PROSPECTS FOR ‘CLOSING THE GAP’ IN SOCIOECONOMIC OUTCOMES FOR INDIGENOUS AUSTRALIANS?

Authors

  • Jon C. Altman,

    1. The Australian National University
    Search for more papers by this author
  • Nicholas Biddle,

    1. The Australian National University
    Search for more papers by this author
  • Boyd H. Hunter

    1. The Australian National University
    Search for more papers by this author
    • *

      The data collection was partly funded through an Australian Research Council linkage project with the Australian Bureau of Statistics (LP0348733). An earlier version of a part of this paper was presented to the 2007 Australian Institute of Aboriginal and Torres Strait Islander Studies conference, Forty Years On: Political Transformation and Sustainability since the Referendum and into the Future, and at a seminar given at the University of Wollongong. Comments were generously given by Mike Dillon, Bill Fogarty, Melinda Hinkson, Melissa Johns, Will Sanders, John Taylor and two anonymous referees. While the support of the ABS is appreciated, this paper should in no way be attributed to the ABS.


Abstract

Practical reconciliation’ and more recently ‘closing the gap’ have been put forward as frameworks on which to base and evaluate policies to address Indigenous disadvantage. This paper analyses national-level census-based data to examine trends in Indigenous wellbeing since 1971. There has been steady improvement in most socioeconomic outcomes in the last 35 years; a finding at odds with the current discourse of failure. Evidence of convergence between Indigenous and non-Indigenous outcomes, however, is not consistent. For some outcomes, relatively rapid convergence is predicted (within 25 years), but for the majority of outcomes, convergence is unlikely to occur within a generation, if at all.

INTRODUCTION

The notion of ‘closing the gap’ between Indigenous and non-Indigenous life expectancies has attracted considerable attention in Australia since the election of the Rudd Government in November 2007. The aim of this paper is twofold. The first is to examine national level data from the last eight censuses in order to document the historical progress in ‘closing the gap’ across a range of broad socioeconomic indicators that are often associated with health outcomes, and which may explain differences in life expectancy. The second is to make predictions about when statistical equality between Indigenous and non-Indigenous Australians might be achieved if the historical trends that we identify were to continue. Before outlining the paper in more detail, it is worth considering the origins of the notion of ‘closing the gap’, especially as it relates to Indigenous policy in Australia.

The precise origins of the term ‘closing the gap’ can probably be traced to the special programs of governments in New Zealand in the 1990s that sought to target Maori and Pacific Islander disadvantaged groups with assistance.1 In Australia, a similar policy of targeted assistance to Aboriginal and Torres Strait Islanders had come to prominence in the late 1980s with the Aboriginal Employment Development Policy of the Bob Hawke Labor government committing to statistical equality in employment, income, and educational status between Aboriginal and other Australians by the year 2000.2 There appears to be no substantive difference between ‘closing the gap’ and ‘statistical equality’; both concepts seek to eliminate socioeconomic disparities. This too was the intent of the John Howard Liberal government's notion of ‘practical reconciliation’ that came to prominence after Howard's election victory speech in October 1998. As a policy framework, practical reconciliation sought to reduce Indigenous material disadvantage in the areas of health, housing, education, and employment.3 This goal was closely aligned with a growing emphasis during the decade from 1998 on mainstreaming or normalisation, but successive Howard governments were careful not to stipulate precise timeframes for achieving such goals.4 A strong case can be made that the goal of practical reconciliation, without any clearly specified timeframes for achievement of such goals, was itself highly symbolic. It was good political rhetoric but did not provide a basis for a comprehensive policy framework.

It was only in the aftermath of the Northern Territory National Emergency Intervention that the Howard government committed, in its last days, to concrete targets with its pledge to ‘stabilise, normalise and exit’ 73 prescribed communities in five years.5 Arguably, this pledge could be interpreted as a commitment to close the gap within that time frame. In response to the federal intervention, the language of ‘Closing the Gap’ was adopted by the Northern Territory Government in August 2007 with its Closing the Gap of Indigenous Disadvantage: A Generational Plan of Action.

While in Opposition the current Prime Minister, Kevin Rudd, had already supported a ‘Close the Gap’ campaign, championed by Oxfam, Get Up, Australian for Native Title and Reconciliation and other Australian non-government organisations, that focused on the 17-year life expectancy gap between Indigenous and non-Indigenous Australians.6 As part of the national apology to the Stolen Generations made on 13 February 2008, specific reference was also made to closing the gap in Indigenous and other child mortality rates within set time frames. In March 2008, Prime Minister Rudd signed a declaration of intent to eliminate the life expectancy gap between Indigenous and other Australians within a generation while in April of the same year he made a commitment to have annual progress reports on meeting this target tabled in parliament.7 While the term initially focused on the life expectancy gap, the concept has gradually changed to encompass other forms of disadvantage; an acknowledgement that social determinants are likely to be crucial factors driving the disparities between Indigenous and non-Indigenous health.8 As mentioned above, this paper sets out to do two things – each with some precedents in the recent research literature.

First, in 2004 we collaborated in an article that provided an historical account of Indigenous socioeconomic change between 1971 and 2001.9 Examining differences between Indigenous and non-Indigenous socioeconomic outcomes has been a policy research issue for one of us since 1971 Census data became available.10 Tracking such change over time has been a core objective of the Centre for Aboriginal Economic Policy Research since its establishment in 1990.11 Such analysis has focused on absolute and relative change according to standard social indicators over time. Our initial aim is to update this time series with 2006 Census data that became available late in 2007.

Second, and more innovatively, we use the emerging trends from data collected in the last eight censuses to extrapolate when the existing statistical gap between Indigenous and non-Indigenous outcomes might close if current trends were to continue. We are not aware of such extrapolation being attempted across a similar number of variables. Previously projections tended to focus on short-run to medium-term demographic change and population growth with most forecasts being used to predict the associated impacts on employment in reports with titles like The Job Ahead and The Job Still Ahead.12 Earlier still, a number of publications had predicted that the Hawke government's goal of statistical equality for Indigenous Australians by 2000 was destined to fail for a number of reasons including historical legacy, cultural difference, diversity of circumstances, geographic distribution, and unanticipated rapid population growth.13

The likely success of the ‘closing the gap’ policy also depends on what happens in non-Indigenous Australia. Obviously ‘the gap’ is directly affected if the non-Indigenous benchmark changes, however, the ability of Indigenous outcomes to reach such benchmarks will also change with variations in social and economic conditions. An important consideration in this regard is variation in macroeconomic growth across the business cycle. Low skilled workers with little experience, a group that includes most Indigenous people, tend to be the last workers hired in a period of macroeconomic growth and the first workers shed in an economic downturn. Businesses often rationalise this behaviour on the grounds that they want to minimise turnover of their most experienced (and usually most high value-added) staff. The crucial point is that the ability to close the gap cannot be sustained indefinitely as it depends on macroeconomic growth, which by definition goes up and down with the business cycle. Therefore relative Indigenous outcomes are likely to improve during sustained periods of economic growth, but all else being equal, relative outcomes tend to stagnate or get worse in recessionary periods of the cycle. This point is all the more important when one considers the current period of uncertain global economic growth.

