72 Varieties of Welfare Capitalism Table 4B. Welfare state incentives and disincentives for working mothers Child benefits as a Percentage benefit loss Percentage extra percentage of AWI to unemployed if marginal tax if (couple) spouse works wife works Year 1990" 199ib 1992Ľ Australia 2.5 100 0 Canada 3.6 0 0 UK 3.9 39 0 USA 2.2 19 0 Denmark 3.1 3 6 Finland 11.4 4 0 Norway 8.0 0 0 Sweden 4.3 0 0 Austria 7.4 56 — Belgium 10.2 10 9 France 5.6 0 3 Germany 4.6 10 5 Netherlands 4.8 0 14 Italy 0.7 — 0 Portugal 3.0 0 0 Spain 1.6 0 9 Japan 2.2 0 — ;l Estimated family benefits + tax relief as a percentage of couple's income (one at 100 per cent average production worker income, the other at 66 per cent). Source: Recalculations from Gauthier(1996, Table 10.1) and OECD (1995t:, Part 1). b Source: OECD (1995c, Table 8.1). c Increment in husband's marginal tax. rate if wife's income equals 66 per cent of husband's. Source: OECD (1995c). Comparative Welfare Regimes-Re-examined Following Richard Titmuss' (1958) pioneering contribution, there have been intense efforts devoted to welfare state classification. Typologies can be useful for at least three reasons. One, they allow for greater analytical ;' \) parsimony and help us see the forest rather than myriad trees. Two, if we can cluster various species according to similar crucial attributes, the ana- :. lyst can more easily identify some underlying connecting logjcjof movement and maybe even causality^ And three, typologies are helpful tools for [ "• ' generating„andtestinghypotheses. Typologies are problematic because parsimony is bought at the expense of nuance, but especially because they are inherently static. They provide -• a snapshot of the world at one point in time and dp not easily capture mutations or.the birth of new species. Any typology of welfare regimes therefore remains valid onlyas long as history stands still. Welfare state classifications mirror a particular epoch, in most cases the status quo of the 1970s and 1980s. The "three worlds' typology of regimes i; that underpinned my earlier work (Esping-Andersen, 1990) has been '''"' questioned on numerous grounds, and some provide compelling arguments for a major reconsideration. It was a typology too narrowly.based 'o..) -on incj^me-maintenance programmes, too focused on only the state- i/...■} marjětjigpis,, and tooone-dimensionally built around the standard male ■ < production worker. There are primarily two avenues of criticism that merit attention. One questions the simple triad, arguing that we should distinguish additional models—a 'fourth world' so to speak. Another questions the t>asic criteria -which were employed in trie construction of the typology. But before we proceed, therejs^one point of potential confusion that \ must be clarified. The bases for typology construction that I am here (as before) examining, are welfare regimes, riot welfare states nor individual <---socialpolicies. 'Regimes' refers lo the ways in which welfare production is allocated between state, market, and households. Some confusion may arise because the word "regime' is often applied 10 all kinds of phenomena: "poverty regimes^ 'pension regimes', or 'male bread-winner regimesr, just to mention a few. Some criticisms of "the three worlds' are, in a sense, irrelevant because they are not addressing welfare regimes but individual 74 Varieties of Welfare Capitalism programmes. Leibfried's (1992) argument that there is a distinct M-editeiTaneaiLmgimeJbLejiause social assistance in Southern Europe is unique may be well-taken. But here Leibfried misses the mark because he is studying a qualitatively different phenomenon. A similar problem pervades some feminist contributions, at least in so far as they have redefined the dependent variable. It is unquestionably relevant to compare 'breadwinner models', and it goes without saying that this has direct relevance for welfare regime comparisons but, again, a welfare regime typology does not stand or fall solely on one social policy dimension; and, again, 'breadwinner regimes' and 'welfare regimes5 are two distinct dependent variables. THE THREE WORLDS OF WELFARE CAPITALISM RE-EXAMINED The private-public mix was the principal analytical axis that underpinned the 'three worlds' typology; the key defining dimensions were degree of de-commodification and modes of stratificaüon or^if you wish, solidarities^ (Esping-Andersen, í990). The regime labels that represent this triad— liberal, conservative, and social democratic—derive from classical European political economy. They reflect the_,poHticaJ____an_dJdeological thrust that was dominant in their historical evolution, climaxing with the mature welfare states in the 1970s and 1980s. The typology, regardless of political origins, becomes static in the sense that it reflects the socio-economic conditions that prevailed then, namely an economy dominated by industrial mass production; a class structure in _ which the male, manual worker constituted the prototypical citizen; and a society in which the prototypical household was of the stable, one-earner kind. Below is a summary presentation of the three welfare regimes.1 The Liberal Welfare Regime Liberal social policy can trace its roots back to nineteenth-century English ö political economy, to its notions of 'Jess^eligibility' and ^self-help'. 11 harboured anunbounded faithin market sovereignty. Liberal welfare regimes ^ in their contemporary form reflect a political commitment to minimize the 1 Although the labels differ, the typology has considerableafftaity to Tirmuss' (1958) original distinction between residuaL institutional, and 'industrial achievement'"models. American react- -ers should be warned that 'liberal' and 'conservative'' are used throughout in their classical, European usage. Liberal therefore does not imply 'leftist', but rather a laissez-faire, neo-liberal view; conservative does not refer to 'whatever is not liberal", be it the Christian right or the Republican party, but to the European usage of the term. Comparative Welfare Regimes Re-examined 75 state, to individualize risks, and to promote market solutions. As such, ■-"•■ they disfavour citizens' entitlements. Liberal social policy prevaiTs"in countries where sociali^roTClTftstian ° democratic movements were weak or de facto absent. Äs Castles (1993) has pointed out, there is a peculiar clustering of the Anglo-Saxon nations around the liberal model. But, as he adds, it is important to distinguish between those societies, like Australia and Britain, where the labour movement played a significant role in social policy formation and those, like the United States, where its role was peripheral. In the former 'lib-lab' case, to use Castles' terminology, the welfare state is more comprehensive and col-lectivist. Disregarding such variation for the moment, there arejhree core elements that characterize the liberal regime. It is, firstly, resjdual in the sense that social guarantees are typically(-\) restricted to 'bad risks'. It adopts a narrow definition of who should be eligible. Liberal social policy is therefore very much the child of nineteenth-century poor relief, favouring means or income tests so as to ascertain desert and need. Accordingly, the relative weight of needs-based social assistance compared to rights programmes should constitute an excellent indicator of'liberalism'. Indeed, international comparisons show that the Anglo-Saxon welfare states are extraordinarily biased towards targeted sociaj assistance; Australia, New Zealand, the United States, and Canada in particular (Esping-Andersen, 1990, ch. 3). This is amply confirmed in later research (Gough et al., 1997, Table 2): it is either dominant or at least a major element of the total social protection package in Australia, Canada, Ireland, New Zealand, the UK, and the USA; as a percentage of total social expenditures, it now accounts for close to 100 per cent in Australia and New Zealand; about 40 per cent in Ireland and the United States; and 20-30 per cent in Canada and the UK. The closest non-liberal runner-up is Germany with 12 per cent.2 Liberal policy is, secondly, residual in the sense that it adheres to a j ) « narrow conception of what risks should be considered 'social'. The United " States is extremely residual because of its lack of national health care, sick- ° ness and maternity benefits, family allowances, and parental leave provisions. To address market failure in these areas, the approach is targeted aid to 'bad' risks: Medicaid benefits to the poor and Medicare to the aged, AFDCto lone mothers, and tax credits to low-income child families. Yet, if there, is a strong.tradition of selectivity there are also creeping elements; of comprehensiveness. This is no more evident than in the emerging popularity of some form of negative income tax, an approach originally o 2 "There are, to be sure, significant variations in how much targeting is built into social assistance programmes. As Castles (1996) and My les (1996) show, the Canadian and Australian approach to targeting implies screening out the rich rather than, more narrowly, 'screening in" only the abject poor. 76 Varieties of Welfare Capita/ism proposed by Milton Friedman and now being gradually extended in Australia, Canada, the UK, and the United States. --------------Orrorre-eotmtflfbe^^ namely as r^aj^LfeBŠlbLSiÄGes. Trie reasons^" however, are different. Ľi^eitlFvIewlervicmg as^a'n^wcalm^^^ctivity, as an individual 4 responsibility; conservatives insist"that if be tne prerogative of families.3 / , \ .....The third characteristic of liberalism is its encouragement of the market. Nowhere was this more pronounced than in America's promotion of 'welfare capitalism' during the 1920s or in the Thatcher era in Britain. Indeed, it was not until the New Deal in the 1930s that the United States introduced the first national social security programmes. As noted, the residual approach cultivates dualisms: the good risks can be self-reliant in the market; the bad ones become 'welfare dependents'. There are, of course, several ways of playing the market: individually (personal retirement accounts, life insurance, and the like), or collectively (group insurance or occupational welfare plans). What unites them both is that they generally benefit from substantial tax concessions. Welfare regimes are quite polar in terms of the role of market provision: the ratio of private to public pensions is .5 in Australia, .7 in Canada, A in Japan, .3 in the United States, but in Continental Europe or in Scandinavia the ratio is of the order of 0.1 (or less).4 In health care, however, the picture is less clear. Several European nations appear quite private (20 per cent of total in Germany, 36 per cent in Austria) because much of health care is run by 'third sector', non-profit associations. And some 'liberal' regimes, like Canada and Britain, have universal national health insurance. Private health care dominates in the United States (57 per cent of total), but here again non-profit firms (like Blue Cross-Blue Shield)play a decisive role. Private welfare plans have traditionally been more collective than individual in Australia, Britain,1 and the United States, to a large degree because of collective bargaining traditions (Esping-Andersen, 1990, Table 4.2). Declining-union membership and coverage appears, however, to contribute to an erosion of occupational plans and to promote greater individualism. This is especially evident in the United States where coverage under occupational pension plans has declined from about 50 per cent in 1970 to less than a third today.5 3 This was examined in Chapter 4. 4 Calculated from Esping-Andersen (1990, Table 4.3). 