3128.0 - Demography Working Paper 2001/6 - Rebasing Australia's Demographic Estimates Using The 2001 Census Of Population And Housing, 2001  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 12/11/2001   
   Page tools: Print Print Page Print all pages in this productPrint All

Comments on this paper would be appreciated by early January 2002 sent to Matt Berger at matthew.berger@abs.gov.au


INTRODUCTION

1.This paper describes changes to the method that was applied in 1996 and describe the planned output from the 2001 Census rebasing cycle.

2. The completion of the 2001 Census of Population and Housing allows estimation of Australia's resident population using this new source. These estimates (for 30 June of the census year) can be compared with the alternative estimates already produced by updating the previous census date estimate (using births, deaths and migration data). The difference between these two estimates is called the "intercensal error". Of these two estimates, the first is customarily adopted as the true estimate - that is, the true estimate at the census year is the one based on that year's census, not the one carried forward from the previous census. To overcome the break in continuity that this would entail, all population estimates updated from the previous census are then revised. This entire exercise (estimates at 30 June of census year and quarterly estimates back to September of previous census) is known as "rebasing", as the population estimates are determined from a new base - the most recent census. Rebasing is applied to estimates of Australia's population, that of each of the States (which here include the Northern Territory and the Australian Capital Territory), and that of each Statistical Local Area (SLA).


ESTIMATES - PRELIMINARY, REVISED AND FINAL

3. Population estimates appear in three different versions. These versions reflect the availability of different amounts of information regarding the components of the population estimates. The policy of publishing data which is later subject to revision is a compromise between timely availability of estimates for users and the delays in availability of the components.

4. The Estimated Resident Population (ERP) for a reference quarter is published six months after the reference period, at which time the estimates are given the status of "preliminary". When complete information is available on Category Jumpers - which is 12 months after the reference period, the estimates can be updated. Rather than updating every estimate every quarter, an annual revision takes place with the publication of the March quarter ERP. At this time the components of ERP for the previous financial year are revised. For example, when Australian Demographic Statistics, March Quarter 1999 (3101.0) was published in September 1999, the components for the reference period September quarter 1997 through to the June quarter 1998 were revised. At this time, the estimates for quarters which have components updated are given the status of "revised".

5. As noted, following each census rebasing is undertaken, which places estimates of the most recent intercensal period on a new base. At this time the estimates for 20 quarters (September of the previous census year to June of the most recent census year) become "final" - no subsequent revision is planned for the components of these estimates.

6. For census year estimates on the most recent census base, there are two releases of data - "preliminary rebased" estimates, with National and State estimates being available in June of the year following the census, and "final rebased" estimates, with National and State estimates being available in December of the year following the census. Final rebased estimates account for category jumping in the September quarter of the census year.


INTERCENSAL ERROR AND INTERCENSAL DISCREPANCY

7. There are two areas which contribute to the intercensal error:

  • errors in the census-based estimates of the population at the current or previous census date; and/or
  • errors in the estimates of any of the components of population change since the previous census.

8. National and State components of population change (births, deaths, international migration and interstate migration) are all subject to revision using the latest available data. For 1991-1996, the major revision was in interstate migration, with much smaller changes to births and deaths and almost nil for international migration. Information collected in the Census of Population and Housing allows an assessment of how much of the intercensal error was due to inaccuracies in estimates of interstate migration, using data from the two census questions on usual residence 1 year ago and 5 years ago.

9. After the intercensal error is adjusted for components of population change data revisions, the remaining (unattributable) portion is known as the intercensal discrepancy. Thus the intercensal discrepancy acts as a balancing item, that when combined with births, deaths and migration equals the change in the intercensal population estimates.


METHODS USED IN REBASING - IMPROVEMENTS FOR THE 2001 CYCLE

10. The methods used to rebase the population are described in the Demographic Estimates and Projections: Concepts, Sources and Methods, on the ABS web site. Separate outlines are given of the process for National and State population estimates and Statistical Local Area estimates. For the 2001 cycle, a change is planned that will improve the small area estimates of the Indigenous population in the census year. The proposed improvement will incorporate the undercount of the Indigenous population during estimation of the total population, rather than as a later process. This process will take advantage of new developments in sample survey estimation, and will further allow production of consistent population estimates at the SLA level. Second, the 1996 cycle used birth and death registration data on a period of occurrence basis for 1991-1996 (rather than the period of registration basis used prior to 1991). The 2001 cycle will still use period of occurrence data but will attempt to better account for the lag leading up to registration of births.


