Menu lock


Sep 5, 2017


The “old” Amsterdam; cycle city

Here’s a meme that pops up in social media every other week:

When it comes to cycling, the argument your city is not like Amsterdam is invalid. 50 years ago, neither was Amsterdam.

Sometimes the cited city is Groningen or Assen, but in all cases the point being made is that the Dutch were headed down the path of high car use in the 1970s, but deliberately chose as a matter  of policy to step back and put a greater emphasis on cycling (see How come the Dutch got cycle paths and not us?).

Driving is still the majority mode, but cycling has around 27% mode share in the Netherlands, compared to less than 1% in Australia. The mode share of cycling for the journey to work is 40% in Amsterdam, 34% in Utrecht, 24% in Eindhoven, 22% in Rotterdam, and 14% in The Hague. It’s 2% in Melbourne and 1% in Sydney (see How big was cycling in Australia in the past? and Which cities in the world are the most “bicycle-friendly”?).

The message is Dutch cities aren’t unique; there aren’t special or structural explanations for the Dutch cycling miracle. Any city can do what Amsterdam did if it takes action! Unsurprisingly, advocates who earn their livelihood from cycling or urbanism put this meme on heavy social media rotation.

So, can Australian cities “go Dutch”? Is there nothing other than lack of political will stopping Melbourne and Australia’s other cities from achieving Amsterdam-like cycling levels?

I think we can increase cycling significantly in Australia, but it’s usually a mistake to think that we can replicate someone else’s experience. A similar argument could’ve been put in the immediate post-war era – and in many cases was – that every city could choose to rely almost entirely on cars. Yet even in the US where some cities achieved this outcome, the mode share of cars in NYC for the journey to work is small; now around 26%.

Clearly there are other forces at play – like legacy densities and infrastructure – that have a big influence on the extent to which any city can “choose” to change travellers’ behaviour. There are big variations in the level of cycling across European cities too; the mode share for the journey to work is 2% in Barcelona, 12% in Hamburg and 30% in Copenhagen. The fact that 40 years ago “Amsterdam wasn’t Amsterdam ” doesn’t logically mean that other cities can do over the next 40 years what Amsterdam did in the past.

An important factor for Australian cities is we don’t have anything like the tradition of cycling in the Netherlands. Cyclists made up 70% to 90% of traffic in Holland in the 1930s. The Dutch have been pro-bike for 100 years, not just since the 1970s. According to social historian Anne Ebert, the bicycle is an “important object for Dutch national identification” (see How big was cycling in Australia in the past?).

I’m afraid there’s no evidence that cycling was ever particularly big in Australia so it’s never been seen as a natural alternative to driving. Even in World War 2, with fuel rationing and other privations, the Bureau of transport, infrastructure and regional economics estimates cycling’s mode share was under 10%. It’s just never been regarded as being as useful here as it has been in the Netherlands.

The cost of driving relative to incomes was historically much higher in the Netherlands, making cycling relatively more attractive. Australian cities are also larger and less dense. We seem intent on keeping them that way and extending their reach into the regions; that’s made driving relatively more attractive than other options. Aside: it would be interesting to compare the quality of public transport in Australian cities with Dutch cities over the last 100 years or so to see if it was a factor bearing on the relative attractiveness of cycling.

It’s generally argued that pedestrians being run over by drivers in the 60s and 70s was one of the key reasons that political support could be rallied in the Netherlands to take action to reduce the competitiveness of cars and promote cycling. Australia also experienced high casualty rates in that era but the response was to make roads and cars safer. It’s a much less potent issue in Australia now; road deaths peaked at 30 per 100,000 persons in 1970 and fell continuously to 7 per 100,000 by 2009 (see How come the Dutch got cycle paths and not us?).

And here’s one that tends to be underplayed; topography has a big influence on the level of cycling and Dutch cities are very flat – see Is cycling so successful in Amsterdam because it’s as flat as a pancake? While they have flat parts, Australian cities have a more varied topography; even parts of Melbourne’s CBD are steep by the standards of Amsterdam. Riding a Dutch-style 20 kg bike with just three gears – or just one in the case of oBikes – up the northern end of Collins or Bourke St is a big ask for someone who’s a casual rider (see What are the prospects for dockless bike share in Australia?).

The mode share of cycling in Australian cities depends in part on how travellers view factors like the effort involved, exposure to weather, and danger of cycling on roads. It also depends in part on the competitiveness of the alternatives i.e. cars, public transport and walking. The Dutch cycling miracle wasn’t the result of a dictatorship; it happened because of political pressure. A key issue for Australian cities is to identify what would motivate travellers to willingly forego the ease of other modes and take up cycling on the scale of Amsterdam.

Of course Australian cities don’t have to emulate Dutch cities in order to do much better; if cycling were to achieve (say) a 10% mode share across Australia’s capitals that would be hugely beneficial for urban life if it were mostly at the expense of driving. It’s equivalent to what public transport averages across our capital cities at the moment. I think we’ll see a lot more two-wheelers in the future, but I expect that unlike Amsterdam today, they’ll mostly be motorised e.g. electric scooters.


Aug 9, 2017


Average weighted trip distance (kilometres) for all modes, by purpose, for inner, middle and outer Melbourne (source data: VISTA)

The recent refresh of Plan Melbourne by Victoria’s Andrews Government confirmed the goal of creating a 20-minute city is a centre-piece of the city’s strategic plan. Here’s the idea:

The 20-minute neighbourhood is all about ‘living locally’ – giving people the ability to meet most of their everyday needs within a 20-minute walk, cycle or local public transport trip of their home.

