|Page tools: Print Page Print All|
FEATURE ARTICLE 1: FINAL REBASING OF AUSTRALIA'S POPULATION ESTIMATES, SEPTEMBER QUARTER 2006 – JUNE QUARTER 2011
In June 2012 (following the process illustrated in the diagram above) the ABS used the 2011 Census results to produce 'preliminary rebased' population estimates for all previous intercensal quarters (September 2006 to June 2011), and a 'preliminary base' ERP for 30 June 2011 as documented in Australian Demographic Statistics, December Quarter 2011 (cat. no. 3101.0).
In this release, the 'preliminary rebased' population estimates have been updated to produce 'final rebased' population estimates and a 'final base' ERP for 30 June 2011.
Final rebasing is primarily a process of revising preliminary rebased data to allow for the following three factors: the updating of data related to Residents Temporarily Overseas (RTO) at the time of the Census; the application of demographic adjustments designed to resolve any statistical anomalies in the age and sex composition of the derived population; and the accommodation of any further revisions to the components of population growth (births, deaths, net interstate migration and net overseas migration) used to backdate from the Census night to the 30 June 2011 (see diagram above).
The rebasing process also gives the ABS the opportunity to review any of the intercensal components of growth (births, deaths, net interstate migration and net overseas migration) and revise these where additional data is available.
The conclusion of the final rebasing process involves constructing an ERP series for the intercensal period which takes account of the difference between the newly established 30 June 2011 ERP based on 2011 Census, and the original 30 June 2011 estimates based on adding quarterly components of growth to the 30 June 2006 Census base. This difference is distributed evenly over the nineteen intervening quarters of the intercensal period as intercensal discrepancy (discussed further below).
On this occasion the final rebasing process also included an additional one-off 'recasting' of ERP back to September 1991 to accommodate the impact of a methodological improvement in the 2011 estimates of Census undercount.
Due to the introduction of a new, more accurate method of calculating the Census base, the intercensal error for this intercensal period, as reflected in the preliminary rebased estimates, was high. This resulted in an implausible growth series, and inconsistencies in comparing population estimates between this intercensal period (September 2006 - June 2011) and other periods.
To incorporate the impact of the methodological improvement into the ERP series, and to therefore address the issue of implausible growth, the ABS has made a one-off revision to historical ERP data from September 1991 to June 2006. The process used to implement these revisions has been referenced as ‘recasting’. For more information on the ‘recasting’ process, please refer to the feature article: Recasting 20 Years of ERP in Australian Demographic Statistics, Dec 2011 (cat. no. 3101.0).
The remainder of this article provides an explanation of the adjustments made using the 2011 Census results and related rebasing processes to produce a 'final base' ERP for 30 June 2011 and 'final rebased' population estimates for the period September 2006 to June 2011.
Constructing the final ERP figure for Census night
To construct the final rebased ERP for Census night, two primary adjustments have been made since preliminary rebasing. This has involved:
Backdating from Census night to 30 June 2011
A further step in arriving at the final base ERP figure for 30 June 2011, was to revise the components of population change required to backdate the resulting figure from Census night (9 August 2011) to 30 June 2011. This involved subtracting births, adding deaths, and either subtracting or adding as appropriate the net interstate migration and net overseas migration for the period between Census night and 30 June 2011. The adjustments due to backdating from Census night to 30 June 2011 in the final rebasing process (38,800 people) are relatively consistent with those of preliminary rebasing (35,500 people).
Table 1 shows figures that have been produced, following the final rebasing process, which include updated RTOs and demographic adjustments.
These calculations were undertaken to ensure that population estimates for the next intercensal period (i.e. 2011-2016) are as accurate as possible.
REVIEW OF INTERCENSAL COMPONENTS OF GROWTH
As noted above, the final rebasing process affords the ABS an additional opportunity to revise births, deaths, net interstate migration and net overseas migration (NOM) estimates for the 5 year intercensal period with any updated data available to the ABS. In this final rebasing cycle, births, interstate migration and NOM were revised based on updated data.
Revisions to Births Data
The annual births and deaths revision cycle is designed to estimate all 'vital events' occurring in a reference year, including lagged events - that is births and deaths which have occurred but not yet been registered with the relevant state and territory Registrar of Births, Deaths and Marriages. For this rebasing cycle, only births were revised (death figures did not change significantly enough to warrant a revision). In addition to revising births to accommodate lags as indicated above, the ABS has taken the opportunity to incorporate approximately 33,000 unprocessed NSW birth registrations as a result of an ABS systems processing error (see the technical note Effect of previously unprocessed New South Wales births, which appeared in Births, Australia (cat. no. 3301.0)) . Around 10,300 of these relate to the intercensal period (2006-2010), while the remainder relate to earlier periods and will be accounted for in the recasting exercise. The magnitude of the revision is indicated in tables 2-5 below.
These tables indicate that the 2009-10 financial year had the highest level of revision with the previous births figures being understated by 12,800 (4.4%) across Australia. The inclusion of the previously unprocessed births in NSW contributes significantly to the difference over the period. Over the whole 2006-2011 intercensal period, NSW births were understated by a total of 6.4%. Tasmania had the next highest change in births estimates as part of this final revision process, having been underestimated by 1.25% over the period.
