Methodology used in rebased population estimates, June 2021

Methods and quality assurance of preliminary rebased estimated resident population (ERP) at 30 June 2021 and ERP revision back to 30 September 2016


Rebased ERP

Rebasing is the revision of the estimated resident population (ERP) based on the latest Census.

Australia's ERP includes all people who usually live in Australia (regardless of nationality, citizenship or visa status).

Between Censuses, we calculate quarterly ERP by adding the flows of population change (births, deaths and migration) to the stock population estimate (the base).  

The ABS undertakes a Census of Population and Housing (the Census) every five years as required by the Census and Statistics Act 1905.

The Census helps to create a base population estimate, which we update each quarter for the following five years in ERP, until the next Census when we repeat the process.

A regular Census:

  • enables the maintenance of an accurate time series of Australia's population estimates
  • limits the accumulation of errors in ERP.

The difference between this new rebased ERP and the quarterly component-based estimate (known as unrebased ERP) is called ‘intercensal difference’.

To ensure a consistent time series, we revise quarterly ERP between the current and previous Censuses.

The preliminary rebased population estimates are available in National, state and territory population, December 2021.

Final population estimates will be available in June 2023, following final revisions to components in the rebasing process.

Estimation of the base population

We calculate the preliminary rebased ERP at 30 June in a Census year by:

  1. adjusting Census counts of Australian usual residents to account for people missed or counted twice in the Census (based on Post Enumeration Survey results)
  2. demographically adjusting Census counts or net undercount estimates based on coherence with other data sources and to remove implausible demographic features
  3. adding usual residents temporarily overseas on Census night
  4. adjusting for births, deaths and migration from 30 June to Census night.

This new preliminary rebased ERP becomes the base for quarterly estimates going forward from that point until the next Census.

The steps are illustrated in the following diagram.

Diagram of the estimated resident population rebasing process

Diagram of the ER rebasing process
The diagram shows how different elements of the ERP rebasing process are connected. The process is in two parts and begins with the Census numbers for persons at place of usual residence at census night. The first part of the process starts with adding in persons who were missed by the Census by using a number of estimates. PES net undercount estimates are included first. Then Australian residents who are temporarily overseas are added in. Demographic confrontation with other sources of population datasets also informs how many persons were missed. This first process gives us the ERP for Australia at Census night. The second part of the process is about taking the ERP at Census night back to 30 June 2021. Births that have occurred in between this period are removed. Deaths that have occurred in the same period are added back into the population count. Persons who have moved between states and territories from 30 June to Census night are moved back. Persons who have moved into or, out of, Australia during this period are also moved back. That is how we arrive at ERP for 30 June 2021.
Components of rebased ERP, preliminary - 2021
 Persons ('000)
Census count, place of usual residence8,072.26,503.55,156.11,781.52,660.0557.6232.6454.525,422.8
Including people who have been missed
 Plus – net undercount(b)-2.620.950.
 Plus – demographic adjustment13.
 Plus – residents temporarily overseas13.
Equals – ERP at Census night8,095.76,545.95,220.61,802.82,751.0567.8249.0453.425,691.1
Backdating to 30 June 2021
 Minus – births11.
 Plus – deaths7.
 Minus – net interstate migration-0.8-1.22.2-0.10.2-0.1-0.30.1..
 Minus – net overseas migration-1.8-4.6-2.5-0.6-0.9-0.1-0.1-0.5-11.1
Equals – ERP at 30 June 20218,093.86,548.05,217.71,803.22,749.9567.9249.2453.625,688.1

(a) Includes Other Territories comprising Jervis Bay Territory, Christmas Island, the Cocos (Keeling) Islands and Norfolk Island.
(b) Includes Census gross undercount and gross overcount from the PES, and minor adjustments to ensure coherence with ERP.
.. not applicable

Census count, place of usual residence

The ABS held the Census on Tuesday 10th of August 2021. All people in Australia on this night should have been included on a Census form (except foreign diplomats and their families).

The Census count by place of usual residence excludes people in Australia on Census night who were visiting Australia for less than one year. Usual residence is the place a person has lived, or intends to live, for 6 months or more in 2021.

For more information see 2021 Census of Population and Housing.

PES net undercount

Shortly after the Census, the ABS ran a Post Enumeration Survey (PES) to independently measure how well the Census counted people in Australia on Census night. The ABS used the PES results to determine how many people should have been counted in the Census, how many people were missed (undercount) and how many people were counted more than once or in error (overcount). We combined these measures to produce the net undercount of the Census, which is the difference between the PES estimate of how many people should have been counted in the Census and the actual Census count (including imputed persons).

The 2021 Census net undercount was 0.7% (190,044 persons).

For more information see 2021 Census overcount and undercount.

Residents temporarily overseas

Residents temporarily overseas (RTOs) are Australian residents who were temporarily overseas on Census Night and who should be added to the ERP. A departing resident is excluded from Australia's usual resident population only when they leave Australia for 12 months over a 16-month period.