The hoped-for benefits are enormous; it seems there’s nothing that can’t be achieved via land use planning:

A 20-minute neighbourhood can create a more cohesive and inclusive community with a vibrant local economy – reducing social exclusion, improving health and wellbeing, promoting a sense of place, reducing travel costs and traffic congestion, and reducing carbon emissions across the city as a whole.

To give credit where it’s due, the refreshed version of Plan Melbourne gets rid of the jejune idea in the original Plan that jobs must be within 20 minutes travel time by active modes:

Due to the specialised and diverse nature of many people’s work, access to employment will often be outside the 20-minute neighbourhood.

So, is the 20-minute city a compelling idea that warrants being at the top of the policy priority list? Or will it soon be forgotten like the target set in the previous metropolitan strategic plan published in 2002, Melbourne 2030, to achieve 20% travel by public transport by 2020? According to VISTA, public transport’s share of all weekday trips is currently around 9% and its share of motorised trips is circa 11%.

It would be nice to live in a place where you have the option of not having to drive in order to meet your everyday needs (except work!). And there are places where you can e.g. small country towns, tourist resorts. It’s a nice idea, but it’s mostly fluff; it’s more about marketing than serious policy.

Politicians like Malcolm Turnbull love the idea (his version is 30-minutes) because it’s inherently long-term, so it won’t take much from the current budget, won’t inconvenience anyone in the short-term, and the current Government knows it will be long gone before its ever called to account. Key constituencies like the rhetoric around traditional village values and sustainability objectives.

But nice isn’t the same as plausible. You can have anything you want – even congestion-free, untolled motorways – if you’re prepared to cop the financial, social, environmental and economic cost. Few are though. The problems with this idea for real-world policy are numerous.

The first issue is that the defining access limit for active travel with an ageing population must be weighted towards walking. Even in the Netherlands and Denmark, considerably less than half of trips are made by bicycle. Access and wait times make it hard for public transport to be competitive for short time/distance trips.

At present, 60% of walking trips in Melbourne are less than one kilometre and 88% are less than two kilometres, suggesting a limit of around one kilometre, perhaps even less. The average trip distance in the middle ring suburbs by all modes, however, is 5.3 km for shopping, 6.5 km for education, and 8.1 km for social and recreational trips. It’s not much shorter in the inner suburbs (see exhibit).

So the scale of change to existing land use patterns, especially the density of activities and population, would have to be enormous to achieve even a semblance of a 20-minute city in the suburbs where over 90% of the metropolitan population lives. The scale and intensity of opposition to higher density developments in Australian cities suggests that’s optimistic. Something close to it will likely be achievable in the inner city and in pockets of high-density redevelopment around centres, but that’s only a small minority of residents.

Another huge problem is many people would still choose to drive to the local supermarket and the local doctor, or take their kids to the local school in the passenger seat. Three quarters of kids already get driven to primary school. Simply having the option of walking won’t by itself automatically change residents’ preferences away from driving.

In any event, there’s no reason to think that people would necessarily shop, dine or get haircuts locally just because there are service providers within a 20-minute walk. We’ve got a large local centre, but my household largely chooses to travel elsewhere e.g. we regularly travel 9 km to Preston Market, we like to dine at particular restaurants, patronise specialty bookshops, buy hardware at Bunnings, and our children don’t attend local schools/universities. Most of our friends and family don’t live within 20 minutes by walk, bicycle or transit.

What a dense network of walkable local centres can’t offer is the range of services and, in particular, the high level of specialisation, available at a larger centre. Those who’re constrained, or choose, to meet their everyday needs locally, will on average have less choice and/or face higher prices. Those who have a choice will more often than not drive somewhere else.

A big problem is that without the journey to work, the “20-minute city” is a much less substantial idea. The commute is what mostly drives peak hour congestion and determines the capacity and cost of infrastructure. It’s the most important trip most people make. Many work trips are incompatible with the 20-minute city as Plan Melbourne now recognises, but that only goes to show the whole idea is a sideshow.

Plan Melbourne doesn’t acknowledge it, but there are other trips that also don’t fit with the 20-minute limit. For example, trips where specialisation or economies of scale and/or agglomeration are important e.g. to universities, sporting events, medical specialists, hospitals, museums, festivals, private schools, family and friends. Their absence further reduces the importance of the 20-minute idea. It’s nice to have the option of walking to Woolies or Coles, but it’s hard to justify it as a headline objective in a 30 year metropolitan strategy.

And think about this; even in the centre of Paris – a radius of roughly five kilometres – there are many places that can’t be connected within 20 minutes by transit, bicycle, or walking e.g. use Google Maps to get to Sciences Po (university) near the geographical centre from elsewhere in Paris intra-muros. That’s in a city with one of the world’s best metros and one of the highest population densities in a developed country. Then try getting to Science Po from the Banlieues in 20 minutes.

Reducing travel times and improving mode share are important objectives of policy that warrant real action, not spin. It’s time to stop enabling politicians to get away with merely waving the flag; it’s time to stop signalling and start moving.

Update: see later extensions of the arguments made above: Are the politicians trying to con us on this one? and Can public transport define the 20-minute city?


See also: The “20 minute neighbourhood”: does it make sense?Surely the 30-minute city makes sense for primary school trips?Does Turnbull’s ’30-minute city’ work for secondary school trips?Is Turnbull’s “30-minute city” all spin (or a really useful idea)?


