Nov 9, 2015

Is this ranking of “the most liveable suburbs” believable?

Liveability is a handy idea but it’s hard to define it in a useful way. However the media love a list, so we invariably end up with flawed exercises like this one from Fairfax

Alan Davies — Editor of The Urbanist

Alan Davies

Editor of The Urbanist

[caption id="attachment_47743" align="aligncenter" width="552"] The 15 criteria used by The Age to identify and rank Melbourne's "most liveable suburbs" from 1 to 321 [/caption] It's been pointed out many times before that ranking urban areas is a problematic undertaking e.g. see Is being "the world's most liveable city" such a big deal? However Fairfax Media isn’t daunted, even when it comes to ranking very small areas. On Saturday, The Age published a table of “Melbourne’s most liveable suburbs” prepared with help from consultants. It ranks 321 suburbs against 15 criteria, mostly related to access to services (see exhibit). The gold medal “winner” is the near-CBD suburb of East Melbourne, followed at an unknown interval by South Yarra and Toorak. It makes intuitive sense that the most liveable suburbs would be the most sought after. So it’s no surprise the median price of a four bed room house in these suburbs ranges from $2.1 million to a whopping $4.25 million. But there are some curious results that call into question the usefulness of The Age's league table. For example, the suburb of Ivanhoe ranks eighth (median sale price of $1.4 million for a 4 bed house), yet contiguous Ivanhoe East ($1.5 million) comes in way behind at 92 and near neighbour Eaglemont ($1.6 million) at a distant 148. In fact Ivanhoe ranks much higher on The Age’s scale than all inner suburbs north of the river, including the likes of sought-after suburbs like Fitzroy (82) and Fitzroy North (38). Another strange result is outer suburban Caroline Springs; despite being a master-planned community with a median price of $0.52 million for a four bedroom house, it's ranked at 313, just eight places from rock bottom. And Princes Hill, The Age tells us, plunged forty six places from 62 in the last survey back in 2011 to 16 this time around "thanks chiefly to its NBN connection". So merely replacing cable and DSL with NBN increases liveability that much? Something peculiar seems to be going on here, so it’s worth looking at the detail of how the The Age derived its rankings. That presents a problem though. Although Fairfax frequently rails against public agencies for their lack of transparency, it doesn’t deign to explain in detail the methodology underlying the rankings and it chooses to deny its readers access to the consultant's report. We don't have information on vital aspects like how the criteria are weighted; does living close to the CBD get the same weight as living among hills? Nor do we know the interval between rankings; is twentieth place a whisker away from first place or is it a yawning gap? Even so, there are some evident problems with the exercise. A key one is the methodology appears to deal poorly with accessibility to shops, restaurants, culture, parks and schools; they only count if they’re located within the boundary of a suburb. For example, Ivanhoe scores well because it has a number of schools within its boundary; but nearby suburbs like Eaglemont that enjoy high accessibility to these schools by both car and public transport score low on this criterion. (1) That’s a serious failing. The sensible approach would be to rate each suburb according to the access it provides to key services; (say) something like the % of metropolitan jobs accessible within x minutes travelling time by private/public transport from each suburb. But that highlights a related and arguably even more serious failing; the criteria don’t include accessibility to employment despite its obvious importance in residential location choice. There’s a criterion measuring closeness to the CBD but it's not very useful for measuring job accessibility because only around 15% of all jobs in the metro area are in the CBD. Another inexplicable failing is the absence of any measure of the amount of private indoor and outdoor space offered by the housing stock in each suburb. Yet for most Melburnians private space is a key part of the liveability of a location; indeed it's one of the reasons Australian cities developed at low densities. The inescapable problem with these sorts of rankings is that it's necessary, in order to establish a rank order of suburbs, to assume a standard set of preferences defining an average or typical Melburnian. But clearly a city centre suburb like Melbourne (it ranked fifth), where 99% of the housing stock is small apartments, appeals to a vastly different idea of liveability than a suburb like Ivanhoe (ranked eighth) where only 12% of dwellings are apartments. Thus households with children comprise 23% of the population in Melbourne (suburb) compared to 64% in Ivanhoe. These two populations will have vastly different ideas of what constitutes liveability. It didn't stop it, but The Age seems to recognise that a league table of suburbs is a dubious exercise. For those who're prepared to click through, it also offers a "calculator" that enables readers to find their "perfect suburb" by weighting each of the offered criteria in accordance with their personal preferences. Readers can also set a price limit. Although there are still serious deficiencies in the criteria it offers, it's at least recognition that liveability isn't defined in the same way by everybody; it depends on the preferences of individuals (as mediated by prices). (2) If The Age is intent on a universal league table, it should start with the simplest metric of all; prices. Selling prices/rents provide a summary statistic of the complex forces that shape why one suburb is preferred over another. A more useful exercise would've been to work backwards from property prices and figure out the key components explaining why prices differ between suburbs. _______________
  1. Ivanhoe improved from 32 last time to 8 this time,  mainly due, The Age says, to schools. The Age notes Charles La Trobe college and Ivanhoe Grammar School are located within the boundary of Ivanhoe (although it doesn't realise the latter is co-ed) but is unaware (unrelated) Ivanhoe Girls Grammar School is also located within the suburb; if it did, would Ivanhoe score so well?
  2. There are some other oddities too. For example, there are three separate criteria relating to the importance of living near a train stop, a tram stop and a bus stop, respectively; do these "double-count" the other seven criteria on the importance of living close to specific services and facilities? Are suburbs that don't have a tram service penalised, even though they might have high accessibility by other modes?