After discussing relevant methodological issues, the paper provides data on currently observed trends in Indigenous and non-Indigenous outcomes since 1971. It is not possible to assume that there is always going to be convergence in outcomes, so the following census-based analysis merely provides two scenarios for estimating some temporal bounds on when the gap might be closed. Improving the socioeconomic circumstances of a sub-population relative to the rest of the population is a complex policy task. We focus on relatively optimistic scenarios where there is convergence as these provide evidence of what might happen in the future if all the ‘macro’ factors were working in the right direction to close the gap. The concluding section reflects on some of the issues arising from this analysis.

METHODOLOGICAL CONSIDERATIONS

The methodological challenges that arise when using official statistics from the census need to be made explicit. While we have learnt from our experience of describing economic history, there are several issues that arise for any researcher undertaking systematic analysis of the national population of Indigenous Australians.14

These caveats can be categorised into three groups: practical problems, methodological issues, and conceptual difficulties. One practical issue is that, over the past 35 or so years, the size of the Indigenous population has increased dramatically. One reason is that Indigenous people have become increasingly confident about self-identifying as Indigenous since 1971. This means that one cannot be entirely sure that observed trends are due to real changes in outcomes rather than to changes in composition of the population. Hunter argues that one could ignore this issue, at least for some of the period we examine, because it does not have a significant impact on socioeconomic outcomes of Indigenous cohorts relative to similar cohorts in the rest of the Australian population.15 Another practical problem is that changes in census questions and coding make it difficult to get precise comparisons over time. The important thing here is that analysts should be transparent about any assumptions they make in constructing trends. Most of the assumptions used in this paper were outlined in Altman, Biddle and Hunter, but there are a few related issues for the 2006 Census and these will be discussed below.16

A methodological issue is that the Australian Bureau of Statistics (ABS) did not initially have a great deal of experience in collecting comprehensive statistics about Indigenous Australians before the 1971 Census. While the ABS has learnt from its experiences over time, improved data coverage (and hence quality) may itself alter observed trends. Conceptually two main issues need to be emphasised.

First, a person's Indigenous and non-Indigenous status is primarily based on self-identification derived from a census form completed by individuals. Even here there is some variation – in remote regions a special enumeration strategy is used that allows ABS staff to target Indigenous people and assist them in completing specially designed Indigenous personal and household census forms.17 Thus the Indigenous sub-population used here (and generally) for comparative purposes is a statistical aggregation of individuals who identify, or who are identified as Indigenous, rather than a well-defined sociological category. Second, the social indicators that are drawn from census questions reflect the social norms of the dominant mainstream society. Arguably, a number of social indicators that we use mean different things to Indigenous and non-Indigenous Australians.

Following Altman, Biddle, and Hunter, this paper focuses only on national trends.18 The main shortcoming of this approach is that a national analysis loses the diversity of geographic outcomes, like the local labour market conditions that are likely to be particularly important in influencing changes in socioeconomic status. However, the benefit of a national analysis is that the problem of selective migration between regions will not undermine the robustness of the analysis. Another advantage of a national analysis is that it dilutes the possible impacts of changed Indigenous population composition. It is impossible to undertake rigorous sub-national analysis of trends over such a long time period because of the large regional variations in the unexplained component of Indigenous population growth – that is the population growth related to increased self-identification and inter-marriage between Indigenous and non-Indigenous people. Settled parts of Australia are likely to have particularly large increases in the number of people who now identify as Indigenous but did not do so in previous censuses.19

The main rationale for using a national approach is that current and former Federal governments have all set their targets at this level. In terms of targeting policy and service delivery, regional based approaches are likely to be more useful and there are some recent examples of this.20 However, in terms of evaluating past policy and highlighting future prospects, a national approach is arguably more appropriate.

We have selected robust variables based on the availability of historically comparable data and our previous experience in constructing such trends.21 There have been eight censuses between 1971 and 2006, but not all provide reliable information about Indigenous Australians. The 1976 Census only analysed information on 50 per cent of the population (as a cost cutting measure) and is therefore difficult to compare with the other censuses. If nothing else 1976 data will be less reliable and any trend calculated using those data are likely to be problematic. The other consideration is that some data were not collected in all censuses. For example, income data were not collected in 1971.22

A problem for comparative inter-censal analysis is that the questions asked and the coding of responses changed significantly between censuses. Fortunately, the last three censuses have broadly comparable questions, and to a lesser extent comparable coding, and provide the most reliable estimates of short-term trends.23 The down-side of using these short-run trends between 1996 and 2006 is that this period falls entirely within the growth phase of the economic cycle. The extrapolation of trends usually requires more than a few observations, so whenever possible we also use the long-run trends based on information collected as far back as 1971. When no comparable data are available for earlier censuses before 1996, the earliest available valid data are used as the base from which to calculate trends.

Even where data were collected, they are not always available in a usable form. For censuses before 1996, we often had to rely on published data, which were often not sufficiently flexible to allow the calculation of comparable statistics. The 1986 Census is particularly problematic in this regard and was consequently not used here to estimate long-term trends in socioeconomic outcomes for Indigenous and other Australians.

TRENDS IN INDIGENOUS OUTCOMES SINCE 1971

The 2006 Census allows us to update the analysis in Altman, Biddle, and Hunter so that we can track changes in socioeconomic outcomes since 2001.24 This paper augments the tables in the 2005 paper by including comparable data from the 1996 and 2006 Censuses in order to provide detailed information on trends in socioeconomic outcomes for both Indigenous and non-Indigenous Australians.25 The following narrative provides a broad summary of this information documented in Appendix A.

Labour force status

The early 1970s were characterised by some of the lowest unemployment rates in Australia's recorded history. In contrast, the stagflation of the mid-1970s and the international recession of the early 1990s saw large increases in unemployment that were particularly pronounced for the Indigenous population. Since that time unemployment has fallen for Indigenous and non-Indigenous groups in one of the longest unbroken period of economic expansion ever seen in Australia. Indigenous unemployment rates have very similar trends to non-Indigenous rates – the main differences arise from the Indigenous labour force being disproportionately affected by the recessions of the early 1980s and 1990s.

A complicating factor for interpreting Indigenous labour force status is the rapid expansion of the Indigenous-specific Community Development Employment Program (CDEP). CDEP was devised by the Malcolm Fraser Liberal government in 1977 when unemployment benefits were first being introduced to remote Indigenous communities.26 In the early 1980s, some ‘teething’ problems with CDEP were addressed and the scheme began expanding. Notwithstanding its historical connection and early notional fiscal nexus with unemployment benefits, CDEP is regarded as employment for official statistical purposes (although the census does not have a separate category for CDEP scheme participants in the mainstream forms used predominantly in non-remote regions).

The incidence of CDEP increased only gradually and was a relatively minor influence on Indigenous employment until 1991. The scheme then expanded rapidly as a key element of the Aboriginal Employment Development Policy. It reached the height of its significance in the period analysed in the 2001 Census when it incorporated just over one-tenth (10.9 per cent) of the Indigenous adult population aged 15 and over. It has declined quite markedly since that peak, and at the time of the 2006 Census employed fewer than 7.7 per cent of Indigenous adults.27

After 1991, the substantial growth in CDEP participation means that the difference between Indigenous and non-Indigenous unemployment rates appear lower as more Indigenous unemployed are classified as employed once they take up a place in a CDEP scheme. The estimated trends in Indigenous unemployment before 1991, however, are unlikely to be effected. The existence (and growth) of the CDEP scheme means that our scenarios of when the unemployment gap might be closed are relatively optimistic.

