3110.0 - Demography Working Paper 1996/2 - 1996 Review of Demography Statistics, 1996  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 15/07/1996   
   Page tools: Print Print Page Print all pages in this productPrint All

Demography Working Paper 96/2

1996 REVIEW OF DEMOGRAPHY STATISTICS

John Paice and Lisa Anderson
Demography Section
Australian Bureau of Statistics
PO Box 10
BELCONNEN ACT 2616



      Contents
      Page


Attachment A - List of organizations which have provided comments

1 EXECUTIVE SUMMARY

The current review updates one completed in 1990 which set the strategy for future work in the area of demography statistics.

The current review revisited these issues and examined them from a user relevance and priority perspective. The method used a position paper prepared by the ABS which was then circulated to a wide range of agencies and organisations in all three tiers of government and elsewhere. Some 73 organisations responded with their comments. Discussions then occurred with relevant parties before ABS finalised its response.

As a result of the review, the following changes will be made to the ABS's strategy:

POPULATION ESTIMATES

Release of State and SLA Data

The ABS will explore ways to improving the timeliness for the release of State estimates from six months after the reference date to five months and will also seek to improve the timeliness of SLA total population and secondly estimates by age and sex estimates currently published 9 and 15 months after the reference date.

Service Population Estimates

The ABS will prepare an issues paper covering the conceptual and methodological issues involved.

Indigenous Population Estimates

ABS plans to produce experimental estimates of the Indigenous population by ATSIC region for census years. Estimates will be annually updated at the State level for those States with satisfactory registered birth and death data for Indigenous people.

Population projections

ABS involvement in Sub-State population projections

ABS is seeking cooperative arrangements with State planning bodies in which duplication is avoided and in which users of population projections will benefit. It is open to arrangements which could include:

  • subject to ABS policy being adhered to, ABS publication of results, either in conjunction with State planning bodies or in its own right;

  • provision of projections for publication by State planning bodies; or

  • limited to receiving summary data associated with land use plans.

A preference exists for an arrangement in which ABS is involved in publishing the data.

ABS publication of capital city projections

ABS plans to publish capital city projections after appropriate consultation with State planning agencies on the assumptions to be used. Such projections are envisaged to be published every 2-3 years along with the State and Australian projections.

Ethnic population projections

ABS plans to publish population projections based on country of birth at the State level.

Household and Family Estimates

For States and capital cities except Darwin and Canberra, ABS plans to produce household estimates at least annually along with type of household (which will involve estimates of family type).

For Statistical Local Areas, subject to the availability of resources ABS will produce household estimates for census years and explore options for producing such data in non-census years.

Household and Family Projections

Subject to the availability of resources, ABS plans to produce projections of households at State and Sub-state levels following the 1996 Census. This in turn will be followed by projections of families (type of household) once developmental work on family estimates has been completed.

Fertility and Mortality

Movement of registered birth and death statistics to an occurrence basis

ABS plans to move to an occurrence based approached for birth, death and marriage registered data once it is satisfied that registries can provide the relevant data in a timely fashion.

Additional data

ABS will continue to support registrars of births, deaths and marriages in their moves to collect nationally comparable data.

Life tables for Indigenous people

ABS plans to produce State level life tables for Indigenous people following analysis of 1996 Population Census data.

Marriages and Divorces

ABS is concerned about the loss of registered divorce data. ABS considers that it would be best for a major user of the data to co-ordinate efforts to restore the data involved. ABS is prepared to support such a move.

International Migration

To enable the timeliness of population estimates to be improved as well as appropriate account being taken of 'category jumping', ABS will approach Department of Immigration and Multicultural Affairs to assess the feasibility of obtaining overseas arrivals and departure data consistently no later than 6 weeks after the reference month and preferably earlier.

Internal Migration

Alternatives to Medicare interstate migration data sources

A review of alternative to Medicare interstate migration data sources is planned to be discussed at the 1996 Commonwealth-State Population Workshop prior to decisions being taken on the updating source for the 1996 - 2001 intercensal population estimates. ABS will continue to explore other options as appropriate.


2 INTRODUCTION

2.1 Purpose of the Review

The Australian Bureau of Statistics (ABS) has as its mission to assist and encourage informed decision-making, research and discussion within governments and the community, by providing a high quality, objective and responsive national statistical service.

It is essential that ABS periodically undertakes a review of its statistical activities with the aim of defining its future directions and clarifying its own work in relation to the statistical work of other agencies. This review of demography statistics is planned to update the Review of Social and Labour Statistics: Population published in June 1990 (ABS Catalogue No. 4162.0). In this paper, conclusions from the 1990 review are listed, developments since 1990 noted, issues from identified and the ABS response given.

2.2 Conduct of the Review

In November 1995 a background paper was released for written comment by end January 1996. This paper was sent to major users of demography statistics and those responding to a general call for participation in the review made in the June Quarter 1995 issue of Australian Demographic Statistics. The 73 organisations which responded are listed in Attachment A. Following analysis of comments received, discussions were held with major users on the draft ABS response in May/June 1996.

2.3 Scope of demography statistics

Demography statistics can be defined as those concerned with the size, structure and geographic distribution of the population together with the changes in these parameters. Size relates to the number of people within a specified geographical area and geographic distribution refers to the spread of the population across geographic areas. The structure is the age and sex composition of the population and the components of change are births, deaths and migration (both internal and international).

A broader definition is one which includes additional characteristics of the population such as marital status, family type and ethnicity. Characteristics such as labour force participation and education could also be included. In the broadest sense, all demography, labour and social statistics might be considered population statistics. However, for the purposes of this review demography statistics will include the following areas:
      Population estimates
      Population projections
      Household and family estimates
      Household and family projections
      Fertility
      Mortality
      Marriages and Divorces
      International Migration, including short-term movements
      Internal Migration

2.4 ABS responsibility for demography statistics

In deciding which statistical series should be the responsibility of the ABS and which should be left to other agencies, three issues are relevant:

  • the intrinsic value of the data to the community and the extent of its use;

  • legal obligations of the ABS under the Census and Statistics Act 1905 and other legislation; and

  • comparative advantage.

The first is the intrinsic value of the data to the community and the extent to which they are important to a wide range of users. Demographic statistics are of great value to many users. Responsibility for these statistics then falls naturally to the ABS as the central statistical agency. ABS is in a unique position to provide on a continuing basis a high quality statistical service to all users, regardless of political or sectional interests.

The second issue relates to any legal obligation on the ABS to produce a particular statistical series. In the case of population estimates the ABS is obliged by its own legislation (Sub-section 9(2) of the Census and Statistics Act 1905) to compile and make available statistics of the number of the people of each State as on the last day of each quarter (in this paper reference to State includes Territory unless otherwise indicated). These population estimates are then used in deciding the number of seats for each State in the House of Representatives (the Commonwealth Electoral Act 1918) and the allocation of Commonwealth funds for State governments (the annual State Grants (General Revenue) Acts) and local governments (Local Government (Financial Assistance) Act 1995).

The third issue is concerned with the notion of comparative advantage. Does the ABS have a comparative advantage over other agencies which enables it to produce a statistical series more cost effectively? In the case of demographic statistics the ABS has several strengths, some of which are shared with other agencies, which together form a persuasive case for it to assume prime responsibility for demographic statistics.

ABS possesses:

  • stability and position in government which can guarantee continuity of those statistical series which are of fundamental long term importance;

  • a sufficient critical mass of work to enable continuity in skills of the people needed to collect, process and analyse the data;

  • integrity and objectivity arising from its policy neutral role in government and its reputation for maintaining the confidentiality of personal information which is supported by strong legislation; and

  • statistical knowledge and computer skills necessary to maintain a complex database.

In relation to this last point, demographic statistics has as one of its major requirements to bring together data from several sources (e.g. births, deaths, migration, population census, household surveys). These data can then serve as a source for a number of derived series such as population and household estimates and projections. ABS has a significant comparative advantage in this area of work.

For all the reasons outlined above, ABS will continue to accept the prime responsibility for the compilation and dissemination of demographic statistics in accordance with its corporate mission and assume a leadership role.

