Australian Bureau of Statistics
3106.0.55.001 - Information Paper: Review of Interstate Migration Method, March 2009
Latest ISSUE Released at 11:30 AM (CANBERRA TIME) 18/03/2009
|Page tools: Print Page Print All RSS Search this Product|
In accordance with legislative requirements, the ABS provides quarterly estimates of the population of Australia and each of the states and territories. These population estimates, known as the estimated resident population (ERP), are published each quarter in Australian Demographic Statistics (cat. no. 3101.0).
The Census of Population and Housing provides the benchmark for post-censal population estimates once every five years. Post-censal ERP is then calculated forward each quarter by using the quarterly estimates of each of the three components of population change: natural increase (the excess of births over deaths); net overseas migration; and net interstate migration. Post-censal estimates of births and deaths are calculated using birth and death registrations, while overseas migration is calculated using data collected from passenger cards, visas and passports for those persons travelling into and out of Australia. Data on interstate migration, however, are not directly estimated. Instead, post-censal estimates of net interstate migration are modelled using administrative by-product data. This document describes the most recent review of the method used to produce quarterly post-censal estimates of interstate migration.
Currently, the data used by the ABS is information on change of address advised to Medicare Australia and to the Department of Defence in the case of the military. Every five years, after data from the following Census has been finalised, these modelled estimates are then reviewed against, and potentially replaced by, the interstate migration estimates that are calculated from the Census (i.e. rebased to the Census). This is known as the re-derivation of interstate migration and is mentioned later in this paper.
Strength of Medicare data as an indicator
The ABS has evaluated a range of potential sources of administrative data for estimating interstate migration on a quarterly basis. Medicare Australia data supplying change of address information was found to be the most effective source currently available. For more information, refer to the Information Paper: Evaluation of Administrative Data Sources for Use in Quarterly Estimation of Interstate Migration (cat. no. 3127.0.55.001).
Medicare card holders are required to register changes of address when they make claims, or when their cards are replaced. However, it is known that some people, particularly younger card holders, do not register changes of address with Medicare, or do so long after the fact. Comparison of Medicare change of address with census data on the address of respondents one year prior to the Census suggests that the level of under-reporting in Medicare is fairly constant over time. In addition, this under-reporting seems to be similar for interstate arrivals and interstate departures, as well as for each sex by various ages.
Defence force adjustment
Medicare theoretically covers all Australian usual residents as well as those non-Australian residents granted temporary registration. However, there are a range of Australian usual residents who do not access the Medicare system, primarily due to access to alternative health services. One such group is the military. As such, estimates of interstate migration produced from the interstate migration model are adjusted to compensate for defence force movements not covered by Medicare. These adjustments are estimated using counts of defence force personnel by age, sex and state/territory, obtained from the Department of Defence, with 70% of any change in quarterly defence force numbers assumed to be due to interstate migration not otherwise covered by the model.
2. REBASING AND RE-DERIVATION OF INTERSTATE MIGRATION
Due to the non-compulsory and non-universal nature of the available (indirect) data sources, post-censal quarterly estimates of interstate migration have long been considered the weakest component of population change. For this reason, the model for generating post-censal estimates of interstate migration is largely superseded when new census information becomes available (i.e. rebased to the Census).
Part of the process of rebasing census ERPs for states and territories is the re-derivation of interstate migration for the intercensal period. The overall approach is to minimise state intercensal discrepancy using information from the two census questions on usual residence one-year ago and five year ago to estimate interstate movements. Where this census information does not reduce the intercensal discrepancy, the rebased interstate migration estimates remain largely unchanged from the Medicare-based model.
For example, during rebasing of interstate migration estimates to the 2006 Census as seen in Table 1, for New South Wales, Victoria and Queensland the Medicare-based estimate was used in the re-derivation as it reduced intercensal discrepancy. Whereas for South Australia, Western Australia and the Australian Capital Territory the Census-based estimate was used in the re-derivation. For Tasmania and the Northern Territory a combination of Medicare and Census-based estimates were used to reduce the intercensal discrepancy. In these cases where the intercensal error was negative and closer to zero than the difference then the intercensal error was subtracted from the Medicare-based estimate and used for the re-derivation.
Interstate flows however, must always sum to zero, as for each arrival there should always be a corresponding departure. Therefore, an additional change is made to scale interstate migration levels for each state and territory so the total will sum to zero at the Australia level as seen in Table 1. This adjustment also takes into account all movements for each state and territory before prorating.
When the intercensal discrepancy is finalised the difference between the original interstate migration estimates and the rebased estimates is apportioned across all quarters, movement categories, ages and sexes in the intercensal period in order to minimise quarterly change.
TABLE 1. REBASING INTERSTATE MIGRATION ESTIMATES — 2001-06
Post-censal quarterly estimates of net interstate migration are created for the states and territories using interstate change of address advised to Medicare Australia and to the Department of Defence in the case of the military. Medicare data are adjusted by means of expansion factors.
