The NSW Population Projections User Guide is for people using NSW population projection data to understand how the projections are prepared and how to use them. It answers common questions about projection data, the methods used to prepare the population projections, their interpretation and use.
Published: September 2016
Population projections provide a picture of the population as it may develop in future years. They provide an indication of the size and age-sex structure of the future population if specified assumptions about future fertility, mortality and migration are realised.
Population projections are not forecasts and do not attempt to predict the impact that future government policies, changing economic circumstances or other factors (whether in Australia or overseas) might have on demographic behaviour.
The NSW Government has released 2016 population projections for NSW, Sydney and Regional NSW, local government areas (LGAs) and local health districts (LHDs). The projections provide a picture of the size and structure of the population in NSW and its LGAs from 2011 to 2036.
This projection release follows a revision of all assumptions used to produce the projections based on the most up-to-date information from the Australian Bureau of Statistics, the Department of Immigration and Border Protection, and input from stakeholders. Data reflecting NSW Government announcements for urban renewal projects or greenfield site developments in Sydney and other towns and cities are also included.
Population projections showing the impact of different population change scenarios have also been released. These show the impact of higher or lower growth on population size and age structure.
The population projections are updated every two to three years.
A major review is undertaken when results are available from Australia’s five-yearly Census. This includes establishing a new jump-off population a review of the assumptions about the drivers of future population change. This review pays particular attention to migration patterns.
At least one other review of the projections is done between each Census. The intercensal review focuses on changes to births, deaths, and international, interstate and intrastate migration.
Population projections give the NSW government an estimate of the future population size and composition across all NSW local government areas (LGAs). They are the only set of projections prepared across all LGAs using the same method and base data. The projections are used as a common framework to inform:
Projections are used for planning for services and infrastructure across the state and help to ensure resources are targeted.
No. They are based on assumptions by demographers that take into account recent and current trends for births, deaths and migration. The projections show what is likely to happen in the future if the assumptions hold true.
The projected population numbers and the distribution of the population can change due to economic or political factors, new technology and social attitudes to different living arrangements. They cannot be foreseen with any degree of certainty.
Large amounts of data are needed on the drivers of population change, including births, deaths and migration. Compiling relevant statistics and factoring in change over time is important to developing the assumptions.
Once assumptions are set, they are used to calculate the projections from the jump-off year. Outputs are reviewed and validated. Projection outputs, including total population size and growth rates, are calculated once the projections have been run and the numbers validated.
These projections were produced using the Department’s cohort-component population projection model. The cohort-component method takes a population broken down by age groups and moves them forward in time. In each five-year period we make assumptions about:
The cohort-component method is the most common projection method used by demographers. The word cohort refers to the splitting up of the population into age groups. The word component refers to the three elements of population change – births, deaths and migration – that act together to affect the size of a cohort in the future.
Assumptions are critical for the cohort-component method. If changes are made to the assumptions about births, deaths and migration, the result will be a different projected population at the end of the projection time interval.
A multi-regional model produces projections for multiple regions in a single run. Projections for seven projection regions and all local government areas (LGAs) are produced at the same time as the state-level projections.
A top-down approach is used with projection assumptions and outputs having to add up to the state totals. This is called constraining and is different to projection models that produce projections for one small area at a time and add up the outputs to get a region or state total.
The top-down approach also means that interstate and overseas migration is modelled first at the state level. Interstate, intrastate and overseas migration are then modelled for projection regions with assumption totals not exceeding the state totals. Lastly, LGA migration flows are modelled for all internal and overseas migration and constrained to the projection region totals.
Projections for other regional boundaries such as Local Health Districts (LHDs) are based on LGA projections.
The Housing Unit Method is applied for areas where new housing developments and the timing of public infrastructure construction are known and committed. For areas where the population impacts of an infrastructure project are unknown or speculative, no changes are made to the assumptions for births, deaths or migration. Close monitoring of these components takes place so any impacts that may occur can be reflected in future reviews.
This is a tool used to distribute the projected population of the metropolitan region (Sydney, Central Coast, Illawarra and Lower Hunter) into the local government areas (LGAs) within its boundaries.
The population is first projected for the metropolitan area as a whole using the standard cohort-component model. Then the Housing Unit Method distributes growth into each LGA based on the anticipated future number of dwellings in that area, with LGA assumptions for births, deaths and migration applied. LGAs with more anticipated dwellings will have more projected growth than areas with fewer anticipated future dwellings. The method has been developed for areas where there is significant population growth and where housing supply is restricted by land availability and other related issues. It is also good for areas where new housing developments (particularly on the fringe of urban areas) will dramatically change population growth patterns.
