The starting point for the debate about corporatism is the organizations' increasing influence on the state. The Confederation's standpoint is that this influence has now reached such a level that there is need for a change. On various occasions, representatives of other interest organizations have also emphasized the importance of keeping a clear-cut distinction between the area of political responsibility and the area of the organizations' responsibility. In political issues it is the politicians' duty to weigh up the different judgments and standpoints and to come to a common public interest. The interest organizations' task is to pursue their special interests. These interest organizations should not take part in decisions in the political sector. If this were to happen the special interests would have too much influence on the political decisions.

This quotation is taken from an official statement given by the Swedish Employers' Confederation in 1985 (Ds C 1986-4, 147, my trans.). It was part of an answer to a report issued by a Government Commission established to investigate the relation between the Cabinet and the central National Board and Agencies (NBA) in the Swedish public administration. The issue under consideration is the report was, inter alia, the composition of the NBAs' boards.¹ As the major business organization in Sweden

Resumé

In what way does corporatist representation change the behaviour of participating organizations? classical question is whether the interest organizations' representatives are captured or not. Sweden is one of the countries in which the class-based interest organizations are heavily represented in the central public administration. The importance of corporatist arrangement in this area in Sweden is due to the fact that the task of implementing government policy is largely given to semi-independent directorates. Since the early 19705, the major business organization in Sweden has been strongly demanding cuts in public spending. As it is represented in some of the most important, and expanding, public agencies of the Swedish welfare state, it could be expected that their representatives would demand a decrease in the: agencies' yearly budget proposals. During a ten-year period (1974-1983), there is no sign of such behaviour of the business organization's representatives in these corporatist institutions, On the contrary, business in Sweden can be said to have been more generous with the taxpayers' than both Social Democratic and bourgeois governments have found possible Assuming that both its general demand for cuts in public spending and its specific action in the corporatist institutions are rational, it is discussed how this seemingly contradictor)' behaviour of Swedish business can be explained.

(Michelletti 1984; Heclo & Madsen 1987), and represented on the boards of several such NBAs, the Swedish Employers' Confederation (in Swedish 'Sveriges Arbetsgi vareforening, S AF') now stated in their answer that as a principle they thought this system was wrong and that they supported the Commission's view that the interest organizations as a rule should no longer be represented on the boards of the NBAs.² This signified an important change in the organization's standpoint (cf. Peterson 1977). Since 1913, it has without any hesitation accepted representation and has taken an active part in the Swedish public administration. The aim of this study is to try to explain why this change occurred in the policy of business.

The Problem

The political importance of this question is due among other things to the peculiar organization of the Swedish central public administration. In contrast to the rest of the Western capitalist democracies (except Finland), Sweden's civil service is not organized as large ministries with a minister at the top. The ministries are instead rather small units and function not as implementation agencies but more as policy organizations. The task of implementing government policy is given to semi-independent directorates, the NBAs, of which there are about 170 which have their own boards and director-generals. The individual ministers have no formal rights to issue specific orders to the NBAs. Ministers are responsible to the cabinet as a whole and the NBAs are supposed to take all decisions with their own legal responsibility according to the laws and regulations issued by Parliament or the cabinet. Thus, the ministers cannot give direct orders to an NBA about how it should decide in a specific case, they can only, as members of the cabinet, issue general regulations or ask Parliament to issue a general law (Petri & Vinde 1978; Heclo & Madsen 1987,10). This system of public administration stems from the seventeenth century but has survived and also managed the tasks of the welfare state. One of the arguments for its preservation has been that ministers have been able to concentrate on general policy and stay out of everyday administrative tasks, which they could not supervise or control in any case.

However, this 'relatively autonomous' position of the NBAs has made the composition of their boards a controversial political question. As governments' interventionist ambitions increase, the laws and regulations in many policy areas cannot be made very precise, which means that the public administration gets a great degree of discretion in matters of implementation (cf. Rothstein 1985a). In many policy areas this means that the demarcation line between what is politics and what is administration is blurred. As this political role of the administration has become more

obvious, the debates about the composition of the NBAs' boards have
intensified.

Originally, the boards were composed of the senior top civil servants working in the agency. But as early as the beginning of this century, when the first agencies handling the 'social questions' were established, representatives from the major class-based interest organizations were given seats on the NBAs' boards (cf. Government Bill 1912-108). This development accelerated especially during the Second World War, when many 'crises commissions' were formed to handle the specific administrative tasks caused by the war. As a rule, the organized interests that were to be most affected by the rather severe wartime regulations were given seats on the commissions' boards (Heckscher 1958,136f; Olsen 1983,169). Business organizations took great part in this and for example the managing director of SAF headed the National Industrial Commission (Friberg 1973).

This way of seeking political legitimacy through the system of implementation on the whole, to be very successful, leading both to policy efficiency and to popular legitimacy for wartime regulations (Heckscher 1946, cf. Offe 1985, ch. 8; Rothstein 1987a). When, after the war, the Social Democratic government intensified its social policy ambitions and established many of the so-called 'reform bureaucracies', this system of corporatist representation became dominant (Hadenius 1978; Olsen 1983, 166; Rothstein 1985b; cf. Elvander 1974). It should be added that until the mid-1980s, this principle of 'administrative corporatism' was not a controversial question in Swedish politics, neither for the major political parties nor for the main interest organizations. What has now and then aroused political tension is the composition of specific boards, not the principle of administrative corporatism as such. To cut a long story short, the situation in Sweden (until now) is that practically all the NBAs have a board composed of persons that are called laymen (a rather misleading concept), which means persons who do not work as civil servants in the NBA in quesion. The exception from this rule is that the director-general of the NBA in most cases also serves as the chairman of the board.

The majority of board representatives are taken from the political parties (23 percent), from the civil service outside the NBA in question (15 percent), and from the major interest organizations (19 percent; source: Ds C 1985-16:13). But this is not the whole picture. Many of the NBAs also have special subcommittees, advisory councils and regional as well as local boards which are composed of so-called laymen. The Swedish bluecollarConfederation Trades Unions (Landsorganisationen, the LO), which is a more homogeneous and centrally coordinated organization than its British counterpart, has such a number of officials engaged in government committees, boards, councils, etc., that it has begun publishing a book for

internal use listing those from the organization that have what kind of assignment where. Counting only the national and permanent authorities (i.e. not the very important occasional investigatory government commissions,cf. & Madsen 1987, 12), the LO in 1982 had 1,163 representativeson public boards, committees, sub-committees, advisory councils, etc. (LO/Uppdragsregistret 1982). When asked about the total number of representatives that the organization is asked to put into the state at central, regional and local levels, a member of the executive committee of the Swedish Employers' Confederation, Mr Stellan Artin, said that it amounts to about 5,000 persons (cf. Ljunggren 1985). To conclude, it is fair to say that parts of the Swedish system of public administration are managed in a way that can be labelled 'administrative corporatism' (cf. Ruin 1974). It must, however, be added that there is no general principle about the composition of the NBAs' boards. Some of them are composed solely of Members of Parliament (such as the National Police Board and that of the National Central Bank), others solely of representatives from the major (i.e. the class-based) interest organizations (such as the National Labour Market Board and the National Board for Occupational Health and Safety), while others are mixed (cf. Statliga myndigheter 1986).

