Why are some countries more corporatist than others? Since the reintroduction the concept of corporatism in political science in 1973. remarkably attention has been paid to this question (McLennan 1984, 101). The debate has mostly been concerned with denning the concept or with discussions about the impact of corporatist arrangements (cf. Cawson 1986; Williamson 1989; Cameron 1984; Therborn 1985). The problem of degree, i.e., that however measured, corporatism varies considerably among liberal capitalist democracies, has been only superficially discussed in the literature.

For the most part, explanations of corporatism have been functionalists, i.e., that the economy, or the state, or the ever ongoing reproduction of capitalism, 'need' to ensure cooperation from various functionally differentiatedgroups. problem is that the results flowing from this structurally imposed 'need', in the form of different degrees and types of

Resumé

The question addressed in this article is: Why are some countries more corporatist than others? It is argued that neither pure micro-, nor pure macro-explanations can account for the great variation in the degree of corporatism (however measured) among western countries. Instead, an institutional variant is put forward, where the structure of the state at the time of the formation of working-class organizations is taken as the main independent variable. It has been shown that the development of collaborative or confrontational labor movements was decided by the reaction of the existing political elites to the demands from the working class in the pre-World War I period. Where suffrage came late, and where the class system was rigid, a radical/revolutionary orientation would dominate the working-class movement, hindering corporatist arrangements, and vice versa. The problem with this argument is that it does not fit the Swedish case. Although democracy was introduced comparatively late and although the class system was rigid, Sweden has been considered the nearly ideal-typical case of corporatism. It is argued that the deviant Swedish case can be explained by the specific structure of the pre-democratic Swedish state — centralized, but not closed; bureaucratic and professional but not especially authoritarian: differentiated but not without central coordination of policy.

---

* This article is part of a larger research project entitled Organized Interests and the Public Good financed by the Bank of Sweden Tercentenary Fund, lam especially grateful to Jonas Pontusson for making several precise and valuable comments.

corporatist arrangements, seem to vary greatly between otherwise similar capitalist economies (Crouch & Dore 1990, 11 ff.; Schmitter 1981, 294; Lehmbruch 1982, 16 ff.; Schmidt 1982, 245). A pure structural explanation of corporatism thus seems unattainable (Cerny 1990, 150ff.; Rothstein 1987). On the other hand, a pure rational-choice approach obviously cannot explain this variation either. If individuals act to maximize egoistic motives, and if this sufficiently explains social behavior, then why should the inclinations(a) individuals to join interest organizations, and (b) of interest organizations to join with the state in corporatist arrangements, vary to such an extent between otherwise similar countries (cf. Bianco & Bates 1990)?

This problem can be readily seen if one compares two nations, such as Sweden and France. Both societies are liberal, democratic, western and capitalist, which means that the structurally imposed 'need' to establish corporatism should be the same, and the preferences of the agents in the economy ought not vary. Nevertheless, one of these countries has very strong corporatist institutions and about 85 percent of its work-force unionized while in the other country there is practically no corporatism, the union density is less than 10 percent (France). It seems obvious, therefore, that neither of the two main types of social scientific explanations, the macro/structural and the micro/individual, can help to explain variations in corporatism (Birnhaum 1988, ch. 2). As Peter Hall has argued, we are mainly confronted with theories that operate on such a high level of generality that they cannot explain the variation between otherwise similar nations (Hall 1985, 229 f.). Instead we need an analysis in which rationality is situated according to the historical institutional variation (Cerny 1990, 161 ff.).

States, Corporatism and Collective Action

One of the most important efforts to explain variations in corporatism is that of Peter Katzenstein (1985). He argues that small European nations with economies open to the world market have a greater need to be able to adapt to changes in the world market than do larger capitalist countries. The small states' need for flexible adaptation has tended to give rise to collaborative efforts and arrangements between the parties in the economy, which in turn has resulted in what Katzenstein labels democratic corporatism. main independent variable, however, is not size, but rather the specific historical development in these small countries which has opened up possibilities for corporatism (p. 136 f.). It is also this variation in historical circumstances before World War II that accounts for the rather great variation between these countries.

Even if the content of Katzenstein's work is important, I want to question his explanation for why only some (small) European countries became corporatist. Katzenstein's suggestion is mainly the timing and character of the process of industrialization (Katzenstein 1985, 174 ff.). The problem is that this explanation does not fit the Scandinavian countries which, as Katzenstein recognizes, had rather different processes of industrialization but developed similar corporatist arrangements (ibid., 176 f.).

The question of variations in corporatism can be reformulated as follows: why have some labor movements been able and willing to engage in exchange with the state and employers' organizations in corporatist arrangements others have resisted, or been denied, any such collaboration? The variation in corporatism is directly connected to the adoption of radical/revolutionary or moderate/reformist policies by the various labor movements, for only a moderate/reformist labor movement would be willing to engage in corporatist exhange with the state and employers' organizations, while a radical/revolutionary one would have to refuse to take part in any such corporatist collaboration.

In order to explain variations in corporatism, then, it seems necessary to explain why some labor movements developed in a reformist/collaborative direction while others did not. According to a traditional functionalist analysis, Marxist or otherwise, there should be no such variation between national labor movements, as they all exist under the same kind of structural circumstances. Yet such differences did occur. Why?

In a seminal article, Seymour Martin Lipset has argued that the dominant ideological orientation of the labor movement in various countries was determined by the pattern of interaction between the labour movement and the state/bourgeois elite in the decades before World War I. The political interaction between the labor movement and its adversaries that then developed has, according to Lipset (1983), conditioned the views held ever since by both the labor movement and its adversaries about labor's proper role in the public policy process. More precisely, Lipset argues that two factors during this period determined the character of the labor movements. First, the more closed and rigid the class system before World War I, the stronger the radical and revolutionary tendencies in the labor movement. Second, the longer the working class was excluded from political influence - i.e., the right to vote - and the harsher the repression of organization and strikes, the more radical the labor movement would become and the more hostile to any collaboration with the state and employers' organizations. Conversely, in a country with a relatively open class system in which the political elite was willing early on to share some of its political power with the emerging working class, reformist and moderate tendencies in the labor movement would dominate.