5 The Australian model of a "wage-earners welfare stateVasCastles (1993, 1996).calls it, may be considered a special case of market-biased welfare—indeed Castles-insists that Australia (with Eritain) be classified as a Liberal-labour model. What underpins the seemingly prototypicalJy residual, needs-tested Australian welfare state is a tradition of de facto job and high-pay guarantees to the male bread-winner. Hence, welfare guarantees have been implanted in the labour market. Comparative Welfare Regimes Re-examined 11 If we define the liberal model in terms of the weight of residualism (few rights and modest levels of de-commodification) and of mark~ěTš7there is clear"evidence ofTTat1oTrciirstering."TheTwo attributes"are highly corre-lated. The liberal welfare regime cluster is, furthermore, almost invariably Anglo-Saxon: the United States, Canada, Australia, Ireland, New Zealand, and the UK.6 If we take two key measures (means-tested assistance as a share of total social expenditure (early 1990s), and private pensions as a share of total pensions (1980s) and correlate them with a "liberal' nation dummy, the association is strong and positive: +.68 for the assistance variable, and +.52 for the private pension variable. For both the social democratic and the conservative regimes, the correlations are negative (-.31 and -.41 for assistance; -.27 and -.29 for private pensions).7 This can be presented in terms of logistic odds-ratios, as in Table 5.1. Table 5.1 shows that the liberal regime is fairly well predicted by these two features, while the two others are clearly not. It also suggests that the social assistance bias is a more distinctive feature than are private pensions. Table 5.1. Logistic odds-ratios for social assistance and private pension dominance by welfare regime Liberal Social Conservative regime democratic regime regime Social assistance as a 1.185** 0.855 0.922 percentage of total Private pensions as a 1.188* 0.925 0.924 percentage of all Pseudo-R2 0.681 0.088 0.215 * Probability better than 0.05. ** Probability better than 0.01. Sources: Esping-Andersen (1990, ch. 4) and Gough et al. (1997). " Britain is.aparticularlyinterestingcase since, in the 1950s, it would have been difficult to distinguish it from the Scandinavian. Here is a clear-cut case of typologies being undone by historicalchange,-an issue to be addressed, below. 7 The social-assistance data derive from Gough et cd. (1997, Table 2), the private pension data from Esping-A.ndersen (1990.T!able4J). ThefoUowingirations are scored as 'liberal'; Australia, Canada, Irefand, "New Zealand, Switzerland, the UK, and the USA; social democratic include: Denmark, Finland, "Norway,, and Sweden; conservative, Austria, Belgium, France, Germany, Italy, Japan, the Netherlands. See below for a further elaboration. 78 Varieties of Welfare Capitalism The Social Democratic Welfare Regime T'MsTegimtri^virte^^^ an international latecomer. In Denmark, Norway, and Sweden, its cornerstones were laid with the-advent of stable social democratic governance in the 1930s and 1940s; in Finland, twenty years later. Yet, cornerstones and mature form are not the same thing. There are good reasons why we should reserve the social democratic label for the period since the mid-1960s. In fact, the historical roots of Nordic social policy were, with some minor exceptions, quite liberalistic (Kuhnle, 1981). The legacy, as in Britain, was nineteenth-century poor relief. This was gradually transformed into social assistance and then, from the 1940s through the 1960s, into modern entitlement programmes. None the less, universalism was embryonic in the Scandinavian welfare states already from early on. Denmark was a world pioneer when, in 1891, old age assistance (upon an income test) was extended to all aged. Sweden's 1913 pension insurance plan was, at least in spirit, meant to be universal. Besides universalism, the social democratic welfare state is particularly committed to comprehensive risk coverage, generous benefit levels, and egalitarianism (Korpi, 1983; Esping-Andersen, 1990; Hicks et al., 1989; Stephens, 1996). Most studies also concur that this particular package of attributes is very much the political child of decades of strong, even hegemonic, social democratic rule.8 That universalism is the cornerstone of social democratic risk pooling does not suffice to distinguish it as a unique regime. The social democratic 'peoples' pensions' or national health care are only marginally more universalistic than their post-war British or Dutch brethren. True, the Nordic countries have undoubtedly pushed the frontiers of universalism further than anywhere else. And, more importantly, rights are attached to individuals and are based on citizenship (whereas the British and Dutch pensions are contribution-based)—rather than on demonstrated need or on an employment relationship (Palme, 1990). And where the Nordic countries truly stand out incomparison to other tend'entially universalistic systems (like the British) is their deliberate attempt to marginalize the role of needs-based assistance. The social democratic regime is distinct also for its active:and, in a sense, explicit effort to de-commodify welfare, to minimize or altogether abolish K Li is not difficult to trace the historical link-between what social democratic leaders pursued in the name of socialist solidarity, and what eventually emergeäVíspecially in Sweden. It is, however, possible that this historical 'correlation is spurious in the sense that a culture of universalistic solidarities has its roots much further back in Scandinavian society (Baldwin, 1990; Esping-Andersen, 1992). Comparative We/fare Regimes Re-examined 79 market dependency. While the UK (and also the Netherlands) has encouraged private welfare, especially in pensions and care services, the Nordic - -eetH*tries-sto*ggled4e4ibej«^ equality. When, in the 1950s, private occupational pensions began to mushroom so as to compensate for the rather meagre flat-rate public pensions, the Nordic countries (except Denmark) responded with a public second-tier system; Britain, in a similar situation, first vacillated and then eventually allowed the market to reign (Heclo, 1974; Pierson, 1994); the Netherlands encouraged company-based occupational pensions. ....... The closing off of private welfare is feasible^„onlyitheiiefits.are,adequate. There is no doubt that the Nordic social democracies boast very high income-replacement rates across the board. On various programme-specific or synthetic^^ de-commodification indices, Sweden, Norway, and Denmark are the world's three highest scorers (Finland scores somewhat above the mean) (Esping-Andersen, 1990, Tables 2.1 and 2.2). More recent comparisons come to a very similar result (Ploug and Kvist, 1994). Yet, generous income replacement is not a uniquely 'social democratic' attribute. In fact, the Dutch and Belgian de-commodification scores are hardly inferior to the Danish. And, as to pensions, the largesse of Scandinavia pales in comparison to Italian and Greek benefits (Commission of the European Communities, 1993). The point is that if we limit our study of de-commodification to standard income transfer programmes, the great regime divide is not so much social democracy versus the rest, but rather that the liberal regime provides uniquely modest benefits compared to either of the two other regimes. What, then, is uniquely social democratic is, firstly, the fusion of universalism with generosity and, secondly, its comprehensive socialization of risks. By the early 1970s, most (non-liberal) welfare states had arrived at a fairly similar level of comprehensiveness as far as cash benefit programmes are" concerned. It is at this point, however, that the social democratic regime comes into its own by complementing standard income protection with social services and_generous income support for working women. Led by Denmark and Sweden in the late 1960s, the Nordic welfare states became 'servicing states'. On top of health services (where they hardiy distinguish themselves from elsewhere) was built,a huge and comprehensive infrastructure of services especially catering to family needs. As we saw in Chapter 4, care for children and theaged is especially privileged. The net outcome is that government employment now accounts for up to 30 percent of the labour: force, or more than double the OECD average. The social democratic model and egalitarianism have become basieaily synonymous. To many, the egalitarian element is simply the practice of universalism: everybody enjoys the same rights and benefits, whether rich 80 Varieties of Welfare Capitalism or poor. To others, it refers to the active promotion of well-being and life chances—perhaps no more evident than for women. Still others equate l^fiinníäTllšttrwT^^ shall see in subsequent chapters, they are all right. Full employment has surely been a mainstay commitment in the social democratic model—but so it has elsewhere. What separates Scandinavia from most other nations is decreasmgly low unemployment and, increasingly, maximum employment.. As first Denmark, and now also Finland and Sweden,, suffer from mass unemployment, the coincidence of full employment and sociaL democracy appears shattered. This is, none the less, not as clear as it seems. For one, we should distinguish between actual performance and political commitment. As far as the latter is concerned, the coincidence remains. The Nordic countries' expansion of public employment since the 1960s may have been guided by egalitarian concerns, but it certainly was also a means of promoting employment. \ Current Nordic unemployment may not differ much from, say, the \ German or French, but it occurs on the backdrop of an employment rate i of 75-80 per cent rather than 50-60 per cent—principally because full ! female participation was realized. The employment commitment is equally evident in active labour market policies, both in terms of resources spent and number of persons covered under various training, retraining, or employment promotion programmes. Scandinavian welfare and employment policy has always been couched in terms of 'productivism', that is of maximizing the productive potential of the citizenry. Superficially this seems like an echo of what Americans call 'workfare'. In reality the two are different. Workfare in America implies that social benefits are conditional on accepting work whereas Nordic 'productivism' implies that the welfare state must guarantee that all people have the necessary resources and motivation.to work (and that work is available). The social democratic regime, then, is inevitably astate-dominated welfare nexus. But, the Nordic welfare states, and in particular the Swedish, are now experiencing hard times—to use Stephens et al.'s (1994) expression. For budgetary reasons, governments have reduced social benefits: lowering replacement rates, introducing waiting days for sickness benefits, shortening the duration of unemployment pay and, perhaps most ominously, introducing elements of targeting in the 'peoples' pensions'. While cuts at the margin hardly affect the jessence of the social democratic regime, the implementation of an.income test for pensions may do so since this implies a qualitative retreat from the. principle of universalism: the notion of solidarity of risks is being rewritten. Again, let us examine the validity of the Tegimes. Besides its strong accent on de-familialization (which will be examined below), two features Comparative Welfare Regimes Re-examined 81 of the social democratic regime stand out: universalism and the marginal-ization of private welfare. In Table 5.2 I present odds-ratios for the three i:eiiime .types..wilh_JLhe__j.miversi-i1ism mid pjjy^te_pensian-scaĽes-fram-Esping-Andersen (1990, Table 3.1). Tadlií 5.2. Logistic regressions {odds-ratios) of universalism and private pensions and welfare regime types Social Conservative Liberal democratic regime regime regime Universalism 1.321* 0.988 0.938 Private pensions 0.731 0.926 1.185* Pseudo-R- 0.651 0.07S 0.350 Chi-squared 12.41 1.88 8.43 * Probability = 0.1 or less. N = 18. Source: Esping-Andersen (1990, Table 3.1). Table 5.2. confirms more or less what we would expect. The odds that we will find universalism in the social democratic regime are comparably strong and significant. Vice versa, the odds of private welfare are low (however not significant) in the social democratic regime but, as we already have seen, positive and significant in the liberal. On these variables, the conservative regime is, simply, indistinct. The Conservative Welfare Regime Labelling the Continental European welfare states conservative may appear pejorative. The idea, however, is to signal the dominant political thrust behindthe[r architecture. In most of Continental Europe, liberalism played atruly marginal role and, until after World War Two, the socialists typically found themselves excluded. Early social policy was often inspired by monarchical etatism (especially in Germany, Austria, and France), by traditional corporatism, or by Catholic social teachings. Leo Xlľs Papal Encyclical RemmNovarum (189T)~liátiTtemcndous influence in Catholic-dominated countries. Moreover, the passage from origins to post-war welfare capitalism has, in this group of countries, been guided primarily by Christian democratic or conservative coalitions (in some cases with a Fascist interregnum). The essence of a conservative regime lies m its blend of status segmentation and familialism. Most Continental European countries emulated 82 Varieties of Welfare Capitalism Imperial Germany's social insurance reforms and, like Bismarck, their original aims were far removed from any egalitarianism. The