|Page tools: Print Page Print All RSS Search this Product|
The Australian Bureau of Statistics uses the cohort-component method for producing population projections. In this method, assumptions made about future levels of fertility, mortality, overseas migration and internal migration are applied to a base population (applied by sex and single year of age) to obtain a projected population for the following year. The assumptions are then applied to this new (projected) population to obtain a projected population for the next year. This process is repeated until the end of the projection period is reached.
Span of projections
From a base of 30 June 2017, the projections span the period 30 June 2018 to 30 June 2066 for Australia, states, territories, capital cities and rest of state. Estimated resident population for 30 June 2017 for all above mentioned geographies have also been included in the data.
The base population is the preliminary estimated resident population at 30 June 2017, which takes into account the 2016 Census of Population and Housing.
SUMMARY OF ASSUMPTIONS
Assumptions have been formulated on the basis of demographic trends over the past decade and longer, both in Australia and overseas, in conjunction with consultation with experts at the national and state/territory level. They do not specifically attempt to allow for non-demographic factors (such as major government policy decisions, economic factors, catastrophes, wars, epidemics or significant health treatment improvements) which may affect future demographic behaviour or outcomes.
As future levels of fertility, mortality, overseas migration and internal migration are unpredictable, two or more assumptions have been made for each component and projections have been produced for all combinations of the assumptions. These are intended to illustrate a range of possible future outcomes, although there can be no certainty that any particular outcome will be realised, or that future outcomes will necessarily fall within these ranges.
The table below shows how recent demographic trends (an average of the last three years) relate to the proposed assumptions. The projections will show a smooth transition from the most recently observed data to the long-range assumption. This 'phase-in' period is different for each component assumption and so the table also shows the year that each assumption will be phased in by.
Detailed information on the assumptions can be found in the datacube 'Projection Assumptions (detailed)' located in the downloads tab.
The above assumptions can be combined to create 54 sets of population projections. Three series have been selected from these to provide a range, although not the full range, of projections for analysis and discussion. These series are referred to as series A, B and C.
For some states, series A and C do not depict the highest or lowest population outcomes. Where applicable, other series have been included in commentary.
The inclusion of a zero net overseas migration assumption increases the total number of available projections to 72 series. These extra series (series 55 to 72) do not feature in the commentary and analysis but are included in the ABS.Stat datasets attached to this publication.
WHICH SERIES TO USE
Future uncertainty, along with the subjective nature of assessing current trends, means that using a range of possible outcomes rather than a single projection series give a more realistic view of the possible future size, distribution and age structure of Australia's population.
Different series, constructed from varying combinations of assumptions, are appropriate for different time horizons (shorter or longer term), the geographic region(s) considered, and any volatility in the components. Historically, mortality and fertility have been consistent with slow-moving trends in the data. The projections reflect this. Observed levels of overseas and interstate migration have been far more volatile. This volatility can be expected to continue due to future government policies and decision making, and economic, social and other influences in Australia and overseas.
The following table presents the 72 permutations of the population component assumptions, with series A, B and C specifically identified.
Future trends in fertility are an important determinant of Australia's future population size, structure and growth. To produce population projections using the cohort-component method, assumptions of age-specific fertility rates and the sex ratio at birth are required for each year of the projection period.
Using data from the past 20 years, three long-term assumptions have been made regarding Australia's future total fertility rate (TFR): higher fertility (a TFR of 1.95 babies per woman), medium fertility (1.8) and lower fertility (1.65). Under all three assumptions, the trend towards older ages of mothers is assumed to continue to 2027, but at a slower rate than seen historically, and remain constant thereafter. For all years, the sex ratio at birth is assumed to be 105.5 male births per 100 female births.
Trends in the total fertility rate
In 1961, at the height of the 'baby boom', Australia's TFR peaked at 3.5 babies per woman. Since then fertility has declined, falling sharply during the early 1960s, before levelling out at around 2.9 babies per woman in the years 1966–1971. The TFR was last at replacement level (2.1) in 1975, and continued to fall thereafter. Fertility stabilised somewhat during the 1980s, before resuming a more gradual decline during the 1990s. The TFR reached a low of 1.7 babies per woman in 2001, increased again to 2.0, before declining to 1.8 in 2016.
Assumed total fertility rates
The three assumptions for Australia's future fertility levels are made with regard to recent trends in the TFR, especially those of the past 20 years.
The higher fertility scenario assumes that Australia's TFR will reach 1.95 babies per woman by 2027 and remain constant thereafter. This reflects levels of fertility recorded since 1977 of between 1.7 and 2.0 babies per woman, acknowledging the possibility that the TFR could increase more, especially in the short-term.
The medium scenario assumes a continuation of the current TFR, with the TFR to remain steady at 1.8 to 2027 and remaining constant thereafter.
Under the lower fertility assumption the TFR declines to 1.65 babies per woman by 2031 and remaining constant thereafter. Fertility rates have reached low levels in many European countries, and higher-income Asian countries such as Singapore, Japan and the Republic of Korea.
Footnote(s): (a) Babies per 1000 woman. (b) Calendar year rates calculated using births by occurrence, adjusted for registration lag. This differs to the TFRs published in Births, Australia (cat. no. 3301.0) and Australian Demographic Statistics (cat. no. 3101.0).
Source(s): Population Projections, Australia, 2017 (base) - 2066
State/territory and capital city/rest of state observed fertility
The table below shows the TFRs for all states and territories and Australia from 2007 to 2016. Some states have consistently been higher or lower than the national rate, while others have fluctuated over the past 20 years. In recent years, TFRs for Victoria, South Australia and the ACT have been lower than rates for Australia as a whole, while TFRs for the remaining states and territories, particularly Tasmania and the Northern Territory, have been higher.
The ratio of each state and territories' average TFR for the ten years 2007–2016 to that of Australia is calculated, then applied to assumed future Australian TFRs. These ratios remain constant throughout the projection period.
Assumed TFRs for the capital cities and state balances are derived by applying the average ratio (for 2007–2016) of the region to its respective state/territory to that state/territory's assumed TFR. Fertility rates for Australian capital cities are typically lower than rates for their respective states and territories, while rates for state balances are higher.
Age-specific fertility rates
Population projections require assumptions about future age-specific fertility rates, which are derived from assumed TFRs and age distributions of fertility. These rates are applied to the projected female population in each year of the projection period in order to determine future numbers of births, and therefore the size of future projected populations.
Over the past 10 years, age-specific fertility rates have been declining for the younger age groups (women below age 30), whilst increasing among women aged 30 years and over, representing a gradual shift in fertility towards older ages.
The projected age distribution of mothers is based on half the average rate of change in the age-specific fertility rates during the period 2012–2016. The historical rate of change is assumed to slow down due to limits on child-bearing ages. These trends are assumed to continue under all three fertility scenarios until 2027, after which the age pattern of fertility remains constant.
Linear interpolation is used to obtain TFRs for each year 2017 to 2026 for all three scenarios, using the known TFR for 2016 and assumed TFR for 2027. To create assumed age-specific fertility rates, the assumed distribution of age of mothers is then applied to the assumed TFR for the corresponding projection year.
Footnote(s): (a) Babies per 1,000 woman
Source(s): Population Projections, Australia, 2017 (base) - 2066
Sex ratio at birth
Projections require an assumed sex ratio at birth (the ratio of male to female births), so that total projected births can be split into male and female births.
Historically, the sex ratio fluctuates between 105 to 106 male births per 100 female births. A constant ratio of 105.5 male births per 100 female births has been used for the duration of the projection period.
For the population projections in this issue, two assumptions on future life expectancy at birth have been made. Only two assumptions have been made because life expectancy has consistently shown an improving trend since Australian records began.
The higher life expectancy at birth assumption assumes that life expectancy will continue to improve at the average rate observed in 2012–2016. The medium life expectancy at birth assumption assumes that life expectancy will also improve, but at a declining rate.
Life tables for 2015–2017 were not created in time to be included in the mortality assumption process.
Trends in life expectancy
Australian life expectancy at birth has improved steadily for both men and women. Since the early 1920s, life expectancy at birth for both males and females has increased by about 21 years. Since the 1980s, faster increases for males has narrowed the gap between male and female life expectancy at birth, from 7 to 4 years. Recent years has seen the improvement in life expectancy declining for both males and females, with both males and females recording an improvement of 0.06 years in 2014–2016.
Source(s): Population Projections, Australia, 2017 (base) - 2066
Assumed life expectancy at birth
The higher life expectancy assumption assumes male and female life expectancy at birth will continuously increase at the average growth of 2012–2016. The yearly increase of life expectancy of birth for the higher assumption is about 0.14 for males and 0.09 for females. Based on this assumption, male life expectancy at birth reaches 87.68 years in 2066 and female life expectancy at birth reaches 89.16 years.
The medium life expectancy assumption assumes male and female life expectancy at birth will increase at a slightly lower rate than the higher assumption for the first year, and then gradually slow. Based on this assumption, male life expectancy at birth would reach 83.00 years and female life expectancy at birth would reach 86.00 years in 2066.
Age-specific death rates
The inputs of the mortality component of population projections are 'survivorship ratios' obtained from assumed future life tables. Life tables for each year in the projection period (i.e. 2018–2066) are calculated in two steps: (1) life expectancy at birth for each projection year is determined; and (2) a life table is generated which gives the desired life expectancy at birth and allows for a shift in the age curve of mortality over time.
The shifting age curve of mortality over time should ideally represent current trends in age-sex differences continued into the future. To achieve this, rates of change indicative of recent trends for each age-sex group are incorporated in the production of the assumed life tables. Determining assumed rates of change is achieved by observing historical patterns in age-specific death rates.
Between 2006 and 2016, males aged 0–29, males aged 65–79 and females aged 0–9 had the largest relative declines in age-specific death rates. The groups with the least improvement in death rates were males aged 40–49, males aged over 95, females aged 35–49 and females aged over 90. In general, greater change over time has been observed for males compared to females.
These trends are reflected in the assumptions. The assumed rates of change of age-specific death rates over time continues to 2035–36, after which the age-specific death rates are proportionally adjusted to fit the assumed life expectancy at birth.
Assumed age-specific mortality rates
Age-specific mortality rates are assumed to decrease for all age groups for both males and females over the projection period. Very little change is assumed for ages over 90. For corresponding ages, mortality rates for females are always assumed to be lower than for males.
Assumed state/territory and capital city/rest of state mortality
Assumptions of life expectancy at birth by state/territory are derived from the Australian assumptions by applying a ten year average of the ratio of the state or territory life expectancy to the Australian life expectancy. A similar process is used to create the capital city/rest of state assumptions, by using the ratio of capital city/rest of state life expectancy to the state life expectancy for the most recent life table period 2014–2016. These ratios are calculated separately for males and females and remain constant through the projection period.
NET OVERSEAS MIGRATION
Three assumptions have been made about Australia's future levels of net overseas migration (NOM):
In addition, a zero net overseas migration assumption has been included to facilitate analysis of the effect of overseas migration on Australia's future population.
ABS projections have used forecasts produced by the Department of Home Affairs up until 2021–22. Beyond 2021–22, the NOM assumption reflects established trends observed in the period 2008–2017. All NOM assumptions are held constant from 2026–27 onwards.
Annual levels of NOM have fluctuated considerably in Australia over the past 20 years. For financial years, the level has been as low as 80,000 in 1997–98 to a high of 300,000 in 2008–09.
Footnote(s): (a) NOM estimates include a break in series at September 2006
Source(s): Population Projections, Australia, 2017 (base) - 2066
New South Wales was the largest gainer of NOM from 2008 to 2017, with Victoria being the second largest gainer. NOM for Queensland and Western Australia declined after those states recorded strong NOM gains at the start of the period. Tasmania, the Northern Territory and the Australian Capital Territory have seen fluctuations in their NOM between 2008 and 2017, with Tasmania and the Northern Territory recording strong gains in 2017.
Assumed state/territory and capital city/rest of state share of net overseas migration
Previously, overseas migration data was not available below the state/territory level which meant an indirect method was used to calculate the capital city/rest of state level. With the release of sub-state population component data in the ABS publication Regional Population Growth, Australia (cat. no. 3218.0), regional overseas migration estimates have been used to calculate the capital city/rest of state for overseas migration.
Each state and territory's proportion of NOM is based on an average of the last ten years of NOM data. The table below shows the assumed state/territory net overseas migration distribution:
Assumed age structure of net overseas migration
The assumed age/sex structure of NOM for the states and territories is derived from the 2015–2017 NOM. Overseas migrant arrivals and departures by state/territory, age and sex are simultaneously constrained to the total assumed NOM level for Australia and to the assumed state/territory shares of NOM. The assumed age/sex structures are held constant throughout the projection period.
NET INTERSTATE MIGRATION
Interstate migration is a volatile and largely unpredictable component in population estimation or projection. The movement of people between the states and territories of Australia is influenced by many factors such as varying economic opportunities, overseas immigration and settlement patterns, lifestyle choices and marketing campaigns targeting interstate movers by state/territory governments. As the effect of these factors cannot be anticipated, past net interstate migration trends are used as the basis for assuming future levels.
Net interstate migration (NIM) estimates since 1 July 2007 are shown below. These are calculated using Medicare change of address records, defence data and Census data on usual residence one year ago and five years ago.
Victoria has returned to net interstate migration gains after earlier losses, overtaking Queensland in the last four years to be the largest recipient of net interstate migration. Western Australia has experienced a turn around from gains up until 2012–13 to increasingly larger losses. New South Wales and South Australia have continued to experience net interstate migration losses, not recording a net interstate migration gain in the last 38 and 26 years, respectively.
State/territory and capital city/rest of state assumptions
Levels of assumed net interstate migration were derived by analysing trends over the past ten years and constraining them such that they sum to zero. The assumptions reflect the view that each state/territory will trend towards their short term average.
Three assumptions have been made about future net interstate migration levels:
The medium series assumptions are based on NIM averages for the states and territories in the period 2008–2017. The large and small assumptions are based on minimum and maximum share of state values of observed arrivals and departures over the last ten years, with adjustments made based on the trend data, as well as ensuring that each sum of the state/territory NIM is zero.
It should be noted that for some states the large interstate flows assumption corresponds to large net interstate migration losses, therefore the small interstate flows assumption will yield greater population growth in such cases.
All assumptions are separated into arrivals and departures for each state/territory and capital city/rest of state (Greater Capital City Statistical Area). Rates for arrivals and departures for the states and territories are generated from movement data from recent Censuses to obtain age/sex levels. Further, 2011 and 2016 Census data are used to generate age/sex arrival and departure levels for each capital city/rest of state. As a result, all age/sex arrival and departure disaggregations sum to the net internal migration assumptions.
These documents will be presented in a new window.