How America’s Staggering Traffic Death Rate Became Matter-of-Fact

How did more than 30,000 annual motor vehicle deaths become something that most Americans accept as normal? A new paper by Boston University professor Itai Vardi tries to answer that question.

A graph from a 1933 publication by Traveler's Insurance Company pins the blame for traffic deaths on individuals. Image: Vardi, 2014
A 1933 publication by Traveler’s Insurance Company pins the blame for traffic deaths on individual decisions without noting structural factors like the skyrocketing rate of driving itself. Image via Itai Vardi

Vardi reviewed American attitudes toward the problem of traffic deaths, starting in the early era of automobile growth, when there was a great deal of “moral panic” about the carnage on the nation’s streets.

His work is in a similar vein to University of Virginia professor Peter Norton, whose book Fighting Traffic recounts how the forces of “motordom” reshaped American streets by changing how people thought about cars in the city. Like Norton, Vardi has identified key conceptual frameworks that eventually led people to adopt the “matter-of-fact” tone we use to discuss today’s staggering rate of traffic deaths.

Vardi’s research encompasses historical accounts from media outlets, auto and insurance industry publications, activist groups, and, eventually, federal safety agencies. Here are three big factors that, according to Vardi, shaped the modern American view of traffic violence.

1. Thinking of traffic deaths in terms of fatalities per mile driven

Vardi says early, passionate accounts of traffic deaths framed the problem in terms of individual tragedies — the death of a child, for example — or listed the total number of annual casualties. Implicit in these reactions was a fundamental uncertainty about the widespread adoption of cars and their place in cities.

Emphasizing the total number of fatalities or individual cases dramatized the issue and often led to bold headlines. The Los Angeles Times editorial board, for example, told readers in 1922: “Blame the buzzwagon for nearly 15,000 violent deaths in America last year.”

For obvious reasons, car companies and car enthusiasts were threatened by that kind of outcry. Auto industry publications preferred to frame the problem in terms of deaths per mile traveled. This way the industry could emphasize safety improvements and downplay the danger of allowing driving to become widespread.

Eventually auto industry advocates were able to advance their framework and silence “moralist” objections to the expansion of automobility. A big victory for the auto industry came in 1924, Vardi reports, when the National Conference on Street and Highway Safety, convened by President Herbert Hoover, agreed to adopt the National Automobile Chamber of Commerce’s preferred safety metric: fatalities per vehicle.

Later, when federal agencies like the National Highway Traffic Safety Administration were formed to address the problem, they embraced the deaths-per-mile metric — in part, Vardi says, because it helped demonstrate that the agencies’ programs were having an effect.

Vardi says fatalities-per-mile “remained through these years the central statistical means by which to demonstrate progress in safety or counterbalance alarmist claims about accidents.”

There’s a vast gulf between the moral urgency of the Los Angeles Times 1922 headline and a modern NHTSA report. In the 1990s, an update from the agency read: “1992 was a very good year for traffic safety … on a vehicle-miles traveled basis, it was the best ever.” That year, 39,250 people were killed in traffic.

2. “Saving Lives”

Once federal agencies were created and charged with reducing motor vehicle deaths, they began using statistical forecasting to predict how many lives would be lost in the future. This allowed agencies like NHTSA to compare the forecast to actual traffic deaths and claim the difference was “lives saved” by their programs.

Vardi calls “saving lives” — which is actually part of NHTSA’s motto — “a rhetorical device to meet institutional goals.”

Forecasting future deaths, Vardi writes, also sidesteps the tricky question of what is an acceptable number of deaths.

This statistical exercise “advances ideas of normal or reasonable death rates, which… may downplay mass carnage,” says Vardi.

The framework “risks furthering the construction of the problem as one that can only be affected to a measured extent, never systemically,” he writes. “More radical challenges to the socio-technical arrangements that regularly produce mass casualities” — calls for less driving altogether, for example — “are omitted from the conversation.”

