194 The Transition from Capitalism to Socialism power in Swedish civil society put production politics back on to the immediate political agenda. The growth of white-collar labour ^'organisation and the release of non-manual employees from the j hegemony of capital brought first employee power in the enterprise j and, ultimately, collective ownership to the centre of the labour \ movement's programme. Though the question of collective ownership is yet to be decided, we think we have established conclusively that the transition to socialism has not and will not be prevented by the absorption of the Social Democratic labour movement into the capitalist system. The inability of Social Democracy in Sweden to move forward on production politics in earlier times was the result of a deadlock with capital, not a sell-out. Even if the 1979 election does not bring the Social Democrats to power, the future of socialism in Sweden will not be decided. The events of the last few years show that some immediate problems of the Swedish labour movement require a socialist solution, thus the issue is likely to be raised again. Furthermore, the renewal of socialism in Sweden shows that it is at least possible for other reformist socialist parties to bring socialism back on the agenda given the right conditions. The transition to socialism depends not only on what the labour movement does but also on what capital does. Will the introduction of employee capital funds spark capital exit and investment stops by domestic capital? And will international capital perceive the Swedish model as a threat and orchestrate some kind of censure? Any assessment of these future problems involves speculation which we leave to the conclusion. TO SOO/lLiJM/ \f. 495" 7 Conclusion Marx's faulty predictions on the questions of the development of the class structure and consciousness formation led him to think that the transition from capitalism to socialism would occur much more rapidly than it has. The development of the new middle class and the more gradual decline in the petty bourgeois sector, particularly the rural portion, has meant that even under the best of conditions the development of labour organisation and class consciousness has proceeded at a much slower rate than the early Marxists imagined. They also underestimated the strength of religious, ethnic and linguistic ties and, like virtually all academics to date, failed to recognise the substantial differences in the economic structure of various countries. Because of variation in ethnic and linguistic diversity, the strength of religious ties and economic centralisation, Western capitalist countries vary substantially in the level of labour organisation.and the strength of the socialist parties. Consequently, the political economies of various countries vary considerably. Countries where labour and the left are weak are moving toward a corporate collectivist pattern. Where labour and the left are strong, the political economies are transitional forms between capitalism and socialism. Despite the fact that the Western labour movement has not yet fulfilled the historic role assigned to it by Marx, there is no denying that its achievements have been substantial. In almost every country, the working class has been a major actor in the struggle for political democracy. Though the introduction of democracy did not lead to socialism, it did ultimately lead to a compromise on the part of the bourgeoisie, albeit after an authoritarian interlude in some cases. The welfare state was the essential element of that compromise. The expansion of the welfare state was a post-Second World War i ku r i ľ .' i i i i ii—i -f?5 196 The Transition from Capitalism to Socialism phenomenon, but in every country that experienced even medium welfare state build-up, plans were laid before the end of the war and the alignment of political forces destined to be instrumental in welfare state expansion was already determined. In every case except the United Kingdom, the welfare statist coalition contained the organised industrial working class in coalition with urban or rural petty bourgeois elements.1 In the 1930s, the Danish and Swedish Social Democrats formed a stable coalition government with a petty bourgeois party. The Norwegian Social Democrats followed a similar pattern though without formal cabinet collaboration. The Social Democrats in Belgium had already entered a coalition with the Catholics in the 1920s, a pattern which was followed much later by the Dutch. In Scandinavia, the Netherlands, Belginm, and to some extent the United Kingdom, the growing strength of organised labour also necessitated significant concessions from employers. The French situation, as we have seen, was somewhat different. A coalition between working-class and petty bourgeois forces did initiate welfare state development. But the decentralised nature of the French economy and the resultant paternalism on the part of employers and the erratic character of French unionism impeded any compromise in the industrial sector. In the long run, these factors, particularly trade union character, undermined the strength of theFrench left. Similarly, the defeat of fascism and the discrediting of the old ruling coalitions in those countries brought compromises between the labour movement and the formerly intransigent bourgeoisie. The most dramatic case of compromise was Austria, where Catholic conservative forces that had eliminated democracy rather than compromise with the labour movement, now agreed to rule in an uneasy coalition with their former enemies. This coalition, which dominated Austrian politics until the Social Democrats won a majority in the 1970s, was responsible for very substantial expansion of the welfare state. The point of this historical sketch is to reinforce several points made throughout this book. First, the class structure and the weakness of labour organisations prevented the socialist movement from pushing through its more radical programmes in the inter-war period The coalition with petty bourgeois elements forced socialisation off the immediate programme during the inter-war period Conclusion 197 and the 1940s and early 1950s. The cold war contributed to keeping the issue in the background. Second, the social and economic rights associated with the welfare state are products of class struggle. After their attempt to limit the gains of the working-class movement to mere political rights failed, the bourgeoisie either compromised or aided the downfall of democracy. When the latter strategy was discredited by military defeat and the holocaust, the bourgeoisie everywhere settled for a compromise. But the terms of the compromise varied substantially according to the strength of organised labour and the socialist parties. Moreover, the variation among capitalist countries increased in the whole post-war period. Immediately after the Second World War all countries stood at the beginning of welfare state development and the levels of expenditure were similar. By 1970, the differences were striking with some countries spending less than a quarter of GDP on non-military public goods and services and other countries spending almost one-half. Differences can be seen in the variation in control patterns in the economy, too, as our discussion of pensions in the United Kingdom, the United States and Sweden showed. Furthermore, as Table 4.8 showed, the variation in labour organisation indicates that the underlying distribution of power in capitalist democracies is increasingly different from country to country. Thus the character of the political economies has diverged too, some countries moving towards a corporate collectivist pattern, others, towards democratic socialism. The inability of the left to move beyond the welfare state has been due to its weakness more than anything. There is a certain amount of truth to the Downsian model of democracy: parties do compete for the middle-of-the-road voter. But what determines where the middle of the road is - where the mean opinion lies - is the distribution of power in civil society. Yet even accepting the prevailing distribution of opinion, the electoral record of socialist parties has been poor. Looking back at Table 4.9, one can see that only in Scandinavia does the socialist period of incumbency exceed that of the centre and right in the period 1945-70. Given this, it is not surprising that only a few countries have moved much beyond the welfare state. Taking the Swedish case as a model, we would expect / the labour movement to bring production politics to the centre of its j programme when (1) the political necessity of a coalition with petty i bourgeois forces is ended; (2) white-collar employees are 198 The Transition from Capitalism to Socialism í thoroughly organised; and (3) the socialists have experienced a i substantial period alone in the government. The only other country whick fits these three criteria is Norway. Even Denmark does not qualify, because, as Esping-Aridérsen (1978) shows, continuation of thí coalition with the petty bourgeoisie has prevented the Social Democrats from going beyond welfare statism. Thš is not to say that the reintroduction of socialism to the centre of the ieft's programme in the future is limited by the same three