Census of Population and Housing: Understanding the Increase in Aboriginal and Torres Strait Islander Counts

Between the 2011 and 2016 Censuses, the number of Aboriginal and Torres Strait Islander people in Australia increased by 18.4% (100,803 people).

Large increases in the counts of Aboriginal and/or Torres Strait Islander people have been observed over various periods since 1971. Particularly-large increases occurred between the 1991 and 1996 Censuses (33.0%) and the 2006 and 2011 Censuses (20.5%). Factors contributing to these intercensal increases were explored in Occasional Paper: Population Issues, Indigenous Australians, 1996 (cat. no. 4708.0) and in the previous release of this publication after the 2011 Census.

The purpose of this publication is to discuss factors contributing to the increase in counts of Aboriginal and Torres Strait Islander people observed between the 2011 and 2016 Censuses and their implications for interpreting and understanding Census data. This analysis is important because increases such as these have an impact on other statistics, such as population measures and/or performance indicators used for government reporting.

This analysis explores the components of the increase by breaking it down into its demographically-explainable and unexplainable components. Components including births and deaths, migration, a person’s propensity to identify as an Aboriginal and/or Torres Strait Islander person, and Census enumeration and response are all taken into account to help define and understand the intercensal increase.

The chapter Overview of the Increase examines how the Aboriginal and Torres Strait Islander population changed during the intercensal period. It shows that births between the 2011 and 2016 Censuses (0-4 year olds) are driving the majority of the intercensal increase (72.7%). The intercensal growth in counts of Aboriginal and Torres Strait Islander persons between 2011 and 2016 is not consistent across the country, with growth primarily occurring in Major cities and on the eastern coast of Australia. Population growth is significantly higher in Non-Remote areas (22.8%) than in Remote areas (2.0%).

The chapter Change Explained by Demographic Factors quantifies how much of the 100,803 increase in the Census count of Aboriginal and Torres Strait Islander people between 2011 and 2016 can be explained by demographic factors. It finds that most (78.6%) of the increase can be accounted for by explainable demographic factors of population change – births, deaths and migration.

The remaining 21.4% of the increase cannot be explained by typical factors of population change and equates to 3.3% of the total 2016 Census count of Aboriginal and Torres Strait Islander people (Change Not Explained by Demographic Factors). This increase did not occur consistently across the country and was primarily focused in Major Cities and Inner Regional areas, particularly in New South Wales.

Factors of unexplainable change in Census counts include: coverage, response rates and a changing propensity to identify. Analysis in this chapter shows that while the overall increase in Census counts is predominately composed of intercensal births (72.7%), other young children (5-14 year olds) were also a significant contributor (14.7%).

The chapter Changing Propensity to Identify examines peoples’ changing propensity to identify as Aboriginal and/or Torres Strait Islander. It finds that there is a relationship between changes in parental identification and changes in the identification of children. Children of couples where both parents are of Aboriginal and Torres Strait Islander origin – or children in lone parent families with an identifying parent – are more likely to be identified than children of couples where one person is of Aboriginal and Torres Strait Islander origin and the other is non-Indigenous.

The final chapter of this publication Impact of Intercensal Change on Selected Characteristics explores differences in the measures of socio-economic characteristics of Aboriginal and/or Torres Strait Islander people counted in the 2011 and 2016 Censuses, particularly for characteristics used to monitor progress against Closing the Gap targets (Education, Labour Force and Income). The chapter finds that attempting to measure the impact of changing propensity to identify is complex. Across all characteristics analysed, intercensal improvements appear to be a result of a combination of the changing population distribution of Aboriginal and Torres Strait Islander people across Remoteness areas, persons who are socio-economically better off choosing to identify, and actual progress.

There are three technical notes attached to this publication to provide background and context to the analysis. Technical Note 1 Aboriginal and Torres Strait Islander Status and the Census looks at the characteristics of people whose Indigenous Status is unknown. Technical Note 2 The undercount in the Census and the PES outlines Aboriginal and Torres Strait Islander undercount and overcount from the 2016 Post Enumeration Survey. The Aboriginal and Torres Strait Islander undercount increased slightly to 17.5% in 2016 (from 17.2%). Finally, the Technical Note 3 provides information on the Australian Census Longitudinal Dataset. For more information, please see the technical note sections on the Methodology page.

• Counts of Aboriginal and Torres Strait Islander Australians increased by 18.4% between the 2011 and 2016 Censuses. This increase is less than the increase between 2006 and 2011 (20.5%).
• The growth in counts of Aboriginal and Torres Strait Islander persons between 2011 and 2016 is not consistent across the country, with growth primarily occurring in Major Cities and on the Eastern coast of Australia.
   

There were 100,803 more Aboriginal and Torres Strait Islander Australians counted in the 2016 Census compared to 2011, an 18.4% increase. Since the introduction of the Standard Indigenous Question (SIQ) in 1996, the Census count of Aboriginal and Torres Strait Islander Australians increased by 83.9%.