A problem with this otherwise very convincing argument is that, as

Katzenstein rightly points out (1985, 162), it cannot explain the Swedish case. On the one hand, Sweden has been considered the nearly ideal-typical case of a corporatist society, in which labor has long since been firmly incorporated into the political system (Fulcher 1987; Cerny 1990, 172). On the other hand, the Swedish political elite resisted universal suffrage longer than in most comparable countries. This situation should, by the tenets of Lipset's theory, have resulted not in moderation but rather radicalization of the labor movement. Universal male suffrage was established as late as 1917, and general suffrage in 1921. It came, moreover, after a long and bitter political struggle, and was granted out of fear of revolutionary tendencies both abroad and in Sweden (Lewin 1988, ch. 3). Sweden's nearest neighbours - Denmark, Norway - achieved universal suffage much earlier, and without any particularly intense political struggles (Katzenstein 1985, 161 f.). As for Lipset's second variable, the Swedish class system was, at the time of the industrial breakthrough, more rigid than other comparable nations. This should have given rise to a revolutionary-oriented working-class movement. In sum, Sweden stands as a truly deviant case to Lipset's theory on both scores (Katzenstein 1985, 162).

Katzenstein tries to explain the Swedish case by pointing to the collaboration took place between the Social Democrats and the Liberals until 1923 (ibid.). I agree that this alliance, which was caused by the split in the political right, was important for strengthening the moderate wing in the labor movement. However, it is not obvious how an electoral alliance of this kind explains the development of corporatism. The corporatist channel of representation is usually understood as an alternative to the electoral/parliamentary channel; the route from, for example, the electoral alliance of 1917 to the specific kind of corporatist representation so typical of modern Sweden is thus not at all clear (cf. Crouch & Dore 1990, 11). Instead of corporatism, one could very well argue the result should have been a strengthened Parliament and more pluralist political system. Obviously, something more is needed to explain the deviant Swedish case.

The reformist and moderate character of the Swedish labor movement was actually decided as early as 1917, when the more radical forces inside the Social Democratic party were forced to leave as a minority. What needs explaining is why, at precisely the moment when the party, according to Lipset, should have been most inclined to revolutionary tendencies, its leadership was able to mobilize a firm majority to throw the radical left out of the party and to abandon the revolutionary path. The reformist leadership the Social Democratic party, in the decisive years after 1917, was able (a) to keep a firm control over most of the union movement, (b) to remain the dominant working-class party and (c) to engage in a formal government alliance with one of the bourgeois parties from 1917 to 1920 (Elvander 1980).

Alongside Katzenstein's mostly 'society centered' argument, I wish to put forward a type of explanation for the Swedish case, which has been termed 'neo-institutionalist' or 'state-centered' (cf. Skocpol 1985; Hall 1985, 231 f.). Pierre Birnbaum has forcefully argued that, when it comes to explaining variations in corporatism, one very important factor is usually left out-namely, the state. In corporatist theory, the state is usually treated as a general entity, as if there were one and only one type of capitalist state (Birnbaum 1988). Birnbaum, comparing primarily France and Great Britain, argues that this generalized and ahistorical notion of the state is superficial, because it does not take into account that the character of the capitalist state varies greatly, and that 'each type of state generates specific mechanisms for managing social conflicts and directs action of the working class into specific channels' (ibid. 109). Moreover, he agrees with Lipset that the historical character of the state determined (a) the ideology of the labor movement (Marxist, reformist, trade-unionistor anarcho-syndicalist), (b) whether a corporatist political exchange would develop, and (c) what character such a system would have. Comparing Britain, Germany and France, Birnbaum is able to show that the differences between these states, as an independent variable, have primacy over the timing and pace of industrialization in explaining variations in corporatist arrangements (Birnbaum 1982, 497, 1988, 79).

The problem with Lipset's analysis of the development of reformist or revolutionary tendencies in labor movements is that he considers only one part of the state - the input part (suffrage). Katzenstein's analysis also leans strongly to the input side, as he stresses the question of proportional versus majoritarian electoral representation as a factor behind the development political alliances and corporatism. This is not very convincing, however, as he has to make exceptions for France, Italy and Weimar Germany, none of which, although having proportional representation, developed any strong corporatist arrangements (1985, 150-157).

Birnbaum uses a broader notion of the state's relation to the working class. He agrees with Lipset about the importance of the historical legacy in the relation between the state and the working class, but in contrast to the former writers, he points to the specific administrative character of different states as an explanatory factor when it comes to variations in working-class orientation and/or corporatism. Birnbaum argues that corporatismnever in France because of the exceptional strength and administrative centralization of the French state. The French labor movement was 'always excluded from the state; it had immense difficulties in having its voice heard and always had to act conflictively" (Birnbaum 1988, 123). On the other hand, no stable corporatism could ever develop in Britain on account of the very administrative weakness of the British state - the non-differentiated, non-autonomous, non-institutionalised'

character of the British state has prevented it from intervening effectively
in the labor market, and has hindered the integration of unions and business
organizations in the policy process (ibid.).

Like Lipset, Birnbaum fastens on historical factors during the formative period of the labor movement, but in contrast to the former, he does not confine his analysis of the state to the question of suffrage or parliamentary representation. This broader analysis of the state is altogether in order, for particular to corporatism is that it increases the state's ability to invervene in society and this cannot be accomplished through the electoral process, but rather through the state's administrative apparatuses. It is therefore the organization of the latter which should be focused when it comes to using the state as an explanation for variations in corporatism.

As is well known, this line of reasoning has recently been put forward at a more general level by a number of scholars, most profoundly by Theda Skocpol (1985), who argues that it is necessary to 'bring the state back in' as an important independent variable, especially in comparative political analysis. This is also known as the 'new institutionalism' in political science (Hall 1985, 231 ff.). Yet, there are problems with the analysis of 'state capacity' and its relation to the development of different collective movements as the labor movement. In particular, the causal link between administrative capacity and the inclination of social actors to join in collaboration each other and with the state is not fully established (Rothstein 1990). It would seem, in fact, that it is not state capacity in general, but rather the specific character of this capacity that is the crucial factor. The question is this: what kind of state capacity and organization causes what kind of labor movement? And how is this to be explained? As Birnbaum rightly underlines, not only the question of corporatism, but indeed the entire debate about collective action, free-riding, etc., has generally occurred without full consideration being paid to the state (1988, 9 f; cf. Bianco & Bates 1990).

