Your Car's Safety Rating Is Hiding a Second Number

Two metrics define whether a vehicle keeps you alive. One gets published. One doesn't. Death rate per 100 million VMT measures how often a vehicle is involved in a fatal crash per mile driven. Per-crash lethality measures something scarier: when a fatal crash does happen, how likely are you to be the one who dies?^[1]

Four vehicles sit in what I'm calling the false-safety zone: low death rate per VMT, lethality above 0.70. Nissan Kicks (rate 0.69, lethality 0.739). Nissan Versa (rate 0.90, lethality 0.723). Honda Fit (rate 0.72, lethality 0.720). Hyundai Accent (rate 0.80, lethality 0.717).^[1] All subcompacts. All under 2,800 pounds. All looking statistically reasonable until you ask what happens at impact.

Compare those to vehicles that are safe by both metrics. Toyota RAV4: rate 0.19, lethality 0.498. Chevy Traverse: rate 0.20, lethality 0.393. Ram 1500: rate 0.13, lethality 0.341.^[1] Not only do these vehicles crash fatally less often per mile, but when they do, their occupant survives about 60-66% of the time. A Honda Fit occupant survives 28%. Physics is not impressed by your fuel economy.

At the class level, the hierarchy is stark. Sedans: 0.645 average lethality across 89,127 deaths. SUVs: 0.524 across 46,442 deaths. Pickups: 0.489 across 41,593 deaths.^[1] IIHS has documented this for years, noting that driver death rates in mini cars are roughly double those of very large cars.^[2] Weight remains the single strongest predictor of occupant survival in multi-vehicle crashes.^[3]

Why does the death rate flatter these vehicles? Because their owners drive fewer miles, mostly in urban settings, at lower speeds. A Versa commuting 8 miles to work and back accumulates little VMT exposure. That produces a low rate. But VMT is the denominator, not a safety feature. Strip it away, and you see what the structure delivers when crumple zones run out of crush space at 2,500 pounds.

Strongest Counterargument

If Versa owners genuinely drive less and slower, their real-world risk IS lower. Rate captures actual probability of dying, which is what matters to a buyer calculating lifetime risk. That's true. But it creates a false equivalence: a RAV4 at 0.19 rate is safer by BOTH metrics. A Versa at 0.90 rate is safer only so long as nothing goes wrong. When something goes wrong, you're in a 2,500-lb penalty box facing opponents that average 4,000+.

What You Can Do

If you drive a subcompact: Know that your vehicle's low death rate comes from your driving pattern, not from structural protection. Avoid interstates and high-speed roads where the mass mismatch is most lethal. If you're shopping: Don't compare a Kicks (rate 0.69) to a RAV4 (rate 0.19) and think the gap is 3.6x. It's 3.6x on rate AND the RAV4 is 48% more survivable per crash. Both metrics matter. Check your VIN at nhtsa.gov/recalls, and consider that an extra $3,000 for a compact SUV might buy structural margin no safety feature can replicate.

Limitations

FARS captures only fatal crashes (roughly 1 in 180 total crashes). Per-crash lethality measures only the probability of being the fatality in an already-fatal event, not overall crash injury risk. VMT estimates carry approximately ±15% uncertainty for low-volume models. Fleet age differences are not controlled: Honda Fit spans 2007-2020 model years while Nissan Kicks starts at 2018, meaning the Fit's lethality reflects more pre-modern-safety-era units. Crash opponent mix is also uncontrolled: subcompacts may disproportionately encounter larger vehicles, inflating their lethality independent of structural design. Low sample sizes (Kicks: 339 deaths) produce wider confidence intervals than high-volume vehicles.