Personal choice and community impacts
Submission 253
Submission to the Inquiry on personal choice and community impacts
This submission is intended to discuss Australia's all-ages mandatory helmet laws for pedal cyclists.
SUMMARY
Australia was the first country in the world to introduce mandatory helmet laws (MHLs) for
pedal cyclists of all ages. Supporters of these laws quote reductions in cycling injuries coinciding
with the introduction of MHLs throughout Australia as evidence that MHLs were an effective way
of protecting cyclists. However it seems likely that these reductions in injury numbers had more to
do with other road safety measures, as similar reductions can be seen in motorist and pedestrian
injury numbers at around the same time. The term "cycling" covers a wide range of activities from
road racing, to downhill mountain biking, to casual commuter cycling. Arguments for the efficacy
of MHLs often fail to distinguish between the different levels of risk faced by cyclists on main
roads, as opposed to minor roads and off-road paths, in effect arguing that all cycling is equally
dangerous. Furthermore, any changes in the percentages of cycling injuries conceal an important
fact - absolute numbers of cycling injuries are very small, with the odds of serious injury (per
cycling journey) being around one in a million, and the odds of a fatality being around one in ten
million. Any improvements in cycling safety that arise from mandatory helmet use are equivalently
small, calling into question the justification for making them mandatory. Several independent lines
of evidence suggest that mandatory helmet laws discourage cycling, indicating that a large sector of
the Australian public considers MHLs to be an unwelcome imposition on personal choice, and a
strong disincentive to use a form of transport that provides a net social benefit (in the form of
improved health, reduced costs to public health care, reduced transport infrastructure costs, and
reduced congestion and pollution). Measures to protect cyclists that focus on preventing accidents
from happening in the first place (i.e. physically separating cyclists from traffic by the use of
dedicated cycling infrastructure) would reduce all types of cycling injuries, and encourage a greater
range of transport choices, with associated health and environmental benefits. Mandatory helmet
laws do nothing to reduce the odds of accidents occurring, and do nothing to protect other body
parts than the head. Reducing the odds of accidents occurring (through the provision of
infrastructure, driver training and awareness campaigns, etc.) reduces any safety benefit gained
through helmet use, further undermining the argument that helmets should be mandatory. As such,
MHLs can be seen as the least effective, least justified form of intervention to prevent cycling
casualties. It seems likely that MHLs have changed the style of cycling that is most popular from
fairly safe casual commuting, to comparatively dangerous sports cycling, making cycling on
average less safe. It also seems likely that MHLs have contributed to a shift in attitude towards
cyclists which has made them more likely to be targets of road rage. In the discussion that follows I
shall expand upon these points, providing published data to support my statements.
DISCUSSION
There is no argument that all-ages mandatory helmet laws (MHLs) for cyclists do not restrict
personal choice. Cycling without a helmet is a "victimless crime", in which the cyclist does not
endanger any other individual, and so it is reasonable to ask whether outlawing such behaviour is
compatible with the principles of a free, democratic society. However most Australians would
probably agree that some measure of government interference in individual behaviour is acceptable,
if it can be shown to have substantial net positive outcomes for either the individual or society as a
whole, which offset any loss of personal freedom. I wish to argue that this is not the case for
mandatory helmet laws, for several reasons. Firstly by showing that MHLs are less effective at
reducing cycling injury and fatality rates (relative to cycling participation) than other measures
which do not interfere with individual freedom, and any benefit arising from mandatory helmet use
is minimal. Therefore MHLs can not be said to benefit the individual. Secondly, by arguing that
Personal choice and community impacts
Submission 253
MHLs probably reduce cycling participation, reducing the health, environmental, and economic
benefits to society, which have been repeatedly shown to outweigh any injury-related cost to
society. Thirdly I wish to argue that MHLs have created a cultural perception that cycling is a
dangerous sporting activity, rather than a valid form of transport, and bred hostility towards cyclists,
making cycling more dangerous, rather than less. For all of these reasons, HMLs can be seen as
failing to provide benefits to either the individual or society, and that given the imposition upon
personal freedom and the cost of fines borne by cyclists who choose not to comply with these laws
they should be repealed. Furthermore "cycling'' can refer to many different activities, from highspeed road racing and downhill mountain biking, to slow cycling along quiet streets and off-road
bike paths. Calculations of the dangers of cycling based predominantly on cycling on roads, or
cycling for sport, may seriously over-estimate the dangers of cycling in other circumstances.
Therefore even if helmets can be shown to provide some benefits under some circumstances, we
should ask whether they should be mandatory in all circumstances.
