Introduction
The welfare state has been a favored topic of research for many years now. This is not very surprising when we consider its fantastic pace of growth in most countries during the 1960s and 1970s. What once were night-watchman states, law-and-order states, militarist states, or even repressive organs of totalitarian rule, are now institutions predominantly preoccupied with the production and distribution of social well-being. To study the welfare state is therefore a means to understand a novel phenomenon in the history of capitalist societies.
In the league of advanced capitalist democracies, states clearly vary considerably with regard to their accent on welfare. Even if the lion's share of expenditures or personnel serves welfare aims, the kind of welfare provided will be qualitatively different, as will its prioritization relative to competing activities, be they warfare, law and order, or the promotion of profits and trade.
The historical characteristics of states have played a determinant role in forging the emergence of their welfare-statism. In his recent book, Giddens (1985) highlights the causal influence of wars, a factor which has been almost wholly neglected in the large literature on welfare-state origins. In our account, this argument cannot be confronted directly. Yet, it is given some indirect support in our emphasis on the relative strength of absolutist and authoritarian rule. The leading theme in our account, however, is that the history of political class coalitions is the most decisive cause of welfare-state variations.
The welfare state has been approached both narrowly and broadly. Those who take the narrower view see it in terms of the traditional terrain of social amelioration: income transfers and social services, with perhaps some token mention of the housing question. The broad-
2 introduction
er view often frames its questions in terms of political economy, its interests focused on the state's larger role in managing and organizing the economy. In the broader view, therefore, issues of employment, wages, and overall macro-economic steering are considered integral components in the welfare-state complex. In a sense, this approach identifies its subject matter as the 'Keynesian welfare state' or, if you like, 'welfare capitalism*.
In this book, we follow the broad approach; this is why we begin with the issues of classical and modern political economy, and why we devote the last third of the book to issues of employment and general macro-economic steering. This is also why we prefer to employ terms such as 'welfare capitalism* and 'welfare-state regimes'.
'Welfare-state regimes' is, in a way, the organizing concept of the book. The reasons are several. First, as it is commonly used, the concept of the welfare state is too narrowly associated with the conventional social-amelioration policies. Second, what we will show is that contemporary advanced nations cluster not only in terms of how their traditional social-welfare policies are constructed, but also in terms of how these influence employment and general social structure. To talk of *a regime* is to denote the fact that in the relation between state and economy a complex of legal and organizational features are systematically interwoven.
The broader approach implies a trade-off. Since our intention is to understand the 'big picture*, we shall not be able to dwell on the detailed characteristics of the various social programs. So, when we study pensions, for example, our concern is not pensions per se, but the ways in which they elucidate how different nations arrive at their peculiar public-private sector mix. A related trade-off is that large-scale comparisons, such as ours, prohibit detailed treatments of individual countries. I am convinced that readers knowledgeable about any of the 18 nations included in the study will feel that my treatment of 'their* country is superficial, if not outright misrepresentative. This is unfortunately the price to be paid for making grand comparisons, given the intellectual limitations of the author and the page limitations set by the publisher.
This book has been written with two beliefs in mind. The first is that existing theoretical models of the welfare state are inadequate. The ambition is to offer a reconceptualization and re-theorization on the basis of what we consider important about the welfare state. The existence of a social program and the amount of money spent on it may be less important than what it does. We shall devote many pages to arguing that issues of de-commodification, social stratification, and
introduction 3
employment are keys to a welfare state's identity. The second belief is that only comparative empirical research will adequately disclose the fundamental properties that unite or divide modern welfare states. The distant dream of social science is to formulate laws of societal motion. Whether formulated in terms of the logic of capitalism, industrialism, modernization, or nation-building, they nearly always posit similar and convergent evolutionary paths. Obviously, laws are not supposed to have deviant cases.
The comparative approach is meant to (and will) show that welfare states are not all of one type. Indeed, the study presented here identifies three highly diverse regime-types, each organized around its own discrete logic of organization, stratification, and societal integration. They owe their origins to different historical forces, and they follow qualitatively different developmental trajectories.
In the first chapter, our task is to reintegrate the welfare-state debate into the intellectual tradition of political economy. This serves to bring into sharper focus the principal theoretical questions involved. On this basis we will be in a better position to specify the salient characteristics of welfare states. The convention of conceptualizing welfare states in terms of their expenditures will no longer do. In a sense, our ultimate goal is to 'sociologize' the study of welfare states. Most studies have assumed a world of linearity: of more or less power, industrialization, or spending. We will in this book understand welfare states as clustering into three different types of regime that we have labeled conservative, liberal, and 'social democratic'. Their crystallization and subsequent development can hardly be explained with analytical parsimony.
In chapters 2, 3, and 4 we offer a reconceptualization of what we believe to be the salient characteristics of welfare states. The extension of social rights has always been regarded as the essence of social policy. Inspired by the contributions of Karl Polanyi, we choose to view social rights in terms of their capacity for 'de-commodification'. The outstanding criterion for social rights must be the degree to which they permit people to make their living standards independent of pure market forces. It is in this sense that social rights diminish citizens' status as 'commodities'.
Social stratification is part and parcel of welfare states. Social policy is supposed to address problems of stratification, but it also produces it. Equality has always been what welfare states were supposed to produce, yet the image of equality has always remained rather vague. In some analyses it is simply taken for granted that social benefits diminish inequalities. In others, the focus is on the eradication of
4 introduction
poverty or the overall distribution of income. The really neglected issue is the welfare state as a stratification system in its own right. Does it enhance or diminish existing status or class differences; does it create dualisms, individualism, or broad social solidarity? These are the issues of chapter 3.
Both social rights and social stratification are shaped by the nexus of state and market in the distribution system. To a social democrat, reliance on the market for the basic means of welfare is problematic because it fails to provide inalienable rights and because it is inequitable. To a laissez-faire liberal, reliance on the welfare state is dangerous because it cripples freedom and efficiency. In chapter 4, we examine how the interplay of public and private sector has contributed to the crystallization of the pension-mix in different welfare-state regimes. The point is two-fold. First, we cannot grasp the welfare state without locating its activities in relation to the private sector. Second, it is a myth to think that either markets or the state are more naturally equipped to develop welfare. Instead, markets are often politically created and form an integral part of the overall welfare-state regime.
Part I of the book develops the dimensions of comparative welfare states, and demonstrates the clustering of advanced capitalist democracies into three distinct regimes. Part II examines how this came to be. In this analysis we can obviously not limit ourselves to why some welfare states score more or less than others on some attribute. We have to account for why the world is composed of three qualitatively different welfare-state logics. In chapter 5, we adopt the standard comparative correlational approach to identify the relative importance of political forces in the creation of welfare states. In line with the prevailing academic consensus today, we must conclude that politics not only matters, but is decisive. In contrast to most studies, however, it is not necessarily the political mobilization of the working classes that matters here. For some regimes, their role has been marginal and we must instead understand the evolution of welfare states here as the result of the state's history of nation-building and/or the influence of conservatism and Catholicism. We have tried to embed our explanations in the political histories of nations.
The second part of the book broadens the field of investigation considerably. Here the focus is not so much on what created welfare states as on what their effects are on our economies. Specifically, we examine three facets of welfare-state-employment interactions. To begin with, in chapter 6 we lay out an argument for why labor-market structures are closely tied to welfare-state regimes. We show that the coincidence of the two is striking, and that the behavioral characteris-
introduction 5
tics of labor markets cross-nationally depend on how the welfare state is constructed.
