MPIfG Working Paper 04/3, September 2004


'The Good, the Bad, and the Ugly' – Esping-Andersen's Regime Typology and the Religious Roots of the Western Welfare State[1]


Philip Manow , Max Planck Institute for the Study of Societies, Cologne






Esping-Andersen’s ‚Three World of Welfare Capitalism’ has been the most influential contribution of recent years to the comparative welfare state research literature. According to Esping-Andersen, the welfare state basically comes in three variants: as a social-democratic, a conservative, or as a liberal regime. Yet, at a closer look particularly the conservative regime type proves to be a highly problematic category. The article claims that major problems of the ‘three worlds’-typology originate from Esping-Andersen’s sole focus on the class conflict, whereas he only very selectively accounts for the importance of religious cleavages. Major empirical problems of his approach vanish once we take into account not only the impact of the Catholic social doctrine on the development of the welfare state, but consider also the influence of social Protestantism, especially that of reformed, ‚free’, disestablished or dissenting Protestantism. The paper substantiates this claim with data-analysis for the early formative period of welfare state formation (1890-1920) and for its times of growth and crisis (1960s-1990).





Folgt man dem einflussreichsten Beitrag zur vergleichenden Wohlfahrtsstaatsforschung der letzten Zeit, Esping-Andersens ‘Three Worlds of Welfare Capitalism’, so tritt der entwickelte westliche Sozialstaat in drei Varianten auf: entweder als sozialdemokratisches, oder als konservatives, oder als liberales Wohlfahrtsstaatsregime. Ein genauerer Blick zeigt jedoch, dass Esping-Andersens Typenbildung und Länderzuordnungen insbesondere im Fall des konservativen Regimes problematisch sind. In diesem Papier argumentiere ich, dass gravierende Probleme daherrühren, dass seine Theorie vornehmlich den Klassenkonflikt, nur sehr selektiv aber den Einfluss konfessioneller Faktoren in den Blick nimmt. Wesentliche empirische Unstimmigkeiten seines Ansatzes lassen sich lösen, wenn man neben der Bedeutung der katholischen Soziallehre für den institutionellen Entwicklungspfad des entwickelten Wohlfahrtsstaates auch den Einfluss des Protestantismus, und hier insbesondere den Einfluss der reformierten protestantischen Strömungen, in den Blick nimmt. Der Beitrag zeigt dies mit vergleichenden Daten sowohl für die Frühphase (1890-1920) als auch für die Hochzeit des entwickelten Wohlfahrtsstaats (1960-1990).









1     Esping-Andersen and Sergio Leone


The most influential contribution to the comparative welfare state literature in the last two decades has certainly been Gøsta Esping-Andersen's 1990 book "The Three Worlds of Welfare Capitalism". I think it is not unfair to label the typology of OECD welfare states as developed in the book as essentially 'Sergio-Leonesque'. Although Esping-Andersen always attempts to base his comparative assessment of welfare state regimes on quantitative analysis, and although he is explicitly committed to the standards of "comparative empirical research" (1990: 3), the study can hardly camouflage the strong normative loading of its distinction between Social Democratic, liberal und conservative welfare state regimes. According to this perspective, the Social Democratic regime assumes the role of the 'good' regime, since it liberalizes citizens from market dependency via generous welfare entitlements – it 'de-commodifies' labor: "Social rights push back the frontiers of capitalist power" (1990: 16). The liberal regime, in contrast, plays the part of the 'bad' regime, since it leaves the market as the prime institution for the distribution of income and life chances more or less untouched. Liberal welfare states offer, at best, marginal corrections of market outcomes in cases of proven need (means testing): Here the "state encourages the market" and the de-commodifying effect of welfare entitlements "is minimized" (1990: 26 and 27). The Continental-conservative regime, however, is neither truly good, nor truly bad, but simply 'ugly'. It is an ugly regime because it is an 'undecided regime'. On the one hand, it shares the conviction held by the Social Democratic regime that more than marginal market corrections are necessary to secure social peace and equality. Yet, on the other hand, when it comes to welfare state intervention, the conservative regime – according to Esping-Andersen – adheres to the wrong principles. Instead of following the Social Democratic regime by making citizen status the only precondition for welfare entitlements, and by rendering these entitlements generous enough for broad segments of the labor force to break away from market dependency, conservative regimes instead preserve social inequality and limit the re-distributive – and therefore liberalizing – potential of the welfare state through their anachronistic emphasis on status-maintenance, paternalism, and patriarchic social order.[2]


Yet Esping-Andersen's category of the conservative-Continental welfare state regime type is – as I shall argue in this paper – ugly in another sense, namely in an analytical one. We can list three main reasons that support this claim. First, the conservative regime type is essentially a residual category. Within the same category we find welfare states as diverse, for instance, as the German and the Portuguese. In the case of the conservative regime type, variation within the category is much broader than it is for Esping-Andersen's other two regime types (Alber 2001). However, for quite some time, scholars have argued convincingly that at least the Southern European welfare state represents a separate, distinct type with quite distinct features (cf. Ferrera 1996, 1997, MIRE 1997). Re-analysis of Esping-Andersen's data with the help of cluster analysis has supported this claim (Wagschal/Obinger 1998).


Secondly, what is striking in the case of the conservative regime cluster is its lack of geographical fit – especially in comparison with the other two clusters, the Nordic Social Democratic and the Western Anglo-Saxon liberal regimes, which are geographically much more coherent. We don't have to put it as pointedly as Peter Baldwin, who remarked that in Esping-Andersen's account, "proximity to Stockholm" is the main factor "determining the generosity of social benefits in any given country" (Baldwin 1997: 95). But it is true that the European continent in particular harbors some peculiar geographical outliers – with a surprisingly Scandinavian Netherlands, and with liberal Switzerland seemingly located right on the coast of the United States.


Third and most importantly, Esping-Andersen frequently points – if somewhat vaguely – to the influence of Catholic social doctrine on the development and design of the Continental welfare state (1990: 4, 17, 40, 60-61 and passim). Yet he abstains from a more detailed analysis of exactly those European countries of the "dark heartland [...] of religious practice" (Martin 1978a: 271) in which the social importance of religion is most pronounced – a region stretching from Northern Italy to the Netherlands and Great Britain, with Switzerland as its center (ibid: 271-272). It is in precisely this 'religious heartland' that political Catholicism has led to the formation of Christian Democratic parties (Italy, Austria, Switzerland, the Netherlands, Belgium, Germany; Kalyvas 1996: 3), yet in Switzerland and the Netherlands, we also observe the formation of Protestant parties; and in Switzerland, the Netherlands and Great Britain, we see the important impact of reformed and free-church Protestantism. Esping-Andersen's study, however, completely ignores the pronounced anti-welfare state stance of reformed Protestantism. On the other hand, in some of the allegedly 'Catholic' regimes, Christian Democracy has never emerged as a political party (Spain, Portugal, France), or early welfare state legislation had an explicitly anti-clerical motivation (Italy, France, Belgium).


In the following section, I will argue that many of the empirical and theoretical contradictions of Esping-Andersen's regime typology can be avoided if we account for the as yet almost completely ignored role of Protestantism in Western welfare state development (see, however, Heidenheimer 1983). Yet, in order to do so, we must distinguish between the two main currents of Protestantism – state Lutheranism on the one hand and reformed Protestantism and the free-church currents of Protestantism on the other. The existing studies in the field have – if they considered religion at all – focused exclusively on Catholicism, thereby implicitly assuming that the main analytical dividing line runs between Catholicism and Protestantism. As I shall argue in this paper, negative findings with respect to the influence of Protestantism on Western welfare state development can be largely attributed to the fact that studies have failed to account for the fundamental differences between Lutheran and reformed Protestantism. Yet, once we draw this distinction, we have two critical dimensions of variation at hand: the variation between Catholicism and Protestantism and the variations within Protestantism. I maintain that once we place Western welfare states within these two dimensions, we can solve most of the analytical and empirical problems that haunt the conventional 'three-worlds-of-welfare' perspective.[3]


I will present my argument in three steps. In the first paragraph, I shall sketch out the as yet largely ignored influence of reformed Protestantism's social doctrine on the development of the Western welfare state (Section 2). In the following two sections, I will then present comparative data on the formative period of the welfare state (1880-1930; Section 3), as well as on its period of consolidation and crisis (from 1960 until 1990; Section 4). These data support the claim that reformed and free-church Protestantism had a persistent impact on welfare state development in those countries in which reformed Protestantism represented a significant religious current. I will conclude with some reflections on the current regime typologies that inform comparative welfare state research.



2     The welfare state and reformed Protestantism


If we take the religious composition of the population as our – admittedly rough – first indicator of the relative social and cultural influence of different denominations, we move through Europe along the south/north axis from the Catholic to the mixed-denominational to the purely Protestant countries. This would already suggest – as has been frequently proposed in the literature – a distinction between the Southern welfare state model and the corporatistic Continental one (cf. MIRE 1997). Moving along the east/west axis allows us to differentiate within the group of countries with a religiously mixed or with a hegemonic Protestant population: between those countries in which a Lutheran state church dominates (e.g. Sweden, Germany) and – the further West we go – those countries in which the free, reformed, non-conformist, dissenting currents of Protestantism have played an influential role (e.g. Switzerland, the Netherlands, Great Britain, and finally the USA, Australia and New Zealand; cf. Heidenheimer 1983). My main argument is that the critical differences between Esping-Andersen's liberal and conservative models can be attributed to the different impact of reformed Protestantism, whereas the important differences between the Southern, Continental and Northern models of 'social capitalism' must be attributed to the differing roles of Protestant and Catholic social doctrine (and, respectively, of the political parties associated with these doctrines).


