Resumé

The role of the Social Democrats in the establishment of the Scandinavian welfare state has been challenged in recent years. Institutional legacies have conditioned post-war Social Democratic and the bourgeois parties have played a larger role than so far acknowledged. By exploring the origin of five core policies of the early Danish welfare state, it is shown that policy legacy theses cannot account for the pattern of policy organization. Focusing on party and class dynamics, it is demonstrated that the bourgeois parties, and in particular the Liberals, had a crucial influence on the choice of funding model and administrative structure. The distributional and administrative interests of core constituencies shaped the preferences of the bourgeois parties decisively.

Introduction

The dominant perspective on the formation of Scandinavian welfare states stresses the role played by Social Democrats from the 1930s onwards (Korpi 1978; Esping-Andersen & Korpi 1984; Esping-Andersen 1985, 1990; Stephens 1986). In recent years, this interpretation has been challenged in two interrelated ways. First, there is an increasing focus on the plausible role played by bourgeois parties and constituencies in the formation and shaping of welfare states in general (Kuhnle 1983; Baldwin 1990; Kangas 1991; Swenson 1991; Huber et al. 1993; Olsson 1993; Hicks et al. 1995; Martin 1995; cf. Wilensky 1984). Second, the institutional turn in political science has animated scholarly interest in the institutional paths of contemporary welfare states as well as in the origin of these decisive legacies (Heclo 1974; Swenson 1991; Rothstein 1992a; Skocpol 1992; Knudsen & Rothstein 1994; Orren & Skowronek 1994). The character of early welfare state consolidation shapes subsequent welfare policy achievements (Hicks et al. 1995,331).

---

* Asbjørn Sonne Nørgaard, Department of Political Science, University of Aarhus. Universitetsparken, DK-8000 Aarhus C, Denmark.

The purpose of this article is to explain the institutional structure of the early Danish welfare state, i.e. before 1920. Empirically, I shall approach the usual core social security programs - old age, sickness, accident, unemployment and occupational safety and health regulation. The argument is that the bourgeois parties had a crucial influence on the shaping of the early welfare state. My focus is on policy organization; i.e. by whom and how the policy was funded and administered, and in the latter case in particular whether corporatism or a pure public administrative structure ensued. Thus, corporatism is a contingent policy choice; a choice made by the bourgeois parties.

Corporatism, as conceptualized here, is a sectoral phenomenon (cf. Lehmbruch 1984). It is an institutional structure that, by way of legislatively power, privileges certain organized interests in the making and the implementation of policy. Effectively, corporatist administration implies significant policy-making power because of broad delegation, and because later statutory changes frequently are drafted by the corporatist bodies. Corporatism may foster a cooperative policy-making style, but it does not do so by definition. Corporatist administration is widespread in social insurance programs but is also found in other types of welfare policy.

There are several reasons to focus on the institutional aspects of early welfare policy. First, whereas the specific content of welfare policies may have changed considerably since their origin, the organization seems quite robust (Nørgaard 1997; cf. Weir & Skocpol 1985; North 1990; Orren & Skowronek 1994). In the five cases studied here, only sickness insurance has shifted to become a purely tax-financed public program. The development in the other four areas has been path dependent; those that were based on contributions have remained so (although taxes finance much more today); those with a corporatist governing structure have kept it. The early institutional set the parameters for future options.

Second, institutional innovations in one policy area may be copied in other areas at a later date (Heclo 1974; DiMaggio & Powell 1991; Nørgaard 1999). Irrespective of its exact definition, corporatism is a dominant institutional of contemporary Scandinavian welfare states (Cameron 1984; Lehmbruch 1984; Lijphart & Crepaz 1991). Today, privileged organized interests partake in the making and management of a broad range of policies, not least social and labor market policies (Johansen & Kristensen 1982; Rothstein 1992b; Blom-Hansen & Daugbjerg 1999; Christiansen & Rommertvedt 1999; Christiansen & Sidenius 1999; Nørgaard 1999).

Third, the institutional structure of past policies shapes politics in the present (Kangas 1991; Rothstein 1992a; Thelen & Steinmo 1992; Pierson 1993, 1996). By empowering some actors at the expense of others, today's politics is shaped by past policy choices. Several of the organizations

privileged by the early welfare state have been powerful political actors in
post-war Danish political history.

The argument to be unfolded below can be stated very briefly. The organization of the early Danish welfare state in matters of old-age assistance sickness insurance (1892), accident insurance (1898), occupational and health regulation (1901), and unemployment insurance (1907) was not modeled after Bismarck. Nor can it in any simple way be explained by previous policy legacies. And the influence of the Social Democrats been scattered at best. When the result was corporatism and when it was not, the organization of early Danish welfare policy was by and large the making of the Liberals, observing farmer interests, and the Right, trying to strike a balance between farmer and business interests. The chosen funding models were also strongly influenced by the bourgeois parties.

The following section discusses dominant theories on the origin of Scandinavian welfare states and presents three theses. In the next section, the decision-making processes in the five early welfare policies are analyzed. The conclusion summarizes the findings and discusses some of the theoretical of the study.

Theories on the Origin of Scandinavian Welfare States

The debate on the origin of the welfare state has not focused specifically on administrative organization. Policy content and the consequences of different and benefit structures have been the focus for most theorizing (e.g. Esping-Andersen & Korpi 1984; Stephens 1986; Esping-Andersen 1990; Kangas 1991). The organization of policy in terms of administrative structure, and the causes of alternative funding structures have received less attention (cf., however, Skowronek 1982; Kuhnle 1983; 1984; Weir & Skocpol 1985; Levi 1988; Baldwin 1990; Steinmo 1993). However, the preferences a certain benefit structure also depend on the financing of benefits. Besides, often the choice to support a given policy is contingent upon how and by whom policy is managed, because control over implementation is as important for the policy consequences as the minute details of statutory language (Moe 1989; Rothstein 1992b; Nørgaard 1997). In fact, this is a fairly trivial observation in the literature on implementation (Pressman & Wildavsky 1973; Lipsky 1980).

Still, if administrative structure and funding are critical policy decisions, it seems reasonable to explore whether dominant theories of the origin of the welfare state and corporatism can account for the outcome in the five areas studied here.

Some scholars hold that an emulation of the organization of welfare
policy in northern Europe took place after Bismarck's social security

initiatives in the 1880s (Flora & Alber 1984, 60; Kuhnle 1984, 126-31; Petersen 1985, 29-46; cf. Briggs 1961; Collier & Messick 1975). This diffusion thesis, or argument on transnational learning, does not fare well in the Danish case. As early as 20 years ago Stein Kuhnle showed that the essential characteristic of Bismarck's social insurances - compulsion - was absent in most Scandinavian programs, notably in Denmark (Kuhnle 1983; 1984). Besides, core social insurance programs had been considered in Denmark and Sweden years before the Bismarck initiatives.

Kuhnle did not focus on administrative organization per se. Bismarck's inclination to let corporatist bodies manage and supervise the schemes may have been copied, even if policy content differed in other respects. As it was, Denmark did establish corporatist administrative structures some years after Bismarck. However, corporatism had been discussed in details in a Danish governmental report at least 5 to 10 years before Bismarck's initiatives, and there is no mention of German sources of inspiration (Workers Commission 1878). A corporatist organization of policy was suggested in several areas of social and labor market policy (ibid., vi-vii). When Danish welfare programs came to include corporatism, therefore, it did not seem to be in imitation of the German model.

Thus, endogenous rather than exogenous explanations of welfare policy seem more promising in the Danish case. However, there are several competing endogenous explanations of welfare-state origin as well as of corporatism.

The Legacy Theses: Macro or Policy Legacies

Policy legacy theses draw attention to the fact that political conflict and choices make sense only in historical context. No policy is conceived, let alone organized, de novo. Some talk of 'institutional and societal learning' as the critical determinant of policy organization (Heclo 1974, 319; cf. 304-19; Petersen 1985; cf. also Kuhnle 1996). This version of the thesis has an apolitical tinge. However, the argument can be politically underpinned by stressing that privileged state actors have 'organizational and career interests of their own' (Weir & Skocpol 1985, 118). If so, it must be demonstrated those interests are powerful enough to exercise a decisive influence on policy organization. The causal argument has to go through the elected officials who have the power to reject any proposed solution (cf. Kangas 1991).

Leaving aside the most apolitical arguments on myopic learning, the policy legacy theses usually hold that the choices that policy makers face are biased by legacies and tradition (Weir & Skocpol 1985; North 1990). In general, the focus is on two kind of legacies which operate at different levels: macro legacies and legacies found in one specific policy area.

The first argues that new ways of organizing welfare policy are biased by macro legacies. Whether one refers to 'state structures and capacities' (Weir & Skocpol 1985), political institutions such as 'patriarchal states' and 'unitary democracies' (Hicks et al. 1995), 'tradition' (Heckscher 1984), or 'ideological setting' (Kuhnle 1984; cf. 1983), there is much to be said for this argument. Few would argue that all institutions are equally feasible in all settings.

In a similar vein, some theories of the origin of corporatism also stress the impact of macro legacies. Leaving aside the conceptual ambiguities (the privileged inclusion of organized interests in policy making is at least one common denominator in definitions of corporatism), numerous theses have been advocated. The strength of 'consociational' or consensual democracy been stressed by some (Lehmbruch 1979; Lijphart & Crepaz 1991). Others have featured Catholicism (Esping-Andersen & Korpi 1984; Wilensky 1984; Therborn 1989), the strength and centralization of unions and/or labor parties (Cameron 1984; Maier 1984; Katzenstein 1985), or attributes of the party system (Katzenstein 1985; cf. Lipset & Rokkan 1967).

A one-country study focusing on one period cannot give a genuine comparative assessment of the impact of macro legacies and contexts. Still, if significant similarities in policy organization are found in the five areas studied, this indicates a strong effect - in the sense, if X then Y - of macro legacies. Since there is wide disagreement about the essence of the decisive macro legacy, it is not possible to specify in advance which 'significant similarity' can be expected. One astute observer of the history of Scandinavian politics stresses the 'tradition of paternalistic interventionism' (Heckscher 1984, 40; cf. Levine 1978; Hicks et al. 1995, 338). Another, studying the same phenomena in the same period, underscores 'the overall strength of liberal ideology' in Denmark (Kuhnle 1984, 135; cf. Sørensen 1978).