In chapters 7 and 8, we examine in greater detail how welfare states affect employment by selecting a representative country from each of our three types of regime. In chapter 7, the focus is oh nations' capacities to maintain full employment; in chapter 8, it is on the post-industrial transformation of employment structures. In the former chapter, we analyze how welfare states have become key institutions in managing the dilemmas and tensions that emerge with a full-employment commitment. In the latter, we argue that it is false to believe in the emergence of a general post-industrial employment path. We identify three qualitatively diverse trajectories, each of which owes its dynamic to the structuring of the welfare state. We conclude that each trajectory produces its own stratification outcome, and results, therefore, in very different conflict scenarios.
The book, then, sees the welfare state as a principal institution in the construction of different models of post-war capitalism. Hence, the choice of its title, The Three Worlds of Welfare Capitalism.
1
The Three Political Economies of the Welfare State*
The Legacy of Classical Political Economy
Most debates on the welfare state have been guided by two questions. First, will the salience of class diminish with the extension of social citizenship? In other words, can the welfare state fundamentally transform capitalist society? Second, what are the causal forces behind welfare-state development?
These questions are not recent. Indeed, they were formulated by the nineteenth-century political economists 100 years before any welfare state can rightly be said to have come into existence. The classical political economists - whether of liberal, conservative, or Marxist persuasion - were preoccupied with the relationship between capitalism and welfare. They certainly gave different (and usually normative) answers, but their analyses converged around the relationship between market (and property), and the state (democracy).
Contemporary neo-liberalism is very much an echo of classical liberal political economy. For Adam Smith, the market was the superior means for the abolition of class, inequality, and privilege. Aside from a necessary minimum, state intervention would only stifle the equalizing process of competitive exchange and create monopolies, protectionism, and inefficiency: the state upholds class; the market can potentially undo class society (Smith, 1961, II, esp, pp. 232-6).1
Liberal political economists were hardly of one mind when it came
* This chapter is adapted from an article which previously appeared in the Canadian Review of Sociology and Anthropology, Vol. 26:2 (1989) under the title "The three political economies of the welfare state'.
10   the three welfare-state regimes
to policy advocacy. Nassau Senior and later Manchester liberals emphasized the laissez-faire element in Smith, rejecting any form of social protection outside the cash nexus. J. S. Mill and the 'reformed liberals', in turn, were proponents of a modicum of political regulation. Yet they all were agreed that the road to equality and prosperity should be paved with a maximum of free markets and a minimum of state interference.
Their enthusiastic embrace of market capitalism may now appear unjustified. But we must not forget that the reality they spoke of was a state upholding absolutist privileges, mercantilist protectionism, and pervasive corruption. What they attacked was a system of government that repressed their ideals of both freedom and enterprise. Hence, theirs was revolutionary theory, and from this vantage point, we can understand why Adam Smith sometimes reads like Karl Marx.2
Democracy became an Achilles' heel to many liberals. As long as capitalism remained a world of small property owners, property itself would have little to fear from democracy. But with industrialization, the proletarian masses emerged, for whom democracy was a means to curtail the privileges of property. The liberals rightly feared universal suffrage, for it would be likely to politicize.the distributional struggle, pervert the market, and fuel inefficiencies. Many liberals discovered that democracy would usurp or destroy the market.
Both conservative and Marxist political economists understood this contradiction, but proposed, of course, opposite solutions. The most coherent conservative critique of laissez-faire came from the German historical school, in particular from Friedrich List, Adolph Wagner, and Gustav Schmoller. They refused to believe that the raw cash-nexus of the market was the only or the best guaranteee of economic efficiency. Their ideal was the perpetuation of patriarchy and absolutism as the best possible legal, political, and social shell for a capitalism without class struggle.
One prominent conservative school promoted the 'monarchical welfare state', which would guarantee social welfare, class harmony, loyalty, and productivity. In this model, an efficient production system comes not from competition, but from discipline. An authoritarian state would be far superior to the chaos of markets in harmonizing the good of the state, community, and individual.3
Conservative political economy emerged in reaction to the French Revolution and the Paris Commune. It was avowedly nationalistic and anti-revolutionary, and sought to arrest the democratic impulse. It feared social leveling, and favored a society that retained both hierar-
the three political economies of the welfare-state 11
chy and class. Status, rank, and class were natural and given; class conflicts, however, were not. If we permit democratic mass participation, and allow authority and status boundaries to dissolve, the result is a collapse of the social order.
Marxist political economy not only abhorred the market's atomizing effects, but also attacked the liberal claim that markets guarantee equality. Since, as Dobb (1946) puts it, capital accumulation disowns people of property, the end result will be ever-deeper class divisions. And as these generate sharpened conflicts, the liberal state will be forced to shed its ideals of freedom and neutrality, and come to the defence of the propertied classes. For Marxism this is the foundation of class dominance.
The central question, not only for Marxism but for the entire contemporary debate on the welfare state, is whether, and under what conditions, the class divisions and social inequalities produced by capitalism can be undone by parliamentary democracy.
Fearing that democracy might produce socialism, the liberals were hardly eager to extend it. The socialists, in contrast, suspected that parliamentarism would be little more than an empty shell or, as Lenin suggested, a mere 4talking shop' (Jessop, 1982). This line of analysis, echoed in much of contemporary Marxism, produced the belief that social reforms were little more than a dike in a steadily leaking capitalist order. By definition, they could not be a response to the desire of the working classes for emancipation.4
It took major extensions of political rights before the socialists could wholeheartedly embrace a more optimistic analysis of parliamentarism. The theoretically most sophisticated contributions came from the Au-stro-German Marxists such as Adler, Bauer, and Eduard Heimann. According to Heimann (1929), it may have been the case that conservative reforms were motivated by little else than a desire to repress labor mobilization. But once introduced, they become contradictory: the balance of class power is fundamentally altered when workers enjoy social rights, for the social wage lessens the worker's dependence on the market and employers, and thus turns into a potential power resource. To Heimann, social policy introduces an alien element into the capitalist political economy. It is a Trojan horse that can penetrate the frontier between capitalism and socialism. This intellectual position has enjoyed quite a renaissance in recent Marxism (Offe, 1985; Bowles and Gintis, 1986).
The social democratic model, as outlined above, did not necessarily abandon the orthodoxy that, ultimately, fundamental equality requires
12    the three welfare-state regimes
economic socialization. Yet historical experience soon demonstrated that socialization was a goal that could not be pursued realistically through parliamentarism.5
Social democracy's embrace of parliamentary reformism as its dominant strategy for equality and socialism was premised on two arguments. The first was that workers require social resources, health, and education to participate effectively as socialist citizens. The second argument was that social policy is not only emancipatory, but is also a precondition for economic efficiency (Myrdal and Myrdal, 1936). Following Marx, in this argument the strategic value of welfare policies is that they help promote the onward march of the productive forces in capitalism. But the beauty of the social democratic strategy was that social policy would also result in power mobilization. By eradicating poverty, unemployment, and complete wage dependency, the welfare state increases political capacities and diminishes the social divisions that are barriers to political unity among workers.
The social democratic model, then, is father to one of the leading hypotheses of contemporary welfare-state debate: parliamentary class-mobilization is a means for the realization of the socialist ideals of equality, justice, freedom, and solidarity.