Factors that can have an impact on the intercensal increase are:

• The fertility rates of Aboriginal and Torres Strait Islander women, which have traditionally been higher than other Australian women.
• People entering and leaving the population through migration.
• Variation in Census coverage and response rates.
• People changing if they identify as an Aboriginal or Torres Strait Islander person between Census years.
   

This chapter will quantify the intercensal increase between 2011 and 2016. It will be examined by age, state and territory and Remoteness. Change Explained by Demographic Factors and Change Not Explained by Demographic Factors will then break that increase up into its demographically-explainable and unexplainable components.

An age cohort is a group of people with the same age within a defined period. For example, persons aged 30-34 years in 2016 are the same five-year cohort as persons who were aged 25-29 years in 2011. An age cohort analysis is used to analyse changes for a chosen group of people at points in time and is a way of examining Census data longitudinally without linking. Utilising an age-cohort analysis to examine intercensal change between the 2011 and 2016 Censuses can provide further insights into how change has impacted the overall population.

Expressing the intercensal change in each age cohort as a proportion of the total population change allows analysis on how each age group contributed to overall population change. The majority (72.7%) of the 100,803 person increase between 2011 and 2016 can be attributed to intercensal births (0-4 year olds). In a population where people cannot migrate in or out, we would expect 0-4 year olds to be the only cohort to contribute to population growth. However, between 2011 and 2016 almost every five-year age cohort under 70 years old increased in size.

After 0-4 year olds, the biggest contributing age groups were 5-9 year olds (8.3%) and 10-14 year olds (6.4%). In total, children under 15 accounted for 87.4% of the increase. Most other age cohorts had minor contributions of up to 2.9%.

Counts in some age cohorts decreased between the 2011 and 2016 Censuses, resulting in these cohorts making negative contributions to the population. These were primarily older age cohorts (70 years and over). The exception was the 20-24 years age cohort, which recorded a decrease of 1,748 in the count of Aboriginal and Torres Strait Islander persons. Falls in this age cohort were observed in previous Census cycles and possibly reflect a small number of people being missed by the Census, then being picked up in subsequent Censuses later in life.

Across all age cohorts, population changes were generally consistent for males and females. Slightly more males contributed to the increasing population in younger age groups and slightly more females in older age groups. Again, the exception was the 20-24 year old cohort where counts of males fell significantly more than counts of females, with males contributing approximately two thirds of the decrease in this cohort.

The Australian Statistical Geography Standard (ASGS) Remoteness Structure is a geographic classification that divides Australia into broad regions that share common characteristics of Remoteness. Characteristics of Aboriginal and Torres Strait Islander populations can vary significantly between different types of Remoteness areas making this classification very valuable for understanding the overall population.

Remoteness is associated with poorer outcomes on a range of Closing the Gap targets (see Closing the Gap targets: 2017 analysis of progress and key drivers of change released by the Australian Institute of Health and Welfare). Consequentially, changes in the population distribution across Remoteness areas may impact on the progress towards meeting these targets (see Impact of Intercensal Change on Selected Characteristics).

Since the 2006 Census, the distribution of Aboriginal and Torres Strait Islander people across Remote and Non-Remote Australia has shifted. Proportionally, more Aboriginal and Torres Strait Islanders are living in Major Cities and Inner Regional Areas and fewer people are living in Outer Regional Australia, Remote Australia and Very Remote Australia. These shifts are larger between the 2011 Census and the 2016 Census than they were between the 2006 Census and the 2011 Census.

Aboriginal and Torres Strait Islander people living in Major Cities accounted for the greatest proportion (53.6% or 53,981 persons) of the additional 100,803 Aboriginal and Torres Strait Islander persons counted in the 2016 Census. Other major contributions came from Inner Regional areas (34.0% or 34,309 persons) and Outer Regional areas (9.3% or 9,391 persons). Remote and Very Remote areas contributed just 2.3% of the overall increase (2,351 persons).

In terms of population distribution, Outer Regional Australia and Very Remote Australia recorded the largest percentage point falls between 2011 and 2016 (1.9 percentage points each). Percentage growth fell in Outer Regional areas by 12.2 percentage points, and 10.3 percentage points in Very Remote areas.

All states and territories had significant population growth in Non-Remote areas between the 2011 and 2016 Censuses. New South Wales (26.6%), Victoria (26.1%) and Queensland (23.2%) saw large increases in their Non-Remote Aboriginal and Torres Strait Islander populations. The Northern Territory had the smallest (7.8%) increase, while the ACT’s seemingly-large increase was off a low base.

The Northern Territory had the largest proportion of Aboriginal and Torres Strait Islander persons living in Remote areas, which saw essentially no growth (0.1%) between 2011 and 2016. Queensland (5.3%), Tasmania (6.1%) and Western Australia (3.7%) were the only jurisdictions to record positive growth in Remote areas.

Intercensal births (age cohort 0-4) will always show the largest increase between Censuses. For cohorts aged 5 years and over, growth primarily occurred in Major Cities and Inner Regional areas. In Outer Regional, Remote and Very Remote areas there were declines in counts in almost every age cohort.