The terms of reference of this inquiry cover a number of topics, including the provision of drugs
such as alcohol, tobacco, and marijuana. Although it is my intention only to address the matter of
bicycle helmets directly, I feel I should point out an important distinction between this and the other
terms of reference. Drugs such as those mentioned above cause gradual and inevitable health
problems, by their repeated use. More use leads to more serious negative health effects, and laws
which restrict or prevent the use of these drugs aim to prevent these inevitable, and cumulative
effects. By contrast, any negative health effects from cycling are the result of rare, exceptional
events (crashes). In the vast majority of instances, cycling leads exclusively to positive health
benefits for both participants and society as a whole, such as improved cardiovascular fitness,
improved bone and muscle density, and reduced air pollution. Mandatory helmet laws are therefore
the only laws within this inquiry's terms of reference which may discourage an activity which in
almost all cases is entirely harmless and actually beneficial.
Australia was the first country in the world to introduce all-ages mandatory helmet laws for
cyclists, beginning on the 1st of July 1990 in Victoria, and following in other states and territories
over the next few years. It is not unreasonable to review a law which was introduced without
precedent, after twenty-five years, to see whether it actually achieved what it was intended to. If it
did not, it should be counted as a failure and modified or repealed. One of the earliest arguments for
the success of MHLs was a reduction in cyclist injury rates at the time MHLs were introduced.
Supporters of MHLs claim that following the introduction of helmet laws in Victoria the rate of
cyclist injuries fell immediately by 20% to 30%. Figure 1. shows a graph of cyclist fatalities in
Victoria by year, created from data obtained from the Traffic Accident Commission website. There
is a clear reduction around 1990. However correlation does not necessarily imply causation. Figure
2. shows fatality data for cyclists, motorists, motorcyclists and pedestrians over the same period. It
is clear that a comparable reduction in fatalities occurred at the same time for cyclists, motorists
(both drivers and passengers), and pedestrians, with a more gradual downward trend for
motorcyclists. The most likely explanation is the introduction and increased enforcement of random
breath testing, anti drink-driving TV campaigns, reduced speed limits, etc. It does not make sense to
claim that any reduction in cyclist casualties is exclusively the result of helmet laws. Furthermore
the number of cyclist fatalities seems to rise from 1988 to 1989 before dropping again, which may
indicate that the apparently large reduction in casualties at the time MHLs were introduced is
illusory, arising from a statistical fluctuation - an anomalously high casualty count in 1989. We also
need to assess whether this reduction in cyclist fatalities was due to a reduction in the number of
people cycling. If a reduction in cyclist casualties corresponds with a reduction in the number of
cycling trips undertaken, it weakens the case that any reduction in casualties was due to helmet use.
Note that this is not dependent on the reason the number of cycling journeys fluctuates. There is a
long-running debate in Australia about whether MHLs discourage people from cycling. But for the
purposes of the current argument, this is irrelevant. The number of people cycling may vary from
Personal choice and community impacts
Submission 253
Figure 1
Figure 2
year to year because of increased car ownership, economic factors (e.g. the price of petrol),
population growth, demographic shifts that make commuting to and from work by bicycle more or
less practical, etc. The reasons cycling participation fluctuates from year to year are not important.
All that matters is that, all other things being equal, fewer cyclists means fewer cycling casualties.
To assess any effect on casualty rates from the introduction of mandatory helmet laws alone, we
need to compare the number of cycling casualties (taking into account the number of cycling trips
undertaken) against the number of casualties for other modes of transport. Pedestrians are the
natural choice for such a comparison, since fluctuations in the number of people travelling by
bicycle and by walking are likely to be similarly influenced by bad weather, and motor vehicles
may become safer as time goes by due to improvements in brake technology, airbags, crumple
zones, etc. making cycling look more dangerous in later years, compared to driving. Therefore for
each year with available data we should calculate the ratio
(number of bicycle fatalities per bicycle journey)
W = -------------------------------------------------------------------(number of pedestrian fatalities per pedestrian journey)
This ratio can be thought of as the odds of a cycling fatality relative to a walking fatality. If W
remains roughly constant before and after the introduction of mandatory helmet laws, it indicates
that any changes in cycling safety correlated with an equivalent change in the safety of walking, and
therefore is probably not a result of mandatory helmet legislation. Details of the data and
calculations used for such a comparison are given in Appendix A. The results are as follows;
Year
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
======================================================
W value
1.06 0.73 0.82 1.76 0.70 0.81 1.08 0.58 0.74 0.72
It is easy to see that there is barely any difference between pre-law (1987-1989) and post-law (1991
onwards) values, suggesting that pedestrians and cyclists experienced roughly equal changes in the
odds of being killed in a traffic accident. In fact, the average value of W for the three years pre-law
is almost exactly the same as the average W value for the three years immediately post-law (19911993). This strongly suggests that any reduction in cyclist fatalities in Victoria after 1990 was due
more to general road safety measures such as random breath testing, speed cameras, and the like,
and not to mandatory helmet laws. The value of W is at its highest in 1990, the year MHLs were
introduced. Examining the raw casualty data (in Appendix A) indicates that this was due to an
anomalously low number of pedestrian casualties that year. The fact that W is less than 1 in most
years indicates that cycling fatalities are rarer (per journey) than pedestrian fatalities. Discussion in
chapter 2 of the AustRoads publication AP-R155 "Pedestrian and Cyclist Safety - Recent
Developments" (2000) supports this.
