Densification: not much happening in the US

The government’s housing plans –  and, I presume, the Labour Party’s –  seem to make great play of squashing more people, and more dwellings, into much the same space.    And it is certainly true that many of the older state houses seemed to sit on ridiculously large sections, (especially incongruous when the sections themselves are in otherwise very valuable locations).

Increased density appeals to planners, and perhaps even to people in certain demographics.  I wouldn’t want to stand in the way of people who prefer to live more densely. But that is rarely enough for the enthusiasts.  Instead, much rhetoric is aimed at so-called NIMBYs, people who might be reluctant to see a change in the character of their neighbourhood pushed through by bureaucratic fiat.  There was an article in today’s Herald along exactly those lines.

As I’ve noted here previously, over history, as cities have been richer they have tended to become less dense, not more so.     Space seems to be a normal good.

And so I was interested to stumble this afternoon on an article in the New York Times, with a couple of interesting graphics on changes in density in 51 metropolitan areas (population in excess of one million) in the United States over the last few years.  2010 to 2016 is quite a short period, but –  given history –  the results shouldn’t really be surprising.  I can’t cut and paste the graphics (click through to have a look), but what they highlight is two things:

  • only a fairly small number (10) of US cities saw increased density during that period, and the increases in density were typically small (although Seattle stands out with by far the largest increases)
  • the remaining 41 cities (metropolitan areas) saw a shift towards less density over that period, and many of those falls were quite substantial.

In  the chart, there is also some suggestion that areas with faster population growth were more likely to have become less dense over this period (Dallas, Raleigh, Houston, Nashville, San Antonio and Austin stand out).

It looks a lot like a case where cities spread –  people want more space –  when land use restrictions don’t stop them doing so.   It isn’t obvious why New Zealanders’ preferences would be that much different from those of Americans.   And it is hardly as if New Zealand is short of land either.

Land use restrictions may actually stop cities’ populations growing much –  at least in the US where there are many big cities to chose from (some with tight restrictions, others without).  That would seem to be the message in the latest Hsieh and Moretti paper , which highlights how little population growth there has been in the US cities with some of the tightest land use restrictions (San Francisco, San Jose, and New York) relative to other cities.   By doing so, those restrictions may have imposed substantial real economic costs (lost opportunities to take advantage of high productivity opportunities in those cities).   The case that such restrictions might have had a large real cost here is less strong –  numbers in Auckland has grown very fast even with the restrictions.  Perhaps here the cost is “simply” the shockingly high cost of purchasing house+land, a systematic redistribution against the young, the poor and the credit constrained.

UPDATE: For anyone interested, John Cochrane has a nice post explaining clearly, and in more detail, what is going on the the Hsieh and Moretti paper, and commenting on a couple of other papers in a similar vein.

49 thoughts on “Densification: not much happening in the US

  1. Michael cities do not become less dense. Nowhere except for failing cities like Detroit are there neighbourhoods where there are less dwellings now compared to 10, 20, 30….. years ago.

    Average density may be falling if large amounts of low density areas are added to the urban environment. It is a surprisingly hard thing to measure average density. Does average mean density where the average person lives or is it it just dwelling numbers divided by urban area.


    • The cost of America’s inefficient sprawl which have lead to a huge number of bankrupt American cities.

      When sprawling new development happens, it’s easy to mistake that for prosperity. New buildings and wide roads look great when they first meet the eye. But over time, distant development costs more, gradually bleeding taxpayers and putting the hurt on municipal budgets.

      Think about it. Every time a new, spread-out subdivision is built far away from existing infrastructure, somebody has to pay for a bunch of roads that serve a small number of residents.


      • You will find there is no evidence to support this now – this is a 5 year old article.The Smart Growth movement has been discredited – the result has only been increased traffic congestion and unaffordable housing.

        Liked by 1 person

      • One good source that is sensible is better than zero source from you. Also it is low density sprawl that creates traffic congestion.


      • Getgreatstuff says “it is low density sprawl that creates traffic congestion”.

        This is a common misperception, but it is one of the worst falsehoods believed in by people today. It is not supported by the data at all.

        If we take Auckland, for example, most US cities of comparable population are only 1/3 the density. For example, Indianapolis, Nashville, Richmond. So what does the TomTom data on traffic congestion delays, tell us?

        Auckland; 45 minutes per 1 hour of driving at peak. The US cities: 13 minutes (Richmond), 15 minutes for the others.

