Worlds of welfare

Welfare state theory has been dominated for the past 20 years by Reference Esping-AndersenEsping-Andersen’s (1990) ‘three worlds of welfare’ typology. This typology not only sought to capture common policy elements across jurisdictions, but also a common political logic. Echoing Marshall’s earlier argument, Esping-Andersen saw the welfare state as a political strategy to advance social citizenship, not only a mechanism to prevent poverty or address a specific social need. Thus, the division between social democratic, conservative and liberal models followed ideological distinctions as well as distinctions in policy design.

Esping-Andersen initially classified Australia as a liberal welfare state, alongside other Anglo democracies. This reflected the preponderance of means-tested benefits and relatively low taxes. While Esping-Andersen’s approach went beyond simple measures of social spending, his analysis nonetheless focused predominantly on social welfare. As Reference TitmussTitmuss (1958) has argued, if the objective of analysis is to understand the nature of citizenship and social protection, this may be too narrow a gaze (p. 42). Australian scholarship has largely rejected a simple catergorisation of Australia as a liberal regime. Instead, following Reference CastlesCastles (1985), Australian scholars tended to emphasise the crucial role of industrial protections, and what might now be called ‘affluence tested’ payments. This ensured a relative equality of market incomes, and a flat-rate payments system designed to meet the needs of manual workers rather than the professional middle class.

The description of Australia as a ‘wage earners’ welfare state, however, has now been widely challenged following three decades of market restructuring. Reference CastlesCastles (1994) himself defended the initial period of reform, associated with the Hawke and Keating Labor governments, as a period of partial renovation based on expanding the social wage, such as Medicare, and occupational welfare, such as superannuation (Reference Castles, Shirley, Castles, Gerritsen and VowlesCastles and Shirley, 1996). He only said ‘farewell’ to the wage earner description after more radical industrial and welfare reforms under the conservative Coalition (Reference CastlesCastles, 2001). Others see greater continuity through this period, and a dismantling of the core pillars of social citizenship towards a model of workfare clearly orientated towards promoting paid employment, which itself became less secure and less equal (Reference Bryson, Verity, Alcock and CraigBryson and Verity, 2009; Reference JamrozikJamrozik, 2009; Reference McDonald and MarstonMcDonald and Marston, 2005; Reference ShaverShaver, 2002).

The importance of labour market protections to the overall Australian welfare state has made broader macro-economic changes more central to evaluations of social policy. The moves from a broad commitment to full employment towards inflation targeting and away from centralised wage arbitration towards workplace and individual contracts have recast Australian welfare (see Reference Wilson, Spies-Butcher and StebbingWilson et al., 2013). Evaluations are complicated, though, by the social wage compensation offered through the Accord. Thus, there was a rapid increase in female labour force participation (Reference BrennanBrennan, 1998 [1994]) and an increase in social spending targeted at low- and middle-income families (Reference HardingHarding, 1997). This clearly limited the impact of economic restructuring on inequality compared to the more brutal experiences of New Zealand, the United States and the United Kingdom (see Reference Castles, Gerritsen and VowlesCastles et al., 1996). It has led to some confusion within the social policy literature on the meaning of ‘neoliberalism’, given that social spending was rarely itself under direct attack (e.g. Reference Fenna and TapperFenna and Tapper, 2012).

Even so, inequality increased over the period (OECD, 2011). The expansion of targeted social payments was more than offset by growing inequalities in market incomes. The nature of social risk also changed as the labour market became less secure. Without full employment, and with the introduction of workfare measures and the relative erosion in the adequacy of benefits for the unemployed, targeted welfare tended to reinforce precarity (Reference McDonald and ChenowethMcDonald and Chenoweth, 2006). Alongside this, Stebbing and I have argued that this period of neoliberal market restructuring has produced a ‘dual welfare state’ (Reference Spies-Butcher and StebbingSpies-Butcher and Stebbing, 2011; Reference Stebbing and Spies-ButcherStebbing and Spies-Butcher, 2010). Tax concessions for superannuation parallel direct funding of the aged pension. Private health insurance is subsidised alongside Medicare, while childcare funding provides alternative subsidies targeted to groups on different incomes to purchase care in a semi-private market. Thus, there are parallel policies supporting public and private provision in the same policy domains.