Considering growth was much higher in Non-Remote areas than in Remote areas, it is expected that there are very different patterns in their age cohort analyses. In Remote and Very Remote areas, almost every age cohort made a negative contribution to the area’s overall population change. This means that almost the entire population growth in these areas came from births (0-4 year olds) and that non-demographic factors such as changing propensity to identify are not a significant factor in these areas.

In Major Cities and Inner Regional areas, 0-4 year olds were much smaller contributors to the overall increase, contributing just over half of total population change in these areas. This is significantly lower than the contribution 0-4 year olds made to the national increase (72.7%) and suggests an increased propensity to identify has a significant influence on intercensal population change in these areas.

Children aged 5-14 were the second largest contributor to the national increase (14.7%). This can also be observed in Major Cities (17.5%), Inner Regional (19.1%) and Outer Regional areas (8.5%). However, in Remote areas only 5-9 years olds had a positive contribution.

New South Wales had the largest numerical increase (43,551) in its Aboriginal and Torres Strait Islander population between 2011 and 2016, while Victoria had the highest percentage growth (25.8%). The Australian Capital Territory had the smallest numerical increase in population of all states and territories (1,324), but large percentage growth (25.5%).

Across all states and territories, percentage growth in the number of males (19.2%) was slightly greater than that for females (17.6%). The Australian Capital Territory was the only state or territory where there was greater growth in the number of females than males (27.9% compared to 23.0%).

New South Wales (NSW) recorded the largest count of Aboriginal and Torres Strait Islander people in 2016 (216,176) and contributed almost half (43.2%) of the overall increase in counts of Aboriginal and Torres Strait Islander persons.

The majority (57.3%) of population growth in NSW was attributable to 0-4 year olds with 5-14 year olds contributing a further 17.3% of growth. Persons aged 5 years and over contributed more to population growth in NSW (42.9%) than they did nationally (27.4%).

With the exception of intercensal births (0-4 year olds in 2016), there were falls across all age cohorts in Remote areas in NSW. There were increases in all age cohorts in Non-Remote areas.

Approximately two-thirds (71.9%) of population growth in NSW came from the two Indigenous Regions with the largest Aboriginal and Torres Strait Islander populations, NSW Central and North Coast (40.1%) and Sydney-Wollongong (31.8%). Approximately a third of the increase in each of these two regions came from persons aged 15 years and over.

Census counts of Aboriginal and Torres Strait Islander persons in Victoria increased by 25.8% between 2011 and 2016. This increase contributed 9.7% of the national increase in counts of Aboriginal and Torres Strait Islander persons.

Approximately half of population growth in Victoria was attributable to 0-4 year olds with 5-14 year olds contributing a further 14.7% to total state growth. Persons aged 0-4 years contributed less to Victoria’s population growth than they did nationally, while 44.0% of growth came from persons aged 5 years and over compared to 27.4% nationally.

Between the 2011 and 2016 Censuses, there were consistent increases in the counts of Aboriginal and Torres Strait Islander people in all five-year age groups up to 69 years, including those aged 20-24 in 2016. In this respect, Victoria’s population distribution differs from the national distribution, but is consistent with the 2011 Census where an increase was also recorded in Victoria for this age cohort.

The overall increase in Victoria was divided fairly evenly across Victoria’s two Indigenous Regions with Melbourne contributing 58.3% of the total state increase and Victoria (exc. Melbourne) contributing the remaining 41.7%. However, the contribution by age cohort was different between the two regions with persons aged 15 years and over contributing 42.2% of the total increase in Melbourne but only 11.7% of the total increase in Victoria (exc. Melbourne).

Queensland recorded the second-largest count (186,482) of Aboriginal and Torres Strait Islander people in 2016, after NSW.

There was an increase of 21,857 Aboriginal and Torres Strait Islander Queenslanders aged 0-4 years counted in 2016, which accounted for nearly three-quarters of the overall increase in the state (71.3%). Children aged 5-14 years contributed a further 15.0%, meaning the increase in Queensland counts was driven almost entirely by children under 15.

As observed in NSW, the increase in Queensland was entirely focused in Non-Remote areas, where there were increases in almost every age cohort. With the exception of intercensal births (0-4 year olds in 2016), all age cohorts saw a decline in Remote areas.

The population of 20-24 years olds in Queensland decreased in both Remote and Non-Remote areas, although there was a small increase in Major Cities.

Growth in Queensland was strongest in the Indigenous Region of Brisbane. Brisbane added 17,463 people between Census years and contributed 57.0% of the overall state increase. There was also modest growth on Queensland’s Eastern coast with Rockhampton and Townsville-Mackay contributing a further quarter of the state’s overall growth. Brisbane, Rockhampton and Toowoomba-Roma were the only Indigenous Regions in Queensland that recorded positive growth in persons 15 years and over.

In the 2016 Census, South Australia’s (SA) Aboriginal and Torres Strait Islander population increased by 3,752 persons (12.3%), mostly driven by growth in Non-Remote areas. Similarly to the pattern observed in Queensland, SA’s population growth was driven almost entirely by children under 15.

