Coleman and Karagedikli on residential construction and population growth
A new discussion paper (DP2018/02) has been published on the Reserve Bank’s website:
By Andrew Coleman and Özer Karagedikli
Non-technical summary
This paper aims to understand how population growth has affected building activity in New Zealand regions during the
last twenty years. Using panel data regression techniques, we estimate that 0.25 – 0.30 additional houses are built for
every additional person in a region. The additional 0.25 – 0.30 building permits per person equate to about 40 m2 of new
construction, with a value of just over $60,000 in 2016 terms. This construction is in addition to the ‘background’
construction that occurs to replace old houses, which amounts to 2.5 – 3.0 dwellings per 1,000 people per year, or
approximately 11,000 – 13,000 dwellings per year.
The estimates suggest Auckland’s construction shortfall between 1996 and 2016 was between 40,000 and 55,000 dwellings,
or approximately 10 percent of Auckland’s housing stock. The estimates of the shortfall are fairly robust to changes in
the specification of the models; moreover, they all suggest that the shortfall was modest until the end of 2005, when it
increased rapidly.
We also examine the relationship between the size of newly constructed dwellings and population change. Since four of
the sixteen New Zealand regions experienced almost no population growth over the period, it is possible to contrast the
size of newly constructed houses in regions experiencing population change with those that did not. These estimates
suggest that, at least until 2005, smaller houses were constructed in growing regions with above-average incomes,
particularly Auckland and Wellington, than in growing regions with below average incomes or in regions with no
population growth. This difference appears to reflect the much younger age profile of the residents of Auckland and
Wellington. It appears that Auckland’s housing shortfall was less severe prior to 2005 precisely because of the large
number of small apartments that were constructed in the city. Not until apartment construction almost completely ceased
in 2008 did Auckland’s housing shortage started to become acute.
Finally, we analyse the relationship between population growth rates and the number of ‘residential’ construction
workers. Our estimates suggest that a 1 percent increase in population growth rates is associated with a 0.4 – 0.5
percentage point increase in the fraction of the workforce in the construction sector. Since regions with zero
population growth have 4.5 – 5 percent of their workers involved in residential construction, each percentage increase
in the population growth rate increases the number of residential construction workers by approximately 10 percent. This
does not include additional workers in related industries such as building materials. Auckland is again an outlier. For
most of the period Auckland had approximately 9000 fewer construction workers than could be expected from trends around
the rest of the country. Clearly, if this shortfall continues it will be difficult for Auckland to overcome its housing
shortage.
ends