The Lost and the New ‘Liberal World’ of Welfare Capitalism: A Critical Assessment of Gøsta Esping-Andersen’s The Three Worlds of Welfare Capitalism a Quarter Century Later

Celebrating the 25th birthday of Gøsta Esping-Andersen’s seminal book The Three Worlds of Welfare Capitalism (1990), this article looks back at the old ‘liberal world’ and examines the new. In so doing, it contributes to debates and the literature on liberal welfare state development in three main ways. First, it considers the concept of ‘liberalism’ and liberal ideas about welfare provision contained within Three Worlds . Here we are also interested in how liberal thought has conceptualised the (welfare) state, and the class-mobilisation theory of welfare-state development. Second, the article elaborates on ‘neo-’liberal social reforms and current welfare arrangements in the English-speaking democracies and their welfare states. Finally, it considers the extent to which the English-speaking world of welfare capitalism is still meaningfully ‘liberal’ and coherent today.


Introduction
Esping-Andersen's The Three Worlds of Welfare Capitalism (Three Worlds hereafter) has transformed and inspired social research for a quarter of a century.Some contributions have confirmed or developed his typology, while others have challenged it from substantive and/or methodological perspectives.This article contributes to recent debates about Three Worlds by providing an analytical account of the 'liberal' conception of the role and function of the (welfare) state, including that portrayed by Esping-Andersen (1990) himself in Three Worlds, and the subsequent debates that have ensued over the liberal model of welfare capitalism and the classification and positionality of countries that might be said to resemble or form the liberal world (material not covered in the 25th year anniversary reviews, Emmenegger et al., 2015;Powell, 2015).
The next section examines conceptions of 'liberalism' and liberal ideas about welfare provision contained in Three Worlds.The third section turns to challenges and critiques of Esping-Andersen's liberal model, particularly from specialists in Anglo-American countries and in light of intriguing results from the 'welfare modelling business' (Powell and Barrientos, 2011), which seeks to identify welfare state clusters, and with new and competing typologies being discovered all the time, business is booming and shows no signs of abating (Barrientos, 2015).In the forth section, we consider the development of twenty-first century 'neoliberal' reason, and post-welfare state transformations in the English-speaking democracies with the rise of regulatory labour market instruments and market reforms.The final section offers reflections and considers the extent to which the liberal world Esping-Andersen described in 1990 is prescient for that same world we find today.

Liberalism in Three Worlds
'Liberalism', as Kelly (2004) observes, is a deeply contested notion, not least among liberal thinkers (certain major figures command attention, for instance: Locke, the father of classical liberalism, J. S. Mill, Ludwig von Mises, Friedrich Hayek, Michael Oakeshott, Milton Friedman, James M. Buchanan, John Rawls, Murray Rothbard, Ronald Dworkin and Robert Nozick).We find different schools of thought and competing theories about how society should be (from classical liberalism, ordoliberalism or social liberalism and the so-called Chicago School, often associated with neoliberalism).There is also a variety of different national stories told about liberalism in different countries that often defies traditional notions of left and right, from 'Thatcherism' in the UK and 'Reaganism' in the US, to Paul Keating's 'economic rationalism' in Australia and 'Rogernomics' in New Zealand (for example, liberalism advanced by social-democratic leaders).Each of these histories emphasises different concepts, key thinkers and political developments that enhance our understandings of political traditions and arguments (Gray, 1993;Kymlicka, 2002).
Clearly, liberalism is a multi-faceted ideology with wide-ranging implications for society.It usually has individualism, liberty and equality as central guiding principles, while most people who call themselves 'liberals' agree that there is no end state (this is where liberalism differs from social democracy).In the liberal state, the rule of law provides a framework for the pursuit of private ends (individuals should be left to pursue their own goals and purposes).As far as society is concerned, there are no overall goals or purposes. 1That said, the liberal conception of limited government may encompass something akin to a 'welfare state', but political philosophy and liberal thinking are not at all clear what this should look like (Gray, 1986(Gray, , 1989)).In other words, we cannot be at all clear about the 'ideal' and, arguably, the ideal type of 'liberal welfare state' is not something that exists or could exist.In Three Worlds, the liberal model is a reification, pursued for comparative policy purposes.The validity of this construct has been the subject of major debate since the publication of Three Worlds.