2.5 ABS strategy on population statistics

i ABS will continue to accept the role of the agency with the prime responsibility for the compilation and dissemination of demography statistics.

ii ABS will continue to allocate a high priority to demography statistics. Particular emphasis will be placed on the accuracy of State population estimates.

2.6 Role of data providers

Demography statistics are reliant on data that are a by-product of administrative processes and the ABS looks to the organisations involved to recognise the critical role they play in the national statistical system. ABS is seeking to have all administrative data it receives to be in electronic form (standard in the case of multiple sources providing equivalent data) without errors on the basis that data is most efficiently and effectively captured when this occurs close to its source.

One implication of the use of administrative by-product data is that it is important for the data providing organisations involved to consult with ABS when changes to the processing systems producing statistical data are being contemplated.

ABS is very willing to support organisations in improvements to their systems that produce statistical data required by ABS, improvements that can enhance the effectiveness and efficiency of the organisations concerned. Such support can include, for example, advice on form design, feedback on quality, advice on measures to enhance quality and the development and use of computer assisted coding.

In this regard, the ABS is appreciative of the recent actions of registrars of births, deaths and marriages in moves to integrate their activities, in part by standardising their form designs. In NSW the situation has been reinforced by the March 1993 Parliament of New South Wales Legislative Council Standing Committee on Social Issues Report No. 5 on Births, Deaths and Marriages: An Open Register? This report recommended that the functions of the NSW Registry should include the collection and provision of statistical data.

2.7 Data quality

Data quality, and in particular accuracy, is a high priority for ABS given the important uses made of demography statistics. These uses for population estimates include the allocation of the number of seats for each State in the House of Representatives and the allocation of Commonwealth funds for State and local governments.

The census of population and housing provides an opportunity to evaluate accuracy every 5 years, results for the 1986-91 period having been published in Population Estimates: Concepts, Sources and Methods (3228.0). The census post-enumeration survey, review of NT population statistics, assistance to registrars of births, deaths and marriages, and centralisation of vital statistics processing and development work on sub-State population estimates are examples of steps taken by ABS to enhance the accuracy of population statistics.

ABS will maintain contact with international developments in demography statistics to ensure that best practice is maintained.


3 REACTION TO ABS PROPOSALS

3.1 Population Estimates

1990 review conclusions

i Production of accurate population estimates is seen by ABS as a major priority and one for which it accepts prime responsibility.

ii The population estimates currently produced by ABS satisfy most requirements and will continue. The service has been further improved with the recent introduction of the following three annual series -

SLA x age x sex
Marital Status x age x sex (Australia level)
Birthplace x age x sex (Australia level)

iii ABS will maintain its current level of service on population estimates while at the same time pursuing a program of research into improving their quality and availability. Interstate migration and improvements to the census post enumeration survey will be given specific attention.

iv Intercensal estimates of the Aboriginal and Torres Strait Islanders are acknowledged as a major unmet demand. For reasons set out in the paper, ABS sees no prospect for adequately meeting this demand in the medium term. Nevertheless, ABS will continue research into Aboriginal demography, improving the quality of the source data needed for acceptable population estimates and estimation methodology.

Developments since 1990

Production of accurate Australian and State population estimates has continued to be a high priority. The ongoing program of research into estimation methods and the accuracy of the statistics has resulted in a revised model for estimation of interstate migration movements in the 1991 - 1996 period, the introduction of a new estimation method for category jumping and a special investigation in 1993-94 of estimates for the NT.

With an objective of improving the accuracy of sub-State population estimates, development resources on these estimates have been centralised rather than dispersed in each office to enable avoidance of a conflict of priorities between operational and developmental work, standardisation and the advantages of synergy.

In June 1994 population estimates back to June 1991 were changed to take account of final figures for births and deaths on a period of occurrence rather than registration basis. This practice was introduced to avoid anomalies arising from changes in the processing of these data by registration authorities. In particular, such anomalies can significantly distort the distribution of single year of age population estimates at ages zero to four. Preliminary population estimates are also proposed to use birth and death data on an occurrence basis - see discussion under Fertility below.

As suggested by the 1990 review, an examination has been made of the quality of data for Aboriginal and Torres Strait Islander people from the 1991 Census and birth and death registration statistics. Experimental estimates were subsequently published in November 1994 in Experimental Estimates of the Aboriginal and Torres Strait Islander Population: June 1986 to June 1991 (3230.0). In the NT for 1991-1994 it has also produced experimental estimates of Community Government Councils and Incorporated Associations which are predominantly populated by Indigenous people.

From July 1993 population estimates are being made for Cocos (Keeling) and Christmas Island.

A new publication entitled Population Estimates: Concepts, Sources and Methods (3228.0) was released in June 1995. The publication describes each of the regular published series of population estimates with details of how the estimates are compiled. The various data sources used to obtain the base population, births, deaths and migration are also described and an assessment is included of their accuracy.

There has been increasing use of ABS population data by market research organisations for benchmarking surveys they are responsible for. A prime role of population statistics within the ABS is the provision of benchmarks for population surveys and it is apparent that this role is extending outside the organisation which can only help in the provision of comparable statistics.

Plans are being put in place for the 31 December 1996 State population estimates due for release before 10 June 1997 to be based on the 1996 Census.

Issues

Concern has been expressed at the time involved to produce State level estimates. The timetable for population estimates currently calls for preliminary State level estimates to be produced by 6 months after the reference date and preliminary Statistical Local Area level estimates by 9 months after the reference date. Estimates for 31 December are produced before 10 June (5.3 months after the reference date) as a determination on the population of the States is required before that date by the annual State Grants (General Revenue) Acts.

Timing is affected by the need to take into account 'category jumping', an adjustment for the effect of persons whose travel intentions change from short-term to permanent/long-term or vice versa. While 12 months are required to assess the complete impact of category jumping, preliminary estimates are made on the basis of overseas arrival and departure data relating to 3 months after the reference date which covers about 90% of the events involved. Unfortunately, no standard pattern has emerged to facilitate estimation on a lower level of events. Data for the 3 months after the reference date are not available to ABS until 4.5 months after the reference date. Following receipt of these data provision needs to be made for compilation, analysis and consultation.

A second issue relates to the necessity to revise estimates to take account of later data. There are three stages in estimation:

  • preliminary - at the State level produced by 6 months after the reference date;

  • final to take account of revised data - 15 months after the end of a financial year; and

  • rebased - to take account of the following population census.

It is not apparent that reasonable options exist to avoid these revision steps given the high premium placed on accuracy by users of population estimates.

A third issue relates to the provision of service population estimates. Population estimates employ the concept of estimated resident population which links people to a place of usual residence within Australia. Estimates on this basis do not always meet the needs of service providers. The most obvious examples include tourist areas, mining towns and seasonal employment areas while a less obvious example is central business districts within major urban centres. The need for a variety of services including employment related services, health care services, child care, transportation, water and sewerage is not necessarily well reflected by measures of the resident population. State grants commissions are likely to be interested in taking such needs into account in recommendations associated with the distribution of Commonwealth funds for local governments.

The ABS is willing to produce service population estimates on a user pays basis. It would be necessary to identify and implement means by which other data could be obtained to enable adjustment to the usual residence population to accord with some other concept such as average population over a year.

Fourth, there is a degree of inconsistency between the definition of usual residence between the census of population and housing and international migration statistics. A 6 month time period is the basis for the determination of usual residence in the census while 12 months is used in international migration statistics. This issue is planned to be further explored in the development of the 2001 Census.

ABS sought advice on:

  • the extent to which users of population estimates are adversely affected by the current timetable for the release of population estimates and/or revisions; and

  • the need for service population estimates.

The extent to which users are adversely affected by the current timetable for the release of population estimates and/or revisions.