It should be emphasized that the development of the interventionist welfare state, of which Sweden is a prime example, has meant that the scope for precise parliamentary steering of the public administration has been greatly reduced. This means that the power over the implementing agencies has become a much more politically important question as it can be argued that it is in the specific decisions at the administrative level that much, if not most, of what is politically important in the policy area in question is decided (cf. Lipsky 1980; Offe 1985, ch. 8; Lundquist 1987). The question of the importance of 'administrative corporatism' is thus whether the representatives from the different interest organizations really wield any influence, or whether they are just hostages captured by the public agencies (Olsen 1983, ch, 5).

In the discussion about corporatism, the interest organizations' ambitions to wield power and influence over the State have been taken for granted (Olsen 1983, 151). If it is so, as I will take for granted, that many of the NBAs are loci of power in the Swedish polity, then SAF's changed standpoint the system of administrative corporatism needs an explanation. It cannot be explained by a general changed attitude towards these questions among the dominant interest organizations in Sweden because both LO and the white-collar Confederation of Trades Unions (Tjånstemånnens Centralorganisation, the TCO) presented completely different standpoints to that of SAF in their answers to the governmental investigatory commission C 1986-4).

Theories about Corporatism

In a comparative perspective the development of corporatist structures varies to a great extent (cf. Cawson 1986; Scharpf 1987; Lembruch 1984). Even although the notion of corporatism is an 'essentially contested' concept 1985), there have now been many efforts made to operationalize and measure its existence. One can spend a great deal of energy arguing about definitions and measurement, but there seems to be a hard kernel of unanimity about which countries rank high on the corporatist top-ten list. These are foremost Austria, which seems to be the 'model-land' of corporatism (Marin 1985), followed by Sweden and Norway; then come countries such as Denmark, BRD, Finland and the Netherlands (Lembruch 1984, 66, 1985, 13; Schmitter 1981, 294; cf. Scharpf 1987; Scholten 1987; Cawson 1984). Even though Sweden ranks high in almost every comparative study about corporatism, there are few concrete studies about how Swedish corporatism works. As the debate about corporatism does not seem to have come to an end, I hope some empirical insights from the Swedish case might add some new fuel to the debate.

The problem of the discussion about corporatism is that the concept has served as 'ein mådchen fur alles', i.e. as a way to describe very different relations between class-based interest organizations and the State in capitalist Three main interpretations (at least) of what the development corporatism means for state theory have been put forward.

A. One of these is that corporatism is mainly a way of taming the working class, i.e. in the precise meaning of the word, of incorporating the working class into the capitalist system and making it (or at least its organizations) accept the logic of the capitalist economic system. Schmitter, for example, writes that 'the decay of pluralism and its gradual displacement by societal corporatism can be traced primarily to the imperative necessity for a stable, bourgeois-dominant regime ... to associate or incorporate subordinate classes and status groups' (Schmitter 1979, 9; cf. Jessop 1979, 200; Panitch 1980, 175). As I have argued elsewhere, this understanding of corporatism does not give due consideration to the fact that there seems to be a very strong correlation between polities with a strong, not to say dominant, Social Democratic labour movement and polities which rank high on most corporatist indices. Intuitively, there seems also to be something wrong with an argument such that the stronger a national working class is organized, the more it has been tamed by the bourgeois forces. Even so, it might be the case that this taming of the working class is the effect of corporatist arrangements, but then there is the question of intentions versus effects. A more plausible hypothesis would be to argue the other way around, i.e. that corporatism, understood as the existence of corporatist arrangements

such as those mentioned above, can be interpreted as a specific Social Democratic institutionalization of class conflict, i.e. as a result of working class organization strength, not weakness (Rothstein 1987b; cf. Grant 1985).

B. A different interpretation of corporatism is that the strong interest organizations together have been able to colonize parts of the state apparatus. This interpretation indicates that the influence of interest organizations superseded the parliamentary influence, that the modern state is the state of the strong organizations (Lowi 1979; cf. Ruin 1974; Wilson 1973; Olsen 1983). A special variant of this interpretation of corporatism is that the political system has become more and more fragmented and crystallized within specific and rather isolated policy segments. This means that a small number of individuals from political parties, interest organizations, media, the public administration and the academic communities, groups that dominate specific issues. The policies concerning these issues might develop such a complicated structure that only the persons involved can wield any influence (Damgaard 1981).

C. A third interpretation is that the development of corporatist institutional arrangements has implied a statization of the organizations of the civil society. The demarcation line between the state and the civil society has been blurred by the state's ambition to incorporate every organized interest and make it an instrument for state policy. By using the organizations' knowledge, expertise and most of all capacity to legitimatize policy efforts, the State has been able to widen its scope into new areas of civil society (Mann 1986, 127; cf. Offe 1985, ch. 8).

Obviously, a notion of corporatism that contains all these meanings is analytically useless. The reason is of course that, at least to some extent, they respresent diametrically opposed interpretations of the importance and meaning of corporatist arrangements. If corporatism can be anything, then obviously it means nothing. The reason for this messy state of affairs is, I believe, due to a confusion of levels of analysis. The reason is that analyses of corporatism have been made at the level of state theory, about the relations between state and society. I would argue that this is not the proper level at which to start analysing corporatist arrangements (cf. Alford & Friedland 1985, 390f.). On the one hand it is of course an important political fact that different interest organizations are formally represented in a state's policy and implementing apparatuses, or that important political questions are discussed and even negotiated between the government and the major interest organizations. But on the other hand I would argue that the mere existence of this representation does not, as such, imply anything

about the direction in which influence is exerted (Olsen 1983, 152). As Heclo & Madsen have stated in their recent book about policy and politics in Sweden, an administrative technique is not the same as a political regime (1987, 315). State theory operates at a different level from theories about the reasons for formal representation and administrative arrangements. This is the difference between state theory and theory about institutions and organization (March & Olsen 1984). To be able to say something about what corporatist arrangements imply at the level of state theory, one must first analyse its meaning and impact at the institutional level (Schmitter 1985, 40; cf. March & Olsen 1984). This means that one has to separate the question of the general distribution of power in state and society from its institutional arrangements, i.e. so that the latter cannot be seen as a simple function of the former (Olsen 1983, ch. 5; cf. Heclo & Madsen 1987, 315).

When efforts to measure the degrees of corporatism in different countries have been made, the most usual method attempted has been registering different levels of institutional corporatist arrangements. The existence, or non-existence, of central, concentrated interest organizations with a monopoly in their respective areas and their participation in formal political decision-making has been measured. From the level of state theory it is not clear what all this means. Being part of a decision-making institution, or for that matter any informal decision-making group, does not, to my knowledge, guarantee any influence over the political outcome from that institution/group. Political actors might even be forced to take part in decision processes where the final result might be that they have had no influence whatsoever. It might even be the case that participation in such an institution can imply less influence than if the organization in question had stayed out. The reason is that formal participation breeds responsibility that makes forceful opposition impossible or at least very difficult (Olsen 1983, 162).

