Australia's sluggish population growth: Justin Fabo
GUEST OBSERVATION
Australia’s population growth has typically lagged the economy’s fortunes – people vote with their feet. Rising unemployment typically results in a slowing in net immigration and population growth (typically with a lag), and vice versa (Figure 1).
We are seeing this currently as sluggish growth in activity has resulted in a slowing in population growth. We discuss a few reasons why net immigration and population growth may have already slowed by more than official figures suggest.
Australia’s population growth slowed to 1.4% y/y, on current estimates, at the end of 2014 from a peak of more than 2% y/y prior to the GFC. Lower net immigration has contributed most to the slowing but births have also fallen (see page 4) (Figure 3).
Monthly overseas arrivals and departures data suggest that net immigration slowed further in H1 2015 (Figure 4). In fact, on a sequential basis, population growth already looks to have dipped towards 11⁄4% at end 2014 in annualised terms.
The slowing in population growth has been most pronounced in the mining states of Western Australia, Queensland and the Northern Territory (Figure 5). These states have also seen sharp slowings in 457 visa (Temporary Work (Skilled)) applications which allow recipients to stay in Australia for up to four years. This indicates that slower economic activity and reduced job opportunities in these states has been an important reason for the overall slowing in net immigration. This has taken pressure off unemployment from rising faster than it has in these states.
Historically, swings in population growth have tended to lag changes in economic momentum. To illustrate, population growth has typically slowed as the unemployment rate has risen, and vice-versa (the one exception being the late 1980s where population growth slowed in advance of the economic cycle) (Figure 6).
The slowing in population growth underway therefore largely reflects weaker economic activity. From a labour market perspective, slower population growth can be thought of as ‘shock absorber’. It limits upward pressure on the unemployment rate to the extent that it reflects a slower rise in the supply of available workers. A large share of the slowing in net immigration has been among the prime working age (25- 44 year olds). Fewer arrivals British and other European citizens in this age bracket have accounted for half of the decline in permanent and long-term arrivals to Australia since 2012.
Measuring population growth accurately in real time is difficult and we suspect that it could be revised lower due to downward revisions to net overseas migration (NOM) over time.
The ABS count persons as being in the resident population if they are in Australia for at least 12 months in a 16 month period. The problem is that this cannot be known in real time and the ABS cannot produce ‘final’ NOM and population estimates until 21 months after the relevant quarter.1 In the interim its preliminary estimates of NOM are subject to revision which can sometimes be large (Figure 7). In fact, NOM and population growth have already been revised lower since its most recent peak. This could be due to the way the ABS calculates preliminary estimates of NOM.
To estimate NOM the ABS uses passenger card information combined with propensities from the year before for groups of individuals to ‘NOM in’ or ‘NOM out’. Revisions to NOM can therefore occur if these propensities are changing quickly which may occur if net migration flows are changing rapidly as is currently the case.
Net immigration has already been slowing faster than expected. For instance, ABS estimates of NOM have already significantly undershot the latest estimates/forecasts by the Department of Immigration made in mid 2014, particularly on the departures side (Figure 8). Net immigration from New Zealand has slowed more sharply than expected as unemployment had been falling across the Tasman but rising in Australia (Figure 9).
More important, however, has been the much weaker-than-expected decline in net immigration of persons on Working Holiday and Work & Holiday visas and the slowing in net immigration of workers on 457 visas, both reflecting the softness of Australia’s labour market (Figures 10 and 11).
Finally, net immigration may be weaker than official estimates because published (final) NOM figures have historically overstated net immigration. Given births and deaths can be measured relatively accurately, and total population estimates are provided by the Census every five years, a more accurate estimate of net immigration is calculated as population growth less natural increase. Figure 12 shows that this residual measure of NOM has been lower than published NOM. If history is any guide, therefore, the availability of 2016 Census figures could lead to further downward revisions to NOM and population growth.
In summary, these factors suggest net immigration and population growth have already slowed by more than official figures suggest.
