In 2024, 5,340 people died in traffic crashes involving large trucks in the United States. Of those, 920 were occupants of the truck. The other 4,420 were in another vehicle or on foot, which is 83 percent of the total, and that ratio has barely moved in a decade. NHTSA’s figures put the share of deaths outside the truck at 83 percent in 2014 and 82 percent in 2023.
The asymmetry is not a moral point but a mass and geometry point, and everything the federal government does about heavy vehicles follows from it. It is also why the regulatory record here is stranger than the crash statistics, and worth reading directly rather than through advocacy summaries.
The definitions decide the numbers
A large truck in NHTSA’s crash data is any medium or heavy truck with a gross vehicle weight rating above 10,000 pounds, excluding buses and motor homes, and it covers both commercial and non-commercial vehicles. That is a wider net than the phrase suggests, catching heavy pickups, box trucks and privately owned equipment as well as tractor-trailers. Of the large trucks in fatal crashes in 2023, 71 percent were heavy trucks above 26,000 pounds and 29 percent were medium trucks in the 10,001 to 26,000 pound band. That 26,000-pound line returns later as the threshold the speed limiter proposal used.
Exposure is where the picture gets counterintuitive. In 2023 large trucks were 5 percent of registered vehicles, produced 10 percent of vehicle-miles traveled, and accounted for 9 percent of vehicles in fatal crashes. Per mile that is close to proportional. What is not proportional is the outcome once a crash occurs: 80 percent of large trucks in fatal crashes were in multi-vehicle crashes, against 63 percent for passenger vehicles.
A decade of the truck involvement rate
The involvement rate per 100 million miles traveled by large trucks is the cleanest available series, and it runs 1.34 in 2014, 1.68 in 2019, 1.62 in 2020, 1.75 in 2021, 1.77 in 2022 and 1.63 in 2023, against large-truck travel of 329,858 million miles in the last of those years. The rate rose by nearly a third over the decade to 2022, then fell back to essentially its 2020 level.
The counts moved as well. NHTSA’s April 2025 fact sheet reported 5,472 deaths for 2023 and the research note a year later gives 5,478, because the annual file had been replaced by the final file. A truck fatality total is therefore a statement about a file vintage as much as about a year, a point developed in what the fatality data shows.
A standard raised while its label requirement fell
The configuration that drives the most engineering attention is the one where a passenger vehicle goes under the trailer rather than into it, so the trailer bed reaches the occupant compartment above the car’s own crash structure. The 2023 fact sheet quantifies the exposure precisely if indirectly. In two-vehicle fatal crashes involving a large truck, the truck was struck at the rear in 21.2 percent of cases against 6.3 percent for the other vehicle, roughly three times as often.
NHTSA upgraded the rear guard standard by final rule on July 15, 2022, adopting requirements similar to Transport Canada’s and requiring enough strength and energy absorption to protect occupants of compact and subcompact passenger cars striking the rear of a trailer at 56 kilometers per hour, or about 35 mph. It took effect on January 11, 2023, answered petitions from the Insurance Institute for Highway Safety and the Truck Safety Coalition, and carried out a requirement written into the November 2021 infrastructure law.
The rest of the docket is less linear. In June and July 2024 NHTSA denied a petition for reconsideration of the 2022 standard and a separate petition for rulemaking on rear impact protection. On side underride it published an advance notice of proposed rulemaking on April 21, 2023, again under the infrastructure law, framed around research into effectiveness and feasibility rather than a proposed requirement, and it has repeatedly denied defect petitions about van-type semitrailers without side guards, most recently in April 2025.
Then, in February 2026, FMCSA rescinded the requirement that a rear impact guard carry a permanent certification label, on the ground that labels become illegible in service and had proven problematic for carriers. The structural standard was raised and the field-verifiable evidence that a guard meets it was withdrawn. Both actions have coherent justifications. Together they describe an enforcement problem nobody has solved.
The 2009 brake rule that drew no public argument
The most consequential heavy-vehicle safety rule of the past two decades attracted almost no public argument. In July 2009 NHTSA amended the air brake standard to require about 99 percent of new heavy truck tractors to stop within 250 feet from 60 mph at their gross vehicle weight rating, a 30 percent reduction against the level then required, with 310 feet allowed for very heavy severe service tractors and 235 feet required when lightly loaded.
The agency estimated that a fully equipped fleet would save roughly 227 lives and prevent 300 serious injuries a year, along with more than $169 million in annual property damage. The engineering was unexciting: enhanced drum brakes, air disc brakes, or hybrid systems, several already in commercial use. Shortening stopping distance by 30 percent changes the outcome of every rear-end sequence in the record, and it provoked no argument about driver autonomy, because it regulated the vehicle rather than the trip.