Personal choice and community impacts
Submission 253
While looking at the percentage change in accident rates over time is useful, it can obscure the
fact that cycling accidents are rare, and it is not inevitable that a cyclist will have a crash whenever
they go for a ride. “Hospitalised Sports Injury in Australia 2002-2003” by Flood and Harrison
(2006) lists the annual hospitalisation rate per 100,000 participants of various sports in Australia.
This study found that the figure is 97.0 for cycling, 107.7 for field hockey, 121.2 for touch football,
121.3 for cricket, and 150.2 for netball. To be fair, the types of injuries most commonly sustained
vary from sport to sport, but this still serves to illustrate how uncommon cycling injuries are, and
one may still ask why cycling is the only sport on this list for which all participants are legally
required to wear protective equipment.
Even if the odds of sustaining an injury were reduced by wearing a helmet, the absolute
reduction in risk that arises from helmet use may not be very great. We can illustrate this with a
hypothetical example, using light and dark clothing instead of helmets. Suppose 100 cyclists in dark
clothing and 100 cyclists in light clothing go riding. Suppose 1 rider wearing light clothes, and 2
riders wearing dark clothes are seriously injured. At first glance, wearing light-coloured clothing
seems to lead to a 50% reduction in injuries. If we ignore for a moment the important caveat that
correlation does not prove causation, this conclusion would not be incorrect, but it is misleading due
to selection bias - we have ignored the cyclists who were uninjured. If we consider all 200 cyclists
in our hypothetical example, we see that accidents are rare. Only 2% of the cyclists wearing dark
clothes, and 1% of the cyclists wearing light clothes were injured. So wearing light-coloured
clothing reduces the chance of being injured by just 2% - 1% = 1%.
What does this hypothetical example tell us about helmet use? Firstly, it reminds us to be careful
about the risks of selection bias. For instance, doctors who work in hospital emergency departments
may develop an exaggerated view of the dangers of cycling, since they only see injured cyclists
being admitted, and don't see the thousands of others who do not have accidents. Likewise,
anecdotal accounts from people recounting tales of their friends (or themselves) having accidents
provide a biased view of the hazards of cycling, because nobody goes out of their way to tell their
friends about all the uneventful, perfectly safe rides they have had, and so these go unnoticed and
unremarked. We should therefore try to assess not only how many cyclists are injured, but how
many are uninjured, in any given year. Let's look at statistics from South Australia (where good
cycling participation and accident data are available, and which is close to the national average in
cycling participation). According to the Australian Cycling Participation Survey 2011 conducted by
AustRoads, in the week before the survey was conducted 18.1% of the South Australian population
rode a bicycle, with an average of 4.7 trips per week being undertaken. Given a state population of
1.6 million, and 52 weeks in a year, this equates to around 71 million cycling journeys per year.
Reducing this number to account for seasonal variation in the number of cyclists between winter
and summer still leaves around 56 million cycling journeys per year. Over the period 2003-2012
there have been an average of 3.4 cyclist fatalities per year, and 65.3 serious injuries per year (data
from the Road Crashes in South Australia “Pink Book” 2012, published by the Department of
Planning Transport and Infrastructure). This makes a total of 69 fatal-or-serious injuries, out of 56
million cycling journeys. Therefore the odds of sustaining a fatal-or-serious injury every time you
get on a bicycle are a miniscule 0.00012% (or about 1 in 811,000). Let us assume, for the sake of
argument that helmet use results in a 70% reduction in the rate of brain injuries, or in other words a
complete cessation of helmet use would result in a 3.3-fold increase in head injury rates (notice that
we are being generous to the pro-helmet-law case by accepting the 70% figure given in the widelyquoted CARRS-Q monograph on bicycle helmets, by Haworth, Schramm, King and Steinhardt, and
by over-estimating the odds of a head injury by assuming that all serious injuries are head injuries).