The central thesis of Three Worlds is that working-class actions, through the instruments of the democratic process, can modify capitalism to empower workers.Social rights are the result of collective political demands.With the emphasis on the scope and nature of cross-national differences, the Three Worlds typology then helps us to understand the importance of political ideas and causal forces behind 'welfare state' development (and for the thinking about the 'welfare state ' concept itself, cf. Veit-Wilson, 2000). 2 According to Esping-Andersen's approach (building on power resources theory and working-class mobilisation, Korpi, 1983), cross-national welfare state variations are the result of crossclass coalitions and the differing capacity of labour movements to organise and mobilise politically.A weak left class coalition prevented the development of the welfare state in Anglo-American countries, a strong left class coalition led to a social democratic welfare state with universalistic social policies in the Nordic countries (Esping-Andersen, 1992).Some scholars have attempted variations on this (overly monocausal) theme.Francis Castles, for example, points to a weak right being more important to successful The Lost and the New 'Liberal World' of Welfare Capitalism welfare state development than a strong left (Castles, 1978: 131-40).There is also an argument that Three Worlds understated the importance of middle-class actions and political representatives in its historical analysis, with the right behaving opportunistically to secure various sorts of social policy provision as Peter Baldwin (1990) has shown.Middle-class support tended to appear when the middle classes believed that exclusion from social policies was no longer acceptable.Weakness and strength are, of course, relative terms, and a 'coalition' here may resemble a temporary alliance for combined political action.The extent to which 'solidaristic' welfare policies were the result of middle/working class coalitions (Esping-Andersen) and/or the result of middle classes and political parties of the right behaving opportunistically (Baldwin) is an important issue but it goes beyond the scope of this review.Peter Baldwin may very well have underplayed the role of left parties just as Esping-Andersen may have overplayed them.
Different welfare states, the 'real types', exemplified or approximated to 'ideal types' (in the Weberian sense) of theoretically constructed welfare states associated with different political movements of the 20th century.Welfare states are re-specified as 'welfare regimes' by Esping-Andersen, defined in term of the relations between politics and markets, or perhaps more appropriately, between state and economy.More specifically, the regime concept is used to describe and differentiate relations between society's institutions, the state, market and family, that interact to produce the peculiar publicprivate sector mix found to exist within the three distinct worlds of welfare capitalism: social democratic, conservative and liberal (an upper-case form would indicate political parties or other formal organisations, lower-case is used to indicate far broader sets of political ideas and policy approaches).Esping-Andersen's three-way welfare regime 'typology' functioned as a classification device to classify real cases along theoretically relevant dimensions (here: decommodification, stratification and the public-private mix of social provision).In sum, the three regimes or worlds of welfare consist of clusters of countries that, despite their differences, share major similarities.
Circumscribed by capitalist power, the essence of the liberal outlook to welfare (which builds on Titmuss' (1958) 'residual' or social-assistance model of welfare) is thus described: In one cluster we find the 'liberal' welfare state, in which means-tested assistance, modest universal transfers, or modest social-insurance plans predominate.Benefits cater mainly to a clientele of low-income, usually working-class, state dependents.In this model, the progress of social reform has been severely circumscribed by traditional, liberal work-ethic norms: it is one where the limits of welfare equal the marginal propensity to opt for welfare instead of work.Entitlement rules are therefore strict and often associated with stigma; benefits are typically modest.In turn, the state encourages the market, either passively -by guaranteeing only a minimum -or actively -by subsidizing private welfare schemes.(Esping-Andersen, 1990: 26) Liberalisms preference is clearly for private insurance in the market, the work-centred ideology and commodity logic are supreme, and social stratification in society is largely determined by competitive individualism in the market.Esping-Andersen defined the liberal world in this way in order to draw attention to the 'big picture' in Three Worlds, as Béland (2010: 47) reminds us.Countries belonging to the liberal world (with low levels of decommodification and high levels of market stratification) include the Englishspeaking democracies: Australia, Canada, New Zealand, the UK and the US (although Notes: a Esping- Andersen (1990: 52).This score is a measure of the extent to which the welfare state neutralises the effects of the market through its welfare policies.b Esping- Andersen (1990: 74).This score indexes the extent to which welfare state attributes follow the principles of liberalism, socialism and conservativism; the latter is not shown in the table.c Ireland is ambiguous, being in the liberal category based on its low decommodification score but with medium scores on Esping-Andersen's conservatism index.Similarly, Canada and Switzerland had ambiguous arrangements according to Esping-Andersen's classification system.Source: Esping-Andersen (1990), Tables 2.2 and 3.3.
Ireland is ambiguous, see Table 1).Hence, the liberal cluster is strongly associated with the Anglo-American 'liberal tradition' of political philosophy.
