Table of Contents
Over the last two years, Indigenous affairs policy at the national level has shifted direction dramatically: the central tenets of policy have shifted from terms such as self-determination, self-management and national Indigenous representation and advocacy to mainstreaming, mutual obligation, shared responsibility and a whole-of-government approach.
This broad change in direction has been predicated in large measure on a widespread perception that the socioeconomic situation of Indigenous people in Australia has, at worst, been a failure over the past 30 years or, at best, has not improved fast enough.
The new approach has been based on a growing emphasis on what has been termed ‘practical reconciliation’, or the pursuit of statistical equality between the standard of living of Indigenous and other Australians in the areas of health, housing, education and employment. In the foreword to the latest and influential Overcoming Indigenous Disadvantage: Key Indicators Report 2005, Productivity Commission Chairman Gary Banks refers to these issues as follows. Firstly, he notes the new determination by Australian governments to address the root causes of Indigenous disadvantage. Secondly, referring to the Prime Minister’s speech at the National Reconciliation Conference, he identifies the shared goal [sic] that Indigenous people can ultimately enjoy the same standard of living as other Australians, for them to be as healthy, as long-living and as able to participate in the social and economic life of the country (SCRGSP 2005: iii).
We are almost certain that no-one would quibble with the need to address the causes of Indigenous disadvantage or the need for improvements in standards of living. A more lively debate might ensue about whether in the last 30 years any government has ever pursued objectives different to these; or whether the new approach is likely to be more successful than previous approaches. Arguably, it is difficult enough having conversations about what has worked and what has not in the past, let alone predicting what might be the best approach for the future.
Under such circumstances of policy contestation, one might anticipate that statistics collected by the official agency, the ABS, would play a crucial role in clarifying both the causes of Indigenous disadvantage and trends in Indigenous wellbeing, in absolute and relative terms. That is certainly one expectation that animated the conference ‘Indigenous Socioeconomic Outcomes: Assessing Recent Evidence’. Our focus is on the 2002 NATSISS, the second social survey conducted by the ABS between August 2002 and April 2003. It is based on information collected from 9400 Indigenous Australians aged 15 years and over across all States and Territories of Australia (ABS 2004c: v).
The aims of this chapter are three-fold:
To provide some historical background about the emergence of this survey instrument by linking this conference to earlier workshops convened by CAEPR in 1992 and 1996 about the initial 1994 NATSIS.
To broadly examine the role and value of the 2002 NATSISS as one element of a broad ABS strategy to enhance the availability of statistics on Indigenous Australians.
To explore the longer-term role that the NATSISS instrument, now to be conducted every six years, might play in tracking Indigenous policy performance, informing Indigenous policy development, and in meeting the statistical needs of all stakeholder groups, including Indigenous Australians.
These three aims should be differentiated from the overall aim of the conference, which is to examine the 2002 NATSISS in great detail from a variety of methodological (e.g. coverage), thematic, and conceptual perspectives as the latest evidence about Indigenous socioeconomic outcomes. Then we anticipate some discussions of issues for consideration in developing the 2008 NATSISS. For these purposes, a wide array has been assembled of social scientists and Indigenous policy practitioners who are specialists in disciplinary areas.
Conference participants were provided with two earlier monographs published by CAEPR, A National Survey of Indigenous Australians: Options and Implications (Altman 1992) and The 1994 NATSIS: Findings and Future Prospects (Altman & Taylor 1996a). We summarise some salient issues from these monographs to provide a historical backdrop to this conference.
The first of these monographs reports the findings from a workshop, titled ‘A National Survey of Aboriginal and Islander Populations’, that was convened in April 1992. This was held, coincidentally, just a week after the Keating government committed $4.4 million to the ABS to conduct a special national survey as outlined in recommendation 49 in the Final Report of the Royal Commission into Aboriginal Deaths in Custody (Commonwealth of Australia 1991). The workshop provided an early opportunity for the ABS to engage with a diversity of academics and Indigenous stakeholder groups about the conduct and potential content of such a national survey.
The impetus for this recommendation in 1991 came from two directions. On one hand, the Royal Commission itself experienced an acute shortage of information about Indigenous Australians in its deliberations. In particular, its conclusions about the underlying factors that precipitated higher Indigenous incarceration were mainly drawn from the 1986 Census (Gray & Tesfaghiorghis 1991). There was a strong view articulated by the Royal Commission that additional, and more timely, information about distinct and diverse social, demographic, health and economic characteristics of the Indigenous population was needed.
On the other hand, there was an equally strong view at the time that information was urgently needed at the just-established ATSIC Regional Council level to facilitate the development of regional plans and to gauge community infrastructure needs. It is interesting to note other policy drivers for a national survey and more statistical information on the situation of Indigenous Australians in 1992, including:
a need for regional statistics so that Indigenous-specific programs could be better targeted to Indigenous people in greatest need
debates about whether Indigenous-specific or mainstream programs were more efficient and effective
a recognition of the need for more concerted and coordinated inter-governmental effort in Indigenous affairs, including the need for legally-binding bilateral agreements between the Commonwealth and the States, and
a need for more information to inform the reconciliation process and track change in ‘Indigenous disadvantage’ over time.
