May 2, 2017

Are Australia’s big cities dense enough?

The extensive suburbs in the middle rings of Australia's east coast capitals explain why the "missing middle" is so much less dense than in comparable Canadian cities

Alan Davies — Editor of The Urbanist

Alan Davies

Editor of The Urbanist

Percentage of population by density in persons per Hectare (source: Spencer et al)

In an interesting paper to the Australian Transport Research Forum, researchers from SGS Economics and Planning compared the densities of six cities – Brisbane, Melbourne, Sydney, London, Vancouver and Montreal (Urban or suburban? Examining the density of Australian cities in a global context).

Andrew Spencer, Jeremy Gill and Laura Schmahmann used small-area census data (SA 1 in the case of Australian cities) for 2011 to compare the populations and urbanised areas of the cities. They used a cut-off of 4 persons per Hectare to exclude very low density development like peri urban hobby farms.

Brisbane and Melbourne have by far the lowest (population-weighted) average density in this company, with 25 and 32 persons per Hectare respectively, while Sydney has 52 p/Ha. Greater London is in another league at 97 p/Ha, although with a population of 8.2 million in 2011 it’s significantly larger than the other cities. The Canadian cities offer an interesting comparison; Vancouver’s population of 2.3 million is similar to Brisbane’s but at 65 p/H, it’s more than twice as dense. Montreal has nearly as many residents as Melbourne, but at 71 p/H it’s over double the density.

As the first exhibit shows, most residents of Australian cities live at densities of less than 60 p/Ha. Indeed, three quarters of Brisbane’s population and more than half of Melbourne’s live at the lowest density of  just 4 – 30 p/Ha. They’re well behind Canadian cities at densities above 60 p/Ha and, notwithstanding all the current angst about high-rise, above 100 p/Ha.

The authors point to a key disadvantage of ultra low densities:

Coupled with their relatively large metropolitan areas, the distribution of density in Australian cities does not appear conducive to making a viable network of mass public transit infrastructure. 

That’s a legitimate if familiar criticism. They go on to make a number of other acute observations:

  • The construction of Australia as an urban nation – particularly in popular discourse – overshadows the fact that the majority of Australia’s urban areas feature relatively low densities.
  • Despite the claim to being one of the most urbanised countries, Australian cities are, for the most part, relatively low density.
  • Sydney, Melbourne and Brisbane might be characterised as having generally low densities with pockets of high and very high density development focused in particular areas.
  • International comparison cities appear to have been less reliant upon opportunistic high density development as they have more people living in mid-range densities than Australian cities.
  • Canadian cities are significantly denser. They also feature larger proportions of their populations living at above 60 people per hectare.

The authors question the virtue of Australian cities accommodating growth by relying heavily on a combination of city centre high-rise, high density transit-oriented development around rail stations, and fringe areas. They ask:

Should planning strategies seek to provide more housing at intermediate densities, over a larger area, rather than high densities in relatively few, concentrated areas?

The importance of the “missing middle” was explained yesterday by John Daley, Brendan Coates and Trent Wiltshire from the Grattan Institute; they indicate boosting supply in middle ring suburbs is the single most important action state governments can take to improve housing affordability while minimising negative collateral impacts (Options for housing affordability: the good, the bad and the cosmetic).

What’s needed in the middle ring suburbs is mostly 2 – 3 storey in-fill terraces/town houses that provide private outdoor space at ground level for families and are in scale with the existing suburban streetscape. Governments need to look at retro-fitting transport services (e.g. buses) to connect to the main network in many of these areas.

Extent of urban areas (excludes zones with less than four people per Hectare) compared to official/administrative metropolitan boundaries
(Visited 31 times, 2 visits today)


Leave a comment

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

6 thoughts on “Are Australia’s big cities dense enough?

  1. Anthony Y

    This study is of dubious value as the authors have cherry picked their data.
    Melbourne and Brisbane are plains cities with little or no geographical limits on urban sprawl. Both Montreal and Vancouver are on islands that create a geographical demarcation between each city and its hinterland, encouraging higher densities. Planning policies may have contributed to these densities but I suspect geography had a far greater role.
    Sydney’s status as Australia’s densest city bears this out, constrained as it is between the harbor and the mountains.
    I would be curious to see comparisons with Toronto, Edmonton and Calgary which are far more analogous to Australian cities.

    1. Tony Morton

      Hi Anthony. I think you might have a point regarding Montreal, but I’d give Vancouver more credit here. Vancouver’s urban density is most definitely the result of proactive planning policy. In 1990, Vancouver was actually a good deal less dense than Melbourne or Sydney and comparable to Perth; its densification is a good deal more recent than in a lot of other cities and has been particularly successful in its integration with strategic transport policy.
      There are actually a lot of cities that have gone in the reverse direction: they’ve got a lot of high-density development and quite high overall densities but they fail as urban environments because they went wrong with both transport and street-level activation. Suburban Las Vegas is a stand-out example: row after row of mid to high-rise apartments on top of car parks facing multi-lane arterial roads. Planners call it ‘dense sprawl’. It should remind us that there’s a lot more to successful cities than density.
      There’s a fair bit that’s wrong with the SOAC paper in my view, but the density comparisons pretty much stack up.

  2. Anthony

    I live in a duplex in Ermington (a middle ring suburb) in Sydney. My duplex is as large or larger than most houses and has a backyard for my kids. It is also modern with better features than most houses.
    My suburb, which is in Parramatta City Council, have had a lot of duplexes built over the last few years on large lots formerly occupied by stand alone houses, with more being currently built. There is clearly demand for these homes.
    However, in other suburbs (in different councils) there are relatively few (if any) duplexes being built. I understand different councils have different rules concerning duplexes.
    State Governments enabling duplexes to be built on appropriately sized lots with minimum of planning approvals would be a simple measure which would significantly increase the supply of homes. It would also boost small businesses as duplexes can be built by smaller operators.

  3. Jacob HSR

    SYD airport is surely dense enough or busy enough to justify the construction of an Automated People Mover – so that people can get from terminal 1 to terminal 2.

    What is the point of making NSW more dense if infrastructure is lagging behind even 3rd world cities.

    Vancouver has SkyTrain – look at the opposition to building one in Vic!

  4. Guest

    Have to question studies like this that don’t use weighted density.
    The density sounds like an objective measurement, but it is not because it changes depending on where you draw the boundary and how you define the city.

    From the map, it looks like bits of Ipswich and Caboolture are included. The low figure for Brisbane is likely due to them including Brisbane Airport, various hill areas that are parkland (Griffith University Nathan, other hills in the area) and a very large flood plain around Rocklea. Are these really valid to include in density calculations?

    1. Alan Davies

      The authors did use weighted density -see 3rd para of article. Also, the first exhibit shows density against population.

      They excluded any SA 1 with a density less than four persons per Hectare, so non-residential areas are excluded.

      They could’ve done a better job with London, though. There’s another 4 million residents beyond the green belt in the “overspill” towns.

Share this article with a friend

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

Your details

Your friend's details