Another way to assess the labour market outcomes for Indigenous Australians is the percentage of the population aged 15 years and over who are employed. It was only in 2006 that this ratio rose above that measured at the 1971 Census for the Indigenous population. This is not surprising since previous research has shown that the low-skilled Indigenous labour force suffered severely from the collapse in the demand for labour in the period of economic restructuring that occurred after the stagflation of the mid 1970s.28 However, the role of CDEP in causing poor Indigenous employment outcomes in recent public debates seems to be challenged by the statistics presented here. The decline in the Indigenous employment-population ratio occurred almost entirely before 1986; a period before CDEP participation became a significant aspect of Indigenous labour market experience.

If one chose to consider CDEP participation as a form of welfare rather than employment, then Indigenous employment outcomes are substantially worse today than they were in the early 1970s.29 In contrast, non-Indigenous employment improved, largely in the last three censuses. Again, the role of the CDEP scheme in improving Indigenous employment tends to provide an optimistic estimate of convergence in this indicator.

Given that the CDEP complicates our estimates of trends in Indigenous employment, it is instructive to exclude this form of employment and focus solely on private sector employment. Indigenous private sector employment was particularly affected by the structural adjustment in the Australian economy during the 1970s, especially the decline in the manufacturing and agricultural industries.30 However, private sector employment has expanded considerably since 1981 with the increase greater for Indigenous Australians than other Australians, especially in the last inter-censal period 2001–2006. While trends in private sector employment will not be influenced by the rise (and recent decline between 2001 and 2006) in CDEP employment, it will be affected by the increased incidence of privatisation that became popular with Federal and State/Territory governments since the microeconomic reforms that began in the 1980s.31 Unless privatisation disproportionately increases the number of low skilled jobs that employ many Indigenous people, it is unlikely to substantially affect our estimated date for ‘closing the employment gap’ between the Indigenous and non-Indigenous populations.

Labour force participation rates measure the percentage of the population who either work or look for work and is a key indicator of economic engagement. Participation rates increased up to 1991 for both Indigenous and other Australians, largely as a result of increased participation of females. While there was some improvement in relative participation of Indigenous population vis-à-vis other Australians between 1971 and 1991, outcomes have tended to decline since that time as Indigenous participation has fallen at a time when overall labour force participation is static.

Income status

One of the main benefits of employment is a higher income or improved access to economic resources. Appendix A shows the trends in median personal income since 1981, the earliest year that robust income data were available. Median personal income has gradually increased for the Indigenous population since 1981, but the trend is not as consistent as it is for the non-Indigenous population, which saw a combination of stagnation and decline to 1996, but a large increase since that time. Consequently, the relative income measure improved for Indigenous people until 1996, but has since declined. Notwithstanding the increases in private sector employment for Indigenous people, the financial benefits of recent macroeconomic growth in the last 10 years seem to be accruing disproportionately to the non-Indigenous population. This is consistent with the Australian inequality literature that has shown larger increases at the top end of the income distribution, historically high asset prices and sustained, substantial profits for companies.32 Assets and profits are sources of income that Indigenous Australians generally do not enjoy.

Household size, income, and home ownership

The average Australian household has been getting smaller over the period under review. While Indigenous households remain substantially larger than other Australian households, their absolute size declined substantially towards the non-Indigenous average, especially since 1991. The relative decline in Indigenous household size, compared to non-Indigenous households was limited by the ongoing demographic change that has reduced the size of many, or even most, Australian households.33 One potential positive implication of this is a lowering of the pressure on Indigenous housing stock, unless this historic decline is offset in the future by population growth that outstrips growth in new housing stock.

One reason why the decline in Indigenous household size is important is that household income is a major determinant of poverty; the more people in a household, the greater the opportunity to pool resources for living expenses. Poverty studies routinely control for household size and composition to determine changes in standard of living and welfare (using an ‘equivalence scale’ adjustment).34 At a purely mechanical level, average household income will decline with a fall in average household size; to the extent that this fall represents a decline in the number of adults with a potential to earn income. Alternatively, the recent decline in Indigenous fertility may entirely explain the fall in household size.35 If this is the case, then welfare of Indigenous households will increase because of the decline in the number of dependents using household resources.36 While it is beyond the scope of this paper to resolve such issues, they should be borne in mind when analysing changes in household income.

Raw household income fell for both Indigenous and non-Indigenous populations to 1991, and then increased substantially. Given that average personal income increased substantially for Indigenous Australians in the 10 years to 1991 (while non-Indigenous personal income did not change that much), the decline in household income is likely to be due to the declining household size. Another factor may be the relatively depressed labour market during the recession of the early 1990s, which seemed to have particularly affected Indigenous employment. That is, some households would be disproportionately affected by the relatively large number of low-income people. This observation is more germane if one takes into account that 6.8 per cent of Indigenous adults were employed in the CDEP in 1991. The increases in household income in the last two censuses are likely to be influenced by buoyant labour market conditions and historically high real wages, but it may also be associated with new generous transfers to all families.37

The rate of home ownership has been stable for many years amongst the non-Indigenous population at just over 70 per cent. If anything the incidence of home ownership has declined slightly in recent years with the widespread fall in affordability associated with the rising house prices, especially in metropolitan areas.38 The Indigenous population living in a home owned by residents declined substantially in the 20 years after 1971, but there has been steady increase since 1991. Indeed, the incidence of Indigenous home ownership has been equivalent to or exceeded the 1971 level since the 1996 Census.

Education

Overall, Australians are more likely to gain a post-secondary school qualification than ever before. Indigenous Australians have matched this trend quite closely and have even experienced a very small relative improvement in the incidence of post-school qualifications. However, the income outcomes from post-school qualifications depend heavily on the type of qualification, with degrees from universities having the largest overall economic benefits in terms of income and employment.39 Unfortunately, proportionally far fewer Indigenous people secure degrees than other Australians. Hence, the trend in qualifications is not as positive as first appearances suggest. Nonetheless, there has been steady gains in Indigenous education participation and completion over the 35 years studied here.

The complexity of interpreting the change in educational outcomes is illustrated in the recent National Report to Parliament on Indigenous Education and Training. Although this report focuses on short-term trends, Indigenous students studying at higher education declined in 2005 and was at its lowest level since the year 2000 (p. 217).40 That report speculates that some potential students postponed higher education studies to take advantage of the strong labour market and that many students viewed vocational training as a means of taking advantage of job opportunities in buoyant industries (p. xxxi).