Dec 13, 2016


Density of motorways by country- 2009* (km motorway per 100 km² land area) Source: ERF - European road statistics 2011
Density of motorways by country- 2009 (km motorway per 100 km² land area) Source: ERF – European road statistics 2011

One of the more interesting papers on cycling I’ve read in quite a while just appeared in the latest edition of the international journal, Safety Science. In The Dutch road to a high level of cycling safety, Schepers et al investigate why the Netherlands achieved an 80% reduction in the number of cyclists killed per kilometre over the last 30 years. From the abstract:

Factors found to contribute to this improvement include the establishment of a road hierarchy with large traffic-calmed areas where through traffic is kept out. A heavily used freeway network shifts motor vehicles from streets with high cycling levels. This reduces exposure to high-speed motor vehicles. Separated bicycle paths and intersection treatments decrease the likelihood of bicycle–motor vehicle crashes.

The high amount of bicycle use increases safety as a higher bicycle modal share corresponds with a lower share of driving and greater awareness of cyclists among drivers. Low cycling speed was also found to contribute to the high level of cycling safety in the Netherlands.

Traffic-calmed areas are extensive in the Netherlands and extremely important for safety; almost 60% of all cycling in urban areas is done in them. Separation of cyclists from motorists is also critical given that over 80% of all police-reported fatal and severe bicycle-vehicle crashes in built-up areas occur on distributor roads where exposure to high-speed motor vehicles is greatest.

But while it’s common to recognise the benefits of segregated infrastructure, the vital consequential question of what happens with cars and trucks that lose road space to bicycles is usually overlooked.

The authors point out a key way the Netherlands achieves high levels of safety for cycling is via the country’s motorway network. Whereas around a quarter of all vehicle kilometres of travel in Europe (and less than one third in the US) take place on motorways, they say the figure is around a half in the Netherlands. At 57 km per 1,000 sq. km, it has the has the densest motorway network in Europe.

While both are important, they say the level of network separation is a much more important contributor to safety than the level of bicycle modal share. It’s tempting to assume that road space given over to cycling paths leads to vehicle trips being replaced by public transport, but that evidently doesn’t happen to a sufficient extent in the Netherlands to obviate demand for motorways. As dense cities like Amsterdam and Paris that otherwise tick many of the boxes of ideal urbanism show, reducing the use of cars is politically very hard; it doesn’t take many vehicles to generate serious congestion, environmental and safety impacts in dense areas.

The researchers also found that other policies – like intersection design – are important in achieving a high level of cycling safety. However, they concluded some policies don’t appear to help much:

There are also a number of factors…(that make) no or at most a very small contribution to cycling safety, i.e. comprehensive traffic education and training of both cyclists and motorists. There is also a lack of research about the impact of traffic regulations that favour cyclist and pedestrians and we did not find the introduction of strict liability in the Netherlands in 1994 to be associated with cycling safety.

Another interesting point the authors make – it’s not directly related to safety – is cycling was always a popular mode of transport in the Netherlands. Although it’s increased sixfold since then (see How come the Dutch got cycle paths and not us?), even by 1965 there were already 6,000 km of cycling paths in the country.

(Cycling) has been at a high level for almost a century and it was only at the beginning of the sixties that the number of kilometres travelled by motor vehicles exceeded the number of bicycle kilometres (during which time bicycle use started to decline).

This is consistent with the case made by others that the Netherlands has a long history of widespread bicycle use that predates the boom of the 60s and 70s. This is an important insight for assessing the potential of cycling in Australia; contrary to the romantic notions of some, Australia never had a very high level of cycling compared to European countries (other than during WW2) and we never had the cultural identification with cycling as a means of transport (as distinct from a sport) that the Dutch have (see How big was cycling in Australia in the past?).

Cycling has great potential as a mode of transport in Australian cities, but too often it’s blithely assumed everything that happens in The Netherlands or Denmark can be transplanted here willy nilly.


Nov 7, 2016


The more cyclists a country has, the fewer fatalities it will tend to experience
The more cyclists a country has, the fewer fatalities it will tend to experience

This chart from Statista (see exhibit) shows how fatalities from cycling vary inversely with the level of cycling; fewer cyclists die (per kilometre) in countries with higher rates of cycling (km per capita). Statista suggests it’s due to the “safety in numbers” effect:

In an environment where bicycles and the infrastructure to support them are plentiful, awareness will be high. By contrast, a lone cyclist in a traffic choked urban environment is far more likely to end up in an accident, due to motorists who may not be aware/used to his or her presence. 

Some caution is needed in interpreting the numbers in the chart. Reliable comparative data, especially kilometres of cycling, is hard to come by on a country-by-country basis. Also, this is a very small sample of countries and it’s highly selective; it’s likely the relationship would be much less marked given a substantially larger number of countries (or better still, cities). Note it also measures fatalities, not injuries; the former are relatively few in absolute terms e.g. 185 in the Netherlands (and 32 in Australia) in 2015.

But if we accept there’s probably some sort of inverse relationship between these variables, I doubt the explanation is as straightforward as Statista contends. Here are some hypotheses that might explain it:

  1. Infrastructure. Places with high levels of cycling also tend to have a high standard of cycling infrastructure that segregates cyclists from collisions with vehicles e.g. separated cycleways. They also tend to have road rules that support cycling e.g. low speed limits. In the Netherlands, there’s also a presumption that drivers who collide with a child cyclist are at fault (see Are Dutch motorists strictly liable if they collide with a cyclist?).
  2. Safety in numbers. As Statista argues, large numbers of cyclists means motorists expect to see riders more frequently and adjust their driving behaviour accordingly.
  3. Motorists are cyclists too. In places where cycling’s mode share is very high, there’s a higher likelihood that the average driver also cycles and hence can empathise with riders. 
  4. Culture. In countries with a long history of cycling, motorists are more likely to acknowledge and accept the legitimate place of riders on roads and treat them respectfully.
  5. Selection. Risk-takers constitute a high proportion of riders in low cycling countries like Australia. In contrast, the risk profile of riders in places like the Netherlands is necessarily a closer match to the relatively more risk-averse profile of the population.

