Social inclusion, exclusion, and well-being in Australia: meaning and measurement.

Saunders, Peter

Introduction

Throughout the period in office of the ALP Government, much of Australian social policy was shaped by the government's social inclusion agenda (SIA). Drawing on the actions and experience of successive labour governments in the UK, administrative structures were established to implement an agenda which was wide ranging in its aims and had the potential to impact on how policy was designed and implemented and, ultimately, experienced. But throughout the period, concern was expressed about how well this potential was being realised in practice, with some arguing that the agenda achieved little, acting as a smokescreen for inactivity in areas like poverty relief and redistribution that should have been the focus of the government's social policy agenda.

It is too soon to provide a comprehensive assessment of the impact of the SIA, and this paper does not seek to undertake that important task. Instead, it reviews the broad framework established by the government to monitor its direction and impact and contrasts it with those established by two of Australia's leading independent social research institutes. It also asks what difference such an agenda could make as a way of better highlighting its actual impact. These issues overlap with another that is central to the concepts of social exclusion and inclusion that has attracted considerable attention in the academic literature --the idea that social problems are multi-dimensional and interconnected.

This feature raises new challenges for policy design and delivery, as well as for measurement. Is it meaningful, for example, to claim that Australian society has become 'more inclusive' and, if so, what kinds of data and measures are needed to demonstrate this?

The demise of the ALP government in late 2013 has almost certainly seen the abandonment of social inclusion as a policy priority. Even so, addressing the nature, role and impact of the SIA is of more than just historical interest. This is because the discussion raises issues that are of ongoing concern to social policy analysis, including those that relate to important concepts (the relationship between exclusion and poverty), to policy design and delivery (and the need to resolve long-standing federal-state rivalries), to impact assessment (and the role of indicators), and to data coverage and availability (and the need for better reporting of social outcomes).

The discussion proceeds as follows: the ensuing section considers the relationships between social exclusion and poverty and between social exclusion and social inclusion--in brief, because these topics have already generated much discussion in the literature. This is followed by a discussion of the role of indicators and a review of the social inclusion indicators framework developed by the government's Social Inclusion Unit. That framework is then compared with those developed by the Melbourne Institute in conjunction with the Brotherhood of St Laurence (MI/BSL) and the Social Policy Research Centre (SPRC) as a way of demonstrating how broad conceptual similarities can dissipate when the details are compared. Some results that draw on work conducted at the SPRC on recent changes in social exclusion and its association with subjective well-being are then presented as a way of helping to establish the impact of exclusion on those who experience it. The final section draws together the paper's main conclusions.

Exclusion and poverty

Social exclusion has been presented in some contexts as an alternative to poverty; however, the two concepts are best seen as distinct but related. While much has been written about the ambiguities surrounding the concept of exclusion, the debate over poverty has been dominated by measurement issues. Despite these ongoing controversies, a consensus has emerged among poverty researchers over the use and value of a poverty line set at one half of median income (see OECD 2008). In contrast, the debate over the definition of social exclusion continues unabated (Saraceno 2002; Beland 2007; Hick 2012) and has not been able to resolve points of disagreement.

The relationship between poverty and exclusion was recognised in the Australian context over three decades ago when the Poverty Commission noted on its opening page that:

Poverty is not just a personal attribute; it arises out of the organisation of society. Australian society has failed to adapt ... with the consequence that many are ... cut off in many ways from participation in community groups and urban society (Commission of Inquiry into Poverty 1975: 1; italics added).

Shortly after, Townsend proposed a definition of poverty that has since been widely embraced as capturing its main elements, one of which is exclusion:

Individuals, families and groups in the population can be said to be in poverty when they lack the resources to obtain the types of diet, participate in the activities and have the living conditions and amenities which are customary, or at least widely encouraged or approved, in the societies to which they belong (Townsend 1979: 31; italics added).

The rise to prominence of social exclusion, however, reflects the idea that it is more than just a marker of poverty. Its significance reflects the view that it is something more than just one of many consequences of not having enough income to meet basic needs. Without going into the details of the many contributions to this idea, the three main features of social exclusion have been identified by Atkinson (1998) as relativity, agency, and dynamics. (1) The latter two of these features put the focus on understanding the nature and impact of the processes that give rise to different forms of exclusion, including those that may be self-imposed, and how those processes play out over time. (2)

There is, of course, a need to distinguish between what social exclusion actually is and how it emerges or evolves, although both are important if the concept is to be properly understood. In relation to the former dimension, researchers at the University of Bristol (where Townsend was by then based) produced the following 'composite working definition' after reviewing the many options available:

Social exclusion is a complex and multi-dimensional process. It involves the lack or denial of resources, rights, goods and services, and the inability to participate in the normal relationships and activities, available to the majority of people in society, whether in economic, social, cultural or political arenas (Levitas et al. 2007: 9).

The first part of this definition again highlights the close link between exclusion and poverty, although here the careful use of the word 'involves' implies a neutral stance on the nature of this relationship.

Speaking at the Australian Social Policy Conference a decade ago, Bradshaw (2004) voiced the concerns many in the UK felt at the time about the Blair Government's embrace of social exclusion. He has noted that when it first emerged, social exclusion seemed to add little to poverty and in some guises carried a great deal of behaviourist ideological baggage or blamed the poor (Bradshaw 2004: 182).