Comments received

The majority of respondents can manage with the current timetable for the release of estimates. While most users perceive an earlier release of estimates to be highly desirable, they recognize the constraints operating on the production of estimates.
There are several important exceptions. Key users such as the Commonwealth Grants Commission, the Department of Education, Employment and Training, and the Department of Immigration and Ethnic Affairs expressed concern over the timing of estimates. The Commonwealth Grants Commission would prefer June estimates to be available by the end of November, and DIEA by October. DEET requires SLA estimates by age and sex within 12 months of the reference point. The Commonwealth Treasury anticipates the possibility of returning to a May budget in 1997 and sought December 1996 State estimates by late April. Several other organizations also require SLA data generally earlier than is currently released. For instance, the Sutherland Shire Council in NSW requires SLA estimates by age and sex earlier than March for the purposes of the Council's Annual Reporting Process. It advised that other councils face similar difficulties with the existing timetable for release of estimates. The Western Australian Local Government Grants Commission stated that the release of estimates in February would be desirable.

ABS response

For State population estimates, ABS will explore releasing the relevant statistics by 5 months after the reference date. In particular it will approach Department of Immigration and Multicultural Affairs to assess the feasibility on obtaining overseas arrivals and departure data consistently no later than 6 weeks after the reference month and preferably earlier which would enable the timeliness of population estimates to be improved as well as appropriate account being taken of 'category jumping'.

In the event that ABS is not able to obtain overseas arrival and departure data relating to the full 3 months after the reference date, there will be a reduction in accuracy of category jumping estimates and hence State population estimates. Such a reduction is unlikely to have a significant affect on State relativities.

For SLA population estimates, the current timetable calls for intercensal estimates of the total population and secondly estimates by age and sex to be published by 9 and 15 months from the reference date respectively. ABS will explore the feasibility of bringing the timetable forward.

Service population estimates

Comments received

There is strong and widespread support for service population estimates. The agencies that support service population estimates include most State Treasuries and Health and Community Services authorities as well as organizations such as the NSW Department of Bush Fire Services, Tourism NSW, the NSW Environmental Protection Agency, Sydney Water, the Victorian Department of Sport and Recreation and South Australia Police. Service population estimates are perceived as invaluable for the purposes of strategic planning. Estimates based on resident population do not meet the needs of service providers in tourist areas, mining towns and seasonal employment.

Service population estimates are relevant to State and local government planning and resource allocation. They are recognized as important by States Grants Commissions in taking service needs into account for the distribution of Commonwealth funds for local governments. The WA Local Government Grants Commission noted that "... a significant number of councils have expressed dissatisfaction with the ERPs. In part, the development of service populations would address these concerns."

Users claimed that the ABS is in a position of comparative advantage to produce estimates and has ready access to data for employment and tourist accommodation. It is stressed that service population estimates are required for LGAs or even smaller scales. Many agencies argued that the user pays approach to service estimates is no longer appropriate. Rather, estimates should be produced as a matter of course and the cost be absorbed by the ABS.

ANU cautioned that the issue of service population estimates is a complex one. The definition of population will vary according to which service is being considered. For example, commuter populations being present at a particular locality for parts of the day and seasonal peaks of tourists are very different forms of service populations and reflect the issues that need to be addressed. ANU suggested that ABS produce an issues paper which explores the problems involved in service population estimates.

ABS response

The issues raised by the concept of service population are complex. By their very nature, service populations will vary according to the service identified and the time of day and year. As such a single estimate or even a consensus about the range of estimates of the population to be serviced is unlikely to be achieved. ABS does not see its role as putting out a whole variety of service estimates to meet these needs.

ABS will prepare an issues paper in 1997 which will examine user needs, the conceptual and methodological issues involved and the reporting load. For the 1996 Census of Population and Housing the ABS will also code SLA of workplace for persons employed in the Pilbara and possibly other areas to gain a further insight into the issue. In the meantime ABS is willing to assist with the production of service population estimates where the user is prepared to take responsibility for the wide range of explicit and implicit assumptions and cost involved.

Indigenous population estimates

Comments received

There is a demand for ongoing intercensal estimates of the Aboriginal and Torres Strait Islander population. Annual Indigenous population estimates are sought by the Commonwealth Grants Commission, DSS, DEET and a wide range of State departments. The Queensland Statistician's Office deemed the lack of ongoing intercensal estimates to be a significant shortcoming. The NSW Sutherland Shire Council and the WA Local Government Grants Commission sought estimates of the Indigenous population at the LGA level. Northern Territory Health Services considered that separate population estimates were needed to help monitor the "huge differential in the health status of Aboriginal people". The NT Office of Aboriginal Development regarded Indigenous population estimates as critical.

ABS response

ABS plans to produce experimental estimates of the Indigenous population by State and ATSIC region for census years and annually update these estimates at the State level for those States with satisfactory registered birth and death data for Indigenous people. ABS is willing undertake the production of Indigenous population estimates for other areas where the user is prepared to take responsibility for the assumptions and cost involved.

ABS will continue to support efforts by State registrars of births, deaths and marriages to improve the coverage and comparability of data they collect on registered births and deaths and monitor the resulting quality.

Birthplace population estimates

Comments received

DSS requires ABS figures for the region "United Kingdom and Ireland" separately identified for the United Kingdom and the Republic of Ireland as it has separate international social security agreements with these two countries.

Several Victorian agencies such as the Ethnic Affairs Unit of the Department of the Premier and Cabinet and the Department of Health and Community Services seek the compilation of State population estimates by age, sex and country of birth.

ABS response

ABS will provide separate estimates of the Australian resident population born in UK and Ireland relating to June 1996, based on the 1996 Census, and subsequent reference points.

ABS does not consider that sufficient justification has been advanced for it to compile intercensal population by age, sex and country of birth at the State level. National level estimates are compiled. Any intercensal State level estimates would be affected by the lack of country of birth data for internal migrants.

3.2 Population Projections

1990 review conclusions

i ABS will continue to provide a range of Australian and State and Territory population projections for general government and community use, based on a range of realistic demographic assumptions.

ii ABS projections are a consistent set for the States and Territories and as such serve both as a framework and a point of departure for other organisations working in this field.

iii ABS will continue to undertake special purpose projections for other government agencies and will provide advice and data to agencies who are doing their own projections.

iv Given population growth in key geographical areas is often related to State and local government policy and planning decisions to which ABS does not have access, others are better placed to produce sub-State population projections. ABS therefore will not accept direct responsibility for compiling or publishing such projections. However, it will give assistance to other agencies undertaking such projection work on technical matters and by the provision of data. ABS is also prepared to undertake projections on a user pays basis, provided the projection assumptions are specified by the user.

v For projections of the ethnic population, ABS considers the relative priority of this work is not high. It will therefore undertake work in this area in consultation with the Department of Immigration, Local Government and Ethnic Affairs (DILGEA as it then was) only if user funding is available.

vi With respect to population projections work ABS will continue its current arrangements with DILGEA whereby that agency only produces forecasts at the national level for up to five years ahead while ABS produces longer term (20-30 years) national and State/Territory projections.

Developments since 1990

ABS has continued to provide a range of Australian and State projections for general government, business and community use. It has undertaken special purpose projections for other agencies and provided advice and data to agencies doing their own projections.

In accordance with the 1990 review recommendations, ABS continues to provide statistics and advice on assumptions for sub-State projections. ABS also undertakes sub-State population projections on a user pays basis, provided the projection assumptions are agreed with by users. To this end ABS has developed computer software to undertake population projections for small geographical areas.

The policy that ABS should not compile and publish sub-State projections in 1994 was re-examined and amended to allow publication in certain prescribed circumstances.

The total policy statement now reads as follows:

  • It is ABS policy to accept a responsibility for the compilation and publication, on a regular basis, of a consistent set of Australian and State/Territory population projections for general government and community use.

  • It is ABS policy that compilations of population projections may be provided to clients based on a range of assumptions offered to, or by, clients, with or without ABS advice on those assumptions, and with the cost of the ABS service being met by the client.

  • It is ABS policy that the provision of compilations of client-specific projections, for all geographic levels, will be on the condition that client responsibility for the assumptions is to be acknowledged whenever they are used.