Certain macro legacies can also be seen as necessary but not sufficient conditions for policy outcomes. This weaker causal claim - if Y then X - cannot be tested here, since it merely states that a certain policy organization, corporatism, can be found only in some macro settings. One instance of corporatism validates this. Of course, the even weaker probabilistic claims - if X then P(Y) is high - which are usually studied by way of quantified data and correlational evidence for several (country) cases, cannot be tested either (Lijphart & Crepaz 1991; cf. Crepaz 1998). However, both weaker causal arguments leave room for political and actorcentered since they imply that several thinkable and technically feasible alternatives are politically germane (cf. Nørgaard 1997, chapter 2).

In sum, only the strong version of the macro legacy thesis can be tested
here:

Thesis 1: The administrative organization in all the five policy areas studied will either be corporatist or purely public because the policies are passed at approximately the same time within one macro context. The funding model will be the same in the four social security programs.

The second conception of the policy legacy thesis stresses the significance of the institutional heritage in a specific policy area. Most often it is based on the premise that decision makers have limited computational, or cognitive, Thus, they choose from a narrower set of alternatives than those that are feasible and perhaps more attractive had they been pondered (Simon 1976; cf. Nørgaard 1996). Broadly conceived, policy legacies give rise to positive as well as negative feedbacks (Heclo 1974; cf. Kangas 1991), which is a very weak causal claim that reaccentuates the significance of politics and actor dynamics. However, the thesis may also be given a stronger meaning, stating that the organizational development of policy is overwhelmingly path dependent and incremental (North 1990; Pierson 1993, 1996), thus suggesting that the dominant legacy in a policy area tends to be reproduced even when reforms are passed.

This formulation of the policy legacy thesis may rest on the testable assumption alluded to above, namely that already privileged bureaucratic actors will try to persuade and pressure the most plausible political allies in government and parliament to opt for their preferred institutional solution (Kangas 1991; Nørgaard 1997). It may also rest on the idea that the influence policy tradition is so pervasive that nobody genuinely considers alternatives to the prevailing organizational heritage (March & Olsen 1989). Although in the latter case the issue of policy organization is not politicized, both theses yield the following:

Thesis 2: The organization of the five welfare policies will be path dependent.
The administration and, in the four social security programs, the funding
model will build on the dominant legacy in the respective policy areas.

Policy is never constructed out of time and space, as the legacy theses remind us. Still, there may be critical formative moments or junctures in history (Krasner 1984; Rothstein 1992b) when new policy legacies are instituted and the number of embryonic policy paragons to build on is exceptionally abundant. At such times, elected officials can choose among legacies in similar policies in other countries, in other policy areas, in differentrelevant or private initiatives, and in the writings of contemporaryobservers. officials somehow build on the stock of relevant, contemporary experiences and knowledge when designing policy (cf. Olsson, 1993; Kuhnle 1996). However, this does not say much about why

some and not other legacies are imitated. If sufficiently spacious, the policy
legacy thesis is always true ex post facto.

Actor-Centered Theses: The Social Democratic Party Thesis and Beyond

The class - or power resource - thesis emphasizes agency and focuses on party politics and class dynamics when explaining the origin of welfare states (Castles 1978; Korpi 1978; Esping-Andersen & Korpi 1984; Esping- Andersen 1985; 1990; Stephens 1986; Huber et al. 1993). Following the dominant perspective, the institutional contours of Scandinavian welfare states, which include tax-financing and publicly organized schemes, originated the 19305. Allegedly, they were founded by the Social Democratic parties in strategic alliances with parties of the disunited bourgeois block (Esping-Andersen & Korpi 1984, 185).

Paradoxically, it is also argued that the Social Democrats were instrumental securing corporatism (Cameron 1984; Maier 1984; Katzenstein 1985; Lijphart & Crepaz 1991). Focusing on Scandinavia, and in particular on Denmark, it has been argued that the 1930s were the formative period of corporatism. The development of Scandinavian corporatism (denned in the same way as in the present study) is related 'to an early and continuous reach of social democratic governmental power' (Johansen & Kristensen 1982, 194).

Whether Social Democrats advocated public or corporatist administration funding by taxes or a contributory scheme, they cannot plausibly have had pivotal influence on how policy became organized before the 19205: '(T)he virtual absence of leftist governments before the Great Depression' (Hicks et al. 1995, 329) makes it implausible that the SDP, even if pressing vehemently for social reforms (cf. Esping-Andersen 1990, 109-10; Hicks et al. 1995, 330, 344), was allowed a strong influence on how policy was funded and administered. It seems more promising to focus on the bourgeois parties when studying the period 1890 to 1920.

The rediscovery of bourgeois parties and constituencies has been a common theme in recent literature on the welfare state (Baldwin 1990; Kangas 1991; Swenson 1991; Huber et al. 1993; Olsson 1993; Hicks et al. 1995; Martin 1995; Pierson 1996; Green-Pedersen 1999; cf. also Kuhnle 1983, 151-63; Wilensky 1984). Scholars have moved beyond arguments on 'the disunity of the Right' (Castles 1978) or the strategic and pre-emptive measures taken by Bonapartist/Bismarckian governments (cf. Flora & Alber 1984, 70-2; Baldwin 1990, 39-40), and are now inspecting when, why and how bourgeois parties have a positive stake in the welfare state.

In this context Peter Baldwin's open-ended question qui bono may serve
as a point of departure for theorizing the role of the bourgeois parties (1990,
44-54): 'who pays what to whom?' (1990, 48). Although accompanying

industrialization, welfare policies do not only, or even primarily, accommodatethe of urban workers. Depending on how the schemes are organized, old-age assistance and sickness insurance may also benefit selfemployedfarmers, innovations on the funding side may relieve segmentsof tax-paying population of considerable burdens. The point is that constituencies of the bourgeois parties may potentially stand to gain from welfare policy (cf. also Kangas 1991; Huber et al. 1993), and that the scheme preferred by bourgeois parties depends on calculations over the ex ante and ex post distribution of costs and benefits. Thus, sometimes the preferred solution will be a proto-universal, purely tax-financed scheme; at other times it will not.

A similar logic may apply for the administrative aspects of policy. Although this line of reasoning is not pursued systematically by Peter Baldwin, presumably more autonomy and control over policy administration preferred to less (Nørgaard 1997). Following this line of thought, securing core constituents as much unchecked discretion over policy implementation possible will be the party strategy. A certain level of administrative may even be a precondition for party backing.

Baldwin focuses on the interests of risk categories, as defined by the distributional consequences of policy (1990, 48), rather than the interests of class and party constituencies per se. However, his approach begs the question of who is drafting the policy alternatives. In general - and at least in the cases studied here - it is more straightforward and yields the same insight simply to focus on the already constituted dominant social and political actors, i.e. parties and classes. The question then is what policy preferences the dominant political actors have if informed by the quibono logic.

The Qui Bono Thesis and Turn-of-the-Century Danish Party Politics

In Danish politics circa 1900, the dominant political parties were the Right, the Liberals, the Social Democrats, and from 1905/06 the Radical Liberals, who can be ignored in the present context¹ (Lipset & Rokkan 1967; Rokkan 1987; Elklit 1988; Luebbert 1991; cf. Nørgaard 1997, 128-49). From 1890/ 91 to 1910 the agrarian Liberal Party, Venstre, was split into two fractions: the Reformist Liberals and the Moderate Liberals.

Danish welfare politics, at the turn of the century, was embedded in the Systemstruggle. Since 1877 a number of 'provisional laws' had been passed with the backing only of the privileged upper chamber, the Landstinget. However, there were limits to the arbitrariness. All major reforms were passed with the consent of both chambers of parliament. The government of the Right commanded a majority only in the upper chamber. The

Liberals, commanding a majority in the lower chamber, the Folketinget, fought to establish parliamentarism, according to which no government could stay in power without the support of a majority in the lower chamber. The Liberals also controlled most local governments in the rural districts. The Right (together with eight 'Free Conservatives') still had a majority in the Landstinget after the introduction of parliamentarism in 1901.

The ebbs and flows of the Systemstruggle conditioned the possibilities of passing welfare reforms, notably because the Right had a strong incentive to cooperate with the lower chamber and to pass social reforms in order to increase the legitimacy of the 'provisional' governments headed by J.B.S. Estrup. Nevertheless, the struggle did not determine the distributional preferences of the political parties.

The Right was the party in the Folketinget with the most heterogeneous constituency. Originally the party of aristocrats and larger proprietors, it was also the only party representing urban businesses, the upper middle class - notably bureaucrats and the petite bourgeoisie in the urban districts (Dybdahl 1969). Whereas these groups were vastly overrepresented in the parliamentary group, the electoral backing of the Right included a large segment of the working class. In the late 1890s, the Right mustered around 30 percent of the urban working class vote (Dybdahl 1969, 44-5; Elklit 1988, 261, 282). As a share of the Right's total vote, the urban working class made up over 50 percent in many districts. The electoral backing of the Right was halved from 1890 to 1903, but until at least 1910 the relative significance of the urban vote increased .

The constituency interests of the Right were mixed. The distributional interests of the urban districts were most important politically, both because proprietors were willing to pay fairly high property taxes in order to retain their political privileges in the upper chamber, and because the Right increasingly was an urban party. If possible, the Right would observe the interests of urban business before that of the urban working class, but the party could not afford to ignore the demands of the latter altogether. In terms of administrative control, the interests of the Right were best observed by relying on state bureaucrats and/or urban businesses in corporatist bodies.

Although internally split over the issue of how to secure the ousting of the undemocratic Right government, the distributional interests of the Moderate Liberals and the Reformist Liberals were identical. Electorally and organizationally, both wings represented yeoman farmers (Luebbert 1991, 73-80, 133-9). Besides, electoral alliances with the Social Democrats generally implied that the Liberals did not face competition from the left over the votes of the rural working class and smallholders. Until 1909 the two Liberal factions commanded a majority of the seats in the Folketinget (Nørgaard 1997, 135).