This general claim is supported by the fact that once we control for the impact of reformed Protestantism, we can explain many of the institutional and historical peculiarities of the Dutch, British and Swiss cases, all of which have proven highly resistant to any easy categorization within Esping-Andersen's conventional 'three welfare state regimes' typology. It is not by chance, I maintain, that the Netherlands, Switzerland and Britain are the very countries in which reformed Protestantism has had a discernible influence. Initially, Esping-Andersen himself counted the Netherlands among the Social Democratic regimes (1990: 52, 74 Table 3.3, 87). Yet the Dutch 'employment regime', which - according to his own theory - should be seen in close connection with the welfare state regime, exemplifies the classic Continental case of 'welfare without work' (cf. Visser/Hemerijck 1997). This later led Esping-Andersen to characterize the Dutch employment regime as typical of "conservative Continental welfare states" (Esping-Andersen 1996: 84). Today, it seems, he would like to characterize the Dutch welfare state in general as 'conservative' (Esping-Andersen 1999: 77, Fn. 7), yet he hastens to add: "Among the Continental European countries, only the Netherlands deviates markedly from the corporatist mould" (ibid: 82). Given this ambiguity in Esping-Andersen's own account, the confusion in the wider literature is anything but surprising. In accordance with the 'early' Esping-Andersen, Goodin et al. (1999) also treat the Netherlands as a prototype of the Social Democratic regime. Yet they admit that there are serious difficulties with this classification (1999: 11-12). In contrast, Kersbergen (1995: 56), Castles/Mitchell (1993: 123, Table 3.7), Visser/Hemerijck (1997: 127), Scharpf/Schmidt (2000) and Huber/Stephens (2001) all label the Netherlands as a Continental/conservative/Catholic regime. In a detailed comparison, however, Jens Alber (1998) recently showed that the Dutch welfare state deviates in critical respects from the two classical conservative welfare states of Germany and Austria (1998: 57, 10). The Dutch old age pension – based on citizen status without a link to occupational status and financed through general contributions (not by payroll taxes) – has no parallel in Austria or Germany. The same holds true for the Dutch company-based or company group-based health care, unemployment and pension insurance schemes, mainly financed out of employers' contributions, which have no equivalent in the other Continental welfare states. Alber summarizes: "With respect to welfare state structures, Austria and Germany can thus be called birds of a feather, but the Netherlands stand clearly apart as a different kind" (Alber 1998: 57). In light of this dispute, it is somewhat surprising that Kersbergen claims that "mainstream welfare state typologies consistently place Germany, Italy and the Netherlands in the group of conservative welfare state regimes" (Kersbergen 1995: 56; emphasis mine).


With respect to the classification of the UK, Esping-Andersen himself remains undecided (1990: 87, 74, Table 3.3). The British welfare state shares universal features with the Social Democratic regimes (see the NHS, but also the national pension insurance), yet social protection schemes are, in contrast, residual in character and leave a great deal of room for private arrangements. This would suggest placing the UK under the heading of 'liberal regime' instead. The addition of an income and contribution-based 'second tier', which would also have been attractive for the middle class, happened very late in Britain compared to the Scandinavian countries, and was then quickly annulled by Margaret Thatcher (Hinrichs 2000). The Swiss case also sits uneasily with Esping-Andersen's typology. He himself counts Switzerland among the liberal regimes. Scharpf/Schmidt (2000) and Huber/Stephens (2001), however, place it among the 'Christian Democratic welfare states'. Finally, Obinger describes Switzerland as a "predominantly liberal-conservative mixed type, enriched with Social Democratic elements " (Obinger 1998: 18; my translation) – in other words, a little bit of everything.


These problems of categorizing the three European countries in which reformed Protestantism was influential suggests that there might be a neglected dimension in the process of state and welfare state development in Europe, and that Protestantism – in contrast to the received wisdom in the literature (Kersbergen 1995; Huber/Ragin/Stephens 1993; Esping-Andersen 1990; Lagner 1998) – has contributed in a substantial and distinct way to the development of the Western welfare state, even if its contribution was sometimes rather indirect and 'negative'. If we describe reformed Protestantism's contribution to Western welfare state development as 'negative', we are referring first of all to the strongly anti-étatist position of the Protestant free churches and other reformed currents of Protestantism (Dissenters, Calvinists, Baptists etc.). Their emphasis on self-help, the autonomy of the holy local congregation, strict state/church separation (with a church that was conceived to be de-centralized, local, democratic and Congregationalist instead of Episcopal in character), and individual asceticism and prudence propagated a strict anti-state program (often reinforced by traumatic experiences of religious persecution) that - in numerous ways – had a retarding effect on welfare state development. This is true of all the countries in which reformed Protestantism was influential, either because a significant minority belonged to this denomination, and/or because reformed Protestantism made a formative impact on the culture of these societies through the first settler cohorts of 'Dissent' immigrants (Martin 1978a: 237).


The retarding influence of reformed Protestantism is clearly exemplified by the Dutch orthodox Calvinist Anti-Revolutionary Party (ARP). The ARP's Calvinism had a radical anti-state attitude. Its main aim was sovereignty in one's own circle ('souvereiniteit in eigen kring'), in which the state "cannot intervene and cannot command on the basis of its own power" (the Calvinist priest and charismatic leader of the ARP, Abraham Kuyper, quoted from Kersbergen 1995: 60). As a consequence, the ARP frequently voted "against social legislation and as a government party failed to develop social policy" (ibid. cf. also Kossmann 1978: 496-497; de Swaan 1988: 210-216; Cox 1993; Fix 2001: 113-116). This was important given the fact that the ARP, as an electorally successful party, participated in each and every Dutch government between 1901 and the German occupation in 1940, with only two minor interruptions (1905-1908 and 1913-1918). Even before the turn of the century, the Calvinists had been in government as part of a coalition with the Catholics. The ARP thus played a critical role in Dutch politics throughout the entire period from the late 19th century until the 1960s. It is thus not surprising that the Dutch welfare state is a clear laggard among Western welfare states once we account for the already high level of economic development in the Netherlands in the late 19th and early 20th centuries (see below, section 3).[4]


A similar point can be made in the case of Great Britain. Non-conformism had first been influential in the Liberal Party and subsequently achieved prominence in the Labour Party and its social policy program (Parry 1986). Non-conformism's "emphasis on individual freedom and value encouraged an accent on self-improvement and … laisser faire attitudes which were far from compatible with an interventionist approach to welfare" (Catterall 1993: 683). Early British social legislation was influenced by the Liberals' sweeping victory in the 1906 elections, and this victory was due not least to "Nonconformist religious revivalism buoyed up by opposition to the 1902 Education Act" (Powell 1992: 33). By its own definition, the new social liberalism so prominent among the left-liberal circles in Britain around the turn of the century was "practical Christianity" (Freeden 1978: 50-51), combining state welfare with an emphasis on personal responsibility and individual initiative (Freeden 1978). The Protestant ethic of self-discipline, merit, and personal responsibility for the improvement of one's own lot also guided the British working class movement (MacLeod 1984; Pelling 1965). British skilled workers emphasized prudence and foresight in planning for the future, thrift, and discipline – and this was more than a simple emulation of bourgeois values and attitudes. Finally, workers' emphasis on own responsibility intermingled with their more material interest in upholding and protecting workers' voluntary organizations for mutual self-help, such as the friendly societies (de Swaan 1988: 192-197). This ideological and institutional heritage, then, had a long-lasting, restrictive influence on British welfare state development, where universal minimum standards were established to complement the existing voluntary schemes. Even the oft-praised 'path-breaking' Beveridge reforms implemented after WW II have been described as "no more than the administrative completion of the moral programme of British nineteenth-century nonconformism" (Milward 1992: 43).


Compare this to the Swiss case: Qualified workers with a strongly developed occupational ethos emphasized professional status and group solidarity. A crucial element in this was the religiously-influenced opinion that voluntary collective self-help should be called upon first before any state help is implemented (Luebbert 1991: 49). Moreover, in Switzerland, the reformed Conservative Party – and later the Calvinist Evangelische Volkspartei as well – organized successful referenda against the central government's social legislation at several different points in time (Gruner 1977). Thus, it is not Swiss federalism alone which should be held responsible for the laggard status of the Swiss welfare state. There is also ample evidence showing that the roots of the progressive movement in the United States lay in New England's free-church Protestantism (Thomas 1983; Spann 1989), a reform movement which fiercely attacked 'machine politics' – in particular the political clientelism widespread in the practice of awarding veterans' pensions to loyal party followers. This protest effectively retarded national social legislation until the New Deal (Skocpol 1992, 2000). Very similar Protestant ideas about a "Christian capitalism that combined equal opportunity with Christian benevolence" (Thomas 1983: 13) have been influential in New Zealand and Australia. The working class movement here was less inspired by Karl Marx than by Christian Utopians like Henry George or Edward Bellamy, with their reform visions of the 'single tax', land and urban reform, teetotalism, universal school education etc. (cf. Clark 1976; Castles 1985). The fact that progressive personal income taxes play a major role in welfare state financing in these countries (in contrast to social insurance contributions in the Continental welfare states) can be traced back to this ideological background. The Fabians and their variation of Christian socialism have been particularly influential in Australia (Matthews 1993). Generally, the 'liberal' emphasis on individual responsibility should not be mistaken for pure market liberalism. Often individual responsibility was understood in moral, not in economic terms. For instance, in New Zealand, social assistance was refused in cases of divorce, and in the Scandinavian countries, alcohol abuse led to disqualification from welfare benefits.