INTERNATIONAL MIGRATION DATA

11. In August 2000 the Department of Immigration and Multicultural Affairs (DIMA) announced the transition of their passenger card processing system to an imaging and scanning system. Movement to this new system has taken longer than originally planned. Of particular concern is that the passenger card processing will not be completed in time for incorporation of data into the rebasing process. This would impact upon the ability to incorporate residents temporarily overseas on census night into ERP. For preliminary estimation of SLA ERP, a modelling procedure will be needed to place residents temporarily overseas back in the SLAs of their usual residence. Production of final SLA level ERP will be delayed until detailed address information is available for residents temporarily overseas.


PLANS FOR OUTPUT

12. The following demographic estimates are proposed to be based on the 2001 Census of Population and Housing and made available during 2002 and 2003 following the release of relevant census data. All population estimates will be available by age and sex in electronic form. Unless otherwise specified, standard output for age is in 5 year age groups up to 85 and over. For June 2001 population estimates will also be available on for the first time at the SLA level by single year of age up to 100 and over. Small cells will be subject to randomisation to help ensure confidentiality of data provided to the ABS. A description of the method used in the preliminary rebasing of the June 2001 national and State population estimates (including demographic adjustments applied in the assessment of the 2001 Census net undercount) is due to be released in a Demography working paper in July 2002. A description of final rebasing will be provided in an update of Demographic Estimates and Projections: Concepts, Sources and Methods, due for release by June 2003.

13. Population estimates that will be produced are as follows:
  • Total
  • Indigenous
  • Country of birth
  • Registered marital status
  • Household and family

Detailed plans for release of data are as follows.


POPULATION - TOTAL

National and State

Preliminary estimates

14. Preliminary estimates for June 2001 based on the 2001 Census are proposed to be released in June 2002 in Australian Demographic Statistics, December Quarter 2001 (3101.0). September and December 2001 estimates based on the 2001 Census will also be released at the same time. Given that they will be revised 6 months later, preliminary estimates from September 1996 to March 2001 based on the 2001 Census are proposed not to be published but made available on request. Two June 2001 population estimates at the State level would appear - one based on the 1996 Census and the other based on the 2001 Census.

Final estimates

15. Depending on availability of data from DIMA, final estimates from September 1996 to June 2001 are proposed to be released in December 2002 in Australian Demographic Statistics, June Quarter 2002 (3101.0). Components of population growth from July 1996 to June 2001 would be finalised. Population by Age and Sex, Australian States and Territories, June 2002 (3201.0) is proposed to contain final State level June population estimates by single year of age to 85 and over for the years 1997 to 2001 inclusive. The spreadsheets in AusStats associated with Population by Age and Sex, Australian States and Territories (3201.0) will contain these estimates by single year of age to 100 years and over.

Sub-State

16. While standard output is at the Statistical Local Area level on ASGC 2001, Collection District level estimates for June 2001 will also be available on request (Collection District level estimates for June 1996 on ASGC 1996 are available).

Preliminary estimates

17. Preliminary estimates at June 2001, along with June 1991 and June 1996 estimates, on ASGC 2001 are proposed to be released in July 2002 in Regional Population Growth, Australia, 1991-2001 (3218.0).

18. Data cubes in AusStats proposed to be released at the same time on ASGC 2001 are as follows:
  • Regional Population Growth, Australia 1991-2001 - 2001 Census Edition (3218.0.55.001) - contains June 1991, June 1996 and June 2001 total population estimates; estimates for 1997, 1998, 1999 and 2000 will be available on request.
  • Population by Age and Sex, (State) June 2001 (3235.x.55.001).
  • Population by Age and Sex, (State) June 1991 and June 1996 (3235.x.55.001) - these are final estimates, the only change from previous issues of these estimates being the ASGC edition.