The confusion of participation and influence is one of the problems in corporatist analysis. This may be seen both in the usefulness of the concept as such and in efforts to bring any meaning to the importance of the existence of corporatist arrangements. In a recent article, Therborn has showed that the degree of corporatism, however measured, has no significant on such entities as levels of unemployment and inflation (Therborn forthcoming and 1986a). Together with other class analysts (e.g. Korpi 1983, 7), he casts serious doubts about the usefulness of the concept, speaking of its vagueness, arbitrary application and the lack of methodological when used in studies of political economy.

True of not, the question of the importance and meaning of corporatist
arrangements still prevails. It would be foolish to argue that the developmentof
extensive representation of interest organizations in for

example the Swedish state apparatuses has no political importance. On the contrary, it might be taken for granted that no such developments in the state structure occur without any intelligible meaning and important politicaleffects 1985; Olsen 1983, ch. 5). But, it might be necessary to analyse this phenomenon on a more concrete, i.e. institutional level, than has hitherto been made. If correlations are low in comparative statistical analyses of the relation between corporatist arrangements and political economy, that might of course depend on problems of validity and reliability in the independent variable(s). But it might also be that the dependent variables are ones quite different from what has hitherto been expected in the corporatist research industry. If this is correct, then we need to start the analysis of the importance and meaning of corporatist arrangements from a more concrete lever, i.e. to see what the corporatists actually do when they 'corporate'.

Method

To study the power and influence of different political actors in specific decision-making processes is always methodologically difficult. The problem power research is, one could argue, not mainly theoretical but methodological. How do you study such problems as agenda-setting and the kind of power that is wielded by ideological domination (what is usually called the second and third dimensions of power) (Lukes 1977, 3-30; cf. Gaventa 1980). The question here is what the participation in corporatist arrangements as such does to affect the preferences and/or strategy of the actors involved. A well-known leftist argument has been that the mere participation in corporatist arrangements by the leaders of working-class organizations would change them in such a way that they would deviate from the pure class interest. The problem with such an argument is of course first establishing what the pure class interests are. The whole discussion is still more complicated, because even if it were possible to operationalize what the 'objective interests' are of this or that class/group, this can only be a part of the formulation of the actor's preferences because they have to take into consideration both strategic (cf. Elster 1982) as well as organizational issues. Even if you know the pure class interests you are supposed to represent, you cannot always (if you are rational) act only according to that because this might threaten the organizational resources you once established as a means of furthering the (pure) class interest (Rothstein 1987b).

How then are the effects of corporatist arrangements on the behaviour
of interest organizations to be studied? One way was mentioned above,
i.e. to create national indices of levels of corporatization and seek significant

statistical variance against variables such as economic growth, inflation, unemployment and the like. This has not brought forth any impressive results (Therborn forthcoming). One reason for this is simply that the theoretical underpinnings might be wrong, that corporatist arrangements have totally different aims and effects than influencing these macro-economicvariables, even variables that are not easy to quantify. Another method is simply to ask the individual representatives from the interest groups, by questionnaire, if they wield any power of any kind, and if participation changes their behaviour when they engage in corporatist arrangements. The problem with this method is of course to get any reliable and intelligible answers (cf. Wennerberg 1983).

Another method which will be used here is carrying out case studies, i.e. following specific issues on their way through the corporatist institutions and trying to measure who influences what, when and how. The problem here is first to select the right policies/issues and second to measure who influences what in those policies. The number of issues passing through a Swedish NBA's board room is very high and their political weight varies a great deal. Another problem in this specific case is that the minutes from the NBAs' boards are very meagre, stating only the final decision and possible reservations but in general none of the arguments from the discussion the specific decision. It has been argued that these problems make it impossible to use material from the boards' minutes to analyse the importance of their composition (Wennerberg 1983). As I will try to show, this conclusion is wrong. The problem may be solved by a careful theorybased of issues/cases, or, in another way, the validity of the independent variables can be increased by theoretical clarification (cf. George 1979; Pedersen 1977).

The problem stated is to analyse the effects that participation in corporatist has on the behaviour of interest organizations, in this case the primary business organization in Sweden. Then we need to select cases where (a) the organization in question does have a general policy, and (b) where it has to act upon this policy in the board rooms of the NBAs where it is represented. The aim is of course to see if there are any deviances from the general policy of the organization and the specific decisions/standpoints that its representatives take when participating in such corporatist arrangements. Does any such case exist? The answer is yes. What will follow below is a comparison between the general and official policy of the Swedish Employers' Confederation towards the level of government spending and the decisions that the organization's representatives when they have to act upon this issue as members of the boards in the NBAs where the organization is represented.

Generally, this is to say in the public debate, at its conferences and in its
journals, the Swedish Employers' Federation has been very outspoken in

its demands for a decrease in public spending and taxes. In the mid-1970s the organization changed its policy in the direction of playing a more politically outspoken and ideological role than before (Heclo & Madsen 1987, 126; Jerneck 1987, 164). Resources were directed to finance special policy institutes, publishing companies and ideological campaigns to promotea pro-business, anti-solidaristic and anti-welfare state political scenario in Sweden (Hansson 1984; Bresky et al. 1981; cf. Micheletti1984). will not discuss the reasons behind this development, but it may suffice to say that it was a product of, as well as an underlying cause of, a much tougher political climate in Sweden between say 1975 and 1985, comprising for example the much debated LO and later Social Democratic proposal to introduce some sort of economic democracy through wageearnerfunds 1987; Åsard 1986). Another reason is that from a position in the middle in the mid-19605, Sweden is now in the lead in the public spending of the OECD countries (OECD 1987). Anyhow, from a Marxist understanding of these things, demanding cuts in public spending seems to be in line with what could be expected to be the pure class interest of the capitalist class (Therborn 1986b; Korpi 1983).

Be that as it may, the formal procedure is this: Each year, every NBA has to deliver a proposal for its budget for the following year to the government. This is, by law, always a question to be decided by the agency's board.³ The national budget is then presented to Parliament by the government in January every year. It states, area by area, both the demand from the NBA(s) in question and the suggestion from the government. The latter is usually somewhat lower than the former (Amnå 1981).

The budget process analysed in this study covers the period from 1974 to 1983. The proposals concern the financial years 1975/76 to 1984/85 (the Swedish state has a split budget year). The reasons for choosing this period are: (a) to capture the period of the ideological offensive of the business community in Sweden; (b) to see if a change in parliamentary majority/ government has any importance in this connection. During this period, Sweden first had a Social Democratic government (to 1976), then four different bourgeois governments comprised of one, two or all three bourgeois (1976-1982) and then again in 1982 a Social Democratic government; and (c) ten is a neat figure to handle.