Then the odds of sustaining a fatal-or-serious injury every time you get on a bicycle would increase
to 0.0004%. This is a tiny difference in injury rates, of only 0.0004% - 0.00012% = 0.00028%, or
just less than three ten-thousandths of a percent! Such a small improvement in safety to the
Personal choice and community impacts
Submission 253
individual arising from mandatory helmet use is far from significant. The benefits of wearing a
helmet become even less significant as accidents become rarer, as is the case for cyclists in low-risk
areas, such as off-road bike paths, and cyclists travelling at low speed on bicycles with an upright
riding posture. For instance, if the odds of being involved in a crash were halved, the odds of
sustaining a fatal-or-serious injury (in both the helmeted and unhelmeted cases) would be halved,
and hence the improvement in safety arising from wearing a helmet would also be halved. For the
sake of comparison with other states, an equivalent calculation for Western Australia shows an
average of 3.3 fatalities per year, and 109.7 serious injuries, out of 114.5 million cycling trips
annually, resulting in a 1 in one million chance of a fatal-or-serious injury per journey. In 2011
Adelaide had 38 reported cyclist injuries per 100,000 population, versus 32 in Melbourne, 21.5 in
Canberra, 19.4 in Darwin, 17.1 in Perth, 15.1 in Sydney, and a capital city average of 22.7 for the
whole of Australia. Although these figures are per head of population rather than per cyclist, they
should emphasise the fact that the calculation given above for South Australia does not overstate
how safe cycling is in the rest of Australia).
Every activity people perform carries a measure of risk. Some people are seriously injured while
skydiving. Others are seriously injured while surfing. Still others are seriously injured as a result of
slipping over in the shower. At what level of risk is an activity deemed sufficiently safe that people
can freely choose to participate in it, making up their own minds about the dangers they face by
doing so? Since no activity will ever be entirely safe, there must be some threshold level of risk
below which government regulation of an individual's behaviour is deemed unnecessary (if this
were not the case, then it would be logical to mandate the use of all forms of protective equipment
for everyone participating in any activity at all). Given that being a cyclist appears to be no more
dangerous than being a pedestrian (per journey), that the odds of hospitalisation for cycling are less
than for several other sports which do not require mandatory protective equipment, and are
incredibly low (estimated at 0.0004% for each helmetless journey), it seems that the risk involved
with cycling is already below such a threshold, and mandatory helmet laws should be repealed (or
at least amended) to reflect this. One must wonder why a person should be penalised and charged
with a criminal offence for failing to do something (wear a helmet) that only makes them 0.00028%
safer. To put these figures in a different perspective, if an unhelmeted cyclist rode twice a day, 365
days per year, every year, they would expect to sustain one fatal-or-serious injury every 336 years!
On this basis it seems doubtful that cycling accidents are common enough to justify mandatory
helmet use in all circumstances.
It should also be noted that some states (though not all) allow exemptions from MHLs on the
basis of religion, to allow the wearing of religious head coverings such as turbans while cycling.
Perhaps conscientious objection, on the basis of an assessment of empirical data pertaining to
accident rates in different cycling circumstances, should also be allowed.
We can reasonably ask if mandatory bicycle helmets provide a net benefit to society. Some
studies have claimed large reductions in accident rates, however other more recent studies taking
changes in cycling participation into account suggest that the reductions in accident rates from early
studies are overestimates. Furthermore there were few assessments of cycling participation during
the 1980s, and so many studies claiming that helmet laws were beneficial did not have a good
baseline of pre-MHL data to compare against. Since 2011 AustRoads has compiled a survey of
cycling participation every two years. No such consistently collected data for all cycling journeys
exists for the pre-MHL era (the ABS census only counts cycling to work). Federal government
statistics show that each cycling trip to work and back saves the national economy $21 in reduced
pollution, noise, congestion, infrastructure construction costs, and improvements in health. The
study "An economic evaluation of the mandatory bicycle helmet legislation in Western Australia"
by Hendrie, Legge, Rosman and Kirov (1999), found that even under the most favourable
assumptions, mandatory helmet use would have only saved the Western Australian economy $2
Personal choice and community impacts
Submission 253
million per year (and may even have cost the state economy $10 million per year), and even then
only if the healthcare costs associated with decreased cycling activity were ignored. Research from
Macquarie University (De Jong, 2009) indicates that mandatory helmet laws cost Australia around
half a billion dollars per year, due to reductions in cycling participation leading to reductions in the
health and fitness of the average commuter. A number of other international studies have found that
the social, economic and environmental benefits of cycling outweigh the costs of cycling injuries,
with large benefit-to-cost ratios. These include Hillman (1992) with a benefit-cost ratio of 20:1,
Rabl and de Nazelle (2011) with a ratio of 24:1, and de Hartog, Boogaard, Nijland and Hoek (2010)
with a ratio of 35:1. Each of these studies indicates that even if injury rates increased due to helmet
laws being repealed (and the experience of jurisdictions which have had MHLs and repealed them,
discussed below, suggests they would not), the societal benefit from cycling would be comfortably
maintained, and that any reduction in cycling that arises as a result of MHLs induces a net cost to
society. Furthermore around two-thirds of the moderate and severe traumatic brain injuries that
occur in Australia each year are a result of motor vehicle accidents. Motor vehicle use does not have
the associated health benefits of cycling which offset the injury-related costs to the health care
system. Therefore any economic argument about the costs of brain injury seems to present a much
stronger case for mandatory helmet use for motorists than for cyclists. Mandating helmet use for
cyclists appears to be an unjustified imposition on a small segment of the population, out of
proportion to the actual number of injuries resulting from cycling.