The preliminary estimate of RTOs on 2021 Census night was 36,629 people. The preliminary estimate models the duration of overseas stay of departing residents. We will calculate the final estimate of RTOs when data for actual arrivals and departures up to 16 months after Census are available, and include this in the final rebased ERP.

Demographic adjustment

Demographic adjustments are a range of adjustments made to the 2021 Census counts and PES estimates to ensure better internal consistency and coherence with:

  • ERP based on the previous Census and PES
  • other population estimates
  • administrative datasets.

The adjustments address issues such as:

  • age misreporting
  • unreliable estimation of the centenarian population
  • implausible sex ratios
  • implausible features due to sampling error in the PES.
Demographic adjustment of PES estimates for selected age groups
 Original sex ratioAdjusted sex ratioTotal population adjustment (‘000)
0 to 4107.2105.816.8
5 to 9106.6105.98.0
25 to 29100.4101.78.5
30 to 3499.098.08.4
PES estimates of the Northern Territory by sex

We adjusted the sex ratio for the Northern Territory from 101.5 to 103.3, with no change to the total population size.

Estimates of population aged 95 and over

We adjusted the population estimates of persons aged 95 and over using a survivor ratio method (see New estimates of Australia’s centenarian population). The survivor ratio method uses highly reliable deaths data from administrative sources, and combined with survivorship trends, produces more useful estimates of population at the highest ages. The adjustments were minor, also demonstrating the high quality of Census 2021 data at the highest ages.

Data confrontation and coherence sources

The data confrontation and coherence sources used to inform demographic adjustment were:

Component based population estimate

The ABS maintains a population account independent of the Census, based on administrative data: birth and death registrations, and net overseas migration (NOM) derived from overseas arrivals and departures data. The component based population estimate was formerly referred to as the ABS demographic databank.

2021 Census counts

The 2021 Census counts on a usual residence basis are useful for analysing demographics that are not influenced by undercount adjustments.

Unrebased 30 June 2021 ERP

This ERP is based on the 2016 Census, which has been updated with components of population change between 2016 and 2021.

Composite population estimate

A population estimate combining:

  • The unadjusted, rebased population at Census night 2021, weighted at 60%, and
  • The unadjusted, unrebased population at Census night 2021, weighted at 40%. We derived this from ERP at Census night 2016, based on PES 2016 without demographic adjustments, and updated with components of population change between 2016 and 2021.
Multi-Agency Data Integration Project (MADIP)

MADIP data is a secure data asset combining administrative data on health, education, government payments, income and taxation, employment, and population demographics (including the Census) over time. It includes people actively interacting with government administrative systems, scoped for usual residents. More information in MADIP data and legislation | Australian Bureau of Statistics (

Australian Electoral Commission (AEC) enrolments

AEC enrolments data is the count of enrolled citizens aged 17 years and over and is continually updated.

School enrolments

Data from ABS' publication Schools is based on enrolments in all government and non-government schools in Australia during 2021.

Intercensal difference

Intercensal difference is the difference between the new Census-based estimate (known as rebased ERP) and the quarterly component-based estimate (known as unrebased ERP). We calculate it by state, sex, and cohort. 

Intercensal difference may result from error in the start population estimate, the end population estimate and/or in the quarterly estimates of births, deaths, or migration. It is generally not possible to attribute intercensal difference to a particular source, or to quantify the error that each source contributes.

We revised quarterly population estimates between the current and previous Censuses by assuming intercensal difference accumulated evenly across the five-year period, except for certain cohorts in Tasmania and the Australian Capital Territory. For these cohorts, we allocated more intercensal difference to quarters before June 2020.

Our data confrontation showed that error in the state and territory allocation of migration estimates contributed to the relatively large intercensal difference in Tasmania and the ACT, for cohorts aged 20 to 35 at 30 June 2021. Given the international border closure in February 2020, and state border crossing restrictions, this error was more likely to have accumulated before then. The adjustment makes the growth rates over time more plausible.

For the 2016–2021 period, the preliminary intercensal difference was 52,211. We will produce final intercensal difference when we produce final rebased ERP, and revise all components of growth for the 2016-21 period.

Final intercensal difference, by state/territory - 1991 to 2021 (a)
 Persons ('000)
 Percent of population

(a) Negative numbers mean that the unrebased ERP underestimated the population. Positive numbers means that the unrebased ERP was an overestimate.
(b) Includes Other Territories. Excludes Other Territories prior to 2016.
(c) Preliminary intercensal difference.

Preliminary intercensal difference, by five-year age groups, 2021(a)
 Persons ('000)Percent of population
85 and over14.52.6

(a) Negative numbers mean that the unrebased ERP underestimated the population. Positive numbers means that the unrebased ERP was an overestimate.

Back to top of the page