Growth in SA was almost entirely attributed to the Indigenous Region of Adelaide, which grew by 3,535 people, contributing 94.2% of the overall state increase. The remaining population growth was in Port Augusta, with Port Lincoln-Ceduna recording small negative change. Population growth in both Adelaide and Port Augusta was primarily driven by 0-4 year olds.

Western Australia’s (WA) Aboriginal and Torres Strait Islander population had one of the smallest percentage increases between the 2011 and 2016 Censuses (9.1%). The state had the fourth-highest increase in counts, adding 6,314 persons. Similar to Queensland, all population growth in WA came from children aged 0-14 years with a decline across all other age cohorts.

Keeping with national trends, the increase in WA is focused in Non-Remote areas. With the exception of intercensal births (0-4 year olds in 2016), the only Remote age cohort that saw an increase was children 5-9 years. In contrast with other states and territories, there were decreases in counts recorded in many age cohorts in Non-Remote areas as well as in Remote areas.

Growth in WA was driven by growth in the Indigenous Regions of Perth (56.8%), South-Western WA (19.7%), South Hedland (17.6%) and West Kimberley (14.5%). South Hedland and Western Kimberly were the only two Indigenous Regions in WA to record an increase in counts for persons 15 years and over.

In the 2016 Census, Tasmania’s Aboriginal and Torres Strait Islander population increased by 3,947 persons (20.1%). Due to the small size of the state’s population, the change in the counts of Aboriginal and Torres Strait Islander people aged 15 years and over between 2006 and 2011 was minor for most five-year age cohorts.

The distribution of the population changes in Tasmania in 2016 is consistent with the national trend. Aboriginal and Torres Strait Islander children aged 0-4 years in 2016 accounted for the majority (61.8% or 2,439 people) of the total increase with other children aged 5-14 years contributing a further 19.6% of total population change.

The Aboriginal and Torres Strait Islander population in the Northern Territory (NT) increased by 1,469 people between the 2011 and 2016 Censuses. This increase was driven by births between the 2011 and 2016 Censuses. Apart from a very small rise in the count of women in the 50-54 year cohort, there were population decreases in every other cohort (excluding intercensal births).

The decrease in the 5-9 and 10-14 year olds old cohorts is particularly notable, since these cohorts saw increases in every other jurisdiction. Decreases in most age cohorts occurred in the NT in 2011 as well.

A large proportion of the NT’s Aboriginal and Torres Strait Islander population were living in Remote areas and it was Remote areas that made the major contribution to overall population change. Small growth in some age cohorts was observed in Non-Remote areas in the NT; however this was offset by larger declines in Remote areas.

Four of the Indigenous Regions in the NT recorded intercensal population growth with the Darwin region (63.2%) and Nhulunbuy (20.3%) the main contributors to the Territory’s population growth.

The Australian Capital Territory (ACT) recorded the smallest increase in person counts (1,324) of all states and territories reflecting the small total population of Aboriginal and Torres Strait Islander persons. Small increases were recorded for people aged 15 years and over between 2016 and 2011 for most five-year age cohorts.

Children and youth under 25 years accounted for most of the population growth (83.9%) in the ACT, with growth primarily driven by children age 0-4 years. The ACT was the only jurisdiction other than Victoria to record growth in the 20-24 years old cohort.

Examining population change within Indigenous Regions shows significant variability across the country. The largest population increases occurred on Australia’s east coast with the Indigenous Regions of Brisbane (17,463 persons), NSW Central and North Coast (17,452 persons) and Sydney-Wollongong (13,852 persons) contributing 48.3% of the total national increase.

While most Indigenous Regions recorded population increases between 2011 and 2016, some did record small decreases. These regions were primarily in Remote Australia with the biggest decrease recorded in the Kununurra region (–745 persons), followed by Tennant Creek (–179 persons) and Katherine (–166 persons).

1. Usual Residence Census Counts. Excludes overseas visitors. 

Note: Totals and components may not be consistent within and between tables due to introduced random error to protect confidentiality of Census respondents - see Census Dictionary, 2016 (cat. no. 2901.0). 
See Data downloads section Overview of the Increase, Table 1.
Source: Australian Census of Population and Housing, 2011-2016

The Estimated Resident Population (ERP), is the official population estimate for Aboriginal and Torres Strait Islander Australians. It is calculated using the results of the 2016 Census as a base with adjustments applied during the intercensal period to account for births, deaths and overseas migration over time, together with an adjustment for the net Census undercount (as measured by the Census Post Enumeration Survey (PES) (see Technical Note 2 – The Undercount in the Census and the PES).

The final estimate of the Aboriginal and Torres Strait Islander population at 30 June 2016 was 798,365 people, up from 669,881 in 2011. Breaking this estimate down into age cohorts and by state/territory reveals how the Aboriginal and Torres Strait Islander population is distributed once adjusting for the net Census undercount. By comparing this breakdown to Census counts, we can then see whether the age cohorts that are contributing to the increase in Census counts are consistent with the estimated number of people in that age cohort based on ERP.