The difference between the Moderates and the Reformists centered on strategy. The Reformist hardliners wanted to secure parliamentarism first and social reforms next. The Moderates took the opposite stance. The different on strategy was partially a result of divergent opinions on the need to contain the Social Democrats (Petersen 1985, 84-95). Whereas the internal factionalism paved the way for social reforms, both Liberal factions had the same distributional and administrative interests, namely to privilege the rural economy and independent farmers.

The preferences of the Social Democrats are also unambiguous. As an almost exclusively urban party with a rapidly increasing share of the working class vote, the distributional interests of core constituencies were to disperse the costs of programs benefiting the urban working class to other social groups and the rural population. Because of strong union-party relations (Bryld 1992; Nørgaard 1997), and because union members more often voted for the SDP than the Right, limiting benefits to union members and ensuring union control in the administration was the best organization of policy for the Social Democrats. The interests of organizational and electoral constituencies could then be joined.

Following the qui bono logic, all parties pursued the administrative and distributional interests of core constituencies. Since the bourgeois parties controlled majorities in both chambers, the expectation is that their preferences be strongly reflected in the organization of policy:

Thesis 3: The organization of the five welfare policies will strongly reflect the policy preferences of the Liberals and the Right as informed by their constituency interests. Corporatism as well as pure public administration and the funding model, whether tax based or based on contributions, are of bourgeois origin.

Like political parties in general, the dominant bourgeois parties circa 1900 were not entirely unconstrained in their political maneuvering. At the time, as today, there was a strong norm saying that administrative autonomy and control over a scheme presumed at least co-responsibility for its funding. In contemporary social policy discourse this norm was also informed fortified by the dominant doctrine of 'help to self-help' (Sørensen 1978; Petersen 1985). At the time, as today, politics and policy making were possible only because there were normative limits on the arbitrariness of'powering' (Heclo 1974; Nørgaard 1997).

The Organization of the Early Danish Welfare State

In the following, I examine (1) how the five welfare policies were administeredand
(2) whether the chosen organization reflected an already

dominant policy legacy in the area; and (3) the extent to which the chosen organization reflected the preferences of the bourgeois parties. Furthermore,I whether these preferences are informed by the above mentioned qui bono logic. As far as possible the policy preferences of the burgeoning Social Democratic Party (SDP) will also be analyzed to examine whether the chosen policy organization incorporated SDP demands.

Since old-age assistance has been studied intensely, the analysis of this topic is mainly based on secondary sources. In the four other cases the analysis is based on contemporary public documents, notably commission reports and parliamentary records.

Old-Age Assistance of 1891: Tax Funding, Local Management, and State Supervision

The tax-financed old-age assistance of 1891 was funded by the state and the localities. The 50 percent state reimbursement relieved the financial burden of yeoman farmers, who up to that time had shouldered all the costs of aiding poor elderly people in the countryside. A purely public old-age pension meant that poor farm laborers and smallholders without the means to make contributions to a savings or insurance scheme were eligible for benefits. A tax-financed pay-as-you-go system did not create any intergenerational problems, and the state subsidy immediately helped the farmers to support their elderly.

The universal pension scheme was managed by the localities, and all Danish citizens over the age of 60 who were 'unable to provide . . . what is necessary for their livelihood' (Petersen 1985, 33) had a right to a pension. A set of supplementary conditions was designed to keep the demand at an acceptable level. The assistance did not result in the loss of civil and political rights accompanying poor relief. Indirectly, the requirements bolstered the already time-honoured principle of 'help to self-help'. But it was also an expedient way to keep down the rural costs to poor relief as well as old-age assistance.

The local governments could decide benefit levels. However, the extent to which the level of assistance varied among the localities is uncertain because steps were taken to align benefits to avoid incentives for migration (Petersen 1985, 106-7). The farmer-dominated local governments were also granted wide discretion in deciding whether a person was eligible for assistance (Rigsdagstidende, Tillæg C 1890/91, 545-8). Complaints over denied assistance were decided by the regional state representative, the Overøvrigheden, but the localities could appeal the regional decision to the minister of the interior. The regional authorities assumed only a few other supervisory functions (ibid., 549-52).

Since access to central government power was stalled during the

Systemstruggle, placing administrative control in the localities effectively
privileged the Liberals. The involvement of the state was kept at a
minimum.

Prior to the introduction of old-age assistance several legacies can be identified. From 1856 The Funds for the Poor, De Frie Fattigkasser, provided help for old people. The Funds were based on voluntary contributions, subsidies from the localities (Petersen 1985, 72-3, 162; Nørgaard 1997, 165-7). However, as early as 1842 middle-class people had insured themselves through private annuity schemes. In 1871 a more extensive and state-guaranteed, but still voluntary, annuity scheme was established. A number of other rudimentary schemes also existed (Petersen 1985).

From 1870 to the late 1880s, several schemes were suggested (Petersen 1985, chapter 3). The Workers Commission of 1875 recommended a contributory but a host of competing proposals were advanced in the same period. The idea of some state funding gained strength but the state could do this in numerous ways. The contributions could be compulsory or voluntary; pensions could be contributions-based or determined independently of contributions; the state subsidy could be targeted or could include all; the program could be a bifurcated system combining elements of insurance - or savings - and assistance; and so on. All these forms were contemplated, but until the mid 1880s most agreed that any prospective system ought to be insurance-based in some way.

In the late 1880s the SDP wanted state-run homes for the elderly, where assistance was to be given predominantly in kind. The party opposed any locally managed scheme (Petersen 1985, 221-2; Baldwin 1990, 70-1). The government and the Right also opposed state subsidies to schemes run by the local governments, but they had proposed state subsidies to an old-age insurance a few years earlier (Petersen 1985, 801, 231-3).

The break in the negotiations on old-age pensions was spurred by the escalating conflict between the Moderates and the Reformists over political strategy in relation to the struggle for parliamentarism. The Right had a strong political incentive both to drive a wedge between the two Liberal fractions and to demonstrate that the provisional government could cooperate the lower chamber and pass social reforms. To achieve these goals the government had to accept the Moderate Liberals' proposal on state subsidies to a locally run scheme.

The Liberals had to concede on some of the accompanying regulations and accept some supervisory control by the state. In particular the self-help requirements were strengthened. For this, they obtained an old-age assistancethat was organized in accord with the economic and institutionalinterests independent farmers. The fiscal burden partially shifted to the state, the revenue of which was increasingly collected in the urban

areas, and the Liberals 'preserved farmers' dominance in the administrative
machinery' (Baldwin 1990, 74; cf. Petersen 1985, 103).

A purely publicly funded and managed scheme was passed because this served the interests of the Liberals best. A contributory system would not have been able to attract the poorest groups in the rural districts, which so far had been supported by the independent farmers through poor relief. Not surprisingly, the Liberals preferred to disperse costs to the whole economy by way of the state subsidy and to retain most of the administrative control in the localities.

Sickness Insurance of 1892: State Subsidies, Self-^Management, and Corporatist Supervision 3

By the early 1890s many poor people had already insured themselves against sickness. In the mid 1880s smallholders, farm labourers, and servants made up around 60 percent of the voluntarily insured in the rural districts (Sickness and Accident Commission 1887,29).

In accordance with the Liberals' preferences, the public funding of sickness insurance included only state subsidies. And yet, indirectly, the localities obtained considerable institutional control over the insurance when voluntary insurance associations were guaranteed a high degree of administrative autonomy. Besides, together with a state-employed sickness insurance inspector, a new corporatist body including members of the associations became responsible for coordination, supervision, and other aspects of administrative control.

The institution of voluntary insurance was quite strong by 1885. Almost 1000 voluntary associations insured 120,000 people (ibid., 23-4). By 1891, the number was probably 160,000-170,000. Although other policy legacies can be identified - especially that the local governments so far had assumed some responsibility for taking care of the sick (which the 1892 act ended) - it may be argued that the voluntary insurance institution was the dominant legacy around 1890 (Sørensen 1978; Kuhnle 1984).

Still, other legacies did exist, and a number of other proposals were discussed. A publicly subsidized voluntary insurance could be organized in numerous ways. The company-based insurance that existed prior to 1892 could have been promoted or at least allowed to flourish along with the associations organized by territory. However, the law prohibited this (Rigsdagstidende, Tillæg C 1891/92, 1123-4, §5). Besides, several ways to fund and manage the voluntary insurance were contemplated, and the one Liberal farmers preferred was passed. The Social Democrats wanted the state to run and pay for hospitals and all care services for the elderly (Rigsdagstidende, Tillæg B 1890/91, 1739-43), but nothing came of this proposal.

The sickness insurance also included genuine institutional innovation. The insurance included both a local corporatist board and a central corporatist superstructure dominated by the boards of the insurance associations. no single legacy can explain the chosen organization of the sickness insurance in toto. Although they had to compromise with the Right, again the Liberals had the decisive say on the choice of policy organization.

The Sickness and Accident Commission of 1887 proposed a fixed-permember subsidy which privileged the associations in the rural districts where member contributions were smallest (Sickness and Accident Commission 39-42). In the first bill presented by the Right the subsidy was turned into a percentage share of member premiums {Rigsdagstidende, Tillæg A 1888/89, 1585). This would have privileged the urban districts.

Following the commission's recommendations, the localities were to provide easy and cheap hospital care or a fixed subsidy per resident insurance if the localities did not run hospitals (Sickness and Accident Commission 1887, 45). The rationale for the co-responsibility for funding was that local governments 'obtain a very considerable benefit when the sickness associations exist and thrive in a satisfactory way, . . . because the burden on their poor relief reserve in all likelihood will be so much smaller' (ibid., 39; cf. Rigsdagstidende, Tillæg A 1888/89, 1584). Notwithstanding this rationale, in the final bill the local co-responsibility for funding was removed, except a provision on free transportation to hospitals and doctors.

