Mar 27, 2017

How dangerous is cycling?

Statistically, cycling is much safer than prospective riders imagine, but it's what they imagine that matters; it still seems too dangerous to generate widespread uptake

Alan Davies — Editor of The Urbanist

Alan Davies

Editor of The Urbanist

Comparison of injury risks and accident compensation claims, cycling for transport and other activities in NewZealand, 2001–2013 (source: Chieng et al)

Three researchers from the University of Auckland have just published a paper examining how cycling for transport in New Zealand compares in terms of safety with snow sports, rugby, horse-riding, quad bike, and home DIY. The authors conclude:

In terms of moderate injury, cycling is no more dangerous in a statistical sense than many recreational and every day activities, and in some instances is a good deal safer.

The paper, How dangerous is cycling in New Zealand?, by Michael Chieng, Hakkan Lai and Alistair Woodward, is published in the Journal of Transport & Health. It looks at injuries requiring an Emergency Department visit or leading to an accident compensation claim, but not fatalities (summary of paper here).

I’m not convinced the comparison with activities like rugby and snow-boarding is very illuminating and there are in any event serious measurement problems. The researchers are on firmer ground, though, when they argue the absolute risk of injury while cycling is very small:

Taking injuries that lead to claims (for accident compensation), we found these occur roughly 9 times in every 100,000 short urban bike trips; the chance of receiving an injury sufficiently severe to cause a visit to the hospital was similar.

They say that “if you rode a bike three times a week, most weeks, the chances are you would suffer one moderately severe injury every 70 years”.


So why, they ask, is fear of injury such a barrier to people in New Zealand and other car-dominated countries taking up cycling, given the statistical risk of injury is “unremarkable”? The answer, according to one of the authors, is cultural:

The bicycle has literally been pushed to the margins and the environment sends a powerful message that such use of the road is unusual, different, and is not valued. The transport norm is reinforced in other ways. For example, cycling promotion campaigns with safety-oriented messages such as “Share the road” have, perhaps unwittingly, strengthened the social framing of cycling as an activity that is inherently dangerous.

It’s an interesting question and I’d say the authors are on the right track. But there’s more to it.

I think the public debate in Australia in the 1980s around the helmet law also contributed to a heightened perception of cycling as an especially dangerous activity. The key thing here though is that it was the debate that moulded attitudes and ‘did the damage’, not the law per se; the controversy cemented the idea of cycling as risky. Even if the law hadn’t passed, I think cycling on roads would still be almost universally seen as dangerous.

Another reason cycling is seen as more dangerous than the numbers suggest could be the high frequency of relatively minor injuries suffered by cyclists – like grazes and bruises from falls – that hurt but don’t necessitate a visit to the emergency department or a compensation claim. There are very few cyclists who haven’t suffered an event like this. Perhaps these small injuries remind cyclists of their vulnerability and amplify the idea that it could’ve been a lot worse.

The behaviour of hostile and unthinking motorists towards cyclists is likely also part of the explanation. But in my view the main reason is simply that we’ve come to see roads as very dangerous for all users.


Consider that travelling by car is a lot safer than cycling; according to the authors, “the injury risk per million hours travelled is 75% less for motor vehicles compared with bicycles” (see also Is cycling getting safer?). Yet we still don’t think driving is anywhere near safe enough; we continue to spend huge amounts on initiatives designed to further improve safety.

These initiatives include designing roads and managing traffic to make motoring safer e.g. motorways, black spot programs. We also give enormous attention to improving the safety of motor vehicles e.g. seat belts, multiple air bags, antilock brakes, traction control, electronic stability control, crumple zones. Now there’s a new generation of intelligent safety features being introduced like forward collision warning, automatic brake-assist, blind-spot warning, lane-keeping assist, active head restraints, and automatic tyre pressure monitoring.

It seems our threshold for an acceptable level of motoring safety can never be low enough i.e. it should be zero. So, if we’re prepared to go to inordinate lengths to protect motorists who already have the advantage of metal cages and various safety technologies, is it any surprise that cycling on roads is seen as too dangerous? Is it a surprise when the estimated relative risk of being killed while cycling on Sydney’s roads is around 11-19 times higher than it is in a car (see Is cycling more dangerous than driving?)?