He went on to observe, however, that subsequent developments allayed some of these fears, including: the renewed focus given to poverty in the UK exclusion debate; the embrace of exclusion as a central plank of EU social policy that committed member states to draw up national action plans on poverty and exclusion; and work by the UK Social Exclusion Unit and by researchers generally that opened up new avenues of study into issues such as the impact of mental illness, and the relationship between exclusion and poverty.

Although it cannot be claimed that the Australian SIA placed poverty alleviation at the forefront of social policy, it did make it harder for policy makers to ignore the issue because of its close association with exclusion--as both cause and effect. However, this did not always lead to a positive policy response, as the debate over the adequacy of Newstart Allowance (NSA) illustrates. The decision to index NSA to consumer prices, implemented in the early-1990s, led to a gradual erosion in its value relative to average earnings and hence to median incomes and thus to the poverty line. (3) Calls from the welfare sector (ACOSS 2013) for an increase in payments were echoed by Australian researchers (e.g., Whiteford 2010) and--in an unusually frank assessment--by the OECD (2010). These arguments had no impact on the government, with Ministers arguing that providing the unemployed with employment opportunities was the preferred option, basing this assessment on the implicit view that increased benefits would reduce the incentive to seek work even though (as Gregory 2013 has recently noted) there is little or no evidence to support such a claim. The general point, however, was that when it came to conflict between adequacy and incentives, or between poverty alleviation and inclusion promotion, the rhetoric and action favoured the latter in both instances.

Having said this, the government's social inclusion indicator framework (see below) did--for the first time since Bob Hawke's infamous child poverty pledge of the late-1980s--give some prominence to several features of poverty, including its incidence, persistence, and overlap with financial stress. However, although poverty rates were included in the initial report produced for the government by the Australian Social Inclusion Unit (Australian Government 2009b: Table 5), the word poverty--or poverty line--was not used in the second report, reference being made instead to 'households with low economic resources' or 'low income households', which were identified as those in the lower deciles of the distributions of income and wealth (see Australian Government 2012: 28-35). (4)

It is important to note the contrast between the conceptual literature discussed above, where the focus has been on unpicking the concept of social exclusion and the policy debate where--at least in Australia--the emphasis has been on social inclusion. This raises the issue of whether and how the two concepts are related or whether they are simply opposites. When South Australia introduced its social inclusion agenda in 2002, Premier Rann favoured the term inclusion 'to reflect the fact that "inclusion" is what we are about' (quoted in Buckmaster & Thomas 2009: 6), and a similar approach can be found in the ALP's social inclusion policy document, released in the run-up to the 2007 federal election (Gillard & Wong 2007). This approach sees social exclusion as 'the problem' and social inclusion as 'the response', with the relationship between them unexplored but unproblematic. It thus implies that the definitional ambiguities surrounding social exclusion also exist for social inclusion, which thus 'limit its capacity to serve as a framework for the development of social policy' (Buckmaster & Thomas 2009: 37). Others see the links between the two concepts as raising more fundamental issues. Thus, Daly and Silver (2008) argue that the shift from describing a problem (exclusion) to pronouncing a goal (inclusion) shifts the emphasis in ways that can restrict the policy focus (e.g., away from the role of social structure and institutions and onto economic participation and employment) by downplaying the role of issues like discrimination, social disorder, and incoherence that often feature prominently when analysing exclusion. At the same time, the use of an inclusion approach may embody an underlying moral narrative that shifts the blame onto those affected and away from the processes and actions that give rise to exclusion. The term social exclusion helps to keep the focus firmly on who is responsible in a way that social inclusion does not.

This discussion provides background to the analysis that follows, but also indicates how the Australian SIA diverged from that adopted in the UK in some important respects. On the positive side, the SIA (with the support of the Australian Social Inclusion Board, ASIB) was used as the vehicle to promote a number of important reforms in areas such as mental health, homelessness, affordable housing, child care, 'closing the gap', disability services, school funding, and tertiary education. Against this, the failure to redress long-standing failings such as the inadequacy of NSA is disappointing because it occurred in a period when the government implemented a highly successful policy response to the global financial crisis that allowed real incomes to continue to grow as most other economies fell into recession. (5) The fact that the resources generated by this growing material prosperity were not used to rectify one of the most obvious social policy shortfalls is indicative of the inability of the SIA to free itself of the behaviourist and ideological baggage referred to by Bradshaw.

It has already been noted that it is too soon to provide an assessment of the impact of the SIA. However, this does not preclude a discussion of the criteria that should be used in making such an assessment. The key question here is whether or not the agenda has made a difference--a seemingly straightforward task, but one that raises a series of supplementary questions, including:

1. Has it led to new and better ways to conceptualise social problems and policy issues?

2. Has it led to different or new policy objectives or priorities?

3. Has it led to different or new policy approaches?

4. Has it led to different or new political arrangements and new administrative and advisory structures?

5. Has it led to new data and indicators?

6. Has it led to different or new ways of monitoring and assessing the impact and effectiveness of policy?

Opinion is probably still split on the first three of these questions, and while progress was made on question 4 (through the work undertaken as part of the broader COAG reform agenda--see COAG 2010), the lack of coordination between Commonwealth and State Governments remains a major obstacle to the development of the kind of holistic, integrated approach that a social inclusion focus requires. More progress has been made in relation to the final two questions (see below), although how resilient these will be in the face of the change of government remains unclear.