  • Where State/Territory governments continue to take responsibility for sub-State population projections, it is ABS's policy to support this work by providing statistical advice in the development of projection assumptions and being available to undertake the projections according to assumptions agreed to by State/Territory Governments.
    In a State/Territory where a demand for ABS to publish sub-State population projections arises either because the State/Territory Government has invited such involvement or because other demand is unsatisfied, it is ABS policy to consider doing so where the relevant Deputy Commonwealth Statistician/ Territory Statistician has arranged guaranteed access to State/Territory Government planning parameters, where the State/Territory Government identifies with the assumptions made and such identification is made explicit in the ABS publication.

        It is ABS policy that any sub-State projections it publishes must

        • when aggregated to the State level be within the boundaries of ABS projections last published for the State/Territory; and

        • involve at least two projections being published at the same time for the areas involved to ensure that ABS projections do not become ABS endorsed forecasts as well as to encourage an understanding of the assumptions behind the projections.
    No work has been undertaken on projections of ethnic populations. This was to have been done only with user funding and this has not been forthcoming even though it is apparent that there is some need for these data.

    Indigenous population projections were produced in May 1996 and published in Experimental Projections of the Aboriginal and Torres Strait Islander Population, June 1991 to June 2001 (Catalogue No. 3231.0).

    Issues

    Issues have arisen in relation to population projections relate to:

    • dissemination;

    • frequency;

    • content in terms of capital city projections;

    • accuracy; and

    • roles of ABS and State governments.

    The issue of dissemination arose from the December 1994 House of Representatives Standing Committee on Long Term Strategies report on Australia's Population 'Carrying Capacity' : One Nation - Two Ecologies.
      The Committee suggested that the ABS consider "... developing packages which make such (projection) tasks straightforward and inexpensive." The ABS has developed such a package but, while not excluding the possibility of making the process more efficient, the ABS believes that this suggestion put insufficient weight on the need for skilled input to the generation of projections to ensure realistic outcomes. Although commercial and public domain projection software currently exists, the ABS considers that because of the complexities involved, the best option to assist informed debate is the continued provision of a population projection service and the wide dissemination of ABS standard and client sponsored projections.

      Every 2-3 years the ABS releases projections of national and State populations. The ABS will consider options to achieve wider dissemination of these projections, not only by simplifying the presentation but also by using such means as the Environmental Resource Information Network. Second, there may be justification for increasing the frequency of population projections from every 2-3 years to, say, annually to draw to the public attention the impact of changing international and internal migration on Australian and State populations. This could take the form of a simplified analysis compared with that included in the major reviews conducted every 2-3 years, most recently published in Projections of the Populations of Australia, States and Territories 1993-2041 (3222.0). Effectively, annual projections would involve the transfer of responsibility for short-term projections from the Department of Immigration and Ethnic Affairs to ABS.

      Third, it was apparent in the House of Representatives Standing Committee on Long Term Strategies' contact with ABS that a major focus was in relation to the population of the capital cities, which comprises 63% of Australia's population. Projections for capital cities made by State Governments are inconsistent in terms of their timing and their assumptions. Because land use constraints at the capital city level of aggregation are a much less significant factor than at the Statistical Local Area level of aggregation in affecting population growth, demographic factors are a primary consideration. ABS, after appropriate consultation with State governments, would be in a good position to include capital cities in its normal round of State projections thereby ensuring provision of an Australia wide consistent set of projections for capital cities in a form readily available to the public.

      Fourth, concern has been expressed that ABS small area projections take insufficient account of land use considerations, even though it is not uncommon for these considerations to change associated with changes in factors such as the economy and the level of net overseas migration. Other changes in intentions can even result from projections with better information leading developers to, for instance, not proceed with particular developments.

      Currently land use considerations are indirectly taken into account by reviewing existing State government projections where they are available when setting migration assumptions. ABS proposes to take account of land use considerations in future by seeking advice of relevant data directly from planning authorities.

      Fifth, the issue of land use considerations raises the question of ABS and State government responsibilities for small area population projections. ABS has prepared the most recent Tasmanian and ACT government small area projections using assumptions agreed with by the relevant bodies. A similar arrangement is planned for the NT.

      Below the State level ABS produces small area projections where the client takes responsibility for the assumptions. These projections draw upon ABS's strengths in demography in that:

      i they are generally up to date being based on the latest estimated resident population for current SLA boundaries. ABS is planning to produce standard annual updates of SLA population projections for those clients prepared to agree with the assumptions involved. State government projections often use the estimated resident population in the last census year for SLAs that then existed.

      ii ABS projections are able to take a more objective and global view than is the case with State government planning authorities. This is particularly the case with interstate migration where the assumptions of different State governments cannot always be reconciled with one another.

      iii not all State government projections are available by age and sex.

      iv unlike ABS, not all State government projections use detailed SLA-specific fertility, mortality and migration assumptions.

      State governments other than Tasmania, NT and ACT should consider whether they wish to continue preparing small area population projections given ABS plans to further increase the accuracy of its projections, its comparative advantage in this area and the potential for savings involved. State governments could continue to be involved in the dissemination of the projections

      ABS sought advice on

      • whether Australian and State projections should be produced more frequently than every 2-3 years;

      • whether State planning authorities would have any significant difficulty in supplying to ABS summary data associated with land use plans; and

      • the desirability of ABS producing capital city population projections;

      • whether State governments other than Tasmania, NT and ACT wish to continue preparing small area population projections.

      Whether Australian and State projections should be produced more frequently than every 2-3 years.

      Comments received

      While a number of agencies support annual projections, many do not see the value in more frequent projections than the 2-3 year interval.

      Some interest is expressed in annual projections at a Commonwealth level. BIMPR indicated that it would "welcome an ABS decision to produce Australian and State population projections on an annual basis." ANU also suggested that official ABS projections be produced on an annual basis. The Department of Social Security recommended that annual population estimates include a set of national level population forecasts to cover the year following the one for which the last set of preliminary estimates were available.

      Lack of agreement is prevalent between departments in each State. Submissions from agencies in Western Australia reflect the range of views. For example, the WA Ministry for Planning considered more frequent projections would be a disruption to planning and noted that users want a set of projections that remain common currency for five years. The WA Treasury considered that it might be useful to publish projections more frequently, so long as it can be achieved with minimal effort, by retaining underlying assumptions and methodology, but updating calculations to reflect current data. The WA Electoral Commission sought a frequency greater than every 2-3 years.

      ABS response

      ABS does not consider that sufficient justification has been advanced for it to produce Australian and State projections more frequently than every 2-3 years. ABS will continue to produce short term projections for those users who accept responsibility for the assumptions and costs involved.

      Whether ABS should continue to be involved in Sub-State population projections given current State planning agency involvement

      Comments received

      Concern was expressed at the 1995 Commonwealth-State Population Workshop about ABS's involvement. The State government planning bodies involved defended their role in the production of small area projections. They questioned the "comparative advantage" of the ABS in this respect. The NSW and Queensland planning authorities have reservations about the cohort-component method as the most appropriate method for projections at the SLA level. Land supply and local government planning policy were seen as the main drivers behind small area growth. State officers have access to local knowledge which cannot be easily quantified nor written down in simple formulae. The QLD Department of Local Government and Planning argued that for the ABS to acquire local intelligence, greater resources would need to be allocated to State offices. The South Australian planning authority view was that the projections for SA should be coordinated down to the Statistical Division level with the planning body responsible for SLA level projections.

      Some smaller state agencies consider the ABS to be the most appropriate agency to undertake small area projections. ABS production of sub-State projections is seen to offer greater consistency of data and assumptions.

      ABS response

      The ABS considers it strategically important for it to be involved in the production of sub-State projections to enable it to meet the needs of users for Australia-wide consistent sub-State projections which reflect prevailing trends and which both are and are seen to be independent. This is particularly the case for short term projections undertaken for electoral purposes. Such users include the House of Representatives Standing Committee on Long Term Strategies inquiry into Australia's Population 'Carrying Capacity' : One Nation - Two Ecologies, Australian and State electoral commissions and the Indicative Planning Council for the Housing Industry.

      While land supply and local planning policy is a critical component in determining future population (and associated with it, migration) levels, equally important is the demand for land. The present dwelling oversupply in many areas of Australia is an illustration of how both demand and supply need to be taken into account in determining population of and net migration to an SLA.