As stated above, the reformist development of the main part of the Swedish labor movement would seem to be an anomaly. The rigid class system and the late introduction of universal suffrage should have given rise to a more radical/revolutionary labor movement, but what came about was precisely the opposite - a strong but unusually moderate and collaborative labor movement giving rise to a highly corporatist political system.¹ If Katzenstein and Lipset are correct about the pre-World War I period as the 'formative moment', and if Birnbaum is right about the importance of the character of the state, then the relation of the pre-World War I Swedish state to the working class should come into focus.

In the rest of this article, I will argue that what explains the deviant
Swedish case is in fact the structure and function of the Swedish state at
this 'formative moment'. The Swedish working class was indeed denied

suffrage and the class system was certainly very rigid; however, reformism nonetheless came to dominate the major part of the Swedish labor movementon of the early organization of a corporatist system of representation that was established by the Swedish state before World War I. Even if the Swedish working class did not get the right to vote until 1917/ 21, this system of 'early corporatism' gave the organized working class a channel to the Swedish state, and showed the state to be not entirely hostile to working-class demands. Moreover, this system institutionalized corporatism as a natural and workable political system in Sweden, with the capacity to solve conflicts between the parties to the advantage of both sides. What follows is a description of when and where this system came about, the reasons that were given for it and, lastly, what can be said about its long-term effects on the Swedish political system.

Background: A More Popular State

In a comparative perspective, the 19th-century Swedish state may be characterized as less repressive, as containing a less important feudal heritage and as more open to popular demands than was the state generally in continental Europe. This took many forms: an (internationally unique) independent political representation of peasants in Parliament, no legal hindrance of unions or strikes, a rather low ratio of police to citizens, and local instead of central control over a part of the police force. Moreover, violent repression of the working class by the military and police was uncommon, and general public education and literacy developed earlier than in most comparable countries (Therborn et al. 1978; Therborn 1989).

The reform of Parliament in 1866 - when the ancient representational system with four estates was replaced by a two-chamber Parliament (with very restricted suffrage) - actually increased the representation of (landowning) It is also important to note that, because of reforms in the early and mid-19th century, the civil service of this state was rather well paid, politically independent, tenured, professional and uncorrupt. Moreover, it had some political influence, although it was not politically dominant. The Italian or US system of political patronage and bureaucratic clientelism did not prevail in Sweden at this time. Rather, one could speak of a classic Weberian type of nationally unified 'recht-staaf at the turn of the century. The decisive break with aristocratic-feudal modes of state organization had been made between 1850 and 1880. From the 1860s several highly professional government agencies were set up to deal with the need for infrastructure (roads, energy, telecommunications). In sum, this was not a society ruled by state bureaucrats; instead, the Swedish state was rather well integrated with society (Rothstein 1982a; Therborn 1989. 19 ff.).

The Preludes to Corporatism

How did this state react to the emergent 'labor question'? As in all other western countries, capitalism gave rise to the question on what was to be done about the new and problematic class of free-floating workers in society. The reaction of the Swedish state was to set up two investigatory commissions - the Workers' Insurance Committee in 1884 and the New Workers' Insurance Committee in 1891. These were composed of experts and individuals representing various interests (Heclo 1974, 180; Englund 1976). In view of Sweden's late industrialization (from 1870 onwards), this must be considered a rather early effort. It should be noted that this was well before any national union organization had been established in Sweden (1898), and before the establishment of the Social Democratic party (1889), but just recently after the first major strike (1879).

The Workers' Insurance Commissions argued that the state, after the end of the patriarchical guild system, had the responsibility of reducing the insecurity of workers who were 'solely dependent on their labor' (AFK 1888 no. 1:4). They made a rather straightforward anti-liberal, anti-market argument stating that, although compulsory insurance schemes would interfere personal freedom, property rights and personal responsibility, such freedom was only justified if the interests of the individual and society coincided. If this was not the case, the state had both the right and duty to intervene (AFK 1888 no. 1:4 f.).

The proposals of these commissions did not arise from any Bismarkian authoritarianism, nor were they joined to any form of legal repression against the labor movement. On the contrary, they were proposed by Liberal politicians with a humanitarian interest in social issues (Heclo 1974, 179 ff.; cf. Olsson 1990). The major argument was that a system of workers' insurance would help to contain revolutionary tendencies in the labor movement, because if would 'make the great majority of the working population jointly interested in upholding the existing social organism' (quoted in Heclo 1974, 180). The committee of 1884 did in fact contain one laborer, although he does not seem to have played an important role (Heclo 1974, 180 f.).

What is especially interesting in this case was that the first of these government commissions argued that the insurance system they proposed should be administered by a new government agency, to which a corporatist advisory board should be attached. Moreover, disputes concerning the implementation of the insurance should, according to the commission, be settled not by the usual courts of law, but rather by a new kind of corporatively structured institution, which would consist of experts and of representatives for workers and employers, the latter two groups enjoying equal representation. The second commission argued that its proposed

workers' insurance scheme should be administered by a central government agency, and by local 'corporations or so-called pension boards, in which all interests would be represented' (NAFK 1893 no. 1:105). These local corporatively organized boards would decide all the crucial issues connected with the case-by-case implementation of the reform - fees, benefits, and so on (ibid.; cf. Englund 1976, 44-84).

The proposals of neither commission were accepted by the Swedish Parliament. This was not, however, because these commissions recommended insurance schemes be administered by corporatist institutions. the time was not yet ripe for social insurance as such in Sweden (Olsson 1990). No criticism was directed against the principal of corporatist representation. The point is that while suffrage was still very restricted and the subject of hot dispute, corporatist solutions including labor enjoyed legitimacy and were plainly seen as very natural forms of political representation in late 19th-century Swedish political culture.

The first government proposal to establish a corporatist institution came from a Conservative government, which in 1902 put forward a Bill for workers' occupational accident insurance. According to the Bill, a council consisting of five workers' representatives and five employers would be established to oversee the implementation of the insurance (RP 1902 no. 15). The reform as such was accepted by Parliament, being the first modern social legislation in Sweden, but the corporatist council was rejected. The reason for this action was not any argument against corporatist arrangements such, but rather concerned legal technicalities about how the representatives would, according to the Bill, be chosen. In fact, several MPs, argued that a corporatist council of the type suggested would be of great value for the successful implementation of the reform (RPFK 1902 no. 22:23 ff.; RPAK 1902 no. 31:2 ff.).

