test without the single object most likely to kill it. Position the seat behind an unoccupied front seat when you can. Push the front seat forward when you cannot. Inspect the shell after any rear-end collision, even a fender bender, because the damage that matters is the kind you will not see until the next crash.

Source: SAE International Paper 2026-26-0554 and NHTSA FARS 2014–2023. Rear-impact fatality share is approximate from FARS multi-year data. The SAE paper’s full methodology is behind a paywall; this article relies on the published abstract. Redesigned infant seats referenced in the paper may not yet be commercially available. See methodology for caveats.