        This is not cherry-picked examples to fit my case, either. The trend lines in the correlation between density and congestion run in this direction. The myth was never supported by actual data, and in fact until TomTom and other organisations started providing GPS-based analysis, the charlatans who were promoting the myth were able to exploit the lack of good data. The lack of interest in assembling such data, on the part of well funded bureaucracies at all levels, local, national and transnational, was matched only by their zeal for promoting a falsehood that to their warped perceptions, was so self-evident it did not require actual analysis.


  2. I’m not sure how this NYT guy did his numbers, but average density (people per sq km) seems relevant to me (rather than noting that say downtown Bruges probably has the same number of buildings as hundreds of years ago).


    • Michael the cities around the world which had the biggest reductions in density (residents per hectare) were the ones with large inner city slums, tenement buildings etc. As those cities became more wealthy those slums emptied out.

      That never really applied to NZ. We never had much of that slum like conditions -so the reduction in density (people per hectare) as incomes increased story does not play out so much in NZ.

      I would say that most NZ suburbs had a decrease in the number of people living per dwelling -on average over time -as family size became smaller. Countering this declining suburban population from the smaller family size effect was a compensating gradual increase in the number of houses -the infilling effect.

      You can see from Devonport’s population history -these two effects have cancelled themselves out -population and therefore density has been stable for 70 years.

      I suspect the normal good aspect of housing -is mainly about how much residential floor area per person and not so much about garden area per person. This makes intuitive sense to me -if the big demand driver in cities was more garden area -well residents could achieve that by moving to the countryside and we wouldn’t need cities at all!

      Finally I would say that trying to make some sort of policy analysis from US city density data (especially as there is questions about that data) to the question of whether Auckland should densify (dwellings per hectare) its state housing stock is a pretty long bow.


      • Brendon

        Re your normal goods point, no I don’t agree about gardens. after all, if everyone moved to the dispersed country there would not be good jobs for them. Cities exist for a reason, and individuals have to make tradeoffs (for most people two acre gardens aren’t feasible – economic – in the city, in the way they might be where farmland was $10000 a hectare). Of course, different people have different tastes – some positively value apartment living (I heard Peter Nunns give a lecture the other day extolling high density communities for their dating-supply advantages), others want gardens, others just want physical distance from other people.

        On your final para, I’m not trying to do policy analysis at all here, just suggesting what the US experience – with much more variegation in planning controls – might say about preferences of people not unlike NZers (New World residents). I actually draw the same lesson from NZ provincial cities – where there is no sign at all of dense inner city living. Of course, in growing cities individual suburbs will change character over time – Mayfair was once in the country – but high density living in rich countries looks as though it is mostly a function of planning restrictions, which make the amount of land people might normally consume unaffordable to most. Personally, I’d favour more or less full freedom to choose even if in fact everyone wanted to live in a high rise apartment.


      • Michael -sure there are lots of effects going in cities. My point about demand for garden space in cities is that it might be a weak ‘normal’ demand compared to a stronger ‘normal’ demand for residential floor space in a city.

        WRT density -sure there is the well documented effect that when incomes and mobility have allowed it -cities have spread out -I agree with that. But the even longer story (over centuries) is population (and even more so the density of built space) of city neighbourhoods have increased -if that wasn’t true then cities would not exist. If the human desire to disperse was greater than the desire to agglomerate then cities would not exist.

        The story of the world for the last few centuries is that people have moved from living in spacious (less people per hectare) rural communities to living in less spacious (more people per hectare) cities. That is the bigger historical story of density -to not acknowledge that effect is to miss one of biggest economic stories in the world. Humanities move to high density environments is probably as big a story as the shift from hunter gathering to settled agricultural communities.

        Michael there is some economic analysis of intensive versus extensive marginal growth in cities -perhaps you should review that. There is also a lot of empirical data showing that US cities are outliers -in having unusually low density compared to cities in Australasia, Europe and Asia. Because of this outlier fact -it is debatable how relevant US city data is to NZ.

        I find it perplexing that in rural areas, NZ allows intensification (land to change its use) -dairy conversions for instance -with very little limits -despite the negative externalities. Whereas in NZ cities intensification (changing land use) is much more limited -despite the (in my opinion) lesser negative effects on the community.

        Michael as an economist I am sure you would think that a fixed labour supply -where workers are prevented from changing jobs is inefficient. Imagine if the only way to shift labour resources was by employing new workers (immigrants and the young) in new areas that needed more workers -how inefficient would that be? Economists spend a lot of time advocating for flexibility of the labour market.