Further evidence that mandatory helmet laws are not the most effective or important way of
reducing cycling injuries comes from the Monash Alfred Cyclist Crash Study (MACCS) conducted
by P. Biegler, S. Newstead, M. Johnson, J. Taylor, B. Mitra, and S. Bullen, for the Monash
University Accident Research Centre, published in July 2012, which assessed the characteristics of
cyclists admitted to the Alfred and Sandringham hospitals in Melbourne. This study found that the
only significant predictor of head injuries was speed;
“ A cyclist travelling at 30kph or over prior to the crash was estimated to have
nearly 5 times the odds of sustaining a head injury in the crash compared to a
cyclist travelling below 20kph. This was statistically significant. Even
cyclists travelling at 20-29kph before the crash were estimated to have 2.7 times
the risk of a head injury compared to those travelling below 20kph.”
The MACCS study also found that;
“ The odds of the crash involving a travel speed above 20kph were over 80% lower
for off-road crashes compared to those on-road.”
Taken together these results suggest that cycling on off-road paths is around five times less likely to
result in a crash at high speed (compared to cycling on-road), and any crashes that do occur are
around five times less likely to result in a head injury. indicating that riding off-road (i.e. most
probably at less than 20km/h) could lead to a twenty-five-fold reduction in the odds of sustaining a
head injury in a crash. The most optimistic estimates from helmet law supporters claim only a threefold to five-fold reduction in head injuries due to mandatory helmet laws. The MACCS results are
derived from a thorough analysis of empirical data, pertaining to the time, location and
circumstances of a number of accidents resulting in hospital admission. Significantly, helmet use (or
non-use) was not found to be a significant predictor of head injuries in the MACCS study.
It is sometimes claimed that the cultural differences between Australia and Europe (with our carcentric transport culture) make helmetless "Copenhagen-style" cycling impossible or impractical in
Australia. In 2013 New York city, one of the most car-centric cities in a first world nation, began a
bike share scheme. At the time the scheme was introduced, critics claimed that it would lead to a
Personal choice and community impacts
Submission 253
public health disaster, with numerous fatalities and injuries, and called for helmets to be made
mandatory. In the end no such law was introduced, and after 23.5 million km of total travel, the
New York bike hire scheme had recorded zero fatalities and only 40 people requiring medical
attention, making the users of hire bikes statistically safer than the average New York cyclist.
Analysis of bike share schemes across the whole USA (as of August 2014) shows that the 36 cities
with bike share schemes have achieved seven years of operation and 23 million rides with no
fatalities at all. This safety record is attributed to the fact that hire bikes tend to be ridden slowly,
with an upright posture (unlike racing bikes that have a hunched forward posture that reduces the
cyclist's view of their surroundings). This is consistent with the findings of the MACCS study
which indicates that other factors such as speed are far more important that helmet use in predicting
the chance of head injuries.