The NSW population projections use a top-down approach working down from the state, to region, to local government area (LGA). State and region projections have not been affected by changes to the LGA boundaries. LGA projections for the new boundaries have been created by adding together projections where whole Councils have been amalgamated. Where boundaries have been changed to create new Councils, populations were distributed based on information about housing completions and short-term housing supply forecasts.
The 2016 projections are for new Councils that have been announced. Data for new Councils that are pending will be released when they become official.
The projections are based on (or jump-off from) the 30 June 2011 Estimated Resident Populations (ERPs) supplied by the Australian Bureau of Statistics (ABS).
The ERP is the official measure of the Australian population, based on usual residence. It refers to all people who usually live in Australia, except foreign diplomatic personnel and their families. The ERP includes usual residents who are overseas for less than 12 months but excludes overseas visitors who are in Australia for less than 12 months.
The 2011 ERPs are derived from 2011 Census counts of usual residents and are adjusted for:
Since 2011, preliminary ERPs and preliminary estimates of what has caused populations to change (births, deaths and migration) have been released by the ABS based on information from:
The most up-to-date data available at the Local Government Area (LGA) level are included in the 2016 revision of the NSW population projections.
The births (or fertility) assumptions are based on data that use the usual residence of the mother, not the place where a birth takes place. Two data sources are used:
1. The NSW perinatal data collection from NSW Health. This data set records information about all births in NSW. Births to NSW mothers that take place interstate in Victoria, ACT and Queensland are also used. Perinatal data include information about birth order. This helps understand the timing of first births and spacing to subsequent births. Timing of first births is an important predictor of life-time fertility levels.
2. ABS data based on information from birth registrations. These include all births registered throughout Australia to mothers who are usual residents of NSW. This information has been used to set the age profiles for childbearing across all local government areas (LGAs).
Data on dwelling completions published by the Department were used for the period 2011-2015. For the period 2016-2021, short-term housing supply forecasts based on known current activity and land zoned for housing have been used (Department of Planning and Environment, unpublished data set). This includes dwellings where construction has started or development approvals are in place.
The Department of Immigration and Border Protection (DIBP) forecasts net overseas migration (NOM) for NSW, Sydney and Balance of State for a four year period on a quarterly basis. These forecasts are the starting point for the overseas migration assumptions. The DIBP forecasts are based on the latest data on visa grants, past behaviour of NOM flows by visa group, existing migration policy decisions and official economic outlooks.
Beyond the four-year forecast period the permanent migration policy position of the Australian Government is used to set NOM. That is the total number of permanent migration places set by the Minister for Immigration and Border Protection. These places include the skilled, family and refugee streams. Extra places are allowed for New Zealand migrants who do not need a visa.
The distribution of NOM assumptions across each local government area (LGA) in NSW is based on 2006 and 2011 Census data and the DIBP settlement database. The settlement database provides information about the intended place of residence of people granted a permanent or provisional visa. Census information tells us about people who were living overseas five years prior to the census and their place of residence on Census night. The average from both data sources was used to show the share of overseas arrivals that go to each LGA in the state. The NOM for each LGA is adjusted so that totals for Sydney and Balance of State correspond to the regional forecasts by the DIBP.
Interstate migration assumptions are informed by the 2006 and 2011 Censuses as well as annual interstate migration estimates published by the Australian Bureau of Statistics (ABS).
The Census question asking respondents where they were living five years ago gives an indication of how many people have moved during the five-year period (but not the total number of moves in that period).
The Census data also tells us the age of people who move. This lets us calculate the likelihood that someone of a given age will move into or out of an area. These age profiles are applied to all migration flows.
The annual interstate migration estimates are based on change of address notifications from Medicare Australia and the Department of Defence (for military personnel).
Assumptions about the number of moves into and out of each local government area (LGA) has been informed by the 2006 and 2011 Census and annual Regional Internal Migration Estimates published by the ABS.
The Census question asking respondents where they were living five years ago gives us an indication of how many people have moved during the five year period (but not the total number of moves within that period).
The Census data also tells us the age of people who move. This lets us calculate the likelihood that someone of a given age will move into or out of an area. These age profiles are applied to all migration flows.
The Regional Internal Migration Estimates are based on change of address data from Medicare Australia and the Department of Defence (for military personnel).