In contrast, in countries in which a Lutheran state church dominated – in the Scandinavian countries[5] – not much stood in the way of the government taking over responsibility for the welfare of its citizens. Lutheran Protestantism did not claim supremacy over the nation state like the Catholic Church did in southern Europe – giving rise to fierce conflicts between clericals and anti-clericals once these societies switched from dynastic to democratic rule (Flora 1986: XVIII). Nor did Lutheran Protestantism – in contrast to the free churches and reformed sects – want to assign local congregations maximum autonomy with full responsibility for the welfare of their members, as small 'Sondergesellschaften' (as nations within the nation) (see Blückert 2000). To put it very pointedly: If the Church of England has been called the Tory Party at prayer, we might describe the Swedish Lutheran Church as the nation at prayer – as "the public agency for moral and religious nurture" (Gustafsson 2003: 54), with its central gospel being, 'Pray, pay and obey'. In societies where "identification between church and state" was "total" (ibid: 51) there was not much reason to protest against the central state taking over responsibility in the welfare arena. In fact, the notion of separate spheres for church and state was not very well developed: Priests were civil servants; as late as 1858, anyone who converted to Catholicism could be exiled. It was only in the late 1920s that parishes had to give up their monopoly on education, and it was only in 1952 (!) that the Swedish state granted full religious freedom as a matter of law. "Nobody was henceforth forced to belong to the Church of Sweden against his or her own free will" (Gustafsson, 2003: 55; Petersson 1994: 196) – but even then, almost 99% of the population remained members of the Church of Sweden. Thus, it is more than evident that Protestantism Swedish (or Scandinavian) style was very different from Anglo-American Protestantism, where religious liberty – combined with passionate resistance against any state involvement in private religious affairs – was the prime goal of the first cohorts of Dissent immigrants. It is therefore not at all surprising that studies which lump these two fundamentally different strands together under the simple label 'Protestantism' cannot find much evidence for a discernible religious influence on welfare state development.


In fact, the close relationship between state and state church did provoke protest by a free church movement in Sweden, as it did in Britain. In Norway, a similar line of conflict was blurred by its overlap with a center/periphery conflict between the urban Danish elite and the Norwegian rural population (Rokkan 1967). Here, a genuine Protestant Christian Democratic party (Kristelig Folkeparti) was formed, but only comparatively late, in the 1930s. In Sweden, Protestant revivalism became manifest in a strong temperance movement that exerted a significant influence on the Liberal and Social Democratic Parties, as did British nonconformism on the Liberal and the Labour Parties. To some extent, this may explain the similarities between the two welfare states up until the 1950s. Yet the historical evidence suggests that Swedish free churches remained less influential than their British counterparts; and a comparison of the conflicts over the 'disestablishment' of the church and over church control of the school system in Britain and Sweden makes it clear that similar struggles were waged with quite different intensity. On the other hand, it should also be stressed that Britain – with its Anglican state church – cannot be counted as an example of the purely 'liberal' model, even if we take into account the stronger role of British disestablished churches.


Although Lutheranism also dominated in Germany, the constellation there was very different due to the Reich's substantial Catholic minority (Manow 2000, 2001).[6] Germany shared this constellation with the Netherlands (although the two countries differed with respect to the importance of reformed Protestantism), and they also shared the central strategy of handling religious conflicts or other social cleavages via the welfare state. The welfare state was used to consolidate and stabilize political and religious camps and to resolve conflict through consociational techniques of 'amicable agreement' and parity representation. The different pillars, or Lager, all formed their own welfare organizations, which gained privileged status in the field of welfare provision and which became stabilized and subsidized through 'corporatistic' welfare state programs (for Germany see Sachsse/Tennstedt 1992: 166-185, 1998). Conflicts over education such as the Dutch school strijd were resolved using a method which also became dominant in the fields of charity and welfare: The state subsidized religious schools and social services, and charity provided by the churches was financed by the new welfare state programs. As a consequence, today, a large third sector is responsible for much of the welfare provision in these Continental welfare states.


The influence of religion had at least two dimensions (institutional and cultural), and we are well advised to take both into account. On the one hand, religious cleavages not only refracted the formation of societal interests along class lines, but they also constituted a conflict line in their own right – especially in all questions regarding the 'division of labor' between state and church in the field of welfare services. On the other hand, we should be aware of the fact that religion played – and still plays – an extremely important role, often being the most prominent element in the national "vocabulary of legitimation" (Martin 1978a: 242). One example is the case of the USA, where religious pluralism prevented the church(es) from taking sides in the capital/labor conflict. The consequence was that public debates could be cast in profoundly religious terms, which, in the American context – with its influence from Protestant sects and free churches – meant deeply individualistic terms (Martin 1978a: ibid.). The consequences are well known: Individual self-help and local charity had the right of way; the state – if it played any role at all – could only be a sponsor and supporter of the 'liberté subsidiée' in the area of welfare. Yet, the recipients of this 'state help for self-help' were mainly voluntary organizations in the US, whereas they were the churches in Continental Europe – a crucial distinction, as the Dutch case exemplifies. The system of state-subsidized welfare provision through third parties of mostly religious pedigree led – in a religiously mixed country like the Netherlands – to "consolidated competition between rival religious and secular oligopolies" (Heidenheimer 1983: 15; Martin 1978a: 240). The consequence was strong welfare state growth after 1960 - which was also intended to compensate materially for the pillars' loss of social cohesion. But the dual structure of state-subsidized, group-specific voluntary programs and universal, citizen-based state provision of social protection also had the potential for a long-lasting restrictive impact on overall social spending. Switzerland and the UK would be cases in point. In Switzerland, competitive overbidding for more generous social protection between the Lager was prevented by federalism; in Great Britain, there were no different religious camps to compete with the labor movement for higher social protection standards.


In contrast, workers in Southern Europe - when they fought against the ancient regime - always fought against the clergy as well. The supra-national Catholic Church had been closely attached to the international dynasties that ruled absolutist Europe, which automatically brought the church into conflict with the new republican nation states. Therefore, it is not surprising that liberal parties in these southern countries often introduced new social programs against the explicit will of the Catholic Church (for Italy cf. Sellin 1971). In fact, much of the new social legislation had quite openly anti-clerical motives. This already speaks against the general assumption that Catholic social doctrine was dominant in the development of the Southern or Continental welfare state. Where Catholicism played a role, it was a more indirect and partially refracted one. For instance, it was only later that Christian Democratic parties such as the Italian Democratia Cristiana used the welfare state as a clientelist resource in their efforts to mobilize voters (Lynch 2004; Flora 1986: XVIII). But this was primarily a means of becoming more independent from the official church hierarchy – whereas the Christian Democratic parties themselves were the rather unintended outcome of the church's fight against liberalism (Kalyvas 1996).


Very similar starting conditions were able to trigger quite distinct national developments, as the brief comparison between the Netherlands, Switzerland and the UK above has shown. Nevertheless, closer inspection reveals many common features and structural similarities in the welfare state in countries influenced by reformed Protestantism. These common features, I maintain, are evidence of their common religious roots (cf. Sections 3 and 4). And although social Protestantism – in contrast to Catholic social doctrine – did not develop a strong positive legitimation for social policy intervention, this must not lead us to the conclusion that Protestantism can be ignored as an explanatory variable in the analysis of welfare state development (cf. Kersbergen 1995: 254-255, endnote 1). Of course, I do not propose to explain Western welfare state development exclusively in religious terms. Important variables such as the strength of the labor movement, the level of industrialization, constitutional features, etc., all remain important. Yet, as I will show in the following sections, the influence of the churches, of parties with a religious leaning and of the different social doctrines crucially influenced welfare state development in the Western world – and, in critical respects, altered, modified, and refracted the influence of the conventional explanatory variables such as working class strength, level of economic development, and strength of non-majoritarian institutions.


In the following sections, I shall present a first test of my argument in two steps. One central implication of the argument relates to the question of the timing of welfare state development. My argument will pass a first test if, indeed, all the countries with a discernible influence from reformed Protestantism can be shown to have been welfare laggards. This is what I will show in Section 3. In Section 4, I will then explore the question of whether we can observe a sustained influence of Protestantism that reaches beyond the early formative years of welfare state development.



3     Has reformed Protestantism delayed welfare state development?


The argument as presented above suggests that in nations in which reformed Protestantism was of more than marginal importance, the welfare state developed relatively late. In contrast, in countries where a Lutheran state church dominated, we can expect to see that the central state took over responsibility for the 'labor question' at a comparatively early point in time. In other words, the argument predicts that European countries such as Switzerland, the Netherlands or Great Britain – plus the USA, Canada and Australia/New Zealand should be found among the late-developing welfare states; whereas countries like Germany and Sweden, with their dominance of a Lutheran national church, should be expected to have taken a pioneering role in welfare state developments.


Relative 'earliness' or 'lateness' can be understood in two ways – in a chronological sense or in an economic one. Chronological time can be measured as the date of the first social legislation;[7] economic time can be measured, for instance, as the GDP per capita at the time of the first social legislation. Countries, then, can be late in one, two or in no dimensions (see Figure 1, cf. Wagschal 2000: 49): They can be later than average in a chronological sense (quadrant I) and/or later than average in an economic sense (quadrants II and III), or they can be late in none of these two dimensions (quadrant IV). Switzerland, Canada and the United States are clear welfare state laggards, since they are late in both dimensions. Our other reformed Protestant countries, such as New Zealand, Great Britain, and the Netherlands are 'early' in a chronological sense of the word (Australia is a borderline case) – yet without exception, all of them are late in an economic sense.





A more precise picture emerges if we do not restrict our analysis by looking only at the introduction of the first social program. 'Time' becomes a less arbitrary indicator if we look at all major social programs – old age, accident, health, and finally unemployment insurance. The relative economic timing of social legislation can then be measured in two ways: either as the GDP per capita at the time of introduction of each for these social protection programs averaged over all four dates, or as the level of economic development at the 'average' (hypothetical) introduction year calculated as the mean of the different years of major social legislation. With respect to the group of welfare state laggards, the two procedures lead to almost identical results (Table 1), whereas the composition of the groups of early and 'normal' welfare states varies a bit more depending on which of the two methods of calculation is used.