The Political Economy of the Welfare State
Our forebears in political economy defined the analytic basis of much recent scholarship. They isolated the key variables of class, state, market, and democracy, and they formulated the basic propositions about citizenship and class, efficiency and equality, capitalism and socialism. Contemporary social science distinguishes itself from classical political economy on two scientifically vital fronts. First, it defines itself as a positive science and shies away from normative prescription (Robbins, 1976). Second, classical political economists had little interest in historical variability: they saw their efforts as leading towards a system of universal laws. Although contemporary political economy sometimes still clings to the belief in absolute truths, the comparative and historical method that today underpins almost all good political economy is one that reveals variation and permeability.
Despite these differences, most recent scholarship has as its foca] point the state-economy relationship defined by nineteenth-century political economists. And, given the enormous growth of the welfare state, it is understandable that it has become a major test case for contending theories of political economy.
the three political economies of the welfare-state 13
We shall review below the contributions of comparative research on the development of welfare states in advanced capitalist countries. It will be argued that most scholarship has been misdirected, mainly because it became detached from its theoretical foundations. We must therefore recast both the methodology and the concepts of political economy in order to adequately study the welfare state. This will constitute the focus of the final section of this chapter.
Two types of approach have dominated in explanations of welfare states; one stresses structures and whole systems, the other, institutions and actors.
The systems/structuralist approach
Systems or structuralist theory seeks to capture the logic of development holistically. It is the system that 'wills', and what happens is therefore easily interpreted as a functional requisite for the reproduction of society and economy. Because its attention is concentrated on the laws of motion of systems, this approach is inclined to emphasize cross-national similarities rather than differences; being industrialized or capitalist over-determines cultural variations or differences in power relations.
One variant begins with a theory of industrial society, and argues that industrialization makes social policy both necessary and possible -necessary because pre-industrial modes of social reproduction, such as the family, the church, noblesse oblige, and guild solidarity are destroyed by the forces attached to modernization, such as social mobility, urbanization, individualism, and market dependence. The crux of the matter is that the market is no adequate substitute because it caters only to those who are able to perform in it. Hence, the 'welfare function' is appropriated by the nation-state.
The welfare state is also made possible by the rise of modern bureaucracy as a rational, universalist, and efficient form of organization. It is a means for managing collective goods, but also a center of power in its own right, and it will thus be inclined to promote its own growth. This kind of reasoning has informed the so-called 'logic of industrialism' perspective, according to which the welfare state will emerge as the modern industrial economy destroys traditional social institutions (Flora and Alber, 1981; Pryor, 1969). But the thesis has difficulties explaining why government social policy only emerged 50 and sometimes even 100 years after traditional community was effectively destroyed. The basic response draws on Wagner's Law of 1883 (Wagner, 1962) and on Alfred Marshall (1920) - namely that a certain
14    the three welfare-state regimes
level of economic development, and thus surplus, is needed in order to permit the diversion of scarce resources from productive use (investment) to welfare (Wilensky and Lebeaux, 1958). In this sense, this perspective follows in the footsteps of the old liberals. Social redistribution endangers efficiency, and only at a certain economic level will a negative-sum trade-off be avoidable (Okun, 1975).
The new structuralist Marxism is strikingly parallel. Abandoning its classical forebears' strongly action-centered theory, its analytical starting-point is that the welfare state is an inevitable product of the capitalist mode of production. Capital accumulation creates contradictions that compel social reform (O'Connor, 1973). In this tradition of Marxism, as in its iogic of industrialism' counterpart, welfare states hardly need to be promoted by political actors, whether they be unions, socialist parties, humanitarians, or enlightened reformers. The point is that the state, as such, is positioned in such a way that the collective needs of capital are served, regardless. The theory is thus premised on two crucial assumptions: first, that power is structural, and second, that the state is 'relatively' autonomous from class directives (Poulantzas, 1973; Block, 1977; for a recent critical assessment of this literature, see Therborn, 1986a; and Skocpol and Amenta, 1986).
The 'logic of capitalism' perspective invites difficult questions. If, as Przeworski (1980) has argued, working-class consent is assured on the basis of material hegemony, that is, self-willed subordination to the system, it is difficult to see why up to 40 percent of the national product must be allocated to the legitimation activities of a welfare state. A second problem is to derive state activities from a 'mode of production' analysis. Eastern Europe may perhaps not qualify as socialist, but neither is it capitalist. Yet there we find 'welfare states', too. Perhaps accumulation has functional requirements no matter how it proceeds? (Skocpol and Amenta, 1986; Bell, 1978).
The institutional approach
The classical political economists made it clear why democratic institutions should influence welfare-state development. The liberals feared that full democracy might jeopardize markets and inaugurate socialism. Freedom, in their view, necessitated a defence of markets against political intrusion. In practice, this is what the laissez-faire state sought to accomplish. But it was this divorce of politics and economy which fuelled much institutionalist analysis. Represented best by Polanyi (1944), but also by a number of anti-democratic exponents of the historical school, the institutional approach insists that any effort to
the three political economies of the welfare-state 15
isolate the economy from social and political institutions will destroy human society. The economy must be embedded in social communities in order for it to survive. Thus, Polanyi sees social policy as one necessary precondition for the reintegration of the social economy.
An interesting recent variant of institutional alignment theory is the argument that welfare states emerge more readily in small, open economies that are particularly vulnerable to international markets. As Katzenstein (1985) and Cameron (1978) show, there is a greater inclination to regulate class-distributional conflicts through government and interest concertation when both business and labor are captive to forces beyond domestic control.
The impact of democracy on welfare states has been argued ever since J. S. Mill and Alexis de Tocqueville. The argument is typically phrased without reference to any particular social agent or class. It is in this sense that it is institutional. In its classical formulation, the thesis was simply that majorities will favor social distribution to compensate for market weakness or market risks. If wage-earners are likely to demand a social wage, so are capitalists (or farmers) apt to demand protection in the form of tariffs, monopoly, or subsidies. Democracy is an institution that cannot resist majority demands.
In its modern formulations, the democracy thesis has many variants. One identifies stages of nation-building in which the extension of full citizenship must also include social rights (Marshall, 1950; Bendix, 1964; Rokkan, 1970). A second variant, developed by both pluralist and public-choice theory, argues that democracy will nurture intense party competition around the median voter which, in turn, will fuel rising public expenditure. Tufte (1978), for example, argues that major extensions of public intervention occur around elections as a means of voter mobilization.
This approach also faces considerable empirical problems (Skocpol and Amenta, 1986). When it holds that welfare states are more likely to develop the more democratic rights are extended, the thesis confronts the historical oddity that the first major welfare-state initiatives occurred prior to democracy and were powerfully motivated by the desire to arrest its realization. This was certainly the case in France under Napoleon III, in Germany under Bismarck, and in Austria under von Taaffe. Conversely, welfare-state development was most retarded where democracy arrived early, such as in the United States, Australia, and Switzerland. This apparent contradiction can be explained, but only with reference to social classes and social structure: nations with early democracy were overwhelmingly agrarian and dominated by small property owners who used their electoral powers
6   the three welfare-state regimes
to reduce, not raise, taxes (Dich, 1973). In contrast, ruling classes in authoritarian polities were better positioned to impose high taxes on an unwilling populace.
Social Class as a Political Agent
We have noted that the case for a class-mobilization thesis flows from social democratic political economy. It differs from structuralist and institutional analyses in its emphasis on the social classes as the main agents of change, and in its argument that the balance of class power determines distributional outcomes. To emphasize active class-mobilization does not necessarily deny the importance of structured or hegemonic power (Korpi, 1983). But it is held that parliaments are, in principle, effective institutions for the translation of mobilized power into desired policies and reforms. Accordingly, parliamentary politics is capable of overriding hegemony, and can be made to serve interests that are antagonistic to capital. Further, the class-mobilization theory assumes that welfare states do more than simply alleviate the current ills of the system: a 'social democratic' welfare state will, in its own right, establish critical power resources for wage-earners, and thus strengthen labor movements. As Heimann (1929) originally held, social rights push back the frontiers of capitalist power.