All of these contribute independently to varying degrees but it’s complicated by feedback loops galore. It’s unclear how important each one is, but I think number two, safety in numbers, is probably the least important. In any event it’s probably the density of cyclists that matters more than the numbers (see Cycling: is the Safety in Numbers effect all about the numbers?).

On the other hand, I think the selection effect is under-appreciated. Risk takers are over-represented on Australian roads. We know this because riders who suffer serious injury are mostly “lycra louts”; almost 70% were wearing cleated shoes at the time of their crash (see Should cyclists need a licence to ride on public roads?). Moreover, close to a third of cycling fatalities are due to ill-health e.g. heart attacks triggered by exertion (see Is it just vehicles or are MAMILs killing themselves too?).

Compare that with the Netherlands. In a city like Amsterdam where 38% of all trips are made by bicycle, the vast majority of travellers aren’t likely to be risk-takers; they’re people who cycle on roads in order to get places, not for fitness or sport. The average Dutch rider is accordingly more cautious than the average Australian rider.

While it’s hard to disentangle the variables that make for safer cycling, I think improving infrastructure is the key action (see Do more bikeways mean more bicycle commuting? and Does cycling infrastructure reduce serious accidents?). By reducing the scope for conflict between motorists and riders, infrastructure increases riders’ sense of subjective safety and encourages new cohorts to take up riding for transport. That in turns puts pressure on politicians to regulate negative driver behavior and invest in more infrastructure. New cohorts of riders will necessarily be more risk-averse than current generations.


May 17, 2016


Vale Alan Parker OAM
Vale Alan Parker OAM (image source: Yarra BUG)

One of the legends of utility and recreational cycling in Australia, Alan Parker OAM, died on Easter Monday. Alan was a co-founder of the Bicycle Institute of Victoria (now Bicycle Network) in 1974 and more recently a regular commenter on these pages on cycling and other topics.

Alan was way ahead of the times in his vision for cycling in the 1970s and 1980s. Although it was a key mode in the pre-war years, by the 1980s its share of adult travel in the inner and middle suburbs of Australia’s capitals was close to non-existent (see How big was cycling in Australia in the past?).

While there were residual pockets in some outer suburbs and provincial centres – mainly blue collar riders – hardly any adults commuted to the CBD by bicycle in those days (Did Sydneysiders cycle more in the 1980s?). Utility cycling was something mainly done by school children (see Did students cycle to school back in the day?).

Thanks to the efforts of activists like Alan, significant improvements were made to the infrastructure for cycling in the decades since then, especially the extensive network of shared recreational cycling paths built in cities and towns across the country. That’s contributed directly to the rising popularity of cycling for commuting and recreation (see How important is cycling in Australian cities?).

You can read more about Alan’s life and his contribution to cycling in this obituary written by his wife, Doreen Parker, published in The Age last Friday (Alan Parker obituary: Sprocket Man fought for cyclists’ rights on the road). There’s also this post by The Yarra Bicycle User Group (Vale Alan Parker OAM).

It’s potentially useful for policy to think about why, despite the best efforts of Alan Parker and others, cycling as a means of transport (as distinct from recreation) was largely ignored by policy-makers and travellers alike in Australian cities until relatively recently. As welcome as it is, the recent gains made by cycling are modest compared to the significant increase in many Dutch cities starting from the 1960s (see How come the Dutch got cycle paths and not us?)

If back in the 1970s, cities like Sydney and Melbourne had the same level of cycling infrastructure as they have today, would cycling have been as popular as it is now? I expect it would’ve done better than it did, but I doubt it would’ve approached today’s levels.

Why not? A key factor, I think, is simply that driving was easier then. While it was more costly in real terms than it is today, it was getting cheaper relative to incomes compared to prior decades. Traffic congestion was less of an issue, too.

Driving was still a plausible option for near-CBD and inner city peak hour travel. There was also less competition for parking. Driving to near-CBD jobs and universities wasn’t easy but it was more feasible than it is today. (1)

Outer suburban and regional city blue collar workers who were (relatively) large users of bicycles at the time for commuting started buying cars. They cycled because it was cheap; driving was regarded as infinitely superior and preferred as the real cost dropped (see Did the helmet law reduce commuting by bicycle? and The helmet law and commuting in Sydney and Melbourne).

Another factor is Australia didn’t have anything like the historic levels of cycling that countries like the Netherlands did. As driving became progressively harder, there was no cultural tradition nudging travellers toward bicycles (see How big was cycling in Australia in the past? and How come the Dutch got cycle paths and not us?).

Cycling for transport is largely an inner suburban phenomenon today (see Do political values help explain high cycling levels?). But back in the 1970s the CBD was declining as an employment and residential location. The great wave of services employment in the city centre was in its infancy. The knowledge workers who today populate the inner city – and account for a large proportion of utility cycling – were starting to arrive but outside some pockets the conquest was far from complete.

Today, cycling is seen as a way of getting fit and as a means of minimising environmental harm. Both these factors were much less prominent in the 70s and 80s; it seems astonishing today, but 49% of males aged 18- 44 years smoked in 1977!