The ABS has reviewed the interstate migration model using finalised data from the 2006 Census of Population and Housing. The outcome of this current review is similar to the previous model used to estimate interstate migration from 2001 to 2006. This review has focussed on the expansion factors applied to Medicare data. The defence force adjustment applied to interstate migration estimates was not reviewed.
Calculating expansion factors for Medicare data
Expansion factors applied to Medicare data were based on the estimated proportion of the population covered. To calculate this proportion, Medicare data for the period October 2005 to September 2006 was compared with 2006 Census data on persons' usual residence one year ago as seen in the graph below. The period of Medicare data chosen reflected the assumption that there is a lag in the registration of change of address through Medicare.
Expansion factors, as shown in the equation below, were estimated for each state and territory, by sex, single year of age and movement category (arrivals/departures). Analysis showed that for all states and territories expansion factors on average, were greater than 1 (i.e. suggesting 'undercoverage' in Medicare data) for males aged 17-30 years and females aged 17-25 years. These age ranges were chosen as they decreased intercensal discrepancy the most when compared to any other range. Expansion factors are therefore applied universally for all states and territories to Medicare data with these age-sex characteristics. For all other age groups, it is assumed that Medicare data provides a full coverage (i.e. an expansion factor of 1 was used). The equation used to calculate the expansion factors was:
Multiple movers and census data
Some people will move more than once during a given time period, including some who move to one location, then return to their original location. However, the Census asks for each person's address one year ago without reference to multiple moves which could have occurred over the one year period.
Since Medicare data do include information on multiple/return movements, these were used as a proximate to adjust data from the 2006 Census. The proportion of multiple movers identified using Medicare data was calculated by dividing all quarterly movements by the number of final annual movements for each state or territory, by sex, single year of age and movement category (arrivals/departures). Calculations showed that 7% of all movements were multiple movements. Those who had moved twice represented 6% of all movements whereas those who moved more than twice only represented 1% of all movements.
The use of census data on address one year ago does introduce some potential problems because a relatively short (and therefore potentially more volatile and less representative) time period is used to estimate expansion factors. Smoothing used in generating the expansion factors, as seen in the many scenarios analysed, helps to address these problems.
4. SCENARIOS ANALYSED
A number of scenarios were developed (although not all are listed in this paper) using Medicare data and data from the 2006 Census. The complete process of estimating interstate migration for the intercensal period (2001 to 2006) and final rebasing was replicated for each scenario. These scenarios vary in their use of expansion factors, lagging, smoothing and other adjustments, and were each used to estimate June 2006 ERP by state/territory. The extent to which each scenario matched final June 2006 ERP, or in other words reduced intercensal discrepancy, was used as an indication of its accuracy in modelling interstate migration (refer to Scenario outcomes below). The reasonableness of assumptions in each scenario was also taken into account.
Some of the scenarios used in the interstate migration review are summarised in the following table with their components described in more detail further in this paper.
TABLE 2. INTERSTATE MIGRATION REVIEW, Scenarios analysed
Lagging of Medicare data
Analysis has shown that registration of changes of address through Medicare generally occurs some time after the actual move. The interstate migration models for both 1996-2001 and 2001-2006 assumed an average registration lag of three months (one quarter), so Medicare information for a particular quarter was used to estimate interstate migration for the previous quarter. It is not possible to lag this data further (i.e. to assume a delay in registration of more than three months) as this would impact on the production and publication of population estimates which are released within 6 months after the reference period.
The first scenario considered (S1) assumes that Medicare changes of address perfectly capture all interstate movements (i.e. no undercoverage), and that there is no delay between persons moving address and registering a change with Medicare (i.e. no lagging). These assumptions are known to be false, but are used to provide a baseline for later scenarios. While the results from S1 appeared reasonable for some states when compared with the actual intercensal discrepancy as seen in Table 3, this may be due to errors in interstate arrivals and departures (such as those due to undercoverage and lagging) cancelling each other out.
Comparison of the outcomes of most scenarios tested (although not all are listed in this paper) indicated that the use of lagging reduced intercensal discrepancy particularly at the overall level as seen in the sum of absolutes in Table 3. Later scenarios tested alternative methods of applying expansion factors (smoothing input data and/or smoothing the factors produced), adjusting census data for the age at move, as well as the impact of imposing an upper limit (or 'cap') on the factors.
It was assumed that much of the variability in data from one single-year age group to the next was due to 'random noise', so smoothing was incorporated in some form in many of the scenarios analysed. Two options for smoothing were used in the production of some expansion factors. Firstly, for each state/territory, sex and movement category, all the input data used to calculate expansion factors were smoothed across single years of age, using a three-term moving average as applied in scenarios S5 to S9. The second option was to smooth the actual output (or expansion factors), again using a three-term moving average as applied in scenarios S8 to S10.