The question of why the welfare state itself is a power resource is vital for the theory's applicability. The answer is that wage-earners in the market are inherently atomized and stratified compelled to compete, insecure, and dependent on decisions and forces beyond their control. This limits their capacity for collective solidarity and mobilization. The social rights, income security, equalization, and eradication of poverty that a universalistic welfare state pursues are necessary preconditions for the strength and unity that collective power mobilization demands (Esping-Andersen, 1985a).
The single most difficult problem for this thesis is to specify the conditions for power mobilization. Power depends on the resources that flow from electoral numbers and from collective bargaining. Power mobilization, in turn, depends on levels of trade-union organization, share of votes, and parliamentary and cabinet seats held by left or labor parties. But the power of one agent cannot simply be indicated by its own resources: it will depend on the resources of contending forces, on the historical durability of its mobilization, and on patterns of power alliances.
There are several valid objections to the class-mobilization thesis. Three in particular are quite fundamental. One is that the locus of
the three political economies of the welfare-state 17
decision-making and power may shift from parliaments to neo-corporatist institutions of interest intermediation (Shonfield, 1965; Schmitter and Lembruch, 1979). A second criticism is that the capacity of labor parties to influence welfare-state development is circumscribed by the structure of right-wing party power. Castles (1978; 1982) has argued that the degree of unity among the conservative parties is more important than is the activated power of the left. Other authors have emphasized the fact that denominational (usually social Catholic) parties in countries such as Holland, Italy, and Germany mobilize large sections of the working classes and pursue welfare-state programs not drastically at variance with their socialist competitors (Schmidt, 1982; Wilensky, 1981). The class-mobilization thesis has, rightly, been criticized for its Swedocentrism, i.e. its inclination to define the process of power mobilization too much on the basis of the rather extraordinary Swedish experience (Shalev, 1984).
These objections hint at a basic fallacy in the theory's assumptions about the class formation: we cannot assume that socialism is the natural basis for wage-earner mobilization. Indeed, the conditions under which workers become socialists are still not adequately documented. Historically, the natural organizational bases of worker mobilization were pre-capitalist communities, especially the guilds, but also the Church, ethnicity, or language. A ready-made reference to false consciousness will not do to explain why Dutch, Italian, or American workers continue to mobilize around non-socialist principles. The dominance of socialism among the Swedish working class is as much a puzzle as is the dominance of confessionalism among the Dutch.
The third and perhaps most fundamental objection has to do with the model's linear view of power. It is problematic to hold that a numerical increase in votes, unionization, or seats will translate into more welfare-statism. First, for socialist as for other parties, the magical '50 percent' threshold for parliamentary majorities seems practically insurmountable (Przeworski, 1985). Second, if socialist parties represent working classes in the traditional sense, it is clear that they will never succeed in their project. In very few cases has the traditional working class been numerically a majority; and its role is rapidly becoming marginal.6
Probably the most promising way to resolve the combined linearity and working-class minority problem lies in recent applications of Barrington Moore's path-breaking class-coalition thesis to the transformation of the modern state (Weir and Skocpol, 1985; Gourevitch, 1986; Esping-Andersen,  1985a; , Esping-Andersen and Friedland,
18    the three welfare-state regimes
1982). Thus, the origins of the Keynesian full-employment commitment and the social democratic welfare-state edifice have been traced to the capacity of (variably) strong working-class movements to forge a political alliance with farmer organizations; additionally, it is arguable that sustained social democracy has come to depend on the formation of a new-working-class-white-collar coalition.
The class-coalitional approach has additional virtues. Two nations, such as Austria and Sweden, may score similarly on working-class mobilization variables, and yet produce highly unequal policy results. This can be explained by differences in the history of coalition formation in two countries: the breakthrough of Swedish social democratic hegemony stems from its capacity to forge the famous 'red-green' alliance with the farmers; the comparative disadvantage of the Austrian socialists rests in the 'ghetto' status assigned to them by virtue of the rural classes being captured by a conservative coalition (Esping-Andersen and Korpi, 1984).
In summary, we have to think in terms of social relations, not just social categories. Whereas structural functionalist explanations identify convergent welfare-state outcomes, and class-mobilization paradigms see large, but linearly distributed, differences, an interactive model such as the coalition approach directs attention to distinct welfare-state regimes.
What is the Welfare State?
Every theoretical paradigm must somehow define the welfare state. How do we know when and if a welfare state responds functionally to the needs of industrialism, or to capitalist reproduction and legitimacy? And how do we identify a welfare state that corresponds to the demands that a mobilized working class might have? We cannot test contending arguments unless we have a commonly shared conception of the phenomenon to be explained.
A remarkable attribute of the entire literature is its lack of much genuine interest in the welfare state as such. Welfare-state studies have been motivated by theoretical concerns with other phenomena, such as power, industrialization, or capitalist contradictions; the welfare state itself has generally received scant conceptual attention. If welfare states differ, how do they differ? And when, indeed, is a state a welfare state? This turns attention straight back to the original question: what is the welfare state?
A common textbook definition is that it involves state responsibility
the three political economies of the welfare-state 19
for securing some basic modicum of welfare for its citizens. Such a definition skirts the issue of whether social policies are emancipatory or not; whether they help system legitimation or not; whether they contradict or aid the market process; and what, indeed, is meant by 'basic'? Would it not be more appropriate to require of a welfare state that it satisfies more than our basic or minimal welfare needs?
The first generation of comparative studies started with this type of conceptualization. They assumed, without much reflection, that the level of social expenditure adequately reflects a state's commitment to welfare. The theoretical intent was not really to arrive at an understanding of the welfare state, but rather to test the validity of contending theoretical models in political economy. By scoring nations with respect to urbanization, level of economic growth, and the proportion of aged in the demographic structure, it was believed that the essential features of industrial modernization were properly considered. Alternatively, power-oriented theories compared nations on left-party strength or working-class power mobilization.
The findings of the first-generation comparativists are difficult to evaluate, since there is no convincing case for any particular theory. The shortage of nations for comparisons statistically restricts the number of variables that can be tested simultaneously. Thus, when Outright (1965) or Wilensky (1975) find that economic level, with its demographic and bureaucratic correlates, explains most welfare-state variations in 'rich countries', relevant measures of working-class mobilization or economic openness are not included. Their conclusions in favor of a 'logic of industrialism' view are therefore in doubt. And, when Hewitt (1977), Stephens (1979), Korpi (1983), Myles (1984a), and Esping-Andersen (1985b) find strong evidence in favor of a working-class mobilization thesis, or when Schmidt (1982; 1983) finds support for a neo-corporatist, and Cameron (1978) for an economic openness argument, it is without fully testing against plausible alternative explanations.7
Most of these studies claim to explain the welfare state. Yet their focus on spending may be misleading. Expenditures are epiphenomenal to the theoretical substance of welfare states. Moreover, the linear scoring approach (more or less power, democracy, or spending) contradicts the sociological notion that power, democracy, or welfare' are relational and structured phenomena. By scoring welfare states on spending, we assume that all spending counts equally. But some welfare states, the Austrian one, for example, spend a large share on benefits to privileged civil servants. This is normally not what we would consider a commitment to social citizenship and solidarity. Others spend disproportionately on means-tested social assistance. Few contemporary
20   the three welfare-state regimes
analysts would agree that a reformed poor-relief tradition qualifies as a welfare-state commitment. Some nations spend enormous sums on fiscal welfare in the form of tax privileges to private insurance plans that mainly benefit the middle classes. But these tax expenditures do not show up on expenditure accounts. In Britain, total social expenditure has grown during the Thatcher period, yet this is almost exclusively a function of very high unemployment. Low expenditure on some programs may signifiy a welfare state more seriously committed to full employment.
