The HR-V and Ram 2500 Both Kill at 0.13 per 100M VMT. Only One Is Actually Safe.

Two vehicles with identical death rates arrived at that number through completely opposite mechanisms, and the difference between them contains a lesson the auto industry would prefer you never learn.

The Honda HR-V records 30.4 fatal crashes per 100,000 registered vehicles across a decade of FARS data.^[1] That is the lowest crash-involvement rate of any mass-market vehicle in the dataset. Its drivers almost never appear in fatal collision records, a population overwhelmingly sober, suburban, and statistically invisible. But when an HR-V finally does show up in a fatal crash, someone dies 53.4% of the time.

The Ram 2500 crashes 2.8 times as often, its drivers hardly paragons of caution, but its crash lethality sits at 20.5% — the lowest of any vehicle with a 500,000-plus fleet. Four out of five Ram 2500 fatal crashes produce survivors, compared to roughly one in two for the HR-V. Same death rate, yet one vehicle earns it through 6,000 pounds of structural engineering while the other borrows it from its owners' behavior and then hopes nobody notices the difference.

This decomposition works across the entire fleet because death rate is really the product of crash frequency (how often a vehicle's drivers appear in fatal crashes per 100K registrations) and crash lethality (what percentage of those crashes actually produce a fatality). Splitting them reveals which vehicles are genuinely protective and which just attract people who don't crash much.

When you pull class averages, the pattern turns structural. Sedans crash at 262 per 100K and kill 63.6% of the time, while pickups crash at 250 per 100K (nearly the same frequency) but kill at only 46.6%.^[2] SUVs land at 161 per 100K and 50.6% lethality. Mass advantage doesn't prevent pickups from crashing, but once the collision happens, all that steel and frame rail absorbs energy that a sedan's unibody transfers directly into its occupants.

Individual vehicles sit in even more alarming positions. The Chevrolet Blazer records 99 fatal crashes per 100K (below median) but converts 71.0% of them into deaths, the highest lethality of any modern SUV with adequate sample size. Buyers who chose a Blazer over a Traverse got a vehicle that crashes at similar rates but kills 81% more often per crash.

Subaru's entire lineup clusters in similarly uncomfortable territory, though you'd never know it from the marketing. The Forester logs 58 crashes per 100K fleet at 55.7% lethality, and the Outback comes in at 95 per 100K and 58.7%, both sitting above the SUV class lethality average of 50.6%. What Subaru owners bring to the equation is discipline; what Subaru doesn't bring is a structural advantage over the segment. The brand's safety reputation is, to a measurable degree, a demographic artifact from the kind of people who comparison-shop all-wheel-drive crossovers in Vermont and then drive them at 4 mph under the posted limit forever after purchasing them.

Run the counterfactual and the illusion evaporates. If you took the HR-V's actual lethality (53.4%) and gave it the median crash frequency for its class (157.7 per 100K), its death rate would explode to 84 per 100K. That is not a safe vehicle wearing a safe death rate; that is a fragile vehicle wearing a safe driver's clothes.

Strongest counterargument: FARS only captures fatal crashes, and vehicles with lower lethality per fatal crash may still produce worse injuries in non-fatal collisions that never enter this dataset. The Ram 2500's mass advantage protects its occupants partly by transferring crash energy to the other vehicle, raising third-party lethality in a way this single-vehicle decomposition cannot capture.^[3] A vehicle that "protects" its own occupants while exporting force into a Civic's passenger cell is safe only in the narrowest actuarial sense, and an honest accounting of that dynamic would shift the rankings substantially.

Limitations: Crash frequency per 100K fleet functions as a proxy rather than a direct behavioral measure, meaning vehicles with higher annual mileage will appear more crash-prone even if their per-mile risk is identical. FARS captures all fatal crashes, not just single-vehicle events, so lethality percentages include multi-vehicle collisions where the other vehicle's mass and speed matter as much as the subject vehicle's design. Fleet estimates carry ±15% uncertainty for lower-volume models.^[4]

What to do with this: Stop reading death rates as a safety score. A vehicle with a low rate might be genuinely well-engineered, or it might just attract cautious owners. If you want to know whether a car will protect you in a crash, not just help you avoid one, look at IIHS crashworthiness ratings independently.^[5] If you're shopping for a subcompact crossover, know that the HR-V's 53.4% crash lethality sits above its class average. And if you drive a Subaru because it's "the safe choice," understand that the math is mostly about you, not the car.