Life expectancy at birth assumptions are set using analysis of deaths data collated by the ABS from death registrations. These include all deaths registered to people who were usual residents in NSW.
Throughout the projection period (2011 to 2036), the numbers of deaths for each age group are based on information from the Human Mortality Database (www.mortality.org). The Database is an international collaboration created to make detailed mortality and population data available. NSW mortality rates are very close to those for Australia as a whole so the national data informs these projections.
Population projections are derived from an assessment of historical, current, and likely future trends in births, deaths, and migration. These are the three components of population change. Assumptions about future fertility (births), mortality (deaths), and migration are formulated after analysis of short-term and long-term historical trends, government policy, information provided by local planners and other relevant information. Assumptions are set first at state level and used as a constraint for assumptions set for regions and local government areas (LGAs). This top-down approach means sub-state projections must always add up to the state totals.
The assumptions used for the 2016 NSW population projections are:
Assumptions about future fertility (births), mortality (deaths) and migration are formulated by the Department of Planning and Environment after analysis of best available data. The assumptions are discussed with and endorsed by the Population Projections Group, a whole-of-government technical and advisory body.
Assumptions about potential housing supply are based on short-term housing supply forecasts to 2021, and estimates of housing supply capacity from 2021 to 2036. Housing supply assumptions are used across metropolitan areas (Sydney, Central Coast, Illawarra and Lower Hunter). Data on housing approvals for 2011 to 2015 are included. Short-term housing supply forecasts based on known current activity and land zoned for housing were used for 2016 to 2021. Housing supply capacity was determined by the Department’s Regional and Urban Planning Branch based on announced major projects and local developments that will affect housing supply.
Births are one of the two demographic factors that make a population grow. An upswing in NSW birth rates over the last decade is assumed to continue into the future, with the Total Fertility Rate (TFR) rising to 1.96 births per woman during 2031-2036. Differences across the state in the average number of children that women have are also taken into account. In Sydney, for example, women have fewer children on average than women in regional or rural areas. These differences are assumed to continue into the future. Assumed TFRs are underpinned by the age profile of childbearing as indicated by age-specific fertility rates.
NSW fertility rate trends are consistent with those for the whole of Australia. The post-war Baby Boom, when women were young mothers and had three children on average, was followed by steep fertility declines during the 1970s. Fertility rates slowly declined further over the next two decades. By the beginning of the 21st century NSW women had fewer than two children on average. Since 2004, fertility rates have increased (Figure 1, Births Section).
The main factor affecting the annual total fertility rate has been the age at which women have their children. The increase in fertility rates from 2004 was unexpected. It began when a greater proportion of births occurred to women in their 30s. Recent increases in fertility rates are considered unlikely to continue because there is only a finite age that childbearing take place. Additional analysis was also done based on a measure called the tempo-adjusted total fertility rate. This is a measure designed to remove the impact of the changing age that women have children. Analysis for NSW showed relative stability in the state-level fertility rate over time once the effects of age were removed.
Looking at recent generations of NSW women, once they begin childbearing there is little difference in the spacing or the number of births that they have. The average length of time between the first and second birth has been consistently around 2.5 years (see Figure 2, Births section).
NSW birth assumptions are made on the basis that:
Assumed fertility rates for NSW are shown below:
|Period||Assumed total fertility rate|
Local government area (LGA) birth rates are set in relation to the state assumptions. The relationship between the state and each LGA is held constant for the projection period. This means if an area has a higher birth rate than the whole of NSW at the start of the projection period, it is assumed the birth rate will remain higher. A ratio was calculated between each region and NSW using Australian Bureau of Statistics (ABS) births data 2006-2011. This ratio was then applied to the projected total fertility rate. A similar approach was used to set the age profile of childbearing using the same ABS data.
The Australian Bureau of Statistics (ABS) released projections in November 2013 for Australia, the states and territories. The fertility assumptions used by the ABS are lower than those used by the Department for the 2016 NSW population projections. The ABS assumes the total fertility rate in NSW will decline from 1.93 births per woman in 2013 to 1.79 births per woman by 2026, and then remain constant.
The fertility assumption set by the ABS is based on recent trend data. It reflects a move to older childbearing but assumes a return to a lower overall number of children per woman. The ABS assumption is not informed by analysis of parity data (information about birth order) which shows a relatively stable number of births per woman over her life time in recent years.
This means the 2016 NSW population projections show a larger number of births over the next 20 years compared to the ABS projections. This has flow-on effects for population growth over the projection period when the projected babies start school and enter early adulthood.