Health

One of the best measures of health status is life expectancy. Unfortunately, it is not possible to present updated trends because the official method for estimating Indigenous life expectancy has recently changed.41 The new method controls for changing rates of Indigenous self-identification between censuses, but it is difficult to estimate long run trends because Indigenous life expectancy has to be re-estimated for each inter-censal period. Despite this, Altman, Biddle, and Hunter collated the extant comparable historical data to make an assessment of long-run trends in life expectancy before 2001.42

Indigenous male life expectancy generally did not improve at the same rate as that for non-Indigenous males; especially after 1991. Life expectancy for Indigenous females was more variable. After substantial improvement between 1971 and 1981, the relative life expectancies for Indigenous females declined gradually from 1981 while the life expectancy for non-Indigenous females increased steadily.43

Health status can also be measured by a rough proxy that is assessed consistently for all the relevant censuses: the proportion of the population aged over 55 years.44 The proportion of the population who can be characterised as ‘elderly’ is an imperfect measure of health because it is also affected by the fertility rate, which determines the total population numbers in the denominator. Notwithstanding this, an increase in the proportion of the population who are elderly can be construed as enhancing welfare if it represents either a decline in fertility or a reduction in adult mortality. If fertility is the driving factor, then Indigenous households will benefit from a long-term reduction in age-related dependency rates, which ultimately reduce the demands on household financial and infrastructural resources.

The percentage of the Indigenous population who were elderly between 1971 and 1991 decreased, signalling a worsening outlook. After 1991, Indigenous outcomes did not change relative to the non-Indigenous outcomes until the latest census. While the Indigenous demographic profile is not substantially different to that in 1971, the non-Indigenous profile has changed to reflect the aging of the population in the last three and a half decades.

Summarising long- and short-run trends

Given that there are well-founded concerns about the quality and consistency of historical data on Indigenous people we use two methods to estimate the period until the various ‘gaps’ in outcomes might be closed. The first set of estimates is based on long-run trends that use as much information as is available and which we judge to be reliable and comparable. This uses data back to 1971. The second set of projections estimate trends based on three observations from the post-1996 period. These short run trends are provided because we can be reasonably confident that such data are reliable and relate to an Indigenous population which is enumerated relatively comprehensively.45

The results from Appendix A are also summarised in absolute and relative terms in Table 1. This summary is provided according to changes observed for the long series, 1971–2006 and for the short series 1996–2006. Absolute differences are based on Table A1 data, and relative differences from Table A3 that provides Indigenous to non-Indigenous ratios.

Table 1. Absolute and relative change in Indigenous socioeconomic outcomes, long and short series
 Long seriesShort series
1971–2006 absolute change1971–2006 relative change1996–2006 absolute change1996–2006 relative change
  • Based on Tables A1–A3. Note that where data are not available for one census the latest available data are used.

  • Some improvements are due to minor positive changes not evident in Appendix A, where ratios have been rounded to two decimal places. + and − denote an increase or decrease (respectively) in the relevant outcome. NA, not applicable.

Unemployment rate (% labour force)++
Employment to population ratio (% adults)++
Private-sector employment (% adults)+++
Labour force participation rate (% adults)++
Median weekly personal income $A (2006)+++
Household size++++
Median weekly household income $A (2006)++++
Home owner or purchasing (% population)++++
Never attended school (% adults)++++
Post-school qualification (% adults)++++
Degree or higher (% adults)NANA++
Attending educational institution (% 15–24-year-olds)NANA++
Male life expectancy at birth (years)++
Female life expectancy at birth (years)++
Population aged over 55 years (%)+++

Rather than distract the reader with excessive detail, the outcomes from these statistics are presented in Table 1 with a plus (+) to indicate improvement and with a minus (−) to indicate decline in socioeconomic outcome. Our findings can be briefly summarised as follows. For the longer series there are improvements in 12 of 13 variables in absolute terms and 10 of 13 variables in relative terms. For the shorter, more statistically reliable series, there are improvements in absolute terms for 12 of 15 comparable variables and in relative terms for nine of 15 variables. In overall terms, both series suggest that in absolute and relative terms Indigenous socioeconomic outcomes have improved at the national level for most variables.46

These findings are very much at odds with the dominant discourse of policy failure in Indigenous over the past 35 years. For example, there has been an unambiguous improvement (in both absolute and relative terms) for both the long-run and short-run series in the areas of household size and income, home ownership and all the education outcomes. Table 1 indicates that improvements were not always consistent for the short-run and long-run series for two crucial areas: labour force status and health. As discussed above the labour force trends are affected by the growth of the CDEP scheme, but they are also affected by the large-scale structural adjustments in Australia which followed the reductions in tariffs and other trade barriers that began in 1974 and which have continued with the continuing internationalisation of the economy. Indigenous workers were and are disproportionately employed in the manufacturing sector that was severely affected by such structural adjustment.47 It is also probable that there was some residual structural adjustment from the decisions in the equal pay cases in pastoral awards that occurred during the mid-1960s.48

There is also a certain lack of concordance between the absolute and relative changes in the short and long run for certain health indicators. Recall that the changes to the method of calculating life expectancy means that the short run series is based on a brief 5-year period between 1996 and 2001. Given that part of the reason for changing the methodology was dissatisfaction with the assumptions that underpin the calculations, it is advisable to be particularly cautious about these estimates. However, even if one focuses solely on the health proxy of the incidence of elderly population there is some inconsistency in the short-run and long-run series. While there were absolute improvements in both series, the long-run trends in relative outcomes are actually negative. The demographic bulge of the non-Indigenous baby boomers is working its way through Australian society over this period, and in this context the relative improvement in the number of older Indigenous people is particularly noteworthy.

Table 1 provides a summary of the broad trends since 1971 (and 1996) to assist the reader to distinguish between absolute and relative changes or trends. There was obviously an even greater variation in trends for the various inter-censal periods, but it is beyond the scope of this paper to elaborate on all noteworthy variations in trends. Interested readers are referred to Altman and Hunter and invited to draw their own inferences from the data provided in the Appendix.49

CONVERGENCE AND THE PROSPECTS FOR CLOSING THE GAP

As indicated above, some of the data in Appendix A were not collected in 1971 (e.g. income), while other data were either unavailable publicly or were not available in a consistent form for all censuses (e.g. private sector employment or home ownership data). Notwithstanding such difficulties it is possible to estimate trends for many of the data sets reported in Altman, Biddle, and Hunter.50 In order to estimate the prospect for closing the gap we have estimated the number of years before a gap could be closed using trends identified in Appendix A over two time periods, 1971–2006 and 1996–2006.51 That is, using a simple (linear) projection of observed changes in relative outcomes, we estimate when Indigenous indicators will be the same as those for other Australians.

Where the time series is not complete we use the trends based on the longest available period. Convergence is estimated as the number of years required to eliminate the gap between Indigenous and non-Indigenous Australians (where the gap is measured in relative terms) following current trends. In order to account for the variation in the longest period for which consistent data are available, the trend estimates are calculated as the improvement in relative outcomes per year. This trend (or gradient) is projected forward to identify when the ‘gap’ between Indigenous and non-Indigenous indicators will be removed.

From an analytical perspective it is always risky to make excessively precise claims about long-range forecasts. While our analysis here does not present a formal time series analysis, readers should bear in mind that forecast errors always increase as the lead time becomes longer.52 Most forecasts become unreliable quite quickly and are rarely reliable beyond a decade. The unreliability of our estimates of convergence is underscored by the fact that there is a divergence between Indigenous and non-Indigenous outcomes for several of the variables examined. Accordingly, our estimates must be interpreted as the possibility of closing the gap, if current trends continue. Our estimates are not a prediction of what will happen, but rather a description of what may happen if current trends continue. Policy shifts, unpredicted demographic changes and economic circumstances (including the current international financial crisis) are all likely to impact on what will actually occur into the future.