conditions. There is no reason to think that past incumbency perse should limit future possibilities to that extent. And most structural trendi favour the labour movement. Religious divisions in the workňg class are declining because of the reorientation of the Catholic Church in the wake of Vatican n, the jettisoning of militant seculirism by the socialist parties and the general process of secularisation. Where it is not supported through subsidisation for politkal reasons as in France and Italy, the petty bourgeoisie, particularly in the rural sector, is declining rapidly. And capital centralisation is increasing everywhere. On the other hand, the cases of Belgium and Canada show that ethnic and linguistic divisions are still strong bases for group loyalty. This brings up a topic which we can only touch on here. It has often been claimed that a number of ongoing changes in the class strucJure, the most important of which is the relative decline in the proportion of manual industrial occupations, will undermine the socialbases of support for socialist parties. Our own research on the Swedish and UK cases as well as that of a number of other authors on olier countries show that this argument received little support when tested empirically.2 For one thing, forced mobility into the non-aianual ranks created by changes in the class structure increases the number of non-manuals from socialist backgrounds and thus,given the strong intergenerational correlations in party preference, the proportion of non-manuals voting socialist tends to increase. Furthermore, the universalisation of secondary education is reduting the educational differences between lower non-manuals and workers and thus decreasing closure of mobility between the manial and non-manual working class. The emergent social division's both our work and Hamilton's show, is between the upper midie class and the working classes as defined in Chapter 2. The declňe in status consciousness, an essentially pre-capitalist element of class relations, contributes to this process. In Sweden the decline Conclusion 199 in closure has been furthered by an equalisation of consumption levels and class desegregation of living patterns. This has contributed to the decline in class voting. The socialists have lost some support in the working class and gained substantially among non-manuals. Consequently, the expansion of non-manual ranks is not likely to weaken the socialists' electoral base. ' The increasing labour force participation of women also favours the left. The expansion of the white-collar ranks in the post-war period led to an inflow of women into active economic life. This is the material base on which the women's liberation movement grew. The increase in sexual equality, the greater contact with work life and thus the labour movement, and the general process of secularisation has eliminated sex differences in voting in Sweden and promises to do so elsewhere. In the Swedish case, the growth of sexual equality has already made crucial electoral contributions to the cause of socialism: the Social Democrats would certainly have lost the 1973 election and probably the 1970 election had the small sex differences in voting that characterised the country in the 1940s not been eliminated by the mid-1960s. With these trends as a background, it seems at least possible that the left in the other 'corporatisť democracies could follow the Swedish pattern and bring socialism back on the immediate agenda. In Norway and Austria, there are no obstacles to doing this. In the Netherlands, the limited white-collar organisation is a problem, but the recentideological renewal of the Labour Party and the break-up of the politics of accommodation are encouraging signs in the other direction. In fact, the Labour Party has shown some interest in instituting collective funds along the lines of the Swedish plan. Given the co-operation between the Social Democrats and the People's Democratic Union (the Communists) and the continued decline in the agricultural population, the Finnish left's future prospects look good. The Danish Social Democrats have the support of a fairly strong labour movement, but the severe economic problems combined with political fragmentation of the left's support make the future of Danish Social Democracy hard to predict. The Danes are discussing the introduction of a collective ownership scheme similar to the Swedish plan with the exception that it is to be developed through a payroll tax (Lykketoft 1977). Among the other small democracies, the possibility of ideological renewal comes into question only in the cases of New Zealand and í iccign-iiiicsled ľirms controls labor organization, cullecüvc bargaining, and siřeti «!•«■ strieiK' üian in domestic firms to accelerate the influx of foreign t.-.i)iii.U. 72. NaiJiii Kau and J. Keinnitzer, "Join the Future Now! Women in the F.lec-tiv.nirs Industry-: .Sexism, Structure and Profit," in Wurncn, Men and the Inter-•niim-ial IH-isinn oj Labor, ed. June Nash and Maria l'alricia Fernandcz-Kclly i.\lli,iii\: Stale Univcrsitv of New York I'ress, 1983): 332-45 /:>. Sec Koben Thomas, "Citizenship and Gender in Work Organization: Some lliiiisiJcraiiuns fur 'ITieories of the Labor Process," in Marxist Inquiries, ed. Michael llurawoy and Theda Skocpol (Chicago: Univ ersity of Chicago Press, 1981); and I Alejandro Portes and John Walton, Labor, Class anil the International System (Neu-Yorl: Academic Press, 1981). 7-1. \i. Skinner, "Wanted: Managers for the Factors' of the Future," Annals of the American Academy of Political and Social Science, Special Issue on Robotics: I-it tu re f-.i;tories anil Future Workers, 1983, 102-14. 7:"i. Dankbaar, "Maturity and Relocation," 243. 7 (1985): 1 Sfi-'Jlili. I ) > 1Ut'l,l^ "T "~~*-. ^ /nrW^.n^/ ,vo^f . ~-//A Power and Distributional Regimes (i () S ľ A ľ. S ľ I N G - A N D F. R S F, N THIS essay presents an analysis of the relationship between political-power structures and variations in welfare-state regimes within eighteen advanced capitalist democracies. The leading question is whether, and under what conditions, the mobilization of working-class political-power resources affects the distributional and institutional characteristics of welfare-slate development. In comparison to most studies in the literatuře, I his one attempts to rcconceptualizc both the concept of power and the characteristics oT welfare states to be examined. l'he general hypothesis is that the degree of power mobilization genuinely affects distributional outcomes.1 Trade-union organization and the strength of left parliamentary-cum-cabinct control serve as the basic power resources in affecting state policies. Yet, it is important to recognize that the relationship cannot be strictly linear. Similar levels of power mobilization may yield widely different outcomes given the structure of power. We must, in other words, locate working-class power within a broader power matrix that minimally takes into account long-standing,divisions within either the right or the left, and that is sensitive to the patterns of class-alliance building. It is hypothesized that the effectiveness of working-class power resources depends on: the extent to which the nonsocialist parties are mutually divided; and the conditions open to working-class parties to forge alliances with other social classes. The lack of political unity among nonsocialist parties 'Hiis article is an interim product of a research project on comparative social policy and distributional conflicts, centered at the Swedish Institute for Social Research. The research has been funded by grants from the Swedish Delegation for Social Research, the liank of Sweden Terceiitenniary Fund, the German Marshall Fund, ami a grant from Harvard University. LT-% per/nils working-class parties greater scope for asserting their objectives .* against rightist resistance.2 And the working class's ability to forge j'liam.