        Surely by the same argument a fixed land supply is also inefficient -land use in cities should be able to change in response to economic forces. Surely economists should be explaining the benefits of cities decreasing the marginal cost of intensification (and extension)?

        I find it disturbing the superficial discussion about density and land use for NZ cities -where most of us choose to live.


      • Brendon

        This seems very strident and rather tetchy response to a post that was mostly a link to a couple of overseas pieces – one purely illustrative, and one suggestively illustrating exactly the potential costs of land use restrictions (at least in the US case – since those guys first came out with their research I have consistently cautioned that the results probably aren’t valid for NZ).

        And yes, of course, i agree that the rise of cities is a remarkable phenomenon. So average density across whole countries is growing (both trivially in that we have more people and the same land, and substantively in that a smaller proportion of the people are in scattered villages and farmhouses, and more live quite close to each other in large numbers). It doesn’t change the fact that space – floor space and garden/land-between-houses space – is probably a normal good. We’d like more of it, but we respond to price. In a well-functioning market, price wouldn’t adjust that much (except perhaps v close to the centre of the city), but in our rigged markets it does.

        I also agree that land use in cities should be able to change. But I also think that existing users have rights, which should be protected. In new subdivisions people do it routinely, through covenants, which seem to be an efficient market response. Thus I’ve argued for something similar in existing neighbourhoods, while being symapethic to your own proposal for smaller-scale collective action options.


      • I recall an urban planner telling me that in many inner city suburbs (like old tram suburbs 2-5kms from the city) in Australia the number of residents per dwelling had fallen dramatically in the last hundred years, as a function of both the steep fall in the number of children per family unit as well as the reduction in the cost of transport. Old shotgun cottages which used to house 6-8 people now house well to do DINKy couples. And I don’t know that those suburbs ever really constituted slums, but the effect still stands.

        That sentence “Space is a normal good” I couldn’t reiterate more. And I don’t see the complication in understanding the dynamics here – virtually everyone would choose to live in a grand mansion on several acres in the middle of the CBD if given the choice, and this behaviour drives prices for land. The only complication in the last twenty years is that land is generally harder to come by than built space, due to green belts and supply restrictions compared to a relatively competitive market for building supplies (and marginally fewer restrictions on build space, particularly in newer subdivisions).

        Let’s put this another way – if you’re looking at an advertisement to sell a house or a unit, and it has space in either the garden or the dwelling itself, does this get a high priority in the marketing? If the answer is yes, then I think you know where people’s preferences lie. In fact, I’d argue that space, or the right to more space, probably comes at the highest priority over the build quality, appliances, or anything else at all.


      • Ryan what would you (or some representative person) prefer more, what would you be indifferent to, what would you be willing to trade -wrt a larger house or a larger section? I think the demand for floor area (per person) is greater than the demand for site area (per person). But I do acknowledge they are both normal goods -that all other things being equal -as incomes rise more will be purchased.


      • Brendon – space vs built floor area? To be honest, not sure. Probably land, because I could turn it into built space later if I wanted to.

        This question makes me think of all those McMansions that planners hate, built in the last 20 years, occupying all but 10sqm of the block. One might look at this and conclude that people don’t care about land. Alternatively, you could point out that section size has fallen from 1000sqm to 400sqm and people aren’t that willing to reduce the size of their preferred house accordingly (and can’t afford the modern 1000sqm block).

        Of course everyone’s preferences here are going to fall differently though. Ideally the planning rules, and by extension the market supply, will permit that diversity of tastes.


      • Brendon, what you say about the old city cores, with their “slums”, reducing in density, is actually true of most cities, at the same particular stage in the process of the economic development of each of them. Whether it was slums or formal-housing-market tenements, crowding has always been the mechanism by which high densities were achieved in the early stages of national economic development and urbanisation. If NZ was different, then NZ is pretty unique. Probably because it was always so under-populated, that it never needed crowding to enable a balance between local mouths to feed, and the food-producing space available to supply it without refrigeration or high-speed transport. Horse ownership was far more common for NZ colonial citizens than it was in the old countries.

        At the presentation Michael refers to (sorry I failed to spot you and introduce myself, Michael) Chris Parker also presented a chart from a forthcoming work by Solly Angel et al to show that almost all urban areas in a massive international data set had reduced in density in recent decades. Developing nations cities are going through a similar process that first world cities went through a century ago. The tiny minority of cities that are getting denser are a few in the first world with strangulatory planning and lack of competing options for their population – including Auckland and some UK cities. All these cities have a housing affordability crisis. In contrast, many of the developing-nation cities that are spreading out at reduced density, have falling land values and improving housing affordability.