early social TeronnCTS^werelypicatry^^ era, the imprint of social Catholicism and its doctrine of subsidiarity has been particularly strong in Southern Europe, the Netherlands and, to an extent, also in Belgium and Germany (van Kersbergen, 1995). French social policy has, in contrast, been guided primarily by a republican, anticlerical spirit. France's (and Belgium's) membership in the conservative cluster is, as we shall discuss, problematic in that familialism is less dominant. Yet, both welfare systems display strong corporatist traits. There are several reasons why we might also include Japan in the conservative model. The powerful presence of Confucian teachings throughout Japanese social policy is a functional equivalent of Catholic familialism, and also Japanese social security is highly corporatist.9 The conservative imprint is most evident with regard to risk pooling (solidarityTimd familialism. In both cases, the historical lěgacywíís lypi-callTcäTried over in the making of the post-war welfare state. The etatist legacy remains strong in the privileged treatment of thepublic civil service, especially in Austria, Belgium, France, Germany, and Italy.10 The civil service benefits not only from having its own scheme but also from vastly more luxurious eligibility and benefit rules. Also, despite some attempts to consolidate the myriad occupational schemes, corporatist status divisions continue to permeate social security systems. Of course, the accent differs between individual schemes and countries. Germany is a case of modest corporativism in pensions (the principal distinction is between blue- and white-collar workers), while health insurance is a labyrinth of 1,200 separate regional, occupational, or company-based funds. Italy, in contrast, has a unified health programme -while pensions are divided into more than 120 occupational plans (Castellino, 1976). Both France and Belgium combine fragmented pension systems with a national health insurance that is divided along broad occupational classes. Among the Continental European countries, only „the Netherlands deviates markedly from the corporatist mould. Public pensions are organized more or less around the Beveridge principle of uni-' versal flat-rate benefits, and other programmes, such as heahh, education, and services generally, were "pillared', that is, split along denominational -and non-denominational lines (van Kersbergen,. 1995). "9 Japan deviates from this cluster only in terms of the large role of employer-provided occupational benefits. 1,1 Government employee pensions account For30 per cent of t lie total m Austria; 35 per cent in Belgium; 27 per cent in France and Italy; andZl perceniiii Germany-This dsr on average, two orthree times as high as in Scandinavia or in the Anglo-Saxon countries (calculated from Esping-Andersen, 1990, Table 4 3). Comparative Welfare Regimes Re-examined 83 The accent on cqmpujsory social insurance, complemented with more or less aälwFrésidiväl schemes for strata without a 'normal' employment -n^hrtiüti sMpr-lras-^ remains marginal. Granted, a significant part of health care is, in some countries non-state but this is clíiěfiýdue to the role played by non-profit, •voluntary' associations, frequently affiliated with the Church (such as Caritas). As to pensions or other welfare provisions, both individual and collective occupational plans are generally of marginal importance, in some countries basically non-existing. A notable exception is, again, the Netherlands where (mandated) company pension schemes play a non-trivial role in the labour market.'' The third important attribute of conservatism is its familialism, especially in Southern Europe and Japan. As shown in Chapter 4, familialism is a composite of the male bread-winner bias of social protection and the centrality of the family as care-giver and ultimately responsible for its members' welfare (the subsidiarity principle). What unites Austria, Germany, Italy, and Spain is the continued legal prescription that parents (or children) be responsible for their children (or parents) in case of need. Social assistance, for example, will not be granted even to adults if their parents can support them. Besides legal obligation, there is a systematic disinclination to provide care services, and the more familialistic the welfare state, the less generous are family benefits. Family transfers are often regarded as redundant given the practice of a family wage. But, since the model assumes the standard male bread-winner family, provision for 'atypical' households, such as lone mothers, tends to be residual. There is, then, a modicum of residualism in the conservative model that seemingly parallels the liberal. Yet, its target is very different: liberal residualism means picking up bad risks left behind by market failure; conservative residualism, in contrast, is primarily a response to family failure. In both cases, none the less, the approach favours social assistance over rights, such as the German Sozialhilfe, the Italian and Spanish social pension, or even the French RMI., Again for different reasons/both conservative and liberal social policies inherently favour a passive approach to employment management) The liberal model simply prioritizes unregulated labour markets; the conservative, strong job protection for already employed adult, male householders. Active empLoyment or training, policies tend to be marginal in both 11 Labour market pensions do play a role also in France, but their status should, best be regarded as compulsory. Here Japan is of course deviant because so much of the entire welfare package is provided by employers within the context of lifetime employment. As in the USA, Japanese occupational welfare is mainly the privilege of core workers—about one-third of alL In the Japanese case, it is debatable whether this kind of private welfare conforms to market principles (Esping-Andersen, 1997a). 86 Varieties of Welfare Capitalism Universalist: Denmark, Finland, Norway, Sweden, the Netherlands (and to a degree, the UK). ~~SoTiäl insurance: Án5tria7ÍfelgŤtrm7~fr^ and Spain. C: FAMILIES Familialist: Austria, Germany, Italy, Japan, the Netherlands, Portugal, Spain (and, less so, Belgium and France). Non-familialist: Australia, Canada, Denmark, Finland, New Zealand, Norway, Sweden, the UK, and the United States. We turn now to the question of whether a 'three worlds' typology remains robust and valid. THE HISTORICAL AND COMPARATIVE ROBUSTNESS OF REGIME TYPOLOGIES Since typologies refer to one time-point, we shall miss out on possibly decisive transmutations. And since they are, in a sense, ideal types there are bound to be ambiguous cases. Some critics have, for example, pointed out that the Netherlands and Britain, both in their own way, fit poorly in any of the three clusters. It is also possible that the criteria employed to demarcate a regime may err: if alternative attributes were considered, the classification might break down or, at least, require additional regimes. For different reasons it has been argued that the Antipodes, East Asia, and Southern Europe all merit a 'fourth world'. It has also been argued that the entire typology is problematic because it is built on the experience of the standard male breadwinner; it is not gender-sensitive. Getting the welfare regime typology straightrmatters not just for the historical record but also for further analytical progress. At the core of a welfare regime study lurks the presupposition that institutional configuration matters for how risks are absorbed and distributed, for social stratification and solidarities, and also for the operation of labour markets. Hence, regimes should display some degree of congruence and commonality in how they adapt to massive social and economic change. This is a major issue of this book and we need, therefore, to establish just how robust is the working typology. If there are ambiguous cases in a typology, the question is how much they matter. The goodness-of-fit of the three-way regime typology lias been tested utilizing several methodological techniques, including cl Lister analysis, Boolean algebra, factor analysis, and more conventional correlational analysis (Kaňgas, 1994; Ragin, 1994; Shalev, 1996). While these Comparative Welfare Regimes Re-examined 87 studies find that some countries fit more poorly than others (for different reasons, Belgium, France, Britain, and the Netherlands are frequently ------rjt^eü^es-)7-thei:e-is~ii-lso-seme-s-ti ppe r-t-fo rthe-t h-Fee-el u^te-ra----------------- Britain is mainly a problem because the typology does not take into account mutation. Had we made our comparisons in the immediate postwar decades, we would almost certainly have put Britain and Scandinavia in the same cluster: both were built on universal, fiat-rate benefit programmes, national health care, and a vocal political commitment to full employment. Moving ahead into the 1970s and beyond, the two clearly part ways: Britain failed to uphold its full-employment commitment and to supplement modest flat-rate benefits with a guarantee of adequate income replacement (Heclo, 1974; Martin, 1973). Failure to keep up promoted a gradual privatization that was, no doubt, accelerated by concerted de-regulation, more targeting, and privatization during the 1980s: sickness and maternity benefits were transferred to employers, council housing was sold off, the earnings-related pension (SERPS) was 'privatized' through opting-out, and both private pensions and health insurance have been nurtured via tax subsidies (Taylor-Gooby, 1996).14 In a contemporary comparison, then, Britain appears increasingly liberal. Britain is an example of regime-shifting or, perhaps, of stalled 'social democratization'.15 The original 'three worlds' typology focused rather one-sidedly on income maintenance. Herein lie perhaps the ambiguities of the Dutch case. When we study income maintenance, the Netherlands appears 'social democratic' in the sense of strong universalism, comprehensive coverage, and generous 'de-commodifying' benefits (van Kersbergen, 1995). But when we include social service delivery—and when, more generally, we examine the role of the family—the Netherlands becomes squarely a member of the 'conservative', Continental European fold. Like Britain, this would not have been immediately obvious in the 1950s because then, also the Nordic welfare states were service-lean and transfer-biased. It is in its sustained inattention to social services that the Netherlands emerges as a prototypical example of Catholic familialism (Bussemaker and van Kersbergen, 1994; van Kersbergen, 1995; Gustafsson, 1994). What is more, the generosity of Dutch income maintenance is chiefly the expression of a pervasive male bread-winner assumption. M It is also telling that public residential care for the aged declined by 4 per cent, 1976-86, while private care grew by 363 per cent (Evers and Svetlik, 1991: 130). 15 The bulk of the data used to identify regimes in my Three Worlds referred to c.1980. The true liberal nature of Britain's transformation only became fully visible later. It may be true that Britain and Sweden were rather alike urthe 1950s, but this does not imply that Britain was then "social democratic'. What í define as the essence of social democratic welfare regimes emerged in Scandinavia later. 88 Varieties of Welfare Capitalism The Dutch enigma, then, highlights the need to reconsider, once again what must be compared and measured. Income-transfer programmes rcrptuTeHrot-one~side-^^ democratic (or the conservative) welfare states lies not so much in their de-commod'ifying income-maintenance guarantees as in their approach to services and sponsoring women's careers. In any case, the Netherlands remains a Janus-headed welfare regime, combining both social democratic and conservative attributes. The point here, as in other cases, is that we must weigh the relative importance of different, possibly conflicting attributes. No regime, let alone country, is pure. The United States epitomizes liberalism and, yet, the Social Security pension scheme has broad coverage and benefits that approach adequacy levels. Does this push the United States out of the liberal fold? No, because even if one programme deviates from the 'ideal type', the over-dominating character of the entire welfare package remains 'liberal'. Or consider Denmark which, like Britain, failed to implement a universal, second-tier pension system. Does this imply that Denmark, with Britain, is a failed social democratic model? On this count, again the answer is no. The benefits in the Danish 'peoples' pension', unlike the British, were systematically upgraded so as to uphold their universalistic appeal; and on virtually any other criteria, the Danish welfare state is pro-totypically 'social democratic'.16 There will always be slippery or ambiguous cases, and one programme does not define a regime. The real problem is how to deal with systematic deviants. The issue here is whether a three-way typology adequately exhausts the variance. If there are cases that follow a wholly different underlying logic, we would have to construct yet another, separate ideal-type—a fourth 'world of welfare capitalism'. Three-plus Regimes! There are in particular three cases that arguably call for an additional, fourth 'world': the Antipodes (Castles and Mitchell, 1993), the Mediterranean (Leibfried, 1992; Lessenich, 1995; Ferrera, 1996), and Japan (Jones, 1993; Rose and Shiratori, 1986). Assuming the validity of all three claims, we will find ourselves with a total of six models for a total of 18-20 nations. The desired explanatory parsimony would be sacrificed, and we might as well, return to individual comparisons. 