In practice, then, the liberal welfare state is typically defined as a minimum welfare state, or a safety net of services for the poor population, as told in Three Worlds (1990: 41-4), but, as we shall see in the next section, the English-speaking democracies continue to use a variety of instruments and expenditures to correct for disadvantageous market outcomes.

Liberal welfare regime debates
For Esping-Andersen (1990), the theorisation of the social democratic and conservative worlds is always clearer than the liberal one, which is more widely debated.Many critics claimed Esping-Andersen misunderstood the development of social policy in the English-speaking nations with his European-centred theory, and that he over-estimated the congruence between linguistic and cultural heritages and accompanying welfare arrangements in Anglo-American nations.Certainly, English-speaking democracies were social policy innovators in different ways during the course of the twentieth century for example, as a range of country specialists have been prepared to testify.
In another celebrated work of comparative social policy, The Working Class and Welfare, published just five years earlier, Francis Castles also developed the The Lost and the New 'Liberal World' of Welfare Capitalism class-mobilisation theory (the original thesis from 1985 was revisited and developed many times, for example, Castles, 1994).The outcome of cross-class coalitions in the Antipodes was not welfare provision through the state as seen in (north) European social policy, but an attempt at a broader category of 'welfare society' (of which 'welfare states' may be but variants).The answer to the problem of inequality in capitalist society lay in the regulation of the wages structure and the notion of a legally guaranteed 'living wage' for unskilled labour, the welfare state remained a second order priority.The Antipodean model of 'welfare society' in the post-war period in question was essentially 'social democratic', with its emphasis on egalitarianism and wage-earner rights.Castles' original thesis (virtually ignored in the tale told in Three Worlds) points to working-class action and coalition strategies with middle-class support influencing the path of Australasian social policy during the twentieth century (Deeming, 2013).Social democratic efforts focused on the primary distribution of income generated by the capitalist market economy to secure socially protective social policy, rather than, say, the redistributive mechanisms at the command of the British 'welfare state' in Europe.Asa Briggs (1961), for example, speaks of the 'liberating effects of legislation' introduced by the Labour government following the Beveridge Report (Cmd 6404, 1942).Clearly, then, it might be argued, for example, that much of the English-speaking world during the post-World War II period should not be considered 'liberal' in any theoretical or classical sense, if the state was meant to be minimalist or government limited (see also Ann Shola Orloff's (1988) account of the political development of the welfare state in North America).Reality rarely conforms to the idealised pattern of the liberal model, and liberal principles rarely apply across all social programmes in market liberal societies.The British NHS provides another example of governmental intervention based on social democratic founding principles of universality and equity.Thus, in the field of healthcare, the UK, which belongs to the liberal welfare regime in Three Worlds, is more similar to social democratic Sweden than it is to the US, another member of the liberal welfare regime.
The suggestion that the UK is best described as a 'liberal welfare regime' is particularly problematic, given that there have been lengthy periods of what might be described as 'social democratic' government, which, in theory at least, might give rise to a conflicting set of policy prescriptions, as in the case of the New Labour government under Tony Blair (Powell, 2008).The same is true of the Australian Labor government in the Rudd-Gillard era (Deeming, 2014).Given the evidence for encompassing welfare functions, these nations could conceivably be considered 'non-liberal' and the fit with the 'liberal' model presented in Three Worlds would certainly be questionable.The US (often considered a 'welfare laggard') also provides another example.In the US, public social spending (i.e., the monetary size of government welfare) remains relatively low, but has been rising.All told, the US government now spends one-fifth of GDP, more than ever, on social welfare.However, this depiction of a nation's 'welfare state effort' misses important and well-established information.Christopher Howard (1997) points to a 'hidden welfare state' in North America.According to Howard (2008), policy tools such as 'tax credits' and 'tax expenditures' serve the same or similar social welfare functions found in other welfare states. 3With this expansion of welfare capitalism, North America can no longer be regarded as exceptional perhaps in the manner it was over a quarter of a century ago when Three Worlds was published.