It is difficult to ascertain for certain whether the introduction of MHLs reduced cycling
participation. Determining cause and effect is not simple in such a case. However there are several
indications that mandatory helmet laws discourage a large number of people from cycling. A survey
conducted in Sydney (Rissel & Wen, Health Promotion Journal of Australia 2011; 22: p178-183 )
found that about one in five respondents (22.6%) said they would cycle more if they did not have to
wear a helmet, particularly occasional cyclists. Furthermore the study by Clarke (New Zealand
Medical Journal, Vol 125 No 1349 , 2012) found that mandatory helmet laws in New Zealand
reduced the average hours cycled per person by 51%. The comparison of cycling participation in
Australia between 1985/86 and 2011, by Gilham and Rissell (World Transport Policy and Practice
Volume 18.3, May 2012) found that growth in cycling participation after the introduction of
mandatory helmet legislation was only around one-third of the rate of population growth, indicating
a significant decrease in the popularity of cycling. The Northern Territory relaxed its helmet laws in
1994 to make helmet use optional for adult cyclists on off-road paths and footpaths (while retaining
mandatory helmet use for minors and on-road cycling). Today the Northern Territory enjoys higher
cycling participation than the rest of the country (26% of the NT population had ridden in the week
before the 2011 AustRoads cycling participation survey, while the national average was 17.8%) and
cycling casualty rates in the Northern Territory are no worse than the national average (in fact they
are lower than in South Australia, for instance) despite the NT having worse transport accident rates
for pedestrians, motorists and motorcyclists. The fraction of all cycling trips undertaken by women
in the Northern Territory is about 1.5 times higher than the national average. Given that cycling
participation in the NT is above the national average, this probably indicates that relaxation of
helmet laws encourages greater use of bicycles as a means of personal transport, especially by
women (the NT's high rate of cycling participation may be partly due to a warm climate, however
the NT rate is so much higher than in other states and territories that it seems unlikely this is the
whole story. Furthermore, cycling participation in the ACT is higher than in Queensland, suggesting
that there is more to people's transport choices than climate alone). In recent years Israel, Mexico,
and the US cities Austin and Dallas, all jurisdictions that previously had mandatory helmet laws
equivalent to Australia's, have relaxed their laws in an effort to encourage more people to cycle, and
in some cases specifically to encourage greater use of public bike share schemes. Doing so has lead
to an increase in cycling participation (Tel Aviv saw a 54% increase in the number of cyclists riding
to school or work between 2010 and 2012, after the helmet law in Israel was relaxed), and no
evidence of a significant increase in injury rates per cyclist. The bike share schemes in Brisbane,
Melbourne, and Adelaide are utilised significantly less than in equivalent cities overseas, where
helmets are not mandatory. The National Cycling Participation Survey 2015 shows a decrease in
cycling from 2011 to 2015 and concludes that "the [National Cycling Strategy] target of doubling
the number of people cycling between 2011 and 2016 is unlikely to be achieved". These several
independent lines of evidence, taken together, suggest that mandatory helmet laws do indeed serve
as an impediment, and reduce cycling participation rates.
Increasing the proportion of commuter journeys taken by bicycles can be seen to have safety
Personal choice and community impacts
Submission 253
benefits for all commuters by reference to a simple model. Suppose for simplicity that there are a
fixed number of commuter journeys per day, taken either by car or by bicycle. Suppose that A
represents the percentage of journeys undertaken by car, and hence (100 - A) is the percentage of
journeys taken by bicycle. Then if we create a diagram with A on the vertical axis, and (100 - A) on
the horizontal axis, the point corresponding to any choice of car-cyclist proportions will lie on the
diagonal line, shown in Figure 3. The number of car-cyclist collisions will be proportional to the
number of cars, multiplied by the number of cyclists (that is, the area of the yellow rectangle), since
more cars means higher odds per cyclist of a collision with a car, and more cyclists means more
collisions involving cyclists. By a similar argument the number of car-car collisions is proportional
to the number of cars, multiplied by the number of cars (the area of the orange square). The odds of
a car-cyclist collision per cyclist is simply proportional to the number of cars (the height of the
yellow rectangle, illustrated here with a blue arrow). It is easy to see by comparing Figure 3 and
Figure 4 that as the percentage of commuter journeys taken by bicycle increases, the odds of car-car
collisions decreases, as do the odds of car-cyclist collisions per cyclist. In other words, increasing
the number of commuter journeys taken by bicycle makes both driving and cycling safer. This
should be enough reason to pursue any course of action which may increase cycling participation.
Figure 3
Figure 4
Australia's mandatory helmet laws were introduced two-and-a-half decades ago. Since then
Australians have begun to embrace the idea of living closer to or within the CBDs of major cities,
rather than in sprawling suburbs. At the same time petrol prices have increased, while the health
benefits of gentle regular exercise, and the environmental costs of fossil fuel use have come to be
better recognised. This should make cycling a more practical and appealing form of transport over
short to medium distances. In the 1980s and early 1990s the idea of creating networks of dedicated
cycle paths as part of a city's transport infrastructure would have seemed impractical. Yet today
major cities around the world, such as Barcelona and Paris are excluding cars from once-busy
thoroughfares and creating pedestrian and bicycle corridors in their place. Off-road paths are
effective at preventing cycling accidents from occurring, thereby protecting a cyclist's entire body
from harm - something helmets cannot do. Mandatory helmet laws, in contrast, are based on an
outdated approach to safety that views accidents as inevitable, and seeks to reduce their severity
rather than their frequency. In doing so, MHLs have become a way of blaming the victim. If a
jogger were attacked by an aggressive dog, the owner of the dog would be considered to be at fault.