The speed limiter proposal, as the record actually reads
The contrast could not be sharper. In September 2016 FMCSA and NHTSA jointly proposed requiring heavy vehicles above 26,000 pounds to be equipped with a speed limiting device maintained at a set speed, with 60, 65 and 68 mph discussed as candidate settings. FMCSA announced in May 2022 that it intended to proceed. On July 24, 2025 both agencies withdrew both actions.
FMCSA and NHTSA withdraw the September 7, 2016 joint NPRM that proposed to require that heavy vehicles (those with a gross vehicle weight rating (GVWR) of more than 11,793 kilograms (26,000 pounds)) be equipped with a speed limiting device that is maintained at a set speed.
The stated grounds were significant policy and safety concerns and continued data gaps creating considerable uncertainty, and the reasoning deserves engagement rather than dismissal. The baseline number of crashes a limiter would prevent is harder to estimate as automatic emergency braking and forward collision warning spread through the fleet, and those systems may reach the same goal more effectively. A set speed below the maximum posted daytime limit on many roads also raises questions about federal displacement of state speed setting.
What the withdrawal does not settle is the speed differential question, which is the actual safety argument. A truck held below the prevailing speed of surrounding traffic generates passing manoeuvres that would not otherwise occur. Whether the severity reduction from lower impact speeds exceeds the crash increase from differential speed is an empirical question nine years of docket did not answer, and pointing at data gaps rather than at a finding concedes it.
What infrastructure can do that regulation cannot
Three physical treatments address heavy vehicles directly, and each has a written warrant that shows how narrow the intervention really is.
Climbing lanes are triggered by performance rather than preference. Washington State’s design manual sets two warrants, and generally both must be met. A grade that costs a typical truck 10 mph below the posted limit satisfies the speed reduction warrant, and on a two-lane highway the level of service warrant needs upgrade traffic above 200 vehicles per hour with more than 20 trucks per hour among them. The worked example shows the scale: at 60 mph posted, a 4 percent grade drops a truck to 50 mph within 1,200 feet and to 35 mph by 4,000 feet, and the resulting climbing lane runs 4,800 feet. Roughly a mile of pavement to recover 25 mph of truck performance, which is why these lanes appear on mountain corridors and almost nowhere else. The wider effect of truck performance on alignment and cross section is covered in how trucking shapes highway design.
Escape ramps address the downgrade case, and the manual ranks the types by how well they work. Arrester beds of smooth free-draining gravel are the most desirable, sand piles less so because deceleration is high and the sand is weather-dependent, and gravity ramps least of all, because their length and grade create control problems and rollback after stopping. Length comes from a stopping equation using entering speed, rolling resistance and ramp grade, with a 200-foot minimum and a desirable entering speed of 90 mph, on the stated basis that out-of-control truck speeds rarely exceed that.
Parking is the treatment where the shortfall is quantified. FHWA’s survey under Jason’s Law counted 36,222 truck parking spaces at 1,908 public facilities and 272,698 at 6,376 private truck stops, a national total of 308,920. From the information FHWA received, 36 state transportation departments, 72 percent, reported a truck parking problem. More than three-quarters of commercial drivers reported regularly having difficulty finding a safe place to park when rest or sleep was required. Over 90 percent of surveyed drivers and industry professionals reported difficulty finding safe parking between 7 pm and midnight, and over 60 percent between midnight and 5 am. The consequences are traced in the truck parking shortage.
Who is regulated, and who is at risk
One last figure sets the frame. In 2024, 4 percent of large truck drivers in fatal crashes were alcohol-impaired, the lowest share of any vehicle type, against 25 percent for passenger car drivers and the same for motorcycle riders. The count did rise, from 178 to 224, a 26 percent increase against a 3 percent decline across all drivers.
A year earlier, in 2023, large truck drivers in fatal crashes had the second-highest rate of previously recorded crashes among vehicle types, at 19.4 percent, and were less likely, at 6.4 percent, than other vehicle types to have a previous license suspension.
That describes a population under more scrutiny than any other road user, driving vehicles whose braking has been federally upgraded and whose rear structures now have to survive a 35 mph impact, and which still kills four people outside the cab for every one inside it. The risk falls overwhelmingly on people the system does not regulate, and neither the parking survey nor the withdrawn speed limiter docket nor the safety section contains a mechanism for handing it back to whoever created it. Work zones compress the same mismatch into a few hundred feet, as work-zone crash data shows.