        I get your point about the core versus the “newgrown” urban area around it, but I think you are wrong to suggest that there is not also density reductions in the core at the same time as the growing / “urbanising” population is choosing suburban living. The only data I have seen on this, from Angel et al again, does show that urban cores in western cities had extremely serious crowding in the Victorian era, comparable to that in Dhaka and Lagos today. Regardless of the absence of tall buildings, the highest recorded densities in the world and in individual cities histories, relate to pre-automobility-era crowding.

        Not every city ends up building “up” impressively in its modern phase of development so as to provide decent floor space for a population level reversal; it depends entirely on the nature of the city’s economy. New York is different to Detroit for a perfectly logical reason. Had New York’s collapsed garment industry not been replaced by global finance as the primary reason for the city’s existence, New York would have been just another bigger Detroit. That garment industry was in its time, bigger than automaking was in Detroit. I believe that most cities in Europe have not rebounded to the core densities that existed in 1880. A few exceptional ones might have – but you are aware, are you not, that even Manhattan does not have as many people now as it had in 1900? I believe that would be true of most European cities too – none of them have added floor space in tall buildings like Manhattan has.

        The oldest cities in the USA are not that different to Europe’s cities, in terms of their pre-automobility form. It is a mistake to categorise the USA’s cities as uniformly different to those of the rest of the first world. In fact things were very much of a muchness around 1880. What does make the US unique, is that it does have cities which were almost nonentities in the pre-automobility era, and which evolved since. The absence of comparable cities in the rest of the first world is probably because of more prescriptive central planning for growth, and lack of market freedom. The Economist did a good essay called “Concrete Gains” a few years ago, where they pointed out that Europe greatly lacks cities of the kind that the US has many of, that are smaller or medium size and high-growth. Europe tends to have more “primary cities” that hog all the growth (if there is any), while everywhere else stagnates.

        But where the USA’s “old” cities differ from the rest of the world, is their absurdly low density sprawl in the automobility era. See Fulton, Pendall et al, “Who Sprawls the Most”? Ironically, it is the oldest cities in the USA with their dense and strong pre-automobility cores, that have had the lowest density suburbs; while Los Angeles, that had almost nothing “old”, but rapidly sprang up into a superstar city in the automobility era, had suburban densities very similar to those in Germany, France, Canada and NZ. Lot sizes were typically 1/8 of an acre. The 1/4 acre was mostly a myth in NZ and is probably not something that most people ever did or do aspire to.

        The massive and absurd large lot sizes in many US cities new suburbs since the 1960’s, are a perverse consequence of Federal education policy combined with local zoning policy and local taxation / school funding policy. Glaeser discusses this in “Triumph of the City”. No-one actually wants a 1 acre lot, or a 2 acre lot, or a 4 acre lot – they are a default mechanism to the end of making the suburb “exclusionary”. People want and are paying for the “exclusionary” aspect, the large lot is just an unfortunate “necessary evil” to them, as other mechanisms for “exclusion” have been ruled against by the courts.


  3. Houses in Mt Eden including double grammer zones are starting to fall quite dramatically due to the single house zones that Nimby groups have been fighting for. Not sure if they are now happier with the lost equity. Houses that have accepted the higher density appear to hold its values quite well due to the development opportunities now presented.


      • Actually there were 2 points in that paragragh.

        1. Nimby’s were under the impression that higher density would preserve their equity and loss of their leafy suburb from higher density would adversely affect their wealth. But they are being proven quite wrong in that assumption. Once the emotional chinese money has dissipated we are back to values being supported by fundamentals.

        2. Increased density provides a logical and substantive support for higher prices sustained in recent years.
        higher density sites therefore justifies higher values than single house sites.


      • Correction: Nimby’s were under the impression that lower density would preserve their equity and loss of their leafy suburb from higher density would adversely affect their wealth.


      • David George, it is quite interesting that you would call a counter factual argument with supporting references as trolling. Would it not be better for you to counter with a robust factual comment with supporting references rather than resort to name calling?


      • Getgreatstuff actually has a point there. The assumption that allowing higher density will make housing more affordable, is entirely unsupported by the evidence. In fact “site values are elastic to allowed density”, and the elasticity is so great, in combination with the way urban land markets are distorted by prescriptive planning, that the smaller you can force the average new dwelling size to fall to, the more expensive the average dwelling is.


        The assumptions on which modern urban planning is being conducted are basically wrong about everything.