The probability of suffering a moderate to severe injury from cycling three times a week might well be only once every 70 years as the authors contend, but it seems that virtually all of those who don’t currently cycle still see that as a risk too far.

The implication for policy-makers is there’s little value to be gained from telling existing and prospective cyclists their fear is irrational. The proper basis for policy is to focus on their sense of perceived safety, just as we do for motorists. That requires actions like constructing separate cycle networks (see Is it time our cities got cycle superhighways?) and, where road space is shared, placing restrictions on the behaviour of motorists e.g. one-metre overtaking laws (see Shouldn’t every state have a one-metre cycling law?).


Leave a comment

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

17 thoughts on “How dangerous is cycling?

  1. Marty Wallace

    The graph is meaningless in real terms as it is a graph on compensation claims, not accidents, nor does it indicate the severity of the accidents. I’ve had many bike accidents and I’ve never made a compensation claim. Also how many people play rugby or do snow skiing compared to cycling? In the event that a cyclist is involved in an accident it’s rare for the police or authorities to actually follow it up.
    The whole article seems pointless as it is based on useless statistics.

    1. Alan Davies

      The exhibit shows both crash risk (grey) and compensation claims (white). The former is based on bicycle emergency department statistics. Both are de facto measures of severity i.e. serious enough to require an ED visit or to justify a compensation claim. The researchers were restricted to serious injuries by the available data, but it seems reasonable when discussing how dangerous cycling is to focus on serious injuries.

      1. Rob Anderson

        There are significant issues about how/whether cycling injuries are properly counted, even here in San Francisco, where cycling is fashionable:

  2. Letterboxfrog

    Perception of safety extends to school kids.
    Despite the fact I was allowed to ride my bike to school with stackhat back in the 80s in suburban Brisbane on my own, my wife (who was raised on a property 60km from town) with my mother’s support, will not let my son walk, let along scoot or cycle unsupervised to school in suburban Queanbeyan (near Canberra), due to stranger danger. Add to that notices from Queensland Police on Facebook stating that children cannot commute to school unsupervised, we are going to end up with a bunch of “poo slugs” who are too scared to use a bike. We need authorities to give us confidence that it is safe for cyclists to walk and ride, and the long term consequences of not doing so are far worse than the short term risks.

  3. Anthony Young

    I don’t think the belligerence of a lot of motorists towards cyclists should be underestimated as a deterrent to cycling. I don’t like driving my car on the roads these days because of the amount of agro. I haven’t ridden for years because of the intimidation you’re subjected to.

  4. [email protected]

    I don’t think that the perception that cycling and motoring are risky activities is excessive. Turn around the quoted estimated and it says “if you cycle regularly, you will probably suffer a moderately severe injury sooner or later”. That’s not going to stop me cycling, but it’s not acceptable either

    1. Tony Morton

      No, it’s not acceptable. However, it could be at least partly due to the contemporary cycling population being weighted much more heavily toward young, impetuous males in this country, relative to other places. If we could somehow replicate the cycling demographics of the Netherlands I suspect that even with no other changes you would see a lower injury rate at the population level.

  5. Tony Morton

    Michael has a point, which builds on your own point I think Alan. When Dorothy Robinson gathered the stats from the late 1980s on relative risk, she found that when looking specifically at traumatic brain injury per hour of activity, the risk ratio for motorists and cyclists was close to 1:1.

    If it’s higher today, there are at least two factors at work. One is the well-documented improvement in motorist safety features. But there’s doubtless also an effect from self-selection in the cycling community: the population of cyclists is more biased toward risk-takers and against the risk-averse compared to 30 years ago.

    Naturally it’d be very difficult to quantify a selection effect like this, but we should keep in mind that ‘safety in numbers’ has many dimensions to it.

  6. Michael

    Cycling became very dangerous in the early 1990’s when the Government spent a lot of advertising money convincing us that it was so very dangerous that we all needed helmets like motor bike riders do, and in droves parents stopped buying bikes for kids and the damage was done – a whole generation that will not ride because it’s too dangerous.
    Witness our local high school bike parking area – in 1990 there were 200 bike spots and they were always full, now there are 20 and often not even 5 are used, a sad loss of exercise and environmental benefit.