The subsequent bills and amendments included different funding models, and in all cases the Right wanted the localities to bear a larger share of the costs than the Liberals, who preferred state subsidies (cf. discussion in Rigsdagstidende, Tillæg B 1890/91, 1735-9; Rigsdagstidende, Tillæg A 1888/ 89,1585-6).

As late as in March 1891, the Reformist Liberals in the committee did not venture down the route of cooperation. But after the passing of the oldage the Moderate Liberals negotiated a truce with the Right on the sickness issue. The final bill included only a public subsidy paid out of state coffers, but it was still a compromise. The public support of the associations was composed partly of a fixed sum per member, which privileged the rural associations, and partly of a subsidy of 20 percent of member premiums, which benefited the more prosperous urban insurance associations {Rigsdagstidende, Tillæg C 1891/92, 259-60, 1127-8). On balance, rural associations obtained the largest state subsidy - as a fraction of total expenditures (cf. Sickness and Accident Commission 1887, 42; Rigsdagstidende, paa Folketinget 1891/92, 2715).

The final compromise was made in the lower chamber after the Moderateshad
committee under control {Rigsdagstidende, Tillæg B 1891/92,

550). In the upper chamber the Right and the Liberals simply made a joint statement, which agreed with the Folketinget: 'it is probably not advisable at the present to increase the local governments' burden to purposes like those presently on the table, even if the localities can expect to harvest significant indirect gains from the sickness insurance movement' (ibid., 1667). Thus, the Right agreed to pay the price demanded by the Liberals.

Having won the battle on funding, some administrative alternatives were foreclosed for the Liberals. The norm saying that administrative control and funding responsibility go hand in hand implied that the local governments not demand a privileged role in the administrative structure. The question for the Liberals was how to ensure that their core constituents obtained institutional control by other means. A two-tiered corporatist administration privileging the insurance associations and the localities was one way to observe these interests.

To prevent company-based insurance, the final bill recognized and subsidized associations by territory - which was synonymous with local districts - or trade. Territory-based associations were dominant in the rural districts, and the boards of the associations were probably dominated by Liberal independent farmers.

Because of the considerable public subsidy, the issue of free access to the associations and the question of how to supervise the associations became pertinent. As to the first issue, the first two government bills suggested that a new state bureaucracy headed by an inspector for sickness insurance should decide in all cases where the right to membership was uncertain. But the Liberals wanted more institutional autonomy. In the last government bill, the provision had already been changed. Now a local committee of three members (a doctor, an association representative, and a local government appointee), which in all likelihood would become dominated Liberal farmers, assumed the authority to decide whether applicants could join an association {Rigsdagstidende, Tillæg A 1891/92 1857-8, §5). However, the local corporatist bodies were only one way to ensure that the Liberals obtained institutional control.

Following the government bill, supervision of the extent to which the associations fulfilled the legally mandated articles of association and regulations to be placed in the hands of the new state bureaucracy. By way of delegated authority the inspector could regulate the associations in a number of areas not regulated by law {Rigsdagstidende, Tillæg A 1891/92 1855-70).

The subsequent lower chamber committee report was sceptical of 'letting the public interfere too strongly in this free movement' of sickness associations Tillæg B 1891/92, 546). The Liberal majority on the committee hoped that the development of the necessary standard rules would 'continue on the road of voluntary settlement' (ibid., 549). Thus, the Liberals wanted state subsidies but no state control.

If, on the other hand, the sickness associations were given more autonomy,
instance by creating a central corporatist board, the Liberals endorsed
even stronger delegation of authority to the administrative level:

We therefore recommend including some provisions mandating the organization of yearly
conventions for the associations and letting these meetings provide the standard articles of
associations, rules for member transferals etc. ... It would be most fortunate if in this way
an advisory board for the inspector of the sickness associations could be established to advise
him in the many questions in which his judgement is definitive. Finally, a lot could be said
for giving such a board a certain discretionary authority to intervene in cases of misuse and
unfair competition when the practice is unregulated by law. (Ibid., 550)

In this case too the Liberals achieved their goal. The boards of the associations should now appoint a central corporatist board, the Sickness Association Board, which was given a number of rather substantial powers (Rigsdagstidende, Tillæg C 1891/92, 1133-4, §24).

In sum, the policy legacy thesis has some bearing on sickness insurance. However, it cannot explain why only some legacies were exploited. For instance, traditionally local governments had subsidized the sickness associations, in the final act this never became mandatory. Both funding and administrative control were extremely politicized. The policy legacy thesis cannot account for institutional innovation. An administrative organization on corporatist principles, locally and centrally, was not included the original government bills.

The preferences of the Liberals, as derived from the interests of their core constituents, can explain why only some legacies were exploited, why the state was to pay the full subsidy, and why corporatist administrative structures were incorporated into the final bill in the way they were.

Accident Insurance of 1898: Employer-Paid Private Insurance, Corporatist Supervision

The Liberal influence on the organization of workers' accident insurance was strong, even if the final bill did not include agriculture. Many of the proposals on policy organization tabled by the Liberals were put forward at a time when independent farmers were supposed to be covered by the law, but they remained in effect when the law became restricted to workers employed in dangerous trades. The first bill in 1888 was drafted on the basis of the Sickness and Accident Commission's report from 1887, but its passing was delayed until 1898.

The 1898 law covered most urban trades using mechanical power, and it instituted a main principle of employer liability for accidents occurring at work (Samling af Love og Anordninger 1897-1899, 192-203). Almost all employers insured themselves in private companies or in mutual liability funds. To determine the merit of compensation claims, the size of compensations,etc,

sations,etc,a new corporatist body, the Worker Insurance Board, was established. The board included a chairman and two state officials (including at least one doctor), two employer representatives, and two employee representatives.Observing few procedural guidelines laid down in the law, the new board made its own rules of procedure. If the procedural roles were observed, the board's decisions were final.

In accident insurance there were neither pre-policy legacies of public involvement nor self-organized private initiatives to consider. 'We are not here on historical ground,' the Sickness and Accident Insurance Report said (Sickness and Accident Commission 1887, 71).

The commission majority recommended an insurance model, although somewhat different forms of insurance were proposed. However, the full commission proposed a state-run insurance institution managed by a traditional Some suggested state and local government subsidies to the employer-paid insurance (Sickness and Accident Commission 1887, 71-6,92-100). commission report included agriculture under the law.

Including fishing and shipping, and introducing a different way of calculating contributions, the organization suggested in the government bill resembled the one recommended by the commission (Rigsdagstidende, Tillag A 1888/89, 1605-40). Fishing, agriculture, and the urban trades were to be organized in three main sections, or risk communities. Within the urban trades three risk classes were denned. These groups differentiated employer premiums to some extent, but they were still based on a principle of solidarity within risk groups.

In the upper chamber, a few notable changes were made. Now, selfemployed could also get insured, and the compensations determined the director of the state insurance institution should be appealable to quasi-corporatist regional committees (Rigsdagstidende, Tillæg A 1890/ 91, 2439; cf. 1889/90, 1722-3).

Two employer and two employee representatives should sit on the committees, with two regional state officials. However, the employer representatives were appointed by the local governments and the employee representatives by the injured worker himself in each case. Since committee members were not organizational representatives, it is questionable whether we can talk of corporatism.

By and large the lower chamber endorsed the amendments (Rigsdagstidende, B 1890/91, 1377-94), but party politics suddenly interfered. The report was published three days (6 March 1891) before the committee on old-age assistance submitted its report (9 March 1891), which included the truce between the Moderate Liberals and the Right. The Reformist Liberals probably already knew of the compromise between the Moderates and the Right, and therefore wanted to obstruct the passing of another significant reform.

On 12 March, the Reformist leader, C. Berg, took the unprecedented step of introducing his own bill in parliament after the committee report had been published (Rigsdagstidende, Forhandlinger paa Folketinget 1890/91, 4645-60). Since he spent most of the time on the rostrum taunting the Moderates for their political maneuvering behind closed doors, it is questionable whether Berg actually endorsed his own bill whole-heartedly (cf. ibid., 4676-702).

The introduction of this new bill became significant in two ways. First, it deadlocked the negotiations between the Moderates and the Right. Second, due to the insightful analysis and critique that underlay the bill, it influenced the subsequent policy-making process significantly.

The Berg bill included all farmers. Compensation was to be paid out by the state, but all costs were fully refunded, some by employers in dangerous urban trades, who could insure themselves against this cost privately, and the rest by local government (ibid., 4647-8). The idea of state-run insurance was shelved.

When the Berg bill was reintroduced in 1895 by the new Reformist leader, I.C. Christensen (cf. Rigsdagstidende, Tillæg A 1895/96, 2913-30), it soon became evident that the Liberals did not want (yet) an encompassing social assistance, granted as a right and paid for by local government.⁵ They backed an amendment which took 'out all the provisions which concern the public assistance' (Rigsdagstidende, Tillæg B 1895/96, 1883). For the same reason, accident insurance once again became a targeted liability scheme for workers in dangerous trades.

Berg's proposal on the administrative organization at the central level was extremely cumbersome, indicating that it was made in a rush with the purpose of blocking the negotiations between the Moderates and the Right (Rigsdagstidende, Forhandlinger paa Folketinget 1890/91, 4678).

Like other Liberals, Berg wanted as much administrative power for his own constituents as possible. Besides, because compensation to farmers and farm laborers was paid out of local taxes, it was important to secure his constituents a say in the settlement of claims. A central quasi-corporatist board which should decide all compensation claims served this purpose (ibid., 4651-60). Eight employer and eight employee representatives were to sit on the board, and they were to be selected by two colleges of 400 electors appointed according to different principles (ibid., 4701-2).

When Berg's bill was reintroduced in 1895 the sickness insurance act had been passed. Therefore, the appointment of worker representatives to the board could be made more easily. An amendment recommended that the employee representatives were appointed by the Sickness Association Board rather than an electoral college (Rigsdagstidende, Tillæg B 1895/96, 1893, 1905). This was fully in accord with the institutional preferences of the Liberals.