      In relation to reservations expressed over the cohort-component method, ABS notes that the Indicative Planning Council for the Housing Industry has commissioned a project addressing small area population projections and awaits with interest its outcome. In the meantime ABS plans to continue to use this method directly taking into account land planning data rather than indirectly as currently occurs (see next issue). ABS is open to further discussion on the issue in professional forums.

      ABS is seeking cooperative arrangements with State planning bodies in which duplication is avoided and in which users of population projections will benefit. It is open to arrangements which could include:

      • subject to ABS policy being adhered to, ABS publication of results, either in conjunction with State planning bodies or in its own right;

      • provision of projections for publication by State planning bodies; or

      • limited to receiving summary data associated with land use plans.

      A preference exists for an arrangement in which ABS is involved in publishing the data.

      For Tasmania, NT and ACT there exists currently or planned close relationships between ABS and State planning bodies. In discussions associated with this review, the NSW, QLD and SA planning bodies have indicated a wish to explore cooperative arrangements with ABS.

      ABS will explore means of producing population projections for varying geographical areas to facilitate it meeting the needs of users.

      Whether State planning authorities would have any significant difficulty in supplying to the ABS summary data associated with land use plans.

      Comments received

      Willingness to supply land and planning information is again related to broader views on ABS production of small area projections. On the one hand, reservations were expressed about supplying to the ABS summary land use data. The NSW Department of Urban Affairs and Planning explained that essential local knowledge comes mainly from the experience of regional managers and senior executives through continuous liaison with councils over a long period of time. "This information is not in publication format and cannot be easily quantified nor written down in simple formulae." The WA Ministry for Planning stressed the confidentiality of land supply data, equivalent to the concept of the ABS unit record file. On the other hand, the Tasmanian Department of Environment and Land Management noted that it would be happy to provide information of land use to assist with capital city and small area projections, given limited resources in that State to generate its own projections.

      ABS response

      The type of land planning information being sought by ABS, the expected number of dwelling units in local areas over time, should already be in the public domain, albeit on some occasions in a disaggregated form. The data represents the outcome of land planning and not land planning data in its own right. ABS accepts that it would be generally inappropriate to use non-public domain data where this may lead to an inadvertent disclosure of land development plans.

      ABS will seek the cooperation of State planning agencies in the provision of the relevant data in a summary form noting that part of such data may originate from local government sources.

      ABS publication of capital city projections

      Comments received

      Many Commonwealth departments such as Health and Human Services, Social Security, the Attorney-General's Department, AIFS, BIMPR and several State agencies identified a deficiency in the lack of a comparable set of capital city projections and looked to the ABS to provide these. Several Tasmanian departments argued that capital city projections would be less relevant in that State due to its decentralized population. The Tasmanian Department of Treasury and Finance would prefer projections for all major cities including Launceston, Devonport and Burnie.

      Larger State government planning agencies expect little benefit from ABS production of capital city projections. Capital city growth is seen as inextricably linked to wider strategic planning programs and State information systems. The WA Ministry for Planning commented that, "Due to the scale and complexity of the task, ABS is not in a good position to include capital cities in its normal round of projections."

      ABS response

      In discussions associated with this review there was little concern expressed over ABS publishing capital city projections.

      ABS notes that:

      • a need exists for comparable and current population projections for capital cities as reflected in the lead up work to the House of Representatives Standing Committee on Long Term Strategies report on Australia's Population 'Carrying Capacity' : One Nation - Two Ecologies and in the needs of Commonwealth agencies; and

      • population growth of capital cities, which comprise 63% of Australia's population, is closely linked to trends in State population growth.

      Given these considerations, ABS plans to publish capital city projections after appropriate consultation with State planning agencies on the assumptions to be used. Such projections are envisaged to be published every 2-3 years along with the State and Australian projections.

      ABS is also generally open to publishing projections at a sub-State level as reflected in the following policy:
          In a State/Territory where a demand for ABS to publish sub-State population projections arises either because the State/Territory Government has invited such involvement or because other demand is unsatisfied, it is ABS policy to consider doing so where the relevant Deputy Commonwealth Statistician/ Territory Statistician has arranged guaranteed access to State/Territory Government planning parameters, where the State/Territory Government identifies with the assumptions made and such identification is made explicit in the ABS publication.
      Whether ABS should undertake small area projections according to user specified assumptions

      Comments received

      There is an adverse reaction to the concept of the ABS undertaking sub-State population projections based on users' assumptions. The QLD Department of Local Government and Planning argued that ABS policy to compile projections based on assumptions specified by clients is "disturbing and puzzling....very few users would have sufficient understanding of population dynamics to be in a position to formulate relatively complex mortality, fertility and migration assumptions. Indeed, if they were in possession of such knowledge, it is quite likely that they would produce the projections themselves!" The WA Ministry for Planning considered ABS production of small area projections to be a "policy for intra-governmental confusion", creating a situation where departments in the same State have generated different projections. The Ministry stated, "I would recommend that you do not carry out any sub-State projections and that you focus your resources in improving timely census output data and expand demographic research. This would be a real service to users and the real practitioners of sub-State projections".

      ABS response

      Implementation of cooperative arrangements with State planning bodies previously referred to is likely to reduce demand for special projections. Nevertheless, given its mission to 'assist and encourage informed decision making, research and discussion ....', ABS considers that it should make its methodological expertise available to users when requested but that the resultant projections need to be under the auspices of the user.

      While the vast majority of users of ABS's small area population projection service adopt ABS advised assumptions that reflect prevailing trends and known constraints, there are cases where the user wishes to assess the impact of alternative scenarios. ABS considers that it would be inappropriate for it to refuse to support such analysis.

      Ethnic population projections

      Comments received

      There is a demand for ethnic population projections from a range of State and Commonwealth agencies which need these statistics for the planning and delivery of community services for ethnic minority groups. Expressions of interest have been raised by Social Security, BIMPR, the Attorney-General's Department and many welfare oriented State government departments. The SA Health Commission advised that projections were needed 'because of the differential ageing patterns evident among the population born in predominantly non-English speaking, notably European, countries. These population groups have specific requirements which are largely different from those of other Australians. Another group of importance are the (mainly) younger group with birthplaces in Asia'. It noted that responsibility for aged care programs will be moved to State Governments from the Commonwealth Government in January 1998. BIMPR noted that it receives many enquiries from government agencies, the media and public about the future ethnic composition of Australia. Users expressed concern that ethnic projections are deemed to be a low priority by the ABS. Australian governments at all levels are committed to the provision of culturally appropriate services and making those services more appropriate.

      ABS response

      A standard set of projections based country of birth would be helpful for health authorities in planning the provision of ethnic specific services.

      ABS plans to publish population projections based on country of birth at the State level. It is prepared to undertake other projections where the users of the data are prepared to define the concept of ethnicity required, agree to the assumptions involved and pay for the associated costs. 'Ethnicity' is a concept that can be defined in a variety of ways including by country of birth, birthplace of parents, language spoken at home and religion.

      3.3 Household and Family Estimates

      1990 review conclusions

      i ABS attempts to meet the current demand for estimates of the numbers and types of households and families through the population census and household based social surveys. However, there are differences in concepts, definitions and classifications from survey to survey and from the population census.

      ii ABS will move to standardise household and family concepts across surveys and between surveys and the census. This will enable the production of a range of statistics on a more consistent basis.

      iii Once progress has been made on establishing a more consistent body of survey and census data, ABS will review the quality of the household and family estimates and the degree of unmet demand. It will then consider whether new methodology needs to be developed for meeting this demand. It will also investigate the possibility of publishing regular household and family estimates.

      Developments since 1990

      The 1990 review recognised that before the demand for household and family estimates could be assessed, the statistical concepts used to define households and families needed to be reviewed and standardised.

      This work has now been completed and published in Standards for Statistics on the Family (1286.0) released in April 1995.

      ABS has produced household estimates for use as benchmarks for the 1993-94 Household Expenditure Survey. It is planned to publish estimates in Household Estimates (3229.0) scheduled for release shortly. This is planned to be followed by family estimates.

