tip off

The “20 minute neighbourhood”: does it make sense?

The idea that every home should have local services and facilities within a 20 minute walk, cycle or public transport trip seems a no-brainer. But is it a worthwhile and realistic goal, or a dream?

The 20 minute neighbourhood

The draft metropolitan planning strategy released earlier this year for Melbourne proposes a city made up of 20 minute neighbourhoods. As the exhibit shows, the idea is that every home will be within 20 minutes travel time of jobs, shops, cafes, schools, parks and community facilities.

Of course every home already is if you’re driving, but the important innovation here is Plan Melbourne proposes it will be 20 minutes travel by active modes i.e. by public transport, walking and cycling (1).

It’s certainly an appealing idea. It promises to draw suburban Melburnians out of their cars en masse. That would provide environmental benefits, better health outcomes, improved local amenity, stronger communities and reduced traffic congestion.

Who wouldn’t want to have local services within an easy walk? And residents would still retain the option of driving! It seems there are no losers here, only winners.

Is it too good to be true? There are a number of issues I’d like to see explored first before I’d be convinced this is the clear and obvious winner Plan Melbourne would have me believe it is.

Google Maps tells me I can cycle to my nearest supermarket in 7 minutes, which is reasonably competitive with driving (4 minutes). It says it’ll take me 20 minutes to walk. It won’t give me an answer for public transport but I can see it’d take longer.

Which mode is the limiting bound? If its cycling then we’ve pretty much got 20 minute neighbourhoods already. A cyclist can travel 5 km in 20 minutes, which yields a radius of almost 80 sq km.

If it’s public transport then in many suburbs 20 minutes will be far too restrictive. The attractiveness of public transport for short trips is limited because walk and wait times make up a relatively large part of the overall travel time (2). In addition, public transport evolved to serve the CBD; it isn’t set-up as a series of small-scale ‘hub and spoke’ systems serving each neighbourhood centre or local facility, and nor should it be.

While cycling will be an attractive option for some people, walking must be the default active mode because it’s the only one that will suit the majority of travellers (3). But after allowing for differences in factors like age and topography, walking implies all the nominated services and facilities would need to be located within 1 – 1.5 km of each home i.e. a radius of 3 – 7 sq km.

A key issue with a relatively dense service delivery grid is the trade-off between economies of scale and travel time. For example, it would undoubtedly be nice to have a State secondary school within 20 minutes active travel of all homes (as Plan Melbourne specifically envisages), but some economies of scale would necessarily be foregone. Big high schools offer more courses, more extra-curricular activities, and more specialist staff than smaller schools (4).

Even assuming services and facilities could be retro-fitted in the suburbs to provide the full-service neighbourhood centres envisaged in Plan Melbourne, the optimal spacing is not the same for all services and facilities. They don’t all fit neatly within a 20 minute access time. Users are also prepared to travel further for some than others; indeed some actively prefer the more distant option e.g. private schools.

Nor can all activities be co-located easily; for example, many existing primary schools are not part of neighbourhood centres. The task of giving each dwelling access to at least two (and very possibly more) geographically separate local destinations within the 20 minute time frame would be extraordinarily hard.

While State and local governments might see political advantage in increasing the density of public services, private providers won’t show any interest unless there’s a clear commercial imperative (5).

Only a very limited range of businesses can flourish with a small, defined catchment. Most seek to draw customers from as large an area as possible. Many businesses see an advantage in co-locating in a limited number of locations.

It might well be feasible to establish something  that resembles a network of neighbourhood centres every 2-2.5 km in both directions. But the liklihood is a large proportion will be very small and offer only a limited and lop-sided range of services. They won’t be attractive to the locals.

Even in locations where local services are good, some residents will nevertheless want more choice. They’ll want to travel to larger centres – like Preston Market (say) – that offer a bigger range and/or a more exciting experience than they can get with a 20 minute walk.

The idea that residents could find, or would seek out, a job within a 20 minute walk of home misunderstands the nature of cities. A key reason they’re attractive places to live is they offer opportunities for workers to specialise and firms to agglomerate. That’s a key reason why the median journey to work in Melbourne is around 15 kilometres for both car and public transport commuters.

Given that 90% of households have at least one car (and over half have two or more), Melburnians have ample scope to travel beyond their immediate neighbourhood to access better work and non-work opportunities.

I’ve no doubt the idea of 20 minute neighbourhoods will appeal reflexively to many planners (see Are big cities just collections of country towns?). But it seems to me it’s more about good politics than good policy.