The administrative and financial organization agreed upon in 1896 was echoed in the final bill (cf. Samling af Love og Anordninger 1897-1899, 192-203). Other unsolved issues - notably the scope of accident insurance, which also came to include (some big) farmers using machines - prevented the bill from getting enacted until 1898. But the institutional organization, which involved the invention of the Workers Insurance Board, was exactly the same as the one the committee in the lower chamber had suggested two years earlier.

No policy legacies can explain the organization of accident insurance. The scope of accident insurance was narrowed during the process, and, except for the state's administrative cost, employers in the dangerous trades covered by the law were liable individually for all benefits paid to injured workers. In practice, employers insured themselves privately against this cost. Constitutional politics implied that only a modest accident insurance policy was passed in 1898 (Baldwin 1990, 81-2). When it was decided that farmers would not be covered by the insurance, the Liberals vetoed public co-responsibility for funding.

Also the corporatist administration had a Liberal origin. The idea of having a quasi-corporatist board as an appeal authority was fostered in the upper chamber. However, it was the Liberals who replaced the state bureaucracy with a centrally organized, corporatist administration, whose decisions were final. Berg's proposal provided the platform for the amendment introduced a genuine principle of organizational representation by the sickness associations on the new Accident Insurance Board.⁶

Again, the Liberals had a strong influence on the organization of the scheme. Corporatism was suggested when the bill included self-employed farmers. When the scope of the bill was narrowed the main tenor of the proposed corporatist organization endured.

Occupational Safety and Health Act of 1901: National Regulation, Corporatist Supervision 1

Factory legislation does not involve considerable budgetary commitments. Thus, corporatism in occupational safety and health regulation is not conditioned the norm saying that funding responsibility and administrative authority go hand in hand. The strong corporatist organization instituted in 1901 was a bourgeois initiative. This time it was the Right and the DA (Dansk Arbejdsgiverforening, the federation of Danish employers) who came up with the idea of a tripartite corporatist body to oversee factory inspection and assume a number of delegated powers.

The Liberal influence on the bill was also evident. The Liberals succeeded
in a general exemption of all 'agro businesses, including dairies' from
national inspection (Rigsdagstidende, Tillæg C 1900/01, 865-82, §1). This

clause narrowed the existing act, which had not exempted dairies or other
agro-businesses, at least not explicitly.

Under pressure from the Social Democrats the government introduced a bill less than a year before the introduction of parliamentarism. Prior to 1901, the national factory inspectorate was a classical bureaucracy, although under two different ministers regulated its authority. Factory legislation dates back to 1873, when child labor in the still nascent manufacturing became regulated and the national inspection was founded (Samling af Love og Anordninger m.v. 1870-1874, 1875, 523-9; see also Nørgaard 1997, 158-61). In 1889 a Machine Protection Law was added. This law also covered farmers using machines. However, rather than accepting by the national inspection, the Liberals managed to create a corps of local layman inspectors; i.e. fellow farmers who should secure the implementation of the law in the countryside (Samling af Love og Anordninger m.v. 1886-1890, 1891, 482-91; cf. Nørgaard 1997, 161-3).

The only pre-reform institutional legacies were the bureaucratic national inspection and the local layman inspection. The government bill did not include a corporatist superstructure either. The administrative control structure became politicized as part of a bourgeois strategy to avoid inspection small, 'safe' firms.

Contention about the scope of the bill was strong in the lower chamber. The committee member from the Right, Niels Andersen, who was also chairman of the DA, unsuccessfully tried to narrow the scope of the bill to firms with more than ten employees. When this demand could not be met, he turned his attention to institutional engineering. The Right and the employers would accept inspection of firms with more than five employees if an exemption clause for small, 'safe' firms was included and the authority to make these exemptions was delegated to a tripartite corporatist body:

after we have proposed to institute a Work Council in a - in my view - very sensible way, a majority in the committee has maintained that there was an additional guarantee that these small businesses were not to be regulated unless it was absolutely necessary. (Rigsdagstidende: paa Folketinget 1900/01, 3756)

In fact the Work Council was more than an appeal authority. It also assumed various advisory functions vis-å-vis the minister. According to the minister this did not infringe upon the authority of the government, which could still decide against the council's advice. The advice of the council would primarily be of value 'regarding technical matters' (Rigsdagstidende: Forhandlinger paa Landstinget 1900/01, 1455-6).

Evidence indicates that the council assumed a much larger role. Almost all regulations issued by the minister involved the Work Council and the organizations (Nørgaard 1997, 244-5; 1999, 43). However, the de facto delegation of decision-making authority to the new corporatist body went

even further. When in 1911 a general revision was proposed by the government,the was a verbatim replica of a proposal drafted by the council {Rigsdagstidende, Tillæg A 1910/11, 3271-506; cf. Rigsdagstidende, Tillæg A 1912/13,2153-94).

The Right's initiative to propose a corporatist council in 1901 was a way to observe employer interests in administrative control to the highest extent possible (excluding workers altogether was impossible). But it also served broader employer interests. Developments in the structure of industrial made it rational for the DA chairman to propose a central corporatist body representing the federations of unions and employer associations. Empowering the central federations - in public policy as well as in industrial relations more generally - was the only way for employers to curb union power at the regional level where employer associations were much weaker (cf. Galenson 1952, 71-2; Swenson 1991; Due etal. 1993).

The Social Democrats welcomed the first institution, which recognized union officials as legitimate worker representatives in policy implementation Forhandlinger paa Folketinget 1900/01, 3741). The SDP member of the committee, Jens Jensen, was chairman of the LO. In person, he embodied the close ties between party preferences and constituency

Still, it was the Right who proposed the corporatist council, and a privileged corporatist administration could not have been created against the will of the party commanding a majority in the upper chamber. The Liberals kept a low profile during the decision-making process because agriculture was shielded from the intrusive inspection by the national inspection corps.

Unemployment Act of 1907: State Subsidies, Self-Management, and Corporatist Supervision *

Unemployment insurance was based on voluntary membership of tradebasedunemployment and could receive state recognition if no other association existed in the trade.⁹ Apart from member contributions, the associations obtained a mandatory state subsidy and a voluntary local subsidy. The subsidy was a fixed percentage of member premiums, and not a fraction of expenditures like in the Belgian Ghent system.¹⁰ Although formally independent of the unions, the associations were de facto union controlled. An inspector and a union-dominated corporatist body, the Unemployment Council, supervised the corporatist self-administration by unions. Although there were no employers on the council, the bourgeois parties still influenced policy outcome. Observing the distributional interests of core constituents, the Right managed to avoid employer contributions to

unemployment insurance. However, with no funding responsibility administrativecontrol
unattainable as well.

Before 1907, unemployment assistance was scattered and insufficient, but there were several rudimentary legacies (Unemployment Commission 1906, 14-16; Vater 1932, 10-12). The most impoverished workers received assistance through poor relief, which carried with it the degrading loss of civil and political rights. Apart from charity (Sørensen 1978, 38-48), since 1856 'worthy' workers had been able to get extraordinary help from the semi-public Funds for the Poor, De Fattiges Kasse (the so-called 'free poverty assistance'). In 1877, and on a number of occasions in the following years, the Ministry of the Interior urged the local councils to contribute to their local Fund for the Poor, so that it could aid those who inadvertently had become unemployed. A few times, the state made interest-free loans available for the local councils contributing to the Funds for the Poor.

After 1900 the state became more involved in supporting the local assistance through the Funds for the Poor. In 1904 a law decided that local councils contributing to the Funds could get a third of the costs refunded by the state (Unemployment Commission 1906, 17).

From the late 1890s, in the wake of a period of strong unionization, and in particular after the turn of the century, a number of unions provided some assistance for their unemployed members. After 1900, union help was probably the most important source of unemployment assistance (Nørgaard 1997, 185).

It may be argued that elected officials chose to build on the most recent and possibly the most elaborated unemployment scheme, which existed in 1907. However, a number of legacies did exist; they were all inadequate; and they could all have provided the paragon for a state-sponsored unemployment The Unemployment Commission pondered a myriad of ways to organize unemployment help, e.g. public assistance and mandatory insurance programs.

All parties were represented in the parliamentary commission, which drafted the bill that reflected a genuine quid pro quo compromise involving all of them. With the backing of a quarter of the electorate, the demands of the Social Democrats could no longer be ignored (Nørgaard 1997, 135). It is hard to know exactly what happened in the commission, since no meeting protocols exist. However, reasonable guesses can be made in the light of the qui bono logic and on the basis of the floor debates in parliament.

The interests of better-off independent farmers, both as employers and as the largest tax payers in the countryside, were still critical for the Liberals. With few unemployed in the rural economy and the still widespread practice of paying wages partially in kind, the Liberals wanted to minimize the rural contribution to what was seen largely as an urban social policy. The biggest risk was the creation of a public assistance scheme like the old-age pensions

or a mandatory insurance system. Both these forms of unemployment help would undermine the incentives for poor farm laborers to accept poorly paid work. Their products being sold at world market prices, the upward pressure on wages could have been debilitating for the rural economy.

The Liberals had a distributional interest in maximizing member contributions the unemployment associations. This alone would make it hard for most poor farm laborers to insure themselves. Besides, if unions controlled insurance few would be able to become members because unionization the rural economy was modest. Finally, if a public subsidy was unavoidable, a local subsidy was preferable to a state subsidy. This would minimize the redistribution from the rural to the urban economy.

The Liberals in the commission did not agree on all questions, but later Minister of the Interior, Jensen-Sønderup, lent his voice to the rural demands. the absence of unemployment insurance, the localities somehow had to support the poor and unemployed themselves. Unemployment insurance was 'as much an insurance for the specific local governments as for individual workers' (Unemployment Commission 1906, 50). Therefore, it was fairest if the local governments paid the full public subsidy or at least co-subsidized the scheme on equal footing with the state (ibid.). A similar argument, but to the detriment of the rural localities, could have been made in the case of old-age assistance and sickness insurance.

'(T)hough partially admitting the validity of the advanced viewpoints' (ibid., 51), the SDP and the Right would not accept any mandatory local subsidy, whatever its size. Representing a majority in the commission, the Social Democrats and the Right had a shared urban interest in dispersing the costs to the rural economy, i.e. in creating a full state subsidy.