The Development of Local Corporatism

Corporatist institutions were first established at the local level in Sweden. not at the center. From 1902, city governments started to organize public labor exchanges, which were directly governed by corporatively organized boards. The boards had not merely an advisory role, but took full responsibilityfor the labor exchanges under the city councils. Half the representatives were taken from the labor movement and half from local employers, while the position of chairman was given to a neutral local official. Although this development resulted from local political initiatives, a nationally unified system quickly emerged. By 1907. all major cities had established corporatively organized public labor exchanges. The general principles were as follows: the labor exchanges were not to be restricted to

specific trades or occupations, but should supply labor for all kinds of work, free of charge for both workers and employers. They were to operate on a strictly neutral basis, and were not to side with any of the parties in industrial disputes. A so-called 'golden rule' was established to guide their operations, which stated that employers should get the best available workforceand workers, in turn, the jobs best suited to them (Skogh 1963, 177; Rothstein 1982b). The most vexing issue concerned how the exchanges should behave in industrial disputes. A compromise was struck so that while employers hit by strikes were allowed to apply for workers at the exchanges, the officials had the right and the duty to inform workers if the job suggested was under a union blockade (Rothstein 1982b).

This peaceful development differed markedly from the general European pattern, in which control over labor exchanges usually rested either with employers' organizations or with trade unions (Schiller 1967, 9-36; cf. Pelloutier 1971). While a public and neutral public/corporatist employment exchange system rapidly became dominant in Sweden, it was usually the exception in continental Europe. In Germany, for example, the question of control over the labor exchange system had become a major source of conflict between labor and capital (Schiller 1967, 8, 31 f.). The reason, of course, is that in a capitalist society, control over the supply of labor is of the greatest importance to the parties on the labour market and in the class struggle. If the unions are able to gain control of the labor exchange system, strikes and blockades can be made much more effective. Unions can also use labor exchanges in more regular bargaining with local employers, as they can discriminate between employers when supplying labor. Moreover, they can require workers to become union members if they wish to use the services of the labor exchanges, indeed a powerful selective incentive. In short, if unions control the labor exchanges, their major power resource in the class struggle - control over the supply of labor - can be greatly enhanced. This is of course also why employers' organizations would try to seize control over the labor exchange system. Strikes can be fought much more effectively, union leaders and strike activists can be forced out of the work-force, and union power generally can be minimized.

The labor exchange system is thus of pivotal importance to the parties on the labor market. In some parts of Germany, the unions had gained the upper hand, while in most areas the employers' organizations had been able to seize control. Just why events developed altogether differently in Sweden is less than plain. Neither the Swedish Employers' Federation (SAF), nor the Confederation of Trade Unions (LO), nor the Social Democratic party propelled this development. When the establishment of a public/corporatist scheme was proposed in Parliament in 1903, the chairmanof Social Democratic party, Hjalmar Branting, opposed the proposal.He that, because universal suffrage to the city councils was

not yet established, the exchanges would probably side with the employers despite the corporatist principle. The main Social Democratic newspaper also warned that the exchanges might be used to recruit strike breakers. On the other hand, Social Democrats in some city councils, while not proposing such a system, sometimes argued in favor of it (Rothstein 1982b).

From 1903, the newly established employers' federation showed an interst in establishing an employer-controlled scheme, and sent officials to Germany to study such a system in operation. Although they came back with enthusiastic reports, and argued such a system could be a decisive weapon against the union movement, the leadership of the SAF hesitated to put the plan into operation. The reason is probably a matter of timing, for when the SAF, after some internal turbulence, finally had made up its mind in 1907 to accept the public/corporatist labor exchange system, it was already established, and employers at the local level had not hesitated to support it. Another reason, given by the chairman of the employers' federation in 1907, was that a 'German' system could only be established after a major defeat of the labor movement.

In 1906, a liberal government had obtained the approval of Parliament for government grants to the corporatist local labor exchange system. The conditions for receiving such grants were that the principles established locally in 1902, i.e., unconditional neutrality, corporatist boards, services free of charge to all occupations and trades, and the 'golden rule', be respected (Skogh 1963, 179 f.). It is important to note that these principles, while established at the local level, had been confirmed at the national level, from 1906 onwards, by a series of conferences on the labor exchange question organized by the National Board of Trade, to which representatives the local public labor exchanges were invited. Although a temporary phenomenon, these conferences were the first corporative institutions established on the national level by the Swedish state.²

If neither the employers' federation nor the labor movement stood behind the development of public labor exchanges, who did? The answer is twofold: First, at the local level it was primarily what can be called liberal forces in the Swedish bourgeoisie that were interested in social policy and a peaceful solution to the 'labor question'. Some of these persons also had strong connections with the temperence movement, at this time an important force in the Swedish society (Rothstein 1982b). Secondly, at the central level it was the Liberal government that played the critical role. In particular, the civil servants at the Board of Trade, by initiating the series of conferences on the public labor exchange system, contributed greatly to making this system dominant in Sweden. It should be noted that the national public administration in Sweden at this time had (and still has) a rather unusual organization. While the ministries were quite small and dealt mostly with policy issues, responsibility for actually implementing

public policy rested with semi-autonomous national boards and agencies. These government organizations were not under the direct command of any minister but had to implement policy under their own legal responsibility.This the top civil servants enjoyed a considerable amount of discretion and could take initiatives of their own. Whether the original initiative to arrange the conferences that from 1906 on established the corporatist principles of the labor exchanges was taken by the bureaucrats themselves, or after hints from the liberal government, is unfortunately not known. Officially, it was the Liberal government which in 1906 requested the Board of Tråde to take this initiative.³ It is also important to note that while the Liberal government was replaced by the Conservatives as early as 1907, the Board of Tråde continued its work for the establishment of a public labor exchange. It is likely that without its semi-independent position from the government, this would not have taken place (Rothstein 1982b).

The Development of Corporatism at the National Level

At the national level, several initiatives to establish permanent corporatist institutions were taken before 1914, as stated above. In 1909, a government commission of inquiry dealing with occupational safety also suggested a corporatist advisory board be established to counsel the work inspectorate. In 1908, the National Insurance Board (which had been established in 1903, see above), and which was led by civil servants, sent a letter to the government asking it to establish a corporatist council attached to the Board.⁴ Corporatism was finally established as a permanent principle of political respresentation in 1912 when Parliament decided to set up a new agency, the National Board for Social Affairs. In the commission planning the organization and tasks of the Board, two labor representatives were given seats, one being the chairman of the LO, the other a prominent member of the party leadership. The new Board was to be governed by its top civil servants according to the common rule. But what was new was that the board would have two outside representatives, one from the organized working class and one from the employers' federation (RP 1912 no. 108).