Table 1

Rank order according to level of economic development at the moment of introduction of accident, sickness, old age and unemployment insurance as well as family support (1 = lowest, 21 = highest level)

Average per capita GDP in the 'introduction years' of the welfare state

Per capita GDP in the average 'introduction year'

1. Greece ($ 10,077)

1. Portugal ($ 1,780)

Mean - 1 Standard Deviation ($ 10,783)

2. Norway ($ 2,079)

2. Spain ($ 10,812)

Mean - 1 Standard Deviation ($ 2,188)

3. Portugal ($ 13,204)

3. Finland ($ 2,476)

4. Norway ($ 13,278)

4. Greece ($ 2,526)

5. Italy ($ 13,421)

5. Spain ($ 2,525)

6. Austria ($ 13,386)

6. Sweden ($ 2,584)

7. Ireland ($ 13,540)

7. Italy ($ 2,703)

8. France ($ 16,850)

8. Japan ($ 2,709)

9. Finland ($ 16,951)

9. Ireland ($ 2,800)

10. Germany ($ 17,031)

10. France ($ 3,217)

11. Japan ($ 17,075)

11. Austria ($ 3,312)

12. Denmark ($ 17,763)

12. Germany ($ 3,416)

13. Sweden ($ 17,885)

13. Denmark ($ 3,417)

14. Belgium ($ 19,490)

Mean ($ 3,836)

Mean ($ 20,958)

14. Belgium ($ 3,887)

15. Netherlands ($ 21,750)

15. Netherlands ($ 4,324)

16. New Zealand ($ 25,274)

16. UK ($ 5,032)

17. UK ($ 25,277)

17. Australia ($ 5,095)

18. Australia ($ 29,464)

Mean + 1 Standard Deviation ($ 5,483)

Mean + 1 Standard Deviation ($ 31,134)

18. New Zealand ($ 5,670)

19. USA ($ 35,955)

19. Switzerland ($ 6,026)

20. Switzerland ($ 43,816)

20. USA ($ 7,018)

21. Canada ($ 47,829)

21. Canada ($ 7,929)

Source: own calculations based on Schmidt (1998: 180 Table 5) and Maddison (1995: 194–201, Table D-1a and D-1b).



The cluster of late welfare states is apparently quite stable, and comprises all those countries that already qualified as 'late' in our first analysis, which was restricted simply to the introduction of the first major social protection program (with Belgium being a borderline case; see above Figure 1; quadrant II and III).[8] Needless to say, the Commonwealth countries of Australia, New Zealand and Canada, as well as the USA, share a substantial influence of reformed Protestantism with our European countries – the UK, the Netherlands and Switzerland (Mol 1972; Barrett 1982).[9] The data therefore quite clearly confirm a pattern of a delayed "westward spread of the welfare state", which Arnold Heidenheimer explained on the basis of religious factors as early as 1983.


This pattern is confirmed if we substitute per-capita GDP with other variables measuring economic development, such as the share of first sector employment or degree of urbanization at the moment of the first major social legislation (cf. Alber/Flora 1981; Alber 1982). We could expect Switzerland to be an outlier with respect to urbanization, given that much of Switzerland's early industrialization took place in rural areas[10] – but apart from this aspect of the Swiss case, Tables 2 and 3 confirm that we have to include Switzerland, the UK and the Netherlands among the group of welfare laggards (Tables 2 and 3; for similar results see Flora/Alber 1981 and Alber 1982; Obinger/Wagschal 2000 do not include Britain and the Netherlands in their group of late welfare states, which then, of course, renders their explanation for relative welfare state timing highly problematic).



Table 2

Rank order, level of economic development at the moment of first introduction of a major social insurance program for 12 Western European countries (1st sector employment share of total employment; Rank 1 = highest share, Rank 12 = lowest share)

1. Finland
2. Italy
3. Austria
4. Denmark
5. Sweden
6. France
7. Norway
8. Germany
9. UK
10. Switzerland
11. Netherlands



Some of the data had to be estimated, since I could not find precise information on agricultural employ-ment for all points in time.
Source: Own calculations, based on Flora (1983 and 1983a).



Table 3

Rank order, level of urbanization at the moment of introduction of the first major social insurance program for 13 Western European countries (share of population living in cities with more than 100,000 inhabitants; Rank 1 = lowest share, Rank 13 = highest share)



2. Finland
3. Norway
4. Sweden
5. Switzerland
6. Belgium
7. Italy
8. Germany
9. Ireland
10. Denmark
11. Austria
12. Netherlands
13. UK

Sources: Own calculations, based on Flora (1983 and 1983a).



Still another way of proceeding would be to look at the coverage rate of the new social insurance programs with respect to those people employed dependently outside of agriculture. One could well argue that the existence of legislation itself does not provide us with enough information about the 'real' impact of the new welfare state and that instead, we need to ask how relevant and encompassing the new social insurance schemes were in terms of coverage. Table 4 compares welfare state coverage rates for workers (in the 2nd and 3rd sector) for four groups of countries: the Nordic (Norway, Sweden, Finland and Denmark), the Continental (Belgium, Austria and Germany), the Southern (Italy and France)[11] and – for lack of a better name - the 'belated-Protestant' welfare states (the Netherlands, Switzerland and the UK) at five points in time (1890, 1900, 1910, 1920, 1930). Again, we can confirm the previous findings that the Netherlands, Switzerland and Great Britain clearly belong to the group of Europe's 'late' welfare states.



Table 4

Coverage of social insurance schemes as a % of those not employed in agriculture, 1890–1930

  1890 1900 1910 1920 1930

Nordic welfare states
(Nor, Dnk, Fin, Swe)






Continental welfare states (Austria, Bel, Ger)




46.7 64.9

Southern welfare states
(Ita, Fra)

1.5 9.7 12.2 39.9 52.6

'Belated-protestant' welfare states (Swz, Ntl, GB)






Sources: Flora (1983, 1983a), own calculations.



A further – rough – confirmation is provided by a simple OLS regression. The literature provides us with some standard explanations for the timing of welfare state development. Most prominent among the factors that are said to be responsible for an early takeoff of the welfare state are the strength of the labor movement and the authoritarian character of the state in the last quarter of the 19th century (cf. Alber 1982). To test for these factors, I took the following as my independent variables: union density before 1913 (Stephens 1979: 115), left vote before 1918 (Bartolini 2000), and the 'authoritarian legacy' index of Huber/Ragin/Stephens (1993). My dependent variable is the level of economic development at the moment of the first introduction of a major social protection program (Schmidt 1998: 180). The inclusion of a dummy for those countries in which more than 15% of the population were members of a reformed Protestant church around 1900 (Mol 1972) renders all three other variables insignificant, while the dummy has a high and robust explanatory power (see Table 5).



Table 5

Determinants of the relative timing of welfare state development in 16 countries (OLS regression)

Dependent variable: level of economic development at the moment of introduction of the first major social protection program

Model 1
coefficients (standard error)

Model 2
coefficients (standard error)

Union density (1913/14)




Left vote share (before 1918)  



'Authoritarian legacy' –139.42




Reformed Protestantism





adj. R2



Sources: Economic development, Schmidt (1998: 180); union density, Stephens (1979: 115); left vote share, Bartolini (2000); 'authoritarian legacy', Huber/Ragin/Stephens (1997); reformed Protestantism, Barrett (1982).



But what happens if we take a look at the postwar period? Can we detect an influence of Protestantism that reaches beyond the early formative period of welfare state development?



4     Social Democracy, Christian Democracy – and social Protestantism?


Apart from the substantial delay in welfare state formation, can we detect institutional similarities inherent in the free-Protestant welfare state type? Does developmental delay translate into specific dynamics of subsequent welfare state expansion after 1945? Given the diverging postwar developmental paths of, say, the Dutch and the Swiss welfare states, we may harbor serious doubts about our ability to identify common features among our group of 'reformed Protestant' regimes. Difficulties are already apparent from the fact that even today, these states are often seen to belong to entirely different regime types (see above). Moreover, secularization has significantly reduced the impact of religion on Western postwar societies. Ironically, insofar as the welfare state has helped to reduce the "sense of existential insecurity", of personal "vulnerability …. and survival-threatening risks" it has presumably also helped to dry out one of the strongest sources of religiosity and longing for religious guidance (Inglehart/Norris 2004: 2). Welfare state growth itself can therefore be seen to be (at least partially) responsible for the loss of the moral and cultural strength of religion in modern societies. We would then also expect that the more mature a Western welfare state has become, the less visible a religious influence on its development should be.


Yet, whereas secularization theory is about attitudes and behavior, my argument is about institutions. And institutions tend to outlive the motives and interests that were important at the moment of their establishment. Given what we know about the importance of historical sequencing (cf. Pierson 1999; Kamens 1976), we should also expect long-term consequences related to the substantial differences in the timing of welfare state development. And, in fact, the literature frequently refers to these sequence effects to which Abram de Swaan, for instance, has referred as the "law of arresting advance" (de Swaan 1988: 157). If the state takes over social responsibility rather late, private arrangements are already in place. Once these private, voluntary arrangements meet with a crisis, the state intervenes – but in a specific, restrained way: not by substituting for the existing arrangements entirely, but by regulating and subsidizing them more strongly, as well as by providing for complementary protection with additional schemes built around the existing arrangements. This usually means that welfare state provision comes in the form of universal basic protection built around group-specific programs, because "where the voluntary collectivization of [welfare] provision had preceded very far and mutual societies or unions had built up extensive institutional networks of their own, they tended to resist any takeover by the state and to oppose compulsory arrangements" (de Swaan 1988: 157).