The majority of age cohorts made similar contributions to changes in ERP as they did to the change in Census counts. In both datasets, intercensal births, or young children aged 0-4 ,were the main source of intercensal population change, contributing 73.0% of the increase in ERP and 72.7% of the increase in Census counts.

There was a noticeable change between Census counts and ERP for young adults aged 20-24 years. This cohort made a negative contribution to the overall Census difference (–1.7%) but a positive contribution to the overall ERP difference (1.0%). Because the ERP adjusts for undercount, this suggests the difference may be attributable to undercount for this age cohort in the Census rather than a change in identification.

When comparing ERP and Census counts by State and Territory, Australia’s larger States – NSW, Victoria, Queensland, WA and SA – showed a similar distribution across age cohorts to that observed nationally, with the majority of population growth from intercensal births (0-4 year olds) and young children (5-14 year olds).

The most notable exception could be seen in Victoria where 25-29 year olds contributed much less to changes in ERP estimates than changes in Census counts (3.0% compared to 0.2%).

In NSW, the contribution of age cohorts 15 years and over to population change was 26.9% - higher than all other states and territories and the national contribution (13.0%).

Following the 2011 Census, the ABS projected that Australia’s Aboriginal and Torres Strait Islander population would grow to 744,956 people in 2016. The 2016 Estimated Resident Population (ERP) shows the population is 7.2% bigger than the projected population at 798,365 persons.

Comparing ERP to the population projections, we can see the difference in the projected and actual 2016 populations has primarily been driven by children under 15, with 5-9 year olds making up 20.8% of the difference and 10-14 year olds contributing 13.4%. The large contribution of 5-14 year olds appears to be driven mainly by the unexplained increase in the Census counts for children in these age groups as, outside of new births, they were the largest contributors to intercensal growth.

• Most (79,272 or 78.6%) of the 100,803 increase in the Census counts of Aboriginal and Torres Strait Islander people between 2011 and 2016 can be explained by demographic factors. That is, births, deaths and overseas migration.
• The proportion of the increase that is explainable is greater than it was between 2006 and 2011 (70.2%).
• The remaining 21.4% (21,531) of the increase in Census counts is unexplainable change; that is change that is not explained by demographic factors of births, deaths and overseas migration.
   

As outlined in Overview of the Increase, there was an increase of 100,803 persons (18.4%) between the 2011 and 2016 Censuses. There are components of this increase that can be explained by demographic factors of births, deaths and overseas migration (referred to as demographically-explainable change) and a component that cannot (referred to as unexplainable change). The proportion of the increase that can be attributed to each component can be quantified by comparing Census data with administrative datasets of births and deaths where Aboriginal and/or Torres Strait Islander status is collected and is considered of acceptable quality. This chapter will:

• Quantify the explainable change due to the demographic factors of births, deaths and migration between the 2011 and 2016 Censuses.
• Explore the relative contribution of those factors.
• Compare the explainable changes with those observed between the 2006 and 2011 Censuses for states and territories, Remoteness Areas and Indigenous Regions.
   

The analysis presented in this chapter is based on the 649,171 people who identified as being of Aboriginal and/or Torres Strait Islander origin in the 2016 Census. This analysis does not consider Census records where a person did not provide their Indigenous status.

In order to measure explainable change in Census counts between 2011 and 2016, the net effects of births, deaths and overseas migration are added (or subtracted) to the 2011 Census count of Aboriginal and Torres Strait islander people to produce an expected 2016 count.

If it were possible to achieve complete coverage in Census enumeration, and consistency in the reporting and recording of each person's Indigenous status, change in Census counts between the 2011 and 2016 Censuses would be entirely attributable to births, deaths and migration combined. In reality there will always be an element of the total change in counts between Censuses that cannot be explained, such as the change that occurs as a result of individuals identifying, or who are identified, as an Aboriginal and/or Torres Strait Islander person in one Census but not in another, as well as errors in the measurement of births, deaths and migration.

The framework for measuring explainable change can be represented as:

The proportion of intercensal change attributable to explainable change is calculated on the difference between 2011 and 2016 Census counts.

Natural increase is the net effect of deaths minus births between two Census periods. At the national level, 76.7% of the increase in Census counts of Aboriginal and Torres Strait Islander people between 2011 and 2016 can be attributed to natural increase. In comparison, 58.7% of the increase between 2006 and 2011 was attributable to natural increase.

Between states and territories, there is variation in the contribution to the intercensal change in counts of Aboriginal and Torres Strait Islander people that can be attributed to natural increase. In New South Wales and the Australian Capital Territory, just over half of the increase can be attributed to natural increase (57.0% and 58.7% respectively), while in Queensland, natural increase accounted for three-quarters of the increase (75.1%). In contrast, natural increase was greater than the overall increase in Census counts for Aboriginal and Torres Strait Islander people in the Northern Territory, Western Australia and South Australia with the population growth among 0-4 year olds offsetting declines in numbers for all other age cohorts.