3. Seatbelts and Drunk Driving

Finally, once highway safety was placed in the hands of “dispassionate” federal agencies, they framed the problem as one of individual mistakes or mechanical failures, rather than systemic flaws. This paradigm was, ironically, advanced by the Ralph Nader-led reforms of the 1960s aimed at car manufacturers, Vardi says.

For example, the top chart, published in 1933 by the Travelers Insurance Company, omits structural contributions to the high rate of traffic deaths — such as street design and poor non-automotive travel options.

As Vardi writes, absent from these charts is “the growing reliance on automotive transportation at the expense of other modes of conveyance, the culture of speed, the marketing messages that encourage risk-taking, the mythologization of the automobile as a sublime technology, the constructed symbiosis between personal identity and one’s car.”

41 thoughts on How America’s Staggering Traffic Death Rate Became Matter-of-Fact

  1. Vardi’s paper (which sits behind a paywall, by the way) appears to be in the tradition of Wolfgang Sach’s masterful history of the automobilization of Germany, “For Love of The Automobile.”

    Sachs’ book, written in German and published in 1984, and republished in English translation in 1992, was life-changing for me when I read it the following year. You can buy it from U-C Press, My review of it, published in 1993, is here:

  2. Hi, Charlie. I’m at work, so I can fetch the Vardi paper courtesy of Big Government.

    The Vision Zero paradigm-shift holds the built environment accountable with a pithy expression “…where the individual fails, the system should not…” or some such. It holds the traffic engineers, as subject matter experts, to a higher standard. That isn’t quite enough because in many VZ nations, drivers are held to much higher standards than in the States. I think one needs to attack the crash and fatality problem from the standpoint of how we design transportation systems, how we sell and market cars, and how we hold individuals accountable for their behavior.

    That said, one has to measure fatality rates in some way to get a number and ask how we are doing. Maybe deaths per exposure hour would be better, as it better reflects time at risk rather than how fast one is going (a bicyclist may be on the road the same number of hours as a motorist, but will have fewer miles total unless the average speed limit for all vehicles in the cyclist’s environment is that of a bicycle).

    I’ve run into this question in the context of gun deaths and gun control. Why are 30,000+ gun deaths (20k suicides, 10k homicides, a few k other) a national horror but 30,000-plus traffic deaths a reasonable price to pay for individual mobility? Because unfettered car use that kills 30 thousand is acceptable, but unfettered gun use that causes 30k deaths is not? Heck, there isn’t even a Bill of Rights statement about the right to keep and drive cars….and dead is dead. Its not just the gun rights people who will have their guns pried from their cold, dead hands. Try prying the car keys out of those motorists hands. We have accepted carnage on the roads as a necessary evil.

  3. I’m not sure whether “acceptable” is the right word for our national attitude toward automotive-related fatalities. “Tolerable” might be closer. Drivers vote, and any politician advocating stricter licensing standards or liability laws might find that he or she would be out in the next election. Californians may remember the long-drawn out fight for stricter laws and punishments for drunk driving. Comparing motor vehicles to firearms is a bad analogy–guns are weapons, and cars are a means of transportation. Police officers in the quieter areas may go from rookie to retiree and only fire their service weapons at the gun range, but they typically spend hours a day in their patrol cars. Streetsblog emphasizes non-automotive transport, but for many, if not most Americans, their daily commutes and errands are too far to walk, too tiring to bicycle and too time consuming to use transit. I remember one official quoted as saying, “Taking away a person’s driver’s license is like sentencing them to house arrest.”

  4. Guns and cars are both hazardous to the public when in the wrong hands. The fact that our entire transportation system is based on cars, and dependent entirely on individual responsibility coupled with near universal access, has a price. As does near universal access to guns.

  5. The real question should be why did we ever move to a more inherently dangerous system like automobiles driven mostly by barely trained novices in the first place? Our present built environment never would have existed if we had not accepted that paradigm. Even in the 1920s we saw the handwriting on the wall, namely the fact the vast majority of people just cannot safely drive, and therefore shouldn’t be allowed to. Perhaps steps should have been taken to limit automobile ownership back then, or better yet limit production of automobiles to just enough for things like taxi services. Just the fact automobiles used smelly, noisy internal combustion engines should have been enough to prohibit their use in and around major population centers.