The 2016 population projections take into account the increasing longevity of the NSW population. A baby born in 2011 is expected to live, on average, to 81.0 years for males and 85.0 years for females. By 2036 this may have increased to 85.2 years for males and 88.2 years for females. People living in metropolitan areas currently live longer than people in regional NSW and this pattern is expected to continue.
The projection model uses measures of life expectancy to remove people at certain ages due to death. Life expectancy is an estimate of how many years on average a person can be expected to live when they reach a specified age. Life expectancy at birth is how we summarise the range of deaths that happen at different ages throughout life. It allows us to make population adjustments at all ages throughout the projection period to take account of death rates.
Deaths are one of the most stable component of population change because changes to death rates over time are relatively slow. Historical trends are a good guide to the possible future trajectory of deaths in NSW. People are now living longer than ever before. Male and female life expectancy at birth has steadily increased since 1975, with gains for males slightly greater than for females (see figure below). These increases are because death rates across all age groups have declined.
Trends in death rates and life expectancy have followed a fairly steady path. In the short term, future life expectancy increases can be modelled with some confidence. There is debate about the future of human life expectancy in the longer term, particularly if there is an upper limit. The optimistic view is that gains in life expectancy show no sign of slowing and will continue to increase due to continued improvements in preventative and curative medicine. Another view says that there is a natural limit to life expectancy and that recently emerging health trends such as obesity together with rising atmospheric CO2 levels may slow life expectancy improvements or even lead to decline. Current data do not indicate that the trend of improving life expectancy at birth will reverse, so this perspective is not reflected in the NSW population projection assumptions.
For NSW as a whole, age-specific death rates have been extrapolated to calculate future life expectancies. It is expected that females will continue to have higher life expectancy at birth than males, but males are expected to make life expectancy gains faster than females. On this basis it is assumed that by 2036-41 male life expectancy at birth in NSW will reach 86.1 years and females will reach 88.8 years.
Regional scaling factors, based on regional life expectancies published by the Australian Bureau of Statistics (ABS) for the period 2007-11, have been applied to the state-level assumptions to produce life expectancy at birth assumptions for Local Government Areas (LGA). These factors adjust the NSW-level assumptions up or down depending on local mortality trends. While there are local variations within each region, these are relatively small and regional life expectancies are a good composite of local area differences.
The Australian Bureau of Statistics (ABS) released projections in November 2013 for Australia, the states and territories. The ABS life expectancy assumptions for NSW match those set for Australia as a whole. They show male life expectancy at birth of 79.8 years in 2011 rising to 85.2 years by 2061, and female life expectancy at birth rising from 84.2 to 88.3 years. This is slightly lower than the assumptions set for the 2016 NSW population projections. These variations result from slightly different methods of applying the continued decline in mortality rates at all ages. For the 2016 NSW population projections, the continued mortality rate trends were used to calculate projected life tables and to determine life expectancy at birth. In contrast, the ABS method continues mortality rate trends to 2031 and after that age-specific mortality curves are fitted to the assumed life expectancy.
Net overseas migration (NOM) is the difference between the number of immigrants arriving in NSW and the number of emigrants leaving. It is a major contributor to NSW population growth. Recent high levels of net overseas migration are assumed to continue, gradually increasing from an annual net gain of 67,300 migrants in 2015-2016 to reach a long term annual net gain of 73,950 migrants by 2036-37.
Net overseas migration (NOM) is an important input to the calculation of official population estimates. For this purpose, the Australian Bureau of Statistics [ABS] defines an international migrant as a ‘usual resident’ if they live in Australia for at least 12 months over a 16 month period, or for emigration, a usual resident lives outside of Australian for at least 12 months over a 16 month period; this is widely known as the 12/16 rule. The12/16 rule has been in place since the September quarter 2006. Prior to this, an international migrant had to reside in Australia or be outside of Australia for 12 months continuously to be defined as an immigrant or emigrant respectively.
International migration flows are highly variable due to a large number of factors that impact on migration decisions. These include government migration policies, domestic and global economic conditions and the global economic outlook. Historically, NSW has received the highest proportion of migrants into Australia compared to other states and territories. Throughout the 1970s, 1980s and 1990s, NSW received over 40 per cent of all net overseas migration (Figure 1). This share fell to approximately 30 per cent by the mid-2000s. This decline may be attributable to the resources boom, which created better employment opportunities in the resource rich states of Queensland and Western Australia. This labour-intensive phase of the resources boom has come to an end in recent years and NSW’s share of Australian NOM has been increasing again. NSW’s share of Australia’s NOM has been upwards of 35 per cent since 2014.