Cycling was also portrayed as a dangerous activity, especially in the 80s. There was a consensus that it’s a risky activity, ultimately leading to the mandatory helmet legislation passed in all Australian States in the early 1990s with hardly any opposition. (2)

Despite the logic of their case, the early cycling activists faced significant structural obstacles. That cities like Melbourne nevertheless built a major network of shared trails which today play a major part in the city’s cycling network was in large part due to the persistence and political skills of activists like Alan Parker.


  1. More speculatively, I expect that public transport was more comfortable back in the day for CBD workers. On the other hand, perhaps it was less reliable due to more frequent industrial action?
  2. The idea that cycling is dangerous was so firmly embedded in the Australian psyche that even if the helmet laws were repealed tomorrow it mightn’t have as big an effect as opponents imagine on either cycling levels or helmet-wearing rates. The greater part of the “damage” was done by the pro-helmet campaigns of the 1980s; the legislation was the cream on the cake.


Nov 24, 2015


Amsterdam’s cycle paths attract, ahem, a diverse range of users

I was in Amsterdam for six days last month. This is the city where cycling averages a phenomenal 38% mode share across all trip purposes.

It’s good to be reminded of what can potentially be achieved and to try and figure out how they did it. So I took a (completely unscientific) straw poll of ordinary Amsterdammers, asking them why they thought cycling levels are so high in their city.

They gave reasons like the high standard of cycling infrastructure, the high cost of driving, the relatively short cycling distances, and the long tradition of cycling in the Netherlands.

All eminently plausible reasons. But they also prefaced their explanations with words to the effect “of course it’s flat” as if it were so obviously the most important explanation it didn’t need further elaboration.

Flatness doesn’t usually get much emphasis – and in many cases doesn’t even get mentioned – in discussions about the potential of cycling as a mode of transport in car-oriented cities like Australia’s.

It’s clearly not a sufficient condition for high levels of cycling; plenty of cities with relatively flat topography nevertheless have levels of cycling an order of magnitude below Amsterdam’s. There’s obviously more to it.

But is it a necessary condition for the extraordinarily high mode shares seen in places like the Netherlands? Unfortunately, I haven’t seen any data that isolates the effect of topography.

This writer says it’s a myth cycling can’t do well in cities with hills. He reckons even San Francisco isn’t a problem:

I rolled up and down the hills of San Francisco on a one-speed Biomega, together with friends on upright bikes. I was unimpressed. And I’m just a normal schmuck in normal clothes, not some Captain Spandex MAMIL.

Seriously? I lived in downtown San Francisco when I was a student; those hills are bloody steep (c.f. Bullitt). Stick to the water’s edge and it’s flat but those hills are daunting for anyone other than sport cyclists.

There’s a sound argument that cycling can nevertheless still grow in hilly places if the right policy choices – like improved infrastructure – are made, but winning a really high mode share in such locations seems unlikely.

I don’t think topography is given enough attention. All the famous Dutch cycling cities are very flat. So is most of Copenhagen. I don’t know for sure how much flatness contributes to the extraordinary success of cycling in those cities but I suspect it’s a lot.

Flatness means less effort, so there’s less sweat and less need to wear special clothing. It helps explain the preference  for heavy bicycles in some places.

They’re more comfortable because they have big, fat bump-absorbing tyres. An upright riding posture makes more sense because obtaining mechanical or aerodynamic advantage isn’t as critical in a sympathetic environment.

Bikes can be loaded up with enhancements like baskets, chain guards, bike stands, and dynamos. The option of carrying shopping, children, a passenger, or everyday items like a change of clothes and a bundle of textbooks isn’t as off-putting as it is in undulating places.

I expect flatness is a key reason why cycling in the immediate post-war era in Australia was strongest in certain regional centres like Bundaberg in Qld and Grafton in NSW but not in others.

The role of topography is often downplayed or dismissed by cycling advocates because, I suspect, it doesn’t fit with the dream that all cities can potentially emulate Amsterdam, or at least Copenhagen.

Those with a stake in promoting cycling prefer to focus on matters where changes can be made.  Topography and climate are inconveniently immutable; they’re ‘nature’ whereas something controllable like infrastructure provision is ‘nurture’.

The lay of the land isn’t destiny; it doesn’t mean cycling can’t win a significant mode share in non-flat places. But my intuition is topography imposes a limit; even with the same suite of policies, localities with an undulating landscape aren’t likely to come close to emulating Amsterdam.

Having said that, the future might be very different. The increasing popularity of affordable, power-assisted bicycles should make topography largely irrelevant for cycling.

Transport - general

Aug 17, 2015


Mode share for the journey to work, 1951 and 2011, Melbourne (source data: Thompson and Maginn; ABS Census 2011 via .id)

The exhibit shows public transport was very popular in 1951; the majority of workers in Melbourne travelled to work on trains, trams and buses.

That’s probably not surprising; what’s perhaps less expected is the relatively minor role of cars. Even decades after cars became a familiar site on Australian roads, they accounted for only a fifth of trips to work in Melbourne by 1951

The data is from Australia’s first travel survey, conducted by the Melbourne and Metropolitan Board of Works. The 2011 data was extracted from the Census by .Id. (1)

According to the authors of Planning Australia: an overview of urban and regional planning, Susan Thompson and Paul Maginn, driving was still an expensive option:

Melbourne in 1951 was a city dominated by public transport, with walking and cycling also playing significant roles; car use was confined to the wealthy.