Union-run friendly societies are a good example. As Jens Alber has shown, the countries in which unions developed early are those countries in which compulsory social insurance came rather late (Alber 1982: 145-146, Table 17). Yet, it was the Protestant countries in which workers organized themselves at a relatively early point in time. This sequence then leads to a dualism between private, group-specific insurance schemes with a varying degree of public regulation and imposed obligation on the one side and a comparatively basic state provision of social protection on the other (cf. Esping-Andersen 1990: 24-25). This dualism characterizes the Swiss, Dutch and British welfare states equally. It was also those unions that tried to defend their own schemes of voluntary self-help that at the same time fought for state intervention to increase security at the workplace. This explains why the UK and Switzerland were laggards with respect to social protection programs, but forerunners in the field of work safety regulations.


The long-term effects of the institutional decisions made during the formative period of welfare state development become clearer if we take a look at specific welfare state programs – for instance, old age insurance. Here we can observe some striking similarities between the Netherlands, Great Britain and Switzerland, which these countries share with our other 'belated Protestant' welfare states. One common feature is that all the countries in this group put a much stronger emphasis on fully funding public pensions, whereas all the other mature welfare states follow the pay-as-you-go method of pension financing. Furthermore, we can observe that the reformed Protestant countries have established a dual pattern of old-age protection which combines a basic, universal, tax-financed public scheme with a private second pillar (see Table 6). Admittedly, the Social Security system in the United States follows classical 'Bismarckian' insurance principles, yet low income ceilings force broad segments of the labor force to complement Social Security entitlements with private/company pensions. Complementary schemes come in the form of company pensions in Switzerland, the UK and the USA, whereas in the Netherlands, they are organized at the level of the industry sector. Despite these institutional differences, similarities predominate.[12]



Table 6

Pension insurance in the OECD, share of GDP and importance
of the 'second pillar'

  '1st Pillar'
Public Pension
'2nd Pillar'
Company Pensions
  Expenditures in % of GDP (1995) Assets as % of GDP (1993)





5.2 25
Netherlands 6 85
New Zealand 5.9 n.a.
Switzerland 6.7 70
UK 4.5 82
USA 4.1 72
Ø 5 63.1
Austria 8.8 n.a.
Belgium 10.4 10
France 10.6 3
Germany 11.1 6
Italy 13.3 4
Spain 10 2
Ø 10.7 5
Denmark 6.8 21
Finland 11.5 0.5
Sweden 11.8 16




Source: Disney (2000).



With respect to postwar development, we can observe that in all of these countries, an income-related second tier as a complement to the basic protection scheme was introduced only relatively late (Hinrichs 2000: 356-366). In his study of pension reforms, Karl Hinrichs identified five 'latecomers' in this respect - the Netherlands, Australia, Great Britain, Switzerland and New Zealand, which did not introduce an income-related second tier at all. Moreover, if a second tier was introduced, it was introduced in a specific way. Whereas the Swedish ATP pensions followed Bismarckian principles, governments in our group of countries regulated and subsidized existing private or company pensions and made them obligatory. Furthermore, delays can be observed not only with respect to the introduction of the first welfare programs, but also with respect to their subsequent expansion. This had important consequences for the responses of these countries to the times of austerity after 1973/74 (Hinrichs 2000), given that entitlements recently granted are usually easier to cut than mature, well-entrenched entitlements (Pierson 1994).


Another – and very different – welfare state sector which underlines the impact of religion is child care and family policy. The fact that this is an important welfare state sector is beyond dispute, given that a child care infrastructure – via its impact on female labor force participation – has had a profound influence on the political economies of the OECD-countries, with consequences for the employment structure (part-time/full-time, industry vs. service employment) and the overall employment rate, and profound consequences for the productive use of human capital investments. As the low-fertility equilibrium of the Continental welfare states shows, this feature of the Western welfare state has also had substantial feedback effects on the long-term fiscal sustainability of welfare states themselves (Esping-Andersen 1999: 67-70). Yet, once we investigate the 'women-friendliness' (Hernes) of welfare states, the traditional welfare state typology does not seem to be of much help. If, as Esping-Andersen claims, "the social democratic regime's policy of emancipation addresses both the market and the traditional family" (Esping-Andersen 1990: 28), we should expect high female labor force participation in combination with encompassing child care provision in all countries that saw an impact from the Left similar to that in the Scandinavian countries – e.g. in Austria or Australia. The data, however, do not support this expectation. Moreover, Esping-Andersen's claim that "day care, and similar family services, are conspicuously underdeveloped" in conservative-Continental welfare states (1990: 27) is clearly contradicted by the French, Belgian and – to some extent – the Italian cases (Morgan 2002, 2003, 2004). These are countries we might expect to have patronizing conservative-Catholic welfare states with a dominance of a traditional familialistic ideology and the male-breadwinner role model as the regulative idea of the labor market. Yet, in France, Belgium and – less strongly – in Italy, child care has been 'de-familialized' as strongly or even more strongly than in the Scandinavian welfare states, and labor force participation by mothers is high (see Table 7). At the same time, these 'Catholic outliers' were not strongholds of a feminist movement, to say nothing of strong unions (whereas the claim that the labor movement was always in favor of a high female labor force is historically untenable).


The feminist welfare state literature, in turn, which rightly criticized Esping-Andersen's 1990 contribution as largely 'gender blind', has not yet come up with a plausible alternative explanation for why some welfare states are much more women-friendly than others (Lewis 1992; Orloff 1993; Lewis/Ostner 1994; Sainsbury 1994, 1996). Only recently, Kimberly Morgan has presented a highly convincing account of the French, Belgium and Italian 'exceptionalism' (Morgan 2002, 2003, 2004), and her argument is supported by Birgit Fix's comparison of family policies in Belgium, the Netherlands, Germany and Austria (Fix 2001). Both authors agree that it were intense church/state conflicts over education which finally led to the strong role that the French, Belgian and Italian states play in child care and early education and for the strong role that the 'third sector' plays in child care service provision in the Netherlands and Belgium. The varying importance of mostly religious, third sector organizations in the provision of care for the elderly has been interpreted along similar lines (Alber 1995).



Table 7

Child care and female labor force participation rates in Western Europe
in the 1990s


Percentage of children < 3
years in public child care

Percentage of children 3-6 years in public child care

Labor force participation of mothers, married/ cohabiting Labor force participation of mothers, single mothers
Austria 6 80 (part time) 46 58
Belgium 30 95 61 68
Denmark 48 83 84 69
Finland     70 65
France 34 99 68 82
Germany 2 78 (part time) 41 40
Ireland 2 55 (part time) 32 23
Italy 6 91 41 69
Netherlands 10 71 (part time) 52 40
Norway 23 61 77 61
Portugal 12 48 55 50
Spain 2 84 (part time)    
Sweden 40 83 80 70
United Kingdom 2 60 (part time) 62 41

Source: Morgan (2003: 265, 267, Tables 1, 2, and 5).



Still another welfare state sector in which the church/state relationship was very influential is social assistance. Here, the fact that it is the Catholic countries of Europe's South in which national social assistance schemes have either not been introduced at all up to the present day (Portugal, Spain, Italy, Greece) or have been introduced extremely late (France; cf. Matsaganis et al. 2003; cf. Kahl 2004), again suggests that religion in general and the role of state/church relations in particular are important for the institutional setup of today's welfare state. The largely varying size of the third sector may be mentioned in this respect as well (Evers/Laville 2004).


As a further test of the more long-term institutional effects of the – religiously influenced – timing of welfare state development, we may finally switch again from an institutional perspective towards a comparative-quantitative analysis by asking: Is there a measurable influence of 'reformed Protestantism' on aggregate indicators of welfare state activity such as 'total social expenditure'? Table 8 shows the results of a pooled analysis in which social expenditure (as a share of GDP) in 19 OECD-countries from 1962 to 1989 figures as the dependent variable. The independent variables comprise the 'usual suspects' of the comparative welfare state literature, such as cabinet share of Left and Christian Democratic parties, the unemployment rate, the dependency ratio (population share of those under age 14 or over age 65), GDP per capita, union density and importance of non-majoritarian institutions (cf. Huber/Ragin/Stephens 1993). A dummy controls for the influence of reformed Protestantism in those countries in which more than 15% of the population belonged to a reformed Protestant denomination in 1960 (or close to 1960, depending on data availability; Mol 1972; Barrett 1982). We cannot measure the partisan-political influence of reformed and free church Protestantism directly, because religious parties of this leaning formed in only a few countries (New Zealand, Switzerland, Norway, the Netherlands). Therefore, we cannot simply measure the relative vote, seat, or cabinet shares of Protestant parties. Moreover, as I have argued above, in many countries, reformed Protestantism exerted a profound influence on the programmatic stance of left and liberal parties. It would therefore be too restrictive to focus solely on explicitly/programmatically Protestant parties.



Table 8

Social expenditures as a share of GDP, 19 OECD countries 1962-1989

Dependent variable: Social expenditures as a share of GDP Model 1
Non-standardized coefficients (standard error)
Model 2
Non-standardized coefficients (standard errors)
Share of population above 65 and below 14 years 0.1243
Unemployment 0.5276
Union density 0.0734
Per capita GDP 0.0005
Non-majoritarian institutions –0.2586
Cabinet share, left parties 0.3269
Cabinet share, Christian Democratic parties 0.2448
Reformed Potestantism   –4.658
R2 0.26 0.26

Prob > c2






NObs 481 481
Remarks: Combined cross-section and time series analysis with OLS, no constant, panel-corrected standard errors and correction for serial autocorrelation in the residuals. ** p < 0.05; *** p < 0.01. Source: Own calculations based on Huber/Ragin/Stephens (1997).