This perspective provides a less rosy assessment of Australian social policy than often given in comparative international assessments. For example, OECD reports have praised Australia’s ability to target social payments effectively, doing more to reduce inequality per public dollar than anywhere else (Reference WhitefordWhiteford, 2008). While the model was largely pioneered by Labor governments, conservatives have now also embraced what they call a ‘platypus model’ that allows relatively free markets without the extreme inequality of countries like the United States (Reference AlexanderAlexander, 2010/2011). The redistributive strength of Australia’s tax and spend system has begun to recede in recent years, particularly due to changes in taxation (Reference WhitefordWhiteford, 2013: 40). However, it remains far more egalitarian than elsewhere.

This article attempts to provide an initial step in reconciling these competing accounts, which variously see Australia as part of the neoliberal vanguard, or a creative experiment in combining markets and equity. It does this by focusing particularly on Labor attempts to draw on neoclassical economic methods while advancing social equity. It argues this marketisation has some undesirable longer term political implications consistent with both the dual welfare state thesis and growing welfare paternalism. The argument echoes the concerns of Korpi and Palme that by promoting welfare targeting, social policy can erode the political coalitions needed to achieve equity.

Making markets

The adoption of market-orientated economic restructuring during the 1980s and 1990s is a more complex story than can be addressed here. It clearly involves a global shift in economic power, as well as national interventions from more organised business groups (See Reference Hacker and PiersonHacker and Pierson, 2010; Reference HarveyHarvey, 2005). However, in Australia, Labor’s acceptance of the economic logic behind market reforms was significant. Labor came to rely heavily on orthodox neoclassical economic advice. As Reference PuseyPusey (1990) and others have pointed out, the internal economic orthodoxy within the state was reshaped during the 1970s and 1980s in ways that substantially shifted policy making in favour of markets rather than nation building.

The embrace of ‘economic rationalism’ reflected changes in the academy, which increasingly rejected more Keynesian and institutionalist understandings of the role of the state in developing the economy (Reference Groenewegen and McFarlaneGroenewegen and McFarlane, 1990; Reference StrettonStretton, 1987). But unlike some more orthodox understandings of neoliberalism, this embrace did not necessarily mean marginalising the state. Instead, the economists Labor worked with were often committed to some sense of social equity – even a form of social democracy – and saw the state playing a central role. That role, however, was now to shape competition, rather than to replace markets with direct provision.

The trend towards using markets as a mechanism to advance the broader social goals of Labor was already emerging under the Whitlam Government. This was reflected not only in reforms traditionally associated with the rise of economic rationalism – such as the 25% tariff cut or the last Hayden Budget – but also in reforms usually thought to be part of Labor’s social democratic legacy. Two of those reforms stand out: the introduction of the Trade Practices Act and Medibank (the forerunner of Medicare). Both were modelled on traditional neoclassical reasoning, and both explicitly used state regulation to impose competition in markets where producers exercised significant market power. This not only disciplined producers in the sense usually associated with consumer sovereignty, but because competition was imposed by regulation, the impact was often more selective, reflecting state imperatives. This is perhaps more obvious with the Trade Practices Act, which built on a tradition of anti-monopolistic legislation consistent with neoclassical reasoning but usually opposed by strong market advocates, such as Anti-Trust legislation in the United States.

Australia’s universal health insurance system reflected similar economic logics. Prior to 1969, Labor’s health policy, inherited largely unchanged from the Chifley Government of the 1940s, was to nationalise health care. State Labor governments had achieved this to some extent in Queensland already, and only the High Court prevented Chifley from doing likewise (Reference GillespieGillespie, 1991). There remained some concerns that the policy would continue to face constitutional barriers. However, the impetus for a shift in policy came from a proposal by two health economists, John Deeble and Richard Scotton. Their plan, which became Medibank, was a response to inefficiencies and inequities in the subsidised system of private health insurance (Reference Scotton and MacdonaldScotton and Macdonald, 1993). Their analysis differed from most traditional neoclassical approaches in that it accepted medical treatment was a ‘need’ rather than a normal consumer good, and therefore that failure to access needed medical treatment could be considered a failure of provision. But otherwise, their approach was consistent with mainstream analysis.