Not all trends in Indigenous and non-Indigenous outcomes converge. Indeed, divergence in outcomes can even occur if Indigenous outcomes are improving in absolute terms, but such increases are not as large as those evident for other Australians. The main divergence in outcomes occurs in labour market data where Indigenous outcomes are improving slowly than the rate for other Australians. However, employment prospects is complicated by the effect of CDEP and the disproportionate location of Indigenous people in regional and remote areas where labour market conditions are relatively depressed (aside from mining activity).

For those outcomes that are converging, Table 2 reports when the expected number of years until convergence might occur if policy settings do not change significantly. If the number of years to convergence is greater than 100 years, the entry is shown as 100+. This was done in acknowledgment of the fact that any forecast over this time frame is highly uncertain and hence unreliable. Having regard to the commitment to close the gap within one generation, the truncation of the convergence estimates at 100 years may be viewed as indicating whether the gap might be closed in just over three generations.53

Table 2. Number of years till convergence of Indigenous and non-Indigenous outcomes
 Convergence based on long run trends since 1971Convergence based on post-1996 trends
  1. †Indigenous and non-Indigenous outcomes are diverging. The trends are based on the maximum period for which comparable data were available. For example, the long run convergence for income calculated from 1981 as there were no available estimates for 1971. If the number of years to convergence is greater than 100 years, then the table entry is shown as 100+.

Unemployment rate (% labour force)28
Employment to population ratio (% adults)
Private-sector employment (% adults)23
Labour force participation rate (% adults)100+
Median weekly personal income $A (2006)100+
Household size100+100+
Median weekly household income $A (2006)94100+
Home owner or purchasing (% population)100+100+
Never attended school (% adults)214
Post-school qualification (% adults)4425
Degree or higher (% adults)NA100+
Attending educational institution (% 15–24-year-olds)NA63
Male life expectancy at birth (years)100+
Female life expectancy at birth (years)47
Population aged over 55 years (%)100+

The long-run trends in unemployment rates indicate that Indigenous and non-Indigenous outcomes will converge in 28 years around the year 2035. This is a potentially excessively optimistic scenario because it may be associated with the rise of the CDEP scheme. Indeed, the more recent inter-censal trends show that Indigenous and non-Indigenous unemployment rates are diverging – an observation that might be associated with the relatively stagnant number of CDEP participants (at least in net terms between 1996 and 2006).

The employment to population ratios are diverging for both trend estimates. Convergence in employment and unemployment are obviously dependent on where Indigenous live, the structure of incentives facing individuals and organisations (who are making decisions on where to locate), as well as the state of national and regional economies. With respect to the latter, with the global economic crisis and end to the long national economic boom, then low qualified and short-term workers are likely to be the first shed by firms whose profit margins are being squeezed.

Indigenous private sector employment was been particularly depressed until the late 1980s. However, private sector jobs increased substantially for Indigenous people after 1996. This may reflect the tendency to privatise low skilled jobs in the public sector, but it has been achieved despite the depressed labour market conditions in regional Australia. While this recent improvement of Indigenous participation in the private sector is noteworthy, the estimated year to convergence of 23 years is likely to be excessively optimistic as it assumes sustained economic growth at ‘pre-crisis’ levels.

Relative labour force participation rates improved for Indigenous Australians after 1971, but recent trends show that Indigenous people have been rather less successful in this crucial dimension of economic engagement since 1996. The most optimistic scenario for the rate of long-run improvement in relative outcomes shows that the gap in labour force participations will not be eliminated within the next century.

The convergence of individual income is only evident in the long-run trends. Even then convergence tends to be rather slow, with an estimated closing of the gap projected to take at least another 100 years. The lack of income convergence evident from recent trends may be a reflection of growing national income inequality where low-income groups are not faring as well as the rest.54

The decline in Indigenous household sizes vis-à-vis other Australians offers a rather more positive story. On current trends it will be at least three generations before this gap is closed. However, this is one gap that we would be reasonably confident can be closed because the demographic and economic factors will tend to be mutually reinforcing. High educational outcomes and income tends to be associated with smaller families. For example, human capital models make a firm prediction of this result, largely due to the increasing opportunity cost of female time making work more attractive relative to home production activities.55

Median household incomes appear to be converging albeit rather slowly. The long run trends indicate that the gap will be closed in 94 years. In contrast, recent trends seem to indicate that convergence will not occur for at least another century. While the truth is likely to lie somewhere in between, the recent trends illustrate the main problem with a paradigm based on closing the gap – namely that the goal posts continually move. Indigenous household income has improved, but not as fast as that of non-Indigenous Australians. Note too that raw household income will converge faster than predicted in Table 2 because the size of Indigenous households fell by more than non-Indigenous households thus reducing the number of people who could assist to close the gap in household income.

The predicted convergence in home ownership is estimated to take place in around 100 years.56 It is probable that the rate of convergence will be closely linked to the ability to improve Indigenous household incomes. The other relevant factors are the ability to resolve current transactions and administrative costs that seem to impede the individuation of leases on Aboriginal-owned land. Unless such tenure issues are resolved, institutional barriers might result in the home ownership gap being intractable.

The most optimistic prediction is for convergence in educational outcomes. The gap in never having attended school is predicted to close in between two and 14 years depending on which trend is used. In recent years, almost all Indigenous children attend school at some stage. The main concerns are now about retention rates and the quality of schooling and outcomes for those who are attending.

The predicted convergence in post-school qualification is also quite optimistic for both projected trends; the gap may be closed within 25 and 44 years. One reason for this relatively positive prediction is that such outcomes are amenable to fast policy action. For example, the mutual obligation regimes, which have become increasingly popular with government as a means of providing conditional welfare payments, reward recipients who return to their studies. However, it may become more difficult to improve outcomes as the quality of the initial educational attainment becomes more central to performance. That is, it may be possible to improve vocational training through Technical and Further Education Institutes in the short-run, but the ability to enhance the participation in universities and other tertiary institutes may be limited by the quality of initial schooling.

Although we only have short-run trends for attendance of youth (aged 15–24 years) at educational institutions, convergence in rates of attendance would not occur for another 63 years. If we exclude secondary students to focus on people studying at tertiary educational institutions, convergence would not take place for over 100 years.57 The difficulty in achieving convergence for the more demanding tertiary qualifications is confirmed by the gap in the incidence of degrees amongst Indigenous and non-Indigenous people is not predicted to close for at least another century.58

As indicated above, the ABS recently changed the method of predicting life expectancy for Indigenous people to take into account changing Indigenous identification.59 Whatever its merits, the resulting figures are not strictly comparable with previous estimates and cannot be used to calculate long-run trends. Hence, we use the older unadjusted series to estimate trends in male and female life expectancy. On historical trends, and in the absence of significant and successful policy innovations, it is difficult to see the gap in male life expectancy being removed for at least another 100 years. Current trends are more optimistic in terms of closing the gap in female life expectancy with the estimated convergence taking 47 years. The more recent trends, however, seem to indicate that this is exceptionally optimistic as the relative gap increased between 1996 and 2001.

Our earlier analysis used the demographic proxy for health of the proportion of population aged 55 and over. This proxy is easy to calculate, though rather more difficult to interpret. Nevertheless, the recent trend in this proxy indicate the estimated time to convergence of female life expectancy is too optimistic and the resources required to close the gap in life expectancies within one generation will not be trivial.