Critics also questioned the internal coherence of the liberal regime visà-vis the positionality of the English-speaking nations within Three Worlds.One of the earliest challenges came from Francis Castles and Deborah Mitchell in a series of works (1990,1992,1993).According to them, Australia, New Zealand and the UK were 'misclassified' in Three Worlds (Castles and Mitchell, 1993: 106).These countries belonged to a fourth 'radical' category of welfare capitalism within the alternative 'families of nations' classification posited by Castles and Mitchell, based on their index of benefit equality and welfare expenditure.With strong labour movements, countries like Australia and New Zealand focused their redistributive efforts through wage-fixing instruments, rather than high levels of welfare expenditure.Australia, for example, has high benefit equality, and payments are not time-limited.Interestingly, Esping-Andersen was later persuaded by the case made for Australasian exceptionalism (Esping-Andersen, 1997), and indeed for 18 distinct worlds of welfare capitalism (Esping-Andersen, 1993: 136), before the restatement of the three worlds typology (Esping-Andersen, 1999: 94).Some of the typological disagreement gently subsided, however, when high-profile economic reforms during the 1990s signalled the end of the Keynesian chapter and the wage-earning model of welfare, according to Castles (2001), and Australia and New Zealand came to be defined more in terms of their family resemblance to other English-speaking nations (Castles and Obinger, 2008).Nevertheless, this 'radical' positioning of Australasian social policy has attracted many followers over the years (Deeming, 2013, provides a review), who also agree that the Australian model of welfare cannot easily be accommodated within the original three-world typology presented by Esping-Andersen.As Peter Whiteford recently observed: the characterization of Australia as a radical welfare state is apt.Australia is the strongest example of a country using the 'Robin Hood' approach to the Welfare State, relying more heavily on income-testing and directing a higher share of benefits to lower-income groups than any other OECD country.(Whiteford, 2013: 65, emphasis added) At the same time, there is an ongoing debate in the literature over the number of 'worlds' and the types of countries that might (or might not) cluster within them, as predicted or otherwise, at a particular point in time (Ferragina and Seeleib-Kaiser, 2011;van Kersbergen and Vis, 2015).Typologies and country arrangements are largely dependent on normative judgements, as Danforth (2014) illustrates (to do with the selection of countries, methodological techniques and the assemblage of macro-level indicators gathered together).One such example, based on 'new social risk' indicators (for example, the number of young people in education and female labour market participation rates), suggests that Austria and the UK now form the liberal group of nations (see Ferragina et al., 2015, for their proposed four worlds of 'welfare reality'). 4Another attempt to include both production and welfare regimes (drawing in Varieties of Capitalism (VoC)), broadly confirms Esping-Andersen's original rendering (Schr öder, 2013).Thus, Australia, Canada, New Zealand, Ireland, the UK and the US belong to the liberal market economy/liberal welfare regime in Schr öder's analysis.Classifying countries in this fashion, with an appeal to 'ideal types' of welfare and market society, is a way of mapping the messy terrain of social policy in comparative research.However, as these models are not in any way predictive, there is no real test of the model and we are left wondering whether they reveal anything more than we already know.We might also question the continued relevance of ideal-typical concepts in this field too.
The Lost and the New 'Liberal World' of Welfare Capitalism Straightforward resolution of such a debate, over the liberal model of welfare capitalism and the degree of fit between the English-speaking democracies and their welfare states, is not easy, nor is it confined to history.The debate over the presumed 'liberalism' has again resurfaced, as scholars have turned their attention to the developmental trajectory of the welfare state in the English-speaking nations.They are now asking whether governmental policy (based on ideas about the supremacy of free markets, the privatisation of public services, deregulation and the efficiency of market forces) might more accurately be described as 'neoliberal' capitalism (for example, Higgins, 2014).We return to this issue next.For now, let us accept that 'neoliberalism' is another fiercely contested concept in the social sciences (cf.Dardot and Laval, 2014), one which was barely a topic for discussion in Three Worlds.

Reforms since Three Worlds
In this section, we are concerned with social policy developments in the English-speaking nations since the periodisation of Three Worlds (i.e., post-1980s). 5The issue we are concerned with here is whether, or the extent to which, the liberal model of welfare described in Three Worlds still resembles features of liberal welfare state formation today.In other words, when we are speaking of a liberal welfare state today, are we using the same category or notion as the one introduced by Esping-Andersen in 1990, and, if not, in what ways has the welfare state in the English-speaking democracies changed?Has political liberalism and its concretisation in welfare provision changed over time and in which ways?
Radical reforms were ushered across the English-speaking democracies during the 1980s and 1990s, as welfare states began to adapt to the new social and economic conditions of the post-industrial era.The effect of the limited or minimalist 'liberal' welfare state in Anglo-American nations was to distance people away from the world of work, or so the emerging 'neoliberal' critique of the welfare state would claim (Peck, 2001, provides an overview).It is also worth noting here that 'neoliberalism', as a social policy concept, does not really feature in Three Worlds; it is mentioned, but only twice (pages 9 and 42), and it is not indexed.As Figure 1 suggests, 'neoliberalism' was not really a concept of the social sciences in the late-1980s when Three Worlds was written.However, interest in the language stream of 'neoliberalism' has gained strength in the last two decades (Figure 1), particularly with key texts developing the concept in the fields of sociology (e.g., Dean, 2014;Gane, 2014), human geography (e.g., Harvey, 2005;Peck, 2010) and political philosophy (e.g., Plant, 2012).Social policy has yet to subject the concept to such critical scrutiny (Béland and Petersen, 2014).