-es with other social classes helps to pre-empt the establishment of a unified bourgeois block and, in any case, to increase its chances of success, because a majority based solely on traditional working-class vote is unlikely to occur and equally unlikely to be stable. This study takes an unusual approach to the study of welfare-state variations in that the welfare state is viewed as an articulation of distributional conflicts, in particular, a conflict between market principles and political-allocation principles. Furthermore, the decisive distributive nature of the welfare state cannot be grasped with conventional indicators such as level of social expenditure. Rather, historical conflicts have centered around the institutional arrangements of social policy. This includes, for one, the capacity of social policy to decoin-módily wage earners, that is, to endow individuals with a relative independence of the cash nexus and work compulsion. It also includes the stratification attributes of social policy-docs it cultivate-universal solidarity, status segmentation, or individualism? And it includes the decree to which guaranteed employment is institutionalized. Kxamincd as institutional arrangements of distribution, it is difficult to examine welfare states as simply arranged on a scale of "more or less"; instead, then- emerge among our nations distinct clusters distinguished by nni(|tie types of welfare-state regimes. As will be discussed below, it is possible to identify an essentially "social democratic" welfare state regime. The question of power and distribution may accordingly be icphrased as the relationship between power mobilization and the "sueid democratization of capitalism." Šime welfare-state activities presumably grant entitlements outside iniiket relations, institutionalize collective political responsibilities lor individuals' living standards, redistribute income and resources, and provide goods and services outside the cash nexus, they are likely to be lcsisH-d liv the privileged classes. This is not merely because they in-IV i nur on tlie prerogatives of private-property ownership, but also liecaiisr t he v hold the potential of altering the balance of class power m lawn of the disadvantaged. This was the central point in Michael Kalecki's and in Kdward lleimann's analyses.'1 To both, social reforms .nul fill! employment would push the boundaries of "socialism" lor-w.nduiul weaken the market and the authority of the employer. K.iiYcki and lleimann are sophisticated representatives of prewar Mn-i.il democratic thought. Social reformism is not regarded as merely futile amelioration but as a strategic objective of political struggles: s«-ci.il citizenship may alleviate pressing needs, but it also provides a means of "social democratizing" capitalist society. A theoretical focus on working-class politics in welfare-state evolution is therefore warranted. liven a cursory review of the dominant empirical" studies on the question suggests a severe mismatch between theoretical intent and research practice. Virtually all studies examine the welfare state 'from the point ol view of social-expenditure levels or its redistributive egalitarian ellects.'1 However, spending levels and redistributive effects are, as such, not especially fruitful for theoretical verification. Granted, an advanced welfare state will absorb a good deal of the Gross Domestic Product (GDI'), and it may also promote a modicum of income equalisation. Still, huge expenditures can also reflect a preponderance of unsolved social problems, such as unemployment—something that an effective "social democratization" of capitalism presumably would eradicate. And, more importantly, expenditures mask basic institutional (liflerences in how social policy is applied: social transfers may be disproportionately targeted to privileged civil servants; they may be granted only when accompanied by more or less stigmatizing means tests, they may require lengthy individual contributions that typically exclude the weaker groups from participation; or, to the contrary, they may be granted on a universal basis to all as a citizen right. They may further status cleavages or they may engender solidarity. Measures such as spending easily obscure the presence of distinctly different welfare-state regimes. Any historian of social policy will tell us that the early social reform initiatives were spearheaded by representatives of the ruling classes, (hi the European continent, welfare-state foundations were erected by arch-conservative, statist, and paternalistic reformers such as ľ.duard, Count Taaffe in Austria, Olio Bismarck in Germany, and J. B. I'.strup in Denmark. These pioneers were quite prepared to build extensive schemes for social protection as a means to arrest class conflicts and subordinate workers to the state. Organized working-class participation in this era of welfare-state formulation was, at best, marginal; often it was oppositional. But liberals, too, played a pioneering role, although their primary efforts were dedicated to the establishment of social programs that strengthened the market mechanism, such as voluntary contributory insurance and occupational contractual plans.5 With few exceptions the working classes were objects rather than subjects of social legislation until the 1930s. And the design of welfare programs typically reflected this. Thus, in Germany and Austria, where levels of social expenditures, compared to that of other countries, were very high in the 19150s, welfare-state schemes certainly did not conform to labors wishes; they were deliberately designed to fragment wage earners into status-distinct corporative identities with visibly different financialresponsibilities and benefit privileges. To put the problem in different words, we cannot assume that social-expenditure levels alone can capture theoretically relevant variations in welfare-state activities. Welfare-state regimes with only marginal advances in the direction of social rights may be comparatively high spenders. :is was the case with the United States during the 1930s. Different nations may spend approximately the same, but what they spend on may be more important, or, still more important, the institutional properties of their programs may differ significantly. No political actors have, historically, struggled over spending levels as such. To test theories of working-class power and welfare-state development we need first to specify concretely what were the working-class objectives for »Tuch power mobilization was instrumentally important. The iteniture is marred by an additional problem. Its conceptualization oJ the relationship between explanatory variables, such as left power oi economic development and distributional outcomes, such as spending, is almost invariably linear: the more power, the more redistribution, the more economic development, the more spending. This shows a disregard for the sociological meaning of a term such as power. Power must be explained as a relational phenomenon. The effects of similar levels of working-class mobilization—percent trade unionism, parliamentary strength of labor parties—depends greatly on the overall power matrix: the chance of producing radical change is much greater if the bourgeois forces are split and incapable of mobilizing a unified counterforce. For example, the basic power resources of Austrian labor (trade-union organization and party strength) are closely similar to that of Scandinavian labor. Yet, in Austria, the bourgeois block is consolidated into the Österreichisches Volksparici (OVP); in Scandinavia the various bourgeois parties have a long history of mutual opposition, which, in turn, opened opportunities for cross-class political alliances. A powerful labor movement is easier to oust into marginalily il the right canmuster effective majorities. The problem of power is additionally complicated by the constellation of parties in the social structure. Harold Wilensky, among others, has brought attention to the central role played by Christian, and especially Catholic, parties in social reform.6 Catholic movements are required to appeal to their frequently large working-class constituencies; historically they have also been dedicated to humanitarian social amelioration and have resisted, together with traditional conservative forces, tie consolidation of a purely cash-nexus-based form of social organization. In nations where Christian movements predominate, the working-class organizations are likely to be split along professional lines, and their social reformist goals are more likely to be thwarted by the early institutionalization of Catholic-conservative legislation. Generally speaking, the translation of power resources into political outcomes must depend on two decisive conditions. First, a high level of power mobilization is likely to be effective only if it is durable. A brief .itint of lei t government, preceded and then followed by conservative cabinets, is unlikely to produce lasting institutional changes. Second, Kduard Bernstein's predictions concerning class structural development have largely proven correct: few labor movements have ever been able m count on working-class electoral majorities alone. The translation of labor-party votes into power will almost invariably require the formation of cross-class alliances: in two-party systems through incorporation .ind in multi-party systems through, more likely, governing coalitions. I he conditions under which working-class parties have been capable of forging alliances have varied sharply among nations, and this has indisputably affected their fate as long-term governing parties.' To summarize, we must pay attention to what specifically are working-class objectives in social policy, and we must be capable of identifying the structural conditions under which power mobilization is likely to affect social policy. As will be briefly argued next, we must be able to recognize qualitatively distinct social policy regimes and power matrices. IHK SOCIAL DEMOCRATIZATION OF CAPITALISM Any meaningful analysis of working-class power and welfare-state outcomes must begin with a specification of what, concretely, workers demand and can be mobilized to struggle for. The "welfare state" is one particular outcome of demands that logically flow from the position in which wage earners find themselves; it is, as I shall argue, the most likely outcome where wage-earner demands are given a collective, political expression. To the individual wage earner, social policy may primarily be a means of relief, a source of immunity from the wicked whip of the market. For