The two economists argued that a single national health insurer would solve many of the problems in the health market. These included high administrative costs, a lack of price control and asymmetric information between patients and doctors (Reference Scotton and DeebleScotton and Deeble, 1968). Importantly, they did not advocate eliminating the private sector entirely, and indeed their plan envisaged the sector would remain sizeable. They saw the private sector playing a legitimate role in allowing patients a choice of doctor and accommodation (Reference Scotton and DeebleScotton and Deeble, 1968: 11–12). This was not a plan for gradual socialisation, nor did it respond to class segregation within the health system. Instead, a central concern for both authors was efficiency and cost control. And both spent considerable energy demonstrating that a single public insurer was more effective at containing costs than was the previous private market (Reference DeebleDeeble, 1982; Reference Scotton, Scotton and FerberScotton, 1980). Indeed, reading their work gives the impression that the scheme was as much a micro-economic reform as it was a form of expanded social provision.

There was significant resistance to the new health policy within the Labor Party. The move from nationalisation to public insurance of private provision was, rightly, seen as a significant move away from the socialisation goals of the party’s leftwing. Even as the left within the party and union movement campaigned for Medicare in the 1980s, it continued to demand future reforms to extend public control of health care, and the expansion of community health (see Reference Boxall and GillespieBoxall and Gillespie, 2013). Yet, the model’s market orientation was also a significant internal advantage. Initially, it helped to overcome constitutional barriers to medical conscription. Later, its impact on inflation and the Budget, alongside ongoing support from unions (Reference Boxall and GillespieBoxall and Gillespie, 2013: 114–142), was central in Labor determining to commit to its reintroduction in the 1980s, when otherwise it was shying away from large spending commitments (Reference SaxSax, 1984: 241).

A similar approach emerged in higher education, where another economist, Bruce Chapman, developed a scheme for more equitably charging university students for tuition. Chapman’s scheme, introduced as the Higher Education Contribution Scheme (HECS), involved the reintroduction of student fees alongside a system of public loans repayable through the tax system (Reference Chapman and RyanChapman and Ryan, 2005). Again this followed a neoclassical logic, which attempted to separate out the ‘private’ and ‘social’ benefits of higher education. Many in Labor were relatively hostile to subsidising higher education, given it primarily worked to benefit middle-class students, but upfront fees also prevented working-class students entering the system. Chapman’s scheme attempted to solve the ‘capital market’ problem these students faced by guaranteeing access to loans and allowing repayments to be deferred (perhaps indefinitely) until the student had greater ability to pay.

The scheme has largely proved a success on the equity grounds its creator suggested and has now been taken up in other jurisdictions. There is little evidence that the loans discouraged poorer students from enrolling (Reference Chapman and RyanChapman and Ryan, 2005). However, it is also clear that reintroducing fees appears to have lowered the political barriers to further deregulation, which now extends to higher and differential fees. This process has been accelerated to the more radical deregulation of international student fees, which became a central source of new university revenues as their operations expanded but public funds stagnated.

Student fees initially provided a modest boost to overall funding for the sector and were levied at relatively low rates. Since then fees have increased and repayment thresholds have changed several times. There have recently been calls to ‘privatise’ student debts ( The Sydney Morning Herald, 2013). In this sense, fees have facilitated the commodification of higher education and potentially the trends towards a greater emphasis on vocation among students and faculty (Reference Marginson and ConsidineMarginson and Considine, 2000). As with more recent reforms to increase access to universities through the deregulation of student places (Reference Stokes and WrightStokes and Wright, 2012), Labor has consistently used marketisation alongside its promotion of equity, while also advancing a process of commodification.

Labor’s embrace of competition as an organising principle has been much broader than these more targeted schemes. These examples do suggest, however, an attempt to separate the relational aspects of commodification that result from competition from the distributional consequences of free markets. Labor has consistently viewed many areas of policy making in this way, conceiving of the policy domain as a market, attempting to improve the operation of competition in a neoclassical sense of pareto optimality and then also attempting to limit inequality. This, for example, describes almost exactly the way Labor has understood climate policy through the process of the Reference GarnautGarnaut (2008) review and the Clean Energy Future package (Reference Spies-ButcherSpies-Butcher, 2010: 59–60). It highlights not only how neoliberal thinking has conditioned Australian politics as a whole but also how an egalitarian impulse of sorts has remained.