We would not blame the jogger for failing to wear body armour. If someone was shot by an
assailant, we would not chastise the victim for failing to wear a bullet-proof vest at all times in
public. Yet if a cyclist is harmed while not wearing a helmet, they are more often than not regarded
as being in the wrong for failing to "protect" themselves. Placing the onus for cycling safety on the
use of personal protective equipment has probably slowed the adoption in Australia of other more
effective measures of protecting cyclists from harm, such as off-road cycling infrastructure, by
reinforcing the mistaken belief that accidents are inevitable and that injuries are the fault of the
Personal choice and community impacts
Submission 253
injured. Mandatory helmet laws should therefore not be viewed as a successful government policy,
but as a failure of policy, which has absolved governments of the responsibility to provide effective
safety measures that individuals can not provide for themselves (i.e. infrastructure).
On the 29th of November 2013 a Queensland inquiry into cycling issues by the Transport,
Housing and Local Government Committee concluded that
•
•
•
•
There was not “sufficient evidence of the safety outcomes of compulsory helmet wearing
to justify the mandating of helmet wearing for all cyclists of all ages regardless of the
situational risk.”
“relaxing mandatory helmet laws in specific circumstances is likely to increase
cycling participation rates with a range of associated health benefits and
economic benefits.”
“there is sufficient evidence provided by the Northern Territory example that a
relaxation of mandatory helmet laws in lower risk situations (such as cycling
on footpaths and on dedicated cycle paths), does not inevitably reduce the
safety of cycling.”
And recommended “a 24 month trial which exempts cyclists aged 16 years
and over from the mandatory helmet road rule when riding in parks, on
footpaths and shared/cycle paths and on roads with a speed limit of 60 km/hr
or less”.
Although these recommendations were not acted upon by the Queensland government, it is clear
that many people, upon assessing the evidence, have come to the conclusion that all-ages mandatory
helmet laws are not an effective or desirable policy, and one which has the unfortunate effect of
discouraging people from cycling, hence the repeal of MHLs in the Northern Territory, Israel,
Mexico, Dallas and Austin, and the reluctance of other countries to introduce MHLs.
In this discussion, I have tried to stick to arguments based on published research and empirical
facts. However I would also like to share some personal experiences and observations. I have
cycled as a regular means of transport for most of the past thirty years. I have travelled and cycled
extensively in Australia, North America, and Europe. I have noticed that in Australia cycling is
regarded more as a sporting activity requiring specialised equipment than as a normal, convenient
means of getting around. I have also notice that in Australia people tend to regard cycling as
dangerous. Often when discussing MHLs people will ask me questions that indicate they believe
cycling must be dangerous because helmets are compulsory. The belief that cycling is dangerous,
bolstered by the existence of MHLs, may go some way towards explaining why cycling is not
frequently chosen as a means of transport here as it is in other countries. Furthermore, while
cycling as a sport seems to have gained popularity, the safest form of cycling (riding slowly, for
transport or relaxation, on quiet roads and off-road paths) has been discouraged, further cementing
the idea that cycling is a dangerous activity. There is also evidence of motorist hostility towards
cyclists. In the Adelaide hills recently, a downhill stretch of a road frequented by sporting cyclists
was booby-trapped with oil and tacks designed to pop tyres and cause cyclists to lose control of
their bikes, a premeditated act which poses an obvious threat to cyclist safety. Cyclists are
frequently dismissed in online discussions as being illegitimate road users. Infamously, radio talkback host Derryn Hinch has referred to cyclists as "cockroaches on wheels". This same comparison
was used by a Queensland police officer in on online discussion on the cycle.org.au facebook page
(http://www.qt.com.au/news/police-label-cyclists-cockroaches-sparking-online-/2443076/). It seems
entirely possible that mandatory helmets act as a sort of "uniform" that marks cyclists as different
from motorists, thereby reducing motorists' empathy for cyclists. This is a well-known sociological
phenomenon, associated with the formation of "ingroups" and "outgroups". As a result motorists
tend to view any bad behaviour by a small number of cyclists as typical of all cyclists, while
Personal choice and community impacts
Submission 253
viewing bad drivers as rare exceptions. This lack of empathy may lead to increased incidents of
road rage directed at this group of vulnerable road users and go some way towards explaining why
the countries with lower rates of helmet use tend to also have low rates of cycling accidents.
Although I have emphasised that cycling accidents are rare in Australia, they are still much more
common than in European countries where helmet use is almost non-existent.