Frameworks and indicators

The definition of social inclusion adopted by the Australian Government was informed by 'a vision of a socially inclusive society' in which '[b]eing socially included means that people have the resources, opportunities and capabilities they need to: learn, work, engage and have a voice' (ASIB, 2010: 15).

This was articulated more specifically in the first social inclusion indicator report:

Social inclusion means building a nation in which all Australians have the opportunity and support they need to participate fully in the nation's economic and community life, develop their own potential and be treated with dignity and respect (Australia Government 2009a: 2).

Reports produced by the ASIB in 2009 and 2012 set out a list of indicators that were used to ascertain 'How Australian is faring'. The later report was described in a Foreword by Minister Mark Butler as one of the most crucial pieces of Board research, which

sets out our progress as a nation across a range of indicators and provides vital insights into what drives social exclusion and how it can be addressed (Australian Government 2012: iii).

The first report identified 33 indicators--13 of which were described as supplementary--spanning 6 broad domains: poverty and low income; lack of access to the job market; limited social supports and networks; effect of the local neighbourhood; exclusion from services; health; and contextual. (6) Of these, 20 were used by the EU and allowed Australia's performance to be benchmarked against the European experience. By 2012, the framework had been revised to include 22 'headline' and 22 supplementary indicators across three broad areas --participation, resources, and multiple disadvantage--spanning 12 domains that aligned more closely with policy goals: work; learn; engage; have a voice; material/economic resources; health and disability; education and skills; social resources; community and institutional resources; housing; personal safety; and multiple and entrenched disadvantage. (7) Poverty and income inequality no longer appeared as indicators and the explicit EU comparisons were dropped, although a number of international comparisons were presented.

Seven priority areas were identified along with six 'strategic change indicators' that were the focus of government policy attention. However, the relationship between the strategic change indicators and the headline and supplementary indicators is unclear, as indeed is the need for the latter, which are bureaucratically driven and do not always have a clear link to social inclusion as such, or to the framework itself. (8) This might seem like a minor issue, but it suggests a weakness in the coherence of the framework that reflects an underlying bureaucratic resistance to engage with a process that would require individual departments and agencies to reformulate the indicators used to measure their performance. In light of the emphasis given in the social inclusion literature on the need for an integrated--'joined-up, government'--approach, it is symptomatic of a fundamental failure to translate the agenda into effective action at an administrative level.

Despite these limitations, the social inclusion indicator framework provided the basis for an informed--evidence-based--debate about the nature of social exclusion and how well government action was promoting inclusion. It allowed researchers and stakeholders to assess progress and provided a basis for identifying where there are gaps that needed filling with new or better data. (9) Significantly, the framework covered both the economic and social dimensions of exclusion/inclusion and thus redressed the imbalanced focus on economic issues that had previously prevailed--in Australia and elsewhere. The concept of multiple disadvantage introduced in 2012--identified as experiencing at least three of six indicators, covering income, joblessness, self-assessed health, educational attainment, safety, and support--was an important initiative that shed new light on the nature and dynamics of social exclusion. It was the forerunner of the concept of deep and persistent disadvantage given prominence subsequently by the Productivity Commission (McLachlan et al. 2013).

But there are other areas where the framework is lacking. One of these is the link between some of the supplementary indicators and social inclusion. Relative income inequality, for example, is an important social indicator in its own right, but the relationship between the degree of inequality and the extent of inclusion--or changes in both--is not immediately obvious and is not explored in the report (see pages 36-7). A similar point can be made about the indicator 'life expectancy at birth', which is described as providing 'a useful summary of the health of the population' with no attempt to explore whether--and how--a healthier population is also more inclusive. These criticisms are not intended to imply that income inequality and life expectancy are not important social issues that need to be tracked and tackled, but rather to question their role as part of an agenda that is focused specifically on social inclusion.

One final comment about the indicator framework relates to the relative neglect of subjective indicators. There are exceptions, including the use of self-assessed health status, subjective life satisfaction, and perceptions of safety at night, but they are relatively few--far fewer than current thinking requires (see CMEPSP 2009). This may reflect the absence of relevant data as the ABS--the main source of data to support the framework--has been reluctant to collect subjective data, but perceptions and feelings are critically important in providing information about people's quality of life, and thus establishing whether and how people are excluded.

The government's social inclusion indicator framework can be compared with those developed by researchers working on social exclusion at the Melbourne Institute and Brotherhood of St Laurence, (MI/BSL) (Scutella, Wilkins & Horn 2009; Scutella, Wilkins & Kostenko 2013) and the Social Policy Research Centre (SPRC) (Saunders, Naidoo & Griffiths 2008; Saunders & Wong 2012a). (10) The MI/BSL framework covers seven domains and includes 29 'measureable' indicators derived from the HILDA survey, while the SPRC framework includes 27 indicators across three broad domains estimated using data from specifically designed national surveys (further details are provided in Saunders 2013). (11) An advantage of the use of HILDA data is that the dynamics of exclusion can be examined (see Horn, Scutella & Wilkins 2011; Scutella, Wilkins & Kostenko 2013), shedding important light on the factors associated with entries and exits into, and durations of, spells of exclusion. Against, this, the use of a survey designed specifically for the purpose overcomes the constraints imposed by lack of relevant data, although at a cost in terms of the smaller sample sizes--and response rates--that such 'unofficial' surveys inevitably involve.