Therborn (1983) is right when he holds that we must begin with a conception of state structure. What are the criteria with which we should judge whether, and when, a state is a welfare state? There are three approaches to this question. Therborn's proposal is to begin with the historical transformation of state activities. Minimally, in a genuine welfare state the majority of its daily routine activities must be devoted to servicing the welfare needs of households. This criterion has far-reaching consequences. If we simply measure routine activity in terms of spending and personnel, the result is that no state can be regarded as a real welfare state until the 1970s, and some that we normally label as welfare states will not qualify because the majority of their routine activities concern defence, law and order, administration, and the like (Therborn, 1983). Social scientists have been too quick to accept nations' self-proclaimed welfare-state status. They have also been too quick to conclude that if the standard social programs have been introduced, the welfare state has been born.
The second conceptual approach derives from Richard Titmuss's (1958) classical distinction between residual and institutional welfare states. In the former, the state assumes responsibility only when the family or the market fails; it seeks to limit its commitments to marginal and deserving social groups. The latter model addresses the entire population, is universalistic, and embodies an institutionalized commitment to welfare. It will, in principle, extend welfare commitments to all areas of distribution vital for societal welfare.
The Titmuss approach has fertilized a variety of new developments in comparative welfare-state research (Myles, 1984a; Korpi, 1980; Esping-Andersen and Korpi, 1984; 1986; Esping-Andersen, 1985b; 1987b). It is an approach that forces researchers to move from the black box of expenditures to the content of welfare states: targeted versus universalistic programs, the conditions of eligibility, the quality of benefits and services, and, perhaps most importantly, the extent to which employment and working life are encompassed in the state's extension of citizen rights. The shift to welfare-state typologies makes simple linear
the three political economies of the welfare-state 21
welfare-state rankings difficult to sustain. Conceptually, we are comparing categorically different types of states.
The third approach is to theoretically select the criteria on which to judge types of welfare states. This can be done by measuring actual welfare states against some abstract model and then scoring programs, or entire welfare states, accordingly (Day 1978; Myles, 1984a). But this is ahistorical, and does not necessarily capture the ideals or designs that historical actors sought to realize in the struggles over the welfare state. If our aim is to test causal theories that involve actors, we should begin with the demands that were actually promoted by those actors that we deem critical in the history of welfare-state development. It is difficult to imagine that anyone struggled for spending per se.
A Re-Specification of the Welfare State
Few can disagree with T. H. Marshall's (1950) proposition that social citizenship constitutes the core idea of a welfare state. But the concept must be fleshed out. Above all, it must involve the granting of social rights. If social rights are given the legal and practical status of property rights, if they are inviolable, and if they are granted on the basis of citizenship rather than performance, they will entail a de-commodification of the status of individuals vis-a-vis the market. But the concept of social citizenship also involves social stratification: one's status as a citizen will compete with, or even replace, one's class position.
The welfare state cannot be understood just in terms of the rights it grants. We must also take into account how state activities are interlocked with the market's and the family's role in social provision. These are the three main principles that need to be fleshed out prior to any theoretical specification of the welfare state.
Rights and de-commodification
In pre-capitalist societies, few workers were properly commodities in the sense that their survival was contingent upon the sale of their labor power. It is as markets become universal and hegemonic that the welfare of individuals comes to depend entirely on the cash nexus. Stripping society of the institutional layers that guaranteed social reproduction outside the labor contract meant that people were com-modified. In turn, the introduction of modern social rights implies a loosening of the pure'commodity status. De-commodification occurs
22    the three welfare-state regimes
when a service is rendered as a matter of right, and when a person can maintain a livelihood without reliance on the market.
The mere presence of social assistance or insurance may not necessarily bring about significant de-commodification if they do not substantially emancipate individuals from market dependence. Means-tested poor relief will possibly offer a safety net of last resort. But if benefits are low and associated with social stigma, the relief system will compel all but the most desperate to participate in the market. This was precisely the intent of the nineteenth-century poor laws in most countries. Similarly, most of the early social-insurance programs were deliberately designed to maximize labor-market performance (Ogus, 1979).
There is no doubt that de-commodification has been a hugely contested issue in welfare state development. For labor, it has always been a priority. When workers are completely market-dependent, they are difficult to mobilize for solidaristic action. Since their resources mirror market inequalities, divisions emerge between the 'ins' and the 'outs', making labor-movement formation difficult. De-commodification strengthens the worker and weakens the absolute authority of the employer. It is for exactly this reason that employers have always opposed de-commodification.
De-commodified rights are differentially developed in contemporary welfare states. In social-assistance dominated welfare states, rights are not so much attached to work performance as to demonstrable need. Needs-tests and typically meager benefits, however, service to curtail the de-commodifying effect. Thus, in nations where this model is dominant (mainly in the Anglo-Saxon countries), the result is actually to strengthen the market since all but those who fail in the market will be encouraged to contract private-sector welfare.
A second dominant model espouses compulsory state social insurance with fairly strong entitlements. But again, this may not automatically secure substantial de-commodification, since this hinges very much on the fabric of eligibility and benefit rules. Germany was the pioneer of social insurance, but over most of the century can hardly be said to have brought about much in the way of de-commodification through its social programs. Benefits have depended almost entirely on contributions, and thus on work and employment. In other words, it is not the mere presence of a social right, but the corresponding rules and preconditions, which dictate the extent to which welfare programs offer genuine alternatives to market dependence.
The third dominant model of welfare, namely the Beveridge-type citizens' benefit, may, at first glance, appear the most de-commodifying.
the three political economies of the welfare-state 23
It offers a basic, equal benefit to all, irrespective of prior earnings, contributions, or performance. It may indeed be a more solidaristic system, but not necessarily de-commodifying, since only rarely have such schemes been able to offer benefits of such a standard that they provide recipients with a genuine option to working.
De-commodifying welfare states are, in practice, of very recent date. A minimal definition must entail that citizens can freely, and without potential loss of job, income, or general welfare, opt out of work when they themselves consider it necessary. With this definition in mind, we would, for example, require of a sickness insurance that individuals be guaranteed benefits equal to normal earnings, and the right to absence with minimal proof of medical impairment and for the duration that the individual deems necessary. These conditions, it is worth noting, are those usually enjoyed by academics, civil servants, and higher-echelon white-collar employees. Similar requirements would be made of pensions, maternity leave, parental leave, educational leaye, and unemployment insurance.
Some nations have moved towards this level of de-commodification, but only recently, and, in many cases, with significant exemptions. In almost all nations, benefits were upgraded to nearly equal normal wages in the late 1960s and early 1970s. But in some countries, for example, prompt medical certification in case of illness is still required; in others, entitlements depend on long waiting periods of up to two weeks; and in still others, the duration of entitlements is very short. As we shall see in chapter 2, the Scandinavian welfare states tend to be the most de-commodifying; the Anglo-Saxon the least.