      Issues

      While an end is near of development work associated with household and family estimates, ABS has yet to commit itself to the continuing production of household and family estimates. Advice sought advice on the frequency to which household and family estimates are needed.

      The frequency to which household and family estimates are needed.

      Comments received

      For some agencies, such as the Australian Housing and Urban Research Institute, the number of families/households is more important than estimated resident population. A wide range of agencies involved in urban and welfare planning and the provision of services linked to the family unit support the production of household and family estimates. State planning agencies expressed a significant interest in household estimates at the SLA level which take account of the changing structure of households. Most users interested in these estimates expressed a need for annual estimates.

      ABS response

      For States and capital cities except Darwin and Canberra, ABS plans to produce household estimates at least annually along with type of household (which will involve estimates of family type). For Statistical Local Areas, subject to the availability of resources ABS will produce household estimates for census years and explore options for producing such data in non-census years.

      3.4 Household and Family Projections

      1990 review conclusions

      i While head of household has ceased to be a meaningful statistical concept and ABS no longer collects this data in its population census, ABS recognises the importance of household projections in urban planning. Accordingly, ABS is prepared to make available to specialist users Census output for household reference persons on a user pays basis.

      ii ABS will re-examine the demand for household and family projections to assess whether there is a need for it to go beyond its current level of involvement (ie to not only provide supporting data to other users but also to develop its own series).

      Developments since 1990

      Statistics relating to "reference person" have continued to be made available from the Census to support the compilation of household projections for use in urban planning. While household and family projections are a natural corollary of population projections, work on these aspects has been put on hold until estimates have been established. Where these projections have been needed by specific clients, they have been provided on a user pays basis.

      Comments received

      The Department of Social Security is interested in family level estimates and projections rather than household level estimates as families are the units towards which payments are generally made. DSS would welcome ABS giving close consideration to the possibility of producing its own series of family projections.

      ABS response

      ABS notes the importance identified by DSS of these data. Subject to the availability of resources, ABS plans to produce projections of households at State and Sub-state levels following the 1996 Census. This in turn will be followed by projections of families (type of household) once developmental work on family estimates has been completed.

      3.5 Fertility

      1990 review conclusions

      i ABS recognises the importance of fertility data in the compilation of population estimates and projections. It will maintain its current tabulation program and dissemination service.

      ii ABS will continue work on analysing fertility trends and will publish the results in social reports.

      iii ABS will continue its central role in the production tab of accurate aggregate statistics on fertility (for total Federal, State and SLA populations and particular population groups) for use in population estimates and projections, government health and education programs and in epidemiological research.

      iv ABS will continue negotiations with the Registrars on the inclusion of some extra data items on the birth registration form in order to obtain national statistics ie Aboriginal identification, number of previous births (present marriage and other), date of birth for "other" births, birthweight, standard occupation descriptions and occupation of mother. ABS will cooperate in studies of the quality of such data and will process and publish statistics for these extra data items.

      Developments since 1990

      ABS has continued its central role in the production of accurate aggregate statistics on fertility and in analysing fertility trends. Fertility trends are now published in various social reports. As the topic "children ever born" was not included in the 1991 Census, data on differential fertility (by variables such as income, education level, indigenous origin) were not directly available. To meet the need for these data synthetic estimates of "issue" were generated from census data using the "own child" method. Estimates for the Aboriginal and Torres Strait Islander population were published in Births, Australia, 1993 (3301.0). Estimates for other populations are available on a user pays basis. A question on issue will be included in the 1996 Census.

      ABS is attempting to advance the provision of additional registered statistical data through its support of a project being undertaken by the registrars which aims to integrate their systems of birth, death and marriage registrations, most immediately by the development of a standard registration form for use in each State and Territory. The inclusion of extra items on registration forms (eg Indigenous origin, number of previous issue, birthweight, occupation of mother) is subject to this development although an Indigenous origin identifier will be included on birth and death registrations forms in all States/Territories from January 1996 when the relevant Queensland forms were amended.

      Issues

      First, ABS is seeking to have all registered birth, death, marriage and divorce data it receives to be in a standard electronic form without errors on the basis that data is most efficiently and effectively captured when this occurs close to its source. Ideally the same registration computer system should be used throughout Australia to facilitate the production of standardised data. Support will continue to be given to the registrars of births, deaths and marriages to achieve this objective. Of most concern is that registered Aboriginal and Torres Strait Islander birth and death data for only South Australia, Western Australia, Northern Territory and the Australian Capital Territory are currently of publishable quality. These data are considered to be at least 90 per cent complete.

      Second, ABS currently produces as standard output birth and death registration data on a period of registration basis rather than what is really needed, a period of occurrence basis.

      In quarterly estimates of Australian and State populations, birth and death registration data are used in preliminary estimates and occurrence data in final estimates. An increase in the average time taken between births occurring in a State and births being registered would currently result in the population of a State being initially underestimated.

      In most States over 98% of births and 99% of deaths occurring in a quarter are registered by 4 months after the end of the quarter. Allowing 0.5 months for the data to reach ABS and for initial ABS processing, this would mean that estimates on an occurrence basis could be readily included in quarterly Australian and State population estimates. For accuracy reasons, registries for those States where registration of events does not achieve the registration rate indicated above would need to adjust their processing arrangements to achieve more timely processing.

      Third, a need has been identified to link birth and death data, particularly for perinatal and infant deaths so that statistics associating causes of death with risk factors can be produced. This linkage already occurs in Western Australia and while it is now not proposed to pursue an Australia-wide linkage, the option exists for other States to adopt this approach. Such linkage would not only involve registered data but also hospital based data.

      ABS sought advice on any difficulties a move to an occurrence basis would cause users of registered birth and death statistics.

      Movement of registered birth and death statistics to an occurrence basis

      Comments Received

      No organization reported difficulty with a transition to compiling births on an occurrence basis. The move is generally welcomed the proposal as it would improve population estimates and more closely reflect reality. Users recognized that occurrence based reporting would remove the inconsistency involved in current methods of compilation which involve preliminary and final estimates. State government departments anticipated the opportunity to compare births data with their own collections. The WA Department of Family and Children's Services suggested that access to both registration data and occurrence data for a period might be useful. Several agencies noted that support for the change would be withdrawn should the change affect timeliness and result in further delays.

      ABS response

      ABS plans to move to an occurrence based approached for birth, death and marriage registered data once it is satisfied that registries can provide the relevant data in a timely fashion. Registered divorce data is already on an occurrence basis.

      This move has implications for the timely processing of registrations by the registrars of births, deaths and marriages. In most States over 98% of births and 99% of deaths occurring in a quarter are registered in the subsequent 4 months. Allowing 0.5 months for the data to reach ABS and for initial ABS processing, this would mean that estimates on an occurrence basis could be included in quarterly Australian and State population estimates. For accuracy reasons, registries for those States where registration of events does not achieve the registration rate indicated above would need to adjust their processing arrangements to achieve more timely processing. Based on the 1994 situation this will affect the NSW, QLD and NT registries for birth data and the NT registry for death data. It may also mean that ABS will need to receive data from smaller registries more frequently than quarterly.

      Additional data

      Comments received

      There is general support for registrars of births, deaths and marriages to include extra data items onto birth registration forms, such as the occupation of mother and all previous births.

      Many agencies valued the inclusion of an Indigenous origin identifier on the death registration form. There is general demand for the proposed coding system that provides information on multiple causes of death, especially from Commonwealth and State health departments.

      The AIFS is interested in trends in mortality according the family status of the deceased.

      ABS response

      ABS will continue to support registrars of births, deaths and marriages in their moves to collect nationally comparable data.