1,1 Two very recent reforms may, however; very well push Denmark towards a more liberal fold. One,, the peoples' pension has lost some of its 'rights" character with the introduction of aa income test in 1996; two, legislation has just been passed to build a private (collectively negotiated) second-tier earnings-related labour market pension. Comparative Welfare Regimes Re-examined 89 f lie Antipodean Fourth World _9isttes-4ind~-MitdieU-(lí^^ favour of a fourth, Australian or Antipodean, welfare regime (what they term the 'wage-earne£sj_welfare_s.ta.te'.).i At first glance, Australia's and pjew Zealand's rather modest and targeted welfare state benefits conform to the residual, liberal model. All income maintenance-schemes are now premissed on an income test, but 'Medicare' in Australia is essentially a rights programme. While needs-tesfěď benefits are generally much lower than equivalent insurance benefits in other countries, they are also more 'needs-sensitive' than elsewhere. Thus, assistance benefits to child families are double those to single persons (Castles, 1996: 109). Moreover, the income ceiling for eligibility (in Australia, but less so in New Zealand) is drawn at middle incomes, not at a poverty line: assistance is more inclusive. Pensions are said to cover two-thirds of the aged; family benefits reach most middle-class families.17 The point that Castles and Mitchell make is that we err by focusing solely on state activity because in Australia (and once also New Zealand), strong and functionally equivalent welfare guarantees were implanted in the labour market via the wage arbitration system. What seems like an extreme case of a liberal, means-tested system when studying only state welfare is, in reality, something essentially social democratic with its emphasis on egalitarianism and wage-earner rights. If valid, the argument is theoretically fundamental because it compels as to reconsider markets. In Australia, as the argument goes, the labour market is a welfare producer. Hence, it may be a fallacy to simply equate markets and liberalism. There is no doubt that the wage arbitration system in Australia implanted strong and egalitarian guarantees, at least as far as the male bread-winner was concerned. There was little need for a welfare state because male full employment was de facto 'full5, because earnings differences were highly compressed, and because the employment relationship furnished general welfare guarantees, such as home-ownership and adequate pension income. However, as Castles (1996) himself points out, these very same guarantees inevitably eroded when, during the 1980s, the Australian economy was liberalized; ffiey were effectively eliminated in New Zealand. With unemployment rates hovering at 10 per cent, and with heightened wage inequalities, the 'wage-earners'welfare state' in the market is, almost by definition, being dismantled. rT For a period, old-age pensions became universal rights benefits, but are now again income-tested. For a general, comparative presentation of the Australian system, see also Saimders (1994). 90 Varieties of Welfare Capitalism It is possible that the Antipodean model provided a package of welfare guarantees that was essentially 'social democratic1 in the 1960s and 1970s. -Like^Brito-i-n^h^tw^v^r^-the-passage of time is pushingj\ustralia-—and cer-tainly New Zealand—towards what appears as prototypical liberalism: minimal state and maximum market allocation of risks, and the market side of the coin appears increasingly genuinely market. The Mediterranean Fourth World It has been argued that the Mediterranean countries should be considered distinct from Continental Europe (Leibfried, 1992; Ferrera, 1996; Lessenich, 1995; and Castles, 1996). Ferrera's point has mainly to do with distributive practice—the pervasive use of social benefits, especially in Italy, for purposes of political clientelism. Invalid pensions and public jobs are notorious ways in which the Christian Democrats (but also socialists) maintained their grip on the electorate. A perverted use of welfare programmes and public bureaucracies may define the character of a polity, but il is difficult to see how it defines a welfare regime unless the entire system was from the very beginning specifically designed for the purpose of clientelism rather than social protection. Such an argument would be very hard to sustain. Leibfried's (1992) call for a distinct Mediterranean regime is, as previously discussed, limited to one programme—social assistance. Bui even if social assistance practice is distinct, this hardly merits a distinct regime type. It would be a different matter if such deviance forms part and parcel of a broader complex of attributes. Does it? To an extent yes, because the extremely residual nature of Southern Em^ one face of its strong familialism. Unlike elsewhere, social assistance was never upgraded because of two assumptions: it is assumed (and legally prescribed) that families are the relevant locus ofspcialaid; and it is assumed that families normally dö not 'fail'. The acid test of a distinct Mediterranean model depends therefore on the issue of familialism, to which will shall return shortly. The East Asian Fourth World Japan, possibly with Korea and Taiwan, poses a particularly intriguing challenge to welfare regime typologies because it is such a unique version of capitalism to begin with: sustained full employment, highly regulated internal labour markets and industrial structure, compressed earnings, and a relatively egalitarian distribution of income, all overlaid by rather, authoritarian employment practices, a conservative . 'ar^-y^ty1 democracy, and 'corporatism without labour (Pempel, l:989)vls 1K The following treatment is based on Espmg--Andersen (1997;/). Comparative Welfare Regimes Re-examined 91 If we confine our examination to Japan, one aspect is immediately obvious: it blends vital regime attributes in a way which makes it either unique or hybrid. This, indeed, is the fundamental issue to be resolved. State wel-fare combines liberal residualism with conservative corporatism. Social insurance is, like in Europe, status-segmented along broad occupational lines. In part due to lack of system maturation (at least as far as pensions are concerned), insurance- benefits are. quite modest and attached with prohibitive eligibility criteria—this is especially the case in unemployment insurance. Levels of de-commodification are modest indeed. The modesty of benefits reflects a certain in-built residualism. State benefits assume that the core (male) labour force will have private benefits via the employment relationship, but also ample family support. The former undoubtedly holds for the roughly one-third employees in the large corporate sector—but much less so for the rest of the labour force. Employer occupational welfare includes not merely standard health and pension coverage, but also a vast array of services from sports clubs to funeral services. And herein lies one important source of inequalities, not so much according to class or occupation, but because access and generosity depends on educational level and, of course, on whether one is employed by a large corporation or not. Japan's social assistance system is yet another example of liberalism: strictly means-tested and highly targeted as well as stigmatizing. The take-up rate is extremely low (below 30 per cent). The welfare state, then, fuses conservative and liberal elements which is hardly surprising when we recall that the insurance component was inspired by German practice, and that assistance schemes were designed by the American occupation forces in the aftermath of World War Two. However, if we examine more closely the welfare market it would appear somewhat less than 'liberal'. Occupational welfare forms part of the consensual mode of labour regulation, but it also mirrors a conservative, paternalistic practice. It is none the less arguable that the market 'works': it provides (until now) a genuine employment guarantee and, for many, welfare benefits. The rather undeveloped and residual public welfare commitments could—like once in Australia—be ascribed to the fact that needs and risks are contained because the labour market functions so well. In fact, several scholars maintain that Japan is an extremely highly developed 'welfare society' and that a large welfare state is therefore less pressing (Vogel, 1980). The state is not needed because the market and the family are sufficient. There is little doubt as to the importance of families in the overall Japanese welfare mix. Public social services, be it to the aged or children, are truly marginal because it is institutionally assumed that 'thefamily must carry the real responsibility. Thus, even today 65 percent of the aged live with, and are cared for by their children—that isr the wife of the oldest son. The Confucian tradition of familial piety and loyalty 92 Varieties of Welfare Capitalism has, like the Catholic subsidiarity principle in Europe, been the ovei powering force behind Japanese welfare policy (Jones, 1993). —TheTeŤrclrarly-Troth^^ appears therefore a hybrid case. Yet, it is precisely this particular combination of hybrid attributes that, some say, warrants a distinct regime label. True, Japan does manifest a mix of liberal and conservative traits and, unlike other 'mixed cases', they appear to form an internationally rather peculiar combination. Yet, the liberal side of the equation is less liberal than appearances suggest. There are also trends that point towards a strengthening of the conservative attributes. The corporatist social insurance system is now rapidly maturing and will, in the coming decades, dominate the pension mix. This, together with Japan's unusually accentuated familialism, makes a strong case for assigning Japan squarely to the conservative regime. In brief, it is inescapably true that Japan, like Australia and Southern Europe, manifests features that are not easily compatible with a simple trichotomy of welfare regimes. Yet, we must also ask ourselves what would be gained from adding a fourth, fifth, or sixth regime cluster? We would probably benefit from greater refinement, more nuance, and more precision. Still, if we also value analytical parsimony, neither Japan nor the Antipodes warrant additional regimes. The peculiarities of these cases are variations within a distinct overall logic, not the foundations of a wholly different logic per se. The case for a unique Southern Europe regime depends ultimately on the centrality of families. This was the weak link in the original 'three worlds' model and therefore deserves special attention. Families and Welfare Regimes To what extent, then, does more systematic attention to the family alter our political economy of welfare regimes? Does it call for a fourth regime or even a radical rethinking of what constitutes a welfare regime in the first place? Those who .follow .a more explicitly gender-based analysis often arrive at typologies that are. at odds with the 'three worlds' model (lor an overview, see 0!Connor, 1996). A similar lack of fit emerges from some studies of the 'family benefit package' (Bradshaw and Ditch, 1995). From the evidence-presented in Chapter 4, there is much to be said for an additional-Southern European regime (with Japan). Of course, in the final analysis it all comes down to the choice-of indicators and measurement. The weight of the evidence in Chapter 4 addressed families (and not gender perse) and their role in welfare production (rather than family benefits provided by the welfare state). In fact, previous work on airing and servicing sometimes squares with the vthree