Policy prescriptions arising out of the 'dependency diagnosis' also began to gather shape during the late-1980s.Influential neoconservative thinkers such as Lawrence Mead began to argue that modern welfare systems had become too oriented to rights and entitlements (Mead, 1986).Although Mead accepted the collective obligation to provide a social minimum, as Hayek makes clear in The Constitution of Liberty (1960), welfare services must now become conditional on the discharge of obligations.The other more libertarian alternative policy prescription proposed at this time came from Charles Murray (1988), who argued for the dismantling of the welfare state, as Nozick had done before him in Anarchy, State, and Utopia (1974).In that case, welfare goods and services are just like other goods traded in a market.A removal of the welfare safety net would also force benefit recipients to rely on resources given by family, church and the charitable sector.We might think of David Cameron's rhetoric on the 'big society' in this context, for example.We expand on these neoliberal conceptions of the role and functions of the state below.

Activation and the workfare state
In the English-speaking countries at least, the 'welfare state' appears to have been transformed into a 'workfare state'.Although 'Work for the Dole' schemes can be traced back to the post-war period in the US (Fox Piven and Cloward, 1972), they have gained considerable momentum in the English-speaking democracies since the publication of Three Worlds (Robert Walker describes this as the 'Americanization' of social policy, Walker, 1999).In the modern-day context, workfarist regimes regulate labour, in essence by enforcing 'Work for the Dole' and low-waged and insecure work more generally (characteristic of the post-industrial labour market) while residualising state welfare.The strong libertarian arguments about a welfare-dependent 'underclass' made by Lawrence Mead and Charles Murray during the 1980s and 1990s appealed to economic liberals and parties on the right (Mead, 1986(Mead, , 1997;;Murray, 1990). 6It was claimed that the post-war liberal welfare state offered support, in the form of 'unconditional' benefits but expected little in return.Accordingly, 'Work for the Dole' is seen as perfectly and morally justifiable as the social contribution now owed to society by those receiving out-of-work benefits.Conditionality thus restores their right to equal citizenship status because, as Mead has argued, 'only those who bear obligations can truly appropriate their rights ' (1986: 257).
The US was the first of the English-speaking nations to introduce the new wave of workfare measures enacted under Democratic Party President Bill Clinton, although similar reforms soon gained purchase in the UK, Australia, New Zealand, Canada and to a lesser degree in Ireland.Clinton's election promise 'to end welfare as we know it' set the stage in 1992.Then in August 1996, the US Congress passed the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA).From this point, workfare schemes in which state welfare benefits are dependent on recipients accepting attendant statedefined work became a central feature of US social policy (King, 1999;Peck, 2001).Work requirements in section 407 of the 1996 welfare-reform law were subsequently strengthened under the Republican administration of George W. Bush (Deficit Reduction Act of 2005), only to be reversed again by the Obama administration in July 2012.
In the UK, for example, Margaret Thatcher introduced 'Work for the Dole' plus a £10 top-up in 1988, during her third term, based on the US workfare model.Although Tony Blair's 'New Deal' and 'welfare-to-work' policies (introduced in 1998) aimed for universal activation for all, Labour stopped short of US-style workfare programmes (Powell, 1999).Under the present Conservative administration led by David Cameron, however, new mandatory work conditions will apply to long-term unemployed claimants.Those receiving a Jobseeker's Allowance (JSA) will be required to give something back to their community under the Conservative's new workfare programme to maintain eligibility to benefit: making meals for older people, clearing up litter or working for a local charity are among the options currently being discussed (Deeming, 2015a).Similarly, in Australia, the Liberal-led coalition government under John Howard introduced 'Work for the Dole' legislation in 1998, which removed the provisions that prevented unemployed workers from being required to work for the dole.The Labor government (2007-13) strengthened activation in order to address 'welfare dependency' in Australia, although 'Work for the Dole' was no longer mandatory under Labor.However, all unemployed Australian residents were actively encouraged to participate in work-for-the-dole schemes.With the election of the Liberal-National coalition government in 2013, however, community workfare programmes are once again being revived and expanded (Deeming, 2015b).