      3.6 Mortality

      1990 review conclusions

      i ABS recognises the importance of mortality data in the compilation of population estimates and projections. It will maintain its current tabulation program and dissemination service.

      ii ABS will continue work on analysing trends in mortality and will publish the results in social reports.

      iii ABS will continue its central role in the production of accurate aggregate statistics on mortality (for total Federal, State and SLA populations and particular populations groups) for use in population estimates and projections, government health programs and in epidemiological research.

      iv The ABS will aim to meet the demand for information on multiple causes of death as part of its development of an automatic cause of death coding system.

      v In co-operation with the Australian Institute of Health (now the Australian Institute of Health and Welfare), ABS will continue to press the Registrars for the inclusion of an Aboriginal identifier on death registration forms. It will also pursue the inclusion of a standard occupation question. ABS will also co-operate in studies of the quality of such data.

      vi ABS will support the establishment of a National Death Index by the AIH.

      vii The understanding between ABS and the Government Actuary on the publication of life tables will be re-examined to ensure there is no duplication of effort. ABS is prepared to publish life tables that can be used for actuarial purposes.

      Developments since 1990

      ABS has continued to analyse trends in mortality and now publishes them in various social reports.

      Subject to a current trial, multiple causes of death will be available for 1996 death registration data.

      ABS plans to publish 1993-95 tables in the 1995 edition of Deaths, Australia (3302.0) on a consistent conceptual basis with those produced by the Australian Actuary.

      Issues

      See Fertility above.

      ABS also plans to produce State level life tables for Indigenous people following analysis of 1996 Population Census data. ABS will be willing to undertake the production of life tables for smaller areas where the user is prepared to take responsibility for the assumptions and cost involved.

      3.7 Marriages and Divorces

      1990 review conclusions

      i ABS will continue its central role in the production of aggregate statistics on marriages and divorces for use in the formulation of family welfare policies, the analysis of fertility, the calculation of population estimates by marital status and for general community information.

      ii Standard classifications of families and living arrangements will be developed for use in ABS collections, including household surveys.

      iii Opportunities for including questions on current and retrospective living arrangements in ABS household surveys (either in a special supplementary survey on families or in other surveys) will be examined.

      Developments since 1990

      The 1990 review stressed the importance of developing standard classifications of family type and living arrangements. Both standards have now been developed and promulgated.

      Data on cohabitation outside marriage was seen by the 1990 review as a major unmet demand. It was recommended that questions on current and retrospective living arrangements be included in the next Family Survey. Such questions were subsequently included in the 1992 Family Survey. For people currently married, data are available on whether married more than once, date of first marriage, and whether cohabitated before marriage. Current defacto relationships are recognised and data on whether or not such couples were previously married are available. For people previously married but not currently in a relationship, data on date of previous marriage and whether cohabitated before marriage are available.

      A significant analysis of divorce trends was included in Marriages and Divorces, 1994 (3310.0) which involved a method for deriving divorce rates by length of marriage. Unfortunately, action by the Family Court of Australia in ceasing to collect data on marital status at time of marriage will mean that it will no longer be possible to measure how many first marriages end in divorce, the standard international measure of divorce.

      Issues

      ABS is concerned about the state of registered marriage and divorce data.

      The costs involved in ABS processing of marriages are relatively large which may result in a long term reduction in data availability. With the exception of NSW and WA, all other States and Territories provide data completely in paper format necessitating lengthy and costly loading and editing procedures. ABS is seeking to have all registered marriage data it receives to be in a standard electronic form without errors on the basis that data is most efficiently and effectively captured when this occurs close to its source. Ideally the same registration computer system should be used throughout Australia to facilitate the production of standardised data. This objective will be pursued with the Attorney-General's Department, which is responsible for the administration of the Marriages Act 1961, and the registrars of marriages. A cost-benefit study is planned to take place in 1996-97 on this issue.

      Registered divorce data have been affected by the poor quality of data received from the Family Court of Australia and reductions in data availability. While it was expected that data quality should improve as a result of the Family Court introducing a new registration form from 1 February 1995, such an improvement has not yet been in evidence. Unfortunately, associated with this new form has been the loss of statistics on:

      • marital status at time of marriage;
      • State and Territory of usual residence of people separating;
      • occupation at date of separation; and
      • number of children affected by divorce over age 18.

      ABS is concerned about the loss of these data, a situation which cannot be remedied unless the Chief Justice of the Family Court changes his decision on this issue.

      Loss of data

      Comments received

      The ABS is strongly supported in its efforts to have the Family Court restore data item. Organizations such as the Attorney-General's Department, the WA Department of Family and Children's Services, Relationships Australia and ANU endorsed ABS moves to have the collection and publication of this data reinstated. The data items are considered to be important social indicators, essential to family welfare planning and service provision. As Relationships Australia (SA) noted, the absence of data on marital status at the time of marriage means that comparison cannot be made between first and subsequent marriages. The usual residence of people separating is more important than the location of application. Occupation is a significant feature of divorce and the number of affected children over the age of 18 is a valuable perspective. ANU comments that, "A continuing effort should be made to have the Family Court restore the question on marital status at the time of marriage. Its removal is an international embarrassment".

      ABS response

      Given the circumstances surrounding the loss of the data, ABS considers that it would be best for a major user of the data to co-ordinate efforts to restore the data involved. ABS is prepared to support such a move.

      3.8 International Migration, including Short-term Movements

      1990 review conclusions

      i ABS recognises the importance of data on overseas migration in ABS population estimates, in the calculation of the balance of payments and in the services component of the national accounts. Data published by ABS is used extensively by other government departments and private agencies in migration and ethnic affairs policy and tourism.

      ii Timely and comprehensive data on overseas movements are therefore seen as of high priority. ABS will work with DILGEA to ensure the present statistical service is maintained.

      Developments since 1990

      Recognising the importance of data on overseas migration ABS has maintained close contact with DIMA to ensure the continued availability of timely and high quality data. Despite new DIMA computing systems, reviews of administrative arrangements and reviews of trans-Tasman travel arrangements this goal has been achieved.

      Issues and ABS response

      As outlined in section 3.1 above, to enable the timeliness of population estimates to be improved as well as appropriate account being taken of 'category jumping', ABS will approach Department of Immigration and Multicultural Affairs to assess the feasibility on obtaining overseas arrivals and departure data consistently no later than 6 weeks after the reference month and preferably earlier.

      3.9 Internal Migration

      1990 review conclusions

      i Internal migration statistics are an essential component of ABS State and sub-State population estimates. They are also needed by a range of users in urban and regional planning, housing policy, etc.

      ii ABS accords a high priority to internal migration questions being included in future population censuses, to support the production of accurate State population estimates.

      iii ABS will continue to actively pursue administrative by-product collections as a source of interstate migration data and will monitor their accuracy.

      iv An investigation will be undertaken to find ways to improve the accuracy of data from surveys. A recommendation to repeat the ABS Internal Migration Survey will only be made if there are prospects of data of sufficient accuracy to aid the estimation of State populations.

      Developments since 1990

      Despite actively pursuing other administrative data as a source of estimates of internal migration no alternate to Medicare has been found. Relevant questions to assess internal migration were included in the 1991 Census and, in response to significant user demand, address for both 1 year and 5 years ago will be coded to the SLA level in 1996.

      Issues

      A need has been registered to collect data on reasons why people move, ideally at a small area level. The basis for this need and options to meet it are due to be considered at the 1995 Commonwealth-State Population Workshop. Such data would complement Statistical Local Area of usual residence statistics five years and one year ago data to be available from the 1996 Census (at a cost of about $3m to $5m for SLA of usual residence five years ago and $0.5m to $1m for one year ago).

      A review of alternative to Medicare interstate migration data sources is planned to be discussed at the 1996 Commonwealth-State Population Workshop prior to decisions being taken on the updating source for the 1996 - 2001 intercensal population estimates.

      Reasons why people move

      Comments received

      Agencies generally seek this information which is seen as vital for policy development by those responsible for the management of urban areas and is useful for needs based planning. The Department of Economics at The University of Melbourne claimed however that it would be an exercise of doubtful validity and usefulness to attempt to collect reasons on why people move their household location. ANU questioned the methodology used to obtain this information and claims that the, "...investigation of this issue is not straight forward. A simplistic approach (Why did you move? with closed response categories) may be more misleading than useful". ANU argued that the task requires a full-scale population survey on the topic of internal migration.