There were 5,849 more Aboriginal and Torres Strait Islander 0-4 year olds counted in the 2016 Census than in the 2011 Census (73,265 compared to 67,416). One possible explanation for this increase would be a higher fertility rate for Aboriginal and Torres Strait Islander women between 2011 and 2016 than that observed between 2006 and 2011. However, the average number of children ever born across all age groups decreased slightly between these periods. This suggests that the increase may be a result of increased propensity for parents to identify their children as being of Aboriginal and/or Torres Strait Islander origin in the 2016 Census than in the 2011 Census. For more detailed analysis of fertility rates and parent/child identification please see Changing Propensity to Identify.

Some information about overseas migration is collected in the Census, in response to the question "What was your place of usual residence five years ago?" People who answer this question by stating they were overseas are assumed to have migrated to Australia after the last Census and before the current Census. Persons who were overseas at the time of the 2016 Census will not have a Census record for 2016 so these will be an unquantifiable net loss between the two Censuses.

In the 2016 Census, the proportion of Aboriginal and Torres Strait Islander people who indicated that they had been living overseas at the time of the 2011 Census accounted for 2.0% (1,969 persons) of the increase in Census counts of Aboriginal and Torres Strait Islander people during this period (compared with 1.5% or 1,375 persons between the 2006 and 2011 Censuses).

Administrative data sources of overseas migration (arrival and departure information from the Department of Home Affairs) do not collect the Indigenous status of persons arriving to or departing from Australia, so it is not possible to compare Census migration information to administrative records to explore possible differences.

Within Australia, migration between the states and territories provides an important insight into population changes at the jurisdictional level. Information on internal migration is collected in the Census, in response to the question "What was your place of usual residence five years ago?"

Between the 2011 and 2016 Censuses, the Australian Capital Territory, Victoria, South Australia, Queensland and Western Australia recorded a net gain in Aboriginal and Torres Strait Islander residents from interstate migration. While most rates of net gain remained largely similar for these states and territories over the 2006 to 2011 intercensal period, the Australian Capital Territory recorded a net gain of 19.9% between 2011 and 2016 compared with 11.4% between 2006 and 2011.

The Northern Territory, Tasmania and New South Wales experienced net loss from interstate migration between 2011 and 2016. The net loss was highest in New South Wales (2,012 Aboriginal and Torres Strait Islander people).

In Major Cities and Inner Regional areas, natural increase was substantially lower than relative change between Censuses and accounted for a smaller proportion of the overall increase in these areas compared to other parts of Australia.

Whilst the number of Aboriginal and Torres Strait Islander people who migrated to, or returned to, Australia between 2011 and 2016 is small (1,969 persons), these persons were most likely to be returning to Major Cities. The number of people returning to Major Cities from overseas between 2011 and 2016 was more than double the number returning to Regional areas and significantly higher than Remote and Very Remote areas.

Major Cities and Inner Regional areas were the main beneficiaries of inter-regional migration between 2011 and 2016 with the vast majority of movements into these areas.

Across the 37 Indigenous Regions (IREGs) in Australia, 21 had natural increases greater than the overall increase in Census counts of Aboriginal and Torres Strait Islander people between 2011 and 2016. Kalgoorlie, Jabiru-Tiwi and North-Western New South Wales had the highest proportion of intercensal change due to natural increase.

There were four IREGs where less than half of their increase in Census counts between 2011 and 2016 was attributable to natural increase:

• New South Wales Central and North Coast (42.3%)
• West Kimberley (45.6%)
• Melbourne (46.3%)
• Brisbane (46.4%)

When natural increase is calculated as a proportion of the 2016 Census count in each IREG, Perth (16.1%), Geraldton (15.2%) and Cairns-Atherton (14.7%) had the highest proportion of the 2016 Census count attributable to natural increase. In contrast, the contribution of natural increase to the 2016 Census count was lowest in Apatula (4.5%), Nhulunbuy (5.0%) and Torres Strait (5.9%).

Small changes in the number of persons migrating to an IREG in 2016 from overseas can appear as large proportional changes in IREGs with small population change. As such, for overseas migration, we’ve looked at overall numbers rather than proportional change or contribution. In 2016, the IREGs with the highest numbers of Aboriginal and Torres Strait Islander persons migrating to Australia from overseas between 2011 and 2016 were all IREGs within close proximity to Australia’s three largest capital cities: Brisbane (377), Sydney-Wollongong (348) and Melbourne (216).

As was the case with overseas migration, the proportional change due to inter-regional migration appears greater in IREGs where there was small change in counts between the 2011 and 2016 Censuses.

In the IREG of Brisbane, which had the largest count of net inter-regional migration (3,305 persons), also had the highest proportion of positive change due to inter-regional migration (18.9% of relative change between 2011 and 2016). Apatula recorded the highest proportion of relative change attributable to inter-regional migration however the actual increases in counts in Apatula as a result of net inter-regional migration are relatively small (108 persons).