labor movements, social policy has always been viewed as a means for cultivating class unity and collective solidarity in the struggle against competitive market atomization and narrow corporate loyalties. Be they union funds or benevolent societies, their purpose went beyond the immediate cry for relief; they were institutions for building collective communities, often presented as islands of socialist solidarity. Tlir exercise oľ solidarity requires not only communal organization, hut also resources. Social policy emerges as an important instrument to uplili the working classes, to distribute resources more equally among (lirm, and to grant individuals the strength and capacity for communal action. Social policy is therefore integral to the labor movements' process of class formation. It is a means through which the balance of class power can be shifted in lavnr of labor, as viewed by Kalccki and lleiiisaim. Social policy, therefore, becomes an arena for the accumulation of working-class power resources; the overriding principle is lo substitute market exchange with social distribution and properly rights with social light«. I lie "social democratization" of capitalism implies that social pn!ii\ involves a lour-pronged agenda: lbe decommodification of wage earners and of consumption, a restratificalion of society along suliibristic principles; redistributive corrections of market-induced inequalities: and, above all, the institutionalization of sustained full employment. /Vr iillllll ililifiťllllliH As individuals and collectivities, wage earners will logically strive in dn iimmodily their status.s for the individual, the issue is primarily one »i securing adequate means of sustenance that are independent of niarkei chances. For the labor movement, the strength and solidarity o| lise collectivity depend on ils capacity to provide workers with an ai i e|il.ible exil from the cash nexus. The problem group was mainly the weakest workers, the migrant rural workers, the lumpen proletariat, and i lie unemployed are the most compelled and tempted to underbid wages. M lake the place of striking workers, or to flood the labor marke!, ihrieby provoking bitter rivalries and antagonisms within the working classes. Tlie social wage that decommodification connotes is ihrrrbirc paramount for labor-movement formation. The degree of dironiinodification is a function of the extent to which citizen rights Mipplant market distribution. .S< '//T obvious importance. It strengthens the bargaining position ní both, unemployment clearly being a basic impediment to unity. Full employment is also a fundamental precondition for the viability of sncial-deiiBHTalic strategy. It constitutes the basic financial underpinnings for dccomniodificalion and solidarity and enhances the chances inr radical income redistribution in general. And to echo Raleiki ajain, it is the chief means by which the labor movement can shilt the lulance of class power in its favor. l-'ull employment on a sustained basis depends on effective power mobili/.alii'i, especially where it jeopardizes price stability, profits, and internátů «a! competitiveness. This means that the labor movement must lie [xnvcrful enough to override the priority for price stability, but it must ;iik>> be sufficiently unified that wage earners can be made to sacrifice individual wage gains for the good of the collectivity. The labor movement must be able to persuade wage earners to help shoulder the financial responsibilities for maintaining full employment, especially when the going gels tough. Not Italy is full employment in practice (if not in theory) a deeply eiuitested issue, but the means by which it can be upheld are relatively lew in a jcivately controlled economy. DISTRIBUTIONAL RECIMF.S 1 he viriciy of contemporary welfare states cannot be viewed as a simple question of more or less "social democratization." This is so lor tun reasons, first, the "social democratic" model of a wellare state is, ideally at least, characterized as integrative, comprehensive, and societal in -.rope. In contrast to the narrower, or matginalist, approach, typical in the liberal tradition, it recognizes no boundaries for social policy. In Richard Titmttss's formulation, it conforms to the institutional welfare state model.11 Second, as Gaston Rimlingcr has shown, the nature ol weHarc-sUte evolution has varied considerably depending on political context.12 It is possible to recognize two kinds of welfare-state regimes that are categorically at variance with the social democratic model. Indeed, a major historical impediment to postwar social democratic reformism is the degree to which alternatives were institutionalized at an early date. The two models—conservative and liberal—that contend with social demoracy cannot be identified so much in levels of social expenditure as in their distinct institutional properties. I'liv Cuii.ifrvulivc Model The conservative welfare-slate regime, prominent in continental ľ.uropean nations such as Germany, Austria, France, and Italy, has its loots in a blend of statist, paternalistic reformism and in social Catholic influence. The principles, as expressed in Bismarck's and Taaffe's early legislation and in the key Papal Encyclicals (Rerum Novarum and Quad-ragesimo Anno) were not so much guided by desires to establish market hegemony and rugged individual self-reliance as they were by the goal of arresting the onward march of socialism and capitalism. To this effect, Bismarckian initiatives favored an active interventionist state that promoted individual subordination and loyalty to the state. He wanted "Soldaten der Arbeit," not atomized commodities.13 Catholicism, in turn, favored a traditional corporativist social policy that promoted occupational- or status-exclusive schemes. Both favored an organization in which the employer occupies a paternalistic guiding role. As their early socialist critics were quick to point out, conservative policies were "class politics." By explicitly targeting legislation to workers and by promoting shaqo status distinctions, the political aim was to build Jaad consolidate status cleavages among the wage-earner population. The strategy, in other words, was to thwart broader class iormalion. The legacy of conservative social policy came to haunt social democracy when, subsequently, the labor movements attempted to reform social insurance. In Austria and Germany, both the trade unions and the social democratic- parties attempted after the war to substitute hierarchical and corporative (Berufsstandiche) organization with Scandinavian-style universalist "Volksversichcrung." Of particular importance was the status equalization between salaried and manual wage earners with respect to financial obligations, eligibility conditions, and benefits. In both nations, the socialists were only partially successful in overriding their conservative adversaries. Against OVP objections, the Austrian Sozialistische Partei Österreichs (SPO) managed in the 1955 reform to reduce the number of existing separate insurance bodies ffuisitlcnibly, and a common legislative frame was imposed on both wnrkcr aur/ salaried pension schemes. The German Sozialdemokratische l'.jy'.c! Deutschlands (SPD) had some success too, along similar lines, but in neither ease was anything close to Volkversichcrung accomplished— .in obvious reason being that the privileged salaried employees and ii ill it status groups had gained, over many decades, a vested interest in ictaining corporative exclusiveness. Over the postwar era, some of the jjn-atcii battles, and also victories, of German and Austrian social democracy occurred in the context of "Lohnfortzahlung" legislation, thf reform to equalize conditions for full and automatic sick pay. Conservatism shares with liberalism a concern for actuarialism; however, it is a corporative rather than individualistic actuarialism. This has been another of the major points of conflict in both Germany and Austria. Although the socialists have had some success in establishing closer financial integration among separate insurance schemes, they Ikivc by and large failed to override conservative vetoes against general »iivernmcnt financing or subsidization. in summary, "conservative" welfare-state regimes differ from the social democratic variety not so much in their willingness to grant social rights, as in their corporative structure. It is, however, a structure where rights ,iiul duties arc attached to occupation and status, not to citizenship. Tit, Ub,r,il Model In its ideal-typical form, classical liberal social policy is designed to maximize the commodity status of the individual wage earner.1,1 The cmiccpt of "Less Kligibility" that underpinned the liberals' poor relief schemes served to eliminate exit opportunities from themarket and to punish those who could, or would, not make do with the pure cash ni-xiis. Liberalism, in contrast to conservatism, has maintained a powerful attachment to means-tested, targeted forms of social policy. The traditional practice of ruthless deterrence has, perhaps, vanished, but the principle is sustained as a means of securing optimal help at minimum cost or as a means of selecting the deserving from the nondescrv-iiis; poor. Liberal social policy is residual in that it establishes'narrow