Table 8 provides confirmation with regard to the explanatory strength of the standard variables – the only exceptions being the role of non-majoritarian institutions and the impact of the dependency ratio. What is striking, however, is the strongly negative, robust coefficient of the Protestantism dummy. It is also remarkable that the inclusion of the dummy renders the variable 'cabinet share of Christian Democratic parties' insignificant, even though the Netherlands – with its generous and encompassing welfare state – figures here among the free-Protestant countries, and even though the period covered by the panel excludes the first postwar decade. Therefore, the panel actually discriminates against a hypothesis which predicts delayed welfare state development in the countries with a significant influence of reformed Protestantism.[14] Against this background, the results in Table 8 might lead one to suggest that some of the influence on welfare state development previously ascribed to Christian Democracy might have been falsely attributed. The same might be said to be true for the variable 'non-majoritarian institutions', given that federalism, a second chamber and strong and active constitutional courts are highly collinear with the reformed Protestantism variable. The degree to which 'Protestantism' explains away the variation previously ascribed to 'Christian Democracy' or federalism remains an open question. Yet, in my view, one conclusion clearly follows from the above analysis: The literature's neglect of the impact of Protestantism on modern welfare state development cannot be defended.




Social Protestantism as a neglected variable in the comparative literature on the welfare state


The preceding sections have presented empirical evidence supporting the claim that 'religion' accounts for a substantial part of the institutional variation between Western welfare states. Fierce church/state conflicts over questions of responsibility in education, charity and welfare provision, the influence of Catholic and Protestant parties or, respectively, the absence of any opposition against the state taking over responsibility in the new domain of social protection combined to shape the institutional design and the program features of Western welfare states in crucial respects. However, insofar as previous studies have attempted to account for the religious factor, they have focused exclusively on the critical role of Catholicism, Catholic social doctrine and Christian Democracy (Wilensky 1981; Esping-Andersen 1990; Boswell 1993; Castles 1994; Kersbergen 1995). Here, I have argued that it is also necessary to analyze the impact of 'social Protestantism' on Western welfare states. I hold that this impact has been substantial both with regard to the delayed 'takeoff' of the welfare state in countries in which reformed Protestantism was of some importance, as well as with respect to the further institutional development of these 'belated' welfare states. Common features, such as the dominance of the fully funded method of pension financing and the dualism between basic flat-rate but universal entitlements in pension insurance combined with private arrangements point to the common roots of the 'belated Protestant' welfare states.


Why, then, has this Protestant heritage been largely ignored in the literature up to now? In my view, two causes in particular need to be mentioned in this respect:


First, a frequent, misleading assumption in the literature is that each plea for a strict separation between state and society automatically indicates a secularized neoliberal orientation. But not each and every instance in which the central state abstains from intervening into society should be interpreted as a hailing to free-market principles. Not Adam Smith or John St. Mill – but Calvin and Zwingli were the central intellectual reference points in 'liberal' Switzerland.[15] Closely related to this is the widespread perception that the "dominant pattern" in the relationship between political parties and religion is the "tight association between religion and the Right" (Berger 1982: 1). This idea ignores the Protestant roots of radical liberalism. Andre Gould identified three core elements of historical liberalism: the fight for parliamentary democracy, the desire to create and expand markets, and opposition against the states' administration of religious affairs (Gould 1999: 3-4; cf. Madeley 1982). It is therefore false to equate liberalism only with the second, the economic goal. To restrict the role of the state to providing society with only the "pure human necessities" (Troeltsch 1912 [1994]: 954, freely translated) was a doctrine that had primarily religious, not economic grounds in the countries of reformed Protestantism. Charity and caring, in contrast, were seen to fall under the responsibility of the parish or congregation, whereas the individual was responsible to avoid dependence on welfare in the first place. Individual thrift, self-discipline and the fight against idleness and drinking were the central elements in this 'liberal' program – with often not-so-liberal moral overtones and a high degree of expected conformism. Of course, we can label this model as 'liberal', yet we should keep in mind that this kind of liberalism (with its often harsh moral consequences) was and is deeply rooted in reformed Protestantism. This, at least, can explain the specific geographic pattern of dispersion for this variation of liberalism.

Secondly, another misleading assumption is that Protestantism could not have had a discernible influence on the development of the welfare state given that it was dogmatically fragmented and organizationally weak, and – in contrast to political Catholicism – formed political parties only under exceptional circumstances (Switzerland, the Netherlands, Norway; New Zealand).[16] It is true that reformed Protestantism had an anti-étatist and therefore an anti-welfare state program, yet this must not lead us to the conclusion that we can afford to neglect the role of Protestantism in modern welfare state development. Such a judgment is often based on a comparison between Catholicism and Protestantism that fails to account for the profound differences within Protestantism (cf. Castles 1994). Yet, if we want to investigate the role of religion in welfare state development, it is critical to differentiate between the Lutheran and reformed currents of Protestantism.


What does all this mean for Esping-Andersen's welfare state typology? The work of Stein Rokkan has shown that one centrally important cleavage in Europe is of a religious character (Flora et al. 1999). This cleavage dimension should be taken seriously, and should not be treated as a kind of side aspect ('Nebenwiderspruch') of the labor/capital conflict. I believe that it is false to explain modern welfare state development exclusively in terms of workers' mobilization and to account for the role of religion only insofar as workers also organized along denominational lines – i.e. where political Catholicism successfully mobilized Catholic workers to vote for Christian Democratic parties and to join Christian unions. As Jens Alber has already argued, the religious cleavage dimension is central to the understanding of those social programs that do not directly touch upon the labor/capital conflict, such as caring for children or the elderly (Alber 1995). As I have argued in this paper, there is ample evidence that the religious cleavage dimension was also critical to the way that the capitalist conflict was pacified by means of the welfare state. Especially if we account for the crucial differences between Lutheran and reformed Protestantism, we can solve many of the empirical and theoretical problems in Esping-Andersen's approach. In particular, we can better distinguish between the Southern and the Continental, as well as between the conservative and the liberal welfare state regimes.





Alber, Jens, 1998: Recent developments in continental European welfare states: Do Austria, Germany, and the Netherlands prove to be birds of a feather? Contribution to the 14th World Congress of Sociology, Montreal, 29 July 1998.

Alber, Jens, 2002: Die Modernisierung des Wohlfahrtsstaates: Eine Neumischung moderner und traditionaler Elemente? In: Peter Flora/Wolfgang Glatzer/Roland Habich/Karl Ulrich Mayer (Hg.): Sozialer Wandel und gesellschaftliche Dauerbeobachtung. Festschrift für Wolfgang Zapf. 2002, Leverkusen: Leske + Budrich, 59-111.

Baldwin, Peter, 1997: State and Citizenship in the Age of Globalization. In: Peter Koslowski (Ed.): Restructuring the Welfare State. Berlin: Springer, 95-118.

Barrett, David B. (Ed.), 1982: World Christian Encyclopedia. A comparative Study of Churches and religions in the modern World 1900-2000. Nairobi: Oxford University Press.

Bartolini, Stefano, 2000: The Political Mobilization of the European Left, 1860-1980. Cambridge, UK: Cambridge University Press.

Beck, Nathaniel/Jonathan N. Katz, 1999: What to do (and not to do) with Time-Series-Cross-Section Data in Comparative Politics. In: American Political Science Review 89, 634-647.

Bieber, Ulrich/Karin Henzel, 1999: Die Niederlande. In: Verband Deutscher Rentenversicherungsträger (Hg.): Rentenversicherung im internationalen Vergleich. Frankfurt a.M.: WDV Wirtschaftsdienst, 131-157.

Blückert, Kjell, 2000: The State as Nation. A Study in Ecclesiology and Nationhood. Frankfurt a.M./Bern/Berlin: Peter Lang.

Boswell, Jonathan, 1993: Catholicism, Christian Democrats and 'reformed capitalism'. In: Colin Crouch/David Marquand (Eds.): Ethics and markets: Cooperation and competition within capitalist economies. Oxford: Blackwell, 48-65. 

Castles, Francis G., 1985: The Working Class and Welfare. Reflections on the Political Development of the Welfare State in Australia and New Zealand 1890-1980. Wellington: Allen & Unwin.

Castles, Francis G., 1994: On Religion and Public Policy: Does Catholicism make a Difference? In: European Journal of Political Research 25, 19-40.

Castles, Francis G., 1996: Need-based Strategies of Social Protection in Australia and New Zealand. In: Gøsta Esping-Andersen (Ed.): Welfare States in Transition. National Adaptions in Global Economies. London: Sage, 88-115. 

Castles, Francis G./Deborah Mitchell, 1993: Worlds of Welfare and Families of Nations. In: Francis F. Castles (Ed.): Families of Nations. Patterns of Public Policy in Western Democracies. Aldershot: Dartmouth, 93-128.

Catterall, Peter, 1993: Morality and Politics: The Free Churches and the Labour Party between the Wars. In: The Historical Journal 36, 667-685.

Clark, Norman H., 1976: Deliver us from Evil. An Interpretation of American Prohibition. New York: Norton & Company.

Cox, Robert H., 1993: The Development of the Dutch Welfare State. From Workers' Insurance to Universal Entitlement. Pittsburgh and London: The University of Pittsburgh Press.

Daalder, Hans, 1966: The Netherlands. Opposition in a segmented society. In: Robert A. Dahl (Ed.): Political Opposition in Western Democracies. New Haven: Yale University Press, 188-236.

Disney, Richard, 2000: Crisis in Public Pension Programmes in OECD: What are the Reform Options? In: The Economic Journal 110, 1-23.

Esping-Andersen, Gøsta, 1990: Three Worlds of Welfare Capitalism. New York: Polity Press.

Esping-Andersen, Gøsta, 1996: Welfare states without work: The impasse of labor shedding and familialism in Continental Europe. In: Gøsta Esping-Andersen (Ed.): Welfare States in Transition. National Adaptions in Global Economies. London: Sage, 66-87. 

