Last week, I introduced the concept of elasticity of supply with respect to price as a useful measure of housing market dynamics. Supply elasticities measure how responsive builders are to an increase in demand. In other words, when people turn up wanting dwellings, how quickly do the tradies start building more?
Supply elasticity can in turn have a big, long-run effect on prices. If the building sector is consistently slow to respond, it creates the condition for an ongoing shortfall in supply, which means that people will bid up prices more.
My post last week took a look at some of the (limited) international comparisons of planning regulations, which seem to indicate that New Zealand is not an especially poor performer. For example, consent processing time is relatively fast and efficient compared with other OECD countries.
However, regulations are only part of the picture. For example, Patrick wrote a good post a while back looking at Auckland’s geographic constraints:
Intuitively, we’d expect Auckland’s limited supply of developable land to have an effect on housing supply dynamics. But how much of an effect should we expect?
The empirical literature provides us with a reasonable estimate. A 2010 paper by MIT economist Albert Saiz (Massachusetts, not Manukau) measures constraints on land availability in large US cities and uses them to estimate the effect on housing supply.
Saiz finds that there are large differences in land availability between different cities. For example, “flatland” cities like Atlanta or Houston have very little area constrained by lakes, rivers, oceans, or steep slopes. Over 90% of the area around these cities is available for development. Coastal cities like San Francisco, San Diego, or Miami, on the other hand, might be able to develop less than 1/3 of the surrounding area.
Saiz concludes that:
Quantitatively, a movement across the interquartile range in geographic land availability in an average-regulated metropolitan area of 1 million is associated with shifting from a housing supply elasticity of approximately 2.45 to one of 1.25. Moving to the ninetieth percentile of land constraints (as in San Diego, where 60% of the area within its 50-km radius is not developable) pushes average housing supply elasticities down further to 0.91.
Translated from economese, this means that cities with less developable land have housing markets that respond more slowly to increased demand. (Or, as non-economists might say, duh.) For context, an elasticity of 0.91 indicates that a 10% increase in house prices is met by a 9.1% increase in housing supply. Even if regulations are held constant, a “flatland” city is expected to have a more responsive housing market than a coastal city with lots of hills.
In other words, when people compare Houston’s house prices with San Francisco’s or New York’s, they’re not comparing like with like. Geography matters quite a lot!
So what does Auckland’s geography look like? A 2014 NZIER paper modelled the effect of geographic and regulatory barriers on the city’s house prices. The authors conclude that: “relative to even Australian cities Auckland’s twin harbours severely restrict the availability of well-located land close to the city centre.” Overall, they estimate that less than one-third of the area around Auckland is available for development – most of the rest is water:
In other words, Auckland has very severe geographic constraints. In terms of the availability of developable land, it’s similar to hilly coastal cities like San Diego. Saiz estimated that a city of around Auckland’s size with an average level of planning regulations would have a supply elasticity of 0.91. So: does Auckland perform better or worse than this in practice?
A 2010 study by Arthur Grimes and Andrew Aitken provides some relevant data. Using data at a district council level, they looked at how quickly new dwellings were built in response to “shocks” in demand such as increases in net migration. Their key conclusion was that housing supply in New Zealand’s urban areas tends to be a little bit more responsive than supply in rural areas:
If we divide regions into urban and rural, we find faster adjustment in urban areas (average γ1i = 0.0093) than in rural areas (average γ1i = 0.0064). This result is consistent with an active development industry, based principally in cities, facilitating new construction.
In other words, the authors estimate a supply elasticity of around 0.93 for NZ’s urban areas (principally Auckland). This is almost exactly what we would predict based on Auckland’s geography. The implication of this is that Auckland’s housing market functions more or less as expected given its geography – we don’t have to assume unusually restrictive planning regulations to explain the observed outcomes.
There are a couple lessons we can draw from this.
First, Auckland’s geography is a primary driver of the city’s housing supply dynamics. If we have higher house prices than we’d like, it’s partly because we have less land for housing. As I’ve written before, some analyses of Auckland’s high house prices fall prey to omitted variable bias – i.e. ignoring important causal variables and thus over-estimating the impact of specific policies. This can result in flawed policy recommendations.
Second, we shouldn’t compound constrained geography with bad policy. Because Auckland doesn’t have much developable land, there is an even stronger incentive to use land efficiently. (A fact with implications for transport policy, planning policy, tax policy, and publicly-owned land.) Land-hungry policies might not be too bad in a land-abundant place like Houston, but requiring Auckland to follow a similar pattern is economically calamitous.
As most New Zealand cities are also heavily constrained by geography, this challenge isn’t unique to Auckland. But it’s also not all bad: the interplay of mountains, volcanoes, harbours, and oceans is what makes New Zealand such a beautiful place to live. Let’s build cities that enable us to get the best out of it.