NHTSA data shows drop in traffic and bike deaths — and cyclists fare as well in collisions as motorists

I’m stunned.

Like just about everyone else, I have always assumed that the lack of protection afforded cyclists meant that we fare far worse in collisions than the occupants of motor vehicles.

After all, we don’t have seat belts and airbags — let alone a couple tons of steel and glass — to protect us. Just a thin shell of foam covered in plastic and a maybe bit of chamois between our legs.

But I was wrong.

During an email exchange with fellow cyclist and KCRW chief engineer Steve Herbert, he posed an intriguing question.

For all the cycling deaths we are seeing and the lack of protection a bicycle provides us in a crash with another automobile, I wonder if fatality numbers are proportionally higher than that of motor vehicle occupants?

Fortunately, the answer was readily at hand.

Just yesterday, the National Highway Traffic Safety Administration released the latest traffic fatality statistics for 2010, showing an overall drop in traffic deaths from 1.13 deaths per million vehicle miles traveled in 2009 to 1.09 fatalities per million miles in 2010. And a drop of over 1,000 traffic deaths over the pervious year, from 33,808 to 32,788.

And yes, that’s a significant improvement.

Even if an average of 90 traffic deaths a day is hardly good news.

The news is also better for cyclists, as biking deaths have dropped to 618 — the lowest total in 35 years — despite a dramatic upsurge in ridership.

That’s still an average of 1.7 riders dying on our streets everyday. Nearly 12 every week. Over 51 every month.

And it is still far from acceptable.

The real surprise came when I dug a little deeper into those figures.

According to the NHTSA figures, excluding motorcyclists, roughly 2,009,000 motor vehicle occupants — drivers and passengers — were seriously injured on American roads last year, compared to 23,946 fatalities. That gives a ratio of 83.9 motor vehicle injuries for every death.*

For the same year, roughly 51,000 cyclists were seriously injured compared to 618 deaths, for a ratio of 83.5 to one.

Look at that again — 83.9:1 for motor vehicles, compared to 83.5:1 for cyclists.

In other words, you have virtually the same risk of dying in a traffic collision riding your bike, with little or no protection, as you have in a car or truck surrounded with safety features.

Of course, that does not take into account the frequency of collisions. While the NHTSA can cite a rate of 1.09 deaths per million miles of vehicle travel, no such figures exist for bikes, as there is no quantifiable method of determining how many miles are travelled by bike each year; any estimate you might see is nothing more than an semi-educated guess at best.

But those figures clearly show, once a wreck severe enough to cause serious injury occurs, you face no statistically greater risk on a bike than you would in a car.**

Don’t know about you, but I’m pretty damn shocked.

* Motorcyclists face a significantly greater risk, with 82,000 injuries compared to 4502 fatalities, for a ratio of 19:1.

**Update: One important distinction I failed to make. As maxutility pointed out, the data doesn’t show the same injury to death ratio for all car and bike collisions, but only those severe enough to result in injury. I’ve adjusted the copy to reflect that. The data does not show whether you are more likely to be seriously injured in a collision riding a bike or in a motor vehicle, just the ratio of serious injuries to fatalities.


  1. It would be great if your interpretation held true, but exposure is a really big factor. Per mile traveled, bicycling is definitely more likely to get you killed. (And I bike to work!)

  2. maxutility says:

    “But those figures clearly show, once a wreck occurs, you face no statistically greater risk on a bike than you would in a car.”

    Not to be negative, but the statistics you cite don’t really show that. What they show is that once you are in a wreck in which there is significant injury, you are no more likely to be killed rather than just injured bike vs. car. The question is once a wreck occurs, are you more likely to be injured/killed on a bike than you are in a car. Common sense says yes, but there are arguments why that may not be true. This data just doesn’t answer that question though.

    That all said, you would expect the injury to death ratio to be higher for bikes. This is a good reminder that the general assumptions about the danger of cycling are often wrong and overstated.

  3. Mike Llindblom says:

    The conclusion doesn’t follow from the statisitc. If you’re hit riding a bike, the speeds necessary to cause serious injury are lower than if you’re hit while inside a steel car. But once injured, the world-class Harborview Medical Center in Seattle, and the other first rate emergency rooms around here, will patch me up equally well regardless of whether bicycling or driving.

  4. @Phyllis: “Per mile traveled” favors speedy conveyances over the slow. “Per hour of activity” favors the slow over the speedy. I think you use the word “exposure” in a distance sense, but it can also be seen in the time sense. An hour spent cycling is not more dangerous than an hour spent driving; in fact, the opposite is true, bicycles are safer than cars per hour of activity:


  5. Mark Elliot says:

    I think that the collected statistics are not comprehensive enough to draw any conclusions but only infer with some imprecision the relative risk.
    The question of relative risk as calculated against time versus distance traveled is an interesting one. What about risk relative not to distance traveled but proximity to home? It’s an often-cited statistic that car insurers seem to value (or at least promote) because it’s counterintuitive: we estimate risk based on distance, but it may be frequency of travel that figures more directly. More trips simply originate and terminate near home.
    I agree the data is interesting, but maybe most value as a though experiment. How could we call for data collection that will give us more precision?
    I’m curious about the thresholds. I’d expect a higher rate of deaths when cycling too. But death is a catastrophic collision event, of course. With cycling, I imagine it’s more likely to lead to a variety of serious injuries short of death (by whatever measure we use), whereas a motorist, more protected, is less likely to suffer a serious injury, but when it’s serious, more likely to be a fatal injury.
    Does that make sense? If the threshold for a serious motoring injury is much higher, i.e., closer to actual death given the steel wrapper, would ‘serious injuries’ be more likely visited on a cyclist?
    Would that suggest that it’s not the unexpectedly low relative risk of death while cycling that’s at issue, but the unexpectedly high risk of death in a vehicle because ‘serious’ injuries are more likely to be life-threatening?