Esping-Andersen, Gøsta, 1999: Social Foundations of Postindustrial Economies. New York: Oxford University Press.

Evers, Adalbert, Jean-Louis Laville (Eds.), 2004: The third sector in Europe. Cheltenham: Elgar.

Ferrera, Maurizio, 1996: The Southern Model of Welfare in social Europe. In: Journal of European Social Policy 6, 179-189.

Ferrara, Maurizio, 1997: Introduction général. In: MIRE (Ed.): Comparer les systèmes de protection sociale en Europe du Sud. Paris, 15-26. 

Fix, Birgit, 2001: Religion und Familienpolitik. Deutschland, Belgien, Österreich und die Niederlande im Vergleich. Opladen: Westdeutscher Verlag. 

Flora, Peter, 1983: State, Economy, and Society in Western Europe 1815-1975. Volume I: The Growth of Mass Democracies and Welfare States. Frankfurt a.M.: Campus/London: MacMillan/Chicago: St. James Press. 

Flora, Peter, 1983a: State, Economy, and Society in Western Europe 1815-1975. Volume II: The Growth of Industrial Societies and Capitalist Economies. Frankfurt a.M. Campus/London: MacMillan/Chicago: St. James Press. 

Flora, Peter, 1986: Introduction. In: Peter Flora (Ed.), Growth to Limits. The Western European Welfare States since World War II. Volume 1: Sweden, Norway, Finland, Denmark. Berlin/New York: de Gruyter,  XII-XXXVI. 

Flora, Peter/Jens Alber, 1981: Modernization, Democratization, and the Development of Welfare States in Western Europe. In: Peter Flora/Arnold J. Heidenheimer (Eds.): The Development of Welfare States in Europe and America. New Brunswick: Transaction Books, 37-80 

Flora, Peter/Stein Kuhnle/David Urwin (Eds.), 1999: State Formation, Nation Building and Mass Politics in Europe: The Theory of Stein Rokkan. Oxford: Oxford University Press. 

Freeden, Michael, 1978: The New Liberalism. An Ideology of Social Reform. Oxford: Claredon Press. 

Goodin, Robert E./Bruce Headey/Ruud Muffels/Henk-Jan Dirven 1999: The real worlds of welfare capitalism. Cambridge: Cambridge University Press.

Gould, Andre C., 1999: Origins of Liberal Dominance: State, Church and Party in Nineteenth-Century Europe. Ann Arbor: Michigan University Press.

Gruner, Erich, 1977: Die Parteien in der Schweiz. Bern: Francke.

Heidenheimer, Arnold J., 1983: Secularization Patterns and the westward Spread of the Welfare State, 1883-1983. Two Dialogues about how and why Britain, the Netherlands, and the United States have differed. In: Richard F. Tomasson (Ed.): The Welfare State 1883-1983 (Comparative Social Research, Vol. 6). London: JAI Press, 3-38.

Hinrichs, Karl, 2000: Elephants on the Move. Patterns of Public Pension Reform in OECD Countries. In: European Review 8, 353-378.

Huber, Evelyne/Charles Ragin/John D. Stephens, 1993: Social Democracy, Christian Democracy, Constitutional Structure and the Welfare State, American Journal of Sociology 99, 711-749.

Huber, Evelyne/Charles Ragin/John D. Stephens, 1997: Comparative Welfare States Data Set. Northwestern University and University of South Carolina.

Huber, Evelyne/John D. Stephens, 2001: Welfare state and production regimes in the era of retrenchment. In: Paul Pierson (Ed.): The New Politics of the Welfare State. Oxford University Press, 107-145.

ILO [International Labour Office], 1954: The Costs of Social Security. Genf: ILO. 

Inglehart, Ronald/Pippa Norris, 2004: Sacred and Secular. Religion and Politics Worldwide. New York: Cambridge University Press.

Iversen, Torben/Anne Wren 1998: Equality, Employment, and Budgetary Restraint: The Trilemma of the Service Economy. In: World Politics 50, 507-546.

Kahl, Sigrun, 2004: The history of poor relief in Catholic, Lutheran and reformed Protestant countries: some observations and hypotheses. Paper presented at the workshop 'Religion and the welfare state', 30 April-1 May, Max-Planck Institute for the Study of Societies, Cologne.

 Kalyvas, Stathis N., 1996: The Rise of Christian Democracy in Europe. Ithaca, NY: Cornell Univ. Press.

Kalyvas, Stathis N., 1998: From Pulpit to Party: Party Formation and the Christian Democratic Phenomenon. In: Comparative Politics, 293-312.

Kaufmann, Franz-Xaver, 1988: Christentum und Wohlfahrtsstaat. In: Zeitschrift für Sozialreform, 65-89.

Kersbergen, Kees van, 1995: Social Capitalism. A Study of Christian Democracy and the Welfare State. London: Routledge.

Korpi, Walter, 1983: The Democratic Class Struggle. London: Routledge & Keegan.

Kossmann, E.H., 1978: The Low Countries, 1780-1940. Oxford: Claredon.

Langner, Albrecht, 1998 Katholische und evangelische Sozialethik im 19. und 20. Jahrhundert. Paderborn: Schöningh.

Luebbert, Gregory, 1991: Liberalism, Fascism and Social Democracy. Social Classes and the Political Origins of Regimes in Interwar Europe. New York: Oxford University Press.

MacLeod, Hugh, 1984: Religion and the Working Class in Nineteenth-Century Britain. London: Macmillan.

Maddison, Agnus, 1995: Monitoring the World Economy. Paris: OECD.

Manow, Philip, 2000: 'Modell Deutschland' as an interdenominational Compromise. CES working paper 01/2000. Minda de Gunzburg Center for European Studies, Harvard University. Cambridge, MA.

Manow, Philip, 2001: Ordoliberalismus als ökonomische Ordnungstheologie. In: Leviathan 29, 179-198.

Manow, Philip, 2002: 'The Good, the Bad, and the Ugly': Esping-Andersens Sozialstaats-Typologie und die konfessionellen Wurzeln des westlichen Wohlfahrtsstaats. In: Kölner Zeitschrift für Soziologie und Sozialpsychologie 54(2), 203-225.

Martin, David, 1978: A General Theory of Secularization. Oxford: Blackwell.

Martin, David, 1978a: The religious condition of Europe. In: Salvador Giner/Margaret Scotford Archer (Eds.): Contemporary Europe. Social Structures and cultural Patterns. London etc: Routledge & Kegan, 228-287.

Matthews, Race, 1993: Australia’s first Fabians. Middle-Class Radicals, Labour Activists and the Early Labour Movement. Cambridge: Cambridge University Press.

Matsaganis, Manos/Maurizio Ferrera/Luís Capucha/Luis Moreno, 2003: Mending Nets in the South: Anti-poverty Policies in Greece, Italy, Portugal and Spain. In: Social Policy & Administration 37, 639-655. 

Milward, Alan, 1992: The European Rescue of the Nation State. London: Routledge. 

MIRE [Mission de Recherche et Expérimentation] (Ed.), 1997: Comparer les systèmes de protection sociale en Europe du Sud. Paris. 

Mol, Hans (Ed.), 1972: Western Religion. A country-by-country sociological inquiry. Den Haag/Paris: Mouton.

Morgan, Kimberly J., 2002: Forging the Frontiers between State, Church and Family: Religious Cleavages and the Origins of Early Childhood Education in France, Sweden, and Germany. In: Politics & Society 30(1), 113-148.

Morgan, Kimberly J., 2003: The Politics of Mothers' Employment. In: World Politics 55, 259-289.

Morgan, Kimberley, 2004: The religious origins of the gendered welfare state. The example of work-family policies in Western Europe. Paper presented at the workshop 'Religion and the welfare state', 30 April-1 May, Max-Planck Institute for the Study of Societies, Cologne.

Narud, Hanne Marthe/Kaare Strøm, 2000: Norway: A fragile coalitional Order. In: Kaare Strøm/Wolfgang C. Müller (Eds.), Coalition Governments in Western Europe. Oxford: Oxford University Press, 158-191. 

Obinger, Herbert, 1998: Politische Institutionen und Sozialpolitik in der Schweiz. Frankfurt a.M.: Peter Lang. 

Obinger, Herbert/Uwe Wagschal, 1998: Drei Welten des Wohlfahrtsstaates? Das Stratifizierungskonzept in der clusteranalytischen Überprüfung. In: Stephan Lessenich/Ilona Ostner (Hg.): Welten des Wohlfahrtskapitalismus. Der Sozialstaat in vergleichender Perspektive. Frankfurt a.M.: Campus, 109-135. 

Obinger, Herbert/Uwe Wagschal (Hg.), 2000: Der gezügelte Wohlfahrtsstaat: Sozialpolitik in reichen Industrienationen. Frankfurt a.M.: Campus. 

Parry, Jonathan P., 1986: Democracy and Religion: Gladstone and the Liberal Party 1867-1875. Cambridge: Cambridge University Press. 

Pelling, Henry, 1965: Origins of the Labour Party. Oxford: Clarendon Press. 

Pierson, Paul, 1999: Not just What, but When: Timing and Sequence in Political Processes. Forthcoming in Studies in American Development. Mimeo, Harvard/Cambridge, MA. 

Plümper, Thomas/Vera Troeger/Philip Manow, 2004: Panel Data Analysis in the Comparative Political Economy of the Welfare State: A Note on Methodology and Theory. European Journal for Political Research, forthcoming. 

Powell, David, 1992: British Politics and the Labour Question, 1868-1990. London: MacMillan. 

Richardson, Len, 1992: Parties and Political Change. In: Geoffrey W. Rice (Ed.): The Oxford History of New Zealand. Auckland: Oxford University Press, 201-229. 