During these ten years, SAF has been represented in five of the most
important NBAs in Sweden.4 These five state apparatuses are:

• The National Labour Market Board

• The National Board of Occupational Safety and Health
• The National Board of Education

• The National Board for Universities and Colleges
• The National Social Insurance Board

As indicated above, this does not come anywhere near the real figure of the total amount of business' representation in the Swedish state. First, SAF has representatives on many public committees, councils, etc., and in many regional as well as local state authorites. The five agencies above, however, are the only ones that count as NBAs and that decide about budget proposals as stated above. Second, there are many other business organizations in Sweden, such as the Federation of Swedish Industries and organizations for the smaller industries, branch organizations, etc., that also have representatives in several public organizations (Hadenius 1978). Usually companies that are members of the former organizations are also members of SAF. There is thus no business organization competing with SAF as the major spokesman for business in Sweden. The existence of many different business interest organizations is the effect of a division of labour inside the Swedish business community. The reason for confining the analysis to the SAF representatives is first that SAF is without a doubt the main business organization in Sweden, and second that it has been at the forefront of increased political and ideological activity by business interests against the level of public spending since the mid-19705.

These five state apparatuses operate in different policy sectors, so it can hardly be argued that the behaviour of SAF representatives can be explained by the specific character of any one policy sector. However, what they do have in common is that they can all be defined as welfare state apparatuses. By this I mean that they all, in a wide sense, deal with human reproduction. None of them has tasks such as traditional law enforcement (e.g. police or jurisdiction) or physical infrastructre (roads, ports, railways) or the direct promotion of business interests (such as export credits). As; business organizations can be expected to take a more negative stand towards the former than towards the latter of these government activities, it makes this a more profound test of the theory of the corporatist institutionalization question. Moreover, the share that these five state apparatuses of the total national budget was as high as about 43 percent in 1974/75 and 34 percent in 1983/84.⁵ The reduction cannot be explained by a decrease in the importance of the policy sectors concerned but by an increase in government spending on interest payments due to a dramatically increased budget deficit during this period. If interest payments are excluded, the proportion of the five NBAs' spending increased during the period (i.e. as a part of the total substantial government spending). In any case, I think it is fair to argue that no important decrease in government spending in Sweden would take place without affecting the budgets of these five state apparatuses.

In many cases the NBAs' expenditures are of course ruled by precise
laws and regulations issued by Parliament or the government. But the
administrations' independence and discretion in these five policy areas are

definitely sufficient to make their demands and ambitions important for the level of public spending. The pure administrative costs are anyhow not given by any specific or precise rules. An NBA's budget is of course rather complex and the administrative capacity of the agency is of course overwhelming, but there is no doubt that the Swedish political culture is such that it is a lot easier for Parliament and the government to reduce the budget of an agency if there is a reservation on this matter in its budget proposal.

Even if this is not the place to present the workings and importance of these five state apparatuses 'in toto', a few words might be appropriate. The National Labour Market Board has been identified as lying at the heart of the Social Democrats' main economic cum social policy - the 'active labour market policy' (Hanf et al. 1978; Therborn 1986a). This policy is also known as the Rehn-Meidner policy, named after the two LO economists invented it during the late 19405. Compared with labour market authorities in the OECD countries, it is extremely well financed and politically important, and wields a large amount of power over the functioning the labour market in Sweden. Moreover, organizationally it can be defined as a Social Democratic cadre organization, having recruited its personnel mainly from the union movement and breaking with most of the organizational rules of the classic bureaucratic model (Rothstein 1985a, 1986).

As more than 99% of Swedish pupils go through the comprehensive public school systems, the National Board of Education is of course politically (Boucher 1982). The Board is also responsible for the youth vocational training system as well as the high schools and adult education schools. As there are no private universities or colleges in Sweden (and as the autonomy of the Swedish universities is very low, cf. Svensson 1987) the National Board in that area also plays an important role.

Rules and regulations concerning occupational health and safety in Sweden is a question neither for Parliament nor for the courts. The law concerning questions like this is one of the prime examples of what in Sweden is called a 'framework' law. Apart from some very general and legally unimportant statements, this law contains first and foremost the 'rules of the game', i.e. how the formal procedure regulating occupational safety is to be dealt with. The specific regulations concerning occupational health and safety in different branches are thus issued by the National Board in question, which is composed solely of representatives from the unions and the business organizations (Kelman 1981).

Finally, as is well known, Sweden has a very encompassing and general social insurance system (Olsson 1988). However, although it is in many ways steered by general and precise rules (e.g. pensions and child allowances) and thus not plagued by bureaucratic discretion, there are still many

programmes in this area which must be decided selectively, case-by-case (e.g. questions about early retirement pensions, disablement pensions and the like, cf. Hadenius 1986). The National Social Insurance Board plays an important role issuing regulations to the regional and local social insurance agencies in these matters.

It is of course always difficult to state anything more precisely about the political importance of a public agency in general. It can be argued that in some matters (especially educational matters) the NBAs have lost power both to the government departments and to the municipalities since the mid-19705. But on the whole I think it is fair to say that in addition to being examples of administrative corporatism, these five public agencies comprise an important part, not to say the most important part, of the national welfare state organization in Sweden built up largely by the Social Democrats.

If the reasoning so far is correct, these five NBAs can serve as theoretically important cases of corporatist institutionalization. As an interest group. SAF serves as a particularly significant case, because the organization has launched a general policy about the level of public spending that is contrary to the overall development during this period. We can thus measure the impact of corporatist arrangements by comparing the general policy of the Swedish Employers' Confederation towards government spending (the input value) with how the representatives from the organization have acted when having to take a stand on the boards of the five NBAs about their yearly budget proposals (the output value). The greater the difference between these two values, the more important the corporatist arrangements in question. This means that the problem of selecting issues/cases for studying the distribution of power in the NBAs' boards, which was said to be insurmountable, is now solved fairly easily. These decisions are now reduced to the yearly board decisions about the administrations' budget proposals, which are about 50.

Results

All the records from the NBAs in question concerning budgets proposals
have been analysed with the following results.

There is no case where the representatives from the Swedish Employers Confederation have demanded a vote for or made a reservation in favour of a decreased level of spending compared to the previous budget, neither in absolute currency, nor in inflated currency. Furthermore, there is no case where the SAF representatives have demanded a vote for or made a reservation in favour of a reduction of the NBAs' ongoing activities. There are, however, five cases where the SAF representatives have made a

reservation for a lower level of increase in spending by demanding that the
agency should not take on a specified new task.6

This result is even more interesting when the following development is taken into consideration. During this period these five NBAs have increased their staff from 8,970 to 14,385 persons. Taking the depreciation of the value of money in Sweden into consideration, their total budgets have still increased by 46 percent during these ten years.⁷

I have also compared the budget proposals from the NBAs with the government's budget proposal to Parliament for two years during the period under consideration, one with a bourgeois and one with a Social Democratic government (1981/82 and 1982/83).⁸ If the budget proposals from the NBAs are compared with what the government suggests to Parliament, the result is as follows. Both governments suggested to Parliament that the NBAs should be given less money than what they, and thus the SAFrepresentatives, asked for in their budget proposals. The difference between the agencies' and the governments' budget proposals is not dramatic but significant.