More recently, NSW NOM has fluctuated considerably between a high of 96,400 people in 2008 to a low of 50,000 people in 2010 (Figure 1, NOM section). This variability makes future levels of NOM difficult to predict.
The current projection model uses ABS estimates of net overseas migration for NSW from June 2011 to June 2015. For the rest of the projection horizon, the model requires assumptions about the annual level of NOM for NSW and the age profile of immigration and emigration flows. The following assumptions are used:
The Australian Bureau of Statistics’ (ABS) latest population projections published in 2013 applied DIBP forecasts available at the time for the short-term projected period through to 2016/17. From 2017/18 the ABS assumed annual NOM of 66,200 people for NSW. This is slightly lower than the long-term assumption used for the Department’s population projections. The difference is in part because the ABS set an Australian NOM first and then distribute it to all of the states and territories, while the 2016 NSW population projection assumption is based on NSW data.
Net interstate migration (NIM) is the difference between the number of people moving to NSW from other places in Australia, and the number leaving NSW for other places in Australia. The longstanding trend of net interstate migration loss is expected to continue for NSW, gradually increasing from an annual net loss of -10,300 persons in 2011-2016 to reach a long term annual net loss of -17,000 people by 2026 and onwards.
The most unpredictable and also the most critical demographic process affecting population change in NSW is migration – this includes the movement of people between NSW and other parts of Australia. The large net gain of migrants to NSW from overseas is offset, in part, by the continued net loss of people from NSW to other parts of Australia. The majority of interstate moves are between the more populous metropolitan areas.
NSW has a well-established pattern of interstate migration losses to other parts of Australia. Interstate migration estimates published by the Australian Bureau of Statistics (ABS) show that net interstate migration losses from NSW have ranged between a small net gain of 2,000 (in 1977/78, 1979/80) to a peak migration loss of almost -38,000 (1988/89) (Figure 1, NIM section).
The net results shown in Figure 6 hides considerable variation in the flows of people into and out of NSW (Figure 2, NIM section). Annual variations in the number of people leaving the state are the key drivers of the irregularity in the overall level of interstate migration loss from NSW. These flows are dominated by movements between NSW with Queensland.
Data from the last six Censuses show that net interstate loss from NSW during 2001-2006 may have been a singular extreme. The net loss during 2006-2011 is more consistent with the usual level of interstate flows seen since the early 1980s. As was shown in Figure 2 (NIM section), over the last two financial years from 2013-2015, NSW’s estimated NIM loss has been at a record low. This unusually low NIM loss aligns closely with changes in interstate migration in Western Australia and Queensland at the end of the mining boom in 2013/2014.
The current projection model uses ABS estimates of interstate migration for NSW from June 2011 to June 2014. For the rest of the projection horizon, the model requires assumptions about the annual level of NIM for NSW and the age profile of interstate migration flows:
1. Short term assumptions for 2014-2016:
A linear forecast was applied to interstate in-migration and out-migration for the 10 fiscal years from 2005 to 2014. The balance of the in- and out-migration for the years 2014-15 and 2015-16 was used as the assumption for those two years. This means a NIM loss of -51,600 people for NSW over the five-year period 2011-16, or an average of -10,300 per annum.
2. Long term assumptions for 2016-2041:
The record lows of NIM loss estimated for 2011-2016 period are unusual. They correspond to the winding down of the mining boom, which impacted interstate migration patterns between Queensland, Western Australia, and NSW. These low levels are expected to be temporary. To reflect this, the assumed level of NSW NIM trends down from -10,300 per annum in the 2011-16 period to -17,000 per annum by 2026 and onwards.
3. Age-profiles for interstate migration
People are more likely to move at certain ages. In NSW, those leaving the state are most likely to be in their early twenties, and those arriving from interstate are generally older (Figure 3, NIM section). The NSW population projections are based on interstate migration age profiles by sex and five-year age groups from the 2006 and 2011 Censuses.
4. Local Government Areas
Migration assumptions for local government areas are based on an analysis of flows that combine both interstate and intrastate movements within Australia. For more details see intrastate migration.
The ABS released projections in November 2013 for Australia, the states and territories. The level of net interstate migration loss used by the ABS for NSW is higher than those used in the 2016 NSW population projections. The ABS assumes initial net interstate migration losses for NSW of -18,000 persons, then reducing to -17,000 by 2015 and then staying constant. The ABS assumption is based on an analysis of trends over the period 2007-2012 and ensures interstate migration for all states and territories sum to zero. Interstate migration losses from NSW over that period have been smaller than in years prior to that interval (Figure) and these are reflected in the 2016 NSW population projection assumptions.