Rokkan, Stein, 1967: Geography, Religion, and Social Class: Crosscutting Cleavages in Norwegina Politics. In: Seymour Martin Lipset/Stein Rokkan (Eds.), Party Systems and Voter Alignements: Cross-National Perspectives. New York: Free Press, 367-444. 

Scharpf, Fritz W., 2001 [1997]: Employment and the welfare state. A continental dilemma. In: Bernhard Ebbinghaus/Philip Manow (Eds.): Comparing Welfare Capitalism. Social Policy and political Economy in Europe, Japan, and the USA. London: Routledge, 270-283.

Scharpf, Fritz W./Vivien A. Schmidt, 2001: Introduction. In: Fritz W. Scharpf/Vivien A. Schmidt (Eds.): Welfare and Work in the open Economy. Vol. II: Diverse Responses to common Challenges. New York: Oxford University Press, 1-18. 

Schmidt, Manfred G., 1998: Sozialpolitik in Deutschland. Historische Entwicklung und internationaler Vergleich. Opladen: Leske + Budrich. (2. Aufl.). 

Sellin, Volker, 1971: Die Anfänge staatlicher Sozialreform im liberalen Italien. Stuttgart: Klett. 

Skocpol, Theda, 1992: Protecting soldiers and mothers: the political origins of social policy in the United States: Cambridge, MA: Belknap Press. 

Skocpol, Theda, 2000: Religion, Civil Society, and Social Provision in the U.S. In: Mary Jo Bane/Brent Coffin/Ronald Thiemann (Eds.): Who will provide? Boulder, Col.: Westview, 21-50. 

Spann, Edward K., 1989: Brotherly tomorrows: movements for a cooperative society in America 1820-1920. New York: Columbia University Press. 

Stephens, John D., 1979: The Transition from Capitalism to Socialism. Urbana and Chicago: University of Illinois Press. 

Stephens, John D., 1979a: Religion and Politics in Three Northwest European Democracies. In: Comparative Social Research 2, 129-157. 

Swaan, Abram de, 1988: In Care of the State: Health Care, Education, and Welfare in Europe and the USA in the Modern Era. Cambridge: Polity Press. 

Thomas, John L., 1983: Alternative America. Henry George, Edward Bellamy, Henry Demarest Lloyd and the adversary tradition. Cambridge, MA: Harvard University Press.

 Troeltsch, Ernst, 1912 [1994]: Die Soziallehren der christlichen Kirchen und Gruppen. Tübingen: Mohr (Siebeck).

 Visser, Jelle/Anton Hemerijck, 1997: The Dutch Miracle. Amsterdam: Amsterdam University Press.

 Wilensky, Harold L., 1981: Leftism, Catholicism, and Democratic Corporatism: The Role of Political Parties in Recent Welfare State Development. In: Peter Flora/Arnold J. Heidenheimer (Eds.): The Development of Welfare States in Europe and America. New Brunswick: Transaction Books, 345-382.






Comments by Jens Alber, Sabina Avdagic, Thomas Ertman, Sven Jochem, Franz-Xaver Kaufmann, Kees van Kersbergen, Bernhard Kittel, Herbert Obinger, Achim Schmid and Christine Trampusch are gratefully acknowledged. Thomas Plümper and Vera Troeger assisted me in all questions of data analysis. For those who think that Sergio Leone is a small state in West Africa: My title refers to the quintessential spaghetti western from 1966, starring Clint Eastwood and directed by Sergio Leone.


Anyone who thinks that this is caricature is too extreme should take a look at the caricature that the original work offers. Whereas the Social Democratic regime transforms "market dependence" into "individual independence", the conservative regime simply substitutes one dependency for the other: It trades market dependency for "dependence on family, morality, or authority" (Esping-Andersen 1990: 47).


Most of these problems could be solved because we can now avoid discussing whether country x belongs to regime type y or z. Instead, denominational affiliations can be treated as "variables, which may be present to a greater or lesser degree in any case" (Goodin et al. 1999: 13, Fn. 35; emphasis mine).


The Dutch welfare state is, to my knowledge, the only welfare state in which a person can be exempt from compulsory insurance if he or she thinks that the notion of social risk is in contradiction with the notion of God’s absolute providence (Bieber/Henzel 1998: 138). The strong emphasis on God’s providence has, of course, always been a distinguishing feature of Calvinism.


The Scandinavian countries are countries "largely without Protestant dissent" (Martin 1978a: 237). Norway is the exception, although the Norwegian reformed sects remain small. However, dissent became politically influential with the formation of the Protestant Christian People’s Party in 1933 (Kristelig Folkeparti, KRF; see Narud/Strøm 2000: 159-160). It is interesting to note that among the Scandinavian countries, Norway was definitely a welfare state late-comer (see ILO, various years).


In 1910, the German population was to 61.2% Protestant, and 36.7% Catholic (in 1925, 64.1 to 32.4%; after the Second World War the ratio changed to 50.5 to 45.8% due to the loss of the Eastern states).


In accordance with the literature, I take the introduction of the first major compulsory social programs as an indicator, not including state funding of voluntary private programs (cf. Schmidt 1998: 180).


The Belgian case points to one potential measurement problem: The GDP per-capita measures wealth, but not necessarily the level of industrialization. In both the Dutch and the Belgian cases, one might argue that the countries’ relative wealth at the turn of the century was due rather to trade than to industrialization. (I am grateful to Thomas Ertman for pointing this out to me.) Yet, industrial employment was already very high in Belgium, Switzerland and the UK in 1890 and 1900 (considerably higher than industrial employment in Germany, for instance). Dutch industrial employment in 1900 lagged somewhat behind, but was still higher than in Austria, Denmark, Finland, France, Italy, Sweden or Norway (Flora 1983a: 449-597).


One objection to my argument might be that the USA, Canada, Australia and New Zealand were, and still are, immigration countries, and that a clear trade-off exists between political participatory rights and social entitlements: "The simple rule governing the interaction of social benefits and political citizenship is that the wider the scope of welfare, the less encompassing social citizenship can be; the more open membership in the political community, the more restricted access to social measures must be" (Baldwin 1997: 111). Hence, the common laggard status of these welfare states might be due to immigration rather than to Protestantism. Yet, to the extent that welfare state development (and welfare state structures) is (are) similar in some European countries and in the New World, this should point to a common Protestant background, given that Australia and New Zealand were much more successful in the restriction of immigration, and that the UK, the Netherlands and Switzerland were emigration – not immigration – countries. However, successful regulation of immigration was certainly a decisive background condition for the development of the 'radical' antipodean welfare state variation in Australia and New Zealand, with minimum wage legislation, compulsory arbitration and a protectionist consensus (Castles 1985: 56-60, 1996).


Swiss industrialization was concentrated in the small-enterprise semi-artisanal sector (printing and clocks, later machine tools): "Rural property ownership was especially widespread and most industry was found adjacent to hydroelectric power in rural valleys" (Luebbert 1991: 48). Compared to France, for instance, Switzerland was much less urbanized, yet much more industrialized at the turn of the century.


The result does not change substantially if we classify France as a Continental-conservative welfare state.


Today, there is not much justification in categorizing Dutch company pensions as public pension expenditures and Swiss company pensions as private (see Esping-Andersen 1990: 81, 87). The Dutch Labor Minister has been able to declare membership in a sector fund to be obligatory for a company since 1949, whereas his or her Swiss counterpart has had this right since 1982.


I have used the welfare state data set of Huber/Ragin/Stephens (1997). To account for serial autocorrelation in the residuals, I applied an error-correction model (accounting for AR (1) dynamics in the errors; cf. the r-value in Table 8). I calculated panel-corrected standard errors to account for heteroscedasticity in the errors (Beck/Katz 1995; a more detailed justification for the method applied is provided in Plümper/Troeger/Manow 2004). I tested for structural breaks, but the coefficients for Protestantism did not change much across the periods of expansion (»1960-1973), stagnation (»1974-1981), and contraction (»1981-1990). The same is true for the coefficients of the Left and Christian Democratic variables. In other words, we are justified in estimating one coefficient for the entire panel period 1962-1989.


Since we have no complete time series reaching back to the 1950s, we cannot extend our analysis any further backwards. Yet all we do know about the expenditure profiles of our country group (Switzerland, the Netherlands, Great Britain, the USA, Canada, New Zealand and Australia) between 1950 and 1960 is that expenditures were clearly below the OECD-average. On average, these countries spent 6.9% of their GDP as social expenditures in 1950 compared to 8.7% in the rest of the OECD (ILO 1954).


Of course, Scottish (market) liberalism itself had Protestant roots – and was influential for religious mobilization in other countries. For instance, the Calvinist "Anti-School Law League" in the Netherlands was consciously designed after the British "Anti-Corn Law League" – itself a movement "with much dissenting support" (Martin 1978: 187; Daalder 1966: 200).


In 1917, the Protestant Political Association was founded in New Zealand and had 200,000 members by 1919 (Richardson 1992: 218-219). The association’s program emphasized the fight against "rum, Romanism and rebellion" (ibid.: 218). See also the example of the Norwegian Kristelig Folkeparti. We must also mention the Evangelische Volkspartei in Switzerland, which ran in 19 Bundesrat elections between 1919 and 1987. Of course, the Swiss Freisinn has always been strongly influenced by reformed Protestantism as well. Kalyvas’s claim that Protestant parties were only founded in the Netherlands is simply wrong (1996: 3, Fn. 6).

Copyright © 2004 Philip Manow

No part of this publication may be reproduced or transmitted without permission in writing from the author.
Jegliche Vervielfältigung und Verbreitung, auch auszugsweise, bedarf der Zustimmung des Autors.

MPI für Gesellschaftsforschung, Paulstr. 3, 50676 Köln, Germany





MPIfG:  MPIfG Working Paper 04/3

[Zuletzt geändert am 04.04.2014 12:07]