3228.0 - Demographic Estimates and Projections: Concepts, Sources and Methods, 1999  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 30/08/1999   
   Page tools: Print Print Page Print all pages in this productPrint All  
Contents >> Chapter 3. Statistical Local Area population estimates >> Disaggregation of post-censal SLA population totals by age and sex

3.21. Post-censal estimates of the age and sex distributions of SLA populations are made by updating the population by age and sex for the census year using annual births (by sex), deaths (by age and sex) and derived age and sex profiles of migration.

3.22. While annual data on births and deaths by age and sex are available for each SLA, data on migration into and out of SLAs for post-censal years are not available and have to be derived indirectly. This is done as follows:

Step 1.

3.23. The estimate of total population growth for each SLA (see above) for the twelve months is split into natural increase and net migration components. Natural increase is derived for each SLA from birth and death registration statistics. Net migration is derived for each SLA as the difference between total population growth and natural increase. Net migration is then split into internal and overseas migration components. This is done by apportioning according to the relative contributions based on 1996 Census one-year migration data, in conjunction with State-level estimates of interstate migration and passenger-card-derived overseas migration.

3.24. The SLA age/sex profiles of internal migration are derived from 1996 Census data on the SLA of usual residence one year ago. These profiles were produced for:

      • inter-SLA arrivals - persons residing in the SLA whose usual residence one year ago was in another SLA;
      • inter-SLA departures - persons whose SLA of usual residence one year ago was that SLA but whose residence at the date of the census was another SLA.

Step 2.

3.25. These SLA age/sex arrival and departure profiles were then constrained so that for each age and sex the net effect across all SLAs in a State equalled the finalised 1995-96 interstate migration estimate.

Step 3.

3.26. An SLA's age/sex overseas arrivals profile was derived from 1996 Census counts for that SLA of people whose usual residence one year ago was overseas. The departure profile for each SLA is assumed to be the same as the arrival profile in the absence of age/sex data on overseas departures at the SLA level from either the census or outgoing passenger cards. For arrivals, the total of all SLAs within a State is constrained to the age/sex profile of permanent and long-term arrivals for 1995-96 for the State. For departures, the total of all SLAs within a State is constrained to the age/sex profile of permanent and long-term departures for 1995-96 for the State.

3.27. As census data on usual address one year ago cannot cover those born less than one year ago, all migration for those aged zero was assumed to be half that of one year-olds.

3.28. In greater detail, estimating SLA migration involves the following procedures:

Procedure 1 - Estimation of net migration totals during year t to t+1 for each SLA

3.29. This step requires population estimates for each SLA as at year t and t+1, and recorded births and deaths occurring during the period t to t+1.



(1)
where P denotes the total population, B births, D deaths, and N net migration during the one year from t to t+1.

Procedure 2 - Split of SLA net migration into four components

3.30. Net migration obtained in equation (1) is then split into internal arrivals and departures and overseas arrivals and departures. Thus,



(2)
where IA and ID are internal arrivals and departures respectively, and OA and OD are overseas arrivals and departures respectively.

3.31. The first estimate of the four components of migration is obtained by multiplying the total population at time t+1 by a 'movement rate' specific for each component and calculated from 1996 Census movement data for SLAs for the period 1995-96. Thus,


(3)
where is the 'movement rate' for internal arrivals and is obtained from

(4)
3.32. ID, OA and OD are calculated in the same way. Once a first estimate of each of the four components of migration has been obtained (equation (3)), a plus-minus iterative proportion fit (IPF) procedure is used to satisfy both equation (2) and to ensure that when all SLAs in a State are added, the four components equal the State levels. Specifically, SLA OA and OD levels must total respective State overseas migration levels, while net SLA internal migration (IA minus ID) must sum to each State's net interstate migration. The plus-minus IPF procedure is explained in Appendix 7.

Procedure 3 - Determination of the age-sex profile for each component of migration

3.33. The census-based SLA age-sex profiles for overseas arrivals and departures and internal arrivals [see paragraph 3.26] are prorated to the SLA OA, OD and IA totals for the year t to t+1 calculated in the previous step.

3.34. However SLA departures are a function of existing SLA population, and so census-based age-sex specific departure rates are employed. As census data, by definition, excludes residents who have left Australia, the requisite data on usual address one year ago is only available for internal (ie inter-SLA) departures, not overseas departures.

3.35. SLA age-sex internal departures for the year t to t+1 are obtained by multiplying the population for each single year of age and sex, who survived to the year being estimated, by the age-sex specific departure rate. Thus for each age and sex,



(5)
where P is population and IDi is the internal departure rate and is obtained from the latest census.

3.36. was initially calculated using 1996 Census-based data (so, t = 1996) as follows:



(6)
3.37. The departure rate in (6) is 1995-96 internal departures as a rate of survived 1995 population (ie. the 1996 population with migration removed). This can then be used in (5) to create, say, 1996-97 internal departures by applying it to survived 1996 population.

3.38. Once an initial age-sex estimate of IA, ID, OA and OD has been obtained for year t+1, an IPF procedure is used to satisfy both the SLA migration component totals [equation (3)] and to ensure that when all SLAs in a State are added, the four components equal the State age-sex levels. Specifically, for each age and sex, SLA OA and OD levels must total respective State overseas migration levels, while net SLA internal migration (IA minus ID) must sum to each State's net interstate migration.

3.39. At the conclusion of each annual iteration of the SLA age-sex estimation process, the SLA-specific OA, OD, IA and ID age-sex migration profiles are refined to correct unsustainable migration patterns.

Procedure 4 - Age-sex SLA population estimates

3.40. Having established estimates of the migration component, the census date population estimates for each SLA by age and sex are then updated in the normal way, (ie. after converting to financial year of birth, when necessary - by adding births, subtracting deaths and adding/subtracting migration). A more detailed account of this procedure at the national/State level is given in Chapter 2.






Previous PageNext Page