The Welfare State as a System of Stratification
Despite the emphasis given to it in both classical political economy and in T. H. Marshall's pioneering work, the relationship between citizenship and social class has been neglected both theoretically and empirically. Generally speaking, the issue has either been assumed away (it has been taken for granted that the welfare state creates a more egalitarian society), or it has been approached narrowly in terms of income distribution or in terms of whether education promotes upward social mobility. A more basic question, it seems, is what kind of stratification system is promoted by social policy. The welfare state is not just a mechanism that intervenes in, and possibly corrects, the structure of inequality; it is, in its own right, a system of stratification. It is an active force in the ordering of social relations.
24   the three welfare-state regimes
Comparatively and historically, we can easily identify alternative systems of stratification embedded in welfare states. The poor-relief tradition, and its contemporary means-tested social-assistance offshoot, was conspicuously designed for purposes of stratification. By punishing and stigmatizing recipients, it promotes social dualisms and has therefore been a chief target of labor-movement attacks.
The social-insurance model promoted by conservative reformers such as Bismarck and von Taffe, was also explicitly a form of class politics. It sought, in fact, to achieve two simultaneous results in terms of stratification. The first was to consolidate divisions among wage-earners by legislating distinct programs for different class and status groups, each with its own conspicuously unique set of rights and privileges which was designed to accentuate the individual's appropriate station in life. The second objective was to tie the loyalties of the individual directly to the monarchy or the central state authority. This was Bismarck's motive when he promoted a direct state supplement to the pension benefit. This state-corporatist model was pursued mainly in nations such as Germany, Austria, Italy, and France, and often resulted in a labyrinth of status-specific insurance funds.
Of special importance in this corporatist tradition was the establishment of particularly privileged welfare provisions for the civil service (Beamten). In part, this was a means of rewarding loyalty to the state, and in part it was a way of demarcating this group's uniquely exalted social status. The corporatist status-differentiated model springs mainly from the old guild tradition. The neo-absolutist autocrats, such as Bismarck, saw in this tradition a means to combat the rising labor movements.
The labor movements were as hostile to the corporatist model as they were to poor relief - in both cases for obvious reasons. Yet the alternatives first espoused by labor were no less problematic from the point of view of uniting the workers as one splidaristic class. Almost invariably, the model that labor first pursued was that of self-organized friendly societies or equivalent union- or party-sponsored fraternal welfare plans. This is not surprising. Workers were obviously suspicious of reforms sponsored by a hostile state, and saw their own organizations not only as bases of class mobilization, but also as embryos of an alternative world of solidarity and justice; as a microcosm of the socialist haven to come. Nonetheless, these micro-socialist societies often became problematic class ghettos that divided rather than united workers. Membership was typically restricted to the strongest strata of the working class, and the weakest - who most needed protection - were
the three political economies of the welfare-state 25
most likely excluded. In brief, the fraternal society model frustrated the goal of working-class mobilization.
The socialist 'ghetto approach* was an additional obstacle when socialist parties found themselves forming governments and having to pass the social reforms they had so long demanded. For political reasons of coalition-building and broader solidarity, their welfare model had to be recast as welfare for 'the people*. Hence, the socialists came to espouse the principle of universalism; borrowing from the liberals, their program was, typically, designed along the lines of the democratic fiat-rate, general revenue-financed Beveridge model.
As an alternative to means-tested assistance and corporatist social insurance, the universalistic system promotes equality of status. All citizens are endowed with similar rights, irrespective of class or market position. In this sense, the system is meant to cultivate cross-class solidarity, a solidarity of the nation. But the solidarity of flat-rate universalism presumes a historically peculiar class structure, one in which the vast majority of the population are the 'little people* for whom a modest, albeit egalitarian, benefit may be considered adequate. Where this no longer obtains, as occurs with growing working-class prosperity and the rise of the new middle classes, flat-rate universalism inadvertently promotes dualism because the better-off turn to private insurance and to fringe-benefit bargaining to supplement modest equality with what they have decided are accustomed standards of welfare. Where this process unfolds (as in Canada or Great Britain), the result is that the wonderfully egalitarian spirit of universalism turns into a dualism similar to that of the social-assistance state: the poor rely on the state, and the remainder on the market.
It is not only the universalist but, in fact, all historical welfare-state models which have faced the dilemma of changes in class structure. But the response to prosperity and middle-class growth has been varied, and so, therefore, has been the outcome in terms of stratification. The corporatist insurance tradition was, in a sense, best equipped to manage new and loftier welfare-state expectations since the existing system could technically be upgraded quite easily to distribute more adequate benefits. Adenauer's 1957 pension-reform in Germany was a pioneer in this respect. Its avowed purpose was to restore status differences that had been eroded because of the old insurance system's incapacity to provide benefits tailored to expectations. This it did simply by moving from contribution- to earnings-graduated benefits without altering the framework of status-distinctiveness.
In nations with either a social-assistance or a universalistic Beveridge-
26    the three welfare-state regimes
type system, the option was whether to allow the market or the state to furnish adequacy and satisfy middle-class aspirations. Two alternative models emerged from this political choice. The one typical of Great Britain and most of the Anglo-Saxon world was to preserve an essentially modest universalism in the state, and allow the market to reign for the growing social strata demanding superior welfare. Due to the political power of such groups, the dualism that emerges is not merely one between, state and market, but also between forms of welfare-state transfers: in these nations, one of the fastest growing components of public expenditure is tax subsidies for so-called 'private* welfare plans. And the typical political effect is the erosion of middle-class support for what is less and less a universalistic public-sector transfer system.
Yet another alternative has been to seek a synthesis of universalism and adequacy outside of the market. This road has been followed in countries where, by mandating or legislation, the state incorporates the new middle classes within a luxurious second-tier, universally inclusive, earnings-related insurance scheme on top of the flat-rate egalitarian one. Notable examples are Sweden and Norway. By guaranteeing benefits tailored to expectations, this solution reintroduces benefit inequalities, but effectively blocks off the market. It thus succeeds in retaining universalism and also, therefore, the degree of political consensus required to preserve broad and solidaristic support for the high taxes that such a welfare-state model demands.
Welfare-State Regimes
As we survey international variations in social rights and welfare-state stratification, we will find qualitatively different arrangements between state, market, and the family. The welfare-state variations we find are therefore not linearly distributed, but clustered by regime-types.
In one cluster we find the 'liberal' welfare state, in which means-tested assistance»modest universal transfers, or modest social-insurance plans predominate. Benefits cater mainly to a clientele of low-income, usually working-class, state dependents. In this model, the progress of social reform has been severely circumscribed by traditional, liberal work-ethic norms: it is one where the limits of welfare equal the marginal propensity to opt for welfare instead of work. Entitlement rules are therefore strict and often associated with stigma; benefits are typically modest. In turn, the state encourages the market, either
the three political economies of the welfare-state 27
passively - by guaranteeing only a minimum - or actively - by subsidizing private welfare schemes.
The consequence is that this type of regime minimizes de-commodification-effects, effectively contains the realm of social rights, and erects an order of stratification that is a blend of a relative equality of poverty among state-welfare recipients, market-differentiated welfare among the majorities, and a class-political dualism between the two. The archetypical examples of this model are the United States, Canada and Australia.
A second regime-type clusters nations such as Austria, France, Germany, and Italy. Here, the historical corporatist-statist legacy was upgraded to cater to the new 'post-industrial' class structure. In these conservative and strongly 'corporatism welfare states, the liberal obsession with market efficiency and commodification was never preeminent and, as such, the granting of social rights was hardly ever a seriously contested issue. What predominated was the preservation of status differentials; rights, therefore, were attached to class and status. This corporatism was subsumed under a state edifice perfectly ready to displace the market as a provider of welfare; hence, private insurance and occupational fringe benefits play a truly marginal role. On the other hand, the state's emphasis on upholding status differences means that its redistribute impact is negligible.
