30 Comments
User's avatar
Samuel Hughes's avatar

I do think that technological change was a driver of suburbanisation, although I'm not sure it is as central as sometimes supposed. London already had vast elite residential suburbs in 1840, at which time it had essentially no transport except carriages. My own sense it that the most universal driver of the planned suburb was simple growth in demand and deeper capital markets.

Expand full comment
Thomas L. Hutcheson's avatar

"But [planners] have had very limited success in reforming rules on suburban densification."

And less than zero on land use and building codes reform that would, permit greater urban density.

Expand full comment
Ben Southwood's avatar

A top YIMBY told me recently that the battle on zoning had been won in his US state, at least in principle, but that the battle on costs had only just begun (and indeed, that we have lost ground in recent years).

Expand full comment
Thomas L. Hutcheson's avatar

Do you want to expand on the differences. Clearly reform of zoning and building codes is just a sliver of making regulation pass cost benefit tests.

Expand full comment
Richard Moore's avatar

Fantastic piece!

Expand full comment
Ben Southwood's avatar

I think so too!

Expand full comment
Sam Penrose's avatar

Outstanding work, thanks so much!

Expand full comment
Christian Scholz's avatar

Brilliant and eye-opening article that provides context to many of my casual observations.

Expand full comment
Nicholas Weininger's avatar

Here are a few things this piece left me curious about:

1. Japan. How did its famously density-friendly zoning come about? Why wasn't it subject to the same political-economic dynamics that obtained in Europe and North America?

2. The two-way amenity-density relationship. On the one hand, yes, less dense neighborhoods tend to be quieter, greener, prettier, and at least *feel* safer. On the other hand, walkability, bikeability, neighborhood "vibrancy", and pleasant small local shops (vs drab suburban big box stores you have to drive to) are amenities that increase with density. How do these vary with density level in practice, and how has that variation changed over the past 100-150 years?

3. The difference between density limits that restrict to 10-15 homes per acre and those that restrict to much less than that. 12 homes per acre, which you cite as the historical target of the downzoners, would actually be a good deal higher than most US cities (not to mention suburbs) today if I'm calculating right, and into a range of "gentle density" which a lot of modern American YIMBYs would be quite happy to achieve. Is that difference just an artifact of US post WWII car culture?

(One reason for asking (2) and (3) together is that I suspect that the amenities of higher density increase a *lot* from US-style 2-5 homes/acre suburbs to 10-15 homes/acre, but then not as much as you go higher than that, while the disamenities follow a quite different curve. But I could well be wrong!)

4. What has been tried to mitigate both the actual disamenities of density and the public perception of those disamenities, and what do we know about how well it works? Architectural pattern books, tree planting mandates, noise ordinances, intensive policing... what's the really effective stuff here?

Expand full comment
Samuel Hughes's avatar

These are excellent and highly sophisticated questions. I don't have a complete answer to any of them.

1. I definitely do not have a complete theory of Japan's distinctive urbanism, although like everyone I find it intensely interesting. The one thing I would say, though, is that the countries that diverged from the suburbanisation/downzoning pattern usually diverged by *never developing low-density suburbia in the first place*, rather than by not downzoning it after it emerged. I know a bit more about this in the case of Spain: Spain only really began becoming rich enough for suburbia a decade or two before the planning consensus shifted to being pro-density, so it never established the path dependencies of the suburban model in the first place.

2 and 3. I don't really have a theory of this. The theory I offer in this article explains (I think) why we ended up with classic nineteenth-century suburbs, 12 homes per acre, predominantly residential but with shops, restaurants etc. layered in. I'm not sure it explains why the USA ended up with 2-5 homes per acre with absolutely everything except homes excluded. It is possible that this is simply the urban form that homebuyers demanded, though I agree that this seems weird, given how popular the older suburbs are in lots of contexts.

4. Lots! This is too big a question for me to answer here, but much of my work is about this. My guess is that the most important policy levers are (a) getting rid of urban pollution, which by and large we did successfully long ago, (b) effective road management and public transport, which we are very slowly getting better at after a long decline, and (c) reducing crime. I do think culture and fashion has a role too: living in a dense urban neighbourhood is obviously far more fashionable in Spain or Italy than it is most of the USA, although the status of urban life has improved in the USA in recent decades.

Expand full comment
Laura Creighton's avatar

re: Japan ... I think the main reason is that homes are not an investment in Japan. The average home depreciates to zero in 22 years. https://www.economist.com/finance-and-economics/2018/03/15/why-japanese-houses-have-such-limited-lifespans So nobody has incentives to 'protect their investment'. This is causing problems today, in that there are a lot of damaged, abandoned homes. It would cost too much to improve them which you couldn't recover when you sold the fixed-up house. But this, in turn, makes 'bulldoze the lot and replace the neighbourhood' plans a great deal more attractive ... especially when existing property owners get a payout as their share of the improving-the-neighbourhood project.

Expand full comment
Patrick Jensen's avatar

Regarding question no. 2, the lower limit for dwellings per area that can support local commercial amenities rose enormously during the 20th century.