While it was not directly stated in the Bill, the persons appointed were the chairman of the LO and the general manager of SAF. They were not to have merely an advisory role but were to take part in all major decisions on an equal footing with the ruling civil servants. Furthermore, the new National Board for Social Affairs would be assisted by a Social Welfare Council. The council would consist mostly of representatives from the parties on the labor market, and would be divided into four different

assemblies for different issues, each with about 10 to 15 members (DK
1912 no. 5:4 ff.).

Although there was opposition in Parliament to the Bill establishing the National Board for Social Affairs, none of this was directed against the corporatist arrangements. Neither the Conservatives, nor the Social Democrats, the governing Liberals taking part in the debate even mentioned the question (RPFK 1912 no. 34:50 ff.; RPAK 1912 no. 40:11 ff.). The idea of administrative corporatism had obviously become generally accepted in the Swedish political culture by 1912. In its written statement to the government about the proposal to establish this new government agency, the Employers' Federation approved its corporatist institutions. Neither SAF nor LO hesitated to participate in the Board (RP 1912 no. 108: 224 f.).

After this breakthrough, the corporatist principle of representation developed in the pre-democratic Swedish state. In 1913, when a public pension reform was launched, the new National Pension Board also got a corporatist advisory board. One of its corporatist representatives was the chairman of the Social Democratic party (Hjalmar Branting) (RP 1913 no. 298). At the oubreak of the war, the newly established unemployment commission also received a corporatist board (Rothstein 1985). Before 1919, two more important corporative agencies were established in the Swedish state - the Insurance Council and the Labor Council. Neither of these corporatist institutions was limited merely to an advisory role. On the contrary, they were court-like institutions responsible for the establishment legally binding verdicts and precedents in two important areas. The Insurance Council was responsible for solving legal conflicts in the public occupational insurance scheme. The Labor Council had a similar role in the area of implementing the law on the eight-hour working day. Both included two representatives from organized labor and capital respectively, three neutral civil servants, mostly lawyers.

Motives for Corporatist Representation

What motives were given for establishing corporatist institutions in the Swedish state at this time? The task of the new National Board for Social Affairs was not poor-relief, as might have been expected, as this matter was handled by the local authorities. Instead, it was nothing less than the 'labor question". The government commission planning the work and organization of the new Board argued that the reason for it was that "in contemporary society, human labor has become a commodity, so to speak, the supply and demand for which are subject to fluctuations and which consequently has a value that is uncertain and dependent upon shifting

circumstances' (DK 1912 no. 5:4). According to the commission, this caused the problem of this new social category becoming concentrated in the cities, where the masses of alienated and rootless workers became separated from any form of local community, and from the country as such. In the words of the commission:

The more clearly the dangers of industrialism became evident even to the most capable and most irreproachable workers - and the more strongly manual laborers saw themselves as a closed class in relation to employers and other groups of citizens - the more clearly the national dangers of this situation became discernible. The rising level of public education then gave workers the means to clarify for themselves and others the source of their problems, the organizational system gave them the collective power on behalf of their own interests . . . The feeling of solidarity that has emerged among the working masses in itself praiseworthy, is limited to themselves and they do not appear to wish to extend it to the whole society in which they share responsibility and play a part. This obviously poses a national danger, which must be removed in the common interest of everyone. Everywhere the government therefore faces the difficult task of mitigating conflicts of interest and repairing the cracks that are opening in the social structure (DK 1912 no. 5:4).

The National Board for Social Affairs was not established to implement any specific social reform. On the contrary, its task was rather vague, but in any case it is clear it was not established to handle the problem of widespread poverty in Swedish society. Instead, it was to handle the 'labor question' and the way to do this was to incorporate into society this new and clearly threatening social class. Most of all it seems the new Board was established to increase the state's knowledge and information about the many social problems entailed by the 'labor question', i.e., to increase what has been called 'state capacity' in this area (Skocpol 1985).

When the National Insurance Board in 1908 requested permission from the government to establish a corporatively structured council, it argued the agency would thereby 'gain knowledge of different opinions among those members of the public most closely affected by the work of the Board'.⁴ As shown above, the Board was denied such a corporatist council when it was established in 1902, and it should be noted that it was members of the traditional corps of top civil servants in the Swedish state who asked the government for this complement.

The motive for establishing the public/corporatist labor exchanges was to find a way to decrease unemployment and improve the functioning of the labor market (Skogh 1963, 177). However, the reason for their being public and corporatist was above all to prevent a development like that in for example Germany, as described above. When this question was raised publicly for the first time in Sweden in 1895, in the city council of Stockholm, the local commission of inquiry explicitly warned against such a developmentas taken place in Germany.^ When the city councils of Gothenburg in 1901 and of Stockholm in 1902 decided to establish public/ corporatist labor exchanges, the same argument was put forward yet more forcefully. The French 'bourse du travail', which included labor exchanges

under union control, was also criticized. It was argued that when either party had gained control over the labor exchange system and discovered what an excellent weapon it was in the class struggle. it would fight hard to keep that control. Such a development, it was argued, had entirely perverted the idea of labor exchanges as a way of increasing the efficiency of the labor market. Moreover - and this was the most critical argument - improving the efficiency of the labor market benefited not only employers and labor, but also society as a whole.⁶

This argument was put forward not just by the liberal political groups that stood behind the initiative to establish public/corporatist labor exchanges in Sweden. In addition, the local unions in Stockholm, which in 1901 were asked to offer their assessment of the proposal, argued that if the exchanges were to function properly, it would be imperative that these exchanges be trusted by the employers as well as by organized labor, and for this a corporative organization was needed. Moreover, the local commission of inquiry in Stockholm also offered its view of how the representatives should be chosen. According to the commission, it was not enough that workers and employers be equally represented on the board. Instead, it was necessary both parties enjoy the right to choose their own representatives. No outside authority should be allowed to pick representatives because, if this occurred, these representatives would not be seen as legitimate by either unions or employers, and the impartiality of the exchanges would be doubted. This would put the whole institution at risk.⁷ The need for legitimacy is emphasized time and again in the minutes from the city councils, and corporatist representation is held out as the main solution to the problem.⁸