        Intensification happens faster because the price of sites is disciplined, when there is sufficient liberality of greenfields growth. See this excellent assessment and my comments below it:


  4. In 1981 I did some Uni research about the “return to the inner city” using my suburb Ponsonby as the example. In the census data I was surprised that the population had nearly halved in the preceding 30 odd yrs. A lot was the clearing of the Freemans Bay ‘slums’ but also a lot was the “Yuppies” buying up the old villas. There was the perfect example across the street where one old two storey Victorian house had 4 students in the upstairs flat and a Polynesian family of 6 in the lower flat. The next door identical house had been restored by the childless young (and painful) professional couple to have only the two of them as occupants.

    So the density reduces even though the actual dwellings stay the same – or even increase slightly. Councils have contributed to this reduced density – the clearing of the slums but also things like the Wellington City Council grants for renovating inner city houses in the 80’s – and also residents parking to ensure the values are kept high in inner city suburbs. And judder bars to keep traffic out.


    • If they had bulldozed all these low density housing and put in 50 level buildings. Then we could still have the 4 students living in the same neighbourhood. We could also have the 6 polynesians also in the same neighbour hood plus the new yuppies in the same Ponsonby. But instead we just have very very expensive yuppy houses and students and polynesians living in cars.

      Better look at your Uni research again because it looks sadly lacking in substance.


      • No, that Uni research has the same results as all such research. It would have surprised me if it had shown something different.

        And your assumption that building “up” instead of “out” results in affordability that allows the original population to afford the rents as well as the new added gentrified population, is totally wrong. There is no example anywhere in the world to support this. Cities originally have slums and overcrowding; when development and better transport systems and suburbanisation arrives, the people disperse to more decent accommodation and density falls. Most cities do not even reverse this trend at all, in spite of all the hype we hear today.

        The cities that do reverse this trend and start building tall apartment blocks in the modern economic era, are exceptional, and they are all expensive and gentrified and exclusionary. The proles who do the shelf-stocking and cleaning and delivery vehicle driving in these cities, all live somewhere else, somewhere they are “priced out” to, and often have the most arduous commutes to work.


    • Humans are predisposed to highly value pleasure today and to deeply discount future pain, especially the more distant it is. It’s easy today to rationalize that future expense, especially when you feel so assured that new growth will make those future people better off. This thinking is how you end up with two dollars of public infrastructure for every one dollar of private investment. This is how you spend yourself into bankruptcy.

      It’s a predicament that nearly every American city, with very few exceptions, finds itself in. Even if there was enough wealth and productivity to fix all of this — and there isn’t anything close to that amount — we would be fools to spend it so unproductively.

      All this infrastructure is a bad investment. America needs a different model of growth and development.


      • Rubbish. The people either pay for infrastructure for growth, in local property taxes, or if the growth is constrained by planning, they end up paying far greater sums of money in aggregate, in rents and mortgage costs. Of course the property-rentier class loves it this way.

        There is no correlation between the density of a city’s growth, and its fiscal position. The fiscal problems all have their origin elsewhere, such as the local economy being based on a sunset industry.


  5. Given the way houses are currently built in the new suburbs, simply requiring all new housing to be two storied would allow a higher effective density with out reducing the amount of open space.


    • The new Unitary Plan does allow for a 2nd dwelling concept allowing for a separate dwellings attached to the primary dwelling in a dwelling existing when the Unitary Plan was first notified in 2013. But this concept has not really taken off as it is restricted only to older primary residences and you cannot build new dwellings of this nature.

      However dwellings of this nature would only lead to more expensive houses supported by the home and income concept or a 2 rental concept if you don’t inform Council that it is not a primary residence. Higher income equates to higher value and not lower value.

      The current higher density offered by the Unitary Plan, Urban dwelling allows only 3 levels which is just not high enough to lower the unit cost substantially.


    • Site values are elastic to allowed density, when the supply of land for the urban economy is rationed, and the lower value land in rural use is unable to be competitively used to discipline land values in the entire urban area.