Market dynamics

Labor’s attempt to promote markets while being mindful of potential inequities – even to use markets to promote equity – invites a more thorough engagement with the processes of marketisation in social provision. A complete review of this literature – and the related politics of the Third Way – is beyond this article. Instead, it seeks to draw on the analysis of marketisation to understand how this process is related to the broader dual welfare dynamic. Reference GiddensAnthony Giddens (1998) has provided the most influential synthesis of market models and social democratic ideals, although Australian Labor politicians have followed (Reference LathamLatham, 1998; Reference TannerTanner, 1999). In practical terms, implementation of marketisation in social policy has drawn heavily on the work of Reference Le GrandJulian Le Grand (1997) and Reference BarrNicholas Barr (2001). In Australia, Reference KeatingMichael Keating’s (2004) work Who Rules? outlines a similar project. Both Keating and Le Grand see marketisation as a state project. Markets are part of the tool kits of governments, and can and should be used by governments to shape behaviour. As Keating explicitly states, competition can allow states to rule more effectively than direct bureaucratic control.

Le Grand’s framework for incorporating markets into social provision is particularly useful here. Like the earlier work of Scotton and Deeble, Le Grand’s discussion of quasi-markets sets out to correct market failures by better aligning incentives to those one would normally expect from a neoclassical textbook. For example, one account outlines five pre-requisites for managed markets to be successful (Reference Le Grand and BartlettLe Grand and Bartlett, 1993: 19–34). These are some level of competition between providers, appropriate motivations and incentives of buyers and sellers (in a context where firms are not necessarily for-profit and buyers are not necessarily the end recipient), symmetry of information, minimal transaction costs and avoidance of ‘creaming’ (also Reference Davidson, Meagher and KingDavidson, 2009: 51). This ensures efficiency, while aligning incentives to the social good, and ensuring equitable distribution of funding through public spending.

The models developed by Le Grand, and more recently in Australia by Reference Davidson, Meagher and KingBob Davidson (2009), are helpful in highlighting the potential problems in managed markets and how they might be addressed. Their focus is primarily on a micro-level of statecraft – assisting governments to make micro-economic reforms within social policy to improve efficiency and equity. The purpose in this article is more explicitly political economic. It seeks to understand how different forms of marketisation might evolve politically. Central to this analysis is the role the state plays as a disciplinary agent – enforcing behaviour on producers (and sometimes consumers) by setting the rules of the market game. Davidson’s work is particularly helpful here in identifying some of the ways marketisation can go wrong, which often involve competitive norms being too ruthlessly enforced (undermining other pro-social norms among non-profit providers) or allowing greater differentiation (of price and content) leading to a two-tiered model of provision (Reference Davidson, Meagher and KingDavidson, 2009). As an initial step towards developing an understanding of marketisation as a contested terrain, Figure 1 sets out several aspects of different social policy markets currently in operation in Australia. It does not include the details provided by Davidson, as its purpose is more conceptual. It identifies the steps involved in establishing market relations, namely, those that establish a market between the state and producers, those that then extend this market relationship to be primarily driven by consumers, and finally a market that allows intermediaries between the consumer and the service they are wishing to purchase by allowing market-based insurance products. Understanding the different types of market within social policy highlights how marketisation opens up new possibilities for less equitable and efficient policy outcomes.

Figure 1. Modes of marketisation.

Establishing market relations has now become central to the entire operation of the state. New Public Management has seen senior public servants employed via market-style contractual arrangements (Reference ConsidineConsidine, 2001; Reference Ramia and CarneyRamia and Carney, 2001). Most departments are funded based on a set of outcomes and deliverables. This does not necessarily mean explicit competition. It can simply be a mechanism for much closer monitoring of both public and community sector workers and producers. However, in most areas of public provision, there is increasingly some element of competition. Initially, this often comes through a tendering process that forces providers to compete for funding. At state and national level, ‘competitive tendering and contracting’ is now common (Reference Davidson, Meagher and KingDavidson, 2009: 53). This signals the beginning of the state using monopsony power to discipline producers. It sets the criteria for allocating funding, not through negotiation or direction of employees alone but by promoting competition.