Another important motive behind the institutionalization of corporatism in the Swedish state was the need for flexibility in implementing social reforms. This is related to what nowadays is called the implementation problem - that is, the difficulties of regulating and intervening in volatile areas such as the labor market, with precise rules and regulations. The complexity and variety of the questions to be handled make bureaucratic discretion in the implementation process a necessity. Occupational accident insurance is a good example. The workers" insurance commission of 1893 had argued that 'it is not possible to set these limits in such a way that all doubt as to whether or not a person is entitled to insurance is eliminated in every single case' (NAFK 1893:103). As in all other western societies, the interpretation of laws in particular cases was traditionally handled by the civil court system. In opposing this solution, the commission argued that the court system, with its possibilities for appeal, would be too slow for handling questions of occupational accident insurance cases. The time between the work injury and the final settlement of the case would be too long (RP 1916 no. 111:95). Another, and perhaps more important argument

against using the courts was that judges lacked knowledge and understandingof labor question'. Corporatist arrangements were to be preferred, as they would provide the necessary flexibility and speed in processing insurance claims and this would increase the legitimacy of the reform in the target group (the workers) (AFK 1888 no. 1:101). It was also argued that corporatist bodies would 'give increased life and intensity to the administration (of the law) and prevent it from becomming too rigid' and that such bodies would take initiatives to change the reforms based on the basis of experience in actual implementation (DK 1912 no. 5:164).

In another area, the implementation of the law regulating the eight-hour working day, it was argued that the traditional juridical method of law interpretation was not suitable. Instead, a more mediatory method of law interpretation was needed. The law was so written as to give ample room for exceptions. The agency that was to authorize these exceptions could not simply interpret the law, it also had to balance the interests of workers and employers (ibid., and RP 1902 no. 15:26). This was not a question of right or wrong - according to the law - but of more or less. For this new type of law interpretation, or juridical method, it was argued that a corporative body was more suited than were the traditional courts. Of course, this was also related to the problem of legitimacy, because unions with representatives in, for example, the Labor Council would find it easier to accept its verdicts (RP 1919 no. 333).

Finally, the most important argument for corporatism during this period was the notion of the 'public interest'. This is clear from the arguments concerning both the labor exchanges and the National Board for Social Affairs. If the state did not intervene in the 'labor question', the organized classes would simply pursue their narrow class interests, and this would cause serious problems; it might in the end destroy the very fabric of society. The state could not of course solve the class struggle, but this struggle could be transformed from an unsolvable into a manageable conflict - in game-analytical parlance, from a zero-sum to a positive sum game. The price the established society had to pay was to accept workingclass It was underlined that the very implementation of laws could often give rise to misunderstanding, suspicion and social conflicts. To avoid this, it was necessary to increase understanding and knowledge about the reforms, and the way in which they were implemented, among the target groups. If representatives from these groups were drawn into the state, and were granted the possibility of wielding at least some real influence over the decisions, it was presumed they would change their attitudes. Or, as it was stated by the commission proposing the National Board for Social Affairs, the representatives

would behave as guardians not only of special interests but also of the interests of everyone,
of society as a whole ... It should certainly be expected that a representative body

structured according to these principles, official and thus functioning with a sense of
responsibility, should provide valuable support for the new social welfare administration
(DK 1912 no. 5:1666).

The Impact of Early Corporatism

The remaining question is, of course, whether this system was effective - that is, did it deliver the collaboration from the organized working class to the state? And did social reforms, and the state as such, enjoy enhanced legitimacy among the organized working class. The answer to both these questions is yes. To begin with, the local employment exchanges quickly became dominant, and in 1914 almost all important industrial cities had established this type of corporatist body. From 1906 until the exchanges came under direct state control at the outbreak of World War 11. there was never any opposition in Parliament (not even from communist or left socialist MPs) to the state's subsidies to the cities and municipalities for the costs of operating the public labor exchanges. Nor were the organizational principles ever questioned (Rothstein 1982b). In a report from 1916 to the government regarding the operation of the exchanges, the National Board for Social Affairs declared that 'no objection has appeared from any quarter against the organizational principles on which the publicly operated labor exchanges were based'.⁹ On the contrary, the board argued it was these very principles which had made it possible for the system to grow, and which had been pivotal for strengthening the confidence their operations enjoyed among both employer organizations and unions, 'which in our country have fortunately abstained from utilizing the employment service as a weapon in the social struggle, which in Germany has partially distorted the whole issue of labor exchanges'.^ul The Board also observed that:

Despite the sharp social and political conflicts that have emerged in other areas of public life between members of the employer and worker camps, on the boards of the labor exchanges the same persons have, in the experience of the National Board for Social Affairs. continued to cooperate faithfully in the interest of objectivity."

In a 1920 article, the officer at the National Board for Social Affairs responsible for overseeing the labor exchanges further argued that "it appears as if this form of organization has a number of advantages over the majority principle that rules politically elected assemblies ... it is outstandingly suitable for institutions in which society needs the direct participation of the parties in the class struggle' (Järte 1920. 564). Moreover, aside from the corporatist boards, most of the officials at the labor exchanges were recruited from the labor movement. This was probably to increase the legitimacy of the implementation process (ibid.).

In 1926, an overall evaluation of the corporative institutions of the
National Board for Social Affairs took place. Concerning the representative

system, in which the leading persons from the LO and the SAF participated
in the major decisions of the agency's board, the Board declared the
following:

The purpose of establishing the representative system was undoubtedly to give the National Board for Social Affairs the necessary immediate contact with the main organizations its most sensitive field of activity - the labor market and its organizations . . . the Board selected persons who enjoyed a particularly high degree of confidence from employers and workers, respectively, and were suitable to represent their interests. The fact that persons in such a position became representatives with the approval of their organizations, on the other hand, imposed on them an obligation to regard themselves also as representatives the public.'2

The bureaucrats on the Board also argued that the informal contacts which they had been able to establish with the representatives, and through them with other persons in the organizations, had been of the greatest value. Moreover, they pointed to the importance of the informal and confidential deliberations that the Director General of the Board often had with one or both representatives on especially sensitive and delicate issues. The value of the corporatist system of representation to the general public interest was, according to the Board, that it had made it possible to sweep away prejudices and to create an understanding of the measures undertaken by the state in regard to the relationship between employers and unions. For example, conflicts could be avoided as informal contacts could be taken at an early stage of an issue's discussion, thus preventing the issue from becoming burdened with prestige and/or prejudices.¹³

The meetings between civil servants and the representatives of the Board of Social Affairs were very frequent. The records show that during the first three years of the Board's operation, the two representatives met with the Board almost every week, and made formal decisions in about 100 cases a year. In the following three years meetings were limited to about once a fortnight.¹⁴ However, this should not be seen, according to the Board, as reflecting any declining interest on the part of the civil servants or the representatives in the Board's operations, nor should the Board be considered less important than before. Instead, the many meetings during the first years were occasioned by the need to establish practices and precedents while laying the foundations of Swedish social policy in Sweden. According to the Board, the decline in the number of formal meetings was compensated for by an increase in the number of informal contacts between the officials and the representatives.¹⁵ Once again, it should be remembered that national boards and agencies in Sweden had a very strong position both in policy implementation and in policy creation.