Besides this shared interest, the SDP and the Right had conflicting distributional interests. The question of employer contributions, which was known from other countries, divided the parties. A DA official testifying for the commission made it clear that employer contributions were out of the question. Employers were already stretched financially due to contributions to other social schemes such as accident insurance (Unemployment Commission appendix IV, 1; cf. Rigsdagstidende: Forhandlinger paa Folketinget 1906/07, 2407-9). The Right was loyal to this employer demand (ibid., 2461-2).

Following the strong norm saying that funding responsibility and administrative go hand in hand, we could expect that the Right and the employers acquiesced on the issue of institutional control. However, the DA representative testifying for the commission advocated an administrative like the one in accident insurance: a central tripartite council including employer representatives should manage the scheme and decide all unemployment benefit claims (Unemployment Commission 1906. appendix IV, 1).

Apparently, the Right fought for this employer demand in the commission. the floor debate, SDP chairman and member of the Unemployment P. Knudsen accused the Right of having run the errands of business in regard to the issue of administrative control, which he called 'a striking encroachment on the principle of self-administration' (Rigsdagstidende: Forhandlinger paa Folketinget 1906/07, 2507-8). The Right did not refute this accusation.

When the principle of self-administration was invoked, the Liberals had to agree with the Social Democrats. To the Liberals, the combination of privileged influence and no financial responsibility was out of the question. If the Liberals accepted the unsound principle of third-party interference in unemployment insurance, by analogy other parties could argue that external should have a seat on the boards of the Folk High Schools, the sickness insurance associations, and other publicly subsidized schemes of self-administration which were precious to the Liberals.

Having won the battle over funding, employers and the Right had to give in on the issue of institutional control. The only concession obtained by the Right was a clause saying that union funds and the funds of the unemployment associations were to be kept strictly apart. Their boards should also be independent. Examining the floor debate, it is evident that the parties were aware that the unions would dominate the associations (Nørgaard 1997, 190-7).

The twin issues of funding and institutional control continued to be politicized after 1907. The bourgeois parties made some inroads against union dominance of the control structure in 1914 and again in 1921, but the general architecture of a scheme privileging the unions remained intact (Nørgaard 1997, 204-17).

In 1907 the bourgeois parties foremost observed their distributional interests. With considerable success in that respect, they had to give in on the issue of institutional control. Hence the union-dominated corporatist administrative structure. Knowing in hindsight that control of the unemployment was a pivotal resource for boosting unionization (Rothstein 1992a), it is questionable whether the bourgeois parties would have opted for the same bargain could they have chosen anew. However, when they realized this, the 1907 act had already set the parameters for subsequent changes.

Conclusion and Discussion

The five cases studied here are core welfare policies. They provide basic
social security (Flora & Heidenheimer 1984) and intervene 'in a market
economy to meet certain of people's basic needs through relatively direct

means' in order to minimize 'the risk of exploitation of dependencies' (Goodin 1988, 11, 21). Theoretically, they belong to a common universe. Empirically, they were passed in a period of Danish political history when the bourgeois parties dominated both chambers of parliament. Party preferencesand distribution of political power were fairly stable, although the sway of the Social Democrats was increasing from the late 1890s. However, from other vantage points the nve policies are highly different. They addressed different risks and misfortunes; their distributional consequencesvaried; there were different legacies to consider when reform was on the table.

These similarities and differences allow us to draw some fairly solid
conclusions as to the causes of the organization of the early Danish welfare
state (see Table 1).

The policy legacy thesis (Thesis 2) does not find support in the five cases. The dominant legacy was partially exploited only in one case, namely sickness But in this case genuine institutional innovation also took place. Besides, in all cases the issue of policy organization was politicized.

The macro thesis (Thesis 1) gets along somewhat better. In four out of five cases, a corporatist control structure was included in the administrative apparatus. However, the thesis cannot be confirmed unconditionally because assistance deviates from the general pattern. Here, the local governments were given considerable administrative autonomy, and the (confined) control structure included only the state bureaucracy, namely the regional state representatives and the Ministry of Interior.

From a broader state-building perspective and at a somewhat higher level of abstraction, the five policies can be said to share a common institutional trait: all laws involved decentralized administration. To varying degrees, administrative powers were delegated to the localities, private companies, or semi-private organizations. Corporatism and local self-administration, i.e. functional and geographical decentralization, along with pure private solutions, may very well have been looked upon as three species of the same genus. At least for the Liberals, they all embodied the sound liberalist principle of 'help to self-help'. This ideology may even have influenced the other dominant political parties as well.

The Qui Bono Thesis and Beyond

Even allowing for the general similarity in administrative principles, the organization of the five policies did vary in terms of corporatist or public administration. And when funding in the four social security programs is included, the variation is even bigger. The level of public funding varies (from nothing to all) and different forms of public funding were used (local and state subsidies; mandatory and voluntary).

Only Thesis 3, and the qui bono logic on which it rests, can simultaneously account for the administrative organization and the funding model chosen in the five cases. The bourgeois parties, and in particular the Liberals, had a decisive influence on the organization of policy. The (Moderate) Liberals almost unilaterally decided the organization of old-age assistance and sicknessinsurance.

nessinsurance.No principled reasoning can explain why the localities
should manage the former and not the latter; or why the localities should
not contribute to sickness insurance.

Compared to the situation before their passing, the first two reforms had positive distributional consequences for the Liberals' core constituents, and in both cases independent farmers fortified their control positions in the administrative structure.

The Liberals also provided the sketch for the administrative organization chosen in accident insurance. At some point of time in the policy-making process, independent farmers were supposed to be covered by the law, and at that time public subsidies were included. When coverage was restricted, public subsidies were taken out.

In occupational safety and health regulation, the Liberals succeeded in exempting agro-businesses. Having won this crucial concession, the Liberals did not engage in the debate on how to organize an urban policy scheme. The Right suggested a corporatist board that should be given important administrative powers in order to shield small businesses from national inspection. However, the Social Democrats and the LO also welcomed the institutional innovation.

In unemployment insurance both bourgeois parties won important distributional The Liberals obtained a scheme that in all likelihood would not include poor farm laborers, and in this way the upward pressure on wages could be avoided. Facing a united front of urban parties, the Liberals did not succeed in preventing the state subsidy, and the local subsidy was not mandatory until 1921. The Right managed to preclude mandatory employer contributions. When workers themselves assumed the chief economic responsibility for the insurance, the unions and Social Democrats also obtained the administrative control.

The evidence suggests that when distributional gains had to be traded off against administrative control, the bourgeois parties chose the former. Furthermore, the cases show that the Social Democratic influence increased over time. Even if the bourgeois parties, notably the Liberals, still had a decisive say in how to organize welfare policy, the Social Democrats obtained concessions both in 1901 and in 1907.

The legacy theses fail to account for differences across the five policies, and they cannot account for the quite ingenious and highly selective exploitation pre-reform legacies. An explanation focusing on actor dynamics, and in particular on party politics as informed by qui bono logic, fares better.

Focusing on the policy-making processes in five policy areas in one country rather than on broad correlational evidence (Huber et al. 1993; Hicks et al. 1995), the present article provides additional support for the thesis on the influence of bourgeois parties on the shaping of the welfare state.

From finer-grained (but also narrower) analysis one can question whether 'working-class mobilization emerges as a necessary condition for program consolidation' (Hicks et al. 1995,343). When assessed on an individual policy basis rather than as a regime variable including several programs, the picture is more blurred. In old-age assistance, sickness insurance, and accident insurance the Social Democrats did not even set the agenda.

By implication, the study provides (yet) a corrective to the dominant power resource interpretation of the origin of modern Scandinavian welfare states. Elements of the universalist Scandinavian model are not immanent in 'the Social Democratic Road to Power' (Esping-Andersen 1985; cf. 1990). Funding of welfare programs by way of general taxes - thus uncoupling access to benefits from their financing - goes back to 1891; a high degree of universalism was instituted at the same time; and non-discretionary, flat-rate benefits were added in 1921. However, limits to the principle of universal coverage which still exist today, in unemployment insurance notably, are also a result of turn-of-the-century Danish welfare state politics.

Evidence indicates that the bourgeois parties, and in particular those representing agrarian interests, also played a significant role in the development the welfare state in the other Scandinavian countries (Kuhnle 1983; Baldwin 1990; Kangas 1991, Olsson 1993). In the Finnish case, the Agrarian /Center Party championed universalism and high minima in sickness (Kangas 1991). Distributional concerns seem decisive for party preferences. The battle for universalism and tax funding in early Swedish welfare policy was fought primarily by the peasantry (Olsson 1993, 95). Uniform flat-rate benefits were advocated most consistently by the Social Democrats from the 1930s onwards, but the bourgeois parties supported in pension policy (Olsson 1993, 94-107). In Norway, Conservatives a Social Democratic bill to make pensions purely tax funded as early as in 1923. From 1945 welfare reform did not divide the parties: all reforms 'have been enacted without principled opposition, and with little conflict over the proposed bills' (Kuhnle 1983, 155).

Adding the (more impressionistic) evidence from the other Scandinavian countries there is reason to believe that the complex party politics of the Scandinavian welfare states has not yet been fully grasped. Comparative studies of the policy processes in Scandinavian welfare politics may shed more light on the relative importance of the bourgeois parties in various stages of the development of the welfare state. They may also reveal whether different parties systematically have championed some and only some aspects of the Scandinavian model.

Dominant theories on the origin of corporatism also have to be revised. Echoing Bo Rothstein's findings for Sweden (1992b), corporatism in Danish social and labor market policy originated when the bourgeois parties had a firm grip on government power and at a time when there were barriers to

full democratic government. Corporatism in Scandinavian public policy is
not closely related to the Social Democrats' reach of power.

Whether one focuses on the organization of the early Danish welfare state, the origin of the paths that have conditioned the modern welfare state, or on the question of why, when, and how corporatism found its way into Danish public administration, the important role played by the bourgeois parties must be acknowledged. The bourgeois parties also seem to have played a significant role in the other Scandinavian countries. This, of course, may be consistent with the fact that the Social Democrats played a prominent part in post-1945 Danish-cum-Scandinavian welfare politics. But at that time the bourgeois parties had already completed the first act and set the stage for the second.