Together, the three frameworks have generated important new data on the nature of social exclusion in Australia and share many similarities in terms of purpose and content. (12) However, some of these similarities reflect overriding aims that become less apparent when the details are compared, as the following two examples illustrate. The first relates to poverty. As noted earlier, this is not included in the government framework, although the indicators do include the percentage of households that fall within the lowest three deciles of both equivalised income and equivalised net wealth and experienced at least five (out of 15) indicators of financial stress in the previous 12 months (identified using data from the ABS Household Expenditure Survey). In contrast, the MI/BSL framework adopts a conventional definition of poverty, based on equivalised income being below 60 per cent of the median, while the subjective indicator included in the SPRC framework is the percentage of households indicating that they do not have enough money to get by on. (13)

The second example is volunteering, which has been widely regarded as a vehicle for building social capital and promoting inclusion (Leong 2008). Again, all three frameworks include an indicator of voluntary activity, but the details vary. The indicator used in the government's framework is whether or not an individual participated in volunteering over the past 12 months (based on data from the ABS General Social Survey). The indicator used in the MI/BSL framework is whether or not an individual undertook voluntary activity in a 'typical week' (based on HILDA data), while the SPRC framework includes as an indicator of exclusion whether or not survey respondents participated in any of a range of community activities over the previous 12 months, where the activities specified include being a volunteer in health and community services. The differences are not great and the underlying intent is similar, but it is still the case that the three approaches can produce different estimates of the extent of volunteering at a point in time, and how participation in voluntary activity is changing over time.

One issue that has been debated in the social exclusion literature is whether or not it is meaningful and useful to derive an aggregate index that summarises the different indicators and captures overall movements. The main advantage of such an index is its ability to convey simply to a wide audience the overall picture and how it is changing. It is not a substitute for the detailed tabulations relating to each indicator provided under a 'dashboard' approach (see OECD 2011), but it has the ability to draw attention to whether progress overall is moving in the right direction. Against this, it has been argued on conceptual grounds that an aggregate index conceals one of the most important features of social exclusion, which is that its different dimensions are inter-connected and, at times, mutually reinforcing (see Burchardt, Le Grand & Piachaud 2002). Furthermore, there is no agreed way of weighting the individual components when constructing an aggregate index, so that any specific index will always be provisional and contested.

Of the three frameworks reviewed above, only the MI/BSL approach explicitly includes an aggregate index. It is derived using a sum-score approach that expresses the sum of indicators experienced by an individual within each domain as a ratio of the total for that domain, and then summing the resulting domain scores across the seven domains (see Scutella, Wilkins & Kostenko 2013: 279-80). This approach assigns greater weight to those indicators from domains with fewer indicators in total, but assigns an equal weight to each domain in the final aggregation to produce a sum-score index that varies between zero and seven. (14) Since the number of indicators is a reflection of data availability, this has the somewhat perverse effect of giving greater weight to those aspects of exclusion about which least is known and less weight to those where information is most readily to hand. (15)

An alternative aggregation procedure used by SPRC researchers when summarising the extent of social exclusion (see below) involves simply counting the number of indicators experienced by each individual--either within each domain or across all three domains--and using this score to capture the degree --or severity--of exclusion experienced. (16) This approach is used primarily for analytical purposes, including to examine the overlap between exclusion and poverty and the association between exclusion and other socioeconomic variables. However, a summary index can also be used as a descriptive device for illuminating the nature of changes in social exclusion and to examine the association between exclusion and well-being, as the next section illustrates.

Monitoring exclusion and its association with well-being

This section begins by briefly reviewing what the three indicator frameworks reveal about recent changes in social exclusion. Not surprisingly, the government framework reveals a diverse pattern of change over the four years to 2010, with indicators moving in different directions, but rarely changing significantly in the statistical sense. It is noted that the indicators 'relate to complex and persistent problems which require a long-term approach, including to measurement and reporting' (Australian Government 2012: 3). Despite this, there is evidence of overall improvement, with thirteen examples cited where Australia is 'Doing well' and only eight 'Areas for improvement'. The latter list contains some key indicators including the incidence of multiple disadvantage--which declined slightly but not significantly, and remained at 4.6 per cent--and income inequality, which 'increased steadily from the mid-1990s'.

This picture of modest overall improvement is confirmed over the period 2006-10 by both the MI/BSL and SPRC indicator frameworks. The latter shows, for instance, that the mean incidence of the 27 indicators of exclusion declined from 19.1 per cent in 2006 to 16.6 per cent in 2010, mainly as a result of improved economic circumstances associated with falls in unemployment and joblessness, although again many of the observed changes are not statistically significant (Saunders & Wong 2012a: Table 7.2). Even so, these examples highlight the value of the indicators in monitoring change, and their value will increase as the time period covered is extended and there is greater scope to instigate change. Turning to some detailed findings, attention focuses on the SPRC findings, specifically on results from research on social exclusion conducted by the author with colleagues at the SPRC. These results illustrate the value of using simple indicators and sum-score measures to improve understanding of the nature of social exclusion, how it is changing, and its impact on those affected. (17)

Social exclusion has been measured in ways that draw on its conceptual basis as it has evolved in the literature, relying on data from two nationally representative surveys conducted in 2006 and 2010 that provide the basis for examining how social exclusion has changed over time on a consistent basis, using a set of indicators that capture its different dimensions. (18) Both surveys also included a series of questions about different aspects of the subjective well-being of respondents, and these data can be used to examine the association between the experience of social exclusion and perceived well-being.