  6. I Got a fraction of the way into the Hsieh and Moretti paper … then gave up … they didnt provide a strict definition of the various dense terms they use and one had to guess what the ordinary every day meaning was of spatial misallocation and spatial dispersion … not common usage in a New Zealand context


    Hsieh and Moretti compare 220 cities accross America .. which is fine … an underlying unspoken consideration is the mobility of labour and in that regard you implied comparison of the USA 220 city study to a New Zealand context is deficient in the sense that the cost of labour mobility in USA is massively different (cheaper) than NZ where US petrol is 50 cents per litre while NZ cost is $2.00 per litre

    Accepting the Hsieh Moretti proposition that labour drives productivity and that in turn rubs up against availability of housing which in turn rubs up against house prices … what would happen in NZ if the cost impediment to labour mobility was removed


    • I explicitly noted in the post that I do not think the Hsieh and Moretti results would apply to NZ. As I noted, we have one semi-big city, whereas in the US there are many different ones (with different planning regimes) for people and firms to locate in. And our one big city has seen staggering population growth over the decades – since WW2 probably the largest growth rate for any OECD country’s largest city, with the exception of Tel Aviv- casting doubt on the idea that people have been held back in any number from moving to Akld.

      And, as I didn’t mention this time but have previously, the gap between GDP per capita in Auckland and that in the rest of the country is very small by international standards (big cites vs rest of country) suggesting that even if people had been held back from moving to Akld, the GDP losses would have been pretty modest.

      Even since Hsieh and Moretti first came out with their results I have consistently cautioned that although I (a) wished they were true, and (b) wished they were applicable to NZ, I was sceptical on both counts (especially b).


      • Auckland’s growth may be record high for “a nation’s largest city” but it is not exceptional at all for small cities. Many US cities of similar size have started off smaller than Auckland and are now bigger. They are different in that their growth is because they are providing affordable escape valves for population and business from elsewhere in the country.

        And in most cases, “the nation’s largest cities” have been strangled by regulations, which is Hsieh and Moretti’s whole point – those cities might have grown as fast as Auckland had they provided for the growth instead of losing it to cities with more competitive land prices.

        Auckland actually should not be growing as fast as it is at the same time as its regulatory settings endure the growth feeds a land-value Ponzi; it shares with some of the world’s “global superstar cities”, the feature that much of the population growth is immigrants from countries where they are used to disgraceful crowding and substandard housing and gouging rents. This is virtually the only way that any first world cities are “becoming denser” today, and they are exceptions, not the rule.


  7. Despite my rants about density, Michael has put some good links in his article showing the disturbing costs of restricting land supply (excessive barriers for changing land use) -which are well worth reading -they do not mention density…. ; )


  8. Two inventions, in particular, a hundred plus years ago had a big effect on the density of the built environment. One to build cities up -elevators. One to build cities out -automobiles. Density of cities has never been a one-way street…..


    • It costs around $150,000 to $200,000 to put in an elevator system with its regular maintenence costs. After 3 levels it is a requirement to incur that additional cost, however it does not equate to cheaper buildings. The problem with 3 level Urban dwelling zones that allows for 3 level buildings is that no one wants a 3rd level apartment where they have to walk up 3 level flight of steps everyday with a heavy load of shopping or carrying a child and a stroller or even with older folk bound by a wheelchair.


    • Density is a one-way street, actually. Elevators allowed people to live in city cores in substantial numbers and still have a semi-decent amount of floor space. But even so, in what city core has the population in 2000 been higher than what it was in 1900, assuming it was a city of note in 1900? A city might well have 10 times as much floor space, but in every case there are less people.


  9. Michael I apologise for my strident tone. I just think that density is a red herring. There is no optimal ‘density’ -there are forces pushing for densification and forces pushing for dispersion. Density responds to economic forces, preferences and how city land markets are regulated. I suppose that was a point of your article -showing the different density responses of various US cities (although I would like to know how density was measured -Houston for instance inside its motorway ring -an area bigger than Auckland’s isthmus has had significant densification of its traditional standalone suburbia housing).

    The variation of US cities may be less than what could be possible -as US cities do not control their transport budget/infrastructure provision -much of the funding comes from State and Federal sources -so there is to some extent an ‘one size fits all’ effect. Transport infrastructure has a big effect on shaping cities. Also there is a lot of lobbying for a particular ‘desirable density’ because that supports various lobby groups positions. So caution needs to be taken….


    • Brendon

      Yes, I agree there is no “optimal” density. There are historical revealed trends. But my call is for effective choice.

      And yes, as always, caution is needed in interpreting any actual experience, here or abroad.


    • It would be apropos to this thread, to recommend the work of Issi Romem on US cities and their density trends, eg:

      On the influence of Federal Highway spending, I believe that the historical record shows that this had minimal influence on population dispersion rates and falling urban land values. The mere freedom to travel on roads of any kind at a relatively high speed (in human-history terms) is all that ever mattered. Model T Fords on dirt and gravel roads. The dispersionary trend was well in place before the Federal Interstate Highway program even started, and the land value trend had already played out completely. The urban land value collapse was over by 1950 and merely flatlined from there on.