CONCLUSION

For many years, governments in Australia have espoused a commitment to improving Indigenous socioeconomic outcomes relative to the non-Indigenous population using terms such as ‘statistical equality’, ‘practical reconciliation’, and ‘closing the gap.’ The aim of this paper is to examine national level data from the last eight censuses to assess the historical progress in achieving these aims and to make predictions about when convergence between Indigenous and non-Indigenous Australians might be achieved (if at all) were the historical trends we identify to continue.

There has been steady improvement in a number of the socioeconomic outcomes considered. Very few, however, were predicted to converge relative to the non-Indigenous population in the short to medium term. It is clear, therefore, that policies designed to close the gap between Indigenous and non-Indigenous populations are intrinsically difficult to operationalise. This is likely to be due in part to many of the underlying causes being inter-generational in nature. For example, Marmot and Wilkinson and many others have identified that life expectancy depends on what happens in early childhood when crucial decisions are made by previous generations.60 Health outcomes are also likely to be dependent on social, cultural, and community contexts, and hence policies that address these gaps must have a long-term focus and tackle problems at several levels.61

Official statistics based on mainstream social norms arguably do not capture the extent of Indigenous alienation from mainstream Australia. There are many different and interrelated dimensions of deprivation, social exclusion, and poverty facing Indigenous Australians.62 Policies designed to redress the gap between Indigenous and non-Indigenous socioeconomic outcomes need to account for the inter-dependence between extant disadvantages and be informed by evidence on how disadvantage evolves over the ‘life-cycles’ of individuals and communities.63

Ultimately, our analysis must be supplemented with qualitative and behavioural analyses that illustrate the complex developmental challenges facing Indigenous Australians. Forecasts can become redundant quickly. For example, the short-run forecasts of Indigenous labour force status in earlier studies were generally accurate, but the medium term projections (10 years out) failed to anticipate the prolonged economic boom.64 Even if behaviourally informed analyses were conducted, they may demonstrate that process/rights-based issues and symbolism can motivate and de-motivate people and hence lead to relatively better or poorer outcomes than we predict. In that case, we need to expand our conception of disadvantage to include both practical and symbolic considerations rather than view these two elements as constituting some sort of trade-off. In our opinion, there is no evidence of direct trade-off between practical and symbolic issues and such distinctions represent a false dichotomy.

A significant issue emerging in the literature is the difficulty to establish reliable and robust long-run trends. The caveats documented in this paper illustrate why one has to be cautious about making strong predictions about when the gap between Indigenous and other Australians might be closed. The evidence presented here indicates that it is probable that the long-run trends in Indigenous socioeconomic status are positive, and hence, the recent dominant discourse of policy failure in Indigenous affairs is wrong or over-stated.

Demographic transitions and increased identification is the imponderable factor here. The change in official methodology of calculating life expectancy is an implicit recognition of the importance of such changes. Even if one is willing to suspend disbelief in the short-run it is more difficult to do this in the long run. The increased propensity to identify as Indigenous and the substantial rates of intermarriage will mean that people who would historically be characterised as non-Indigenous will be classified as Indigenous. If such people have socioeconomic and demographic characteristics that are closer to the non-Indigenous profile, then there is a built in tendency towards convergence.

Obviously there is likely to be substantial variation in regional outcomes between sub-national populations, but in broad terms there is less difference in the regional trends than one might expect.65 While the magnitude of the original ‘gap’ can be rather different for some outcomes in remote and non-remote areas, especially for the economic indicators, the overall trends are remarkably similar. However, this may be because the standard remoteness classification hides more variation than it explains. Recent analysis at a much smaller regional level shows variation in trends in Indigenous outcomes over the most recent inter-censal period.66 For example, full-time private sector employment increased substantially for the Indigenous population relative to the non-Indigenous population in Brisbane and Perth, but declined in Sydney. For some types of analysis, therefore, it is important to concentrate on the region that delimits labour markets or other outcome measures.67 Whatever the trends, the underlying causes of the gap are likely to be different in remote and non-remote Australia, and hence a regionally differentiated policy response is warranted.

By examining trends along non-Indigenous and Indigenous categories, it is clear that the ability of governments to ‘close the gap’ depends on what is happening in the rest of the community. There are likely to be different and potentially uneven impacts on Indigenous outcomes from the lower macroeconomic growth of the current financial crisis. On the one hand, if unskilled workers are the first to lose their jobs then this is likely to impact disproportionately on Indigenous employment outcomes and associated socioeconomic status. On the other hand, if job or income losses were confined to the financial sector, then there is likely to be a minimal impact on the relative standing of Indigenous Australians. Another consideration is that global warming, and the policy response to curb greenhouse gas emissions, may place additional constraints on economic growth. Any recession, whatever the source of the downturn, is likely to curtail a government's capacity to directly achieve distributional goals such as ‘closing the gap’ between Indigenous and other Australians.

As we enter a new political cycle in Indigenous affairs policy making it is important that some of the historical evidence presented here is not overlooked, especially during a period when the new discourse of policy-making emphasises evidence rather than anecdote or ideology. We concur that the socioeconomic gap between Indigenous and other Australians remains unacceptably high. While our analysis suggests that at the national level there has been improvement in the three and a half decades to 2006, we also accept that the current rate of improvement is too slow. Nevertheless, using recent economic history as our guide, we predict that it will take many years, possibly many generations, before the respective gaps are closed.

To conclude, we make two observations. First, the long lead times that we have estimated to close the gap suggest that some fundamentally new approach might be needed to ensure structural change of such a degree as to significantly alter the time frame we predict. One possibility is to require a fundamental reallocation of property rights in resources that fully acknowledge the original ownership of the country.68 Another alternative is that a massive increase in investment in Indigenous infrastructure and Indigenous people may eventually affect the persistent gaps. Second, we propose that a degree of policy realism and caution is required in new policy commitments. The only way to measure whether convergence in outcomes actually occurs will be with the sorts of statistics that we have accessed here, inadequate as they might be from a public policy or Indigenous perspective. These statistics will form the basis for assessing outcomes in closing the gaps; this is the statistical evidence base that will constitute political accountability. Under such circumstances we would counsel that commitments to ‘reducing disparities’ might be a more realistic policy goal for the Rudd Government than ‘closing the gap’.

Footnotes

  • 1

    Te Puni Kokiri, Gaps Between Maori and Non-Maori.

  • 2

    Altman, Aboriginal Employment Equity.

  • 3

    Altman, Practical reconciliation and new mainstreaming.

  • 4

    Although the Productivity Commission was commissioned to provide biennial reports on progress in addressing Indigenous disadvantage (Steering Committee for the Review of Government Service Provision, Overcoming Indigenous Disadvantage).

  • 5

    Altman and Hinkson, Coercive Reconciliation. Prescribed communities are those identified by the legislation enacting the Northern Territory Emergency Response.

  • 6

    This figure is based on the ABS/Australian Institute of Health and Welfare publication The Health and Welfare of Australia's Aboriginal and Torres Strait Islander Peoples. Another figure is provided in Vos et al., Burden of Disease and Injury. They estimate that the difference between Indigenous and non-Indigenous life expectancy is around 13 years. A recent ABS publication estimates the difference in life expectancy as being 11.8 years and 10.0 years for males and females respectively (ABS, Methods for Developing Life Tables). Clearly, the gap in life expectancy is very sensitive to the method used.