Thus, in the general political debate the Swedish Employers' Confederation launched a very outspoken policy to reduce public sector spending. Furthermore, this organization has, together with other business organizations, invested large sums (the amount is unfortunately a secret) in policy institutes, information campaigns and publishing companies to further an anti-welfare state, pro-free market capitalist policy. As could be expected. SAF's policy has been at the general level and no specific suggestions as to what kind of public expenditure should be cut have, to my knowledge, officially been proposed. But I think it is safe to say that the public agencies under consideration would have been hit by any significant cut in public spending proposed by business. Nevertheless, when they have been in a position to influence the level of public spending in concrete decisions, the representatives from SAF have almost consistently suggested increasing the level of public spending. Moreover, SAF's generosity the tax-payers' money has been more profound than that of both Social Democratic and bourgeois governments. In formal terms, the input value and the output value for SAF in the corporatist arrangements under consideration differ to quite an extent. As we can hold constant the influence from parliamentary majority and from differences in policy areas as well, it can be presumed that it is the corporatist arrangements as such that have produced this difference between what business says it wants to do and what it actually does when having to take a stand.

Discussion

It is of interest that SAF's behaviour did not change in these matters in

accordance with which party formed the government. One hypothesis that seems reasonable is that SAF would activate its representatives under the period of bourgeois rule demanding cuts in the NBAs' budgets, as it could be expected that such a government would be more sensitive to demands coming from business. Another hypothesis is that the organization would activate its representatives to follow up its general policy demanding cuts under Social Democratic rule as a way of performing a sort of opposition, or fuelling the bourgeois opposition towards the government. It must be added that one of the main reasons why the Social Democrats regained power in 1982 and kept it in the elections in 1985 was their forceful campaign against proposed cuts (albeit rather limited) in public spending proposed mainly by the Conservative party (Holmberg 1984; Holmberg & Gilljam 1986). Having support in expanding public expenditure from SAF's representativesin five important NBAs gives of course a Social Democraticgovernment increased political legitimacy. From this point of view, SAF's behaviour in the NBAs' boards seems simply to be irrational. But, as has been stated by Jon Elster (1979), the analysis should not stop here: one should always take rationality in political behaviour for granted and proceed with efforts to explain political behaviour as rational so long as there is any intelligible possibility.

The argument so far is that it is the general policy (i.e. wanting to decrease in public spending) that is the real, genuine policy of the business interest (or, if one prefers, the genuine class interest of the capitalist class in Sweden) and that it is the behaviour of its representatives in the NBAs' board rooms that is deviant, i.e. that has to be explained by some effect of the system of administrative corporatism. However, it might just as well be the other way around. It might be the case that it is the general policy of SAF that has to be explained as deviant from any such pure business interest, not its behaviour in the NBAs. One such explanation is common in some variations of Marxism under the label 'false consciousness'. Usually this notion is used to explain the (non-revolutionary) behaviour of the working class in capitalist societies. But the notion of 'false consciousness' might be turned upside-down and used in this context as well. That is to say, the general demand for cuts in public spending from SAF is a function of ideological manipulation, i.e. of false consciousness, and so the organization's (objective) interest lies in an increase of public spending (at least in the parts under consideration in this study) as it is expressed in the NBAs. This is first realized when the representatives from the organization are confronted with the specific programmes that are managed by the NBAs in question.

There is, however, a general problem with this kind of explanation which
is that they tend to be analytically empty. With a concept such as 'false
consciousness' one might explain any kind of political behaviour, i.e

everything and thus nothing (Elster 1983:143, esp. note 5). It has been argued that much of what occurs in the Swedish policy might be explained by the ideological hegemony of the Social Democrats (Heclo & Madsen 1987). I doubt that this is correct, or more precisely that the ideological hegemony of the Swedish Social Democrats is so great that it can force the primary business organization to irrational political behaviour (i.e. behaviour that conflicts with its basic class interests). But, even so, if we accept the proposition that SAF's demands for cuts in public spending are just a result of 'false consciousness', it still does not explain the deviance in the organization's behaviour between its public statements and its specific decisions in the NBAs' board rooms.

Perhaps speaking with two tongues should be interpreted as a rational strategy for the organization to achieve important but contradictory goals. The general demand for cuts in public spending is then to be understood as a kind of 'symbolic' policy launched by the SAF to raise enthusiasm among the rank and file members of the organization (particularly militant owners of small enterprises, cf. Jerneck 1986, ch. 8), but not as a policy intended to have any concrete results. Strictly agent-oriented (i.e. structureless) of political behaviour are always tricky and this particular kind of explanation cannot be excluded. But in this special case, I would have some doubts about it. First, the demand for consistency in the public debate is great. Second, the organization in question does express some awareness of the problem. From the SAF's journal comes the following quotation:

SAF strongly demands cuts in public spending. This demand is emphasized now, especially before the SAF congress. However, the SAF representatives on the National Labour Market Board have not made a reservation against the demand for 1,350 new civil servants. How does this make sense? (SAF-tidningen 1980-28, my trans.).

This shows that the organization has realized the inconsistency in its
behaviour. In its statement to the government commission investigating the
composition of NBAs' boards (mentioned above), the organization stated:

Even for the interest organizations themselves this participation in the decisions of the public authorities leads to unsolvable problems. The representatives from the interest organizations get double loyalties and ambiguous mandates. The members of the interest organizations have difficulties in calling their representatives to account. (Ds C 1986-4, 147, my trans.)

As this shows, there is now reason to question the accuracy of the inputvalue'in analysis. During the period under study, the Swedish Employers' Confederation has wanted a decrease in public sector spending, and, it might be added, for good reasons. That this demand has not been presented in the NBAs where the organization has seats must then be explained by the effects of the corporatist institutionalization. It is tempting

to stop the analysis here by stating that SAF has, by its incorporation in these corporatist arrangements, not been able to keep a consistent line in these matters. It seems obvious that the Swedish labour movement, by constructing these corporatist institutions, has been able to get the loyal cooperation from its major political opponent in maintaining a very high and, moreover expanding, Swedish welfare state.

