Mode share (%) of journeys to work by cycling 2011 Census (Source: Charting Transport)

The Guardian’s new Australian offshoot ran a story last week In search of Australia’s best cycling city. The writer came to the conclusion that “Melbourne and Adelaide are vying to be Australia’s top cycling city, with Sydney and Darwin making late surges for the title”.

This article doesn’t really make a very convincing case. It relies on some funny numbers and some thin arguments. I’m certainly not persuaded that having a Cycling Film Festival (Sydney) or a major international cycling tour (Adelaide) makes a city a leading place for cycling. Nor do exciting plans and promises for infrastructure (Melbourne) or the number of active bike shops (Darwin) provide a reliable way to assess if a city is the best place in the country to cycle.

As always, these sorts of exercises come up against definitional problems: what does “best” mean? What’s best for one type of cyclist might be very different for another. Is it best for commuters, for shoppers, or for tourists? Someone who wears full lycra is likely to have a very different view of what makes a city good for cycling to that of a middle aged tourist who only rides occasionally.

I think any problem like this has to start with looking at travellers’ “revealed preferences”: what is the actual level of cycling in each city? That will tell us what travellers actually do, not what we expect they’d do.

The only reliable and consistent basis for comparing cities is the journey to work question asked at each Census. It has its limitations (e.g. commutes only account for around a fifth of all trips; the weather on the day affects numbers) but it’s the best we’ve got. Fortunately, Chris Loader at Charting Transport has crunched the numbers for most cities, but not Darwin (F: 1).

The 2011 Census shows Canberra is the stand-out city with 2.8% of workers commuting by bicycle (see first exhibit). It was followed at some distance by Melbourne. Sydney is a distant last in this company.

Of course Canberra is much smaller than the other cities in geographical extent. It also has a much lower population: 0.36 million in 2011 compared, for example, to Sydney’s 4.6 million and Melbourne’s 4.1 million.

When the relatively small populations living in the inner parts of the larger cities are looked at separately, their performance compares more than favourably with the national capital. For example, cycling’s mode share on Census day 2011 was higher than 4% in many parts of Sydney’s inner west (see second exhibit). Its share exceeded 10% in much of Melbourne’s inner northern suburbs and was a key mode in parts of inner city Brisbane (e.g. West End) and parts of Perth (e.g. Fremantle).

There’s something about these inner city suburbs that makes them much more attractive for cycling than middle and outer ring suburbs, or even other inner city suburbs. Note the rapid increase in cycling use between 2006 and 2011 illustrated in the exhibits.

Melbourne has the highest rate of cycling, but the key point of interest in my view isn’t differences between cities, but why the inner city (or parts thereof) differs so much from the rest of the metropolitan area.

A big part of the explanation is undoubtedly better conditions for cycling: inner city areas tend to have superior cycling infrastructure – more marked lanes and more dedicated paths. It probably helps that traffic moves more slowly on average due to more congested conditions.

But it’s not just a matter of engineering. Another important reason is the characteristics of the population; in particular, many workers living in these areas work elsewhere in the inner city or, especially, in the CBD. Cycling provides them with a faster, more reliable, more flexible and cheaper way of commuting than the most likely alternative (public transport) and incorporates exercise to boot.

Another factor is those who’re younger, better educated and have no dependents are over-represented in these areas. Looking at the geography I expect political attitudes matter a lot too and probably explains why cycling infrastructure is better in inner areas.

The differences in the characteristics of the population are important. Even if superb cycling infrastructure were provided in the outer suburbs, it’s doubtful if it would attract anywhere near as many commuters as the inner city currently attracts even with mediocre infrastructure.

That’s largely because the proportion of residents who work in the city centre is lower in the suburbs. That encourages commuting by car: even if it’s very good, it’s very hard for any other mode to compete against cars for the journey to work when residences and (especially) jobs are largely dispersed, as is the case in the suburbs of most Australian cities. Workers who have the option of driving are much less likely to cycle (or use public transport).

It’s hard to talk meaningfully about the “best” city or place for cycling without access to objective data on actual riding levels. It’s hard, in turn, to attribute significance to observed cycling levels without accounting for the characteristics of the population. No wonder The Guardian had trouble picking a winner.

So the exercise of determining which city is “best” is a fizzer. I think it’s possible though to have a sensible discussion about what characteristics would make any city or place attractive for cycling (as distinct from the “best”). That’s a separate and interesting discussion I’ll come back to later.


(F: 1) Mike Rubbo at Sit-up Cycle reckons 4.2% of Darwin commuters cycle. I’d like to see what geographical area he’s used and if it’s comparable with the ABS definition for other capitals. It’d be phenomenol if it’s the entire metropolitan area, albeit Darwin’s population is 129,000.

Cycling's mode share (%) at 2011 Census. Click to enlarge (Source: Charting Transport)