In »titularies for government intervention but maximum scope for markets in the distribution of welfare. For liberalism, the market place is the obvious site of distribution, and its enthusiasm for private contractual insurance and fringe benefits is therefore logical. This model of "welfare capitalism" has been especially dominant in the United States, but also in Australia, Canada, jiid.in Switzerland. And, as several studies suggest, trade unions occasionally not only sanctioned but even favored private, occupational welfare to state insurance.15 On this score, liberalism stands in sharp contrast to both conservatism and socialism. But, an unusually strong market logic is also apparent where liberalism adopted and sponsored state insurance. Liberal social insurance emphasizes voluntary membership, individualistic actuarialism with benefits closely connected to previous contribution and performance, and comparatively meager public benefits and standards so as to encourage private insurance alternatives. By stipulating long contribution requirements, social insurance may actually help encourage individual market participation. It is also easy to see liberalism's commitment to market hegemony in its typical insistence on substantial waiting periods for benefit eligibility, as well as on brief benefit-duration periods. In short, the hallmarks of liberalism, in juxtaposition to the two other contending regime types, are its residualism, its stress on private market provision, the conspicuous role played by targeted means testing, and its sell-reliant individualism. Albeit promoted in different forms, liberalism does, however, share with socialists a principled preference for universalislic ideas; of course, it tends to be a universal-ism of equal opportunity rather than equal rewards. The industrialized capitalist democracies may rank on a smooth, linear continuum in terms of social expenditures, taxation, unemployment levels, and other aggregate indices of welfare-state performance. But they clearly do not in terms of regime characteristics. Indeed, a fairly distinct clustering is immediately evident. By and large, the conservative model is pre-eminent in nations where the church played a powerful role in social reform and in nations where absolutism was strong and slow to give way—in nations, therefore, where the bourgeois revolution was weak, incomplete, or absent, such as in Austria, Germany, France, Italy, Japan, and Belgium. By contrast, there seems to be a surprising synonymity between liberalistic social-policy regimes and Anglo-Saxon "New World" nations, where the bourgeois impulse was especially powerful, such as Britain, the United States, Canada, and Australia. The ideal-typical features of socialist, liberal, and conservative distribution regimes arc depicted in table 1. The capacity of labor movements to substitute "social democracy" for cither liberalism or conservatism is obviously related to their historical position of power. If labor movements can affect fundamental change through participation in parliamentary democracy, we would expect to find a positive relation between levels of working-class political mobilization and the degree of "social democratization" on welfare-state performance indicators. Where labor is politically strong, /villi libera list ic and conservative welfare-state regime features would (hus !)«' comparatively weak.16 TU K BA1.ANCF. OF CLASS POWER AM) DEGREE OF SOCIAL DEMOCRATIZATION Alituiugh traue unions, business organizations, and farmers play liincLimch!;d roles in defining the relative power resources of the major social i:Lts5'.:s in modern capitalism, the proper focus for a "parliamentary hypothesis" is obviously the balance of electoral, legislative, and executive power among political parties. Some studies employ, as a measure «I working-class political strength, an index of percent left votes, llinvcvcr, electoral rules often do not allocate parliamentary seals in proportion to votes. Many studies also make do with an index of |)i-rcrn! left votes, or seats, in one particular year. This risks being further fun fused by a "Mitterand effect," where a left government may recently dive taken office only to be (probably) ousted shortly thereafter. It ii unlikely that the dominance, let alone hegemony, of conservatism, or social democracy will occur unless backed by long-term and dura hl e power. It is equally unlikely that such a dominance can be established unless a party, or a coalition, controls cabinets and parliament over long periods. Consequently, the appropriate measure of ponei is 31/ weight share of parliamentary seats with share of cabinet pi/sts averaged over a relevant time span.17 I'mil the 1930s, and generally not until after World War II, it was rxi crilingiy rare that working-class political parties held office-or even participate! actively in the formulation of social policy. As in Austria and Germany between the wars, or in France in the mid-1930s, socialist i ahinris were shortlived, embattled, and usually in a minority position. The breakthrough of social democracy as a stable and strong expression <>! wiirknsg-class politics occurred basically after World War II, spearheaded in the I 9:50s and 1940s by the Labour party in Australia and Scandinavia, followed by the Labour party in Britain and by the socialist panic,in Austria, Belgium, and elsewhere. With That in mind, it is highly unlikely that left parties would have made any major imprint on social policy until after World War II. Since ii is als» ťvidem thai the variance of left party power increased among our eight rcn nations after Ihe war, the parliamentary hypothesis would lead us tu believe that the nations' social-policy performance was more homngnimis and convergent before the war, and increasingly divergent ijieivjher. In other words, the process of social democratization is primarily a postwar phenomenon, and we would expect it to vary considerably depending on the long-term influence of left parties over government policy. Finally, as a general rule, we would expect that the degree of left party power is especially important on issues that are divisive, that is, where labor movements seek to push the boundaries of markets back decisively. The scope of nonmarket,-collective consumption or the maintenance of full employment during the very severe recessions of the 1970s ought therefore to require especially strong left parly power, compared to less "threatening" issues such as levels of social expenditure. In the following, we present a scries of statistical analyses of the relation between left weighted cabinet share (LCS) and indices of welfare-state performance that should reflect important advances in i he direction of a social democratization of capitalism.18 V'/f Suťitil Wage In contrast to what is often claimed, patterns of social expenditure are not becoming more convergent among the leading eighteen capitalist democracies. The growing divergence, moreover, cannot be ascribed to factors such as economic development or the logic of bureaucratic iiicrcmcntaiisni.19 The long-run trends in social-wage expansion require a political or social structural explanation. Table 2 indicates that working-class mobilization had no influence on pre-war social-wage levels. This is even true for 1950, although at this point the left parties' electoral share—a more indirect source of influence—had some impact. It is, as we would expect, only in the postwar era that left cabinet control affects the size of the social wage. This seems to contradict the commonly held view that the postwar welfare state expansion was a politically consensual issue.20 If we disaggregate the historical trends, we see better why a nation's ranking on social spending camouflages institutional divergences. Although the 191515 data exclude some of the eighteen countries, it is clear that the ranking at that time has little to do with the situation in the 1970s, and it tells us very little about a nation's progress toward effective social citizenship. In 1915:5, the leaders were Germany, Britain, and Austria; the laggards were Finland, Holland, and Italy. The United States, even if at that time it had introduced virtually no social-rights programs, would still have ranked in the middle. Means-tested poor relief, civil-service benefits, and patronage-flavored veterans' pensions dominated. While social-wage expansion in the postwar era is closely related to left cabinet power, the relationship, as Lee Rainwater shows, varied by program type.'21 As we shall see shortly, left power is especially important in accounting for the growth of collective services, medical services in particular. A Catholic/conservative-social-democratic convergence on the social wage is evident for the postwar decades, showing these kinds of nations now to be the leaders and the "libcralistic" and ri'siduulistic Anglo-Saxon nations to be the laggards. Sunul K it;/its or I'oor Relief? The size of the social wage may indirectly capture the level of diaimmodific ation; another measure is the extent to which citizens command basic entitlements as opposed to being dependent on various kinds of means-tested and socially stigmatizing relief programs. The relative scope of poor relief in total social-wage expenditure