In New Zealand, the National government, promising to deliver a 'Decent Society', also introduced 'Work for the Dole' schemes in March 1991, under the 'Community Task Force Scheme'.Workfare was ended by the fifth Labour-led government of New Zealand, which came into power in 1999 under Helen Clark.The Labour Party refused to accept such schemes, arguing that they undercut genuine jobs.However, New Zealand has subsequently adopted stronger workfare obligations to help tackle the problem of 'benefit dependency' with the election of the fifth National government in 2008 under John Key.Canada and Ireland have also both moved towards greater acceptance of activation and workfare.Canadian legislation in 1996 made workfare legal in Canada, which enabled provinces to develop workfare schemes as part of local 'welfare-to-work' programmes.Activation strategies varied; many provinces rejected mandatory workfare, while others such as Alberta and Ontario, influenced by developments in the US, implemented workfare programmes (Herd et al., 2009).Ireland also has its own distinctive style of social policy and Irish social security is still in the process of workfare reform, which is a mix of what Murphy (2012) calls 'supportive conditionality' and 'sensitive activation'.
This shift towards workfare policies in the English-speaking nations, that combine coercive paternalism and conditional entitlement effectively to reduce the rights of more vulnerable citizens, is not without its critics.Desmond King, for example, argues that punitive compulsory policies such as workfare are basically illiberal (King, 1999).In the context of this article, however, it is important to note that these ideas and experiences in the US strongly influenced welfare reform processes across the English-speaking world, as Walters (2000) observes.However, the puzzle offered by the turn to workfare is not simply a question of why the workfare state emerged, but why the left and right chose particular types of activation and workfare.Part of the answer lies in constrained partisan preferences.Because the left and right's long-term partisan goals differed, so did preferences over workfare and activation.The right aimed to enforce market discipline and social control with workfare, as Wacquant (2010) argues.The left, by contrast, embraced activation policies (which have a long tradition in social democratic countries, cf.Bonoli, 2010) in an effort to address public concerns about welfare state spending and efficiency (Gregg et al., 2009).

Making markets
Across the English-speaking democracies, the trend has been towards the increased use of markets in all areas of welfare.Today, we find market-oriented principles and the growing 'mixed economy' of welfare finance and provision in which hospitals compete for patients (McGregor, 2001), schools compete for pupils and universities for studentsincreasingly more of whom will be privately financed (Connell, 2013).In this way, liberal market society is strengthened.
During the 'Golden Age' of welfare capitalism depicted in Three Worlds, the state operated and directly financed public services.With the introduction of what might be termed 'quasi-markets' or 'internal markets' into the delivery of welfare services, the state stopped being both the funder and the provider of services (Le Grand and Bartlett, 1993).Under the new arrangements, competition in the delivery of welfare services by a plurality of independent agencies, voluntary and private providers replaced state monopolies across the English-speaking world.Providers now competed for business from state-appointed purchasers or directly from users, as new markets for private financing developed.The market (or 'quasi-market ' Bartlett et al., 1998) 'revolution' in social policy, as it became known, swept across the whole range of social services, including education, health, community care, pensions, social security and employment services.This revolution is still gaining momentum, with parks and prisons in the US (Fox Piven, 2015) and child protection services in the UK now being opened up to the market (Jones, 2015).The essential feature of professional judgement and trust, once the bedrock of the 'public service ethic' (or 'ethos'), is now regarded as arbitrary power by 'public-choice theorists', such as James M. Buchanan, in the liberal critique of interest group politics (Plant, 2012).In explaining the policy change, most commentators point to the explanatory role of constrained partisan preferences and goals, with the right using markets to reduce or limit the size of the state (Gingrich, 2014) and the left using quasi-markets and choice (the so-called personalisation agenda) to enhance welfare state legitimacy and efficiency (Le Grand, 2007).
The marketisation and privatisation of welfare services in the English-speaking democracies and their welfare states since Three Worlds are further reflected in the expansion of public-private partnerships, compulsory competitive tendering and contracting out.Today, there is also greater emphasis on privatising the cost of welfare with user charges, regressive tax subsidies and tax breaks for private pensions and healthcare, which increase the role of private actors in key policy areas (MacGregor, 2005).This shift towards self-responsibility and private policy arrangements (what Klein and Millar, 1995, call DIY social policy) has reinforced, rather than mitigated, existing forms of social inequality in the English-speaking nations, compared to Nordic and continental The Lost and the New 'Liberal World' of Welfare Capitalism European welfare states (T óth, 2014).Greater reliance on private arrangements in the English-speaking democracies and their welfare states can exacerbate not only class and gender divisions (O'Connor et al., 1999), but also racial and ethnic inequalities, as Banting and Kymlicka (2006) argue.