But the question still remains, how does this state of affairs really come about, and how is it to be explained? If we assume that both SAF's general policy towards the public sector and its specific behaviour in the NBAs is in some sense rational (i.e. in accordance with the basic interests of the organization), and if we further assume that the organization's main preference lies in a decrease in public spending, then how is its apparently confusing behaviour to be understood. In more formal language, if neither rationality nor policy preferences are the independent variables that can explain this behaviour (as they do not vary), then it seems that the answer lies in the institutional setting, i.e. the existence or non-existence of a corporatist arrangement. The theoretical interest here lies in that this touches upon the meta-problem of social science, i.e. the relation between structural and agency oriented explanations. Corporatist arrangements can be understood as administrative structures that are constructed by political forces at time tx with or without the political intention of changing the behaviour of the participating agents at time t₂. Thus we may transcend the debate about micro or macro (or structuralism-voluntarism) by reasoning while preferences are important, they must be explained by their structural setting. This setting, however, might be constructed at time tx by an agent (A) with, or without, intentions to influence the behaviour of agent (B) at time t2t₂ without the agent (B) being aware of this. The latter's behaviour might then be explained both intentionally (as in accordance with her preferences) and structurally, as a result of the institutional setting where she has to act (cf. Rothstein forthcoming, March & Olsen 1984). Thus says Mr Lars-Gunnar Albåge, a member of SAF's executive and the organization's representative on the National Social Insurance Board since 1963:

It's not so very strange that this question has popped up (SAF's intention to withdraw from the NBAs' boards/br) because things change. What was true in the 1950s and 19605, with a relatively small public sector and rather small NBAs with more precise rules and not such very large systems or great costs, it was a different society from the one we have in 1987. . . . But it is still a very difficult question for us. I can confirm that now and then things develop in such a way that you participate in something you would rather not participate in. But on the other hand this must be weighed against the possibility of getting decisions in other questions where things have not become as bad as you might have expected. (Mr Lars-Gunnar Albåge interviewed December 1, 1987, my trans.)

Explanations in the social sciences can take place at different levels of
analysis. Here I will take three such levels into consideration, the individual.

the organizational and the societal. I think of these levels of explanation and their respective theories as Chinese boxes, the higher incorporating the lower but the latter still retain their own explanatory force (Benson 1982; Lundquist 1987, 43 ff.).

The Individual Level

What influences are wielded upon the individual SAF representatives on the board of an NBA? I will, because of lack of competence (and interest), not go into the more or less intrinsic psychological dimensions of this question. The question is, however, do the business representatives act as SAF's representative in the public agency or as the public agency's representative in SAF? One of the directors of SAF who has been a member of the National Labour Market Board since 1982, gives the following answer:

This is an interesting question and there is something about this, about the role one gets into. At times you get into a position in between, so as to speak. You are nominated by your organization that has a standpoint about the policy to be implemented. At the same time, when you are a member of the board, it is your duty to follow the instruction issued by the government. This means that sometimes you have accepted to manage and implement a policy that you do not, at least in some parts, approve of. We who are representatives on the boards have had some rather philosophical discussions about this problem and I have argued that sometimes it it necessary to differentiate between your role as a member of the board and your role as a SAF representative because the possibility for any radical questioning of the policy does not exist . . . You can come into a dilemma where there does not seem to be any point in making a reservation but where we in fact would have preferred a totally different labour market policy . . . You also get a somewhat split attitude because you understand better why the agency acts as it does under the circumstances. This might cause a problem backwards to your organization. You can manage this problem better or worse, but in any case you have these roles. (Mr Stellan Artin interviewed Sept. 2, 1987, my trans.)

To the same question, Mr Lars-Gunnar Albåge (SAF's representative on the National Social Insurance Board) says: 'In one way yes, but I do not think of myself as SAF's representative on the board, but as an ordinary member of the board with special competence from this side. I cannot act as a negotiating party on the board, and it is this dilemma that has now produced a scepticism within SAF. In some ways you are sitting between two chairs' (interviewed Dec. 1, 1987, my trans). This should be compared to what SAF stated in its answer to the government commission mentioned above: 'The federation is the companies' collective spokesman on ... the National Social Insurance Board' (Ds C 1986-4, 148).

There is, however, one more thing that can throw some light on this matter. The mandatory period of SAF's representatives is surprisingly long. The SAF representative on the National Social Insurance Board has held his seat since 1963. In the National Labour Market Board as well as in the

National School Board there are several SAF representatives who have served more than 15 years a members of the boards.⁹ These very long periods will, I believe, bring forth a sort of psychological incorporation of the individual representatives. It is probably difficult to participate on a board, or in any small group, for such a long time and still be very critical of its policy and operations. This might indeed change the role of the representatives such that they will become more of ambassadors for the NBAs in their interest organizations than vice versa.

The Organizational Level

One explanation for SAF's contradictory behavior is that the organization as such as business in general might be dependent on favourable behaviour from the agency's personnel, at both its leading and 'street' level, in their implementation of the policy in question. This is especially true for the National Labour Market Board and the National Board for Occupational Health and Safety that handle laws which give them a large degree of discretion. If SAF were consistently to demand cuts, this might lead to a negative stand from the agencies' personnel at different levels, both to business interests in general and to its member companies. In the short term, this might be difficult for the organization to handle.

Another organizational explanation is that just like any other interest organization, business organizations need to recuit new members and keep old ones. In other words, every business organization needs to present some output that makes it rational for individual companies to join the organization and keep paying the costs of being members of it. One result that a business organization such as SAF can produce is influence over (or perhaps protection from) the allocation of values that takes place in the public administration.

Within some of the policy areas in question, e.g. labour market policy and occupational health and safety policy, the administrations' room for manoeuvre is exceptionally large. In these cases, a business organization can offer protection from some interventions from the agency handling the occupational health and safety programme, and it might offer pieces from the large cake that is distributed by the Swedish labour market authorities. Thus, the organization can offer potential members something which it does not pay anything for, i.e. something for nothing. Influence, or just information in due time, over matters such as these might be useful arguments when SAF recruits new members. For any single company, it might be difficult, or even dangerous, to stay out of an organization that can influence such matters. It might thus be the case that in order to get some influence over the specific case by case (or better company by

company) implementation of public policy, SAF has to accept the level of the administrations' budget proposals (cf. Schmitter 1985, 40 ff.). It might well be the case that this kind of organizational rationality can come into conflict with the basic class interest that the organization is supposed to defend and promote and which forms the base for the official and general ideology of the organization (Cawson 1986; 87; Alford & Friedland 1985, 349; cf. Coleman & Grant 1984).

A third explanation at this level is that an organization such as SAF develops its own internal administrative logic with its own specific segments of interest (cf. Buksti & Johansen 1979). As the organization has to respond to a large number of public areas, it develops correspondingly specialized segments. All the policy areas in question do, in fact, have such corresponding inside SAF. From this it might follow that these internal divisions will adjust to the established views of their corresponding policy areas and thus become spokesmen for their views inside the organization. A more cynical interpretation is that the persons working within these internal divisions of the organization earn their living from their special knowledge about the policy area(s) in which they have specialized, means that a total confrontation between the organization and the public agency in question is thus not in their interests. Even more so, the status and prestige of the segment will increase the more it can argue for the importance of its corresponding policy area/public agency. To quote Mr Stellan Artin once again:

We have tried to produce written policy documents for each policy area, but it is a fairly recent effort. However, it is rather difficult to make these documents precise enough. The changes that we suggest will often be marginal and this is to some extent our problem with the Labour Market Board, that it is their views that become dominant, not only concerning the administration as such but also concerning the whole policy. The problem becomes how to construct the rules about specific labour market allowances rather than ideas about the general economic policy. This might lead to a certain ambiguousness in our message when our economists at the same time argue for a totally different policy for the labour market while we are discussing the specific design of the rules concerning specific allowances . . . It happens that we come back after having agreed to a decision and this leads to an internal debate about why on earth we have agreed to this and that (interview 2 Sept., 1987, my trans.)