Final estimates

19. Final total population estimates at June 1997 to June 2001 on ASGC 2001 are proposed to be released in February 2003 as data cubes in AusStats as follows:
  • Regional Population Growth, Australia 1991-2001 (3218.0.55.001) - contains June 1991, June 1996 through to June 2001 (7 reference dates) total population estimates.
  • Population Growth and Distribution, Australia 1996-2001 (2035.0.55.001) - contains June 1991 through to June 2001 (11 reference dates) total population estimates.
  • Population by Age and Sex, (State) June 2001 (3235.x.55.001).

20. Final total population estimates at June 2001, but on ASGC 2002, are proposed to be released in February 2003 in Regional Population Growth, Australia 2001-02 (3218.0) and the associated data cube in AusStats.

21. Final total population estimates at June 1991 and June 1996 to June 2001 on ASGC 2001 are proposed to be released in May 2003 in Population Growth and Distribution, Australia 1996-2001 (2035.0).


POPULATION - INDIGENOUS

Preliminary estimates

22. Preliminary estimates of the Aboriginal and Torres Strait Islander (Indigenous) population at June 2001 are proposed to be released as follows:
  • SLA level totals in a data cube - released July 2002 in Experimental Estimates of Indigenous Australians, June 2001 (3238.0.55.001); preliminary estimates of the total population will be released for the same areas at the same time.
  • State level totals in Australian Demographic Statistics, March Quarter 2002 (3101.0) - released September 2002 in a special article.
  • Population Distribution, Indigenous Australians (4705.0) may contain estimates at the State level and possibly lower - released August 2002.

Final estimates

23. Depending on availability of data from DIMA, final estimates of the Indigenous population 1991- 2001 are proposed to be released as follows:
  • SLA level by age and sex at June 2001 as an AusStats data cube - Experimental Estimates of Indigenous Australians, June 2001 (3238.0.55.001) released February 2003.
  • SLA level by age and sex at June 2001 in the Integrated Regional Data Base 2003 - released April 2003.
  • State level at June 2001 totals in Australian Demographic Statistics, September Quarter 2002 (3101.0) - released March 2003.
  • June 2001 SLA level totals in Population Growth and Distribution, Australia, 1996-2001(2035.0) - released May 2003.
  • State, capital city/balance of State, major population centre (see table 5 of 3101.0), remoteness area and ATSIC region levels in Experimental Estimates and Projections of Indigenous Australians, 1991-2016 (3238.0) - released August 2003. Only at the State level will estimates be taken back to 1991 on the basis of the 2001 Census. State and ATSIC region level estimates will be by age and sex.

Torres Strait Islander population

24. Separate estimates of the Torres Strait Islander population are proposed to be produced at the SLA level by age and sex during the production of Indigenous population estimates. Appropriate State, capital city/balance of State, major population centre, remoteness area and ATSIC region level estimates will be published in Experimental Estimates and Projections of Indigenous Australians, 1991-2016 (3238.0) along with Aboriginal and non-Indigenous populations - released August 2003.


POPULATION - COUNTRY OF BIRTH

25. Final estimates at June 1997 to June 2001, along with June 1996, for Australia, are proposed to be released in Migration, Australia 2001-02 (3412.0) due for release in February 2003. Capital city and balances of State estimates at June 2001 are also proposed to be published.


POPULATION - REGISTERED MARITAL STATUS

26. Final estimates at June 1997 to June 2001, along with June 1996, for Australia are proposed to be released in Marriages and Divorces, Australia 2002 (3310.0) due for release in August 2003. State level final estimates are also proposed to be published.


HOUSEHOLD AND FAMILY

27. Final estimates of households at the Part of State level for June 1997 to June 2001 are proposed to be released in June 2003 in Australian Demographic Statistics, September Quarter 2002 (3101.0). Final estimates of households and families at June 2001 for capital cities/ balances of States are proposed to be included as a special article in this publication.

28. Final estimates of households and families at June 2001 at the Statistical Local Area level on ASGC 2001 are proposed to be released as an AusStats data cube Household and Family Estimates, Australia, June 2001 (3236.0.55.001) released at the same time.

Demography Program
Australian Bureau of Statistics
12 November 2001