One Corporation, 51,687 Deaths: GM’s Multi-Brand Body Count

FARS doesn’t file deaths by corporate parent. It files them by brand. Chevrolet in one column. GMC in another. Pontiac over there. Cadillac somewhere else. Seven separate brand entries, seven separate body counts, and the impression that General Motors is just one of many contributors to American road death.^[1]

Stack those seven columns together and the number is 51,687 dead in FARS-recorded crashes from 2014 to 2023.

That’s 27% of all FARS fatalities. GM holds 20.7% of the registered fleet. The math: a 1.31x overrepresentation ratio, or 145.9 deaths per 100,000 fleet vehicles against a non-GM average of 102.6.^[1] Adjusted for fleet size, GM vehicles appear in fatal crash records at a rate 42% higher than the non-GM fleet.

But the ratio isn’t the real finding. The real finding is what the multi-brand structure conceals.

The Platform-Twin Shell Game

A Chevrolet Silverado and a GMC Sierra roll off the same assembly line in Fort Wayne, Indiana. Same GMT800 frame, same powertrain, same crash structure. Different grille. Different badge. Different FARS entries.^[3]

Combined Silverado+Sierra deaths: 12,928. That’s a single platform’s body count split across two brands so no single brand-level FARS query catches the full number.

It goes further:

• Tahoe + Yukon (same GMT800/900 frame): 3,706 deaths
• Trailblazer + Envoy (GMT360): 3,461 deaths
• S-10 + Sonoma (S-truck platform): 1,805 deaths
• Cobalt + G5 (Delta platform): 2,448 deaths

The Cobalt line is special. GM’s own ignition switch defect killed at least 124 people by the company’s eventual admission, prompting a 2.6-million-vehicle recall in 2014 and a $900 million DOJ settlement.^[4] The Cobalt shows a 5.1 per-100M-VMT death rate in FARS. That defect wasn’t a Chevrolet problem. It was a GM Delta platform problem — shared with the Pontiac G5 and Saturn Ion.

Chevrolet Alone Outpaces Entire Manufacturers

Chevrolet accounts for 35,905 of those 51,687 deaths — 171.0 per 100,000 fleet vehicles. That single brand kills at a rate 67% above the non-GM average.^[1] It exceeds Toyota’s total (22,019), Honda’s (17,837), and Nissan’s (14,456) by wide margins, despite each of those manufacturers holding less fleet share.

Down the GM brand hierarchy, the rates tell a story about who GM sells to:

Chevy at the top. Saturn and Oldsmobile at the bottom. The dead brands with the smallest surviving fleets show the lowest rates — but that’s partly survivorship: the Oldsmobiles and Saturns still registered in 2014–2023 were overwhelmingly garaged second cars or low-mileage retiree vehicles, not daily workhorses accumulating VMT.

The Counterargument, at Full Strength

GM’s defenders have a point, and it’s a good one: GM dominates the full-size truck and body-on-frame SUV segments, which have structurally higher involvement in fatal crashes — higher speeds, rural roads, longer trip distances, heavier mass differential with smaller vehicles on the other end of a collision.^[2] GM also sold millions of cheap sedans (Cobalt, Cavalier, Aveo) to younger, price-sensitive buyers whose demographic profile correlates with higher crash risk regardless of vehicle.

Subaru’s 0.42x death-share-to-fleet ratio and Mazda’s 0.46x aren’t just engineering wins. They reflect a customer base that skews suburban, educated, and middle-aged. If GM sold Outbacks instead of Silverados, its FARS count would look different.

That argument is real. The FARS data cannot cleanly separate “the truck killed them” from “the person who buys that truck was always going to die at higher rates.” We don’t have per-vehicle odometer data. We can’t control for income, education, road type, or annual miles driven at the model level.

But the 1.31x overrepresentation ratio accounts for fleet size. And Ford — which has a similar segment mix of F-150s, Explorers, and cheap sedans — has the same 1.31x ratio. Toyota, with less truck exposure, sits at 1.04x. The segment argument explains some of the gap. It doesn’t explain all of it.

The Methodology

Overrepresentation ratio = (brand death share) / (brand fleet share). GM combined: 51,687 / 191,450 total FARS deaths = 27.0% death share. GM registered fleet: ~35.4M / ~171M total = 20.7% fleet share. Ratio: 27.0 / 20.7 = 1.31x. Per-fleet rate: 51,687 / 35,437,500 × 100,000 = 145.9 per 100k. Non-GM: (191,450 − 51,687) / (171,000,000 − 35,437,500) × 100,000 = 102.6 per 100k. Fleet estimates are derived from state registration data proxied through R.L. Polk/S&P Global Mobility totals, rounded to the nearest 250,000 for sub-brand figures.^[1]

Platform twin identification cross-references GM’s published platform architectures (GMT800, Delta, GMT360, S-truck) against FARS make/model entries.^[3]

What This Analysis Cannot Prove

FARS records fatal crashes only — the roughly 40,000 annual deaths that represent a fraction of ~6.7 million total U.S. crashes per year.^[1] A vehicle with a low fatality rate could have a terrible injury rate. Fleet size estimates use registration proxies, not actual odometer readings, introducing ±15% uncertainty for low-volume models. Some GM deaths occur in pre-2000 model-year vehicles where any manufacturer of that era would show similar structural deficiencies. And crucially: this is a death-per-fleet-vehicle calculation, not a death-per-VMT calculation for the full corporate aggregate, because VMT data isn’t available at the make level across all seven brands simultaneously.

None of which changes the arithmetic. Fifty-one thousand, six hundred and eighty-seven people died in vehicles that shared a corporate parent, a parts bin, and sometimes an identical frame. FARS doesn’t connect those dots. Somebody should.