I hope the evidence I have presented here will persuade you that mandatory helmet laws do not
lead to a significant safety benefit for individuals. Given the likelihood that they discourage cycling,
leading to increased social costs from poor health, increased pollution and congestion, etc. it also
seems that MHLs do not provide a net benefit to society as a whole. Any assessment of the health
effects of MHLs should not simply consider emergency department admissions, or else there is the
danger of selection bias affecting the assessment. The overall benefits and costs in terms of the low
injury rate per cyclist, and the reduction in heart disease and other illnesses that come with cycling
participation should be properly taken into account. Furthermore, any assessment of the efficacy of
helmets in the case of car-cyclist collisions should not distract us from the importance of preventing
car-cyclist collisions in the first case! Enforced helmet use is a band-aid solution to a deeper
problem, and not necessarily an effective solution, at that. The likelihood that MHLs discourage
cycling suggests that many Australians consider these laws to be an unwelcome imposition on their
personal freedom. The ability of governments to enforce these laws on a small sector of the
community, while failing to enforce equivalent laws on other groups who expose themselves to
equal or greater risks of injury is inherently discriminatory, and seems to be more a result of a lack
of public resistance (imagine the voter backlash if governments tried to make crash helmets
compulsory for all motorists!) rather than a result of good policy-making based on sound evidence.
As such, I believe that while helmets may be desirable in some circumstances, they do not provide
sufficient benefits for all cyclists in all circumstances to warrant the loss of personal freedom that
comes from enforcing mandatory helmet use. I believe that a gradual repeal of MHLs, beginning
with a repeal for adults on off-road paths, to assess the effect upon participation and safety, would
be a positive step towards encouraging more people to cycle, and facilitating a wider range of
transport options for all Australians.
Thank you for considering my contribution to this inquiry,
Dr. Sundance Bilson-Thompson
Personal choice and community impacts
Submission 253
Appendix A - Comparison of cyclist and pedestrian casualties in Victoria, pre- and post-MHL
The most consistent source of information about transport is provided by ABS census data.
Unfortunately this is only available every five years (with 1986, 1991, 1996, 2001, 2006, and 2011
being the years most closely overlapping with the range of TAC fatality data, and of these only
1986 falls before the introduction of MHLs). Travel to work is not a perfect measure of the total
amount of cycling undertaken, but the data is collected in a fairly consistent manner between years.
By comparing the ratios of fatalities per journey to work for different modes of transport, we
automatically cancel out any effects due to population growth and changes in employment levels.
We will only consider fatalities occurring in the 16 - 60 age group, to try to eliminate any
confounding influences from people who were not travelling to work because they were too young
to be employed, or were retired.
The following table lists the number of trips to work by bicycle and walking, for four consecutive
census years
Year
1986
1991
1996
2001
==================================================
Walking
79,580
74,133
63,668
64,732
Bicycle
24,022
18,334
17,190
18,910
The number of walking and cycling trips drops to a minimum in 1996 before making a slight
resurgence (possibly due to an increase in the population of Victoria, counteracting a downward
trend in participation rates). To extrapolate the number of people walking and cycling to the years
between censuses, these data points were used to fit a polynomial of the form
y = C x3 + D x2 + E x + F
Where y is the number of trips, and x is (year - 1986), so that for 1986 data, x=0, for 1991 data x=5,
etc. For walking and cycling data, values of the parameters C, D, E and F were found that matched
the value of y to the census figures. For walking, this yielded the polynomial
y = 22.06266 x3 - 431.3 x2 + 515.533 x + 79580
For cycling, the corresponding polynomial was
y = - 2.24 x3 + 124.48 x2 - 1704 x + 24022
By choosing values of x = 1, 2, 3, etc. these polynomials were used to approximate the number of
walking and cycling trips to work in the years between census years. For instance, when x=1 we
obtain approximations of the number of walking and cycling trips in 1987, the first year for which
TAC fatality data is available. The resulting number of walking trips in each year were
Year
'87
'88
'89
'90
'91
'92
'93
'94
'95
'96
'97
==================================================================
Walking 79686 79062 77841 76153 74133 71911 69623 67397 65368 63668 62429
Bicycle 22330 21094 19970 19054 18334 17795 17425 17210 17136 17190 17359
Although we could include data for later years, the three years before the introduction of MHLs and
the six years after should be sufficient to spot any change in casualty rates.
The number of fatalities per year from the TAC database is
Personal choice and community impacts
Submission 253
Year
1987 `88 `89 `90 `91 `92 `93 `94 `95 `96 `97 `98 `99 `00 `01
===========================================================
Walking
54 62 76 34 46 40 26 27 41 36 41 43 32 27 42
Bicycle
16 12 16 15 8 8 7 4
8
7 4 7
8 9 5
We now wish to find the ratio of cyclist fatalities per journey to pedestrian fatalities per journey for
each year. For instance, for the year 1987 we calculate
(16 / 22330)
0.00071650
W = --------------- = ---------------- = 1.057
(54 / 79686)
0.00067766
or approximately 1.06. A similar calculation for the remaining years from 1988 to 1996 yields the
table of results
Year
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
======================================================
W value
1.06 0.73 0.82 1.76 0.70 0.81 1.08 0.58 0.74 0.72