worlds^ typology {Kolberg Comparative Welfare Regimes Re-examined 93 and Uusitalo, 1992; Gustafsson, 1995), and sometimes not (Anttonen and Sipila, 1996; and Gornick et al., 1997). Both Anttonen and Sipila, and _^.arjiicliL-£tj2/_&h^ servative regime, but only in this particular area. Let us conduct a simple test of whether the conservative regime should be subdivided, whether a fourth 'Mediterranean' regime is warranted. If Southern Europe (Italy, Portugal, and Spain) together with Japan are qualitatively distinct from the remaining Continental European countries, this should show up on two dimensions of familialism: on welfare state policies towards families (welfare state de-familialization), and on the welfare burdens assumed by families themselves. Table 5.5 presents relative odds-ratios from multinomial logit regressions. The 'outcome' (dependent variables) to be predicted is 'high levels of welfare state servicing to families' (column 1) and 'high levels of intra-family caring burdens' (column 2). Each separate regime estimation includes a measure of women's employment rate (an indicator of underlying need for de-familialization, so to speak). The prediction in column 1 is the degree to which welfare states furnish services to families; in column 2 whether traditional family behaviour prevails (share of aged living with their children combined with the 'male bread-winner bias' variable).l9 The evidence does not suggest any great Table 5.5. Multinomial logit regressions. The relative odds of high levels of welfare state family support and of low levels of family-caring burdens by welfare regime and women's employment levels Outcome: high levels of Outcome: high levels of welfare state servicing to family welfare burdens families Southern Europe and Japan Women's employment rate Chi-squared 5.86e-19*** 0.91 '23.04 1.52e+ 16*** 20.44 Continental Europe minus Italy, Portugal, and Spain Women's employment rate. Chi-squared Perfectly predicted 5.34e-17*** 42.05 2.96e+ 13*** 0-34* 18.14 "* Probability = O.iO or better. *** Probability—0.001 or better. N =18. See n. 19 for an.explanation of variable measures. |l) The welfare state de-familiatism, and the household familialism variables have been recoded so as to develop synthetic indicators: Service spending has been coded 0 (if spending is 94 Varieties of Welfare Capitalism difference between our 'Mediterranean' subregime and the remainder of Continental Europe. The signs and relative odds are quite similar. In brief, the^^eHfera-fourthT-nm^ __ not very convincing when, as in this analysis, we focus on high levels of de-" familialization. The difference does, however, become more accentuated -when we relax our criteria. The correlations between regime type and various measures of familialism differ somewhat more between the two -regime subdivisions when we measure familialism as continuous rather than categorical variables (as was done in Chapter 4). None the less, the -signs of the correlations remain identical for the two subregimes.211 CONCLUSION The comparisons undertaken in this chapter are certainly not the last word on the subject. The question of how to identify and classify welfare regimes will remain open because, as noted, researchers differ in terms of what attributes they consider vital and of how to measure them. The results presented here suggest that, as far as my choice of attributes and measurements is concerned, a simple 'three worlds' typology may suffice for most of the purposes that this book pursues. The final judgement is not yet in, and we shall in fact see that the distinctiveness of the Soul hern European countries does make its mark on issues such as postindustrial employment adaptation. It is to this that we now turn. <.5 percent), 1 (if spending is .5-1 percent), 2 (if spending is 1 percent+). Family benefit spending has been coded 0 (if less than 5 per cent), 1 (5-8 per cent), 2 (8+ per cent). Daycare coverage is coded 0 (for < 10 per cent coverage), 1 (for 10-20 per cent coverage), and 2 for 21+ percent coverage. Home-help coverage has been coded 0 (<8 per cent), 1 (9-16 per cent), and 2(17+ per cent). Percentage aged living with children has been coded: 0 (30+ per cent), I (16-29 per cent), 2 (<16 per cent). The male bread-winner bias is a combination of the tax-benefit penalties of working wives and the loss of unemployment benefit if wives work. The coding is a simple dummy: 0(if there is a significant penalty on either), I (if there is no significant penalty). For data sources, see Chapter 4. 2(1 Bivariate correlations (not shown) indicate that the Southern European regime is considerably more familialistic than Continental Europe without the 'South', and this shows up also on consequences, such as women's employment levels. £ARXXI THE NEW POLITICAL ECONOMY A social trend is basically a projection from the past. Many of the early postindustrial theorists, such as Daniel Bell (1976), predicted a future in which most of what they deemed positive in the era of the 'democratic class struggle' would come to full fruition. Bell's vision was a coming society of professionals and technicians, one where 'situs' rather than class conflict would reign. This was a radical reinterpretation in so far as post-industrial society would do away with class altogether. Also Lipset has now embraced this position (Clark and Lipset, 1991; Clark, Lipset, and Rempel, 1993). Today's visionaries find much less cause for optimism; their projections are most likely to range from the sombre to the outright gloomy. The sombre view insists that little of substance has changed (Wright, 1989; Erikson and Goldthorpe, 1992; and Hout et al., 1993); that the cleavages of the past remain pretty much intact. The gloom comes from those who see a new era of polarization. American or British observers see a world with a 'declining middle', job polarization, and a new underclass (Harrison and Bluestone, 1988; Jencks and Peterson, 1991; Levy, 1988; Burtless, 1990). Europeans, in contrast, see a two-thirds society with social exclusion, marginalization, and outsider classes (Van Parijs, 1987; OfFe, 1985; EspingrAndersen, 1993; Brown and Crompton, 1994).1 What the new pessimists see, in brief, is the possibility of a resurgent proletarian underclass and, in its wake, a menacing set of new 'class correlates'. The transAtlantic difference of accent is clearly related to job performance: in North America labour market exclusion is less dramatic than is growing pay inequality, declining real wages, and a swelling army of the working poor. Europe's social safety net manages to stem the tide of inequality but its inferior job performance induces mass exclusion. How do we: iuiderstand the postindustrial employment proWem? Is there indeed any problem? And, if new class configurations are emerging, what is driving them? Besides the radical optimists, such as Lipset et al, who basically deny ihat ihere is any problem worthy of discussion, the 1 The specification of a 'one-third' outsider class in Europe today is surprisingly similar to Max Adlers (1933) analysis of the 1930s depression era. 8 New Social Risks in Old Welfare States INTRODUCTION The idea of a 'mature' welfare state has several distinct connotations. In the first place, it refers to the move from basic, minimal protection (which o characterized the 1950s) to a much more ambitious promise of benefit adequacy and universal coverage; income-maintenance systems matured in this sense around the late 1960s and early 1970s. In the second place, we o can now, with the aid of hindsight, regard the welfare state of the 1970s as mature" basically because few, if any, major changes have occurred since (it ceased to mature even fürther). In the third place, not only did the welfare state stabilize, but so also did the more embracing welfare regime. The o division of welfare between public, private, and family had become consolidated. Altogether, the welfare state flowered at the very same moment that 'Golden Age' capitalism began to wilt. Most pundits, be they foes or friends, believe that the welfare state now faces an irreversible crisis. It was sustainable as long ašThere was strong growth and full employment—hardly a hallmark of present-day capitalism. Indeed, in addition to its proclaimed unsustainability, many also claim that the welfare state has become a millstone around our necks, the principal impediment to urgently needed economic readjustment and brisk job growth: it upholds rigidities where flexibility is needed, and the rights of social citizenship stifle our incentives to work and save. But is the j crisis real? Is the welfare state incompatible with postindustrial society? A paradox of our times is the more that welfare states seem unsustainable, the greater are the demands for social protection. Of course, the paradox resolves itself easily when we keep in mind the distinction between welfare state and regime. As we saw in the opening pages of this book, the core symptoms of the contemporary welfare state crisis were all exoge- «/ nous, shocks on the system coming principally from the labour market graphics of ageing and family behaviour. We can identify clear differences in strategy, differences that largely mirror past institutional legacies and welfare regime adaptations over the past decades. MANAGING THE LABOUR MARKET The new 'equality-jobs' trade-off has generated two distinct strategic responses (Soskice, 1994; Freeman, 1994; and Blank, 1994). In what Soskice calls 'un-coojdjn^ď^conomjes (largely the Anglo-Saxon, liberal regimes), deregulation through wage flexibility, declining trade unionism, and weakenedemployee protection has predominated. As markets have been encouraged to 'clear', inequalitiesjiave risen. According to "Soskice (op. cit), this model also risks rmdingltself in a 'low-skill equilibrium'. The absence of active public training programmes is compounded by lack of employer incentives because of the availability of a mass of low-cost, substitiLtable workers. In contrast, 'co-ordinated' economies have opted for a high-wage, quality production strategy/lbatpiivileges a core of skilled, stable, mainly male workers at the expense or the "outsiders'. In the latter model, which represents most European Community member coun- New Social Risks in Old Welfare States \ 51 tries, early retirement and welfare guarantees have helped downsize the workforce and absorb excess labour. The underlying premiss is that the -brrd^getaTy-burdtJii of earlyreTirement can be coveredTTy the resulting pro-ductivity dividend. This contrast is of course quite stylized. The European countries with a long tradition of worker protection have, to varying degrees, introduced flexibility measures since the 1980s. Almost all countries have made it easier to hire workers on temporary contracts, rules regarding part-time work have been eased, and the cost of employing youth workers has been de facto lowered (either with subsidies or by lowering social contributions).2 As Blank (1994) and Regini (1998) conclude, partial deregulation reforms hardly alter the basic workings of the system. Take the most widespread deregulation measure, temporary contracts. In Germany, despite great fears to the contrary, they are not widely used and, in any case, they usually conclude in a permanent employment relationship (Buechtemann, 1993). The underlying reason is that the entire German labour market is organized around the formal (dual) apprenticeship system. In Italy, it is true, the lion's share of youth hirings are now on a fixed-term basis, primarily with subsidized 'training contracts'. About half eventually result in a permanent contract, and it is evident (as in the German case) that flexibility here mainly serves as a cost-advantageous screening device, not as an incentive to add additional workers to the payroll. Employers' quest for flexibility remains, as always, focused on hours adjustment among their permanent staff (Adam and Canziani, 1998). Spain represents the single most dramatic case of liberalized temporary contracts. Since the mid-1980s, almost all new hirings are fixed-term. Unlike in Germany and Italy, however, only a very small fraction (10-15 per cent) eventually move into a permanent job (Bentolila and Dolado, 1994; Toharia, 1997; Adam and Canziani, 1998). The consequence has been to actually strengthen and deepen the existing gulf that separates insiders and outsiders. Firms use their prime-age, male permanent workers as their dependable and trained core workforce; the heavily youth and female fixed-term employees, in contrast, serve to achieve numerical flexibility (people instead of hours). Since the latter are easily substitutable there is very little incentive to invest in training. Partial deregulation in Continental Europe has, by and large, meant the preservation of traditional protection standards within the prime-age, insider workforce. This bipolar contrast ignores the Scandinavian 'third way', until recently \ regarded_as_.a. model 'win-win' strategy, since full employment went hand- \ in-hand with an unusually high level of equality.3 The distinctiveness of this ' 2 For an overview, see Buechtemann (1993) and OECD (1994í/). -1 Denmark, of course, has suffered from fairly high unemployment rates since the mid-1970s. Until 1991-2, Finland, Norway, and Sweden succeeded in maintaining de facto full employment. 