    1. Horst (Oz) Kayak

      How representative are Michael’s observations of cycling traffic generated by government run secondary schools?
      Princes Hill Secondary College in inner MEL has several hundred bicycles parked outside by students every school day in 2017. In the early 1990s I never counted more than 20 at Princes Hill SC. Are we talking about different world’s?
      Obviously advocacy groups such as the TCPA find it difficult to agree with Michael unless many more examples are given.

      1. Alan Davies

        An important point. See this article I wrote in 2012, Did students cycle to school back in the day?, based largely on a study of cycling to school done by Alan Parker back in 1987. The money quote:

        Perhaps the most interesting finding is the variability in rates of cycling between schools. Mr Parker only shows differences between municipalities (rather than schools) but there are nevertheless some extraordinarily large divergences.

        For example, more than 20% of students attending bayside secondary schools travelled by bicycle. In the old Sandringham municipality the proportion was 48%.

        On the other hand, less than 5% of students attending school in inner city and inner suburban municipalities like Footscray, Northcote, Fitzroy, Richmond, Hawthorne and Camberwell cycled to school (Alan Parker has more detail here).

        A similar high degree of variability between schools was noted by Smith and Milthorpe in three surveys of cycling conducted in NSW in 1991, 1992 and 1993. They noted some Sydney schools had negligible ridership in all three surveys (pre and post introduction of mandatory helmet law).

      2. Michael

        Hi Horst, was talking about Sale High School, 3 bikes there today.
        Friend just counted the bikes at Princes Hill for me – 62 along Arnold St and 12 in under cover on the Arnold St side; granted that’s more than 3 but far short of your “several hundred”.
        Perhaps best to keep in mind that all of us can pick n’ choose “facts” to suit our viewpoint, so putting our “facts” aside the salient point is that if mandatory all ages bike helmet laws were the way to go then the rest of the world would have followed – but of course they have not and Australia with it’s mandatory helmet laws sits at the bottom of the safety rankings for developed countries; we are not so smart as we smugly think we are and the rest of the world are not stupid.

    2. Alan Davies

      The 80s helmet debate had a significant impact on the perception of cycling as dangerous. I think it’s very important to note that even though the helmet law was passed, it didn’t remove the perception of cycling as unsafe. The debate it was what ‘did the damage’. So even if the law were repealed on the grounds it excessively ‘dangerises’ cycling, it wouldn’t improve the now deeply entrenched perception that cycling’s unsafe (which isn’t to ignore that there are other arguments for repealing the law).

      As for cycling to school, the helmet law is only one factor explaining the dramatic drop. Other factors include the equally dramatic rise in children being driven to school, itself a function of wider changes, including in the structure of the labour market and falling car prices. I expect there were other factors too e.g. digital gaming, private schooling.

      1. Jonathan

        I’m not sure I follow that logic Alan.

        If a big campaign around how dangerous cycling is resulted in less cycling, a big campaign about how safe it is surely might have the opposite effect? It won’t reverse the damage completely but might result in some change.

        Especially if partially implemented, e.g. ‘if it’s 40km/h or a separated lane no helmet is required’. That would send a strong message that some types of cycling are very safe.

        It would also have the effect of encouraging those treatments because the benefits to cyclists (not having to wear helmets) would be greater.

        1. Alan Davies

          An anti-helmet campaign would of course have “some effect” but not much. It’s fallacious to assume that what happens in one direction on an issue will be mirrored equally in the other direction; they’re not nessarily symmetrical.

          I’ve cited a survey before showing over 90% of Australians think the helmet law is appropriate. There was an anti MHL rally in Melbourne the other weekend that attracted only 200 participants.

          In any event, whether or not the introduction of MHL in the early 1990s resulted in a sustained drop in cycling is contested; the evidence on that score isn’t conclusive.

          1. Dudley Horscroft

            However repealing the helmet laws would entice us octogenarians who used to cycle to school each day back onto the roads. And with us back, it would make people think – “If those old codgers think it must be safe, then it must.” At present, most of the cyclists I see are the helmeted, lycra clad youngsters in black (ie near invisible) clothing with their heads bent down so that they probably ought to have a mirror on the handlebars so they can see ahead.

          2. Jonathan

            Those over 50 benefit from helmets the most if they’re in an accident so perhaps not the best plan!

Share this article with a friend

Just fill out the fields below and we'll send your friend a link to this article along with a message from you.

Your details

Your friend's details