Dwelling construction long term

ScreenHunter_22 Mar. 06 18.40

By Leith van Onselen

Following on from today’s post on dwelling completions, I want to take readers through some long-term charts that provide insight into how Australia’s home building industry is travelling at present, and whether the RBA’s goal of dwelling construction filling the void left as the mining boom unwinds is likely to be met.

The first chart plots the rolling annual number of dwelling completions for both houses and apartments against the 30-year average:

ScreenHunter_13 Apr. 17 13.45

According to the ABS, the total number of completions in the year to December 2012 were -3% below the long-run average. Importantly, the employment-sensitive detached house segment was -12% below average, partly offset by unit & apartment construction, which was running 19% above the long-run average.

The story is better for dwelling approvals, however, which have shown a modest recovery in the year to February 2013, suggesting that dwelling construction could soon pick-up (see next chart). That said, total approvals continue to track nearly -2% below the 30-year average, with the employment-sensitive detached house segment remaining highly depressed (-17% below the long-run average), whereas unit & apartment segment is booming (34% above average levels):

ScreenHunter_14 Apr. 17 13.50

One of the more disappointing aspects of Australian dwelling construction is that it has failed to respond to the recent surge in population. As shown below, population growth accelerated from late-2004, whereas dwelling construction – as measured by approvals, commencements, and completions – failed to lift (see next chart).

ScreenHunter_16 Apr. 17 14.05

The failure of the construction industry to respond in any meaningful way to the long boom in house prices post-1996, or the more recent surge in population growth, suggests that housing supply has become increasingly unresponsive to demand. Such unresponsiveness also makes it more difficult for the RBA and policy makers to engineer a pick-up in construction, via interest rate cuts and demand-side stimulus, to offset weakness as the mining-led investment boom winds down.

Ultimately, the only sustainable way of getting construction moving is via supply-side reform and improved affordability through lower land prices.

[email protected]





  1. How do we compare these numbers LHS/RHS in the graph dwelling construction vs population change.

    If population grows 390,000 and there are 250,000 completions then at 2 people per household/dwelling that would mean more than enough supply.

    • Sorry brainfade. Mixing up the figures LHS/RHS here.

      LHS 2012 = 140,000, not 250,000.

      So at 2 people per household that would mean a shortage of (290,000-280,000) = 110,000.

      Question remains the same though. I did not mean to suggest there was no shortage.

      • Occupancy rates of new detached dwellings (greenfield) is usually around three. For new attached (medium density) that drops to around 2.3.

        • Thanks. So it is hard to say what the shortage is without further data.

          The 2012 figures (squinting at the graph here) are roughly 145,000 completions and a population growth of 385,000. At an occupancy of 2.3 that would mean a shortage of 51,500 completions but at 3 it would be a surplus of 50,000.

          Also strange that completions do not always seem to lag population growth.

          It would be interesting to overlay this graph with interest rates and house prices.

  2. Changing demographics of huge existing population require extra dwellings.
    Also extra people require extra dwellings.

  3. It is revealing that this study of the elasticity of housing supply in Sydney finds it to be less than “1”; i.e., inelastic.


    This is the problem.

    The other Australian cities need to be analysed in the same way.

    There is good data for US cities that clearly shows that the price bubble cities had elasticities of below 1, while those that did not have price bubbles, mostly had elasticities of 2 plus. The highest were over 5…..! Kansas is the clear winner for supply elasticity, Texas is just middle of the data set for the supply-elastic States.

  4. By the way, the elasticity of supply in the UK is close to zero; that is, little discernible supply response at all to demand shocks. This makes sense, given that the time required to bring a development proposal to market is longer than one phase of the property cycle.

  5. From chart #3 – the wild swings in population growth (particularly since 2004) clearly have an undesirable destabilising effect on the housing market.

    Why do we manage pop growth to swing so violently?

    Surely one essential part of prudent housing market planning and management is to reduce the volatility of this entirely controllable flow.

  6. At least the NSW government has recognised the nature of the problem and made a start on doing something about it. Probably nowhere near enough, but it will be interesting to see the results.