  • 7

    See Oxfam press release, ‘Oxfam welcomes decision by Prime Minister Rudd to sign nine point plan to end Aboriginal health crisis’, 20 March 2008, [Accessed 4 May 2009] Available from URL http://www.oxfam.org.au/media.

  • 8

    Bronfenbrenner, Ecology of Human Development.

  • 9

    Altman et al., Indigenous socioeconomic outcomes.

  • 10

    Altman and Nieuwenhuysen, Economic Status of Australian Aborigines.

  • 11

    See Rowse, Indigenous Futures, and Altman and Rowse, Indigenous affairs.

  • 12

    Taylor and Altman, Job Ahead; Taylor and Hunter, Job Still Ahead; and Hunter et al., Future of Indigenous work.

  • 13

    See Altman and Sanders, From exclusion to dependence; Gray and Tesfaghiorghis, Social indicators of aboriginal population; Sanders, Destined to fail. These earlier analysis were not based on forecasts, but rather argued on a priori grounds.

  • 14

    Altman et al., Indigenous socioeconomic outcomes.

  • 15

    Hunter, Validity of intercensal comparisons.

  • 16

    Altman et al., Indigenous socioeconomic outcomes.

  • 17

    Altman and Gray, Effects of CDEP scheme on economic status.

  • 18

    Altman et al., Indigenous socioeconomic outcomes.

  • 19

    Ross, Population Issues.

  • 20

    Biddle, Indigenous housing need; Biddle et al., Indigenous participation in labour markets.

  • 21

    Altman et al., Indigenous socioeconomic outcomes.

  • 22

    Altman and Nieuwenhuysen, Economic Status of Australian Aborigines.

  • 23

    A possible exception is the data on private sector employment, which is provided in a slightly different format in 2001 compared to either 1996 or 2006. This qualification, however, does not affect the validity of the trend estimated between 1996 and 2006.

  • 24

    One possible exception is the data on private sector employment, which is in a slightly different format in 2001 compared to either 1996 or 2006. However, that does not affect the validity of the trend estimated between 1996 and 2006.

  • 25

    There are minor differences between the pre-2006 estimates in the Appendix and those in Altman et al., Indigenous socioeconomic outcomes. These are rounding errors and are indicated in the text.

  • 26

    See Sanders, CDEP Scheme in Australian Social Policy. The CDEP scheme began as an innovative Indigenous-specific program that converted the notional equivalents of the unemployment benefit entitlements of Aboriginal people in remote areas within the social security system into grants to Aboriginal organisations from the Commonwealth Department of Aboriginal Affairs. These grants were used by Indigenous organisations to employ nominal unemployment benefit recipients in part-time work. The CDEP scheme was a response to the perceived social threat of ‘sit-down’ money to Indigenous communities in the 1970s. Ironically, the scheme is now criticised as one of the factors driving the social effects of prolonged welfare dependence.

  • 27

    The 2006 CDEP participants are estimated from the Department of Employment and Workplace Relations (DEWR) annual report while the population estimate is derived from 2006 Census Estimated Residential Population.

  • 28

    Fisk, Aboriginal Economy; Gray and Hunter, Analysis of labour market outcomes; and Hunter and Gray, ‘Structure of indigenous employment’.

  • 29

    Hunter, Indigenous policy based on evidence and not hyperbole. It is unclear on what basis one would make such a choice as most CDEP participants work for funding provided by the Federal Government. Such a classification would be akin to categorising public sector employment as a form of welfare.

  • 30

    Altman and Nieuwenhuysen, Economic Status of Australian Aborigines.

  • 31

    Quiggin, Great Expectations.

  • 32

    Hunter, Trends in neighbourhood inequality; and Leigh, Long-run inequality series from tax data.

  • 33

    Household sizes are larger in remote areas compared to non-remote areas, but recent declines in Indigenous household size is largest in absolute terms in remote areas; Hunter, Indigenous policy based on evidence and not hyperbole.

  • 34

    For example, Hunter, Indigenous and other Australia poverty; estimate ‘equivalised’ income by dividing household income by household size. There are more sophisticated measures that control for the composition of household, but discussion is beyond the scope of this paper.

  • 35

    Kinfu and Taylor, Components of indigenous population change.

  • 36

    Where household welfare is measured using adjusted (equivalised) income.

  • 37

    The last 10 years have seen a substantial increase in transfers to families, especially families with children (e.g. Family Tax Benefits). However, given that the census measures pre-tax income, this explanation is only valid if the transfers occur outside the tax system (not as tax rebates) and hence could be construed by respondents to the respective censuses as being part of their gross income.

  • 38

    Harding et al., Trends in Spatial Inequality.

  • 39

    Biddle, Does it pay to go to school?

  • 40

    Commonwealth of Australia, National Report.

  • 41

    See the Bhat method in ABS, Altman, Practical reconciliation and new mainstreaming.

  • 42

    Altman et al., Indigenous socioeconomic outcomes; Ross and Taylor, Comparison of Indigenous Australians and New Zealand Maori.

  • 43

    Most of the life expectancy estimates in Appendix A were rounded in Altman et al., Indigenous socioeconomic outcomes. The Indigenous estimates for 1971 were taken from Smith, Aboriginal population of Australia, p. 279. The small difference in male and female life expectancy was probably due to the high rates of Indigenous deaths in child birth.

  • 44

    Altman et al., Indigenous socioeconomic outcomes.

  • 45

    Note, however, that there was a substantial divergence in 2006 between the national census count (about 450 000) and the comparable Estimated Resident Population for Indigenous Australians (512 000). There are even larger divergences at the sub-national level. For example, the Post Enumeration Survey estimated an undercount of 24 per cent in the Kimberley region of Western Australia and 19 per cent in the Northern Territory.

  • 46

    Two possible exceptions to this observation are for the labour force status and health. As indicated above, the estimated trends are complicated by the interactions of labour force indicators with the macro-economy and measurement error for health proxies.

  • 47

    Hunter, Indigenous Australians in Contemporary Labour Market.

  • 48

    Kerr, Northern Territory Cattle Station Industry Award case and O'Shea case; and Rowse, White Flour, White Power.

  • 49

    Altman and Hunter, Evaluating Indigenous socioeconomic outcomes, used the unprecedented alignment of the electoral and statistical cycles between 1991 and 2001 to explore the statistical success of the practical reconciliation espoused by John Howard. The outcomes were politically contentious as they indicated that the earlier Labor Government had outperformed the Coalition in ‘closing the gaps’. This had implications for the broader debate as to whether practical reconciliation yielded superior outcomes to symbolic reconciliation. Subsequent analyses confirmed this finding (see Altman, Practical reconciliation and new mainstreaming). The unusual alignment between electoral and statistical cycles is now less apparent (the last Federal election was in 2007, not 2006).
    The period 1991–96 saw the relative and absolute improvements in Indigenous socioeconomic status (see Appendix A). The first five years of Howard Governments (1996–2001) saw some absolute improvements but fewer relative improvements, primarily due improvement for non-Indigenous Australians. Despite macroeconomic growth, the gaps in relative well-being were relatively intractable. During the second five years of Howard governments (2001–06), there were more relative and absolute improvements than between 1991–96 or 1996–2001. Noteworthy was the absolute improvement in most variables 2001–06. This suggests that the Indigenous population benefited from the overall expansion of the national economy. In relative terms, however, improvement occurred in only seven of 12 variables, suggesting that convergence was made difficult by the higher relative prosperity of the non-Indigenous population.
    The latest improvement in major socioeconomic indicators might be explained by the ‘new mainstreaming’ in Indigenous policy introduced in 2004 having positive effects by the 2006 Census. This is hard to credit because the two-year lead time is short, but this is a possibility. On the other hand, as economic growth continued during the last term of the Howard Government some benefits may have eventually trickled down to Indigenous Australians.