  6. The statistic of 33,000 or so traffic deaths a year is only the tip of the iceberg. American economic life hinges upon hiding externalities of production and consumption. How many additional Americans die each year because of pollution and an extreme sedentary lifestyle fostered by the autocentric American experience.

    But the American lifestyle is not only autocentric but autocratic. I have been driving in the City of New York for 30 years, but started cycling in the City of New York for only 3 years. In the 30 years of driving in the city I have never once encountered police enforcement not once. Since cycling I have been stopped twice, and had to appear in court once (I was found not guilty). Amazing priorities.

  7. There is a gentleman from Toronto, Kevin Love, who often posts here who noted that the direct traffic deaths in North America are probably dwarfed by the indirect ones that are caused by air pollution in our cities. These include stuff like asthma and other cardiopulmonary diseases caused or exacerbated by ground level ozone and other air pollutants associated with restricted air flow and heavy traffic. Kevin has some direct references, if he is reading this.

  8. But death from degenerative disease doesn’t have the same “If it bleeds, it leads” emotional response as a twisted car or bike and a body under a sheet in the street.

  9. “Perhaps steps should have been taken to limit automobile ownership back then” (the 1920s). But by then, Henry Ford’s factories were rolling out Model Ts by the trainload, then General Motors (after a corporate near-death in the early 20s) started GMAC financing and annual model changes. The Model T started to fade as closed cars became more affordable, and its chassis couldn’t be easily scaled up for heavier vehicles. Improved tires, more efficient and powerful engines and other technical improvements kept car sales growing, while the older models formed a secondary market for folks who couldn’t afford a new “ride”. We could ask, “Who would have the authority to limit cars sales?” States issued driver’s licenses, and one would suspect that the automobile industry would do some subtle (or not so subtle) lobbying to make sure that every person who wanted a car could be licensed to drive one. Restricting production would smack of “socialist planning”–anathema to the US even today.

  10. You can restrict access to goods or services in democratic countries for public safety reasons. I tend to think 15,000 annual deaths from motor vehicles in the early 1920s would be sufficient grounds to do that. Unlike today, since cars were new then restricting car access wouldn’t have resulted in any significant mobility restrictions for the vast majority.

    Or you could just tax cars enough so few can afford them. Or add safety requirements which make them unaffordable. Had we required disk brakes, air bags, and a whole host of safety features standard on cars now back then I’ve little doubt it would have rendered them unaffordable, perhaps even technologically unfeasible.

    A third approach would have been towns and cities restricting motor vehicle access. The more places you can’t go, the less point there is to owning an automobile. This third approach might be the best shot we have today for ending the concept of mass motoring. Large cities could outright ban cars, or require zero emissions vehicles, or require special additions on drivers licenses requiring extra training for operation within city limits. The sheer number of deaths/injuries caused by motor vehicles, along with an order of magnitude more affected by air pollution, certainly provide sufficient public safety grounds for this. The auto manufacturers can lobby all they want but in the end universal car access isn’t a right or a necessity. If 30,000+ people were killed on trains or planes annually the NTSB would just shut those modes down indefinitely until a way could be found to make them much safer. Perhaps we need to do this for motor vehicles. Tell GM et al we won’t allow motor vehicles to go another mile until a way is found to bring the number of annual deaths down by maybe 3 orders of magnitude (that includes pollution deaths). If they can’t find a way, so be it. We’ll just all get around by walking, biking, or trains while restructuring society to reflect this new paradigm.

  11. Joe, you don’t have the experience of being around lots of horse transportation to judge cars as worse. Horses can be unpredictable and dangerous, causing many deaths. They are expensive to keep and produce their own pollution. People then were not stupid. Cars were clearly the better option or people would not have overwhelmingly chosen them.

  12. Seriously? You don’t understand how being confined to tracks is a limitation especially at a time when populations and density were lower?

  13. On Streetsblog network there is an article on how car insurers need to be taken to task for not charging more by the mile driven. So which is it? Do miles matter or not?