Before I’ll be persuaded, I want to see evidence that simply reducing walking times in the suburbs will lead, in the absence of severe constraints on car use, to a large mode shift away from cars for neighbourhood trips. I want to understand what the costs would be (they’re not stated in Plan Melbourne, naturally) and how they’d compare to the benefits.


  1. Plan Melbourne doesn’t explicitly state that driving is excluded from the concept of the 20 minute neighbourhood (and the earlier Discussion paper didn’t exclude driving), but it’s certainly couched in a way that implies it’s solely about active modes of travel
  2. Public transport would make even less sense for local 20 minute trips if the high frequency “turn up and go” grid-based system that’s commonly advocated (like this) is assumed. A 10 minute frequency means an average 5 minute wait to board and, if required, to transfer.
  3. Cycling would offer travel times closest to those of driving, but there’s nothing in Plan Melbourne to suggest Melbourne’s suburbs will look like Amsterdam any time soon.Walking isn’t as straightforward as it might at first appear. Google assumes an average speed, but typical walking speeds vary depending on age, sex, and environmental factors. This US study found pedestrians aged under 65 years cross the road at an average speed of 5.5 kph, whereas those 65 years and over average 4.5 kph. Women walk more slowly on average than men. If these speeds were maintained continuously, the first group would be able to walk up to 1,800 metres in 20 minutes and the latter 1,500 metres (although only 1,350 metres for older women). These estimates make no allowance however for constraints like topography (e.g. hills), traffic (e.g. signalized intersections), or weather (e.g. very hot days). In reality, it’s likely the nominated services and facilities would need to be spaced every 2-2.5 km in both directions.
  4. Suburbs aren’t uniform in their demand for State secondary schooling either e.g. there’s less interest in public schools in higher income areas where many students travel long distances to go to private schools.
  5. Plan Melbourne acknowledges that “In some areas…new village-style cafe and shopping areas may need to be created…”, but merely providing the appropriate planning permissions doesn’t mean the sought-after activity will be commercially viable.

Please login below to comment, OR simply register here :

  • 1
    Ian Harvey
    Posted December 11, 2013 at 10:18 am | Permalink

    Perhaps “radius of” => “circular area of”, or similar.

    How does the over 65 female go speed wise on a bike?

  • 2
    Posted December 11, 2013 at 10:43 am | Permalink

    At the local level the planners are talking about, even a straight grid of bus routes would enable most people to access their local shops without a change of service. That is easily manageable in 20 minutes (average case) with a 10 minute service frequency. Indeed I can do that already, because I live in tram-land. Given many Melbourne suburbs are already more dense than Balwyn, why should we not offer this option more generally?

  • 3
    Posted December 11, 2013 at 10:51 am | Permalink

    There are many advocacy groups who support the concept that health and well-being are the ultimate community outcomes to be planned for in integrated land-use and transport systems. The level of service (LoS) concept used to be taught in 101 for those being prepared for the professions associated with engineering type courses.
    Time allocation to regular out-of-home routine activity needs to be part of the LoS provision outcome in community planning.
    Of course it is impractical if not impossible to satisfy all the community LoS expectations all the time. No matter what the available time-slot is.
    However the vision for meeting many community expectations much of the time can be a quantifiable objective for active transport policy support.
    Would meeting 60% of community needs within 60% of a 20 minute travel time-period using active transport be too low a policy and planning target to facilitate more general health benefits to people in Metropolitan Melbourne?

  • 4
    Alan Davies
    Posted December 11, 2013 at 12:43 pm | Permalink

    Strewth #2:

    When looking at public transport need to factor in average walk time too. Also, the grid of routes should prioritise larger centres; some neighbourhood centres will be off-grid. Note also that Plan Melbourne envisages 20 minute access to more than neighbourhood centres e.g. primary and high schools.

    The reason we should think hard about offering the 20 minute neighbourhood option more generally is that active transport will struggle to be competitive in a world where the option of driving remains largely unfettered. Wouldn’t matter if the required changes were cheap and easy, but that’s hardly likely to be the case.

  • 5
    Posted December 11, 2013 at 2:11 pm | Permalink

    Not specifically related but Alan have you read or heard anything about “Happy City: Transforming Our Lives Through Urban Design”? Wondering if it’s worth buying/reading.

  • 6
    Alan Davies
    Posted December 11, 2013 at 2:20 pm | Permalink

    Dylan Nicholson #5:

    Only heard about it at the start of this week and tried to find an epub version. Here’s a reference to it I saw this morning. I don’t have high expectations: an accurate take on the subject might not make for a saleable book.