Revisions to Interstate Migration Data
It is standard practice at final rebasing to revise the modelled interstate migration estimates during the intercensal period with data from Census questions related to an individual's place of usual residence one year ago, and five years ago and at Census night. For more information, see Population Estimates: Concepts, Sources and Methods, 2009 (cat. no. 3228.0.55.001).
Revisions to Net Overseas Migration Data
The ABS has completed work on the rebuilding of its Overseas Arrivals and Departures (OAD) processing system with the aim of improving the quality of OAD data and the quality of the imputations used within the processing system. The primary result of this rebuild is improvements to the preliminary estimation of NOM, which is based on imputations of traveller behaviour. However, the rebuild has also resulted in an improvement to final NOM data and these improvements have been included as revisions in the final rebasing process. The table below indicates the impact. The improvements to preliminary NOM estimation will be released with Australian Demographic Statistics, June Quarter 2013 (cat. no. 3101.0) scheduled for release on 17 December 2013. More detail on the improvement to preliminary NOM estimation will be provided in the Information Paper: Further Improvements to Net Overseas Migration Estimation (cat. no. 3412.0.55.002) scheduled for release on 17 December 2013.
UPDATING THE INTERCENSAL ERP FIGURES DURING THE 'FINAL REBASING' CYCLE
To construct a final ERP base figure for Australia and each of the states and territories at 30 June 2011, the nineteen intercensal quarters preceding June quarter 2011 (i.e. September quarter 2006 - March quarter 2011) were recalculated to produce 'final rebased' population estimates.
Inevitably there is a difference between the newly established 30 June 2011 ERP based on 2011 Census, and the original 30 June 2011 estimates obtained by adding quarterly components of growth to the 30 June 2006 Census base. This difference is known as the Intercensal Error and is distributed evenly over the nineteen intervening quarters of the five year intercensal period. The population at the end of the first quarter (September 2006) is adjusted up or down by 1/20 of the intercensal error, the second quarter is adjusted by 2/20, the third by 3/20, and so on to the 19th quarter, whilst the final quarter (the new base figure) does not require adjustment. Details of the intercensal error for the 2006-11 intercensal period are available in table 7 below.
The main reasons for updating these nineteen quarterly intercensal ERP figures are to provide a more accurate population estimate for each quarter, and also to ensure that the estimates from the 2006-2011 intercensal period will be comparable with all future estimates, thus creating a consistent time series of ERP data.
These adjustments complete the final rebasing process and, following this, it is not expected that any subsequent revisions will be made to these figures.
Intercensal Error and Intercensal Discrepancy
As noted above, the intercensal error refers to the difference between the preliminary rebased (2011 Census-based) ERP figures for 30 June 2011 and the unrebased figures which have been derived through the intercensal period using the components of growth - births, deaths and migration data.
There are two components that contribute to the intercensal error:
After the intercensal error is adjusted for all revisions to component data, the remaining unattributable portion is referred to as the intercensal discrepancy. The estimate of intercensal discrepancy for each state and territory, birth cohort and sex are spread evenly across the intercensal quarters. Thus the intercensal discrepancy acts as a balancing item, that when combined with births, deaths and migration equals the difference between the two 30 June Census year population estimates.
The table below indicates the Preliminary Rebased Intercensal Error published prior to recasting, followed by the Preliminary Rebased Intercensal Error that applies when the effect of the PES methodological changes that led to the recasting process has been removed, and thereafter the Final Intercensal Discrepancy for the 2006-2011 period.
Table 8 shows 'final intercensal discrepancy' by five year age groups.
The age group with the highest level of discrepancy was the 0-4 year age group with a discrepancy of 41,100. This was followed by the 15-19 year age group with a discrepancy of 26,300. The smallest discrepancy occurred in the 60-64 year age group with a figure of negative 800.
Although 'final intercensal discrepancy' is reported as a quality indicator on the unrebased 30 June Census year population estimate, it is considered to have accumulated over the entire intercensal period.
NEW TERMINOLOGY FOR 2016 REBASING
Given the increasing prominence of ERP over time, the ABS is aware that the terms 'intercensal error' and 'intercensal discrepancy' are often misinterpreted. While there is considerable information available to explain what these demographic terms refer to, the word 'error' is too commonly considered to be a synonym for 'mistake'. As a result, the ABS will use the terms 'preliminary intercensal difference' and 'final intercensal difference' in the 2016 rebasing cycle.
PLANS FOR FURTHER OUTPUT AND RELATED PRODUCTS
Following this issue which contains final population estimates for Australia, states and territories based on the 2011 Census, it is expected that no subsequent revisions to the 2006-2011 intercensal period will be made.
Final rebased sub-state estimates for 2006-2011 will be published on 30 August 2013 in Regional Population Growth, Australia, 2012 (cat. no. 3218.0), along with recast estimates for 1992-2006.
Final rebased estimates of the Australian Aboriginal and Torres Strait Islander population at 30 June 2011 will be published in Estimates of Aboriginal and Torres Strait Islander Australians, 2011 (cat. no. 3238.0), on 30 August 2013.
Australian Demographic Statistics, March 2013 (cat. no. 3101.0) will be released on 26 September 2013 and will contain updated summary tables from the two above mentioned releases.
These documents will be presented in a new window.