The Social Welfare Council attached to the Board met less often, on
average six times a year between 1913 and 1919. In the above-mentioned
statement, the Board argued that while the council had only an advisory

role, it was still of great value, 'because a more detailed awareness of the contents and purpose of issues could be communicated through members to circles interested in a particular issue, with the aim of eliminating prejudices, preventing misunderstandings and awakening understanding'.¹⁶

As for the work of the Labor Council, there is nothing to indicate that it did not function to the satisfaction of both the state and the employers and trade unions. While the law mandating the eight-hour work day was regularly questioned by the Conservatives in Parliament until the 19305, neither side put forward any criticism of the Labor Council. On the contrary, when the law was debated in Parliament in 1930, the general manager of the SAF, who was also a conservative MP, said it had been of paramount importance for the employers that the implementation of the law was so flexible, and that not one single meeting with the Labor Council had taken place without applications for exemptions from the general eighthour being granted. He argued further than many industries would not have prospered as well as they had done had not the law been made so flexible, and that the smooth workings of the Labor Council had made the law bearable for employers (RPFK 1930 no. 22:14 f.).

Finally, as for the work of the Industrial Injuries Insurance Court, it may
suffice to say that this corporatist institution remained unchanged until the
1970s in the Swedish political system.

Conclusions

It is clear that, while the Swedish working class was denied universal suffrage until comparatively late, the pre-democratic Swedish state was not altogether hostile to its demands. On the contrary, in place of the Parliamentary channel, this state opened up the corporatist model of political representation for the working class without any hesitation. As indicated above, this seems to have been a very successful way of encouraging and moderation in the labor movement. At the local level, the corporatist boards of the labor exchanges showed the organized working class that, even in a very sensitive area, cooperation to the advantage of both parties, and society as a whole, was possible. To politicians it showed that, if properly organized, class conflicts could be institutionalized in ways that benefit the public interest instead of causing ruptures in the social fabric.

Concerning the central level, it is hard to overstate the importance of the regular, almost weekly, meetings between state officials and the leaders of the LO and the SAF. The effect was that both organized capital and labor loyally cooperated with the state in laying the foundations of social policy in Sweden. For the organized working class, social policy thus was

not purely a device used by the ruling elite to pacify the lower orders. On
the contrary, the leaders of the working class were given a say in the
establishment and implementation of social policy.

Social reforms did not develop early in Sweden compared, for example, to Germany or Great Britain, which means the timing of social reforms cannot explain the moderation and reformism of the Swedish labor movement. was special about Sweden was that corporatist solutions for political representation found legitimacy almost two decades before democracy established. We can thus solve the dilemma that Lipset's analysis created, by pointing to the willingness by the pre-World War I Swedish state to establish corporatist arrangements. Contrary to Katzenstein's societycentred we may observe a direct line between the particularities the pre-World War I Swedish state and its modern corporatism. Birnbaum has suggested, the main independent variable which explains variation in degrees of corporatism among nations seems to be the function and character of the state.

The Swedish state at this 'formative moment' was not as strong and centralized as the French state, which prevented any form of incorporation of the organized working class. On the other hand, it was not so weak and decentralized (like the British state) as to prevent any form of 'early corporatism' (Birnbaum 1988; cf. Crouch & Dore 1990, 11). Moreover, it was not closed in the manner of the German state, which made the German working class contract into its ghetto and fed its unwillingness to have anything to do with the state. Even if the British, French and German states also created corporatist institutions during World War I, these arrangements were, contrary to the Swedish case, clearly exceptional. Most important, the corporatist institutions in the countries at war did not prove enduring (Mair 1984, 43).

Rather, the Swedish state seems to have been 'lagom', compared to the states mentioned above. 'Lagom' is a very common Swedish word, which unfortunately is untranslatable; it refers to something like 'in between', 'not too much, not too little'. The Swedish state was administratively strong and centralized, but not authoritarian or totally closed. The civil service was indeed a professional and independent bureaucracy, but not to the extent that it considered any problem soluble by means of the traditional legal system. When it was forced to act in new territory (the 'labor question'), it asked for and received permission from Parliament and the Cabinet to collaborate with the 'target group', in order to increase its own knowledge, information and, most of all, legitimacy. It is, I believe, this very 'lagomness' of the pre-World War I Swedish state that gave rise to a reformist labor movement which in its turn made corporatism such a political success story in Sweden. Contrary to Philippe Schmitter's notion of corporatism, in the Swedish case it was not a strong state or political

elite that created loyal interest organizations for the purpose of controlling the working class (Schmitter 1974). Nor was it the parties on the labor market, who by their very strength, could penetrate and capture the state at this particular time. Instead, the specific 'structuration' (Cerny 1990) that took place between the state and the labor movement at the 'formative moment' of organized capitalism was made possible by a specific type of state - centralized, but not closed; bureaucratic and professional but not especially authoritarian; differentiated, but not without central coordinationof

---

NOTES

1. There is of course much Swedish debate and research about the emergence of reformism in the Swedish labor movement. A useful overview can be found in Ostberg 1990. Most of this work is of minor interest here, however. firstly on account of its lack of comparative ambition, secondly because it neglects factors outside the labor movement such as corporatist arrangements and the state.

2. Socialstyrelsens arkiv, Arbetsmarknadsbyrån: Arbetsformedlingens historia. Riksarkivet. [Archives of the National Board for Social Affairs. the Labor Market Bureau: History of the Labor Exchanges. National Archives. Stockholm.]