also offers a way of identifying the stratifying properties of a welfare state. Not surprisingly, one of the labor movements' foremost priorities has historically been to supplant means-tested relief with rights programs. This was i he first and foremost goal of the SPD in the Weimar Republic, and it was the most pressing reform task of the Nordic social democrats when they came to power in the 1930s. In short, the social democrati-/.jiion ol capitalism means assigning poor relief very low priority.22 As we know, a great proportion of total social expenditures in the l!):if)s and even as late as 1950, consisted of means-tested benefits. The major rights programs in Scandinavia, for example, were really only Kgislafcd by the social democrats in the period 1945-1955. Table 2 indicates thai, for 1950, there was a substantial correlation between total expenditures and the scope of poor relief; that relationship disappears by the 1970s. The table also shows that there is no significant impact of left power on the scope of poor relief in 1950. But its impact is decisive in explaining poor relief in 1974. Among the eighteen countries, poor relief has been most effectively eliminated in the social democratically dominated nations, especially in Scandinavia, whereas it [liays the greatest role in nations where the left is weakest: the United Slates, Canada, and France. The "Catholic" nations occupy a position in the middle. /ii-ii'iiimnilifii'il Consumption Because collective goods and services threaten to displace the market mechanism with respect to pricing, and private enterprise with rcs|iect to production, collective consumption is likely to be resisted and to generate intense political conflict. The extension of decom-nsMilificd consumption is, accordingly, a distributional battle that may assume substantial left power mobilization. An index of decommodified consumption must exclude government military and defense outlays. This can be done for all eighteen cases in 1950, but for only fifteen in the late 1970s. Generally spean-ing, collective consumption has grown at a less spectacular pace than has the social wage, but cross-national differences are clearly greater in 1979 than they were in 1950. At one extreme lie Denmark and Sweden, with more than 20 percent of Cross Domestic Product (GDP) devoted to nondefense, public consumption; at the other end lie Japan and Switzerland. Again, the strength of working-class parties has no hearing on collective consumption in 1950. But left power is decisive in explaining both the expansion between 1950 and 1979, and the overall situation in 1979. A lull 66.2 percent of the variance in 1979 can be accounted for by LCS. See table 4. The l'uhlic-I'rivate Mix For reasons of both solidarity and decommodification, the social democratization of capitalism will have to supplant private and contractual welfare distribution with universalistic and adequate social rights. Private insurance has, for many reasons, been an historical enemy of labor movements and a favored solution among employers, private insurance and financial enterprises, and bourgeois parties. Private insurance as well as occupational, fringe-benefit programs strengthen the wage earner's attachment to the market and the employer and therefore help consolidate the individual's status as a commodity. They also replicate and possibly strengthen existing inequalities and differentials among wage earners and tend to favor the strongest and exclude the weakest. They may even undermine solidarity toward the welfare slate, since those groups who command a privileged position in the private market are less likely to support, and identify with, social programs. The relative share of private pensions and health care to total pension and health-care outlays constitutes a good measure of the private-public mix, because these two programs, taken together, account for the lion's share of social expenditures in all nations.23 The impact of working-class political mobilization on the relative scope of private pensions is less strong than one would have anticipated, and lor health even weaker. See table 5. The lack of a powerful relationship may be due to the way in which the nations are distributed on the variable. There is a tendency toward a bi-modal distribution with one cluster showing very high ratios of privatc-to-public pensions (United States, Canada, Switzerland, and Australia), and another with low ratios (Norway, Austria, Germany, Italy, and Sweden). It is obvious that the relationship is not linear and that there are crucial statistical outliers, Denmark especially, where left power scores high together with privatization. Herc is a clear instance where the relationship between power and distribution must take interaction effects into account.24 Individualism or Solidarity us Social Stratification Rarely arc social-insurance systems strictly actuarial, even if they .ire outwardly promoted as such. The liberal ideal of a direct correspondence between benefits and individual contributions is therefore, at best, approximated. A key historical conflict has always surrounded the question of financial burdens, labor movements normally espousing a high—if not total—degree of financial redistribution. In contrast, a defining characteristic of liberal reformism was its insistence ) 'Ilie ratit) of private to public pension and health expenditures measures the relative importance of state and market allocation within the two single largest social programs. The index show's the amount of private pension and health expenditures as a percent ol total public expenditures in these fields. Note that for pensions, the public share excludes civil servants' pensions, since they can be regarded as "occupational" in nature; they have also been excluded from private pensions. The data for private and public health expenditures are derived from OECD, The Growth of the Public Sector (Paris: OECD, 1983), while the data on private pension expenditure (including individual and group insurance plans, as well as company and occupational plans) have been compiled from national governments directly (in most cases from the Ministry of Finance). (i) financial individualism (INDE1N/SSE—share of individual contributions to total social-security finance) indicates the strength of the traditional liberal dogma of self-reliance. It is measured by the share of total social-security revenues derived from individual contribu-lions among the insured as a percent of the total. The data derive from lLO's social-security data series. 7) Corporativism (CORP) in the organization of social-security schemes is measured by taking the number of separate, status-distinct programs in pensions and in unemployment and sickness insurance. Where social security is arranged in a large number of status or occupalionally distinct schemes, a nation will score high on the corporativism index. The data arc collected from Social Security Programs lliroitghottl the World, published every two years by the U.S. Social Security Administration. 8) The data on full-employment commitments and performance fall into two categories. The first three (MANPROG, PUBEMP, and (PUBKMP)X(LFPART)) are indices of commitment. MANPROG is expenditures on manpower programs as a percent of GDP (excluding unemployment compensation); PUBEMP is the percent of public-sector employment of the total in 1980; and (PUBEMP)X (LEPAIi-T) is the expansion or decline in public employment between 1970-80, adjusted by overall labor-force participation rates in 1970. (Obviously, public employment expansion is more impressive where overall participation is already high.) The data for these indices are primarily based on OECD National Accounts series and national statistical sources. The second category has two indices that measure actual full-employment performance. The first is the level of unemployment (UNEMPL), standardized according to OECD convention, as a percent of labor force, averaged over the years 1978-83—the deepest [lostwar recession. The second is a composite performance index (explained in the note to the table itself). The data here are based on OECD sources, as above. Note that a presentation of all the raw data would occupy too much space. Interested readers may obtain data directly from the author. 69 S- c w E ! § 5 w .« .TJ *-■ — rt u -. Ü C ■ž % ■a -a 3 E O .2 m D "3 ^ u •H 2 J = .a o-S .