THE OUTLOOK FOR POPULATION GROWTH
Population growth is expected to slow a little further in the near term as lags from slower economic activity feed through. Over the next five years, it seems likely that population growth will average 11⁄4-11⁄2% per annum (the RBA is now assuming 11⁄2% over the next few years). This would be much slower than growth experienced in recent years but still faster than the 1.1% per annum average growth rate from 1992 to 2006. As net immigration slows, nearly half of Australia’s population growth is again likely to come from natural increase, although this is also slowing.
NATURAL INCREASE
The number of births has moderated again in recent years after a brief spurt and is expected to slow further.
The contribution of natural increase – births less deaths – to population growth steadily declined over time until the mid 2000s (Figure 2). More births then supported population growth as the number of 25-34 year old females – those with the highest fertility rates – rose sharply after being little changed for more than a decade (Figure 13).
The number of births and natural increase in the population is already slowing. It will slow further as much weaker growth over the past five years in the number of 20-24 year old females will result in slower growth in the population of 25-34 year old females.
NET IMMIGRATION
Unless there is a persistent improvement in labour market conditions, it is difficult to see net immigration increasing in the foreseeable future. While some further slowing in net immigration is expected in the near term, we expect it to then stabilise due to: (i) fewer outflows to New Zealand; (ii) a still high official immigration programme; and (iii) strong international student inflows.
First, the relative strength of the New Zealand economy, which has resulted in a small outflow of people from Australia across the Tasman, is diminishing. The jobless rate in New Zealand has increased recently amid more moderate economic activity and the increase in labour supply (Figure 14). This could see at least a stabilisation in net immigration between Australia and New Zealand if not some turnaround.
Secondly, the intake under Australia’s official Migration Programme has been kept at 190,000 persons for 2015-16. While this level of official migration in recent years has not prevented net immigration from moderating, it will limit its rate of decline by, to some extent, putting a floor under immigration levels (Figure 15).
Finally, strong growth in international student arrivals is supporting net immigration. As long as students satisfy the ‘12/16’ rule they are counted as residents for population purposes and many stay in Australia after graduation.
The number of international student commencements has reached the previous peak in 2009 when its contribution to net immigration peaked at around 120K persons per annum (Figure 16). Department of Immigration forecasts, which appear broadly on track, suggest a similarly large contribution to NOM over the next few years which would add 0.4ppts to annual population growth. (Students may add less to potential labour supply than other immigrants, however, as they are limited to working 20 hours per week while studying.)
This trend is being supported by the weaker Australian dollar, streamlined policies towards overseas students and rule changes that allow overseas students to extend their stay for two to four years upon graduation under a Temporary Graduate visa.
IMPLICATIONS OF SLOWER POPULATION GROWTH
Potential growth and interest rates
In the medium-term, one implication of slower population growth is lower potential output growth through weaker growth in both demand and supply. Given the aging population, potential labour supply growth would be even slower absent a persistent increase in labour force participation.
Assuming annual productivity growth around its long-run average of 11⁄2%, potential output growth could fall closer to 23⁄4% per annum over the medium term, well below the 3–31⁄4% widely assumed previously. While it’s per capita growth that ultimately determines standards of living, lower potential GDP growth would imply lower interest rates than otherwise (Figure 17).
Housing demand
One source of slower potential output growth is that slowing population growth would, over time, reduce underlying demand for housing. While several years of underbuilding means that there remains excess demand for housing, particularly in Sydney, smaller increases in the population will eventually result in weaker housing demand.
In this regard, relatively small changes in population growth matter for housing demand. For example, annual population growth of 11⁄4% would add around 125k less people per annum to the population compared with growth of 13⁄4% (Figure 18). Assuming 2.5 persons per household, this would mean housing demand drops by roughly 50k dwellings per year.
A complicating factor is that international students are expected to account for a noticeable share of population growth in the near term. As mentioned recently by the RBA, demand for housing from students may be different to that from other immigrants, including because they have lower incomes and are typically younger. They are also likely to demand housing close to universities and apartment living rather than detached housing. The composition of student arrivals, however, may also have a bearing on the type of housing sought.
Justin Fabo is senior economist, ANZ. You can follow him on Twitter here.