The share of work trips taken by car, truck and motorcycle was 74% in 2011 but back then it was only 19%; that was considerably less than the combined 24% taken by bicycle and on foot at the time (2).

The majority of commutes – 57% – were by public transport. Trams were much more important than today, carrying nearly as many workers (22%) as trains (26%).

The greater mode share of trams was in large measure the result of the tram system covering a much larger proportion of the then urbanised area of Melbourne than it does today.

In fact more workers used on-street public transport – trams and buses – than used trains; 31% vs 26%. The efficiency of these modes would’ve been greatly helped by the relative scarcity of cars.

With a mode share of 9%, cycling was an important means of transport at the time. That’s close to the share of commutes trains attract today (11%) and much larger than the share carried by trams today (2%).

The region with the highest level of cycling by far was around Williamstown, Footscray and Sunshine, where 26% of commutes were by bicycle (the next highest region, in the then outer south-east, was 12%).

This was probably mostly because commuting by tram was very low in this area (2%), no doubt due to the limited network in the region. In comparison, 32% of commutes in Sth Melbourne and Pt Melbourne were by tram and 10% by bicycle.

Walking averaged 14% of work trips across the metropolitan area compared to 7% today. It was much higher in industrial areas like Fitzroy, Collingwood and Richmond (29%) but a lot lower in predominantly residential areas like Camberwell, Box Hill and Nunawading (6%) where rail, car and tram were the largest modes.

There are other important and interesting variations at the regional level; I’ll leave them for another time. The pattern of travel in 1951 would of course change massively in the years ahead. Following the abolition of petrol rationing in 1950, Melbourne was, as Thompson and Maginn say, at the dawn of the automobile age.

The big story here is the relatively low mode share of cars, but two other points are worth making.

First, the numbers reinforce other evidence indicating cycling was never in the same league in Australia as it was in countries like the Netherlands. As I’ve noted before (e.g. see How big was cycling in Australia in the past?), we’re not the Netherlands:

The bike has a different history here and occupies a different cultural niche. The role of cycling and the way it’s implemented is bound to be different in Australia’s future.

Second, public transport, cycling and walking dominated at a time when cars were simply too expensive relative to incomes to produce mass ownership.

Building more infrastructure won’t be enough by itself to dramatically increase the share of trips captured by non-driving modes; that will only happen if the competitiveness of private vehicles is reduced significantly.


  1. Melbourne’s population at the time was 1.3 million, compared to Sydney’s 1.6 million and Brisbane’s 0.4 million.
  2. Comparing the data on walking in 1951 (14%) with 2011 is somewhat fraught because the MMBW combined commuters who walked to work with those who worked at home. It’s probably not a big deal though given that prevailing attitudes to work organisation in 1951 and the limitations of communication technology at the time suggest the proportion who worked at home was probably much lower than it was in 2011 (4.1%). There were probably more who lived and worked “on premises” in 1951, though e.g. station masters.


Mar 15, 2015


Modal shares for urban passenger transport 1900–2011 (source: State of Australian cities 2012)

It’s risky when thinking about the potential of cycling in Australia to make simple comparisons with leading cycling countries like the Netherlands.

It would be possible to significantly increase the level of cycling in Australian cities (currently less than 1%) by building better infrastructure like the Dutch have done, but in order to approach the mode share of Amsterdam and Assen (40%) or Groningen (60%) it would also be necessary to make other forms of transport – especially driving – relatively less attractive.

** The winners of The Urbanist’s comp for the Grattan Institute’s new book, City Limits, are Taylor R and James. Go here for details **

But that probably wouldn’t be enough. One advantage we don’t have in Australia is the Netherland’s long history of cycling. According to the Bureau of Infrastructure, Transport and Regional Economics (BITRE), cycling has always made a much smaller contribution to Australia’s transport task than it has in places like the Netherlands (see here and here).

BITRE’s best estimate is cycling accounted for around 3%-5% of all trips in Australia from 1900 through to the 1930s. Its share increased with the onset of the depression and nearly doubled during WW2, then subsequently fell right away from the 1950s (see first exhibit).

In contrast, cycling has a long history in the Netherlands according to author, journalist and publisher, Carlton Reid (see Why is cycling popular in the Netherlands: infrastructure or 100+ years of history?).

While it got off to a relatively slow start, he says the Netherlands has been Europe’s top cycling nation since 1911. He cites a passage from the 1999 Dutch Bicycle Master Plan which says cyclists made up 70-90% of traffic in Dutch cities in the 1930s.

It’s important not to underestimate the popularity of cycling in the Netherlands before the 1970s. The Netherlands hasn’t had 40 years of being pro-bike, it’s had 100 years of being pro-bike…

Mr Reid goes on to argue that the reason for cycling’s popularity is, in the words of social historian Anne Ebert, because it is an “important object for Dutch national identification”. She says:

The tremendous success of the bicycle in the Netherlands can be at least partly explained by the particular way in which the bicycle was constructed and conceived as a promoter of Dutch national identity. To be Dutch meant to cycle, and this viewpoint remained prevalent until the Second World War, and – arguably to a lesser degree – remains so to this day.

Whether it’s cause or effect isn’t clear, but Mr Reid says the Dutch started building dedicated bicycle paths in the 1890s. They were provided in part to separate cyclists from pedestrians and horse traffic, but also to provide routes for what were the fastest vehicles at the time:

By the 1920s it had been laid down in National Law that the construction of these separate cycle paths was mandatory on roads with more than 500 cyclists passing per day. When the cyclists’ union looked back in the 1930s to three decades of practise, they were very satisfied that this solution had also improved overall road safety.

