Major cities in many countries have become progressively less dense

A reader yesterday linked to the recently-published United Nations World Cities ReportSceptical as I am of most UN things, out of curiosity I dipped into a few chapters of the report.

On doing so, I stumbled on this chart


In a quite striking way it makes the same point made in an early post on this blog: as countries become richer, the cities in those countries tend to become less densely populated.  Here was the chart from the earlier post showing data for London as far back as 1680 (just over a decade after the Great Fire).


These numbers shouldn’t really be a surprise.  Space is a normal good –  people typically want more of it, all else equal, when they can afford it –  and technological advances make longer distance commutes feasible.

No doubt there will be some issues with how the data are compiled/estimated –  quite where are the boundaries around the “built up area”, and how well is that known for, say, 1855.  But the general proposition shouldn’t be surprising: it is easy enough to think of the cramped tenement dwellings of New York in the late 19th century.

Of course, these trends aren’t ones that seem to please either the United Nations or many of our own local councils.  This text is from the UN report, just after the chart above

In recent years, UN-Habitat has brought into the forefront of attention the need for orderly expansion and densification so as to achieve more compact, integrated and connected cities. UN-Habitat’s support for planned city extensions programmes as well as promotion of tools such as land readjustment aims to increase densities (both residential and economic) with compact communities in addition to guiding new redevelopment to areas better suited for urbanization. These interventions are suggested to be an integral part of the New Urban Agenda as elaborated in Chapter 10.

And the constant refrain locally is for “more density”, when there is little or no evidence that such densification is what residents would prefer for themselves.  Indeed, it would be surprising if the revealed preferences across time and across countries/cultures had suddenly reversed.

I have no particular problem if people wish to live in high-rise apartments, or in small townhouses with no garden.  And people will choose to do so if regulatory constraints limit their options –  eg if land simply becomes too expensive –  but it doesn’t look like a first-best unconstrained preferred choice for most people.  We don’t, for example, see such bunching in our own provincial cities –  where housing is less unaffordable than in, say, Auckland –  and of course by international standards even our own largest city, Auckland, is not much more than a large provincial city (just a bit smaller than, say, Nashville).

Freeing up the use of land around cities remains the key to making housing affordable again and providing the choices/options that people value.  Experience suggests more populous cities will cover more space.  That isn’t something officials and politicians should be trying to stop.

28 thoughts on “Major cities in many countries have become progressively less dense

  1. Michael what shouldn’t be forgotten is that even though the population density per hectare in cities has generally declined since a peak about hundred years ago. The built up density -the floor space of cities has continued to increase. Manhatten’s skyline shows this effect -where small -incredibly dense tenement buildings were replaced by larger but much less densely occupied buildings.

    The city density story is actually quite nuanced. Check these links out. Manhatten for instance has had a moderate uptick in density since 1980 and of course go back far enough in history -Manhatten’s density increased in the 19th century to peak in 1910.

    Also density decline in land rich developed cities -such as Sydney -so probably the most applicable example for NZ cities has had the most moderate density decline. The city population density decline story probably applies much less to NZ than other places -especially as you say housing space is a normal good that responds to increases in income. NZ’s per capita income had its big gains many decades ago and is now increasing slowly so we can expect the demand for more space per person to also increase slowly.

    Michael I do not disagree with your general point about freeing up land restrictions on the periphery of our cities as a measure to address the housing crisis -it is a point I have also frequently made. But the evidence also points to a continued demand to build more floor space inside our cities too. So restrictions on building in these areas also need to be addressed. Cities build up and out and when this is artificially constrained there are consequences, such as housing becoming less affordable and homelessness increasing.

    I think Michael you could also address the city transport and public right of ways issues as well. Which my post on your previous article on bank stress tests raised.


    • Thanks Brendon. I was flicking thru the UN report you linked to when i found the density chart (i still haven’t found the material you referred to – can you tell me the chapter or page?).

      As discussed previously, I don’t have any problem with denser suburbs or inner cities if that is what people really prefer (and am happy to see regulatory obstacles removed, subject to the caveat that regulatory obstacles often parallel the private covenants often seen now in private developments). Like you, I’m interested in ways of allowing groups of people to voluntarily contract out of those restrictions.

      And yes, floor space per hectare is different from people per hectare. Even in suburbs we see the same thing – i’ve now lived in the same suburb spanning almost 40 years, in three different houses each of which is bigger now than it was when i first lived in the house.


  2. A friend of mine just built a 4 bedroom house in Pukekohe. His is the first house completed by Signature homes nestled on land subdivided to 20 lots. Pukekohe is 50 to 60 kilometres from Auckland central. He paid $300k for 600sqm. The house 180sqm cost him $300k plus variations of $100k. First and last time I will visit him. The drive was incredibly tiring as there was no street lights for at least 10km of road. The area is basically farmland. The developer after 2 years have only built and sold a couple of houses. The rest are bare sections waiting for someone silly enough to go out there.