Overall, smoothing of expansion factors produced only small improvements in the estimated levels of intercensal discrepancy; in some cases, results were worsened when the factors were smoothed. Greater improvements (represented by smaller intercensal discrepancies) were gained by smoothing the inputs of Medicare and census data used to produce the expansion factors. For most states and territories, the best results were found in those scenarios from which expansion factors were created using smoothed input data.
Capping expansion factors
In the interstate migration model used for 1996-2001, expansion factors calculated as being greater than 2 (i.e. less than 50% coverage estimated for Medicare data) were set to 2. The rationale for 'capping' expansion factors was that this would reduce the influence of outlying extreme results, such as unusually low registrations in particular age/sex groups. In the 2001-06 review capping was applied to expansion factors of some scenarios but did not produce improvements in the intercensal discrepancy.
For this review (2006-11) which has used the most recent data available, analysis of the various scenarios and subsequent expansion factors showed that capping produced improvements in most cases. Scenario S7 (the preferred model) is an improvement on scenario S6 by using capping which has decreased intercensal discrepancy. The only outlying group of interstate movers for which capping has applied during this review is males aged 23 to 25 departing the Northern Territory.
Census data adjustment - age at move
The age of interstate migrants taken from census information on a persons usual residence one year ago is the age at census night - not the age at move. Therefore this adjustment assumes half of the population were one year younger at date of move than at date of the Census. It was calculated by pairing each consecutive single year of age and dividing by two as seen in the equation below:
While this adjustment provided similar results to the preferred scenario, it was not chosen as it did not reduce the intercensal discrepancy for the state with the highest counts. For example persons for Victoria was slightly higher in scenario S5 than in scenario S7 (the preferred model) as seen in Table 3.
5. SCENARIO OUTCOMES
Table 3 below shows the actual intercensal discrepancy (both in terms of persons and as a percentage of final 30 June 2006 ERP) for 2001-06. It also shows the intercensal discrepancy produced under each scenario tested. For each scenario the complete process of estimating interstate migration for the intercensal period (2001 to 2006) and final rebasing has been replicated. The following table summarises these outcomes for each scenario.
TABLE 3. INTERCENSAL DISCREPANCY, Actual and Scenario examples — 2001 to 2006
Assessing Model/Scenario quality
The alternative interstate migration models (scenarios) were assessed using two main criteria.
Results from scenarios 1-10 indicated that the preferred interstate model should contain the following characteristics:
This preferred model was reflected in scenario seven (S7) and provided the best overall results. Although the various scenarios display little variation to each other as seen in Table 3, the preferred scenario (S7) was able to reduce the overall intercensal discrepancy while still reducing it slightly for Victoria, the state with the highest count. In addition S7 was able to improve on the actual original intercensal discrepancy by reducing the sum of absolute percentage values from 3.21 (actual) to 3.16 (S7).
The ABS will apply this model to produce interstate migration estimates each quarter for the current intercensal period (i.e. September quarter 2006 to June quarter 2011) and onwards until once again a review is undertaken after data from the 2011 Census of Population and Housing has been finalised. A defence force adjustment will continue to be applied to estimates produced by this model to compensate for movements of defence force personnel not covered by Medicare.
6. PREVIOUS INTERSTATE MIGRATION METHODS USED
For information on the method used for estimating interstate migration for the intercensal period 1996 to 2001 see: Demography Working Paper 1999/2 - Estimating Interstate Migration, 1996-2001 (cat. no. 3116.0). The method used in estimating interstate migration for the most recent intercensal period 2001 to 2006 is detailed in Demography Working Paper: 2004/1, Review of Interstate Migration Method, May 2004 (cat. no. 3106.0.55.001).
For further information on the interstate migration method, contact Neil Scott on Canberra (02) 6252 5959, or email your comments to the attention of Neil Scott at firstname.lastname@example.org.
ATTACHMENT 1: ERP AND COMPONENTS OF CHANGE, 2001 to 2006
Tables A1 and A2 below show how the demographic balancing equation is calculated between the 2001 and 2006 Censuses. Two tables have been presented to be able to clearly show data for all states and territories. Each table contains preliminary net interstate migration (unrebased) and final net interstate migration (rebased). They show how each contribute to ERP and help generate an intercensal error or an intercensal discrepancy. For example, preliminary net interstate migration which has used modelled Medicare data for quarterly interstate migration estimates, will contribute to the intercensal error. Whereas intercensal discrepancy is the remaining unattributable difference between the two different ERP counts at 30 June 2006. One count is after all components of population change (including interstate migration) have been finalised and the other count is the final ERP at 30 June 2006 based on the 2006 Census.
TABLE A1. ERP AND COMPONENTS OF CHANGE — Selected States, 2001 to 2006 ('000)
TABLE A2. ERP AND COMPONENTS OF CHANGE — Selected States, 2001 to 2006 ('000)
These documents will be presented in a new window.
This page last updated 23 May 2011