  6. Steven Vance says:

    You wrote:
    “In other words, you have virtually the same risk of dying in a traffic collision riding your bike, with little or no protection, as you have in a car or truck surrounded with safety features.”

    I wrote recently on my personal blog, Steven Can Plan, to answer the question, “What’s more dangerous, biking with no helmet or driving with no seatbelt?” It’s an odd comparison, but I decided to try to crack the question.

    If your definition of “dangerous” is “the likelihood that you’ll receive an injury while traveling in/on the vehicle”, assuming that the likelihood of being in a crash is the same*, then you are more likely to sustain an injury while cycling while wearing a helmet than while driving or being a passenger in a car while wearing a seatbelt.


    But I agree with Mark Elliot’s first statement: “I think that the collected statistics are not comprehensive enough to draw any conclusions but only infer with some imprecision the relative risk.”

  7. bikinginla says:

    As you all point out, this data is of limited use. And before we get carried away, I’d want to go back and see if the ratio holds for previous years, as well.

    However, simply on the face of it, I find it absolutely stunning that motorists and cyclists face the same ratio of serious injury to death; throughout my entire riding career, I had assumed that our risk of death was significantly higher. I would have expected something between the 19:1 ratio for motorcyclists and the 83.9:1 ratio for motorists, due to the lower speeds compared to motorcycles and the lack of protection compared to cars.

    What it tells me is that cycling offers, if not more safety than I had thought, at least greater survivability if the worst happens.

  8. Mark Elliot says:

    I like Steven’s comparison. It seems like apples/oranges – after all, what does one have to do with the other? – but asking this kind of questions gets us to slice & dice the data differently. It also sets up a finding with a hook, because it’s counterintuitive to compare such rates.
    Ted, the trends might be especially illuminating because of the limitations of the data. Like you suggest, it might surface change/trends that we can then look to associate with other variables out there.
    >Greater survivability
    Within the confines of the data, that’s a good way to think about it. We can (perhaps must) couple that with an analysis that examines incidence of collision relative to distance traveled, time traveled, or perhaps my suggestion of radius from home. In other words, survivability is important but of less consequence if we’re much more likely to encounter a serious collision (per whatever measure you like) on a bike than in a car.
    I suggest that we use the data to begin to frame a more effective data collection protocol or regime. Starting with local governments and scaling up to the feds. So that we can start to get that fine-grained picture for advocacy.
    We need a some realistic way to assess our odds of injury or death. I’m also curious about the role of bike injuries from falls or equipment malfunction, to which Bob Mionske (counterintuitively) attributed a very large proportion of injuries. I suppose this data doesn’t include solo-vehicle crashes, does it?
    Thinking about risk in broader terms, it’s not generically ‘higher’ but rather is associated with other variables. I’d guess experience or bike skills/training is key; perhaps socioeconomic status or other environmental factors (prevailing speed, mode split, etc.). These are aspects of the risk problem that we can move on quite apart from relative risk and it’s role in constructing our advocacy arguments vis-a-vis motoring.

    • Steven Vance says:

      On radius from home:

      I just started reading Tom Vanderbilt’s “Traffic” and he writes very early on: “It is a repeated truism, borne out by insurance company surveys, for example, that most accidents happen very close to home. On first glance, it makes statistical sense: You’re likely to take more trips, and spend more time in the car, in your immediate surroundings. But could there be something deeper at work? Habits, psychologists suggest, provide a way to reduce the amount of mental energy that must be expended on routine tasks. Habits also form a mind-set, which gives us cues on how to behave in certain settings.

      So when we enter a familiar setting, like the streets around our house, habitual behavior takes over. On the one hand, this is efficient: It frees us from having to gather all sorts of new information, from getting sidetracked. Yet on the other hand, because we are expending less energy on analyzing what is around us, we may be letting our mental guard down. If in three years there has never been a car coming out of the Joneses’ driveway in the morning, what happens on the first day of the fourth year, when suddenly there is? Will we see it in time? Will we see it at all? Our feeling of safety and control is also a weakness. A study by a group of Israeli researchers found that drivers committed more traffic violations on familiar routes than on unfamiliar routes.”

      I’m only on page 20, but lots of interesting information. Maybe it should be required for driver’s ed students 😉

      Amazon link to Traffic book.

      • Mark Elliot says:

        I’ll have a look at it myself. Recently, when I took a cycling safety course and found myself disappointed with the level of prep relative to road risk in the urban setting, I had a look at the ‘drivers ed’ offerings to high-schoolers. And they’re paltry. Packaged curriculum stuff. They need this book instead.

        (My post on cycling training: http://tinyurl.com/7us8np7 )

        The Amazon page for it has a nice Q&A with Vanderbilt, FYI. And per our conversation above, it looks like you can fast-forward to chap 9!

        Speaking of risk, the Chicago figures you cite are most illuminating for illustrating the striking difference between riding and driving when both take the prudent step of bucking up or wearing a helmet. The distribution of injury by extent is pretty alarming – if not unexpected: http://tinyurl.com/7fsq3fz

  9. […] NHTSA data shows drop in traffic and bike deaths — and cyclists fare as well in collisions as moto… […]

  10. […] As bikinginla.wordpress.com  points out, 618 cyclist deaths in 2010 makes it the lowest overall figure in some 35 years. The Arizona figure, 19, puts it close to our 10-year average; coming off of a bad 2009 (25). […]

  11. Manolo blahnik online store italia

    NHTSA data shows drop in traffic and bike deaths — and cyclists fare as well in collisions as motorists |

%d bloggers like this: