NSW Department of Planning and Environment

Projections methodology

Exploring the projection method

Population projections are developed using a range of best practice industry methods. Data sources, including the Australian Bureau of Statistics, Commonwealth Treasury and NSW Health are used in the preparation of the projections. We also release annual current population data to show where our population is at a point in time each year.

For more information about population methodology, read the 2022 NSW Population Projections - Technical Methods Paper (PDF, 522 KB)

Trusted approach

Our approach uses the following best-practice population-projections methods.

Cohort component method

This breaks a population into age and sex and sends each group of people with a common characteristic (each cohort) forward in time. The method makes assumptions about what demographic changes (births, deaths or migration) will happen at each age. This is widely considered the international gold standard for population projections and used by the Australian Bureau of Statistics as well as other statistical agencies around the world.

Our assumptions are based on analysis of historical trends from a range of data sources, as well as any announced policies and local intelligence gained from consultation with local councils.

We have excluded any policies that had not been announced or exhibited at the time of production, or that were on exhibition at the time.

The cohort component approach reflects the fact that certain life events – like having a baby, moving interstate, or dying – are more (or less) likely at different stages of life, and that this can change over time.

Housing unit method

The projections also use the housing unit method for the Greater Sydney Region. Using the Housing Unit Method allows us to anticipate future changes in net internal migration patterns over the projection horizon.

This distributes the projected population over time. It builds in expected housing supply, so population projections reflect greenfield, brownfield and in-fill localities. Using this method to supplement the cohort component method is used widely, including New York City and the City of London, It is useful for the department’s Sydney Housing Supply Forecast.

For the Greater Sydney region, we get the future pattern of housing supply from the Sydney Housing Supply Forecast. Planning professionals also identify more strategic interventions.

Multi regional approach

We apply assumptions to individual age groups independently for NSW, regions and Statistical Areas Level 2 (SA2s).

We then apply assumptions about the population by age and sex, and components of population change for regions, to add up to their equivalents for NSW. We repeat this process so the SA2s add up to the regions.

The model runs them in a sequence, starting with the NSW projection, then the region, then SA2.

Projections cannot add up to more than the state total. This is because it is easier to set assumptions for large populations as there are fewer year-to-year fluctuations. There are also not big statistical responses to small changes at a state level compared to SA2s.

These top-down and bottom-up approaches affect assumptions and outputs. Net interstate migration and net overseas migration assumptions for each region must add up to the state total. Net migration within NSW must add up to zero. If one area has more incoming migrants based on internal migration, then the model will take them from somewhere else.

For births and deaths, assumptions are based on a rate, or the number of births and deaths for each age group. We apply these rates to the population and give a total number of births and deaths, and after that we make the number of births and deaths at each age equal the total. Outputs are limited to the NSW totals in a similar way to the assumptions. We do these so the age profiles in each region or SA2 are based on the starting population and add up to the state numbers.

We have used the SA2 outputs to build projections for local government areas (LGA), which have a different boundary. Unlike the top-down/bottom-up approach for SA2s within regions, this is a top-down approach only. We can combine small geographic areas such as SA2s into broader ones such as Statistical Areas Level 4 (SA4s). Likewise, we combine LGAs into Greater Sydney Commission districts, planning regions and functional economic areas. All outputs, both population measures and components of population change, add up to region and state totals.

Geographic base

The projections are based on Statistical Area 2 (SA2), a geography defined by the Australian Bureau of Statistics Australian Statistical Geography Standard (ASGS) 2020. Outputs are available for SA2 and for local government area (LGA) boundaries as at 30 June 2020. These geography levels nest within each other. Added together, they make up the NSW total population.

SA2s are medium-sized, general-purpose areas built up from whole Statistical Areas Level 1 (SA1s). They are a community that interacts together socially and economically. SA2s are generally the smallest areas used for the release of Australian Bureau of Statistics (ABS) data. SA2s are how data are released by the ABS. This means we are using data in the same format as prepared by the ABS for the NSW population projections.

In some cases, we have combined SA2s to avoid cells with zero or very low values. This happens when an SA2 contains little or no population, such as in the Royal National Park and parts of the Blue Mountains.

LGAs are approximations of local government boundaries, as defined by state and territory governments. Using LGAs ensures that the structure of our projections fit together and are comparable over time.

The names of local government areas are abbreviated in the ASGS structure and include a suffix for the LGA’s status. (C) indicates cities and (A) indicates areas. For instance, the City of Albury is called ‘Albury (C)’, and Armidale Regional Council is called ‘Armidale Regional (A)’.

Base population

Base population

The base (or "jump-off") population at the start of the projection period is the Australian Bureau of Statistics' Estimated Resident Population (ERP) data. This is the official measure of Australia's population.

The ERP is Australia's official population count. ERP counts give a more reliable count of the population than the five-yearly Census as it:

  • takes Census counts of people where they usually live (accounting for interstate visitors and removing overseas visitors)
  • adjusts for the people the Census missed, counted twice or counted by mistake, using the Census Post-Enumeration Survey
  • adds in Australians who are temporarily overseas and applies more demographic adjustments.

For more information, refer to the Australian Bureau of Statistics’ explanation of how it calculates the official population figures.

The base population and population projections estimate the number of people who will usually live in a location, over time. ‘Usually live’ means that the person will be in the location for 6 months or more out of a year. This is the same definition used by the Australian Bureau of Statistics across Census and Estimated Resident Population resources. Babies are counted at the mother’s usual place of residence, not where the birth takes place (usually a hospital that may be in a different town or interstate).

Temporary populations

The NSW projections are for usual resident populations. So not all temporary populations are counted. Most temporary populations counted in the projections tend to live in some form of non-private dwelling arrangements, such as boarding schools and prisons, for example.

Some temporary populations which are not counted in the projections include people on short holidays, and people working longer shifts on mining and other infrastructure projects (typically fly-in-fly-out or drive-in-drive-out workers).

Underlying evidence

The assumptions for the population projections are informed by extensive analysis of a range of datasets:

  • Birth rates are based on historical birth data from the Australian Bureau of Statistics and supplemented with data from NSW, Queensland, Victoria and ACT health agencies’ Perinatal Data Collections.
  • Death rates are based on the Australian Bureau of Statistics mortality data.
  • Interstate and intrastate migration come from historical estimates of interstate migration available from the Australian Bureau of Statistics. Age profiles for migration are informed from the Census question ‘Where did you live 5 years ago?’. We have validated internal migration assumptions by looking at NSW drivers’ licence change of address data. The likelihood of people at each age to move in to, or out of, an area is then applied to future populations.
  • Overseas migration assumptions have been informed by expert advice from demographers and information from the Australian Government Department of Home Affairs and Australian Government Treasury. Overseas migration estimates for small areas come from the Australian Bureau of Statistics. They indicate where people leave from or arrive when moving internationally. Age profiles for both arrivals and departures are determined from Australian Bureau of Statistics data on migration, including the 2016 Census.
Reliability

Projections are reviewed every 5 years against Census results. The margin of error for the projections at 20 years has been:

  • plus or minus 2% at the NSW level
  • plus or minus 3% for regional NSW
  • plus or minus 4% for Greater Sydney.

These results represent a high degree of confidence in the projections. Actual populations are only 2% to 4% above or below the projected population from 20 years earlier.

We updated our approach before developing the 2022 projections. The new model based on small geographies allows us to highlight movements between towns and surrounding rural areas.

As the COVID-19 pandemic has shown, there is some degree of uncertainty associated with all projections.

A low, medium and high series have been produced to show 3 possible future population outcomes, based on what we can observe from the past, what we know about the effect of COVID-19 on our population, and assumptions based on advice from Australia's top demographers about possible futures.

The medium series is the most likely demographic future based on analysis of births, deaths and migration trends. This series forms part of the NSW Government Common Planning Assumptions.

The difference between NSW Population Projections and NSW Government’s strategic plans

The NSW Population Projections are a point-in-time estimate of the future NSW population, based on assumptions for births, life expectancy and migration. We reference data from the the Census, various government departments and short-term forecasts of net overseas migration.

The projections also use known new housing supply and infrastructure investments in Greater Sydney to help distribute the projected future population across the city. They reflect announced policies and projects.

The NSW Population Projections are not NSW Government policy. They are one tool used in the planning for land use, service delivery, infrastructure investment and economic development. The NSW Government’s policy or ‘vision’ for the future of NSW is outlined in its strategic planning documents.

Projections give us a longer-term view of the future that governments can use alongside other information to help inform planning and decision-making.

2022 NSW Population Projections assumptions

The population projections reflect assumptions about births, deaths and migration. We apply these to each five-year age group over the projection period. The assumptions drive the projected population size, rate of growth and age profile.

The assumptions are based on the best available data from the Australian Bureau of Statistics, NSW Health, the Australian Government Department of Home Affairs and Australian Government Treasury.

We set different assumptions for NSW, regions and each SA2. This means an area with university students, for example, has a different assumption to a place with no local university.

Assumptions are informed by historical trends. We assess if those trends are likely to continue or if a different trend should be assumed. While some short term assumptions have been impacted by the COVID-19 pandemic, some aspects such as the age profile of people moving has stayed the same. The assumptions for NSW are shown in Table 2 below.

Assumptions are set at the NSW level, followed by regions and Statistical Area 2 (SA2) in relation to the NSW assumptions. This is called a ‘top-down’ approach, meaning any projections below the state level will add up to the NSW total. All published projections figures are rounded to single digits. In some instances, data table components will not exactly sum to published table totals.

In 2021, we summarised some of the immediate effects of the COVID-19 pandemic on NSW's population.

Learn more about these population changes in NSW.

Births

Two things tell us how many babies are likely to be born in the future: the number of women who are the right age to have children, and how many babies they will have. Our assumptions set a total birth rate for NSW, each region and Statistical Area 2, that shows the average number of babies that will be born to women.

Our assumptions also include the age profile of mothers. In some areas, women have babies at younger ages than elsewhere. When this information is applied to the number of women in each age group, we can project how many babies will be born.

Deaths

Population projections must take away those people who will die during the projection period. To do this, the NSW projections make assumptions about life expectancy at birth, or the average age at which people will die.

Behind assumptions for life expectancy at birth is life expectancy for each age and a measure of the likelihood of surviving from one age group to the next. This means even if life expectancy at birth is 80 years, at age 20 life expectancy will not be 60 years but a higher number. These assumptions let us project how many people will die and at what ages. In areas with more older people, we are likely to see more deaths.

Migration

Three types of assumptions are made for migration, or the movement of people in and out of a place:

  • intrastate migration, where people move within NSW
  • interstate migration, where people move to and from other states in Australia
  • overseas migration, where people move in and out of NSW from other countries.

For intrastate and interstate movements, we consider people have moved when they have lived in a place for 6 months or more.

For overseas migration, we count people as an out-migrant or in-migrant once they have left Australia or been in the country for 12 out of 16 months. This includes Australian-born and overseas-born people. People who leave for work overseas and are gone for more than a year are counted as out-migrants. If they return after that time, they become part of the in-migration flows.

We know that people are not likely to live in the same place that they were born for their whole life, so we make assumptions on the numbers of people moving in and out of an area.

For each migration type there is an assumption about the number of people moving, and about the age at which they move. For example, overseas arrivals are concentrated among people in their 20s. This means the projections assume there are more people in those age groups for particular areas. This then has flow-on effects as they are aged forward in the projection model.

Distribution of urban populations

We use the information about the potential supply of housing for the Greater Sydney region to distribute the projected population over time. The department’s Sydney Housing Supply Forecast informs our assumption for housing completions and our predictions for where future housing development will happen.

Table 2 summarises assumptions used in the 2022 NSW Population Projections.

Table 2. Population projection assumptions 2016 to 2041

 
Item Assumption
ItemBirth rate AssumptionTrending from 1.70 births per woman to 1.63
ItemLife expectancy at birth (Males) AssumptionTrending from 81.8 to 86.0 years
ItemLife expectancy at birth (Females) AssumptionTrending from 85.8 to 88.9 years
ItemNet interstate migration AssumptionAnnual net losses going from −94,000 to −85,000, with smaller losses for 2021–26
ItemNet overseas migration AssumptionAnnual net gains trending from 306,600 to 376,000, with smaller gains for 2021–26
What are population projections and why do we need them?

Population projections provide a picture of the population as it may develop in the future. They provide an indication of local areas likely future population size, and their age and sex profiles, if assumptions about future births, deaths and migration trends are realised. 

An understanding of the size and composition of our future population is essential to making informed planning decisions for the State’s future so we have the right number and type of services in the right areas. The projections are used as a common framework across NSW Government to inform planning policy decisions around infrastructure and service delivery such as the provision of hospital beds, school classrooms, roads and public transport. 

Population projections are not government targets. The projections are open to change as future demographic behaviour can be impacted by government policies, changing economic and social circumstances or other factors (both in Australia and overseas). 

How does changing housing supply impact the projected population?

For the Greater Sydney Region the cohort component method is complemented with a Housing Unit Method (HUM) approach. 

The HUM is used to direct a share of overall population growth to areas that are forecast to see increased housing supply. The HUM is applied to anticipate future changes in net internal migration patterns over the projection horizon. For the Greater Sydney Region, the future pattern on housing supply is derived from the Sydney Housing Supply Forecast with additional strategic interventions identified by planning professionals.