The 1992 workshop ended with some recommendations both for the forthcoming 1994 NATSIS and for the collection of Indigenous statistics more generally. In terms of items to be included in the survey, there was a consensus view that a national survey should focus on distinct Indigenous issues not covered in other collections. Even by 1992, the ATSIC Housing and Infrastructure Needs Survey (Australian Construction Services 1993) was addressing many of the topics specified for inclusion by the Royal Commission. And there was a strong view expressed that a one-off survey would be of limited value without commitment to future surveys conducted on a similar and regular basis.
More importantly, it was highlighted that a national survey could not be an omnibus for all statistical needs about Indigenous Australians: there was a need for enhanced use of Indigenous identifiers in regular ABS household surveys, for greater availability of Indigenous identified information in administrative databases, and more information at the State and Territory level. Finally, there was a strong view that, while committing $4.4 million for data collection was welcome, there was also a need for resources to ensure data analysis, publication and dissemination, especially at the regional level, to Indigenous stakeholders. [1]
The second monograph reports the findings from the workshop ‘Statistical needs for effective Indigenous policy: findings from the 1994 NATSIS’. This workshop was convened in late August 1996, just on nine years ago, not long after the new Howard government’s 1996–97 Budget where it signaled its new direction in Indigenous affairs.
The workshop was convened to assess 1994 NATSIS outputs from academic, bureaucratic, Indigenous and ABS staff perspectives with a special focus on issues to be considered in the development of a future national survey. While the 1996 workshop examined NATSIS from a number of disciplinary perspectives that closely correlated with survey questions, it was not an evaluation of the survey, which had already been conducted by the ABS (1996b) itself (with input from CAEPR & other stakeholders).
In considering the need for a repeat survey, the 1996 workshop, like this conference, sought open, constructive and rigorous scrutiny of the 1994 NATSIS. This is something that is not always easy to achieve, as new and innovative surveys evolve over time into statistical institutions with an identity and a sponsoring agency. Then, as now, the ABS, was an active and lively participant in proceedings.
As for the utility of the 1994 NATSIS, participants were acutely conscious of a fundamental dilemma: the more critical they were of particular questions, the greater the likelihood that questions would be modified or omitted, thus undermining the longer-term comparative utility of a future survey. The second workshop also examined what alternate sources of data were emerging that could be excluded from a repeat of the 1994 NATSIS, again highlighting that a national survey should not provide the means to ‘statistically cost shift’ from other household surveys where an Indigenous identifier might be justified (examples provided were the Labour Force Survey and the National Health Survey).
It is again interesting to note the political context of the 1996 workshop, when there was a perception that the new government might move to bifurcate its policy approach: targeting Indigenous-specific programs at remote regions and enhancing access to mainstream Commonwealth and State programs in non-remote contexts.
If this was a policy option then, it was only implemented by cutbacks to ATSIC Indigenous-specific programs as a contribution to the then overall ‘deficit reduction plan’. It was highlighted in 1996 that the absence of statistics about performance at the program level made agencies like ATSIC vulnerable to cutbacks and, conversely, that comparative statistics would be needed to hold governments accountable for their performance.
In this context, it was highlighted that the political economy of statistics—how statistics influence the distribution of public money—was important. [2]This, in turn, influenced what information was to be collected, how Indigenous participation in the national survey was to be encouraged, and how the empowering information from a survey was to be analysed, accessed (taking into account income disparities and the ABS’s user pays approach) and disseminated.
The 1996 workshop raised some important issues. Two that stand out in current policy debates are the need for information at appropriate regional levels, and the links between political and statistical cycles.
In relation to the former, it was noted that ATSIC regional council boundaries were liable to change and, in any case, might have limited utility given that the discretion available to the ATSIC Board of Commissioners to make regional allocations had declined significantly. While ATSIC has disappeared, it is still important to consider the appropriate regional jurisdictions for data collection.
On the latter, it was noted in 1996 that there would be an urgent need to monitor outcome changes in Indigenous affairs as policy takes new directions. While this observation may not have been heeded at the time, it certainly resonates with the contemporary situation. The 1996 workshop made three concluding recommendations:
that there should be a repeat of the 1994 NATSIS, preferably in 1999 after a five-year interval, but that this should be leaner, meaner and sharper
that careful consideration be given to maintaining some questions in the national survey that are comparable over time, while also allowing for new questions that allow comparability with other data sets, and
that consideration be given to enhancing existing ABS household surveys to allow the identification of Indigenous participants in a statistically valid way that reflects their different geographic distribution. This is done, for example, with the Labour Force Survey once a year, but little is heard about the statistical outcomes.
Each of these issues could well be revisited in the context of this conference.
In the not-so-distant past, official processes served to exclude, devalue, and deter full Indigenous statistical representation. By contrast, the contemporary politics of data collection have sought to encourage inclusion via self-identification. This is manifest most recently in the greater involvement of Indigenous personnel in the collection of census and survey data, as well as in ministerial-level agreements for the adoption of a standard self-reported Indigenous status question in administrative collections.