But the corporatist regimes are also typically shaped by the Church, and hence strongly committed to the preservation of traditional family-hood. Social insurance typically excludes non-working wives, and family benefits encourage motherhood. Day care, and similar family services, are conspicuously underdeveloped; the principle of 'subsidiarity' serves to emphasize that the state will only interfere when the family's capacity to service its members is exhausted.
The third, and clearly smallest, regime-cluster is composed of those countries in which the principles of universalism and de-commodification of social rights were extended also to the new middle classes. We may call it the 'social democratic' regime-type since, in these nations, social democracy was clearly the dominant force behind social reform. Rather than tolerate a dualism between state and market, between working class and middle class, the social democrats pursued a welfare state that would promote an equality of the highest standards, not an equality of minimal needs as was pursued elsewhere. This implied, first, that services and benefits be upgraded to levels commensurate with even the most discriminating tastes of the new middle classes; and, second, that equality be furnished by guaranteeing workers full participation in the quality of rights enjoyed by the better-off.
28   the three welfare-state regimes
This formula translates into a mix of highly de-commodifying and universalistic programs that, nonetheless, are tailored to differentiated expectations. Thus, manual workers come to enjoy rights identical to those of salaried white-collar employees or civil servants; all strata are incorporated under one universal insurance system, yet benefits are graduated according to accustomed earnings. This model crowds out the market, and consequently constructs an essentially universal solidarity in favor of the welfare state. All benefit; all are dependent; and all will presumably feel obliged to pay.
The social democratic regime's policy of emancipation addresses both the market and the traditional family. In contrast to the corporatist-subsidiarity model, the principle is not to wait until the family's capacity to aid is exhausted, but to preemptively socialize the costs of family-hood. The ideal is not to maximize dependence on the family, but capacities for individual independence. In this sense, the model is a peculiar fusion of liberalism and socialism. The result is a welfare state that grants transfers directly to children, and takes direct responsibility of caring for children, the aged, and the helpless. It is, accordingly, committed to a heavy social-service burden, not only to service family needs but also to allow women to choose work rather than the household.
Perhaps the most salient characteristic of the social democratic regime is its fusion of welfare and work. It is at once genuinely committed to a full-employment guarantee, and entirely dependent on its attainment. On the one side, the right to work has equal status to the right of income protection. On the other side, the enormous costs of maintaining a solidaristic, universalistic, and de-commodifying welfare system means that it must minimize social problems and maximize revenue income. This is obviously best done with most people working, and the fewest possible living off of social transfers.
Neither of the two alternative regime-types espouse full employment as an integral part of their welfare-state commitment. In the conservative tradition, of course, women are discouraged from working; in the liberal ideal, concerns of gender matter less than the sanctity of the market.
In the chapters to follow, we show that welfare states cluster, but we must recognize that there is no single pure case. The Scandinavian countries may be predominantly social democratic, but they are not free of crucial liberal elements. Neither are the liberal regimes pure types. The American social-security system is redistributive, compulsory, and far from actuarial. At least in its early formulation, the New Deal was as social democratic as was contemporary Scandinavian social democracy.
the three political economies of the welfare-state 29
And European conservative regimes have incorporated both liberal and social democratic impulses. Over the decades, they have become less corporativist and less authoritarian.
Notwithstanding the lack of purity, if our essential criteria for defining welfare states have to do with the quality of social rights, social stratification, and the relationship between state, market, and family, the world is obviously composed of distinct regime-clusters. Comparing welfare states on scales of more or less or, indeed, of better or worse, will yield highly misleading results.
The Causes of Welfare-State Regimes
If welfare states cluster into three distinct regime-types, we face a substantially more complex task of identifying the causes of welfare-state differences. What is the explanatory power of industrialization, economic growth, capitalism, or working-class political power in accounting for regime-types? A first superficial answer would be: very little. The nations we study are all more or less similar with regard to all but the variable of working-class mobilization. And we find very powerful labor movements and parties in each of the three clusters.
A theory of welfare-state developments must clearly reconsider its causal assumptions if it wishes to explain clusters. The hope of finding one single powerful causal force must be abandoned; the task is to identify salient interaction-effects. Based on the preceding arguments, three factors in particular should be of importance: the nature of class mobilization (especially of the working class); class-political coalition structures; and the historical legacy of regime institutionalization.
As we have noted, there is absolutely no compelling reason to believe that workers will automatically and naturally forge a socialist class identity; nor is it plausible that their mobilization will look especially Swedish. The actual historical formation of working-class collectivities will diverge, and so also will their aims, ideology, and political capacities. Fundamental differences appear both in trade-unionism and party development. Unions may be sectional or in pursuit of more universal objectives; they may be denominational or secular; and they may be ideological or devoted to business-unionism. Whichever they are, it will decisively affect the articulation of political demands, class cohesion, and the scope for labor-party action. It is clear that a working-class mobilization thesis must pay attention to union structure.
The structure of trade-unionism may or may not be reflected in labor-party formation. But under what conditions are we likely to
30   the three welfare-state regimes
expect certain welfare-state outcomes from specific party configurations? There are many factors that conspire to make it virtually impossible to assume that any labor, or left-wing, party will ever be capable, single-handedly, of structuring a welfare state. Denominational or other divisions aside, it will be only under extraordinary historical circumstances that a labor party alone will command a parliamentary majority long enough to impose its will. We have noted that the traditional working class has hardly ever constituted an electoral majority. It follows that a theory of class mobilization must look beyond the major leftist parties. It is a historical fact that welfare-state construction has depended on political coalition-building. The structure of class coalitions is much more decisive than are the power resources of any single class.
The emergence of alternative class coalitions is, in part, determined by class formation. In the earlier phases of industrialization, the rural classes usually constituted the largest single group in the electorate. If social democrats wanted political majorities, it was here that they were forced to look for allies. One of history's many paradoxes is that the rural classes were decisive for the future of socialism. Where the rural economy was dominated by small, capital-intensive family fanners, the potential for an alliance was greater than where it rested on large pools of cheap labor. And where farmers were politically articulate and well-organized (as in Scandinavia), the capacity to negotiate political deals was vastly superior.
The role of the farmers in coalition formation and hence in welfare-state development is clear. In the Nordic countries, the necessary conditions obtained for a broad red-green alliance for a full-employment welfare state in return for farm-price subsidies. This was especially true in Norway and Sweden, where farming was highly precarious and dependent on state aid. In the United States, the New Deal was premised on a similar coalition (forged by the Democratic Party), but with the important difference that the labor-intensive South blocked a truly universalistic social security system and opposed further welfare-state developments. In contrast, the rural economy of continental Europe was very inhospitable to red-green coalitions. Often, as in Germany and Italy, much of agriculture was labor-intensive; hence the unions and left-wing parties were seen as a threat. In addition, the conservative forces on the continent had succeeded in incorporating farmers into 'reactionary' alliances, helping to consolidate the political isolation of labor.