Indeed it seems like constructing corporatist arrangements is a successful way for a government to change the preferences/behaviour of powerful interest organizations. As has been pointed out by Nordfors, this might formally look as if power is taken over by the interest organizations, while in fact the outcome from the corporatist institutions is controlled by the forces (in this case the labour movement) that have written the rules of the game. This should thus not be understood as the absence of government steering but as the opposite: a special steering technique, albeit a rather sophisticated and indirect one but perhaps still more efficient than any

possible alternative (Nordfors 1986:22; cf. Lundquist 1987, ch. 6). Furthermore,this technique can cause a split in the interest organizationbetween who want to influence what under given circumstances might be influenced, and 'fundamentalists', who want to follow the basic ideology of the organization without any compromise and at any cost, thus the old and well-known governing technique.

The Societal Level

At this level we are looking for explanations as to how it has been possible for these relations between the state and the social classes to be established at all. The reasoning of a government might be something like this: In this policy area A we have the policy P which, however, is a rather vague one and its actual implementation must necessarily vary a lot. The reason is that it might be impossible in practice to steer the actual implementation of the policy in question by any Weberian bureaucratic precise rules or regulations, because each and every case has to be judged according to its own specific circumstances (Offe 1985, ch. 10; Rothstein 1987b). Instead of giving this discretionary power to certain bureaucrats or professionals, an administrative structure arranged as a corporatist institution is established the government. By controlling the resources and the interest organizations that are to be represented, and also the number of representatives each organization, the government can be pretty sure of a policy output that is in line with its general intentions. The policy area is so to speak handled, but not governed, by the government (cf. Streeck 1984, 145 f.). In addition, the government gets 'responsible' behaviour from the organizations that represent the main 'policy takers' and their consent of the policy, or as Claus Offe has stated:

Formal admission of corporate groups to the process of public policy formation favours the production of decisions that minimize the probability that social power will be used in order to obstruct or resist public policy, because the actual power of labour and capital, respectively are already 'registered' and taken into account in the process of its formation (Offe 1985:246).

The reason for the interest organization accepting to participate might be that even if the general policy cannot be influenced in any substantial way, there is enough room for influence over the specific implementation of the policy to make partcipation worth while. Furthermore, even if the organization strongly opposes the general aim of the policy, a reason to participate in the process of implementation is that this might be a possibility to mitigate its concrete impact. Anyhow, the result of a refusal to participate might be that all the influence would be given to the opposing organization (in this case the union movement). I would say that this hardly gives the business interest organization any choice; it has to participate.

From a pure class interest, there is, I believe, no reason why a major capitalist organization should occupy itself with administrating the implementation of any Social Democratic welfare policy (Therborn 1986b). The only reason why it might be doing this seems to be that it has met a strong labour movement that intends to enforce such policies 'so oder so'. Confronted with such an opponent, business has had to choose between Scylla and Charybdis, to take part in the administration of Social Democratic and thereby legitimate them in order to gain a minor influence in the stage of implementation, or to refuse to participate, thereby risking a more severe implementation of the policies in question, maybe to the point of threatening fundamental capitalist interests. The government has of course paid a price too, i.e. giving business a substantial degree of influence over the implementation of some of its policies. But this price might not be so hard to pay because it is not certain that the government (or Parliament) actually wants to influence the specific details of the implementation process as long as it stays within the bounds of the general policy. The influence from business is anyway balanced by the representation the unions. Furthermore, by giving away some of the influence over the implementation of a policy, the government does not stand alone with the responsibility for the handling of each and every case (Offe 1985, 247). The steering ambitions of governments, and especially parliaments, should anyway not be overestimated.

Epilogue

In May 1987 the Government Bill concerning the composition of the NBAs' boards was discussed in the Swedish parliament. As could be expected, the Bill from the Social Democratic government did not follow the commission's proposals to abolish the interest organizations' representatives from the boards and place them in advisory committees. This aroused some intense public debate between the chairman of SAF and the Minister for Civil Service Affairs, the former publicly accusing the latter of creating a corporatistsociety ¹⁰ The Social Democratic government followed the statementsfrom and the national white-collar union (TCO), which strongly favoured the status quo (De C 1986-4). The three bourgeois parties were (as usual) divided in some important details, but they all wanted the interest organizations out of the NBAs' boards. The Bill emphasized that in labour market questions in particular, it was a great advantage to have the interest organizations' on the boards so that the decisions could be negotiated among those who were to be affected and that the government thus would not have to interfere (Government Bill 1986/87-99, 93). The political role of the NBAs as well as the existence of the 'framework laws' aroused

some debate. As one of the MPs from the Liberal Party (who is also a distinguished professor of political science, Mr Daniel Tarschys) said to a conservative MP: 'If we were to make the Work Environment Act more precise and take back to Parliament all the decisions that are now taken by the National Health and Occupational Safety Board, we would be standing up to our ears in chemical formulas' (Records from the Swedish Parliament 1986/87-122, 30, my trans.).

The Bill was of course approved by Parliament. SAF who, as mentioned above, strongly favoured the solution suggested by the Commission, has hitherto not withdrawn any of its representatives from the NBAs discussed above. And, I strongly believe, they are not likely to do so in the foreseeable future. A one-sided disarmament of corporatist representation is not likely to happen in the Swedish polity. To quote Mr Albåge once again: 'On the one hand we do not want a corporatist society. On the other hand, SAF does not want to lose its influence or withdraw its participation from the boards if the union confederations are there . . . This would be to handle our responsibility towards our members wrongly. This would only be possible if all parties were to withdraw' (interview Dec. 1,1987, my trans.).

---

NOTES

1. The language is a bit confusing. While many of the NBAs have the word 'Board' in their name, this usually means the whole administration (e.g. The National School Board). But the executive committees are also named 'boards', so one could actually speak of 'the Board's board'.

2. This system of having the reports from the government's investigatory commissions sent out to the major interest organization, asking them for their standpoint on the question on hand, is the common practice in the Swedish political system. The Commission's report, together with the statements from organizations and the NBAs, form the basis for the following government bill. For a more detailed account of this process see Heclo & Madsen (1987).

3. 'Allmånna verkstadgan' (General instructions for the NBAs), SFS 1965:600.

4. In 1981, SAF, together with the other class-based interest organizations, was expelled from the National Board of Education by the bourgeois government, a decision which aroused some great opposition from the Social Democrats. When back in power in 1982 the Social Democrats put the interest organizations back onto the Board's board. Government Bill 1980/81-107, Records from the Swedish Parliament 1980/81-156, 6 pp. and Government Bill 1983/84-100 no. 10 p. 81.

