The badly dysfunctional New Zealand housing supply market

This chart has had a bit of coverage in the last few days.  It was produced by Statistics New Zealand, and was included in a useful release last week bringing together dwelling consent and population data over the last 50 years or so.

snz picture

As SNZ noted, there is a bit in the chart for everyone.

The number of new homes consented per capita has doubled over the past five years, but is only half the level seen at the peak of the 1970s building boom, Statistics New Zealand said today.

One sees these sorts of per capita charts from time to time, but I’ve never been sure they were very enlightening.  After all, the existing population typically doesn’t need many new houses built –  it is already housed, and the modest associated flow of new building permits will result mostly from changes in tastes, changes in occupancy patterns (eg more marriage breakups will probably increase the number of dwellings required for any given total population) or perhaps even the age composition of the population.  Even quite big differences in  the number of new dwelling permits per capita don’t, in isolation, tell you much: Marlborough and Gisborne have very similar populations, but over the 21 years for which SNZ provides the data, there were almost three times as many houses built in Marlborough as in Gisborne.

Mostly (at least in countries like this one), new houses are needed for increases in the population.  Marlborough’s population was growing over that period, and Gisborne’s wasn’t.

So we might be more interested in the growth of the housing stock relative to the growth of the population.   Growth in the housing stock is typically more interesting than building permits, because if two old villas are demolished to build six townhouses, it is the net addition to the number of dwellings that is typically more interesting, than the number of new units consented.  In recent New Zealand context, if lots of houses are destroyed by an earthquake, the gross number of new consents won’t offer much insight on the supply/demand balance.

SNZ produces some housing stock estimates.  I’m not sure quite how they do them, but they suggest that each year typically about 2000 existing dwellings are destroyed, a tiny proportion of the (current) stock of around 1.8 million dwellings.  If New Zealand’s overall population was static, there would still be a small amount of replacement activity and –  if the Gisborne numbers are roughly indicative –  perhaps 11000 new dwelling consents a year for the country as a whole would be fine.   Gisborne house prices, for anyone interested, are still lower than they were a decade ago.

Here is the nationwide picture since 1991.  This shows the increase in the number of dwellings per increase in the population  (thus, 0.4 means one new dwelling added for each additional 2.5 people).

housing stock

So, far from  the situation improving in the last few years –  as the SNZ chart above might have suggested (and as SNZ themselves suggested) –  things were worse than ever in the year to June 2016.  The population is estimated to have increased by 97300, and yet the housing stock is estimated to have increased by only 23800.  Talk about dysfunction, and no wonder house prices have been rising strongly.  In 1999, 2000 and 2001, by contrast, the population increased by only around 21000 per annum.

SNZ doesn’t have (or not that I can find) annual housing stock estimates back to the 1960s, but we can still look at the new building permit numbers relative to the change in the population.   Here is the chart showing new dwelling permits per person increase in the population.

housing 60sWhat happened?   Well, in the late 1970s the large scale outflow of New Zealanders got underway, and the number of non-citizen immigrants had also been scaled right back.  In the years to June 1979 and June 1980, the population is actually estimated to have fallen slightly, and yet 18000 and 15000 new dwelling consents were granted in each of those two years.  For the three June years from 1978 to  1980 there was no population growth at all, and yet there were more than 50000 new dwellings consented.  No wonder that over the late 1970s and through to around early 1981, New Zealand experienced the largest fall in real house prices (around 40 per cent) in modern history.

Nothing in the data suggests that the New Zealand housing and land supply market is now even remotely capable of coping with population increases of 2 per cent per annum.  Of course in some sense it should, and could, be fixed.  But there is little or no sign of it happening –  are there any reports of peripheral land prices in Auckland collapsing since the Unitary Plan was adopted? – which makes the continued active pursuit of rapid population growth look even more irresponsible (than it would already be, given the absence of evidence of other real economic gains to New Zealanders from such a, now decades-old, strategy)