---

NOTES

1. All the policies studied here were passed before the Radical Liberals gained representation the Folketinget. Even if unemployment insurance was formally enacted after the Radicals gained representation in 1906, the bill had been prepared and cleared in detail in a parliamentary commission in which all the dominant parties had been represented. This happened before the Radicals won their 11 seats out of 113 in the Folketinget. Thus, the compromise was already made.

2. This section is mainly basedon the study by Petersen (1985) and on Baldwin (1990, 65-76).

3. The analysis in this section is basedon a commission report and parliamentary records: Sickness and Accident Commission, 1887; Rigsdags tidende, Tillæg A 1888/89, 1567-604; 1889/90, 1687-704; idem 1891/92, 1855-70; Rigsdagstidende, Tillæg B 1890/ 91, 1729-66; idem 1891/92, 545-94, 1659-78, 1891-94; Rigsdagstidende, Tillæg C 1891/92, 255-70, 1121-38, 1203-4; Rigsdagstidende, Forhandlinger paa Folketinget 1891/92, 595-622, 667, 2708-91, 2837-68, 5493-531.

4. This section is partially basedon the analysis by P. Baldwin (1990, 76-82). Baldwin focuses on the issues of funding and coverage. However, it is doubtful whether the purely tax-funded and universal scheme which he discusses was ever realistically considered. does not study the corporatist and wider administrative aspects of reform.

5. At this point of time in the process the Social Democrats had become active. They presented a purely public and (earmarked) income-tax-financed assistance scheme (exempting the lowest income brackets) which was meant to cover not only workrelated all accidents for people without the sufficient means to provide for themselves excluding those who after the accident still had an annual income of more than DKK 2000). A state agency was to manage the scheme, along with a repeal board; cf. Rigsdagstidende, Tillæg A 1895/96, 2351-62.

6. Other trades obtained accident insurance in the years after 1898, and in 1916 one comprehensive bill was passed. After 1916 the members of the Worker Insurance Board for the urban trades were appointed by the LO and the DA, the umbrella associations for unions and employers, respectively. However, the LO and the DA had not reached mutual recognition in 1898 when the first bill was passed. Other organizations appointed the members of the new corporatist councils in charge of accident insurance in agriculture, shipping, and other trådes (Aarbog for Rigsdagssamlingen 1914/15, 231—41; idem 1918/19, 272-3). Although the functions of the corporatist boards changed somewhat after 1916, the organization of 1898 remained intact.

---

6. Other trades obtained accident insurance in the years after 1898, and in 1916 one comprehensive bill was passed. After 1916 the members of the Worker Insurance Board for the urban trades were appointed by the LO and the DA, the umbrella associations for unions and employers, respectively. However, the LO and the DA had not reached mutual recognition in 1898 when the first bill was passed. Other organizations appointed the members of the new corporatist councils in charge of accident insurance in agriculture, shipping, and other trådes (Aarbog for Rigsdagssamlingen 1914/15, 231—41; idem 1918/19, 272-3). Although the functions of the corporatist boards changed somewhat after 1916, the organization of 1898 remained intact.

7. This section is basedon: Rigsdagslidende, Tillæg A: Forelagte Lovforslag 1900/01, 1841-52; Rigsdagstidende, Tillæg B: Udvalgenes Betænkninger 1900-01, 853-938, 2029-36, 2111-12; Rigsdagstidende: Forhandlinger paa Folketinget 1900/01, 30-2, 3703-830, 4026-58, 4914-45; Rigsdagstidende: Forhandlinger paa Landstinget 1900/01, 1445-74, 1490-523, 1608-34; Rigsdagstidende, Tillæg C: Vedtagne Lovforslag 1900/01, 355-72, 865-82, 891-2. For a detailed analysis of the origin of safety and health regulations see Nørgaard (1997, chapter 4).

8. The analysis is basedon Unemployment Commission (1906). The analysis also builds on Rigsdagstidende, Tillæg A: Forelagte Lovforslag, vol. II 1906/07, 2721-810; Rigsdagstidende, Tillæg B: Udvalgenes Betænkninger 1906/07, 1831-40; Rigsdagstidende: Forhandlinger 1906/07, 2323-32, 2406-508, 2539-82, 5642-60, 5732-43; Rigsdagstidende, Tillæg C: Vedtagne Lovforslag 1906/07, 461-76, 931-2; see also Nørgaard (1997, chapter 5).

9. Associations by territory were allowed, but everybody knew that this practice would never become widespread; cf. Nørgaard (1997, 190-2).

10. This difference in funding structure is often neglected in the literature on unemployment cf. Rothstein (1992a). Compared to the Belgian system it implied a comparatively higher subsidy to the well-off and less unemployment-prone associations, the craft unions.

ACKNOWLEDGEMENTS

For helpful comments and valuable suggestions I am indebted to Sheri Berman, Jens Blom-
Hansen, Jørgen Grønnegård Christensen, Christoffer Green-Pedersen, Thomas Pallesen, Jens
Peter Frølund Thomsen and two anonymous referees.

REFERENCES

Aarbog for Rigsdagssamlingen, various years. Copenhagen.

Baldwin, P. 1990. The Politics of Social Solidarity: Class Bases of the European Welfare State
1875-1975. Cambridge: Cambridge University Press.

Blom-Hansen, J. & Daugbjerg, C. 1999. 'Magtens organisering', in Blom-Hansen, J. &
Daugbjerg, C, eds, Magtens organisering. Stat og interesseorganisationer i Danmark. Aarhus:
Forlaget Systime.

Briggs, A. 1961. 'The Welfare State in Historical Perspective', European Journal of Sociology
2, 221-58.

Bryld, C. 1992. Den demokratiske socialismes gennembrudsår: Studier i udformningen af
arbejderbevægelsens politiske ideologi 1884-1916 på den national og internationale
baggrund. Copenhagen: Selskab til forskning i arbejderbevægelsens historie.
Cameron, D. R. 1984. 'Social Democracy, Corporatism, Labor Quiescence and the Representation
Economic Interest in Advanced Capitalist Society', in Goldthorpe, J. H., ed.,

Order and Conflict in Contemporary Capitalism. Oxford: Clarendon Press.

Castles, F. G. 1978. The Social Democratic Image of Society, London: Routledge & Kegan
Paul.

Christiansen, P. M. & Rommertvedt, H. 1999. 'From Corporatism to Lobbyism? Parliaments,
Executives and Organized Interest in Denmark and Norway', Scandinavian Political Studies
22, 195-220.

Christiansen, P. M. & Sidenius, N. C. 1999. 'Venner for altid? Samspillet mellem centraladministration
interesseorganisationer', in Blom-Hansen, J. & Daugbjerg, C, eds, Magtens
organisering. Stat og interesseorganisationer i Danmark. Aarhus: Forlaget Systime.
Collier, D. & Messick, R. E. 1975. 'Prerequisites versus Diffusion: Testing Alternative

Explanations of Social Security Adoption', American Political Science Review 69,1299-315.
Crepaz, M. M. L. 1998. 'Inclusion versus Exclusion: Political Institutions and Welfare

Expenditures', Comparative Politics 31(1), 61-80.

DiMaggio, P. J. & Powell, W. W. 1991. 'The Iron Cage Revisited: Institutional Isomorphism
and Collective Rationality in Organizational Fields', in Powell, W. W. & DiMaggio, P. J.,

eds, The New Institutionalism in Organizational Analysis. Chicago: University of Chicago
Press.

Due, J., Madsen, J. S. & Jensen, C. S. 1993. Den danske Model: En historisk sociologisk analyse af det kollektive aftalesystem. Copenhagen: Jurist- og Økonomforbundets Forlag. Dybdahl, V. 1969. Partier og erhverv. Studier i partiorganisation og byerhvervenes politiske aktivitet ca. 1880 ca. 1913, vol. 1. Aarhus: Universitetsforlaget.

Elklit, J. 1988. Fra åben til hemmelig afstemning. Aspekter af et partisystems udvikling.
Aarhus: Politica.

Esping-Andersen, G. 1985. Politics against Markets: The Social Democratic Road to Power.
Princeton: Princeton University Press.

Esping-Andersen, G. 1990. The Three Worlds of Welfare Capitalism. Princeton: Princeton
University Press.

Esping-Andersen, G. & Korpi, W. 1984. 'Social Policy as Class Politics in Post-War
Capitalism: Scandinavia, Austria, and Germany', in Goldthorpe, J. H., ed., Order and
Conflict in Contemporary Capitalism. Oxford: Clarendon Press.

Flora, P. & Alber, J. 1984 [1981]. 'Modernization, Democratization, and the Development of Welfare States in Western Europe', in Flora, P. & Heidenheimer, A. J., eds, The Development of Welfare States in Europe and America. New Brunswick, NJ: Transaction Publishers.

Flora, P. & Heidenheimer, A. J. 1984 [1981], 'The Historical Core and Changing Boundaries
of the Welfare State', in Flora, P. & Heidenheimer, A. J., eds, The Development of Welfare
States in Europe and America. New Brunswick, NJ: Transaction Publishers.

Galenson, W. 1952. The Danish System of Labor Relations. A Study in Industrial Peace.
Cambridge, MA: Harvard University Press.

Goodin, R. E. 1988. 'Introduction', in Reasons for Welfare: The Political Theory of the
Welfare State. Princeton: Princeton University Press.

Green-Pedersen, C 1999. 'The Danish Welfare State under Bourgeois Reign. The Dilemma
of Popular Entrenchment and Economic Constraints', Scandinavian Political Studies 22,
243-60.

Heckscher, G. 1984. The Welfare State and Beyond: Success and Problems in Scandinavia.
Minneapolis: University of Minnesota Press.

Heclo, H. 1974. Modern Social Politics in Britain and Sweden: From Poor Relief to Income
Maintenance. New Haven: Yale University Press.