      In developing nations, motor scooters and the dispersal of illegal slums are having the same effect, without any need for highway building programs.


  10. Hi Michael
    A couple of points on the Hsieh and Moretti paper, which argues that land restrictions in some very high income cities such as New York and San Francisco lower average productivity across the US as they provide a disincentive to move from low productivity to high productivity locations.

    (a) Their estimates are higher than others eg Glaeser and Gyourko, Albouy Behrens et al (2015) – there is a range, some of which suggests these costs may not be as high as HM suggest. Moreover, HM measure the impact on productivity: the impact on “utility” may be lower, due to the negative congestion externalities that people ignore when they move to a city. (This, as you know, is a standard result in the literature; but it is a second best result, due to myriad positive and negative externalities associated with cities and taxes and land markets, so it is pretty important to use the word “may”).

    (b) As Feldstein argued a couple of decades ago, there are other reasons why people may not move, eg the disincentive effects of the income tax system. Why undertake the cost to move to NY to earn $20000 gross if after taxes you only get $12000? (Feldstein argued that this was the major disincentive effect of the tax system, as, for blokes in any case, the tax system doesn’t have much effect on how much you work but may have a disincentive effect on the type of job or the location of the job you choose.) In short, if you believe the numbers with respect to land regulations are large, you should be willing to track the consequences of other economic interventions.

    (3) Cities sometimes have land restrictions because they believe they increase the utility of city incumbents by reducing the negative externalities associated with larger and larger numbers of inward migrants (eg because of the congestion costs they cause.) Without debating whether their beliefs are true or not, is there a good philosophical reason why the incumbents can’t legislate land use? Sure, it may lower aggregate productivity. But I could easily make a case that aggregate productivity would be higher if more people worked at high productivity businesses rather than low productivity businesses. Should the government force high productivity businesses to raise their employment just because they are high productivity? As many businesses aim to maximise profitability not productivity, it is reasonably certain that aggregate productivity could be raised by forcing productive private businesses to raise their output (and lower their prices) and hire more people. The government doesn’t like forcing incumbent businesses to hire more people; should it also be wary of forcing incumbent cities to accept more people? Similarly, should we prevent councils in high visual amenity areas (views of Fiordland, or nice beaches, say) from having land restrictions that preserve the views?. I don’t know the answer to these questions, but for my own personal preference I would not have been adverse to the Buffalo District council (or whoever) having prevented very ugly shopping malls being built within a kilometre of the Niagara Falls.

    (4) World productivity could probably be raised by forcing highly productive countries to accept more people from unproductive countries. Most countries have a tradition of restricting migration at their borders, which suggests they don’t see raising world productivity as a paramount concern, particularly if they believe it comes at the expense of the productivity, wages, or living standards of incumbents. (Again, I shan’t try and argue whether this belief is true or not). Most countries also have a tradition of allowing free migration within their borders, even if cities would prefer fewer people. There is some tension here, but it isn’t obvious philosophically why a city should think it should be welcoming to all migrants, internal or external.

    Anyway, thanks for stimulating lunch-time fare.


    • Thanks for those thoughtful comments Andrew. The Hsieh-Moretti results do sound high, even for the US, and you are right to point to other lower estimates. As I’ve noted here previously I’m sceptical they would apply to NZ, given the very population growth Akld has had over decades and the low productivity premium in Akld relative to the rest of the country.


    • Andrew -using zoning/planning regulations to restrict resident numbers of a city “because of the congestion costs they (newcomers) cause” is a very blunt instrument -with large consequences -it results in rationing which will price out those who are on middle to low incomes. Congestion pricing and infrastructure charges would seem like a better option than the rationing approach. David Lupton -outlines such a approach here.
      View at

      Michael -I have a friend whose ‘take’ on NZ’s housing supply restrictions/unaffordable housing situation is that given enough demand any town or city is the next Auckland i.e. nowhere in NZ has elasticity of housing supply. Even Christchurch, despite much crowing from Nick Smith, only achieved house price stability when the earthquake housing demand shock passed.


      • I certainly say that nowhere in NZ has elastic housing supply. 50 minutes from Manhattan, you can find rural towns with 30,000 to 50,000 population, with houses below $100,000.

        The cheapest old dumps the same distance from Wellington or ChCh, are still $250,000+ and in Auckland’s case, the price is more like $350,000+.

        Even Paris has rural towns around it with housing cheaper than Warkworth or Levin.