PRT and the strategic behaviour of political parties
Clearly, at the heart of Three Worlds was the question of the resources of groups (Power Resource Theory, or PRT) in the labour markets to make demands on the state via politics.Twenty-five years later, what have we discovered about these dynamics in the least hospitable places of welfare state capitalism?What has changed about the liberal model, and what has not in political terms?
The liberal model has always had strong conservative elements, as Iversen and Soskice (2015) observe.The puzzle now is why, in a world of growing inequality, the middle-class voter in the Anglo-American countries remains on the right (Vis and van Kersbergen, 2013).Certainly, 'Third Way' politics, promoted by the likes of Bill Clinton and Tony Blair, appears to have moved the 'traditional' left to the centre as left parties shifted their electoral base away from the shrinking working-class vote (Bonoli and Powell, 2004). 7 Welfare policies of the centre-left have become more similar to the centre-right, and this may, at least in part, explain the rightward drift (Arndt, 2013).The rightward shift has been accompanied by steep declines in union bargaining power in market liberal society, in contrast to the coordinated market economies in continental and northern Europe.With services replacing manufacturing and the decline of union density and the working-class vote, there perhaps is little left to counter the growing power of wealth in liberal market society, as Huber et al. (2015) argue.Lower taxes on the rich are just another aspect of neoliberal redistribution as Fox Piven (2015) observes. 8Similarly, Hacker and Pierson (2011) suggest that rising wealth inequality has led to 'off-centre' politics in the US (i.e., political action has 'abandoned' the middle class).
At the same time, however, welfare expanded during this period, particularly with Labour governments in office in Australia, New Zealand and the UK.Welfare systems were strengthened for more affluent middle-class families through the mechanisms of tax rebates and credits, as the US Democratic Party President Bill Clinton had done previously for low-to-moderate income workers in 1996.Thus, many dual-earner families in the English-speaking nations now receive 'welfare-in-work', including social benefits in the form of tax breaks and credits.Social policy is now more oriented towards meeting their needs and protecting their living standards.The range of initiatives implemented can be seen as forging a new de facto contract, whereby the state has now absorbed greater responsibility for complementing the market wage of middle-class families.The so-called 'hidden welfare state', which benefits middle-class voters, as Michelmore (2014) suggests, has reshaped politics in profound ways.The triumph of middle-class interests in capturing welfare state resources arguably represents a new political compromise (Gingrich and Häusermann, 2015).In response to new voter demands, centre-left parties in the English-speaking democracies have changed their political strategies to appeal to the growing class of white-collar workers, and no longer appear to promote the interests of manual workers.The 'new politics of the welfare state' literature, developed during the 1990s and early 2000s, also claims there is little further potential for social democracy to cater for working-class families in today's austere times.This also implies that class and partisan politics matter less for the development of social policy.However, the highly interventionist social investment perspective in Australia serves to disconfirm the theory of 'permanent austerity'; here political actors rejected such notions, as social policy was shaped in profound ways in the early part of the twenty-first century (Deeming and Smyth, 2015).
In the English-speaking democracies, at least, political gravity appears to have shifted firmly to the right during the last quarter of a century.There now appears to be a growing political consensus around the merits of markets to deliver public services, the (welfare) state is withdrawing and the 'neo-'liberal case for the 'workfare' model of welfare appears to be growing.Indeed, the present patterns of middle-class welfare (which arguably reduces the redistributive impact of the welfare state for the lower classes, see Goodin et al., 1989), coupled with the punitive and coercive state policies for the poor population, suggest a politics perhaps not presaged in Three Worlds.

Discussion and conclusion: the lost and the new liberal world
Three Worlds continues to invite reflection, a work in progress, rather than a definitive one.Half a century on therefore, this article has revisited the theorizing and typologizing of the 'liberal' world, asking how it advanced in Three Worlds and subsequent developments after.On the one hand, the review has questioned the enduring relevance of the (lost) 'liberal' world of welfare capitalism.There is no 'ideal' or typical liberal approach to a 'welfare state' in the English-speaking nations, as the literature makes clear, and we are still left wondering what this 'ideal' might look like a quarter-century on.The original portrayal of the liberal regime presented by Esping-Andersen did not address real-world social policy principles at play in the English-speaking democracies and their welfare states, even accepting his acknowledgment that there are no 'pure cases' here (1990: 28).Subsequently, we find little consensus amongst country specialists on the fit of realworld country examples with notions of liberal political idealism found in Three Worlds.Researchers using hierarchical cluster analysis provoke further disagreement in this field.They continue to classify different countries as 'liberal' using a variety of different indices and measures, as discussed above in the third section.