152 Welfare Capitalism Recast? approach lies m a combination of policies and institutions. The institutional precondition has been strong, nation-wide consensual wage bar- -----gfthiiflj^hiehT-unt-^ wage equality and full employment.4 The main policy principle has been to - permit a rather extensive degree of laboLirjuarket flexibility while guaranteeing the welfare of individual workers through generous income maintenance and active labour market programmes. Welfare state employment constituted, until the mid-1980s in Denmark and Sweden, the principal --,, means to absorb the huge growth of female employment, and to offset job stagnationor decline in the private sector. Since the early 1970s, public employment accounts for roughly 80 per cent of net new job growth in Denmark and Sweden. Active labour market policies have been important in recycling redundant industrial workers and minimizing youth unemployment, either via training or sheltered (mainly public) jobs. Active labour market spending, at 1-2 per cent of GDP through most of the 1980s in Denmark, Finland, and Sweden, is two or three times the OECD average (OECD, 19966). Participation in active measures may, to a degree, imply camouflaged unemployment. None the less, the Nordic countries' active investment in employment is what accounts for their doubly impressive employment record: the world's highest employment/population ratios and, until the 1990s, low unemployment. The contrast becomes even sharper when we compare older workers (55-64), whose activity rate in most Continental European countries has declined to roughly 40-50 per cent but remains at 60-70 per cent in Scandinavia. Similarly, the activity rate of single mothers in Denmark and Sweden hovers around 80 per cent, compared to 50 per cent in Germany, 30 in the UK, and 45 per cent in the United States.5 Why, then, has the Scandinavian 'third way' stranded? Does the arrival of mass unemployment everywhere except in Norway (with its ample oil revenues) give credence to the argument that there is no way to dodge the 'equality-jobs' trade-off; that some are bound to lose? To arrive at an answer we must omit Norway which, with abundant oil revenues, is able to finance itself out of the dilemma. Beyond chance effects (such as the collapse ofthe COMECON market for Finnish exports), there are a number ofexplanations which all, unfortunately, confirm the argument. The first ^ has to do "with the ali-impj^tant consensual bargaining infrastructure: since the early 1980s, the system of nation-wide concertation has decentralized; the capacity to strike mutually binding, long-term consensual 4 The successful performance of the. Netherlands since the mid-1980s is, similarly, ascribed to the capacity ofthesocial partners to holdback real wage growth (Netherlands Ministry of Social Affairs, 1996).JJ>utch real wage growth paralleled the United Stales, Australia, and New Zealand by being negative through the 1980s and early 1990s. 5 These calculations are based on second wave (1985-7) Luxemburg Income Study data files. New Social Risks in Old Welfare States 153 deals has weakened considerably (Pontusson, 1992; Swensson, 1989; Stephens, 1996). In Sweden, one of the most visible effects has been 4ncr-eased_w^ge^ud-mmm£-i^^^ ------ The second explanation is more fundamental, a true Achilles' heel in the system itself. What theJNordic countries attempted was, in a way, to 'cheat' o the Baumöl cost-disease problem through direct public-service provision-—which'-now accounts for a full one-third of total employment. In a compressed wage structure, like the Nordic, the long-run lag in public-service productivity implies that the relative cost rises; the burden of sus-taimngjthjsjaxrfinanced[sector becomes ever heavier. One solution might be to widen the wage g;ap between public- and private-sector.workers, but ■ in a thoroughly unionized labour market this is very difficult to accomplish. Thus, even if these countries were capable of sustaining long-range wage moderation across the board, the problem remains that public ser- — vices put an ever-growing, direct and indirect, cost burden on the labour market. 'Equality and jobs' therefore end up being at loggerheads regardless of strategic response. Still, strategic choice produces different outcomes and therefore also a different set of winners and losers. The Continental European model of privileging the core workforce in the context of jobless growth means the fortification of an insider-outsider divide: the high wages and job security enjoyed by (chiefly male) insiders, in effect, are also what causes the exclusion of their sons, daughters, and wives. As we have seen, Germany is the only Continental European country where unemployment is not concentrated among youth. Otherwise, most ofthe unemployed are female or young, and the duration of unemployment is exceedingly long. This is then the most likely B-team clientele, and lengthy spells of unemployment are the most probable source of entrapment. Flow-data for 1989-90 show that a small minority of those who were unemployed in 1989 had found employment a year later: 23 per cent in Belgium, 36 in France and Germany. This compares very unfavourably with Denmark's superior performance (65 per cent) (European Commission, 1993). The labour-reduction strategy is primarily based on low female labour supply and early retirement. Both contribute to raising labour costs and reproducing rigidities. Pressures for high wages and fixed labour costs come from the rising financial requirements of pension systems and also from families' dependence on one earner. The need to sustain male breadwinners' job protection is similarly over-determined by early retirement and familial dependencies. Rather than the traditional 40-50 years' unin.-terupted employment, European workers now7 start their career much later (longer education and often very long unemployment) and retire early. It is difficult to assume more than 30-5 contribution years, and this means 154 Welfare Capitalism Recast? that any disruption of a career would have catastrophic consequences for their welfare calculus. And with family dependency on the single-income -erM^^TT+he-r-isl^of^ob^seew median voter support for the existing privileges of insiders becomes virtually hegemonic. A--similar scenario obtains neither in liberal nor in social democratic regimes. The liberal deregulatory 'low-wage1 strategy may assure substantial labour market flexibility, but it also creates more poverty and income inequality which is easily compounded by the logic of deregulation itself: if wages, especially at the lower end, decline this will engender poverty traps unless also guaranteed social minima follow suit. A hallmark of American and Antipodean (but not British) adjustment policy is that more people earn low wages and that low-wage workers have lost terrain.7 But (non-pension) income maintenance benefits have eroded also in the UK. Both poverty rates and income polarization have therefore risen over the past decades. During the 1980s, the lowest decile earners have, relative to the median, lost 8 percentage points in Australia, 11 in Canada, 13 in New Zealand, 23 in the UK, and 29 in the USA (recalculations from OECD, 19966, Table 3.1). The liberal, deregulatory strategy is less likely to create entrapment in unemployment—in fact, average duration is generally low—but low-wage entrapment is a real possibility. Recent data on earnings mobility show that the likelihood of moving out of the lowest wage quintile between 1986-91 is much lower in the USA than in any of the European countries: for males, only 38 per cent of Americans escape low-wage employment compared to 55 per cent Germans, 61 per cent Italians, and 64 per cent Swedes. More or less the same pattern holds also for American youth workers, indicating that low-wage employment is not only a youthful stop-gap experience (OECD, 1996Z>, Table 3.9).8 Since the end of the Scandinavian 'third- way5 is so recent, at least in Finland and Sweden, it is difficult to identify trends and consequences, if we judge by Danish experience and recent distributional developments in Sweden, there appear to be no clear losers. There is certainly more unemployment in Finland and Sweden, especially among youth. And the share of long-term unemployed has risen—although nowhere near Continental European levels. One remarkable feature of Scandinavian unemployment is that it is much more 'democratically' spread than elsewhere: there is less 6 As we have seen. Southern European youth unemployment is almost exclusively absorbed within the family. 7 Theminimum wage as a-percentage of average earnings is 39 per cent in the USA and 40 per cent in- the UK_ In most European countries it hovers between 50 and 60 per cent (Dulado et aiT 1996: Table I). 8 Note, however, that there is little low-wage entrapment in the UK (65 per cent move up). New Social Risks in Old Welfare States 155 of a gender and age bias.9 This can be attributed to active labour market policy efforts to reactivate especially the young (OECD, 19956). A second —rerricntabluTerrtir^^ appears stronger than marker forces even when, as in Denmark, heavy unemployment has been a constant feature over the past decades. Thiscomes aut in all studies of income and welfare distribution: whether unemployment rises or wages become more unequal, this has no noticeable effect on final household income or welfare standards. On the basis of 10 welfare resource components, the Swedish Level of Living studies conclude that equality has actually risen, even during the more troubled 1980s and early 1990s (Fritzell and Lundberg, 1994). A similar conclusion derives from the Danish Level of Living studies, comparing the 1970s and 1980s (Hansen, 1995), and from more recent studies that focus especially on the weakest population groups (Ingerslev and Pedersen, 1996; Velfaerdskommissionen, 1995). For Denmark, the only Nordic country with a long experience of unemployment, evidence suggests that very few—mainly the completely unskilled— remain entrapped in marginality for long (Bjorn, 1995). Being all longitudinal, panel-type data, they also suggest that there are no major groups that show any significant degree of entrapment, be it in social exclusion or poverty. In fact, as we shall see, unusually high-risk groups, such as lone mothers, do not have poverty rates that are appreciably higher than the average. LABOUR MARKET RISKS AND WELFARE REGIMES In order to fully evaluate emerging risks, we need more than 'static5 poverty or unemployment headcounts for any given year. We also need to examine trends and, above all, the 'dynamics' of deprivation and entrapment. Table 8.1 presents recent 'static' data on exclusion, focusing on relative youth and unskilled worker unemployment, low-wage and poverty risks. Table 8.2, in turn, examines unemployment and poverty trends among the same groups over the past 10-15 years. Finally, Table 8.3 presents available 'dynamic' data on duration and on chances of mobility out of unemployment, low-wage employment, and poverty. Within our 'static' comparison, it is above all clear that risks of exclusion or poverty fall disproportionally on the shoulders of the young. Yet, the profile of the losers varies significantly across «nr welfare-regimes.10 In 9 A female bias may be evolvinginSweden, not-so much in unemployment as in declining participation- The. female participation rate has dropped from 83 per cent 1990 to "76 per cent in 1996. Male rates have, however, also seen a drop (from 88 to 82 per cent) (OECD, 19976, Table B). 10 Table 8.1 (and subsequent tables) include-only countries for which we have adequate and comparable data). 