Partly this is because wages have risen enormously while capital has become much cheaper. All else being equal, this favours commerce models that have a low employee to floor area ratio.

Another factor is that increased mobility, particularly by car, increased the effective catchment area of stores, bringing smaller stores with limited variety and service into competition with much bigger ones.

The third is that household sizes have shrunk, both because families have fewer children and because adults live longer as couples or alone in their old age. This means that the same number of dwellings per area house much fewer people on average than they used to.

These factors have in many cases led to a hollowing out of amenities in neighbourhoods that originally had quite good ones.

Expand full comment
Jolyon Jenkins's avatar

I don’t dispute any of this, but wasn’t one reason for reducing urban density the incredible overcrowding of pre-modern cities? You can get an idea of this, in the case of Russia, from Crime and Punishment. Paris is another case in point - Haussmann didn’t only want to be able to stop revolutionary activities: the city had become insanitary and dangerous.

Expand full comment
mc's avatar

I agree that the generally excellent article under-discusses the historic downsides of density that have lessened over time. Improved sanitation, water quality and pest control and increased electrification were all important improvements that greatly improved the amenity of denser living. Technological change was a big factor behind the changing trends.

Expand full comment
Samuel Hughes's avatar

There is certainly truth in this, but a lot of the key downsides to density had actually already disappeared before the Downzoning appeared. The 'Urban Penalty', i.e. the tendency for urban people to die younger than rural people, vanished in NW Europe by about 1890 due to the introduction of clean running water and drainage in the second half of the nineteenth century. There actually seems to have been a small 'Urban Bonus' during the years in which the Downzoning unfolded!

Expand full comment
Luke Jones's avatar

London suburbs were seldom rebuilt as flats but generally went ‘seedy’ with age and converted to flats and rooms in situ. Some of the more outré schemes of management were obviously intended to discourage this. There’s obviously also a mirror world case in Scotland where apartment building was the norm rather than the exception but there still isn’t a clear ‘conflict’ between apartment core and suburban fringe — just that their cultural valence is inverted.

Expand full comment
Samuel Hughes's avatar

I think that's right. That is plausibly why density restrictions came in so late in Britain, only in the years after 1947 – there was very little densification going on in the first place.

Expand full comment
Dan Davies's avatar

Isn't there a political economy problem here which is the reverse of the one you identified - that in a lot of the neighbourhoods you want to densify, the voters are tenants who would suffer the disamenities but not get the wealth benefits?

Expand full comment
Ben Southwood's avatar

I think this is right. My feeling is that if this is imposed upon areas with tenants, then in order to make it work – at least in a modern Western country – tenants must also derive benefits from densification. I wonder if you have any ideas about how to make that work. One idea might be to give them a say, so that landlords in those neighbourhoods have a strong incentive to induce them.

Expand full comment
Samuel Hughes's avatar

Yes, I agree. Ben and I have worked on policy solutions to this exact problem, e.g. here https://policyexchange.org.uk/publication/strong-suburbs/

Expand full comment
Patrick Jensen's avatar

Although I've heard of it, the reference to Tel Aviv pops up somewhat jarringly only in the conclusion, without introduction. Is it missing a link, or did the paragraph it's referring to not make the cut?

Expand full comment
Xavier Moss's avatar

Excellent article! A small correction: that photograph of Toronto is not 'from around 1970', the Skydome is visible which was built in 1989. I also see buildings from the later condominium boom that really densified downtown, so I would not put that photo much earlier than the turn of the millennium. Toronto's downtown was full of surface parking in the 70s.

Expand full comment
Unset's avatar

Interesting material but I'm surprised so little attention is given to the changes in mobility that drove these dynamics - the rise first of rail and then of automobiles. Further, the nature of density has changed. Progressive era reformers were encountering large numbers of people in small, disease-ridden flats. It wasn't just that they preferred lawns. Even though we have much better sanitary conditions, no one is seriously proposing the person-to-square-foot-of-living-space that prevailed then, even if the floor to area ratio is comparable.

Expand full comment
Ockey's avatar

Solid work, thank you. I’m interested if there is validity in the perception that younger generations, increasingly frozen out of home ownership, are driving a trend toward acceptance of densification, either urban or suburban; or will do so in the near future. Perhaps this is not unique to just current generational divisions.

Expand full comment
Samuel Hughes's avatar

My impression is that this is true, although I don't think that rising demand for densification among the young will necessarily lead to zoning reform. It might, or it might not – this lies in our hands to determine.

Expand full comment
mc's avatar

Lower birthrates (and children having a greater preference for indoor activities) also lead to a lesser need for backyards.

Expand full comment
Dave Peticolas's avatar

Amazing article, thank you. Are you aware of any analyses of the amount of land value that could be unlocked through density in the major cities of the West?

Expand full comment
Samuel Hughes's avatar

Thank you! I don't think I am aware of such a study. It is complex because of course the price of floorspace will decline even as the quantity of floorspace increases.

I do know of studies on how much economic growth this could unlock, like Duranton and Puga's famous paper.

Expand full comment
Samuel Hughes's avatar

This too: https://www.nber.org/papers/w26591 So many great collaborations between them.

Expand full comment