§ = "3 3 § 3 '3 CT rt . -a ó. ES E o 13 o Q E J2 o o -1 J OJ C 0 pa S Ľ u 0 ^ a. -1 a J =i £ S o o II 11 c o u Left Power and the Social Wage 1933-1977 r R2 T (n) SSE/GDľ (1933) on LCS (1918-32) SSE/GDľ (1950) on LCS (1918-49) SSE/GDľ (1950) on LĽKT VOTE (1918-45) SSE/GDľ (1950) on UNION (1918-45) SSE/GDľ (1977) on LCS (1950-76) SSE/GDľ (change 1950-77) on LCS (1950-77) .635 36.6 3.29 (18) NOTE: SSE/GDľ is social security expenditure as a percent of GDP, LCS is weighted left cabinet sliare. ľor 1977, left vote explains 6.4 percent of the variance uf SSE/GDľ, and union organization explains 8.1 percent: neither is significant at .05 or better. .275 0.0 .95 (13) .247 0.2 1.02 (18) .588 30.5 2.91 (18) .506 21.0 2.35 (18) .527 32.3 2.48 (18) Table 3 The Institutionalization of Social Rights. The Relative Importance of Means-tested Public Assistance, 1950-1974 R2 T (n) i) l'A/SSK (1950) on SSE/GDľ (1950) ii) I'A/SSE (1974) on SSE/GDP (1977) iii) I'A/SSE (1950) on LCS (1918-49) iv) I'A/SSE (1974) on LCS (1950-76) NOTE: I'A/SSE is public assistance as a percent of total social security spending. Public assistance is defined in the narrow term as relief programs, and does not include income-dependent transfers such as pension supplements or family allowances. Nor does it include expenditure on social assistance type pension or related programs, such as the Australian pensions. Since neither unionization or left votes alter the patterns of association, they have been omitted from the table. -.572 10.8 1.64 (18) -.234 0.0 -.93 (18) .177 0.0 .72 (18) -.674 41.8 -3.54 (18) /'ubfc Civil Consumption Expenditure and Working Class Mobilization r R2 T (n) i) PUHCONS/CDP (1950) on LCS (1918-49) -.050 0.0 -.20 (18) ii) PUBCCWS/C.DP (1979) on LCS (1950-76) .828 66.2 5.33 (15) iii) l-ľnCONS/ODP (change 1950-1979) (on 1.SC (1950-76) .745 52.1 4.03 (15) XOTK: l'UBCONS/GDP is public non-defense consumption expenditure as a percent of CI)!'. As in table 3, left votes and unionization correlations have been omitted, since they add nothing to die patterns. Table 5 The Mix of Market and State in the Distribution of Pensions and Health Care, mid-1970s r R2 T (n) private pension expenditure as a percera of Public1 on LCS (1950-76) -.424 12.5 -1.81 (17) Private Heath Expenditure as percent of Public, on LCS (1950-76) -.309 3.5 -1.26 (17) 1. Private plans include individual/group/occupational schemes. Note that public programs exclude civil servants' pensions. NO 11.: flit sample is reduced to 17, since it is impossible to estimate private sector pi'iisinii/litilth expenditures for Ireland. Left votes and unionization variables have been omilttil since they add nothing to the patterns. Financial Individualism 1950-1977 r R2 T (n) .322 4.8 1.36 (18) -.267 1.3 -1.11 (18) -.196 0.0 -.080 (18) i) INIHTN/SSI- (1950) on LCS (1918-49) ii)lNUI-TN/SSF (1977) on LCS (1950-76) iii) INHKIN'/SSF (change 1950-1977) on LCS (1950-76) iv) INDKIN'/SSK (change 1950-1977) on LCS (1950-76) omitting tw< extreme outliers1 -.523 22.1 -2.29 (16) L The two outliers arc Australia and New Zealand. See text for further explanation. NO'l'l'.: IX'DITN/SSľ is share of individual contributions to total social security finance. Left votes and unionization have been omitted, since they add nothing to the patterns. Table 7 Corporative Social Stratification in Social Insurance 1955-1979 r R2 T (n) CORP (1955) on LCS (1918-49) -476 17.8 -2.16 (18) _____ .„.. ,,„E„™i -.329 5.2 -1.39 (18) -.476 CORP (1979) on LCS (1950-76) -.329 CORP (change 19.55^79) on LCS (1950-79)' -.393 10.2 -1.71 (18) CORP (19 79) on CATHOLIC (1950-79)1 .511 21.5 2.38 (18) 1. The influence of Catholic conservative parties is crudely measured by giving a ' score of 3 to nations with strong confessional voting; a score of 2 to nations with split religious voting; and a score of 1 to countries with no religion-based divisions in voting. Trade unionism and left votes are omitted due to their lack of independent association with corporativism. lame b Government Full Employment Commitment and Performance and Working Class Mobilization' r R2 T (n) MANFROG/GDP (ca 1975) on LCS (1950-76) ľURLMP/TOTEMP (1980) on LCS (1950-76) (I'UREMP) X (LFPART) CHANGE 1970-80 on LCS (1950-76) U N »IP L (1978-83) on LCS (1950-76) PERFORMANCE INDEX (1959-83) on LCS (1950-76) .695 44.4 3.49 (15) .710 47.1 3.90 (17) .691 44.5 3.83 (18)2 .384 9.4 -1.67 (18) .557 26.7 2.68 (18) Key 3o Variable Abbreviations: MAXPROG/GDP = Expenditure on active manpower programs (excluding unemployment compensation payments) as a percent of GDP. I'UFSMP/TOTEMP = Public sector employment as a percent of total employment. (I'UH.MP)X (LFPART) CHANGE 1970-1980 = Percentage change in public employment 1970-1980, multiplied by total labor force participation as a percent of age .{roup 15-64 (in 1970). Note here that, for two nations (Canada and U.S.) where public employment declined, a score of zero was given. U.VLMPL = level of unemployment as a percent of labor force, according to OECD defiiilion. PERFORMANCE INDEX = Average level of unemployment 1959-1978 plus jivrnge level of unemployment 1978-1983 multiplied with ratio of inactives (as per cín i of age group 15-64). Note that index has been inverted so that a positive sign indicates good performance. l.'Iae Commitment variables are manpower program expenditures and public employment expansion. The Performance variables are levels of unemployment, adjusted for labor force participation rates. 2. ISita on public employment expansion 1970-80 for New Zealand are lacking in Lbe OECD sources. I have given New Zealand a rough estimate of 20 percent '-•»pansion (or, similar to Australia and the U.K.). NOTE: Left vote and trade unionism variables have been omitted, since they add notling to the patterns. Table 9 Income Equality in the Market and in the State r r2 ' T (n) WAGEDIFF (average 1957-1979) on UNION (1946-1976) -.449 13.5 -1.74 (14) WAGEDIFF (change) on UNION (1946-1976) .034 0.0 .12 (14) WAGEDIFF (average 1957-1979) on LCS (1950-1976) -.525 21.5 -2.14 (14) TAXREDISTR (1978) on LSC (1950-79) .560 27.0 2.70 (18) NOTE: WAGEDIFF is interindustry wage differentials. TAXREDISTR is a measure of the difference in tax burdens between an average worker family, and a family with twice the average worker earnings. NOTES 1. See, e.g., Walter Koqji, The Democratic Class Struggle (London: Routledge & Kcgan Paul, 1983); and Güsta Esping-Andersen and Walter Korpi, "Social Policy as Class Politics in Postwar Capitalism," in Order and Conflict in Western European Capitalism, cd. John Goldthorpe (Oxford: Oxford University Press, 1984). 2. 'Ulis factor is strongly emphasized in Francis Castles, The Social Democratic Image of Sucirty (London: Routledge & Kegan Paul, 1978). 3. Michael Ralecki, "Political Aspects of Full Employment" (1943), reprinted in A Critique of Economic Theory, ed. E. K. Hunt (London: Penguin Books, 1972); and Eduard Hermann, Soziale Theorie des Kapitalismus (Frankfurt: Surkamp, 1980; first pub. Í92"9). 4. For a recent critique of this research tradition, see Jens Alber, Gösta Esping-Andersen and Lee Rainwater, "Alternative Approaches to Comparative Social Policy Research," in Comparative Public Policy Research, ed. K Mayer (forthcoming). 5. For examinations of early social reforms, see: Gaston Rimlinger, Welfare Policy and Industrialization in Europe, America and Russia (New York: Wiley, 1971); Thcda Skocpol and J. Ikenbcrry, "The Political Formation of the American Welfare State" (Unpublished manuscript, University of Chicago, Chicago, 1983); Asa Briggs, "The Welfare State in Historical Perspective," European Journal of Sociology, vol. 1 (1961); and P. Köhler and H. F. Zacher, A Century of Social Insurance (Berlin:Dunckcr & Humboldt, 1982). 6. See Harold Wilensky, "Leftism, Catholicism, and Democratic Corporatism," in The Development of Welfare Slates in Europe and America, ed. Peter Flora and Arnold J. Heidcnheimcr (New Brunswick: Transaction Press, 1981). 7. Scandinavia's virtually permanent social-democratic rule over the past forty or fifty years has not been premised on electoral advances but on a unique capacity to forge alliances. Their level of electoral support has been basically stable at the levels attained during the 1930s; their original ability to monopolize government was premised on the alliance with farmers' parties in the 1930s. In contrast, both .(.Verman and Austrian social democracy were unable to offer an alternative to the i,- rise oľ fascism in the 1930s precisely because they were working-class ghetto move- ments, unable to coalesce with the farmers. For a detailed analysis of social democracy and class alliances, see Costa F.sping-Andcrsen, Politics against Markets: '/'lit- .Social Democratic Road tu I'ower (Princeton: Princeton University Press 1985). 8. The concept of decommodification, in reference to social policies, has previously been developed in the writings of Claus Ofl'e; its analytical origins clearly derive I rum Polanyi and Marx. For a somewhat different discussion than presented here, see Clans Offe, Contradictions in the Welfare State (Cambridge: M.I.T. Press, 19X4). 9. A detailed examination of solidaristic social policy can be found in I%sping-Aiulersen, Politics against Markets. 10. C. A. R. Crossland, The Future of Socialism (New York: Schocken, 1967). 11. Richard Ti.tmuss, Social Policy (New York: Pantheon, 1974). 12. Rimlinger, Welfare Policy and Industrialization. IM. Ludwig Preller, Sozialpolitik in der Weimarer He/iublic (Stuttgart: Mittel-liai-h Verlan. 1949). 14. Karl Polanyi, The Crcal Transformation (Boston: Beacon Press, 1957). 15. See Rimlinger, Welfare Policy and Industrialization; and Skocpol and Iken-ln-rry, '"Political Fonnation of tlie American Welfare State." |f>. We are here formulating a simple, linear relationship between power mobilization and distributional outcome that seems to contradict our basic point that welfare state regimes are categorically distinct and cannot therefore be analyzed as "inure or less." Additionally, our formulation contradicts the earlier assertion that working-class power resources alone are an insufficient indicator of power, because power is relational and so requires of us an investigation of the broader power matrix. The intention at this point is to establish basic statistical relationships between levels of power mobilization and the degree of social democratization. Subsequently, we will return to the more complex patterns of power and distribution. 