On the other hand, the review strongly suggests that the liberal world that Esping-Andersen described in 1990 was, in fact, prescient for the one we find today in the English-speaking nations and their welfare states.It is highly commodifying in many respects, and the state now enforces work (not just for the poor population and lower classes but other groups in society).The legitimacy of welfare markets, market institutions and market discipline has also increased over the past quarter of a century, as has the unchecked power of firms and gigantic corporations (Crouch, 2011).Liberalism's preference, as told in Three Worlds, is clearly for privately organised insurance in the market.From this perspective, the fact that Anglo-American nations are embracing marketbased welfare solutions to deliver more public services may not surprise at all.More state services are being privatised and, increasingly, private-sector solutions insure against risk.However, welfare markets are now combined with punitive regulatory practices and 'active' political programmes like 'workfare' (not welfare) in the Anglo-American nations.This development preceded Three Worlds and makes the 'neo-'liberal characterisation of political policy very persuasive, as Peck (2010) argues.This is because the notion that welfare rights are not genuine rights, like civil or political rights, is, according to Plant (2012), the key defining feature of the modern asocial neoliberal state.Does this analysis imply convergence?Has the liberal world become more unified than in the original 1990 rendering?This is arguably true, as the essential elements of the 'neo-'liberal state based on 'workfare' and welfare markets suggest.Although there is a common trend across the English-speaking democracies and their welfare states, we also observe distinctions and divergence among these nations.These trends emerged at slightly different times and in different political systems.The welfare state also expanded in some of the English-speaking nations during this period.New Labour expanded the British welfare state, as did the Rudd-Gillard Labor government in Australia, and not all of this expansion occurred through low-visibility programs (although much did).Equally, parts of the Canadian and Irish welfare state expanded, as has the US health care system.In some cases, as in the UK, there has been rapid and profound retrenchment in the public sector in recent years, under what one might describe as 'unforced austerity' pursed by the Conservative government as it attempts to embed a radical and divisive liberalism permanently in British public life.While in New Zealand there are backlashes against austerity cuts, in other states, such as Canada (which has just elected Justin Trudeau as Prime Minister), the cuts are much less severe (Taylor-Gooby, 2012).Does this mean that the Anglo-American nations have become a less heterogeneous entity -suggesting 'varieties of liberalism' or 'varieties of neoliberalism' (Connell and Dados, 2014) perhaps?However, if that is so, to what extent is it useful to think about ideal types at all?So much has changed politically over the last quarter-century, so much so that ideal types do not seem to help the study of social policy perhaps as much as they once did.Is it even worthwhile to attempt to refine regime theory at the country-level then; the analysis suggests it might be better to refocus on the cultural and historical antecedents of social policy programmes.As Rice (2013) observes, interest in programmatic regimes is growing (for example, pension regimes, social security regimes, healthcare regimes or workfamily regimes).This continues to preserve 'regime' as the analytical tool.The focus on country-level welfare regimes is often problematic, and this is certainly the case in some of the English-speaking democracies, particularly in federal (Castles and Uhr, 2007) or provincial welfare systems (Béland and Daigneault, 2015), and also in relation to the four countries that make up the UK (Mooney and Scott, 2012).
Since Three Worlds, it has become very clear that the promised road to equality (as well as to prosperity) in the English-speaking democracies and their welfare states (based on strong elements of individualism, free markets and opposition to state intervention, Spicker, 2013) is failing to deliver (cf.T óth, 2014).The belief that there would be a trickledown effect 'raising all boats', with little or no need for state intervention in the market, now appears dated (even the OECD, which has long promoted efficient market capitalism, now rejects trickle-down economics, instead encouraging nations to 'Go Social!' and grow 'Inclusive!').The Anglo-American countries now stand out in terms of the high and growing levels of inequality seen in the OECD group.The concern, at least among many social policy scholars, is that increasing financialisation and wealth accumulation, along with the privatisation and marketisation of risk and further residualisation of state welfare, will not counter current trends (MacGregor, 2005).There is also the question of whether the English-speaking world is unique in these trends; 'neoliberal' ideas appear to have undermined solidarity (Mau, 2015) and social democratic institutions across Europe in recent times (Lodemel and Moreira, 2014).It may be that no welfare regime is able to cope with the power of global capital, and once distinctive worlds of welfare may be collapsing or may have collapsed already.In such a case we would be forced to abandon regime theory altogether.

Figure 1 .
Figure 1.'Neoliberalism' as social science keyword Source: web of science search August 2015.

Table 1
Rank ordering of countries' political culture and institutions from liberal commodified to socialist decommodified