BITRE’s estimate of the level of cycling in Australia prior to the 1950s is approximate because there’s not much data to draw on. But the circa 5% mode share estimate in the 1930s isn’t trivial; it’s the same as the share of all trips carried by trains in greater Sydney today (5.4%).

Of course like Wynyard or Flinders St stations at peak hour, there would’ve been particular locations (and times) where cycling’s share was much higher and others where it was much lower.

Given the shortcomings in the available data, the possibility that cycling did better prior to the 1950s than BITRE calculates can’t be ruled out. (1)

One reason I’m inclined to think it’s probably a reasonable estimate though is that bicycles were very expensive in Australia when they first came on the market. The ABS Year Book 2001 says they were a luxury item for most people at the time of Federation. (2)

A new bicycle…cost the equivalent of more than seven weeks wages (in 1901) whereas today you can buy a good quality bicycle for…less than half a week’s wages.

Another reason is secondary sources like old photos and films don’t provide much evidence that cycling was a bigger force historically than BITRE estimates (e.g. Sydney 1906Melbourne 1910 and Melbourne 1954). I know it was big in some provincial centres but even 5% might be generous in the major cities.

Hopefully readers can help, but the only photo I could find showing a high level of urban cycling for transport purposes is the one in the second exhibit from the book Wheeling Matilda by Jim Fitzpatrick.

It’s from WW2 when BITRE says cycling got up towards 10% mode share nationally. That’s small compared to the claims made for the Netherlands but it’s still very impressive; it’s equivalent to public transport’s mode share today averaged across all Australia’s capital cities.

If Wheeling Matilda is right, cycling seems to have been mostly a sporting, recreational and rural phenomenon rather than an urban transport one and lasted for a relatively short period, yielding quickly to the car after WW1.

According to Jim Fitzpatrick, the decline in rural cycling in Australia began at the end of WW1 due to the “increasing availability, reliability, affordability and comfort of motor vehicles”.

Making cycling in Australia more attractive by an order of magnitude or more will require both better infrastructure and strong constraints on car use. We’re not the Netherlands though; the bike has a different history here and occupies a different cultural niche. The role of cycling and the way it’s implemented is bound to be different in Australia’s future.


  1. While one could instead make a guess (or perhaps more commonly assume what fits one’s priors!), I put greater store by the more rigorous approach adopted by historians when dealing with matters where wholly reliable data isn’t available.
  2. It would be useful to know how the price of bicycles influenced the take-up of cycling in the Netherlands. Carlton Reid reports that the Netherlands lagged the rest of Europe in cycling until 1911.
Why aren't there many more photos like this? A Melbourne street during WW2 rush hour (source: Wheeling Matilda)


Mar 27, 2014


The Dutch PM travelling to work (Source: @GernotWagner)

Guest writer Jeremy Dore is a lawyer with Aboriginal Carbon Fund and a lifelong cyclist:

Okay, let’s get it out of the way: 1. Bikes can overtake on the left, as long as the car is not indicating and turning left (Australian Road Rules regulation 141). 2. People getting out of cars must not cause a hazard to anyone (regulation 269).

So dooring incidents, like the one in Collins St, Melbourne last week, are nearly always the car occupants’ fault. Editors at The Australian who think blame cannot be apportioned take note.

For over 20 years now I’ve pushed the pedals for freedom and fitness – something the prime minister, Tony Abbott, would understand – and also to visit my mum for dinner. Every time I ride around my worrywart mother panics this will be the time I won’t make it over. Oh she loves to worry! I threaten to buy her a violin. Occasionally I take the train.

But she’s right.

The Urbanist keeps on top of the numbers. Kilometre for kilometre cyclists riding on the roads of Melbourne are about 4.5 times as likely to die as a car occupant. And the risk of serious injury is about 13 times the risk for car occupants. It’s sober reading for unavowed cyclists (the risk is reduced by looking from a time basis as bikes are slower to go places).

I fly, I swim in the ocean and occasionally I even walk down Lygon St. But cycling is by far the riskiest thing that I do. My catalogue of lunatic near misses is testament to this risk.

But there’s more.

Since the 1970s, road safety measures have ensured that the road toll has plummeted even as car numbers have skyrocketed – that is, road safety measures have managed to decouple risk of death from the increasing number of road users. Using the federal government’s Australian Road Deaths Database, in 1990 the road toll was 2,331 and last year it was 1,196. In about 20 years it’s basically halved and it’s still dropping.

What about cyclists? The numbers jump around a bit more than for the total road toll, but in the 1990s, there were on average 52 cycling fatalities per year and in the 2000s 36 per year. So far this decade, the average has crept up to 39 (only on 4 years) and last year the total was 50 – the highest total since 1997. Locally here in Victoria, police have reported a 125% increase in dooring incidents between 2000 and 2010. And the rate of serious injuries is also on the move – up 109% in eight years.

So for a while, it was getting safer for cyclists too, although not as much as for cars. But now there is evidence it is getting more dangerous again. Time for action?

Doug Hendrie made the call on the ABC this week to separate cars and cyclists to reduce risk. His call is backed by a 2006 Australian Transport Safety Bureau report which found the most common cause of death for cyclists was being hit from behind by cars. These days I certainly take the bike path to the folks’ place a lot more than I used to. More separation makes sense.

Ultimately, it is the job of governments to identify risks and respond to them. Are they helping?