160 Welfare Capitalism Recast? they will receive either subsidized education or jobs.16 This combination is surely also what accounts for very brief poverty durations among Danish youth._____________________________________________________ . _ / It is on the income and poverty side that we can locate the reasons for why 'jobs and equality' are not automatic opposites.jWhen we examine the income situation of unemployed youth (aged 20-9), we find a puzzling orthogonality: low poverty in Scandinavia and Southern Europe;/high poverty in France and the UK. The low Nordic poverty rates are clearly linked to generous income support from the welfare state; the very high rates in the UK and France reflect very modest social assistance guarantees. But how do we account for the generally low rates in Italy and Spain where, after all, unemployment is concentrated among first-job seekers who have basically no entitlement to social transfers? The answer, of course, lies in familialism. In Denmark, virtually all (80 per cent) unemployed youth receive welfare state benefits; in Italy only 11 per cent. Yet, in both cases their household income position is basically similar (74 and 77 per cent of average household income, respectively). The level of welfare benefits in Denmark allows them to live independently (92 per cent) while, in Italy, the only realistic shield against poverty is to continue residing with parents (81 per cent of all). The worst-case scenario is found in countries like France and, especially, Britain, where neither the welfare state nor the family provide ample welfare. Hence, in comparison to Denmark, the (logisitic) odds of young unemployed living in poverty is five times higher in France and twelve times higher in Britain.17 Familializing social risks like unemployment can be an effective antidote against poverty, but it incurs indirect costs. Familial dependency may actually slow down job search, since the implicit reservation wage, rather than being, the difference between social benefits and expected earnings, now includes the value of free room and board. It also implies delayed autonomy, family formation, and fertility and here we can identify one of the principal reasons for Southern Europe's low-fertility equilibrium. In the study cited above, we found that almost 40 per cent of young Danish unemployed are parents; in Italy, virtually no one has children. 16 It isxlear that education-andtrainingpTevent against entrapment. Bjorn (1995) shows for Denmark that some vocational training raises the chance of exiting from unemployment by 30 per cent. In a comparison between Sweden and the CFK, Allmendinger and Hinz (1996, Table 7) show that unskilled Swedes bave.a. three times greater chance of moving into skilled jobs than do unskilled British workers^ehiefly due to Sweden's accent on active training programmes (see also Esping-Andersen, Rolrwer, anďSorensen, 1994). 17 Calculations based: on the 1994~Europanel. For details, see Bison and Esping-Andersen (1998). New Social Risks in Old Welfare States 161 FAMILY RISKS AND WELFARE REGIMES Family instability is the second subversive soldier in the Trojan horse of the welfare state. If traditional causes of lone parenthood, like death of a spouse, have declined in importance, this has amply been made up for by J marital instability, divorce,, or separation. The new family structure is cer- """ tainly not exclusively a source of risks. Families have fewer children,_many ^ households have none, and this means that there are fewer mouths to feed. The number of households that may find themselves in Rowntree's youthful phase in the poverty cycle is bound to be limited since so few have large numbers of children. Two-earner households have high incomes and are o resource-rich compared to the traditional one-earner model. Their phenomenal rise means that more children grow up in riches rather than poverty. Of course, two-career couples are also the ones most likely to divorce. We have seen how welfare states have responded in radically different ways to the double challenge of women's employment and family instability. Active de-familialization of welfare burdens in the social democratic regime; an essentially passive or, at most, targeted assistance approach in o the liberal; and a pohcyjrfsiigained familialism in Continental Europe— o much less in France and Belgium, much more in Italy and Spain. This should influence the profile and intensity of social risks within families. Duncan et al. (1993) compared the impact of job loss and of divorce on the likelihood of falling into poverty. As would be expected, unemployment is generally the strongest catalyst, but divorce also imposes substan- u tial risks. About 15 per cent of divorces in Canada, Sweden, and Germany result in poverty; 6 per cent in Holland and 8 per cent in the USA. The social protection package that families enjoy should, nevertheless, make a huge difference. This comes out in Table 8.4, where I have estimated post-tax and transfer poverty rates for three household types: one-earner, two-earner, and single-parent child families. There are several important points that emerge from Table 8.4. The first is that the conventional one-earner family is not a very goodhedge against child poverty given that its poverty rate is everywhere much higher than the national average. Of course, it hardly exists any more in the Nordic countries, and is also rapidly disappearing in the United States.18 There are obvious reasons for why the solo bread-winner family is vulnerable,It has only the welfare state (or relatives) to fall back on if child obligations mean reduced hours or make employment altogether impossible 18 The surprisingly high poverty rates among Swedish and Danish one-earner families is almost certainly due to a selection-bias: this kind ofhousehold is now very rare, and most likely found in rural areas. 162 Welfare Capitalism Recast? Table 8.4. Poverty rates in traditionell and new child family types Single-parent One-earner Two-earner Total . family family lar USA (1986) 61 33 14 Denmark (1987) Sweden (1987) 6 5 16 14 5 4 Germany (1984) Italy (1986) "28 23 14 23 4 4 Notes: Data include only non-aged (< 65) population. Poverty is measured as less than 50 per cent of adjusted median household income (using an equivalence scale with an elasticity = 0.5 for each additional member). Source: Luxemburg income Study data files. The two-earner household is vastly superior, with child poverty rates that are three or even four times lower (see also Mclanahan, Casper, and Sorensen, 1995). Here we can glimpse a first possible 'win-win' solution. If the contemporary economy generates severe unemployment or income risks, one counter-strategy is to nurture the double-career family. Two incomes makes the household less vulnerable to unemployment or career change. Being less dependent on male bread-winner job security and entitlements, the two-earner family is also more equipped to adapt to labour market flexibilization. And as we have discussed earlier, two-income families are eminent service consumers and, therefore, a source of job creation. Labour market and family 'failures' may, however, combine to produce zero-earner families, either because of single-person or single-parent households, or when 'unemployment comes in couples'. Due to Southern Europasi familialism and strong job protection of male earners, it is hardly surprising that most unemployed reside in a household with at least one active earner (60 per cent in Italy). This is, however, less the case in Northern Europe (Jacobs, 1998). However, Northern Europe combines countries with strong income guarantees to unemployed" households (Denmark, Belgium, the Netherlands, and Sweden), and countries with a weaker safety net (the UK and France). Thus, male unemployment in a two-adult household with-no other earner increases the odds of poverty by five times in France and the UK, but not at alt in Denmark, Belgium, the Netherlands, or Sweden (Jacobs, op. cit., Tabteróa).19 Rising divorce rates imply more potentiállyprecarious lone-parent families, almost all of which are headed by women. In Scandinavia, as one 19 In Italy, thexelative odds of poverty are much lower (1.8 times). New Social Risks in Old Welfare States 163 would expect, they do quite well with poverty levels that hardly differ from the population average; their mean disposable income falls very close to 4-h#-nati^}ftl^neftnr^ and, in the United States lone parents are more likely than not to be poor-Given the virtual impossibility for solo mothers to combine child care and work in Italy and Germany, we could have expected even more poverty. In part, they are a rather small stratum; in part, it is very likely that they move in with relatives or parents—the familialistic substitute for welfare state support.20 The massive difference between Nordic and American single-parent poverty rates would lead most observers to conclude that, yes, the welfare state is decisive. In the United States, a single parent that relies on AFDC benefits (plus food stamps and Medicaid) will not arrive above poverty level as we measure it. In Sweden, the combined benefit package assures that he or she will. None the less, it is not so much welfare benefits as employment which matters. The employment rate of Swedish lone mothers is twice as high as the American. Since in Sweden they have employment income (in 89 per cent of cases) and receive, in addition, welfare benefits (equal to 30 per cent of their total disposable income), it is in a direct sense the labour market rather than the welfare state which accounts for low poverty. In an indirect sense, of course, this is made possible by welfare state care services. In the United States, only 45 per cent work and they receive very little added public benefits (equal to only 9 per cent of their final income).21 Sweden urges lone mothers to work; America 'taxes away' their incentive to do so. The price of 'urging' may be quite high to the Swedish welfare state, but the hidden cost of not doing so may be even higher. If, like in America (and Britain), lone parents leave the labour market and have little incentive or capacity to return, they are more likely to become long-term welfare clients (as we have seen) and the net loss due to forgone lifetime earnings (and inferior life chances of their offspring) may become extremely high. 20 Own analyses of LIS data and Italian family expenditure data suggest that this is true for Italy. As in Scandinavia, their mean disposable income is very close to population average. It iurns out that a surprisingly large share of Italian solo motherswork (almost 80 per cent). But here there is a possible 'class effect'. Since divorce is very costly, it is probably a luxury of the wealthier/This, in turn, means that divorced women are more educated and perhaps also more likely to already have a career. Until we know more, the best guess is that lone parents in Italy are a^bimodal group with a minority doing poorly, a majority doing very well. 21 The data derive from Gornick (1994). The importance of the Vcn--k-plus--w-elfaxe state' package is even more evident when we examine lone-mother poverty durations. Own analyses of the Danish CLS data show a mean duration of poverty among lone mothers, equal to the population mean. Three-quarters escape within two years. 