In addition, the ABS has embarked on an ambitious Indigenous household survey and census program (see Table 2.1) that can trace its origin to the need for a government response to the findings of the 1991 Royal Commission into Aboriginal Deaths in Custody (Sims 1992).
As with other Indigenous-specific population surveys, such as the Western Australian Aboriginal Child Health Survey (Zubrick et al. 2004), this response is recognition of the need for non-standard approaches to developing census and survey content and methodology with an emphasis on Indigenous participation and, in some cases, control (Zubrick et al. 2004).
Complementing these enhanced census and survey initiatives, we now also have regular reporting of Indigenous outcomes from administrative data. In 2002, the Prime Minister wrote to Gary Banks as Chairman of the SCRGSP to produce a regular report to the COAG against key indicators of Indigenous disadvantage. This is now available on a biannual basis as the Productivity Commission Report Overcoming Indigenous Disadvantage. In addition, the annual Report on Government Services issued by the SCRGSP now includes a separate compendium of Indigenous statistics drawn from the administrative databases of Australian, State and Territory governments.
Table 2.1. ABS Indigenous household surveys program: 1999–2011
|
Year |
Collection |
Indigenous sample |
Level of geography supported |
|
1999 |
Housing Survey (a) |
850–900 households |
National |
|
1999 |
CHINS |
All discrete Indigenous communities (approx. 1300) |
Community level |
|
2001 |
NHS (& Indigenous supplement) |
Indigenous sample of 3400 persons |
National |
|
2001 |
CHINS |
All discrete Indigenous communities (approx 1300) |
Community level |
|
2001 |
Population Census Indigenous Enumeration Strategy |
All persons |
Small geographic regions |
|
2002 |
NATSISS |
9 400 persons |
States/Territories |
|
2004–05 |
Indigenous Health Survey |
11 000 persons |
States/Territories |
|
2006 |
CHINS |
All discrete Indigenous communities (approx. 1300) |
Community level |
|
2006 |
Population census |
All persons |
Small geographic regions |
|
2008 |
NATSISS |
11 000 persons |
States/Territories |
|
2010–11 |
Indigenous Health Survey |
11 000 persons |
States/Territories |
|
2011 |
Population Census Indigenous Enumeration Strategy |
All persons |
Small geographic regions |
Source: Adapted from www.abs.gov.au , Themes—Indigenous: Directions in Aboriginal and Torres Strait Islander Statistics, March 2000
While these developments clearly demonstrate rapidly enhanced collection and publication of statistical information, much basic information remains unavailable – not so much because data do not exist, but rather for want of appropriate planning frameworks. Viewed historically, there appears to be a growing gap between the scales at which Indigenous polities organise and plan, on the one hand, and the scales for which statistics are available, on the other. Thus, while we have never been so data-rich, leaving aside questions about the cultural relevance of the data, the information that is generated at great expense tends increasingly to be only useful for national, State and Territory level analysis, and/or broad remote/non-remote distinctions. There are a number of methodological, practical, and ultimately political dimensions to this observation.
First, from a methodological perspective, sample instruments such as the NATSISS and its companion General Social Survey (GSS) are best deployed to inform high-level policy discussion about the broad nature of inter-relationships between social circumstances and outcomes. Does crime impede employment prospects? Is health related to income? Do educated women have fewer children? For reasons of sample size and non-response error, they are less well suited to establishing absolute levels of need and comparisons over time.
Related to this is the fact that major constraints arise in relation to the geographic disaggregation of data to regional or community levels. Although the sample size of the adult population in the 2002 NATSISS was equivalent to that of the 1994 NATSIS, data from the latter were released at the scale of 36 ATSIC regions, whereas the basic geography for 2002 comprises the remote and non-remote areas of the States and the Northern Territory (NT). The basic rationale for this shift is to be found in the trade-off between spatial detail and robustness of results.
Secondly, from a practical perspective, in terms of gathering administrative data, a combination of under-reporting, confidentiality provisions, and the grinding and increasingly guarded nature of bureaucratic processes render even the acquisition of data at State and Territory levels a major achievement (see Taylor & Stanley 2005).
Finally, from a political perspective, the current paradigm for the collection and dissemination of Indigenous statistics is suited to the measurement and reporting of gaps. For the most part, this assists processes of government by bureaucrats in Canberra and the State capitals. This paradigm strongly reflects a deficit model of Indigenous socioeconomic need as measured by standard social indicators, and not a community development model. This approach suffers from all the pitfalls of averaging diverse circumstances leading to questions about the utility of data for Indigenous regional and community organisations and their members.
Worst of all, it lacks local Indigenous context. Demographers, for one, were warned some years ago of an imbalance between their statistical precision and detail, on the one hand, and their casualness of treatment of non-demographic contextual variables on the other (McNicoll 1988: 20 cited in Riedmann 1993: 107). Yet perennial problems about the ability of such an approach to capture cultural difference and the diversity of Indigenous circumstances looms yet again (Morphy 2002; Peterson 1996). As social scientists and policy practitioners, we need perhaps to revisit the original goal of the NATSISS instrument to capture information that is distinctly Indigenous and that can be made available at the regional level.