As can be expected, IREGs recording a decline in the number of Aboriginal and Torres Strait Islander people between 2011 and 2016 (i.e. negative growth) were more likely to record a loss in population due to inter-regional migration than IREGs where the Aboriginal and Torres Strait Islander population increased. This was particularly the case in Mount Isa where net inter-regional migration saw an intercensal loss of 565 persons, far greater than Mount Isa’s overall change (a loss of 57 persons). Similarly, in North-Western NSW net inter-regional migration contributed a loss of 546 persons, far greater than North-Western NSW’s overall absolute change (an intercensal gain of 34 persons).

• After accounting for births, deaths and net migration, 21.4% of the 100,803 increase in Aboriginal and Torres Strait Islander persons counted in the 2016 Census could not be explained by demographic factors (referred to as unexplainable change).
• 5-14 year olds contributed the most to the unexplainable change.
• Growth in rates of unexplainable change is primarily occurring in Major Cities and Inner Regional areas. There is negative unexplainable change in Outer Regional, Remote and Very Remote areas.
• Across states and territories, unexplainable change is being driven by changes in New South Wales and Queensland. Australia’s two most populous states had age cohort patterns of unexplainable change that closely matched the national distribution.
   

Between the 2011 and 2016 Censuses, there was change that could not be explained by demographic factors. This is referred to in this chapter as unexplainable change. The factors that may contribute to the unexplainable change are discussed in Changing Propensity to Identify, where we explore identification- particularly of children and families where Aboriginal and Torres Strait Islander persons have partnered with non-Indigenous persons. There are a number of other factors in how Censuses are enumerated that can contribute to unexplainable change. These include coverage and response rates. Technical Note 1 – Aboriginal and Torres Strait Islander Status and the Census and Technical Note 2 – The Undercount in the Census and the PES for a discussion of these factors, including the Census undercount for Aboriginal and Torres Strait Islander persons.

At the national level, the proportion of the unexplainable change as a percentage of the 2016 Census count of Aboriginal and Torres Strait Islander persons was 3.3% between 2011 and 2016, slightly lower than between 2006 and 2011 (5.1%).

This chapter will:

• Examine key components of the unexplainable change between the 2011 and 2016 Censuses by accounting for measurable demographic factors of population change.
• Investigate the impact of the 2011-2016 intercensal change on the final Estimated Resident Population (ERP) figures for Aboriginal and Torres Strait Islander Australians.
   

The unexplainable change is the difference between the observed intercensal increase and the proportion of that increase that can be explained by demographic factors.

The proportion of intercensal change attributable to unexplainable change is expressed in two ways:

• As a proportion of the 2016 Census count
• As a proportion of the difference between 2011 and 2016 Census counts

It is possible for the unexplainable change to be negative if the explainable components of the increase are larger than the total intercensal difference. For example, we may expect to see a certain population in an area based on the previous Census counts, natural increase and migration but actual counted less people than this in 2016. This will result in a negative unexplainable change.

As discussed in the Change Explained by Demographic Factors, 78.6% of the intercensal increase can be explained by demographic factors. The analysis in this chapter utilises Census data and ABS Registered Births and Deaths data to approximate unexplainable change for each age cohort. To do this, an expected 2016 population has been calculated based on the 2011 Census counts adjusted for births, deaths and migration. The expected 2016 population has then been compared to the actual 2016 Census population to provide a measure of unexplainable change. This approach, known as the Framework for Measuring Unexplainable Change, is outlined in detail in the Methodology of this publication.

The age cohort analysis discussed in Overview of the Increase demonstrated that 0-4 year olds were the group that contributed most to the overall intercensal increase between 2011 and 2016.

The framework for measuring unexplainable change allows for a comparison of the number of Aboriginal and Torres Strait Islander persons aged from 0-4 as counted by the 2016 Census with registered Aboriginal and Torres Strait Islander births as recorded in the intercensal period. There was a noticeable difference in the 2016 Census count in 0-4 year olds compared to ABS Registered Births (ABS cat. no. 3301.0). Refer to this publication for information on the collection of Aboriginal and Torres Strait Islander status in births data.

Traditionally a net undercount (5.1% in 2016) of 0-4 year olds has been observed in the overall Australian population. One known reason for this is young children being mistakenly omitted from Census forms in responding households. However, it is also possible that we are observing the effects of delayed identification of children. When examining unexplainable change by age cohort, we see fewer 0-4 year olds than expected but also many more 5-14 year olds than expected. How parents and children in the same family identify is explored in depth in Changing Propensity to Identify.

Major Cities and Inner Regional areas recorded positive rates of unexplainable change between 2011 and 2016.

Conversely between the 2011 and 2016 Censuses there was negative unexplainable change in Outer Regional, Remote and Very Remote areas. This means that there were less people counted in these areas than we would expect based on natural increase and migration.

Major Cities have the largest Aboriginal and Torres Strait Islander population and are the major driver of unexplainable change at the national level, contributing close to or more than half of national cohort growth in 13 of the 18 age cohorts.

Whilst there was a small increase in total counts of persons in the 20-24 year age cohort between 2011 and 2016, once factors of explainable change are incorporated this changes to a small decrease. This main driver of the change in this cohort was positive net-migration suggesting that migration is the main influence on the growth in Census counts in this age cohort rather than changes in identification.