  14. Rails ran to a lot more places back then than they do now. How do you think most people got around? Few people took a horse and wagon on trips of more than a few miles, other than in rural areas. They either used their feet, a bike, subways, trolleys, interurbans, or steam trains. There were few places of interest which couldn’t be reached by those modes.

    Of course, we started dismantling this system in favor of an auto-oriented one after WWII but prior to that we could have limited automobile ownership without severely limiting mobility.

  15. As someone who lived next door to an electric railway for the first11 years of my life, and who has been active in railway preservation for over 50 years, I would love to take us back to the days of interurbans and streetcars, but I think that many people have this attitude toward rail transportation: “Good! Get all those other bozos to ride the train so I can have the freeway to myself.”

  16. While I think this is a fascinating article and I’m saving it for inclusion in a public health course I’m planning, I find it face-palm-inducing that, throughout an article examining how the use of quantification has shaped the construction of the problem of traffic-related trauma…the author uses the word “accident,” a word that is both a product of that same social construction and inaccurate. (I’ll admit to skimming, but I saw nowhere where the author addressed this as a conscious choice of word to demonstrate a point).

  17. You are correct that risk per hour of exposure would be a better measure. This is the measure that insurance companies use and they have a strong motivation to accurately access risk.

    The planning community should demand that the NTSB and other such governmental agencies access risk in a more neutral/accurate manner by reporting risk per hour of exposure instead of risk per mile traveled.

  18. Yes, I understand Angie was quoting the original. But why did the original author use “accident” without quotes or comment? In seems strange in an article about the social construction of the traffic crash problem to use a term that is a product of that construct (rather than an accurate term) and not at least note that he is aware of that and choosing to use it for some specific purpose.

  19. interurban electric rail existed in a dense network throughout the country.

    These systems formed a part of a efficient, privately owned, fully integrated transportation network that included local streetcars, subways, regional rail, and long distance high speed rail

    Even a low density region such as Ohio had a fantastic system

    The Southern California Inteturban system also is reknown

    These are two examples of safe regional electric railway systems that connected small towns with large towns and cities.
    They typically had many stops in the towns.

    They had frequent service, door to door travel time was often faster than today’s driving speeds and were safe.

    The German S-Bahn systsm has many similar characteristics

  20. Nobody’s killing *other* people with opiates. You can’t always stop people from killing themselves. You can stop them from running over innocents.

  21. The NYPD’s a criminal gang, and for some reason has decided to go after cyclists (I guess they got bored with harassing and murdering black people).

    At this point, I say shut the NYPD down. Fire everyone. Hire a fresh police department, and start arresting the former NYPD gang members.

  22. All true. Unfortunately, new technology in the form of motor vehicles became more affordable without the confines of rail rights of way and privately owned systems lost riders, became financially unsustainable, and went bankrupt. The new technology of rail similarly put canal traffic out of business in many places too. The CD seriously damaged LP music sales, and the CD itself is falling to downloads and now streaming, yet the old relics still have small places left. LP record playback equipment today is better than ever before and I’m sure trains are too, but that’s not what the public today most wants.

  23. wrong

    unfortunately the motor car was heavily subsidized by tax dollars

    Still true today, Gas Taxes and other user fees cover less than 1/5 the cost of mass motoring

    Wonder how much VMT would decline if Gas cost $10 a gallon ?

  24. Got a citation for your claim, because in the US, I think you have doubled the numbers? If you compare to the subsidy for public transit, you will find that driving is less subsidized, and yet it is still much more popular than transit.

  25. You have really lost perspective on traffic deaths compared to others, like hospital infections, medical errors, and drug overdoses. How many doctors and nurses go to jail for killing patients when they have accidents? Far fewer than drivers, even though they kill in greater numbers, with hundreds of thousands dying from hospital acquired infections and medical errors. In my state, 4 times as many people die every year from opiate overdoses as from all road accidents. So, instead of cops doing traffic enforcement, let’s have them monitor hand washing at hospitals and save more lives for it!