  • 7
    Bill Parker
    Posted December 12, 2013 at 12:35 pm | Permalink

    This is not the way MegaPerth is being built.

    The steps are as follows.

    1. Scrape off the remnant bush after felling any trees and burning them on site.
    2. Spray the cleared land with recycled paper,green dye and grass seeds.
    3. Install the services
    4. Build the roadways.
    5. Start building project homes, concurrently build service station(s), a primary school ( if needed), and a fast food outlet.
    6. Open some Display Homes (usually furnished with reduced size furniture)
    7. Sell from the weekend papers and TV. usually mentioning “only” X minutes from Perth. BUT… just 5 minutes from the beach.
    8. Move on north and south, repeating from step 1.

    This city is nudging towards 120 Kms end to end and I foresee the “connection” between Geraldton and Bunbury. Or, to put it another way, acres of dismal dormitory suburbs built with no regard to energy consumption or use of solar heat and light.

  • 8
    Simon Mansfield
    Posted December 12, 2013 at 9:18 pm | Permalink

    The World Inside by Robert Silverberg (1971). Good holiday reading for anyone needing a vision of their happy happy future brought to you by well meaning town planners.

  • 9
    Aidan Stanger
    Posted December 13, 2013 at 10:03 am | Permalink

    Alan Davies #4
    Can you give an example of a neighbourhood centre that would be off grid?

  • 10
    Alan Davies
    Posted December 13, 2013 at 12:55 pm | Permalink

    Aidan Stanger #9:

    (1) Draw a grid at 2 km centres on your Melways (500 metre av walk times) and see what happens; it’s a geometric impossibility to pick most centres up. (2) You can twist & turn the routes (like bus companies do) to pick up more centres but the limiting factor is the increase in bus travel times that comes with circuitous routes. (3) You can instead have a much denser grid (e.g. at 1 km centres, av 250 metre walk times) but the limiting factor is lower load factors and the cost of providing high frequencies. (4) Most neighbourhood centres are very small and even if they’re on the grid they won’t have the number and range of retailers to be attractive to residents who have the option of driving to a bigger centre.

    You can’t have the “20 minute city” as well as near-universal car access. In those locations where residents really do “walk to the shops”, the competitiveness of cars has been reduced, usually by parking limits and/or congestion e.g. areas very close to the CBD.

  • 11
    Aidan Stanger
    Posted December 13, 2013 at 2:13 pm | Permalink

    Alan Davies #10
    There’s no reason for the spacing to be the same on each axis. How about 1km (or less*) on the dominant axis and 2 to 3km on the perpendicular axis?

    * 250m is the average walk to the route rather than the stop, so even 1km spacing is inconvenient for some.

  • 12
    Alan Davies
    Posted December 13, 2013 at 2:43 pm | Permalink

    Aidan Stanger #11:

    Have to have much the same spacing in each direction because we’re talking here about what are essentially radial movements from home to the nearest neighbourhood centre (and that’s ignoring the complication of 20 minutes active travel to frequently out-of-centre destinations like primary schools).

    With a 1 km grid and stops every 500 metres (say), I’m guessing average walk to a stop would be around 350 metres, but one or two households would have a max walk of 700 metres. If you have an absolute 400 metre max walk rule (commonly advocated), then the grid needs to be even denser.

  • 13
    Peter McArdle
    Posted December 16, 2013 at 5:36 pm | Permalink

    The idea of a 20 minute neighbourhood is not new. It was the basic plan for Canberra developed by the National Capital Development Commission (NCDC) in the 1970s and early 1980s. The plan was (is)for Canberra to be made up of satellite cities of about 100,000 folks (Belconnen, Tuggeranong, Inner Canberra etc.) Within these satellite cities, all civic services are available, including employment. It basically works and has made Canberra a very liveable city (OK stop laughing, this is serious).

    The major problem is that some more recent town planners (encouraged by property developers) have sought to make Canberra “a real city”, “like other cities”. For example they have renamed the oldest commercial area, Civic, as “The City Centre” with an aim of making it a “real city centre”, you know, with traffic congestion and long commutes.

    If you are interested in the concept of the 20 minute neighbourhood, have a look at how Canberra was planned before the more recent crop of town planners started doing their thing.

    & another (very minor) thing: an earlier post asked “how does the over 65 female go speed wise on a bike”. I am near that age group and I am often trying to keep up with lady cyclists of that generation.

Please login below to comment, OR simply register here :

Womens Agenda


Smart Company




Property Observer