This is an important issue that relates to the question (raised by Hick 2012, but also by others) about the role of choice and constraint in creating identified exclusion. If what is being identified as social exclusion is in fact no more than the diversity in selected activities that one would expect given the differences in individual preferences, then to label some of these outcomes as exclusion--and thus implicitly in need of redress--could create a sense of stigma where none exist. If instead, exclusion is imposed on those unable to avoid it, then there is a role for policy in understanding the situation, identifying causes and addressing them. An insight into which of these two situations better fits reality can be provided by examining the association between the severity of exclusion experienced--as identified using indicator sum-scores--and the expressed well-being of those affected. (19)

Table 1 uses an unweighted sum-score index to examine how social exclusion changed between 2006 and 2010--in aggregate and within each domain. Deep exclusion is defined as those who experience at least seven out of all 27 forms of exclusion, while those who experience at least 10 forms of exclusion are defined as being in very deep exclusion. Sample estimates have been weighted by the age composition of the population to adjust for any age bias.

As noted earlier, these estimates imply a slight decline in overall exclusion over the period, the mean score falling by more than nine per cent, from 3.93 to 3.56. However, this decline was not spread evenly across all dimensions or domains, with the incidence of some indicators increasing. (20) Table 1 also indicates that despite falls in the size of both the deep and very deep excluded sub-samples over the period (shown in the first row), the overall severity of exclusion experienced by both groups increased in absolute--and hence relative--terms. (21) This worsening position is apparent across the first two domains (disengagement and service exclusion), but not in the economic exclusion domain, where absolute, but not relative progress was made despite the global financial crisis.

Although not shown in Table 1, the degree of deep and very deep exclusion declined in relation to access to several important social services, including dental treatment and child care, which were a focus of government action over the period. In the domain of economic exclusion, exclusion relating to unemployment and joblessness also declined markedly. These changes can be attributed in part to policy initiatives introduced by the federal government, suggesting that the social inclusion agenda has had some success, although any such conclusion should be treated with a degree of caution because the indicators that underlie the results in Table 1 do not capture the impact of government policy, nor are they responsive to many of the government's identified priority areas. Even so, the results highlight the value of using individual and aggregated indicators to track changes in social exclusion.

Attention now turns to the association between the severity of exclusion experienced and reported levels of subjective well-being (SWB). It is important to emphasise that the results that follow capture the association between the relevant variables and cannot be interpreted as reflecting causation, which may run in both directions: those who are excluded may experience a decline in well-being, while those with low levels of well-being may be less willing or able to participate and hence more likely to experience exclusion.

The association between exclusion and SWB has been examined using the following four indicators:

Assessed standard of living: a 5-point scale rating people's current standard of living, going from 1 (= very low) to 5 (= very high)

Life satisfaction: a 5-point scale measuring how satisfied people are with their overall standard of living, going from 1 (= very dissatisfied) to 5 (= very satisfied)

Happiness: a 4-point scale measuring how happy people feel generally, going from 1 (= very unhappy) to 4 (= very happy)

Perceived autonomy: a 10-point scale measuring how much choice and control people have over 'your own life and the things that happen to you', going from 'none at all' (1-2) to 'some choice and control' (5-6) to a 'a great deal of control' (8-10)

The mean value of each SWB indicator has been estimated for those who experience different degrees of overall social exclusion, identified by the minimum number of indicators (out of 27) on which they are excluded, and the results are presented in Figures 1 to 4. As can be seen, there is a steady and systematic decline in the mean value of each SWB indicator as the degree of exclusion intensifies. Thus for example, the well-being of those in deep exclusion (i.e., experiencing at least seven instances of exclusion) declines by 33.0 per cent (assessed standard of living), 35.2 per cent (life satisfaction), 22.1 per cent (happiness) and 30.5 per cent (perceived autonomy) respectively, compared with those who experience no exclusion.

In light of the consistency in the shape and strength of the four relationships, these results suggest that social exclusion as identified here is associated with a substantial decline in well-being, however it is measured. (22) It is possible to argue that causation goes from well-being to exclusion rather than the other way, although it is unclear what the underlying causal mechanisms are. What seems more likely is that the well-being of those who experience the forms of exclusion identified by the indicators used in the SPRC research is negatively affected by that experience, with the degree of negativity increasing as the severity of exclusion increases.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

Conclusions

This paper has described the Australian experience over the late 2000s when social inclusion formed the core of the government's social policy agenda. The change of government in 2013 has almost certainly seen the demise of social inclusion as a policy priority, but this does not mean that the foregoing discussion has been of only historical interest. It raises important and enduring questions about the nature and goals of social policy, as well as about the ability of a concept like social inclusion to bring about real change. It also illustrates how the use of an indicator framework can exert pressure on data providers to fill existing gaps, raising issues about the selection of indicators and construction of composite aggregates that give an impetus to policy by helping to identify areas of achievement and failure. The forces that have seen these latter factors come to the fore will not dissipate in a world where social policy challenges are becoming more complex and interconnected.

Despite several notable policy reform achievements, relatively little progress has been made in tackling the bureaucratic and administrative obstacles that prevent the kind of holistic integrated approach needed to combat social exclusion. More might have been achieved in a unitary system and there is little doubt that the politics of Australian federalism--spurred on by an unstable governing coalition and a hostile opposition--made it harder to achieve the kinds of coordinated action that was needed. Even within the federal bureaucracy itself, there were signs that departments sought to protect traditional territory and resist the kinds of change that the inclusion agenda demanded. The silos proved to be resilient.