The creation of market relations between producers and the state is taken further when ongoing payments are linked to competition through payment for individual outcomes or services. There are a multitude of forms of these managed markets. Payment for service can be combined with a tendering process, as with the Job Network, or linked to licensing that dictates minimum standards, as with Medicare provider numbers or childcare. These differences are important for shaping the nature of competition, although, for our purposes, the focus is on how governments discipline producers. So long as the government provides all or almost all of the spending power, and does so directly to the provider, the structure of the market is more likely to be determined by the rules set by the state – what services it funds, what rules it sets and what prices it pays.

Competition may rearrange the identities of the producers according to consumer choices, but this process of competition will usually reinforce the underlying structure provided by the payments. This can create serious issues for some disadvantaged groups. The problems of ‘creaming’ – taking on less difficult clients and excluding or not adequately servicing those that are more difficult – have been particularly pronounced in the Job Network (where for-profit and non-profit providers alike have engaged in creaming) (Reference Marsh and Spies-ButcherMarsh and Spies-Butcher, 2009; Productivity Commission, 2002).

The problem of unequal access to services, however, becomes more widespread once consumers are able to supplement public payment for services. Importantly, as Figure 1 suggests, this is part of a broader process of marketisation. Just as corporatisation is needed to allow public entities to be privatised, so too social services must be organised on the basis of payment for service or outcome to then allow individual consumers to directly engage in the market. This occurs now in childcare and health services, where government funding subsidises a consumer choice, rather than paying the provider directly.

Individual consumer purchasing substantially undermines monopsony power. The forms of competition promoted by a system where consumers are able to bid up prices are likely to be less conducive to improved outcomes without other forms of government regulation. This has been the experience in early child education (Reference Whit and FriendlyWhit and Friendly, 2012). It is a particular concern where there are incentives for consumers to pay more to get a particular kind of service – whether one that is better at meeting a direct need (as in a more skilled surgeon) or one that has status benefits (as in a private school). Because many areas of social provision have some aspect of this kind of product differentiation, allowing consumers to pay more often leads to more diverse and less equitable provision (see Reference Spies-ButcherSpies-Butcher, 2008). In medicine, asymmetries of information reinforce these problems. Here consumers explicitly seek advice from producers about what they are to purchase (Reference McAuleyMcAuley, 2005).

Incentive and informational problems become more complex when rather than involving the direct purchase of services, they instead involve purchasing insurance. Insurance can generally provide consumers with incentives to overclaim, but in health insurance, the structure of the market has exacerbated this tendency. Many funds have been captured by producers (doctors) and so have exercised little price discipline (Reference Scotton and DeebleScotton and Deeble, 1968: 9). In superannuation, the problem is slightly different. The public subsidy is for a product (pension) that is relatively distant in time and provided indirectly by subsiding a managed investment fund. Here, low returns are also symptomatic of the funds, reflecting the interests of parts of the finance sector (APRA, 2012: 6–7). The shift from defined benefit to defined contribution schemes also signals the strength of the financial sector to shift risks from the insurer onto the individual worker – a process seemingly systemic to a financialised system of social provision (Reference HackerHacker, 2006; Reference Stebbing, Meagher and GoodwinStebbing, in press).

As the forms of marketisation shift towards the right-hand side of Figure 1, not only are there more significant challenges in managing the incentives of markets to ensure resources are allocated efficiently, these forms of markets also open up greater opportunities for inequity. If public support is offered through a subsidy to private spending – as it is for childcare and private health insurance, and in a different way education – then the level of subsidy is likely to increase with a consumer’s spending power. If government subsidies are supporting private spending, it is also much easier to provide that support through the tax system (as a tax expenditure) rather than directly (through social spending). This is exactly the history of superannuation and private health insurance subsidies (Reference Spies-ButcherStebbing and Spies-Butcher, 2010). Providing support through the tax system compounds inequalities by relating the rate of subsidy to a person’s marginal tax rate. This occurs in superannuation, but also housing, and has led to enormous inequities. Thus, while direct social spending is highly targeted, a second set of subsidies that are outside the normal budgetary framework and are highly inequitable has developed. The result is a ‘dual welfare state’ (Reference Spies-ButcherStebbing and Spies-Butcher, 2010).