    There is so much bare land out here but frankly no one wants to live out here. My friend drives out to New Lynn to get his weekly shopping done. He wakes up at 7am to get to work in Manukau on time at 8.15 am. This so called more land approach is just not working. Central Auckland is Mt Eden, Mt Albert, One tree hill, Mt Roskill must be rezoned highrise. Preventing central Auckland from going 50 to 60 level buildings is costly.

    Taller buildings have a natural resonance that protects against earthquake. Shorter buildings require rubber dampeners which are costly and rubber deteriorates over time which requires expensive replacements.


    • Hmmm… data point of one… and are you seriously suggesting that this same person drives from Pukekohe to New Lynn to do the groceries?? I’m sorry but the last time I was in Pukekohe before Christmas there were plenty of supermarkets and other large format stores to be frequented…

      And the complaint about the drive… obviously not your cup of tea to live there… and obviously not much of a friend if the drive is so tiring for you…


      • Nothing in Pukekohe that looks like a decent mall. Weekend shopping is about more than just clothing, also the ambiance of a larger shopping format. Sure if you like 50km drives but when you are blind in one eye not so easy mate.


      • Just 6 months ago I had 20/20 eyesight but suddenly I have lost vision in one eye. No reasons have been ascertain after a series of medical tests and physical examinations. Doctor says it is a common occrance this blindness issue amongst 50 plus year olds with no apparent medical reason.


  3. It would be interesting to see what correlation (if any) there is with infrastructure spending on a per capita basis (particularly public transport).

    Also of interest is the effect of the city limits expanding over time


  4. Your comments reflect what I have long thought about NZ urbanisation. There is a smallish group of residents that don’t mind higher density living and may prefer it – around 30% over time… but the rest (70%) want detached housing and maybe a section… It is no good to demand that people cram into high rise apartments if that is something they don’t want… why should they?

    The cities would be much better served by allowing natural expansion and less central planning by well meaning but dumb planners…


    • As the population ages health issues arise. Not so easy to have to travel 50 km to get to a decent hospital for your regular checkup. My brother is a fit 53 year old that still climbs rock faces as a hobby. But even he was in the last couple of weeks was on crutches. Not an easy task driving when you have a swollen foot and can’t press down on the brakes. Unbelievable how a fit and strong individual just completely useless with one big swollen foot. Can’t walk, can’t drive. How easily a person becomes unable to do much. Can you imagine being 50 km away from the nearest hospital? Just last month he was thinking of retiring to the Grampians in Australia. Now he is realising he needs to be a lot closer to health services and other amenities and much closer to friends and relatives within city central. The reality of getting old just struck him.


  5. I agree with GGS that the quality of life in outer suburbs of big cities is poor. In my experience, people under about 50 prefer the inner city and people above that age prefer the inner city. But that rather arbitrary number is moving up as people no longer want to move out as they age where I live (Sydney).

    Interesting discussion on density and cars, with some nice photos, here:


  6. I posted the below over on Kiwiblog, will repost here:

    Average density is a bad way of presenting data like this. Low density exurbs can hugely skew results while only accounting for a small proportion of the population.

    If you are interested in revealed preference you need to get an average of the density in which each person lives in. This is called weighted average density.

    Here is a graph that shows weighted average density vs population size showing increasing density for larger cities.

    I haven’t seen a time series of weighted average density, possibly because the data doesn’t exist or would be very time consuming to process.


  7. Thanks for the comment. It is a fair point, but as you note this material doesn’t deal directly with my point – that over time cities have tended to become less dense (on average) as populations have become wealthier and transport technology has advanced. There is no argument that larger cities tend to be denser than smaller cities or towns (altho as your chart shows the relationship isn’t overly strong).

    More generally, the exurbs point is probably less relevant in places like NZ and Australian cities than it will be for places like eg New York with very high density at the core and very low density in many outer suburbs.


  8. It would be interesting to show a series of photos of London from 1680 to the present day to match your population density bar graph Michael. Would the revealed preference, of actions on the ground be that building density had declined? Obviously not. I think the evidence shows that when a city’s built environment grows -it grows both up and out.

    What this post accidentally shows is a division in society on its preference for how cities should grow. A “Happy Days” versus “Friends” built environment preference. Some prefer suburban stand alone housing growth while others prefer closer to the action -apartments, row houses, townhouses… growth. This seems to frequently match personal preference.

    Maybe we shouldn’t generalise our personal preference onto everyone else? If we support people to be free to act on their individual housing preferences without pre-judging them we might not be in the trouble we are.