164 Welfare Capitalism Recast? It may be true, as Rosen (1996) insists, that the public cost of maintaining working mothers in Sweden is excessive in the sense that it far exceeds their relative productivity. However, his estimate that maybe 50 per cent of total Swedish expenditure to support working mothers is 'wasted' should be seriously questioned in light of the foregoing discussion. If viewed as a question of women's (and their children's) dynamic life chances and lifelong earnings power, the 'waste' may as well be considered productive investment.22 This, at least, is the reasoning behind Scandinavia's active family policies. They continue to assure that the family and the labour market remain the principal basis of welfare. They can be seen as a proactive investment against family and market welfare failure. We can examine more comparatively the relationship between welfare benefits, work, and solo-mother poverty. Although the number of available cases is very small (14), OLS regression is robust because of the excellent fit of the model. As shown in Table 8.5, employment reduces poverty by almost twice as much as do welfare benefits. It is furthermore evident that a sharp decline in social benefits when mothers choose to work provides a substantial negative work incentive. A cross-section regression estimate of its effects on solo mothers' rate of employment (where the independent variable is measured as the differential between welfare benefits to non-working and working mothers) yields a coefficient of-.99 (t = -2.63) with R-squared = .464. In other words, the trade-off is almost perfectly one-to-one: a one percentage (-.99) decline in employment for each percentage decline in welfare benefits to working mothers. WELFARE STATE ADAPTATION TO EXOGENOUS SHOCKS AND NEW RISKS The post-war welfare state was everywhere income transfer biased and service-lean, i.e. based on the idea that service needs would be met in the family. It was also primarily directed to the aged or large child families, i.e. based on the assumption that most adults would find their welfare through market employment. The Trojan horse at century's end has, as we — Welfare dependency among solo mothers in America is, for most, quite shortlived. For 4-1 per cent it "lasts but one year. But this coincides with substantial long-term dependency: 23 per cent remain for two years; almost 12 per cent for seven-plus years. (Gottschalk et «/., 1994). Hence, thercumulative cost to the public can be cruiie high.even when, as in the USA, the benefit is meagre and an estimated half of the eligible lone mothers do not even receive benefits. In their classical article Mincer and Polachek (1974) estimated that (American) women lose 1.5 per cent of lifetime income for every year they remain not employed. Exiting for five years implies a 7-5 per cent income loss. More interestingly, a Dutch study estimates that full-withdrawal leads to a 2 per cent loss per annum while remaining in the labour market on part-time leads only to a .5 percent annual loss (cited in Bruyn-Hundt, 1996: 150). New Social Risks in Old Welfare States 165 Table 8.5. The impact of social transfers and employment on single-mother poverty _____________________Coefficient T-statistic R- Constant 146.07 4.87 0.674 Social transfers -0.79 -2.16 Employment -1.29 -4-44 Note: Poverty is, as in previous tables, measured as below 50 per cent of equivalent median household income. Source: data taken from Gornick (1994). have seen, brought with it a host of new social risks. How do different welfare regimes adapt? Synthetically speaking, the inherent logic of our three welfare regimes seems to reproduce itself. The social democratic nations have responded actively with a redirection of welfare state efforts: increasingly emphasizing services and shifting resources towards younger households, both to sustain their incomes and to maximize their employment. The conservative regimes, in contrast, have responded 'passively' in a double sense of the word: firstly, the old transfer bias remains and has in fact strengthened; passive income maintenance has been their major weapon in dealing with new risks. Secondly, the response is passive in the sense that new risks and new responsibilities are largely relegated to the family. The liberal, deregulatory approach is also passive, but in this case because reliance on markets for welfare solutions has been strengthened. The marketization of risks is evident on several fronts. Besides outright privatization, which is limited, the stress on 'workfare' is one; the move towards some form of negative income tax or low-income credit' to working families is another (Myles, 1996). In different words, the liberal strategy combines a scaled-back social wage and the 're-commodification' of labour, albeit with subsidies. These differences are apparent in the emerging shape of contemporary welfare states. The 'proactive' strategy finds expression in a strong move towards servicing and youth welfare in Scandinavia. The 'passive familial-istic' welfare states remain transfer biased and are increasingly being reduced to 'pensioner states'. Trends in the liberal group,, however, give: some support to Castles' (1996) claim that we must distinguish the'hardcore' liberal welfare states, like the American, from the 'lib-lab' model,Jike Australia. Although in all cases there is a sustained move towards more welfare state residualism, targeting, and means-testing, Australia and, less so,. New Zealand, stand out for their growing attention to young families. See Table 8.6. 166 Welfare Capitalism Recast? Table 8.6. Welfare state adaptation: trends in servicing and age bias, 1980-1992 Servicing bias Age bias Ratio Change Ratio Change services/cash 1980-92 youth/aged 1980-92 1992 (%pts) 1992 (%pts) Liberal cluster 0.12 +0.04 0.80 +0.11 Social democratic cluster 0.34 +0.07 1.02 +0.18 Conservative cluster 0.08 +0.04 0.51 -0.08 Japan 0.08 0 0.18 -0.12 Notes: Health care is excluded from the services-cash transfer calculations. Programmes defined as being youth-oriented include: occupational injury, sickness cash benefits, family cash und service benefits, active labour market programmes, and unemployment benefits. Aged-oriented programmes are old age survivors' benefits and services. Disability pensions and 'other1 have been excluded from calculations altogether. Source: OECD Social Expenditure data base (SOCX). The regime means presented in Table 8.6 do not disguise large internal deviations except in very few cases: Australia, Ireland, and New Zealand are quite Scandinavian in their bias towards the young. And in the conservative cluster, the Netherlands stands out as less aged-biased than average (see detailed country data in Appendix Table 8A). In brief, the same cases which earlier posed problems for a regime classification, remain ambiguous. Japan, however, reasserts its affinity with the conservative regime. The startling differences lie not just in the relative weight accorded to youth and services within the three regimes, but also in trends. The liberal group is, as signalled, bifurcated with one youth-oriented group (Australia, Ireland, and perhaps New Zealand) and one that is becoming ever more aged-biased and service-lean (the UK and the USA). The assumption that markets should reign for working-age adults seems especially strong in the latter. The conservative regime (with the possible exception of the Netherlands) is, and remains, service-lean.and the age bias is being systematically reinforced. Since this hardly represents a-push to prioritize markets—except in the sense that households_are assumed to rely on a protected bread-winner—the implicit drift is to reproduce familialism. CONCLUSION A 'win-win' adjustment strategy implies managing the new risk structure in a way that no groups become systematic losers. And if I am right in New Social Risks in Old Welfare States 167 asserting that youth and young families are being disproportionally bombarded from all sides with risks of poverty, low income, unemployment ■Jmd^rjgjdiap_s^ in marginali7atiojvjadmt4S-calkd-4Q^4s-4iQt- more, nor less, welfare state but a major overhaul; a reprioritization of goals, a recast emphasis in favour of young families and, especially, their servicing needs. In order to return to the ideals that Beveridge, Moller, and others implanted in their welfare blueprints, a positive-sum strategy is clearly one in which the state nurtures markets' and families' capacity to maximize welfare. Perhaps the single greatest obstacle to such a reform strategy comes, paradoxically, from the welfare state's most ardent supporters—the evermore aged median voter and, where insider-outsider cleavages are especially pronounced, from interest organizations. It is really only the social democratic welfare states (with Australia) which have responded and adapted themselves in any meaningful way along these lines. Their ability to minimize family 'diswelfare', poverty, and inequalities and to maximize employment during the hard times of the recent decades has not been, as a neo-iiberal might believe, because they 'tax and spend' themselves out of everything. Total public social expenditure as a percentage of GDP has been stagnant in Denmark and Sweden since 1980; in Norway and Finland, there has been some rise but this is largely a 'catch-up' effect. Neither has the reprioritization of public welfare occurred at the expense of others, say the aged. The latest available data on aged poverty confirm Scandinavia's previous record: internationally extremely low poverty among the elderly.23 It is tempting to attribute Scandinavian-cum-Australian welfare state adaptability to their singularly comprehensive and co-ordinated systems of interest representation—especially when we consider that these are, at the same time, comparatively youthful societies. Neither element, taken singularly, seems to offer much of an explanation. Austria and Germany are prototypes of broad, centralized consensus building; they are also unusually 'aged' nations. North America, Britain (and France) are more youthful, but lack any form of comprehensive interest intermediation. In any case, the least adaptable welfare states seem to concentrate in nations, like Italy, that are neither youthful nor co-ordinated. The contemporary debate is pitched in terms of a perhaps lamentable, but sadly unavoidable, trade-off. Yet, the evidence L have examined suggests that the trade-off differs in terms: of intensity ancLconsequences from one regime to another; within regimes, from one country to another. Most 23 Among retired couples, the poverty rates for 1991-2are the following: Denmark = 3 per cent; Finland = 3 per cent; Norway = 2 per cent; Sweden = ! per cenL This compares with 3 per cent in the Netherlands, Germany, and UK and 13 percent in the USA (own calculations from LIS data bases). 168 Welfare Capitalism Recast? nations are currently struggling with how to maximize, or at least maintain existing standards of equality and social citizenship while, simultaneously, generate more jobs and less exclusion. The basic argument I make is that any such objective cannot be attained by simply rolling back or augmenting the welfare state. The contemporary debate is also pitched inj terms of a wrong-headed set of political alternatives. If Lam right, a posi-.J tive-sum 'win-win' strategy means rebuilding the welfare state so that it,; once again, can assume well-functioning, welfare-producing labour mar- : kets and families. This means changing the nexus between state and mar- ; ket as well as between state and families. In order to identify potential positive-sum strategies, we first need a \ basic principle for judging the desirability of one option against a rival ; one. Current thinking is dominated by two rival principles: the more modestly formulated Paretian and the more ambitious Rawlsian. Reduced to the bare bones, a Pareto improvement implies that if someone gains from any change in the status quo there should, at least, be no concomitant loser. The relevant 'someone' may be a group, a social class, a 'chrono-class', or a gender. A Rawlsian improvement (maximum possible justice i and liberty), following his 'difference' and 'priority' principle, implies that í any change of the status quo should be of greatest advantage to the least well off(Rawls, 1972: 83 ff.). A Paretian improvement favours 'equity'—closer to the spirit of liberalism; a Rawlsian favours 'equality'—closer to the social democratic spirit. The former suffers from a basic shortcoming as far as welfare state reform is concerned. Since it is perfectly possible to obtain a Pareto improvement when one millionaire doubles his income while the poor continue to starve, this is clearly a rather irrelevant criterion for any welfare state analysis. We must begin with the assumption that we desire welfare state reform in order to improve its capacity to deliver welfare, not in order to make the rich richer. We must also assume that the basic societal objective is to reconsolidate the 'democratic class struggle'; to prevent the emergence of an ever-deepening abyss between the A-teams and B-teams of postindustrial society. Accordingly, any 'win-win'strategy of welfare state reform must, by definition, lie closer to a Rawlsian principle of justice. It is this which underpins my admittedly more speculative .approach in. the next, and final chapter. The basic question that guides me is this: What kind of welfare regime will potentiallyTTTOvide positive-sum, or 'win-win' solutions to the dominant dilemmas of postindustrial society? New Social Risks in Old Welfare States 169 APPENDIX Table 8A. Servicing and age bias of welfare states Service bias Age bias Ratio Change Ratio Change services/cash 1980-92 youth/aged 1980-92 1992 (% pts) 1992 (% pts) Australia 0.20 +0.17 1.16 +0.69 Canada 0.08 +0.01 0.78 -0.52 Ireland 0.16 +0.10 1.12 +0.31 New Zealand 0.07 +0.01 0.91 +0.33 UK 0.13 -0.02 0.50 -0.10 USA 0.06 -0.02 0.30 -0.05 Denmark 0.33 +0.01 1.37 +0.15 Finland 0.21 -0.04 0.94 +0.34 Norway 0.38 +0.23 0.83 +0.17 Sweden 0.45 +0.08 0.92 +0.04 Austria 0.03 +0.02 0.3.1 -0.09 Belgium 0.08 +0.07 0.61 -0.21 France 0.13 +0.04 0.48 -0.07 Germany (West) 0.12 +0.03 0.43 -0.05 Italy 0.07 +0.03 0.20 -0.09 Netherlands 0.12 0 0.79 -0.28 Portugal 0.10 +0.08 0.55 +0.09 Spain 0.05 +0.02 0.67 +0.04 Japan 0.08 0 0.18 -0.12 Source and explanations, see Table 8.6.