Free Trial

Proudly annoying those in power since 2000.

Sign up for a FREE 21-day trial to keep reading and get the best of Crikey straight to your inbox

By starting a free trial, you agree to accept Crikey’s terms and conditions


Leave a comment

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

14 thoughts on “Is this ranking of “the most liveable suburbs” believable?

  1. Alan Davies

    Update 12/11/15: This is about as close as Fairfax is likely to get to acknowledging it got its league table wrong, Melbourne’s best suburb? Predictably nobody can agree.

  2. JMNO

    For the last couple of weeks, The Age has been bombarding us with housing information – prices, auction results, liveability and all promoting their real estate service, Domain. I have been wondering what their real agenda is, lack of reporters to cover real news, or getting more punters to use their real estate service

  3. RTAkers

    Looking at the Frankston area rankings, based on the higher price means more liveable assumption, the results seem backwards. Frankston South prices would be far higher than Frankston North (an area known for it’s high crime rate), yet Frankston North comes in well above South in the list. The note about Princess Hill’s jump due to the NBN may also play true here too, as would accessibility to the train line.

  4. Tom the first and best

    Hills have far better scenery. Hill suburbs also tens to be greener and leafier (which is a cause of some of said scenery).

  5. Tony Kent

    I find the report’s author’s explanation for the choice of criteria and their metrics unconvincing, for example; why is access & proximity to quality hospital & medical facilities not in the top 15? And what is the authors’ definition of a “park”? Do 2-5 ha municipal gardens equal a 1000ha riverside park network? No didn’t think so.

    This exercise appears to be more to do with selling ads, generating Fairfax click bait and positioning Deloitte as having some previously unknown expertise in urban planning.

  6. Oz (Horst) Kayak

    The Fairfax funded 15 criteria comparison between 321 ABS gazetted MSD suburbs is a commendable start to understanding urban environments better.
    A group of us have allocated health indexes to the same suburbs, and predictably the ranking is very different for some suburbs.
    Another interesting comparison has been to run the Walk Scores used by many in the real estate industry and freely available on the web. The Walk Scores have a strong correlation with Fairfax ratings. Pity no one so far talks about what walk scores can indicate about healthy urban places to live.
    Another extension of Adam Terrill modelling would be apply the 20 minute criteria as currently proposed in Plan Melbourne 2016 for the same 321 suburbs.

  7. Hamish Moffatt

    @lomlate, flat areas are just a bit dull, and tend to be more windy. I’m a cyclist and I like hills, though not always on the ride to work.

  8. lomlate

    Why do people like hills? As a cyclist I hate hills. Can someone explain who exactly would like them?

  9. Dudley Horscroft

    Would most really agree that access to the NBN is really important? I would suggest that in most city suburbs access to the internet is reasonable, as is access to mobile communications via mobile towers. Is there any indication that internet access varies over Melbourne, other than access to the NBN? For shopkeepers and other businesses, high speed access may be desirable, but surely those who insist on being able to download a feature length film in 10 s flat are few and far between. I think most people would be just happy to be able to receive and send emails and download items of interest as and when wanted – the biggest problems seems to be that (a) the original site server is slow to respond (NBN is irrelevant) and (b) that the site is paywalled at a ridiculous figure!