The politics of markets

If marketisation can produce highly inequitable outcomes, then a central question is whether accepting elements of neoliberal politics entails accepting these broader inequities. If not, we might sensibly discuss a hybrid Third Way model that embraces aspects of neoliberalism while retaining equity through its expansion of ‘affluence tested’ welfare. There is clear evidence of Labor attempting to do this since the Global Financial Crisis. Indeed, it has used the fiscal impact of the crisis to enforce targeted norms in a range of areas. In areas of direct social spending, Labor has used tighter means-tests to help fund increases in benefits for lower income groups, as with both family payments and the aged pension. It has also limited government support for the ‘affluence targeted’ aspects of the dual welfare state, imposing means-tests on private health insurance and reducing tax expenditures from superannuation for those with very high incomes (Reference Wilson, Spies-Butcher and StebbingWilson et al., 2013).

It is important to acknowledge this trend towards targeting within social policy, although the affluence-testing model appears to be fraying in two important respects. The first is its ability to protect the most vulnerable members of society. If targeting is to be an effective form of social protection, it should provide protection for those on the margins. The second is the inability, even after a series of hard-fought reforms, to make any real dent on the system of dual welfare. Indeed, other Labor initiatives, such as increasing the rate of compulsory superannuation contributions (and the childcare rebate), will actively expand this form of support into the future (Treasury, 2010, Appendix E). Labor’s acceptance of market forms of social provision seems to be susceptible to less equitable forms of marketisation, and once these are established, it is much less effective at changing this underlying dynamic. Thus, there is a potential corrosion of ‘affluence tested’ welfare in opposing directions, towards stingy forms of ‘means-testing’ and towards inequitable forms of ‘affluence targeting’ (see Figure 2).

Figure 2. Modes of social provision.

Reference WacquantLoic Wacquant’s (2012) recent exposition of the connections between ‘actually existing neoliberalism’ and increasingly disciplinary forms of social policy provides a framework for understanding how the first part of this connection might operate in practice. Wacquant claims neoliberalism is primarily a political project for reengineering the state and identifies the acceptance of market forms and commodification as an integral part of this project. However, in Australia, at least initially, the extension of markets was not always associated with penal or paternalistic forms of social control. Indeed, market restructuring took place alongside increases in parenting payments designed specifically to support single parents and advanced through a discourse of poverty prevention (Reference BrennanBrennan, 1998 [1994]). Given the somewhat extreme example offered by the United States, on which much of Wacquant’s reasoning is based, how might this argument be articulated in Australia?

There clearly has been an expansion of paternalistic welfare in Australia. While the initial steps towards labour market activation policies occurred under Labor’s Working Nation, the most onerous forms were introduced by the Coalition (Reference McDonald and ChenowethMcDonald and Chenoweth, 2006). Starting with very young unemployed persons, the Coalition introduced work-for-the-dole, a scheme that has gradually been extended. It then began to expand the scope of this paternalistic form of welfare by shifting those on less onerous (and more generous) payments into the unemployment system. This included some on disability support pensions and some of those on parenting payments (Reference Wilson, Meagher and HermesWilson et al., 2012: 328–329). Finally, in the lead up to 2007 election, the Howard Government introduced a form of extreme paternalism as part of the Northern Territory Emergency Intervention, using the army to enforce bans on alcohol and pornography, and introducing compulsory income quarantining, which prevented income recipients spending their payments as they wished (Reference Altman and HinksonAltman and Hinkson, 2007; Reference MendesMendes, 2012).

Importantly, these changes were extended by an incoming Labor government. Labor shifted a second group of parents into the unemployment system (Reference Miletic and HarrisonMiletic and Harrison, 2012) and extended income quarantining across the Northern Territory and then into other disadvantaged communities (Reference MendesMendes, 2012: 3–6). With single parents, this has partly been the result of Labor’s commitment to fiscal restraint. Following the trend set by the Hawke Government’s ‘trilogy’ commitment to reduce the budget deficit and taxation as a proportion of gross domestic product and to keep the rate of public spending growth under that of the economy (Reference StilwellStilwell, 1986: 15, 117), Labor again committed itself to restrained tax growth and balanced budgets (Parliament of Australia, 2012: 9840). Confronted with different groups of parents on different payments, Labor determined to shift more parents into the unemployment system, reducing their payments and making them subject to new forms of mutual obligation.