The government's social indicator framework developed over the period is a positive outcome that, if maintained in some form, has the potential to produce enduring benefits. In conjunction with the two other indicator frameworks discussed here, the prospects for better monitoring social change and tracking policy impacts have been improved. A key issue for the future is the extent to which policy relevance and value will outweigh political opportunism. The results presented in the preceding section illustrate the value of indicators in helping to document change, highlight achievement, and identify areas where more action is needed. By helping to reinvigorate the use of indicators, the Australian social inclusion agenda has had a positive impact, even if such limited success could have been surpassed in a period that generated enough economic resources to do more.

Acknowledgement

This paper is a revised version of the paper originally presented to the Workshop held in Adelaide in March 2013. The author acknowledges comments provided by the workshop audience and by two anonymous referees. The paper draws on but does not duplicate material since published in Saunders (2013). Financial support from the Australian Research Council under Linkage grant LP100100562 is acknowledged.

References

Atkinson, A.B. (1998) 'Social Exclusion, Poverty and Unemployment', in A.B. Atkinson and J. Hills (eds.), Exclusion, Employment and Opportunity, CASE paper 4, Centre for the Analysis of Social Exclusion, London School of Economics, 1-20.

Atkinson, A.B. (2003) 'Multidimensional Deprivation: Contrasting Social Welfare and Counting Approaches', Journal of Economic Inequality, 1 (1), 51-65.

Atkinson, A. B., Cantillon, B., Marlier, E. and Nolan, B. (2002) Social Indicators. The EU and Social Inclusion, Oxford, Oxford University Press.

Australian Bureau of Statistics (2002) Measures of Australia's Progress 2002, Catalogue No. 1370.0, Canberra, ABS.

ABS (2011) General Social Survey. Summary Results. Australia 2010, Catalogue No. 4159.0, Canberra, ABS.

ABS (2013) Household Income and Income Distribution, Australia, 2011-12, Catalogue No. 6523.0, Canberra, ABS.

ASIB (Australian Government Social Inclusion Board) (2010) Social inclusion in Australia: How Australia is faring, Canberra, Department of Prime Minister and Cabinet.

Australian Council of Social Service (ACOSS) (2013) Reduce Poverty: Improve Employment. 2013 Election Brief, Sydney, ACOSS.

Australian Government (2008) Social Inclusion Principles for Australia, Canberra, Social Inclusion Unit, Department of Prime Minister and Cabinet.

Australian Government (2009a) A Stronger, Fairer Australia, Canberra, Department of Prime Minister and Cabinet.

Australian Government (2009b) Social Inclusion in Australia. How Australia is Faring, Canberra, Australian Social Inclusion Board, Department of Prime Minister and Cabinet.

Australian Government (2012) Social Inclusion in Australia. How Australia is Faring. 2nd Edition, Canberra, Australian Social Inclusion Board, Department of Prime Minister and Cabinet.

Azpitarte, F. (2012) Social Exclusion Monitor Bulletin, April 2012, Melbourne, Brotherhood of St Laurence and Melbourne Institute of Applied Economic and Social Research.

Beland, D. (2007) 'The Social Exclusion Discourse: Ideas and Policy Change', Policy & Politics, 35 (1), 123-39.

Bradshaw, J. (2004) 'How Has the Notion of Social Exclusion Developed in the European Discourse?', Economic and Labour Relations Review, 14 (2), 168-86.

Buckmaster, L. and Thomas, M. (2009) 'Social Inclusion and Social Citizenship. Towards a Truly Inclusive Society', Research Paper No. 8, 2009-10, Parliamentary Library, Canberra, Parliament of Australia

Burchardt, T., Le Grand, J. and Piachaud, D. (2002) 'Introduction' in J. Hills, J. Le Grand and D.Piachaud (eds.), Understanding Social Exclusion, Oxford, Oxford University Press, 1-12.

Cappellari, L. and Jenkins, S. P. (2007) 'Summarising Multiple Depriviation Indicators', In Jenkins, S.P. and Micklewright, J. (eds.), Inequality and Poverty Re-examined, Oxford, Oxford University Press, 166-184.

CMEPSP (Commission on the Measurement of Economic Performance and Social Progress) (2009) Report on the Measurement of Economic Performance and Social Progress, www.stiglitz-sen-fitoussi.fr/en/documents.htm (accessed 30 April 2015).

Council of Australian Governments (COAG) Reform Council, (2010), COAG Reform Agenda: Report on Progress 2011, Sydney, COAG Reform Council.

Daly, M. and Silver, H. (2008) 'Social Exclusion and Social capital: A Comparison and Critique', Theory and Society, 37 (6), 537-66.

Gregory, R. G. (2013) 'The Henderson Question? The Melbourne Institute and Fifty Years of Welfare Policy', Discussion Paper No. 682, Canberra, Centre for Economic Policy Research, Australian National University.

Gillard, J. and Wong, P. (2007) An Australian Social Inclusion Agenda, Canberra, Australian Labor Party.

Hayes, A., Gray, M. and Edwards, B. (2008) Social Inclusion. Origins, Concepts and Key Themes, Canberra, Social Inclusion Unit, Department of Prime Minister and Cabinet.