  26. You’re comparing apples and oranges. People who go to hospitals are typically very sick. Many of them would have died anyway, even with medical treatment. Try to prove medical malpractice killed a 90 year who likely already had one foot in the grave. On the other hand, when a young, healthy person is suddenly dead for no other reason than being hit by a motor vehicle, not hard to show the cause.

  27. Its precisely older people who are most likely to die as pedestrians, cyclists, and motorists as a result from injuries in collisions also! Care to discount their deaths when they die? How many of them don’t count?

  28. If they’re able bodied enough to walk or ride a bike chances are good they had many years of life left. Take a look one day at your usual hospital demographic. Large numbers of patients are chronically ill. I’ve heard a lot of so-called medical malpractice is just very ill patients reacting badly to established medical procedures. What is accomplished here by punishing doctors? Punish enough doctors for something they may not really have directly caused, you’ll have a lot fewer people wanting to be doctors. End result is lots of people who might have lived die for lack of doctors to treat them.

    On the other hand, if you punish drivers for something they clearly caused the end results are only positive. If the punishment is permanent license revocation you have a proven dangerous driver off the roads for good. If this draconian system results in more people deciding not to drive, you have less congested roads, far fewer deaths. That’s the difference. In one case you can clearly save lives. In the other, the punishment may well have the opposite effect.

    Note that I’m not necessarily a big advocate of jailing drivers who kill people like many others here. In some cases maybe if extreme recklessness was the cause, but in most cases I think taking away their license for good is punishment enough. It should be standard to do so if you kill or seriously injure someone while driving through recklessness, negligence, or incompetence.

  29. I think you missed the point here. We’re not saying motorists are always at fault for being motorists. That’s sorta what NHTSA is trying to say: individuals are to blame. The point of this is there are systemic causes.

  30. Let me tread lightly, but I wonder whether our obsession with “crash not accident” does safe street advocates any good. We burn up resources fighting a term in such common parlance that its original meaning of happenstance and chance comes in a far second to its now-accepted usage of “vehicle crash.” Changing the term means we buck the overwhelming tide of language, with the likely effect of a few newspapers revising their style guides. Meanwhile we take energy away from substantive, on-street changes that would actually save lives.

    Further, to the extent language does matter, a far bigger problem than “accident” is the blithe way most of us talk in terms of “cars crashing,” rather than “drivers crashing.” Almost any headline featuring traffic crashes will banish drivers from the scene. This usage comes so naturally to us that even the horrific terrorist attack in Nice had at least one headline to the effect, “Truck kills dozens.” If we are going to spend our effort on language, let’s at least spend it in a way that maximizes responsibility.

  31. Funny thing about drunk pedestrians is that they’re only killed by car crashes — not by other other pedestrians crashing into them, nor by bicycles crashing into them. Cars are inherently dangerous because of their size and speed, and magnify the consequences of any mistake around them. People make mistakes, and obvious, normal mistakes should not be deadly.

  32. High real estate prices for high density development caused people to seek space more distant. Health concerns and contagious diseases also caused people to flee firetrap tenement buildings. The flight was first enabled by streetcars and trains. Queens used to be farmland prior to elevated train lines. People were eager to leave the congestion and horse excrement filled streets. Automobiles allowed more people to afford to flee unhealthy cities and live where there weren’t train lines.

  33. High property costs drove longer commutes made possible by cars. Low density is safer with smaller fire risks and better for mental health.

Leave a Reply

Your email address will not be published. Required fields are marked *


CDC: America Falling Behind Other Nations on Traffic Safety

How is the U.S. doing on traffic safety? To hear a lot of people tell it, we’re making great strides. President Obama recently referred to the reduction in American traffic deaths as a success story of sorts, contrasting it with the rise in gun deaths. But while traffic fatalities in America are indeed trending downward, […]

How Cars Won the Early Battle for the Streets

Judging by the recent media backlash against a few bike lanes in New York City, you would think that roads have been the exclusive domain of cars since time immemorial. Not so, as Peter D. Norton recounts in his book, “Fighting Traffic — The Dawn of the Motor Age in the American City.” When cars […]