Political dominance was, until after World War II, largely a question of rural class politics. The construction of welfare states in this period
the three political economies of the welfare-state 31
was, therefore, dictated by whichever force captured the farmers. The absence of a red-green alliance does not necessarily imply that no welfare-state reforms were possible. On the contrary, it implies which political force came to dominate their design. Great Britain is an exception to this general rule, because the political significance of the rural classes eroded before the turn of the century. In this way, Britain's coalition-logic showed at an early date the dilemma that faced most other nations later; namely, that the rising white-collar strata constitute the linchpin for political majorities. The consolidation of welfare states after World War II came to depend fundamentally on the political alliances of the new middle classes. For social democracy, the challenge was to synthesize working-class and white-collar demands without sacrificing the commitment to solidarity.
Since the new middle classes have, historically, enjoyed a relatively privileged position in the market, they have also been quite successful in meeting their welfare demands outside the state, or, as civil servants, by privileged state welfare. Their employment security has traditionally been such that full employment has been a peripheral concern. Finally, any program for drastic income-equalization is likely to be met with great hostility among a middle-class clientele. On these grounds, it would appear that the rise of the new middle classes would abort the social democratic project and strengthen a liberal welfare-state formula.
The political leanings of the new middle classes have, indeed, been decisive for welfare-state consolidation. Their role in shaping the three welfare-state regimes described earlier is clear. The Scandinavian model relied almost entirely on social democracy's capacity to incorporate them into a new kind of welfare state: one that provided benefits tailored to the tastes and expectations of the middle classes, but nonetheless retained universalism of rights. Indeed, by expanding social services and public employment, the welfare state participated directly in manufacturing a middle class instrumentally devoted to social democracy.
In contrast, the Anglo-Saxon nations retained the residual welfare-state model precisely because the new middle classes were not wooed from the market to the state. In class terms, the consequence is dualism. The welfare state caters essentially to the working class and the poor. Private insurance and occupational fringe benefits cater to the middle classes. Given the electoral importance of the latter, it is quite logical that further extensions of welfare-state activities are resisted.
The third, continental European, welfare-state regime has also been patterned by the new middle classes, but in a different way. The cause is historical. Developed by conservative political forces, these regimes
32   the three welfare-state regimes
institutionalized a middle-class loyalty to the preservation of both occupationally segregated social-insurance programs and, ultimately, to the political forces that brought them into being. Adenauer's great pension-reform in 1957 was explicitly designed to resurrect middle-class loyalties.
Conclusion
We have here- presented an alternative to a simple class-mobilization theory of welfare-state development. It is motivated by the analytical necessity of shifting from a linear to an interactive approach with regard to both welfare states and their causes. If we wish to study welfare states, we must begin with a set of criteria that define their role in society. This role is certainly not to spend or tax; nor is it necessarily that of creating equality. We have presented a framework for comparing welfare states that takes into consideration the principles for which the historical actors have willingly united and struggled. When we focus on the principles embedded in welfare states, we discover distinct regime-clusters, not merely variations of 'more' or 'less* around a common denominator.
The historical forces behind the regime differences are interactive. They involve, first, the pattern of working-class political formation and, second, political coalition-building in the transition from a rural economy to a middle-class society. The question of political coalition-formation is decisive. Third, past reforms have contributed decisively to the institutionalization of class preferences and political behavior. In the corporatist regimes, hierarchical status-distinctive social insurance cemented middle-class loyalty to a peculiar type of welfare state. In liberal regimes, the middle classes became institutionally wedded to the market. And in Scandinavia, the fortunes of social democracy over the past decades were closely tied to the establishment of a middle-class welfare state that benefits both its traditional working-class clientele and the new white-collar strata. The Scandinavian social democrats were able to achieve this in part because the private welfare market was relatively undeveloped and in part because they were capable of building a welfare state with features of sufficient luxury to satisfy the wants of a more discriminating public. This also explains the extraordinarily high cost of Scandinavian welfare states.
But a theory that seeks to explain welfare-state growth should also be able to understand its retrenchment or decline. It is generally believed that welfare-state backlash movements, tax revolts, and roll-backs are
the three political economies of the welfare-state 33
ignited when social expenditure burdens become too heavy. Paradoxically, the opposite is true. Anti-welfare-state sentiments over the past decade have generally been weakest where welfare spending has been heaviest, and vice versa. Why?
The risks of welfare-state backlash depend not on spending, but on the class character of welfare states. Middle-class welfare states, be they social democratic (as in Scandinavia) or corporatist (as in Germany), forge middle-class loyalties. In contrast, the liberal, residualist welfare states found in the United States, Canada and, increasingly, Britain, depend on the loyalties of a numerically weak, and often politically residual, social stratum. In this sense, the class coalitions in which the three welfare-state regime-types were founded, explain not only their past evolution but also their future prospects.
Notes
1 Adam Smith is often cited but rarely read. A closer inspection of his writings reveals a degree of nuance and a battery of reservations that substantially qualify a delirious enthusiasm for the blessings of capitalism.
2 In The Wealth of Nations (1961, II, p. 236), Smith comments on states that uphold the privilege and security of the propertied as follows: 'civil government, so far as it is instituted for the security of property, is in reality instituted for the defence of the rich against the poor, or of those who have some property against those who have none at all.'
3 This tradition is virtually unknown to Anglo-Saxon readers since so little has been translated into English. A key text which greatly influenced public debate and later social legislation was Adolph Wagner's Rede Ueber die Soziale Frage (1872). For an English language overview of this tradition of political economy, see Schumpeter (1954), and especially Bower (1947).
From the Catholic tradition, the fundamental texts are the two Papal Encyclicals, Kerum Novarum (1891) and Quadrogesimo Anno (1931). The social Catholic political economy's main advocacy is a social organization where a strong family is integrated in cross-class corporations, aided by the state in terms of the subsidiarity principle. For a recent discussion, see Richter (1987).
Like the liberals, the conservative political economists also have their contemporary echoes, although substantially fewer in number. A revival occurred with Fascism's concept of the corporative {Standische) state of Ottmar Spann in Germany. The subsidiarity principle still guides much of German Christian Democratic politics (see Richter, 1987).
4 Chief proponents of this analysis are the German 'state derivation' school (Müller and Neususs, 1973); Offe (1972); O'Connor (1973); Gough (1979); and also the work of Poulantzas (1973). As Skocpol and Amenta (1986) note
34    the three welfare-state regimes
in their excellent overview, the approach is far from one-dimensional. Thus, Offe, O'Connor and Gough identify the function of social reforms as also being concessions to mass demands and as potentially contradictory.
Historically, socialist opposition to parliamentary reforms was motivated less by theory than by reality. August Bebel, the great leader of German social democracy, rejected Bismarck's pioneering social legislation not because he did not favor social protection, but because of the blatantly anti-socialist and divisionary motives behind Bismarck's reforms.
5 This realization came from two types of experiences. One, typified by Swedish socialism in the 1920s, was the discovery that not even the working-class base showed much enthusiasm for socialization. In fact, when the Swedish socialists established a special commission to prepare plans for socialization, it concluded after ten years of exploration that it would be quite impossible to undertake practically. A second kind of experience, typified by the Norwegian socialists and Blum's Popular Front government in 1936, was the discovery that radical proposals could easily be sabotaged by the capitalists' capacity to withhold investments and export their capital abroad.
6 This is obviously not a problem for the parliamentary class hypothesis alone; structural Marxism faces the same problem of specifying the class character of the new middle classes. If such a specification fails to demonstrate that it constitutes a new working class, both varieties of Marxist theory face severe (although not identical) problems.
7 This literature has been reviewed in great detail by a number of authors. See, for example, Wilensky et al. (1985). For excellent and more critical evaluations, see Uusitalo (1984), Shalev (1983), and Skocpol and Amenta (1986).