      ABS response

      Following discussion at the 1995 Commonwealth-State Population Workshop, ABS considers it inappropriate for additional data collection on reasons for moving to occur until its use in planning in Australia is clearer. The reasons why people move are complex comprising individual, household (such as the family has always lived in this area) and contextual (such as the economy) reasons and are unlikely to be readily quantified using standard ABS population survey methods.

      BIMPR has been approached by a non-ABS party to consider a proposal aimed at establishing a conceptual framework which would identify the need for different types of demographic information. It is envisaged that such a framework would be in the context of the social, economic, environmental, etc factors taken into account in planning.

      Alternatives to Medicare interstate migration data sources

      Comments received

      There is general support to improve the quality of interstate migration estimates. Current internal migration data collected by the existing method is deemed to be far from reliable, especially for small areas. Some agencies suggested possible alternatives to Medicare as a source for internal migration data. The Commonwealth Department of Health and Human Services proposed the use of mortgage or taxation data. The NSW Department of Urban Affairs and Planning suggested that the ABS expand the scope and sample size of the labour force survey.

      ABS response

      A review of alternative to Medicare interstate migration data sources is planned to be discussed at the 1996 Commonwealth-State Population Workshop prior to decisions being taken on the updating source for the 1996 - 2001 intercensal population estimates. ABS will continue to explore other options as appropriate.


      4 OTHER ISSUES RAISED OUTSIDE THE SCOPE OF THIS REVIEW

      Aboriginal and Torres Strait Islander Statistics

      Comments received

      It is widely argued that the ABS should promote statistical leadership by ensuring that all relevant ABS data collection seek and produce Aboriginal and Torres Strait Islander details as a matter of course. The further development of Indigenous statistics is promoted by agencies such as the Department of Health and Human Services, DEET, the Department of Social Security, the NT Office of Aboriginal Development, NSW Health and the NSW Department of Local Government. NT Health Services commented that, "The very poor state of population statistics on indigenous people is the most serious inadequacy in Australian population statistics." The ABS should also be looking to develop data on demographic concepts appropriate to Aboriginal culture, especially in relation to family structures.

      ABS response

      ABS recognises the high level of demand for high quality statistics on Australia's Indigenous population and is currently considering guidelines for the most appropriate means of collecting and producing Aboriginal and Torres Strait Islander statistics from its collections. A greater effort than ever before will be undertaken in the 1996 Census to ensure the highest possible quality Indigenous enumeration and to ensure that Census output presents Indigenous population statistics in the most useful form to users. In the coming years further work will be undertaken to refine the methods for estimating Indigenous populations. While ABS is aware of the inadequacies of conventional family definitions when applied to some groups of Indigenous people, there does not yet appear to be a satisfactory way of defining more culturally appropriate constructs in a form suitable for use as a statistical classification. Work will continue in this area.

      Statistical Geography

      Comments received

      Several local governments and NT agencies identify limitations with the size of the SLA level. They propose that the SLA is reduced in size as it currently accommodates too many variations in population and is not responsive to many community's needs.

      ABS response

      The ABS has just commenced a review of the Australian Standard Geographical Classification in which requirements for different spatial geography will be canvassed. As part of that review the above requirements will be considered and proposals for future standard geography outputs developed to address major user requirements.

      Reference person

      Comments received

      The NSW and QLD planning authorities suggest that reference person data from the Census of Population and Housing continue to be made available as most household projection methodologies used by the States rely on such data.

      ABS response

      Reference person data will continue to be made available as a customised services from the 1996 Census even though ABS questions the conceptual soundness of using these data as a proxy for household head data.

      Dwelling estimates

      Comments received

      The NSW Department of Urban Affairs and Planning is interested in estimates of the number of dwellings and suggests that this should be published.

      ABS response

      The Building Statistics Review conducted in 1994 concluded that there is widespread support for data on stock of dwellings at the national, State and small area (SLA/LGA) levels. Stock of dwelling estimates would be based on the latest Population Census dwelling counts plus commencements, plus net conversions, minus demolitions and relocations in the intercensal period. For WA, stock of dwelling estimates are produced annually by ABS at the Collection District level in Perth and major cities. They provide an important base for population estimates.

      In 1997 the ABS plans to explore the feasibility and cost of producing some estimates of stock of dwellings in this way for other States and will consult with major users as appropriate.

      Defacto relationships

      Comments received

      There is interest from organizations such as the Department of Social Security and State departments in rates of partnership formation and dissolution in both registered marriage and informal unions. ANU argues for a Family Survey conducted every five years.

      ABS response

      The ABS currently produces statistics about the formation and dissolution of registered marriage through marriage and divorce registration data. We are aware that there is also considerable demand for data on the formation and dissolution of relationships (as distinct from data on current de facto relationships). In considering unmet user needs for family-related data, the ABS felt that there was insufficient justification for conducting another family survey in 1997, even though a number of user requirements cannot be met through existing data holdings on families. The ABS is currently investigating alternative options for addressing such requirements. These include conducting an annual supplementary survey which collects regular family and relationship indicators, as well as investigating the feasibility of conducting a longitudinal survey on families and children.


      ATTACHMENT A

      List of organizations which have provided comments

      Commonwealth Government
      Australian Electoral Commission
      Australian Institute of Family Studies
      Bureau of Immigration, Multicultural and Population Research
      Commonwealth Grants Commission
      Department of Employment, Education and Training
      Economic Analysis Branch
      Department of Housing and Regional Development
      Department of Human Services and Health
      Department of Immigration and Ethnic Affairs
      Department of the Prime Minister and Cabinet
      Department of Social Security
      Attorney-General's Department
      Legal Aid and Family Services
      The Treasury

      New South Wales Government
      Department of Aboriginal Affairs
      Department of Bush Fire Services
      Department of Community Services
      Department of Corrective Services
      Department of Local Government
      Department of Sport and Recreation
      Department of Training and Education Co-ordination
      Board of Vocational Education and Training
      Board of Studies
      Department of Transport
      Department of Urban Affairs and Planning
      Demographic Unit
      Metropolitan Planning Division
      Strategic Policy and Planning Division
      Environment Protection Authority
      Ethnic Affairs Commission
      Health Department
      Office of the Board of Studies
      Pacific Power
      Road and Traffic Authority
      State Electoral Office
      State Rail Authority
      State and Regional Development
      State Transit Authority
      Sydney Water Corporation
      Tourism NSW
      Transgrid
      Treasury
      Workcover

      Victorian Government
      Department of Arts, Sport and Tourism
      Arts 21
      Sport and Recreation Victoria
      Department of Education
      Portfolio Resource Planning
      Department of Health and Community Services
      Public Health Branch
      Children's and Family Services
      Department of Premier and Cabinet
      Ethnic Affairs Unit
      Development Albury-Wodonga 2000
      Victorian Textile Clothing & Footwear Industry Training Board
      Victorian Health Promotion Foundation

      Queensland Government
      Department of Local Government and Planning
      Government Statistician's Office

      Western Australian Government
      Department of Family and Children's Services
      Local Government Grants Commission
      Ministry for Planning
      Research Branch
      Registrar General's Office
      State Treasury
      Western Australian Electoral Commission

      South Australian Government
      Office for the Ageing
      South Australian Health Commission
      Policy and Planning Division
      Department of Housing and Urban Development
      Planning Division
      South Australia Police

      Tasmanian Government
      Department of Community and Health Services
      Department of Education and the Arts
      Department of Environment and Land Management-
      Planning Division
      Land Information Bureau
      Department of Premier and Cabinet
      Department of Treasury and Finance

      Northern Territory Government
      Department of Education
      Department of Housing and Local Government
      Aboriginal Housing and Infrastructure
      Housing Policy Division
      Department of Lands, Planning and Environment
      Office of Aboriginal Development
      Territory Health Services
      Treasury

      Local Government
      Liverpool City Council
      Penrith City Council
      Sutherland Shire Council

      Other
      Australian Housing and Urban Research Institute
      Australian National University
      Demography Program
      Relationships Australia (SA)
      University of Melbourne (Economics Department)




      Demography Program
      Australian Bureau of Statistics
      July 1996