3. Ibid.

4. Riksforsåkringsanstaltens arkiv: Utlåtande med forslag rorande åndring av vissa delar av lagen angående ersiittning for skada till fdljd av olycksfall i arbetet. 10 jan 1908. Riksarkivet. Stockholm. [Archivesof the National Insurance Board. Statement ... 10 January 1908. National Archives. Stockholm.]

5. Stockholms stadsfullmiiktiges handlingar 1895 bihang 27:158. [Records from city council of Stockholm 1895 . . . ].

6. Goteborgs stadsfullmåktiges handlingar 1901 no. 132: Stockholms stadsfullmiiktiges handlingar 1902 bihang no. 10. [Records from the city council of Gothenburg 1901 . . . Records from the city council of Stockholm city 1902 . . . ].

7. Stockholms stadsfullmiiktiges handlingar 1902 bihang no. 10:5. [Records from the city council of Stockholm 1902 . . . ].

8. See notes 5, 6 and 7.

9. Socialstyrelsens arkiv, underdanigt yttrande angående statens medverkan for befråmjande avdenoffentligaarbetsformedlingen m.m. den 17 april 1916. Riksarkioet. Stockholm. [Archives of the National Board for Social Affairs. statement . . . 17 April 1916. National Archives. Stockholm).

10. Ibid.

11. Ibid.

12. Socialstyrelsens arkiv, yttrande 30 sept 1926 till sakkunniga for verkstållande av utredning angående fullmiiktigesystemet inom statsforvaltningen. Riksarkivet. Stockholm. of the National Board for Social Affairs. statement . . . 30 September 1926. National Archives. Stockholm.]

13. Ibid.

14. Ibid.

15. Ibid.

16. Socialstyrelsens handlingar. Sociala Rådets protokoll 1913-1926. Riksarkivet. Stockholm. of the National Board for Social Affairs. Records from the Social Welfare Council 1913-1916. National Archives. Stockholm].

REFERENCES

AFK (1888): Arbetareförsäkringskomtlés betänkande. Stockholm.

Bianco, W & Bates. R. 1990. 'Cooperation by Design: Leadership, Structure, and Collective
Dilemmas', American Political Science Review 84, 133-147.

Bimbaum, P. 1982. 'The State versus Corporatism: France and England', Politics & Society
11, 477-501.

Birnbaum, P. 1988. States and Collective Action. The European Experience. Cambridge
University Press.

Cameron, D. 1984. 'Social Democracy, Corporatism, Labour Quiescence, and the Representation
Economic Interests in Advanced Capitalist Society', in Goldthorpe, J., ed.,
Order and Conflict in Contemporary Capitalism. Oxford: Clarendon Press.

Cawson, A. 1986. Corporatism and Political Theory. Oxford: Basil Blackwell.

Crouch, C. & Dore, R. 1990. "Whatever happened to Corporatism?' in Crouch, C & Dore,
R.. eds., Corporatism and Accountability. Oxford: Oxford University Press.

DK (1912): Departementalkomiterades belånkande. Stockholm.

Englund, K. 1976. Arbetarforsåkringsfrågen i svensk politik 1884-1901. Uppsala: Acta Universitatis

Fulcher, J. 1987. 'Labour Movement Theory versus Corporatism: Social Democracy in
Sweden', Sociology 21, 231-252.

Hall, P. 1985. Governing the Economy. Oxford: Oxford University Press.

Heclo, H. 1974. Modem Social Politics in Britain and Sweden. London: Yale University Press.

Katzenstein, P. 1985. Small States in World Markels. Ithaca, N.Y.: Cornell University Press.

Lewin, L. 1988. Ideology and Strategy. A Century ofSwedish Politics. Cambridge: Cambridge
University Press.

Lipset, S. 1983. 'Radicalism or Reformism: The Sources of Working-Class Polities', American
Political Science Review 77, 1-23.

Mair, C. 1984. 'Preconditions for Corporatism', in Goldthorpe, J., ed., Order and Conflict in
Contemporary Capitalism. Oxford: Oxford University Press.

McLennan, G. 1984. 'Capitalist State or Democratic Polity? Recent Development in Marxist
and Pluralist Theory', in McLennan G. et al., eds., The Idea of the Modem State. Milton
Keynes: Open University Press.

NAFK (1893): Nya arbetareforsåkringskomiténs belånkande. Stockholm.

Olsson, S. 1990. Social Policy and Welfare State in Sweden. Lund: Arkiv forlag.

RPAK (1902): Riksdagens protokoll. Andra Kammaren. [Procccdings of the Swedish Parliament.
Chamber].

RPFK (1902): Riksdagens prolokoll. Forslå Kammaren. [Proceedings of the Swedish Parliament.
Chamber].

Rothstein, B. 1982a. 'Den svenska byråkratins uppgang . . . och fall'. Hajten for Kritiska
Studier 15, 27-46.

Rothstein, B. 1982b. 'Från det svenska systemet till den svenska modellen". Arkiv for Studier
i Arbetarrorelsens Historia 23, 1-24.

Rothstein. B. 1985. 'The Success of the Swedish Labour Market Policy: The Organizational
Connection to Policy', European Journal of Political Research 13, 153-165.

Rothstein, B. 1987. 'Corporatism and Reformism. The Social Democratic Institutionalization
of Class Conflict', Acta Sociologica 30, 295-311.

Rothstein, B. 1990. 'Marxism. Institutional Analysis and Working-Class Power: The Swedish
Case', Politics & Society 18. 317-345.

Schiller, B. 1967. Storstrejken 1909. Forhistoria och orsaker. Goteborg: Akademiforlaget.

Schmitter, P. 1974. 'Still the Century of Corporatism?', Review of Politics 36, 85-131.

Schmitter, P. 1981. "Interest Intermediation and Regime Governability in Contemporary
Western Europe and North America', in Berger, S., ed., Organizing Interests in Western
Europe. London: Sage.

Skocpol, T. 1985. 'Bringing the State Back In', in Evans, P., Rueschemeyer, D. & Skocpol.
T., eds., Bringing the State Back In. Cambridge University Press.

Skogh, S. 1963. Arbetets marknad. Stockholm: Almqvist & Wiksell.

Therborn, G. 1985. Why Are Some People More Unemployed than Others? London: Verso.

Therborn, G., Kjellberg, A., Marklund, S. & Oklund, U. 1978. 'Sweden Before and After
Social Democracy: A First Overview', Acta Sociologica 21, supplement: 37-58.

Williamson, P. 1989. Corporatism in Perspective. London: Sage.