The previous federal government made a start with the Walking, Riding and Access to Public Transport report released in 2013. It acknowledges cycling as an important mode of transport that promotes healthy living. It notes the lack of continuous, convenient connection as a key barrier. But it is just a report. The new federal government’s Improve Road Safety policy “shares the concerns of cycling bodies across the nation that the increased participation in cycling, for health, recreation and transport, has not been matched with the same degree of improvements in infrastructure and community information about sharing the road environment.” Yes!

But this concern has not yet been matched by spending commitments. The government’s mid year economic statement, MYEFO, found an extra $8.2 billion over 6 years for major infrastructure projects, but no money for cycling infrastructure – it’s all about highways and roads cyclists can’t even use.

In my state of Victoria, things are no better. Here the current government spends a miserly $30 million per year on new cycling projects (I don’t have to tell you how much is spent on roads – some of which cyclists can use).

The fact is people want to ride. But it’s dangerous. And it’s starting to get worse. It’s time for governments to step up and do a lot more to assist this important transport option. Otherwise the menace of power elites will continue to write revolting editorialsand cyclists will die at increasing rates.

More politicians out cycling this week in the Netherlands. Where's Julie Bishop? (Source: @KentsLundberg)


Jun 13, 2013


How liability in collisions between motorists and cyclists is worked out in The Netherlands

It’s widely believed elsewhere that motorists in the Netherlands are strictly liable for damages and injuries if they collide with a cyclist, even if the cyclist is at fault. It’s thought Dutch drivers are therefore more cautious, making cycling safer and resulting in more riders.

Strict liability is an attractive meme but it’s not true, according to Mark Wagenbuur at Cycling NL. The Dutch don’t even have an equivalent term in their language for strict liability. The key reason for the high level of cycling in the Netherlands, he says, is the high standard of cycling infrastructure.

Nevertheless, the Dutch law on who’s liable for the cost of damages and injuries in collisions between motorists and cyclists takes a sympathetic view of the cyclist, especially if he’s a child under 14 years of age.

The objective of (s 185) in the law is to protect vulnerable road users from financial damage caused by drivers of motorised vehicles. Because due to the differences between motorised and non-motorised road users, it is very likely that the latter will suffer more and more severe damage and/or injuries when both are involved in a traffic accident. The law also considers the fact that drivers are obliged to be insured for such damage and non-motorised road users are not.

Mr Wagenbuur’s article, Strict liability in the Netherlands, is lengthy and is translated from Dutch. So here’s my summary of the key points (also see first exhibit) but bear in mind I’m not a lawyer. Those who’re interested in the detail should read the article.

With one relatively uncommon exception I’ll discuss below, a Dutch motorist is liable for all damages/injuries in a collision with a cyclists if the rider is a child under 14 years of age, even if the crash was the child’s fault.

If the cyclist is an adult, the motorist is liable for 100% of the damages/injuries unless she can show the incident was beyond her control, or the cyclist was at fault – that’s not easy however.

If the mistake leading to the incident was made by the non-motorised road user, that mistake has to be so unlikely, that a motor vehicle user could not reasonably have considered it to happen. Failing to give way or jumping a red light (deliberately or by mistake) are not such unlikely events, they happen regularly, so drivers are not granted ‘circumstances beyond control’ very often.

Even if she can show either or both of these conditions apply, the driver is still liable for 50% of the damages/injuries. Responsibility for the other half is determined by the Court according to the degree of fault of the two parties.

The exception is that a motorist has no liability if she can show the rider caused the damage on purpose, or his behaviour constituted “recklessness verging on intent”. That applies even if the cyclist is a child.

Thus for practical purposes, strict liability only applies to children under 14 years of age in the Netherlands. In the case of adults, the motorist isn’t automatically in the wrong. However the law recognises both the cyclist’s greater vulnerability to serious injury and the motorist’s greater capacity to pay by virtue of mandatory personal injury insurance.

Given that in Australia most cyclists are also motorists (car ownership is much lower in The Netherlands), this looks like a solution that would suit Australian circumstances particularly well. Interestingly, a commenter on Mr Wagenbuur’s article who’s from Australia says “in fact we have the same provisions in our insurance legislation, (it’s) just that it’s little known”.

I don’t know if that claim is correct or not, but if it is it doesn’t seem to be making any difference to the way many Australian drivers behave behind the wheel. That’s consistent with Mr Wagenbuur’s argument that the civil liability law doesn’t materially affect the propensity to cycle in the Netherlands because, presumably, it doesn’t signficantly alter the behaviour of drivers. Perhaps Australian motorists would take greater care if there were a comparable sharing of responsibility under the road rules, with offenders prosecuted by police.

The law isn’t the primary explanation for high cycling rates in the Netherlands:

It is extremely unlikely that people will cycle more because they know their damage will be financially compensated in case they are involved in an accident. In that respect it is telling that this law was implemented in the early 1990s: when cycling had been on the rise again for at least two decades. People will only cycle more when first and foremost they feel it is very unlikely that they will be involved in an accident at all.

Mr Wagenbuur finishes his article on a light note. He tells us that in 1997 the Dutch government sought to extend the strict liability that applies to child cyclists to adults. However the move failed.

A bill was initiated in which the age restriction would be scrapped and the ‘under 14’ regulations would extend to non-motorised road users from the age of 14. That would have led to a ‘strict liability’ as many perceive it is: the driver would always be liable. But this proposed law change did not make it. In 1999 the bill was withdrawn.

The second exhibit is a parody of fears the proposed change, if implemented, would’ve led to selfish and reckless behaviour by cyclists.

A parody of motorists' fears of what strict liability would lead to