Hick, R. (2012) 'The Capability Approach: Insights for a New Poverty Focus', Journal of Social Policy, 41 (2), 291-308.

Horn, M., Scutella, R. and Wilkins, R. (2011) Social Exclusion Monitor Bulletin, September 2011, Melbourne, Brotherhood of St Laurence and Melbourne Institute of Applied Economic and Social Research.

Leong, V. (2008) 'Volunteering: Pathway to Inclusion', Australian Journal of Volunteering, 13 (2), 67-73.

Levitas, R., Pantazis, C., Fahmy, E., Gordon, D., Lloyd E. and Patsios, D. (2007) The Multi-dimensional Analysis of Social Exclusion, Bristol, Department of Sociology and School for Social Policy, The Townsend Centre for the International Study of Poverty and The Bristol Institute for Public Affairs, University of Bristol.

McLachlan, R., Gilfillan, G. and Gordon, J. (2013) 'Deep and Persistent Disadvantage in Australia', Productivity Commission Staff Working Paper, Canberra, Productivity Commission.

OECD (2008) Growing Unequal? Income Distribution and Poverty in OECD Countries, OECD, Paris

OECD (2010) Economic Surveys, Australia 2010, Paris, OECD.

OECD (2011) How's Life? 2013--Measuring Well-Being, Paris, OECD.

Saraceno, C. (2002) 'Social Exclusion: Cultural Roots and Variations on a Popular Concept', In A.J. Kahn and S.B. Kamerman (eds), Beyond Child Poverty: The Social Exclusion of Children, New York, Institute for Child and Family Policy, Colombia University, 37-74.

Saunders, P. (2011) Down and Out: Poverty and Exclusion in Australia, Bristol: Policy Press.

--(2013) 'Reflections on the Concept of Social Exclusion and the Australian Social Inclusion Agenda', Social Policy and Administration, 47 (6), 692-708.

Saunders, P., Naidoo, Y. and Griffiths, M. (2007) Towards New Indicators of Disadvantage: Deprivation and Social Exclusion in Australia, Sydney, Social Policy Research Centre, University of New South Wales.

--(2008) 'Towards New Indicators of Disadvantage: Deprivation and Social Exclusion in Australia', Australian Journal of Social Issues, 43 (2), 175-94.

Saunders, P. and Wong, M. (2012a) Promoting Inclusion and Combating Deprivation: Recent Changes in Social Disadvantage in Australia, Sydney, Social Policy Research Centre, University of New South Wales.

--(2012b) 'Estimating the Impact of the Global Financial Crisis on Poverty and Deprivation', Australian Journal of Social Issues, 47 (4), 485-503.

Scutella, R., Wilkins, R. and Horn, M. (2008) 'Measuring Poverty and Social Exclusion in Australia. A Proposed Multidimensional Framework for Identifying Socio-economic Disadvantage', Melbourne, Brotherhood of St Laurence and Melbourne Institute of Applied Economic and Social Research.

Scutella, R., Wilkins, R. and Kostenko, W. (2013) 'Intensity and Persistence of Individuals' Social Exclusion in Australia', Australian Journal of Social Issues, 48(3), 273-98.

Whiteford, P. (2010) 'Why Unemployment Benefits Need To Be Increased', Australian Policy Online, December, http://apo.org.au/ (accessed 30 April 2015).

Wilkins, R. and Warren, D. (2012) 'Social Exclusion in Australia', in Families, Incomes and Jobs, Volume 7. A Statistical Report on Waves 1 to 9 of the Household, Income and labour Dynamics in Australia Survey, Melbourne, Melbourne Institute of Applied Economic and Social Research.

Endnotes

(1) See also the useful review provided for an Australian audience by Hayes, Gray and Edwards (2008).

(2) A useful summary of the debate over the definition of social exclusion is provided by The Productivity Commission (McLachlan et al. 2013: 46-51). The report notes that: 'The multi-dimensional nature of social inclusion (or exclusion) as a concept is one of its strengths but it also poses significant challenges for measurement and evaluation' (McLachlan et al. 2013: 8).

(3) In 2011-12, the ABS estimates that (OECD equivalised) median income was equal to $790 a week (ABS 2013: Table 1). In December 2012 (the mid-point of the ABS survey period), the single rate of NSA was $243 a week or just above 30 per cent of median income. Real median income had increased from $505 in 1994-95, or by 56.4 per cent, while NSA remained constant in real terms.

(4) In contrast to the reluctance to include the poverty rate, both reports provide useful analysis of changes in the Gini coefficient measure of income inequality, including how Australia compares with other OECD countries.

(5) The policy response to the GFC provided the government with an ideal opportunity to increase the rate of NSA (along with other social benefits) as a way of boosting consumption demand to offset the anticipated recessionary impacts of the financial crisis.

(6) It is worth noting at this stage that the use of indicators in the current context reflects both the ambiguities in the concept of social exclusion and its multi-dimensional nature. The latter feature alone does not necessarily dictate the use of indicators, if the dimensions themselves are clearly specified and there is a long history of using such measures in social science and public health research. The properties of indicators are discussed by Atkinson et al. (2002), ABS (2002: Appendix 1) and OECD (2013: Chapter 1).

(7) The 2012 report (page 88) incorrectly lists 'access to health services' as a supplementary rather than a headline indicator.