Net intrastate migration is the difference between the number of people moving to a Local Government Area (LGA) from other LGAs in NSW, and the number leaving that LGA for other places in NSW.
The assumed net intrastate migration level for each LGA is based on:
The assumptions about the number of moves into and out of each Local Government Area (LGA) have been informed by data from the 2006 and 2011 Australian Bureau of Statistics (ABS) Census and the Regional Internal Migration Estimates (RIME) by LGA for 2006-07 to 2013-14. As with the ABS interstate migration estimates, RIME are based on change of address data from Medicare Australia and the Department of Defence.
The assumed level of net intrastate migration for each LGA in the 2016 NSW population projections is based on a weighted average of migration patterns for 2001-2006 and 2006-2011 from the last Censuses and RIME for 2006-2014. Data from these two sources were compared with caution as the two measures are different in definition and scope. In some cases, these assumptions have been further refined with input from local councils regarding factors that may affect future population trends, particularly migration.
The local migration assumptions are modelled so they do not exceed the assumed levels of interstate and intrastate (i.e. within NSW) migration for seven large projection regions. The mobility of one local government area is affected by the flows of another, rather than being modelled independently. This means that net intrastate migration across all LGAs will always sum to zero because it measures all the movements between LGAs within the state. It also means interstate migration across all LGAs will always equal the net interstate migration assumption for NSW as a whole (see interstate migration for NSW).
Adjustments have been made to the assumed level of intrastate migration for LGAs with small populations that are currently static or showing decline. As the population becomes smaller, migration flows will also reduce. Retaining a constant level of net migration (usually a loss) over the projection period would accelerate the rate of population decline to an unacceptably high level.
Migration assumptions at all levels (international, interstate and intrastate within NSW) have the biggest impact on population projections. This is because migrants are concentrated at particular ages that affect the age profile of the places they are moving to and moving from. Migration assumptions have cumulative effects on births and deaths that influence population size into the future. Net migration gains see an increase in the projected number of women of childbearing age and the number of births compared to the number of deaths over the projection period.
We can measure how accurate projections are by comparing the projected population at a point in time with the actual population when we reach that point in time. Evaluation of past NSW population projections has found them to be highly accurate. The error margin increased as the projection year became further from the base year. Levels of error also tended to be higher for smaller populations, which are more likely to be affected by statistical errors that can arise when projections are based on small numbers.
Because the projections include a margin of error, scenarios have been developed to reflect different assumptions about births, deaths and migration. The high and low series reflect what could happen if there were changes to current trends. The show a potential range that future populations could grow to over the next 20 years.
Projection errors happen if there are sudden changes in demographic behaviour because of political, economic or environmental events. Errors can also occur when the data and information used to set the assumptions changes or is of poor quality. A key reason the projections are reviewed on a regular basis is to respond to such events or data issues that affect the population projection assumptions.
The 2016 NSW population projections are different because they have different assumptions. New long-term assumptions have been set for overseas, interstate, and intrastate migration. All short-term assumptions include already published data (2011 to 2014 or 2015 depending on the data source). These differ to assumptions set for the 2014 NSW population projections.
The 2016 NSW population projections also include new information on housing supply forecasts for Sydney. This information has informed Housing Unit Method assumptions and affects the distribution of the projected population across Sydney’s local government areas.
Even with new assumptions, there are key factors that remain the same. Population ageing remains a dominant feature, for example, and the drivers of population change remain a mix of natural increase and net migration.
The next major review will follow the release of 2016 Census results and be ready for release in 2018.
The 2016 projections show more population growth for NSW compared to the 2014 projections. The revised assumptions mean NSW is likely to have 9.9 million people by 2036. This is 225,000 more people than seen in the 2014 projections. This means more people in all parts of NSW.
The 2016 projections have the same age profile as the 2014 projections. Population ageing as the post-World War II Baby Boomers reach the older ages is still important. Babies continue to be born, and migration affects the number of people aged in their 20s and 30s.
Population estimates (estimated resident population or ERP) are prepared quarterly for Australian states and territories and annually for local areas by the Australian Bureau of Statistics (ABS). Estimates show past population change over time. The estimates are updated to account for births, deaths and net migration since the last estimate.