17. Using Korpi's data set on percent left votes and parliamentary and cabinet seats (the latter two weighted with one another), the data can be broken down into an intcrwar period average (1918-46) and a postwar average (1946-76). For the intcrwur era, the simple correlation between left votes and weighted cabinet shares is r = .04K. For the same period, the correlation between trade-union membership anil weighted left cabinet shares is r = .333. 'Hie postwar correlations are stronger. 'Hie former is r = .636; the latter increases to r =.816. This suggests thai, lor ill e pustwur years, one measure generally substitutes for another, but that is not llie rise for [he pre-World War II era. Note that, among the eighteen nations included in the study, postwar trade-union strength explains 64.4 percent of the weighted cabinet share variance (T = 5.64). In this, as in all subsequent analyses, the It-squared arc adjusted for degrees of freedom. For a more detailed description of lli e data, consult Korpi, Democratic Class Struggle. 1 8. See die accompanying appendix for a more detailed discussion of the data and indices used in die following statistical analyses. 11). The standard deviation of social-security-expenditures/GDP is greater in 1977 than it was in 1950. Gösta Ľsping-Anderscn and Walter Korpi, "From Poor Relief to Institutional Welfare States," in Welfare Research and Welfare Society, cd. R. Krikson et al. (New York: M. E. Sharpe, 1985). The zero-order correlation between social expenditures and per capita GDP was .157 in 1950 and .107 in 1974. A bureaucratic incrcmcntalism theory would argue that past social-security performance constitutes a powerful predictor of the future. Yet, the correlation between social expenditures in 1933 and 1977 is a negative:-.120; the correlation increases somewhat for the years 1950 to 1977. Lee Rainwater, "The Growth of Social Protection in the United Slates, 1929-1979." (Unpublished paper, Department of Sociology, Harvard University, Cambridge, 1982). 20. The raw data base for this analysis, and for all subsequent analyses, has been omitted from the tables for reasons of space. For any interested reader, the data can be obtained by contacting the author direcdy. 21. Rainwater, "Growth of Social Protection." 22. For Germany, a good discussion of SPD policy on this question can be found in F. Teunstcdl, "Sozialgcscliichte der Sozialversicherung," in Handbuch der Sozialmedizin, ed. M. niohmke, vol. 3 (Stuttgart: Enke Verlag, 1976). See also Preller, Sozialpolitik; and lisping-Andersen, Politics against Markets. 23. So far, we have only data on private pension and health expenditure for the 1970s. The difficulties of collecting reliable estimates, especially for pensions, are immense (although the OECD furnishes fairly good data on private health expenditures). For most of die seventeen nations included in this analysis, private pension data have been furnished directly from government ministries. For a few, such as Belgium and Austria, we have been forced to make rough estimates. Note that the measures of private/public pension expenditures exclude civil-servant pensions. 24. One of Danish social democracy's greatest failures has been its repeated inability to arrest the rise of a highly inegalitarian and dualistic private-pension system and substitute it with a strong universalistic second tier, as occurred in Sweden and Norway. A somewhat similar situation has prevailed in Australia, a second outlier in the analysis. Sec Esping-Andersen, Politics against Markets, chap. 5. 25. A good general discussion of the politics of liberal actuarialism can be found in A. Ogns, "Social Insurance, Legal Development and Legal History," in Bedingungen fur di'e Entstehung von Sozialversicherung, ed. H. F. Zacher (Berlin: Duncker iL- Humboldt, 1979). The politics of actuarialism in Üie United States are described in M. Derüiick, Policymaking for Social Security (Washington, D.C.: Brookings Institute, 1979;).„And, Tor Germany, consult F. Zöllner, "Germany," in A Century oj Social Insurance, cd. P. Köhler and II. F. Zacher (Berlin: Duncker & Humboldt, 1982). 26. Note that trade unions have often struggled to maintain exclusive control over unemployment, and sometimes sickness, funds, even if this would entail a substantial element of employee self-contribution. Where unions have succeeded m this respect, as with the Ghent System of unemployment insurance in Denmark and Sweden, they have by and large managed to institutionalize heavy public subsidies. 27. Douglas Hibbs, "Political Parties and Macroeconomic Policy," American Political Science Review, vol. 71 (1977). 28. Separate analyses, not shown here, indicate that the strength of full-employment commitments correlates almost equally strongly with trade-union power. In a multivariate regression, entering both union and left power, die explanatory power of unions declines relative to that of die parties. 29 In terms of explaining performance on full employment during the recessionary years 1978-83, I should perhaps not have employed a left-power measure averaged between 1950-76, but I have taken into account the weakening of many labor parties in the late 1970s, especially in nations such as Denmark, Germany, Belgium, and the Netherlands. 30. Among the studies that argue a positive relationship are: D. Cameron, "Inequality and the State: A Political-Economic Comparison" (Paper presented at tlir American Political Science Association [APSA] meetings, Chicago, 1976); and C. Hewitt, "The Kffects of Political Democracy and Social Democracy on Kquality in Industrial Societies," American Sociological Review, vol. 42 (1977). However, the relationship is refuted in I'. Parkin, Class Inequality and ľniitiral OrilcT (New York: I'racgcr, 1972), as well as in R. Jackman, "Political Democracy and Social Equality," American Sociological Review, vol. 39 (1974). 31. A more developed analysis of Ulis issue can be found in Costa Esping-Andcrscn, "After the Welfare State," Working Papers for a New Society, April 1982. 32. Based nn TECD's study for 1978, we take the percentage difference between highest and lowest earners among single-person households, plus the same for families with two children, divided by two. The data arc derived from OECD, The 197ft rAXIHENEFH Position of a Typical Worker (Paris: OECD, 1979). 33. This has also been argued above, yet there exist exceptions. It appears, for example, that the Australian labor movement has never been fully committed to the fight for universalistic rights to income maintenance, and the American Federation of Labor (A FL) trade unions were, until the 1930s, quite hesitant to sponsor or even sanction government social-security legislation. However, given the nature of American industrial relations in that epoch, AFL's stance is explicable; it was an overwhelmingly crafts-worker organization with narrow corporate loyalties and saw in direct bargaining its best chances for social protection for its clientele. I'nlitics & Society 14, no. 2 (1985): 223-56. International Lending and the Relative Autonomy of the State: A Case Study of Twentieth-Century Peru BARBARA STALLINGS Till', debt crisis that has wracked Latin America since August 1982 has rekindled debate on the impact oľ the international economy on Third World nations. A crucial aspect of that debate focuses on the role of the state and how it is affected by the internationalization of capital. Superficially, the crisis seems to have severely undermined the power of the stale as the banks and the International Monetary Fund (IMF) joined forces to demand austerity programs, and a lower level of state participation in the economy, in exchange for debt relief. The current situation thus appears at first glance to support the more extreme versions of dependency theory whereby external actors have a determining and;-detrimental influence on the Third World.1 Probing a bit deeper, however, reveals that foreign loans had actually strengthened the state before the crisis struck and that the powerful state apparatuses that emerged are not likely to fade away. This paper focuses on past experiences rather than future prospects.- Using 1'eru as a case study, 1 will develop an analysis of the mechanisms by which foreign capital can help expand state capacity as well as limit it. By examining a single case over a period of sixty years, 1 hope to contribute to a general understanding of the possible relationships between foreign capital and the state in Latin America and, by extension, in other parts of the Third World. The autonomy, or relative autonomy, of the state will be a central concept in the analysis. This concept has been very important in the This article was originally presented to the Research Committee on Economy and Society at the International Sociological Association CongTess, Mexico City, August 1982. 1 would like to thank participants at that meeting for helpful suggestions and Peter Evans for extensive comments on various drafts.