(8) An example is the cross-cutting strategic change indicator 'obesity rate by Indigenous status, SEIFA and remoteness', which may provide important information on one aspect of social policy, but has limited relevance to the achievement of social inclusion. The strategic change indicators are specified on pages 89-92 of Australian Government (2012).

(9) One example of this was the decision to introduce a special module on social inclusion into the 2010 General Social Survey conducted by the Australian Bureau of Statistics (ABS) that produced valuable information used in the 2012 indicator report (see ABS 2011).

(10) For additional insight into what the HILDA data imply for the profile of social exclusion in Australia, see Wilkins and Warren (2012).

(11) The MI/BSL domains (numbers of indicators within each domain are shown in brackets) are: material resources (4); employment (5); education and skills (5); health and disability (5); social (2); community (5); and personal safety (3). The SPRC domains are: disengagement (9); service exclusion (10); and economic exclusion (8).

(12) Both the MI/BSL and SPRC studies indicate that social exclusion is widespread in Australia. Thus, Scutella, Wilkins and Kostenko (2013: 292) report that between 2001 and 2007; '20 to 30 per cent of the Australian population aged 15 years and over experience what we refer to as 'marginal exclusion' at any point in time, while Saunders and Wong (2012a: Table 7.3) estimate that in 2010, over 37.6 per cent of adults experienced at least four forms of exclusion and 20.8 per cent at least six forms.

(13) The MI/BSL framework also includes indicators based separately on low (below 60 per cent of the median) household wealth and consumption expenditure. The SPRC framework does not include the poverty rate as an exclusion indicator because of its focus on examining the overlap between social exclusion and poverty (see Saunders & Naidoo 2009; Saunders & Wong 2012b).

(14) The former component means, for example, that the indicator 'little social support'--which is one of only two indicators in the Social domain--is assigned greater weight than the indicator 'unemployed', which is one of five indicators in the Employment domain (see Scutella et al.: Table 2).

(15) The degree of social exclusion is then classified into three categories in the MI/BSL framework: marginal exclusion (scores between 1 and 2), deep exclusion (scores above 2) and very deep exclusion (scores above 3) (see Azpitarte 2012: 2).

(16) The use and merits of applying a 'counting' approach when deriving an aggregate index is discussed by Atkinson (2003) and Cappellari and Jenkins (2007).

(17) The conceptual framework and survey data used in this research and in what follows have been described elsewhere (Saunders, Naidoo & Griffiths 2007; 2008; Saunders & Wong 2012a; 2012b) and those details will not be repeated here. Interested readers can contact the author for further details if required.

(18) In fact, a series of new indicators were included in the later (2012) survey to better capture the impact of the global financial crisis and the locational, community-based dimension of social exclusion. The analysis reported here is based only on those indicators that were included in both surveys.

(19) The following analysis draws on results presented and discussed by Saunders (2013).

(20) This picture of diversity is consistent with that produced for the same period by the government's social indicator framework, which was described by Minister Butler as one in which 'we have made good progress in many areas, [but] further work is needed to address some of the complex and persistent issues identified in the report'.

(21) The detailed analysis reported by Saunders (2013) confirms that the overall picture is one of improvement, with the incidence of exclusion declining for 20 of the 27 indicators across the whole sample. However, the incidence of exclusion among the deeply excluded sub-sample declined for only 11 indicators, and among the very deeply excluded sample for only nine indicators.

(22) The same patterns are apparent if exclusion severity is defined in terms of ranges (i.e., 0 indicators of exclusion, 1-3 indicators, 4-7 indicators, and so on). The results also reinforce those reported by Saunders (2013), who shows that exclusion has a larger negative impact on well-being than poverty, defined using an income poverty line.

Table 1. Social exclusion indicator trends, 2006-10 (age-weighted)

                                           2006

Exclusion             Full      Deep     Very deep    Ratio     Ratio
Indicator            sample   excluded   excluded    (b)/(a)   (c)/(a)
                      (a)        (n         (n
                              [greater   [greater
                              than or     than or
                               equal       equal
                               to] 7)     to] 10)
                                (b)         (c)

Sample size          2,626      503         221       0.192     0.084
(unweighted)

Disengagement         1.28      3.19       4.03        2.5       3.1
(1 = 9): Mean
disengagement
score

Service Exclusion     1.56      3.34       4.04        2.1       2.6
(1 = 10): Mean
service exclusion
score

Economic Exclusion    1.08      3.14       4.04        2.9       3.7
(1 = 8): Mean
economic
exclusion score

Mean social           3.93      9.67       12.10       2.5       3.1
exclusion score
(1 = 27)

                                           2010

Exclusion             Full      Deep     Very deep    Ratio     Ratio
Indicator            sample   excluded   excluded    (b)/(a)   (c)/(a)
                      (a)        (n         (n
                              [greater   [greater
                              than or     than or
                               equal       equal
                               to] 7)     to] 10)
                                (b)         (c)

Sample size          2,605      410         186       0.157     0.071
(unweighted)

Disengagement         1.27      3.61       4.54        2.8       3.6
(1 = 9): Mean
disengagement
score

Service Exclusion     1.37      3.39       4.29        2.5       3.1
(1 = 10): Mean
service exclusion
score

Economic Exclusion    0.92      2.99       3.78        3.2       4.1
(1 = 8): Mean
economic
exclusion score

Mean social           3.56      9.99       12.60       2.8       3.5
exclusion score
(1 = 27)

Source: Saunders (2013: Table 1).