Population projections use population estimates as a starting point and are an indication of future population change based on assumptions about future demographic behaviour (births, deaths and migration).
The Australian Bureau of Statistics (ABS) population projections (catalogue no. 3222.0) are prepared once every five years for each state and territory, capital city and balance of state. The ABS projections are not done for LGAs or other small geographies. The 2016 NSW population projections are the only projections done for all LGAs in the state.
The ABS population projections use different assumptions to those used for the NSW population projections. This is because NSW population projection assumptions are informed by:
A comparison of the ABS and NSW population projection assumptions is included in the assumptions section.
The data files that have been published for download have been summarised at the following geographic levels:
The projections are for 25 years from 2011 to 2036.
The following conventions regarding dates are used in these projections.
‘The population in 2011’ is shorthand for the population at 30 June 2011.
‘Population change in 2011-16’ refers to population change over the five year period starting 1 July 2011 and ending 30 June 2016.
The first step is to try and understand why projections are different. Differences between population projections can be due to the following factors:
1. Jump-off year: The 2016 NSW projections have a base year of 2011 and include published data available to 2014 (or 2015 where available) for births, deaths and migration. Projections with a different base year are likely to have a different result.
2. Method: Different methods produce different results. The NSW population projections use a Cohort Component Method. This can produce different results compared with a model that is based on population trend lines. The NSW population projections also use a multi-regional model so that all local government areas (LGAs) are held to within a state total. This can produce different results to a model that is for one area or region alone.
3. Assumptions: Different population sizes can result from using different assumptions about births, deaths and migration. This includes assumptions about the number in each category as well as the age that each life event is assumed to take place.
The second step is to identify which projections are the best fit for purpose. The 2016 NSW population projections are available at the LGA level, but not the suburb level, so a different data source may be needed to understand suburbs.
Projected population growth rates are calculated after the projection modelling is completed. It shows the cumulative impact of the assumptions related to births, deaths and migration on different age groups over time. Growth rates are not assumed prior to the projection modelling being carried out.
Projections give us an indication of how many people will live in different places in the future. When planning for the future the scope of projected change is important, as well as what is driving that change. For example, population growth driven by natural increase will result in different population composition compared with growth driven by intrastate migration.
When planning for the future it is important to evaluate what the age profile of a population will look like. Two populations may be the same size, but one has one-third of all people aged 70 years or older compared to another that has just ten per cent at those ages. The types of services, housing and infrastructure that need to be planned will be different in each case.
Population numbers are rounded to the nearest 50. Publishing unrounded numbers give a false sense of accuracy to the projections. Projections are developed to give an indication, rather than an exact number, of the population size and makeup – for example, what proportion of the population will be 65 years and older, how many people of working age there may be in 20 years’ time.
Projections are provided for five-year age groups. The youngest age group is 0-4 years and the oldest age group is 85 years and older.
We do not prepare projections for single year age groups. We recommend that data for a single year of age be estimated by dividing the projection for the five-year age group by five. For example, it is projected that there will be 223,000 females aged 10-14 in 2016. By dividing 223,000 by five, it is estimated that there may be 44,600 females in each of the ages 10, 11, 12, 13, 14 years old.
This method assumes that the numbers of persons are equally distributed across the single years of age within an age group. This assumptions provides a relatively good estimate in the absence of other information. The projections are not intended to give an exact number for the population in the future.
Department of Planning and Environment (2016), 2016 New South Wales State and Local Government Area population and household projections. Sydney, NSW.
All figures published in the NSW projections are rounded to the nearest 50. Raw numbers from the projection model are rounded to the nearest multiple of 50. This applies to the published projected population for age groups, and for the total populations.
For the total populations, the raw numbers for smaller groups (such as age groups, LGAs) are added together to give the raw total population. These raw totals are then rounded to the nearest 50 and published in the projections.
We calculate the total population by sum of the raw numbers for three areas: 476, 302, and 180. These raw figures sum to the raw total of 958, which is 950 when rounded to the nearest 50. We would publish 950 as the total for the sum of the three areas. On the other hand, the numbers for the areas would be published as 500, 300, and 200 when rounded to the nearest 50. These rounded figures sum to 1,000 and is different to the published total of 950.
ABS Australian Bureau of Statistics
DIBP Department of Immigration and Border Protection
ERP Estimated Resident Population
LGA Local Government Area
LHD Local Health District
NIM Net Interstate Migration
NOM Net Overseas Migration
NSW New South Wales
TFR Total Fertility Rate
Usual Place of Residence:Refers to the place a person usually lives for six months or more. Used to measure births and deaths data, with the place a person usually lives used not the place a mother gives birth or where a person dies.