Inner Regional areas also made a large contribution to the overall unexplainable increase. The age cohort distribution of unexplainable change is similar to that observed for Major Cities. There are however, some key differences when analysing the size and overall contribution of the cohorts.

While there were still less 0-4 year olds in the Census than in the registered births data, the difference was proportionally much smaller than in other remoteness areas. The deficit in Inner Regional areas contributed only 12.0% of the total national unexplainable change for this cohort, which is much lower than the proportion of the overall Aboriginal and Torres Strait Islander population that live in Inner Regional areas (24.0%).

Despite recording a growth in Census counts of 7.9% between the 2011 and 2016 Censuses, there is negative unexplainable change in Outer Regional Areas. The large unexplainable increase in 5-14 year olds observed nationally and in both Major Cities and Inner Regional areas is negligible in Outer Regional areas. While growth can be observed in the actual Census counts, once net gains in inter-regional migration are accounted for there is almost no change.

Remote and Very Remote areas recorded a small increase in counts between the 2011 and 2016 Censuses (2.0%), and a negative unexplainable change is observed in most age cohorts in both Remoteness areas.

Remote areas recorded a negative unexplainable change of 4,057 persons. This was mostly due to less 0-4 year olds than expected due to demographic factors. In older age groups there was very little intercensal movement, with negligible differences for persons 50 years and older.

Very Remote areas recorded a negative unexplainable change of 5,231 persons. Unlike Remote areas this was spread across almost all age cohorts, with negative unexplainable change observed in 16 of the 18 cohorts.

One potential explanation for low population growth in Remote and Very Remote areas is inter-regional migration. However, an analysis of inter-regional migration from the Census 2016 shows that this is not a factor in the population change in these areas. The net migration impact (including overseas migration) on Remote areas was a decrease of 435 persons. In Very Remote areas the migration impact was positive with a net inflow of 202 persons from 2011.

At the national level, the proportion of the unexplainable change as a percentage of the total 2016 census count for Aboriginal and Torres Strait Islander persons was 3.3% between 2011 and 2016, slightly lower than between 2006 and 2011 (5.1%). This proportion varied by jurisdiction, with proportions higher than the national rate observed in New South Wales (9.3%), Tasmania (6.9%) and Victoria (5.1%).

The increase in the Census counts of Aboriginal and Torres Strait Islander people in Western Australia, the Northern Territory and South Australia between 2011 and 2016 was lower than expected. Once measureable change was factored in, unexplainable change was negative in these regions.

The possible reasons for Census counts being lower than expected in some geographic areas include:

• Parents may have identified their children as being of Aboriginal and/or Torres Strait Islander origin in birth registrations but not in the 2011 and 2016 Censuses, or have moved interstate after their child's birth was registered
• A person may have moved between Indigenous Regions between the 2011 and 2016 Census and did not record their previous address in the 2016 Census
• Aboriginal and/or Torres Strait Islander persons may have been missed from being counted in the Census due to variation in Census coverage and response rates.
   

New South Wales has a similar pattern of unexplainable change by age cohort to that observed nationally. It also has the largest population of Aboriginal and Torres Strait Islander Australians of the states and territories.

Minor differences can be observed in the 0-4 year age group with a relatively smaller difference in numbers of 0-4 year olds between the Census and ABS Registered Birth data in NSW. NSW contributed only 20.0% of the national difference in 0-4 year olds whereas NSW contributes 33.3% of the national population.

New South Wales also saw a small positive unexplainable change for the 20-24 year old cohort. This was driven by negative net intercensal migration in this age group which acted to counter the national trend in this cohort towards reduced rates of identification.

There was a small negative unexplainable change observed for 20-24 year olds in Victoria. This is the result of adjustments for positive net interstate migration. All other age cohorts exhibited a similar pattern to that observed nationally.

Queensland is home to almost a third of Australia’s Aboriginal and Torres Strait Islander population. Like NSW, it exhibits an almost identical unexplainable change pattern across all age cohorts to that observed nationally.

South Australia had almost no unexplainable change between the 2011 and 2016 Censuses (a decrease of 1.5%). Young children (0-4 year olds) were the main contributors to the state’s unexplainable change. Overall, there is very little evidence increased propensity to identify as Aboriginal and/or Torres Strait Islander is affecting population growth in SA.

Western Australia had the largest negative unexplainable change (a decrease of 7.3%). This was almost entirely a result of the difference in counts of 0-4 year olds in the Census in comparison with ABS Registered Births data.

Tasmania had the second highest rate of unexplainable change (6.9%) of the jurisdictions. The pattern of unexplainable change in Tasmania is in line with the national trend, with unexplainable increases in adults aged 30 years and over and children aged 5-14 years. This suggests that a changing propensity to identify may be impacting population change in Tasmania.

The Northern Territory recorded a small negative unexplainable change (a decrease of 4.0%). The impact of migration on the NT is notable, with the state losing a net 961 Aboriginal and/or Torres Strait Islander people between 2011 and 2016. Net migration loses were highest amongst persons aged 5-19 years.