  • 50

    Altman et al., Indigenous socioeconomic outcomes.

  • 51

    The choice of base year impacts on the prospects and rate of convergence. For example, while the broad conclusions remain the same whether 1981 is used as the base year rather than 1971, doing so results in convergence rather than divergence in employment, full-time employment and private sector employment. However, this convergence only occurred after a long time (71 years, 50 years, and 44 years, respectively).

  • 52

    Armstrong, Principles of forecasting.

  • 53

    A generation has traditionally been defined as ‘the average interval of time between the birth of parents and the birth of their offspring’. Before the baby boomer ‘generation’, the length of a generation was about 20 years, but this figure has crept up (even in the Indigenous population) and it is now closer to 30 years.

  • 54

    Especially the top 25 per cent of the distribution. See Leigh, Long-run inequality series from tax data.

  • 55

    Becker et al., Human capital, fertility, and economic growth.

  • 56

    Focusing on the proportion of Indigenous households who own their own home, the rate of convergence in home ownership is slightly more optimistic. Indigenous outcomes will be the same as non-Indigenous outcomes in about 80 years. This estimate is closer to the short-run estimate in Table 2 because that indicator is only available consistently since 1991.

  • 57

    Only the long run trends have been calculated for this outcome between 1981 and 2001 because of data comparability issues.

  • 58

    This is also consistent with the post-1996 trends, that there will not be a convergence in year 10 completion for another 17 years. By contrast, the convergence in year 12 completion will take much longer and will not occur for another 96 years.

  • 59

    See the Bhat method in ABS, Calculating Experimental Life Tables, and ABS/Australian Institute of Health and Welfare, Health and Welfare, p. 148.

  • 60

    Marmot and Wilkinson, Social Determinants of Health.

  • 61

    See Bronfenbrenner, Ecology of Human Development.

  • 62

    See Daly and Smith, Indigenous Australian Children; Hunter, Three Nations: Not One; and Hunter, Measuring social costs of unemployment.

  • 63

    Hunter, Overcoming indigenous disadvantage.

  • 64

    Hunter and Taylor, Indigenous employment forecasts.

  • 65

    Hunter, Indigenous policy based on evidence and not hyperbole.

  • 66

    Biddle et al., Indigenous participation in labour markets.

  • 67

    Taylor and Biddle, Locations of Indigenous Population Change.

  • 68

    A philosophical justification for this proposition can be based on Nozick, Anarchy, State, and Utopia, who mounts a case against addressing equity through redistribution. He argues that policy makers should establish a clear set of property rights based on original possession.

Appendix

Table A1.  Socioeconomic outcomes for Indigenous Australians, 1971–2006
Variable197119811991199620012006
  1. Note: ‘n.a.’ means that the data were not available in that year.

  2. Source: ABS Census of Population and Housing 1971, 1981, 1991, 1996, 2001 and 2006.

Unemployment rate (% labour force)9.024.630.822.720.015.6
Employment to population ratio (% adults)42.035.737.140.741.743.2
Private-sector employment (% adults)29.717.220.521.6n.a.32.8
Labour force participation rate (% adults)46.147.353.552.752.151.2
Median weekly personal income ($A 2006)n.a.217.5244.8245.6246.7267.4
Household size4.64.14.03.63.43.4
Median weekly household income ($A 2006)n.a.784.9757.7n.a.913.21,072.7
Home owner or purchasing (% population)26.119.719.126.126.830.0
Never attended school (% adults)22.710.75.13.13.22.7
Post-school qualification (% adults)3.25.09.512.818.223.8
Degree or higher (% adults)n.a.n.a.n.a.2.23.34.4
Attending educational institution (% 15–24 year olds)n.a.n.a.n.a.27.733.434.4
Male life expectancy at birth (years)49.656.057.057.056.0n.a.
Female life expectancy at birth (years)50.064.064.064.063.0n.a.
Population aged over 55 years (%)7.36.46.26.36.78.2
Table A2.  Socioeconomic outcomes for other Australians, 1971–2006
Variable197119811991199620012006
  1. Note: ‘n.a.’ means that the data were not available in that year.

  2. Source: ABS Census of Population and Housing 1971, 1981, 1991, 1996, 2001 and 2006.

Unemployment rate (% labour force)1.65.811.49.07.25.1
Employment to population ratio (% adults)57.858.256.356.458.960.5
Private-sector employment (% adults)45.641.040.646.3n.a.52.1
Labour force participation rate (% adults)58.861.863.662.063.463.7
Median weekly personal income ($A 2006)n.a.395.6396.8382.5440.5465.2
Household size3.43.12.92.72.62.6
Median weekly household income ($A 2006)n.a.1,087.9990.4n.a.1,171.61,375.3
Home owner or purchasing (% population)70.573.470.272.572.973.4
Never attended school (% adults)0.60.71.00.71.00.9
Post-school qualification (% adults)23.727.732.339.141.645.9
Degree or higher (% adults)n.a.n.a.n.a.11.214.818.3
Attending educational institution (% 15–24 yearS olds)n.a.n.a.n.a.49.054.855.2
Male life expectancy at birth (years)68.071.074.075.076.0n.a.
Female life expectancy at birth (years)75.078.080.081.082.0n.a.
Population aged over 55 years (%)17.118.619.620.422.024.7
Table A3.  Ratio of Indigenous to non-Indigenous socioeconomic outcomes, 1971–2006
Variable197119811991199620012006
  1. Note: ‘n.a.’ means that the data were not available in that year. Results have been rounded to two decimal places.

  2. Source: ABS Census of Population and Housing 1971, 1981, 1991, 1996, 2001 and 2006.

Unemployment rate (% labour force)5.634.242.702.522.783.06
Employment to population ratio (% adults)0.730.610.660.720.710.71
Private-sector employment (% adults)0.650.420.510.47n.a.0.63
Labour force participation rate (% adults)0.780.770.840.850.820.80
Median weekly personal income ($A 2006)n.a.0.550.620.640.560.58
Household size1.351.321.381.331.311.31
Median weekly household income ($A 2006)n.a.0.720.77n.a.0.780.78
Home owner or purchasing (% population)0.370.270.270.360.370.41
Never attended school (% adults)37.8315.295.104.433.203.00
Post-school qualification (% adults)0.140.180.290.330.440.52
Degree or higher (% adults)n.a.n.a.n.a.0.200.230.24
Attending educational institution (% 15–24 year olds)n.a.n.a.n.a.0.560.610.62
Male life expectancy at birth (years)0.730.790.770.760.74n.a.
Female life expectancy at birth (years)0.670.820.800.790.77n.a.
Population aged over 55 years (%)0.430.340.320.310.310.33

Ancillary