  9. Interesting suggestion Brendon, and I’m not sure the answer is a clearcut as you seem to suppose. After all, streets would probably have been a lot narrower in those days. And even today London is a fairly low rise city. But, yes, technology has also enabled higher rise buildings than were once feasible/economic

    I don’t have a preference for how cities should grow. I don’t think it is one of those things public policy should try to control one way or the other. As far as possible, and consistent with the idea that landowners should be able to act collectively in deciding how their land might or might not be developed, I think policy should faciltate the up or out options. My prediction is that if it were to do so, we would not see an increase in population density within our cities (although of course as a larger share of the population has moved into cities that average density of all the places people live has probably increased – as a peasant one needed a farm/small holding, whereas suburbanites mostly only want a garden/backyard.


    • London as a result of its rationed land supply since 1947, has relentless unnaturally high density everywhere, as opposed to cities where the density gradually reduces to quite low levels out towards the fringe. London does not have a very efficient “spatial distribution of density”, which is another way of looking at it, even better than “weighted average density”.

      As a rule, cities with efficient spatial distribution of density, are the ones where the allowance of fringe growth at whatever densities people want, has kept land costs low in the entire city, and nodes of density form based on natural evolution of appropriate cluster types.

      Cities with constrained fringes and rationed land supply, always are marked by inflated land costs, and location decisions by all actors being made under highly constrained choice options. Actual clusters are prevented from evolving because of lack of spare space and because of potential participants being “priced out”. Such “clusters” as UK cities might boast, are far more biased in the direction of “old money” rather than new grassroots entrepreneurship of the famous Silicon Valley type.

      Urban areas with predominant lower density everywhere “other than where it has evolved and makes sense”, will on the whole operate far more efficiently than cities where every location is unnaturally crowded due to the sheer cost of space. It is also a dirty little secret about London, New York, Vancouver and others, that the local nationals will not live in the conditions that are tolerated by the actual majority group in society in such cities, of “immigrants within the last generation”.


      • I recall a report being written by a UK think tank last year (I think) about releasing just 2% of London’s green belt to housing development, which (from memory) would allow for development of enough new dwellings to ease demand considerably… I think it was shot down by the usual suspects…

        So London continues to have very high house prices because of supply constraints caused by the green belt… So a lot of people live in Hampshire, Berkshire etc etc and commute in because that is where they can afford housing and ‘pay’ the supply constraint tax in terms of commuting…

        If I can find that report I’ll email Michael with the link… might be a useful read also…


  10. Progressively less dense or just including the surrounding smaller towns to form a larger city. Eg, auckland super city now spread over 5000 skm when previously Auckland city was just 550skm around the harbour.


    • Well the Auckland Urban area is about 1,086 sqkm according to Google, so the ‘city’ is a little over 20% of the total area of the area covered by Auckland City (the legal entity)… so about 100,000 ha….

      Absorbing nearby towns is what happens as cities grow… once upon a time St Heilers, Howick, Manurewa, Otahuhu etc were all small towns… even Mt Albert had farm land between it and the ‘city’ at some point around 100 years ago…

      What makes Auckland seem bigger than it really is, is the two harbours.. without these the urban area would be a lot smaller… 1,086 sqkm will fit into a radius of just under 18.6km, which isn’t that big… becuase of the harbours the urban area is ‘long and narrow’ to fit in with the geography rather than ‘squarish’ or ’roundish’ as you might find with Christchurch or Hamilton.

      The harbours are Auckland’s endowment and add considerable value to the residents and visitors. it ain’t a perfect world and the amenity of the harbours means, other things being equal, a more spread city than we would otherwise have (on account of all the water in the way)…

      If you don’t like it there is always Hamilton… but then it is Hamilton… so… not sure what the answer is there !! 😜😜😜😜😜


    • The problem with comparing cities is firstly geography because a city like Auckland is elongated versus a city like Houston which is semi circular. A long city has longer travelling distance of equal land mass and rather difficult to network transport routes.

      The second difficulty is the definition of Urban. Eg the Auckland Unitary Plan classified as Mixed Housing – Urban which just means 3 level residential buildings.


      • The third difficulty is when does a small provincial town become gobbled up by a larger administration body in the interest of administration cost efficiencies? Eg the Supercity concept in effect distorting historical comparatives. Many of us do not realise a small provincial town like Pukekohe almost 60 km away from the central harbour is actually part of Auckland due to the Supercity merger.

        So when Waikato is merged with Auckland, does it mean that Auckland is now less dense? Physically has anything really changed other than a artificial boundary shift?


  11. Here is some density numbers from Auckland -Devonport Borough which existed from the 19th century to 1989. Transportblog ran a post on it. Density decline for this nearby CBD suburb is negligible. Here is the relevant section.

    “I’m surprised that Devonport’s population has been completely flat for the last 70 years. The population on census night 1945, of 11,662, was still 318 people higher than the population on census night 2013, of 11,346.

    Zoning controls, have certainly played their part in limiting the number of people who can live in a very desirable coastal area. There are probably a few more houses in Devonport today than there were 70 years ago, but any growth in household numbers has been cancelled out by there being fewer people per household.”


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s