It appears that the politics of affluence-tested welfare are constrained by the political resources of the targeted groups. This is much clearer in the case of income quarantining. While there is a small number of high-profile Indigenous advocates of these policies, there has been widespread opposition from many of the Indigenous communities affected (see Reference CoxCox, 2011, Appendix B). Yet rather than being moderated, the policy has been entrenched and extended. The new subjects of this regime are among the most socially marginalised – those on benefits in highly disadvantaged communities. Given that this disadvantage is itself related to changes in labour markets and family composition that have made stable employment more difficult to access, the dynamics of the process are clearly related to the broader acceptance of both competition and fiscal constraints by both major political parties.

How different groups are treated – and moved through – the different forms of social provision reflects their political bargaining power. As single parents were being moved into the workfare system, and having their benefits reduced in the process, aged pensioners were having their payments increased. This reflects a broader shift in the welfare state to better support the aged (Reference Tapper, Fenna and PhillimoreTapper et al., 2013). Likewise, in negotiating changes to schools funding, the Labor government agreed to supplement funding for private schools to ensure no loss of funding (Reference TopsfieldTopsfield, 2013). However, in New South Wales (NSW), the Coalition government appears only to have capped the extent of money lost by some of the most disadvantaged public schools (Reference ToveyTovey, 2013). For a reform aimed at educational equity, it is ironic that disadvantaged public schools now stand as potential losers.

The second dynamic eroding ‘affluence testing’ is more partisan, although the nature of political contestation has changed. In a number of areas, the Rudd and Gillard Labor governments wound back payments where the Howard government had extended them. This was most notable around private health insurance – where the Coalition introduced and extended rebates and Labor then ‘affluence tested’ those rebates – and around superannuation, where the Coalition extended tax concessions and Labor introduced some limitations. Thus, the policy contest has not been about expansion of social provision per se, but about the nature of that provision, with the Coalition expanding elements of ‘dual welfare’ and Labor applying forms of ‘affluence testing’ (see Reference Wilson, Spies-Butcher and StebbingWilson et al., 2013: 365–366). Even so, Labor’s modest reforms have been fiercely resisted and have failed to claw back earlier changes. Superannuation tax concessions are more regressive after the Labor government’s changes than before the Howard–Costello reforms. Private health rebates are more generous after Labor’s changes than when introduced by the Coalition.

This is consistent with Reference PiersonPierson’s (1995) observation that once established, new constituencies and institutional architecture defend the status quo. If this is the case, then the affluence targeting model is producing a potentially corrosive political dynamic. It is facilitating the development of new welfare constituencies who favour private provision, comprising the most vocal, organised and powerful members of the community on the one hand, and a series of highly marginalised social groups incapable of effectively resisting paternalistic interventions, on the other. This is not incidental to the Third Way model of marketised provision. After all, forms of affluence targeted provision are easiest where consumers purchase social services themselves. As Figure 1 suggests, when services become marketised, it becomes easier to expand subsidies for private purchase, while retaining some form of direct public provision for the less well-off. Marketisation provides the institutional architecture necessary for a dual welfare state.

Targeting welfare in a dual system creates political dangers. Reference Korpi and PalmeKorpi and Palme (1998) famously argued that targeting created a paradox, as the middle class came to identify their interests as lying in minimising tax rather than expanding social provision. This logic is likely to be stronger in a dual welfare state. Not only are the middle class likely to become resistant to taxes they see funding marginalised groups, they are likely to become fierce defenders of an alternative set of privatised social spending. Where this is the case, formerly affluence-tested payments may erode over time, as has happened with Newstart, until they effectively become ‘means tested’. Once this has occurred, mobilising the political resources necessary to increase payments becomes much harder, and the prospect of payments becoming dependent on paternalistic conditions becomes much higher.

We are led to a pessimistic conclusion. Australia has avoided some of the most damaging aspects of radical economic restructuring and has even extended social spending. This reflects a continued commitment to addressing inequality. However, the longer term prospects are less promising. The expansion of marketisation has not always been antithetical to egalitarianism. Yet its effects have been increasingly inegalitarian, as it establishes a political dynamic that leaves those marginalised from the labour market vulnerable to paternalism, while consolidating the interests of the more affluent. ‘Affluence testing’ measures may limit the extent of inequality thus produced in the short run, but they leave in place the political dynamic that is gradually expanding a more corrosive dual welfare logic.