Usual Resident:Person who lives in one place for six months or more. Applies to estimated resident population and projected population.
Population projections are reviewed to answer the simple question, ‘how accurate have past projections been?’ Measuring the projections against the Australian Bureau of Statistics (ABS) population estimates indicates the uncertainty of current projections. Research has shown that projections within three per cent of the actual population are considered highly reliable. Across the last ten projection releases, the official NSW projections have been within two per cent after 10 years, making them highly reliable.
We have evaluated the Department of Planning and Environment projections released in 1984, 1987, 1988, 1989, 1992, 1994, 2004, 2005, 2008 and 2010. The main measure used to assess their accuracy is percentage error (PE), which is the percentage discrepancy between the ABS Estimated Resident Populations (ERPs) and the Department’s projections. It is important to note that ERPs are themselves an estimate, and incorporate error; all resulting comparisons should be made with caution.
In 2013 the ABS released rebased and recast ERPs from 1991 to 2011 in response to a new method of estimating how well the Census counted Australia’s population. These new ERPs showed that previous population estimates for NSW had been too high, and consequently meant that the base populations used for previous sets of official NSW population projections were too high. To account for this change in ERPs, corrected percentage errors (CPE) are calculated for each projection set by taking account of the relative difference between projections and ERPs in each time period. Research has shown that projections within three per cent of the actual population are considered highly reliable.
Mean Percentage Error [MPE] is also used in this analysis. It is a measure of bias indicating whether a set of projections was overall too high (positive values), or too low (negative values).
The accuracy of projected total populations for NSW has varied across the last ten cycles of their production. Figure 1 highlights the extent to which past projections of the total New South Wales population have differed from the subsequently published ERPs. Lines above the 0% axis (bold line) indicate the projected population was too high, while lines below the 0% axis indicate a projected population was too low. The blue diamond symbol indicates the Mean Percentage Error 5, 10, 15 and 20 years from the base (or starting point) of the projection, averaged for all projection series.
Figure 1 shows that official NSW population projections have generally over-projected the future population of NSW. Projections prior to 1994 had positive percentage errors 5, 10, 15 and 20 years into the projection horizon. More recently, there has been a trend to under-project the state’s population, with the 2004 and 2005 releases (at the 10-year mark) and the 2008 release projections (at the 5-year mark) receiving negative percentage errors. The unforeseen boom in net overseas migration that has occurred over the last decade may have contributed significantly to this pattern.
The mean percentage error measures show that the level of error for NSW projections increases with time, and the vertical grey lines (Figure 1) reveal greater margins of error over time as well. However, it is important to note that the majority of errors are less than 1.5% at the 10 year mark, with the exception of the 1989 projections. Therefore the NSW population projections can be considered highly reliable.
An evaluation of past projections for local government areas (LGAs) in NSW was limited by past changes to local government boundaries. Mean corrected percentage errors have been calculated for each LGA in Sydney using boundaries as at the time of the 2011 Census at a point five years into the projection horizon. The 2004, 2005, 2008 and 2010 projection releases have been evaluated and the results summarised in Figure 2. The top five over-projected and under-projected LGAs in the Sydney Metropolitan region are listed in Table 1.
|1. Camden (11.4%)||1. Leichhardt (-3.1%)|
|2. Auburn (4.0%)||2. Parramatta (-3.0%)|
|3. Hunters Hill (3.9%)||3. Kogarah (-2.3%)|
|4. Sydney (3.3%)||4. Manly (-2.2%)|
|5. Campbelltown (3.0%)||5. Randwick (-1.9%)|
Click to enlarge
Outside of Sydney, numerous LGAs in western NSW have been over-projected. LGAs surrounding the ACT and in areas with significant mining activity have been under-projected (Figure 2). Across the state it is notable that the smaller the LGA population, the larger the projection error (Table 2). It is inherently very difficult to model the complex nature of population change for small populations.
Click to enlarge:
|1. Murrumbidgee (11.0%)||1. Palerang (-6.1%)|
|2. Carrathool (9.2%)||2. Yass Valley (-5.0%)|
|3. Bourke (8.2%)
||3. Boorowa (-4.6%)|
|4. Wakool (8.1%)||4. Upper Hunter Shire (-4.4%)|
|5. Hay (7.5%)||5. Gunnedah (-4.2%)|
Page last updated: 14/06/2019