Every weekend we dig into the archives. This post by Peter was originally published in June 2016.
Housing is a normal good. That is, it’s something that people tend to want more of as their incomes increase.
“More” doesn’t necessarily mean “larger”. People do tend to prefer larger homes as they get wealthier, but that’s not the only thing that matters. They may be willing to compromise on space in exchange for a higher-quality living space – bring on the granite countertops! – or a home in a better location. A “better location” could in turn mean anything from proximity to jobs (resulting in efficient use of valuable time), proximity to shops or cultural amenities, location in a good school zone, or access to parks or beaches.
One interesting phenomenon is that people seem to be willing to travel further to work than to consumption amenities (ranging from retail to concerts). In their fantastic book Cities and the Urban Land Premium, Dutch economist Henri de Groot and several co-authors provide some data that shows that people are, on average, willing to travel considerably further to work than to consume. They show that this results in a higher urban land premium for accessible inner-city areas, as vibrant downtown areas have the most varied and interesting consumption opportunities.
Furthermore, you’d expect this premium to rise as incomes rise, as people with more disposable income will have an increasing preference for close proximity to consumption and cultural amenities.
Is the same thing likely to be true in Auckland? Nobody’s done a survey, but we’ve got some data on the distance that people actually travel to access jobs and retail.
In a paper two years ago, I analysed Census data on commuting distances in order to understand what Auckland households spend on housing and transport. I went back and re-analysed that data to get an estimate of the distribution of commuting distances in Auckland. This data suggests that 50% of Aucklanders commute less than 9km, while less than 2% are super-commuters travelling longer than 50km.
As a point of comparison, I used data on retail spending patterns compiled by economist Susan Fairgray in a 2013 report on the Auckland retail sector. Based on electronic card spending data, Fairgray estimates that 50% of Auckland retail spending is done within 5km of people’s homes. (See Table 3 on page 58 of her report.)
Here’s the chart. As in the Netherlands, distances travelled to consume drop off more rapidly than distances travelled to produce.
There are several implications for how we build cities. The first is that we should expect retail, personal services, and recreation to be widely distributed throughout the city. Large tracts of houses without good access to shops and recreation are not likely to be awesome in the future. There are various ways to cater to these needs, ranging from mixed-use zoning that allows retail and housing to colocate to distributing small retail centres throughout suburbs (a la Auckland’s tramway suburbs).
The second thing is that we should think more carefully about how preferences for centrality are changing. The consumption amenities that cities offer play an increasing role in their success or failure. Some important consumer amenities tend to be located centrally. For example, nightlife and entertainment districts are almost always located near the city centre – think of Ponsonby or K Road in Auckland. Likewise, museums and public art galleries are usually located downtown – e.g. Te Papa in Wellington or the Auckland Art Gallery – to maximise the number of people that can access them.
As demand for consumer amenities will tend to increase with rising incomes, we’d expect demand to live close to them to increase in the future. Meeting this demand in a growing city will, in turn, mean building more apartments.
But wait! If people also want more living area as they get wealthier, doesn’t that mean that they’ll reject apartment living? Won’t apartments simply be too small to meet their needs, even after taking location into account?
It is the case that new apartments tend to be smaller than new standalone houses in New Zealand. Over the last five years, the average standalone house consented in Auckland was about twice as large as the average apartment consented in Auckland.
However, there’s no universal law that says that apartments have to be small. Policy can play a big role in keeping apartment sizes down, or enabling them to be more spacious. As LSE economist Paul Cheshire observes, planning policies (and other things like tax policies) can have the unintended consequence of discouraging adequately-sized housing:
If you really want to plan to protect and provide better access to green space and open countryside without artificially constraining land supply and forcing up house prices, then Green Fingers (or Green Wedges) would seem to be the best solution. That is what more egalitarian Scandinavians have. Copenhagen has its Green Fingers – really brown urbanisation along the radial routes out of the city with protected countryside each side. Denmark has not just got cheaper housing: according to the Dallas Fed’s data, the real house price has increased by a factor of 1.6 in Denmark compared to 3.4 in the UK since 1975 but new houses in Denmark are a lot bigger: 80% bigger in fact.
As Cheshire’s example of Copenhagen shows, it’s possible to build dwellings that meet people’s needs for living space and preserve usable open space around cities. You just need to be willing to build intensively where you do build – and integrate it with rapid transit.
For a less anecdotal look at the issue, I used Eurostat data to measure the relationship between dwelling size and dwelling type in 29 European countries. Here’s a scatterplot showing the relationship between the share of dwellings that are detached houses (X axis) and average dwelling size (Y axis). Observe how there is almost no relationship whatsoever. If anything, there’s a slight negative relationship – countries with more standalone houses may have slightly smaller dwellings, on average. (There’s probably an income effect in there that I haven’t controlled for – richer countries tend to be more urbanised, which will tend to mean more apartments, and also have larger dwellings.)
But basically, there doesn’t seem to be an inescapable trade-off between dwelling type and size. Apartments can be small… but they can also be large. And cities that are willing to let people more apartments get built will, in addition to being more affordable, give people more opportunities to realise their demands for both space and proximity.
What do you think of this data?