        Nowhere in NZ, are developers buying land at true rural values and building houses on it. Not around Levin, not in Selwyn or Waimak, not around Masterton, not around Pukekohe or Warkworth or Huntly, not even around Palmerston North or Fielding or Ashhurst.

        While some regional councils in rural areas near a city might be doing their best to permit new developments, eg Selwyn and Waimak, they are small enough that the deflected activity from the anti-development city swamps all the land supply they can “enable”. The way out of this impasse would be MUD-type developments with large investors doing the infrastructure themselves as well as the property development. But of course where you have a Super Council or even a Regional one like GWRC, you get deliberate rationing of supply around the rural towns as well.

        Christchurch hasn’t so much achieved “price stability”, as that its bubble inflation hasn’t kept up with Auckland, Wellington, Tauranga, Hamilton, etc lately – thanks to Selwyn and Waimak Councils, not anything that CCC, Brownlee, Smith etc etc have done.


      • Re elastic housing supplies in different cities: which is to say that CHC has all the same constraints on supply that any other city has, but a lack of demand large enough to result in major capitalised scarcity values.

        This all points to a chronic, and very common, failure of governance. No city is currently setting policies to absorb all this pent up demand from currently very expensive cities with a policy framework which is genuinely pro growth. People are simply moving to locations where the scarcity premium is lower (for the moment).

        There was an article in the Guardian back in 2013 where an economist compared the housing supply problem to the corn laws: a stable equilibium outcome with a steep transitional gains trap which is driving society apart. I still think that’s true.


    • Andrew – in fact productivity increases as an urban area grows, regardless of whether the new residents are crammed in to existing built areas, or added in new ex-fringe suburban development. NIMBYism need not be an obstruction to a city growing in size to become more productive.

      In fact there is a myriad of effects from curtailing fringe growth, that run via inflated land values and extractive land rent, that has a significant productivity-reducing effect anyway. Land is a factor of production, and different industries need space at different rates and have different abilities to pay for it. Also, agglomerations are of multiple types, and many new ones need space and low site costs if they are to evolve at all. Silicon Valley evolved on low-cost exurban land. It is common knowledge among UK urban economists, that their planning system and exorbitant site costs and sheer lack of space for new uses, mean that nothing like Silicon Valley can evolve there.

      While some superstar cities are “compact” and have anti-sprawling regulations, and are marked by “high productivity”, that is partly economic-evolutionary luck and partly “exclusionary” effects – the lower-productivity industries that are nevertheless of value to a national economy (and provide employment of kinds that might not exist otherwise) are excluded by high land prices and other local costs (including the cost pressures from workforces housing costs). In every case, such a city could be larger, and would have lower-productivity sectors and workforces in addition to its high-productivity ones, and due to the size-productivity connection, the national economy would be better off than if the lower-productivity industries were elsewhere, or did not exist at all. The danger when a nation is practising urban-planning groupthink and nowhere is “affordable” – eg in the UK – many potential sources of employment and export growth may be foregone altogether.

      Immigration to “more productive countries” would certainly be more advantageous on net if the more productive countries had at least some cities with regulatory freedom around urban growth, rather than the immigrants cramming into high-density housing, forcing property values up, and remaining unemployed because the industries that match their skill levels do not exist thanks to local planning laws.


    • Interesting points here.

      3) I’ve always thought that the case to be made is that incumbent residents collectively deciding on policies that enrich themselves at the expense of new residents is no different from a cartel (in this case cartelising the supply of land in a defined area). Society is comfortable making cartels illegal in most other circumstances, and in the cases of natural monopoly utility network services, this often results in some incredibly invasive policies restricting their behaviour. The obvious difference here is that most people don’t feel a sense of ownership in a monopoly water business, and won’t miss its dividends.

      Further, I think the array of other policies magnifies the impression that new residents will have a negative impact. Free high quality schooling, free roads, subsidised water, and a range of other services which don’t recover the costs of their assets; means that these goods don’t get a price from the marginal user, and they don’t get supplied in greater amounts when demand rises – because there’s no budget to provide it.

      It’s a bit like the members of a relatively busy golf club deciding that it doesn’t want to accept new members: the charges today only cover the costs of the gardening, and any new members are just going to make it harder to get a 7am tee off time on a saturday. Nobody has the money to build a second course to meet demand or its practically impossible. To get a membership you’re waiting for an existing member to give up golf, or you’re paying far in excess of the notional dues to buy a membership off someone else. In this case, yes, the addition of new members is a pretty negative experience, and the existing members are pretty happy with the arrangement.


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 )

Google photo

You are commenting using your Google 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