Hicks, A., Misra, J. & Ng, T. N. 1995. 'The Programmatic Emergence of the Social Security
State', American Sociological Review 60, 329^19.

Huber, E., Ragin, C. & Stephens, J. D. 1993. 'Social Democracy, Christian Democracy,
Constitutional Structure, and the Welfare State', American Journal of Sociology 99(3),
711-49.

Johansen, L. N. & Kristensen, O. P. 1982. 'Corporatist Traits in Denmark, 1946-76', in
Lehmbruch, G. & Schmitter, P., eds, Patterns of Corporatist Policy-Making. Beverly Hills:
Sage.

Kangas, O. 1991. 'The Politics of Universalism: The Case of Finnish Sickness Insurance', in
The Politics of Social Rights. Edsbruk: Akademitryck.

Katzenstein, P. J. 1985. Small States in World Markets: Industrial Policy in Europe. Ithaca,
NY: Cornell University Press.

Knudsen, T. & Rothstein, B. 1994. 'State Building in Scandinavia', Comparative Politics
26(2), 203-20.

Korpi, W. 1978. Arbetarklassen i vålfårdskapitalismen: Arbete, fackforening och politik i Sverige. Kristianstad: Bokforlaget Prisma i samarbete med Institutet for social forskning. Krasner, S. D. 1984. 'Approaches to the State: Alternative Conceptions and Historical Dynamics', Comparative Politics 16(2), 223^6.

Kuhnle, S. 1983. Velferdsstatens utvikling: Norge i komparativt perspektiv. Bergen: Universitetsforlaget.

Kuhnle, S. 1984 [1981]. 'The Growth of Social Insurance Programs in Scandinavia: Outside Influences and Internal Forces', in Flora, P. & Heidenheimer, A. J., eds, The Development of Welfare States in Europe and America. New Brunswick, NJ: Transaction Publishers.

Kuhnle, S. 1996. 'International Modeling, States, and Statistics: Scandinavian Social Security Solutions in the 1890s', in Rueschemeyer, D. & Skocpol, T., eds, States, Social Knowledge, and the Origins of Modem Social Policies. Princeton: Princeton University Press; New York: Russell Sage Foundation.

Lehmbruch, G. 1979. 'Liberal Corporatism and Party Government', in Schmitter, P. C. &
Lehmbruch, G., eds, Trends Toward Corporatist Intermediation. London: Sage.

Lehmbruch, G. 1984. 'Concertation and the Structure of Corporatist Networks', in
Goldthorpe, J. H., ed., Order and Conflict in Contemporary Capitalism. Oxford: Clarendon
Press.

Levi, M. 1988. Of Rule and Revenue. Berkeley: University of California Press.

Levine, D. 1978. 'Conservatism and Tradition in Danish Social Welfare Legislation, 1890-1933:
Comparative View', Comparative Studies in Society and History 20(1), 54-69.

Lijphart, A. & Crepaz, M. M. L. 1991. 'Notes and Comments: Corporatism and Consensus
Democracy in Eighteen Countries: Conceptual and Empirical Linkages', British Journal of
Political Science 21, 235-56.

Lipset, S. M. & Rokkan, S. 1967. 'Cleavage Structures, Party Systems and Voter Alignments:
An Introduction', in Lipset, S. M. & Rokkan, S., eds, Party Systems and Voter Alignments:
Cross-National Perspectives. New York: Free Press.

Lipsky, M. 1980. Street-Level Bureaucracy: Dilemmas of the Individual in Public Services.
New York: Russell Sage Foundation.

Luebbert, G. M. 1991. Liberalism, Fascism, or Social Democracy. Social Classes and the
Political Origins of Regimes in Interwar Europe. Oxford: Oxford University Press.
Maier, C. S. 1984. 'Preconditions for Corporatism', in Goldthorpe, J. H., ed., Order and

March, J. G. & Olsen, J. P. 1989. Rediscovering Institutions: The Organizational Basis of
Politics. New York: Free Press.

Martin, C. J. 1995. 'Nature or Nurture? Sources of Firm Preferences for National Health
Reform', American Political Science Review 89(4), 898-913.

Moe, T. M. 1989. 'The Politics of Bureaucratic Structure', in Chubb, J. E. & Peterson, P. E.,
eds, Can the Government Govern? Washington: Brookings Institution.

Nørgaard, A. S. 1996. 'Rediscovering Reasonable Rationality in Institutional Analysis',
European Journal of Political Research 29, 31-57.

Nørgaard, A. S. 1997. The Politics of Institutional Control: Corporatism in Danish Occupational
and Health Regulation & Unemployment Insurance, 1870-1995. Aarhus:
Politica.

Nørgaard, A. S. 1999. 'Arbejdsmarkedspolitik: Korporatisme til alle tider og alle sider', in
Blom-Hansen, J. & Daugbjerg, C, eds, Magtens organisering. Stat og interesseorganisationer
Danmark. Aarhus: Forlaget Systime.

North, D. 1990. Institutions, Institutional Change, and Economic Performance. Cambridge:
Cambridge University Press.

Olsson, S. E. 1993. Social Policy and Welfare State in Sweden, 2nd edn. Lund: Arkiv forlag. Orren, K. & Skowronek, S. 1994. 'Beyond the Iconography of Order: Notes for a "New Institutionalism"', in Dodd, L. C. & Jillson, C, eds, American Politics: Approaches and Interpretations. Boulder: Westview Press.

Petersen, J. H. 1985. Den danske alderdomsforsørgelseslovgivnings udvikling I: Oprindelsen.
Odense: Odense Universitetsforlag.

Pierson, P. 1993. 'When Effect Becomes Cause: Policy Feedback and Political Change', World
Politics 45, 595-628.

Pierson, P. 1996. 'The New Politics of the Welfare State', World Politics 48, 143-79.
Pressman, J. L. & Wildavsky, A. B. 1973. Implementation. Berkeley: University of California
Press.

Rigsdagstidende: Forhandlinger paa Folketinget, various years. Copenhagen. Rigsdagslidende: Forhandlinger paa Landstinget, various years. Copenhagen. Rigsdagstidende, Tillæg A: Forelagte Lovforslag, various years. Copenhagen. Rigsdagstidende, Tillæg B: Udvalgenes Betænkninger, various years. Copenhagen. Rigsdagstidende, Tillæg C: Vedtagne Lovforslag, various years. Copenhagen.

Rokkan, S. 1987 [1981]. 'Faser og mønstre i fremveksten av massepolitik i Norden', in Stal,
nasjon, klasse. Essays i politisk sosiologi. Oslo: Universitetsforlaget AS.

Rothstein, B. 1992a. 'Labor Market Institutions and Working Class Strength', in Steinmo,
S., Thelen, K. & Longstreth, F., eds, Structuring Politics: Historical Institutionalism in
Comparative Analysis. Cambridge: Cambridge University Press.

Rothstein, B. 1992b. Den korporativa staten: interesseorganisationer och statsforvaltning i
svensk politik. Stockholm: Norstedts.

Samling af Love og Anordninger m.v., various years. Copenhagen.

Sickness and Accident Commission 1887. Betænkning afgivet af den af Indenrigs-ministeriet d. 4. juli 1885 til 'Overvejelse af Spørgsmaalene om Sygekassernes Ordning og om Arbejdernes Sikring mod Følgerne af Ulykkestilfælde under Arbejdet 'nedsatte Kommission. Copenhagen. Simon, H. A. 1976 [1945], Administrative Behavior, 3rd edn. New York: Free Press.

Skocpol, T. 1992. Protecting Soldiers and Mothers: The Political Origins of Social Policy in
the United States. Cambridge, MA: Belknap Press.

Skowronek, S. 1982. Building a New American State: The Expansion of National Administrative
1877-1920. Cambridge: Cambridge University Press.

Sørensen, T. B. 1978. Arbejderklassens organisering og socialpolitikkens dannelse.
Copenhagen: Selskabet til Forskning i Arbejderbevægelsens Historie (SFAH).

Steinmo, S. H. 1993. Taxation and Democracy: Swedish, British, and American Approaches
to Financing the Modern State. New Haven: Yale University Press.

Stephens, J. D. 1986 [1979]. The Transition from Capitalism to Socialism. Chicago: University
of Illinois Press.

Swenson, P. 1991. 'Bringing Capital Back in, or Social Democracy Reconsidered', World
Politics 43, 513-44.

Thelen, K. & Steinmo, S. 1992. 'Historical Institutionalism in Comparative Polities', in
Steinmo, S., Thelen, K. & Longstreth, F., eds, Structuring Politics: Historical Institutionalism
Comparative Analysis. Cambridge: Cambridge University Press.

Therborn, G. 1989. '"Pillarization" and "Popular Movements". Two Variants of Welfare
State Capitalism: The Netherlands and Sweden', in Castles, F. G., ed., The Comparative
History of Public Policy. Oxford: Polity Press.

Unemployment Commission 1906. Forslag til Lov om anerkendte Arbejdsløshedskasser med tilhørende Bemærkninger, udarbejdet af den i henhold til Lov nr. 50 af 27. Marts 1903 nedsatte Kommission til overvejelse af Spørgsmålet om en almindelig Invalide og Alderdomsforsikring. Copenhagen.

Vater, Aa., ed. 1932. Arbejdsloshedslovgivningen i Danmark gennem 25 år, 1907-1932. Copenhagen:
for Labor Exchanges and Unemployment Insurance.

Weir, M. & Skocpol, T. 1985. 'State Structures and the Possibilities for a "Keynesian"
Response to the Great Depression in Sweden, Britain, and the United States', in Evans,
P. B. et al, eds, Bringing the State Back in. Cambridge: Cambridge University Press.

Wilensky, H. L. 1984 [1981]. 'Leftism, Catholicism, and Democratic Corporatism: The Role of Political Parties in Recent Welfare State Development', in Flora, P. & Heidenheimer, A. J., eds, The Development of Welfare States in Europe and America. New Brunswick, NJ: Transaction Publishers.

Workers Commission 1878. Betænkning af den ifolge Resolution af 20de September 1875 til
Undersøgelse af Arbejderforholdene i Danmark nedsatte Kommission. Copenhagen.