  10. Bennopia

    Have to agree that this survey misses the mark significantly:

    * e.g. Kingsville “Its lower overall ranking compared to 2011 can in part be attributed to the increasingly congested roads.” The congestion of the local streets of Kingsville is much better than Yarraville which gets so congested the local community action group is calling for traffic solutions such as on ramps to the Westgate Bridge. Trucks are banned most hours from Sommerville Road and diverted to Francis Street. There is an art gallery (Quinch). I backtrack to Millers Rd to get to the freeway as Williamtownstown Rd Yarraville M1 on ramps take longer. I use west footscray station as it’s better than yarraville (parking, less road congestion, etc).

    * e.g. West Footscray has two primary schools, two train stations, village cafes, restaurants & supermarket, bunnings, yarraville swimming pool, melbourne museum of printing, community gardens, etc, yet the article states “Unlike Footscray, it does not feature the same proximity to cafes, shopping and schools.” Was south of the train line included in Yarraville by accident? Many of the Footscray / Kingsville / Yarraville sports associations (e.g. at Hansen Reserve, West Footscray) are mixed throughout Yarraville and southern West Footscray for historical reasons so this wouldn’t surprise me.

    End result it’s not open to peer review so reads like a subjective marketing blurb that promotes some “favoured suburbs” above others despite the locals knowing that there’s little which separates them.

  11. Daniel Bowen

    Comment #3 – Sorry, I meant to say the infrequent windy bus is less useful than the frequent direct bus for most people.

  12. Daniel Bowen

    Proximity to public transport stops/stations isn’t a great measure of public transport usability. Suburbs along the Stony Point line get only a handful of services each day, whereas some junction stations get trains every few minutes. Likewise, bus services vary widely in service quality (trams less so).

    Counting stops means an infrequent bus that winds its way through a suburb with lots of stops will rate more highly than a frequent main road direct bus, despite being much more useful to people.

  13. Alan Davies

    Adam Terrill #1:

    Appreciate your response Adam but a statement like “the survey does consider things outside the suburb boundary” isn’t very specific.

    The commentary provided by The Age on each suburb suggests that the focus is on what’s within the boundaries of each suburb e.g. in relation to East Ivanhoe, the paper says “where it falls down is in…(having) relatively few cafes and restaurants”.

    I didn’t question the study in relation to closeness to train and tram stops but, in the continuing absence of the report, I would like to know if each suburb is individually rated on its accessibility to all shops, restaurants, culture, parks and schools in the metro or at least regional area? If so, how was that calculated and how was it entered into the overall ranking?

    Re jobs, access to employment is a measure of a suburb’s liveabilty. Why include accessibility to schools (how ever it was done) but not work?

    Always, always better to be transparent and release the report the media story is based on. Given it was used for a public story, the decision to keep it under wraps is bizarre.

  14. Adam Terrill

    Hi Alan,

    Thanks for your comments.

    The survey certainly does consider things outside the suburb boundary, like stations, schools, trams etc. Where the data wasn’t available at point source, like culture, we smoothed results based on cascading proximity. So proximity to trains, for example, considers your closest station, whether that be in your suburb, the next suburb, or many suburbs away.

    As for Princess Hill, and ‘so merely replacing cable and DSL with NBN increases liveability that much?’. This is the first time telecommunications was measured, and it includes both mobile and broadband. It’s a new indicator, and wasn’t measured before. In 2015 I think most would agree that telecommunications it critical to modern lifestyles. As its new, its not surprising that suburbs with excellent telecommunications shot up the list. There were other issues, like shifting ABS suburb boundaries, that caused changes too.

    Re jobs, we made a conscious decision not to include access to employment as this is a measure of a place to live, not work. A index that sought to measure a suburbs value for work would look very different. We did toy with the idea of allowing people to enter into the calculator their suburb of employment.

    Re the subjective nature of liveability, we agree completely, and you’ll note that it’s a strong theme that runs through the entire series, being mentioned many times. Hence also, the new calculator. But is that a reason not to do it ? No, we say that debate is a good thing, especially when it moves beyond the small world of planning and architecture and into the general populous. If people, who wouldn’t normally, reflected on what makes their suburb good, or what things they would like to see in their suburbs, that is a positive thing, especially if that leads to action from Government and business to reduce the difference across Melbourne.

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