5. Calculated from the Government budget proposals, see Government Bill 1981/82:100 and 1982/83:100.

6. All from the National Labour Market Board (1977, 1981, 1982, 1983, 1984).

7. Calculated from 'Statliga myndigheter' 1975 and 1986. The level of depreciation of the Swedish currency is taken from Statistisk Årsbok (Statistical Abstracts of Sweden) 1987, Table 228.

8. I have chosen these two years so as to get as similar a budget and economic circumstances possible.

9. Eg. Mr Gunnar Lindstrom at the National Labour Market Board and Mr Matts Bergom-Larsson at the National School Board.

10. In the major conservative daily 'Svenska Dagbladet, March 27, 1987.

REFERENCES

Alford, R. R. & Friedland, R. 1985. Powers of Theory. Capitalism, the State and Democracy.
Cambridge: Cambridge University Press.

Amnå, E. 1981. Planhushållning i den offentlig sektorn. Stockholm: Norstedts.

Åsard, E. 1986. 'Industrial and Economic Democracy in Sweden. From Consensus to Confrontation',
Journal of Political Research 14.

Benson, J. K. 1982. 'A Framework for Policy Analysis', in Roger, D. L. & Whetter, D. A.,
eds., Interorganizational Coordination. Theory, Research, Implementation. Ames: lowa
State University Press.

Boucher, L. 1982. Tradition and Change in Swedish Education. Oxford: Clarendon.

Bresky, T. & Scherman, J. & Schmid, I. 1981. Med SAF vid rodret. Granskning av en
kamporganisation. Stockholm: Liberforlag.

Buksti, J. A. & Johansen, L. N. 1979. 'Variations in Organizational Participation in
Government. The Case of Denmark', Scandinavian Political Studies 3.

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

Coleman, W. & Grant, W. 1984. 'Business Associations in Public Policy: A Comparison of
Organisational Development in Britain and Canada', Journal of Public Policy 4.

Damgaard, E. 1981. 'Politiske sektorer. Jernhårde trekanter eller løse netværk', Nordisk
Administrativt Tidskrift 62.

Ds C 1985-16. Government report. The Swedish Ministry of the Interior. Fakta om statliga
myndigheters ledning.

Ds C 1986-4. Government report. The Swedish Ministry of the Interior. Remissyttranden
over verksledningskommitténs betånkande.

Elster, J. 1979. Ulysses and the Sirens. Studies in Rationality and Irrationality. Cambridge:
Cambridge University Press.

Elster, J. 1982. 'Marxism, Functionalism and Game Theory. The Case for Methodological
Individualism', Theory and Society 11.

Elster, J. 1983. Sour Grapes. Studies in the Subversion of Rationality. Cambridge: Cambridge
University Press.

Elvander, N. 1974. 'Interest Groups in Sweden', Annals of the American Academy of Political
and Social Science 413.

Friberg, L. 1973. Styre i kristid. Stockholm: Almånna forlaget.

Gaventa, J. 1980. Power and Powerlessness. Quiescence and Rebellion in an Appalchian
Valley. Urbana: University of Illinois Press.

George, A. L. 1979. 'Case Studies and Theory Development: The Method of Structured,
Focused Comparison', in Lauren, P. G., ed., Diplomacy: New Approaches in History,
Theory and Policy. New York: The Free Press.

Grant, W. 1985. 'Introduction', in Grant, W. ed., The Political Economy of Corporatism.
London: Macmillan.

Hadenius, A. 1978. 'Åmbetsverkens styrelser', Statsvetenskaplig tidskrift 81.

Hadenius, A. 1986. A Crisis of the Welfare State? Stockholm: Almqvist & Wiksell
International.

Hanf, K. Hjern, B. & Porter, D. O. 1978. 'Local Networks of Manpower Training in
the Federal Republic of Germany and Sweden', in Hanf, K. & Scharpf, F. W., eds.,
Interorganizational Policymaking. London: Sage.

Hansson, S-O. 1984. SAF i politken. En dokumentation av nåringslivsorganisationernas
opinionsbildning. Stockholm: Tiden.

Heckscher, G. 1946. Staten och organisationerna. Stockholm: Kooperativa forbundet.

Heckscher, G. 1958. 'The Role of Voluntary Organizations in Swedish Democracy', in
Lauwerys, J. A., ed., Scandinavian Democracy. Copenhagen: The Danish Institute.

Heclo, H. & Madsen, H. 1987. Policy and Politics in Sweden: Principled Pragmatism.
Philadelphia: Temple University Press.

Holmberg, S. 1984. Våljare i forandring. Stockholm: Liber.

Holmberg, S. & Gilljam, M. 1986. Våljare och val i Sverige. Stockholm: Bonniers.

Jerneck, M. 1987. SAFs framtidssyn: Forutsågelser, målsåttningar och dilemman. Stockholm:
SAFs forlag.

Jessop, B. 1979. 'Corporatism, Parliamentarism and Social Democracy', in Schmitter, P. C.
& Lembruch, G. eds., Trends Toward Corporatist Intermediation. London: Sage.

Kelman, S. 1981. Regulating America, Regulating Sweden. A Comparative Study of Occupational
and Health Policy. Cambridge, Mass: The MIT Press.

Korpi, W. 1983. The Democratic Class Struggle. London: Routledge & Kegan Paul.

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

Lembruch, G. 1985. Neocorporatism in Western Europe: A Reassessment of the Concept in
Cross-National Perspective. Working paper, IPSA XIII Congress, Paris.

Lipsky, M. 1980. Street-level Bureaucracy. New York: Russel Sage Foundation.

Lowi, T. J. 1979. The End of Liberalism. New York: Norton.

Lukes, S. 1977. Essays in Social Theory. New York: Columbia University Press.

Mann, M. 1986. 'The Autonomous Power of the State', in Hall, J. A., ed., States in History.
Oxford: Basil Blackwell.

March, J. & Olsen, J. P. 1984. 'The New Institutionalism. Organizational Factors in Political
Life', American Political Science Review 78.

Marin, G. 1985. 'Austria - The Paradigm Case of Liberal Corporatism?', in Grant, W., ed.,
The Political Economy of Corporatism. London: Macmillan.

Micheletti, M. 1984. The Involvement of Swedish Labour Market Organizations in the Swedish
Political Process. Working paper, Business & Social Research Institute, Stockholm.

Nordfors, L. 1986. State and Organized Interests. Abdication or Regulation? - the Case
of Swedish Work Environment Policy. Research report 3, the RiO-project, Dept. of
Government, University of Uppsala.

OECD 1987. Financial Accounts of the OECD countries. Paris: OECD.

Olsen, J. P. 1983. Organized Democracy. Political Institutions in a Welfare State - The Case
of Norway. Oslo: Universitetsforlaget.

Olsson, S-E. 1988. 'Social Welfare in Economically Advanced Countries: Social Service and
Social Security in Sweden', in Dixon, J. & Scheurell, R., eds., Social Welfare in Developed
Market Countries. London: Croom Helm.

Panitch, L. 1980. 'Recent Theorizations of Corporatism. Reflections on a Growth Industry',
British Journal of Sociology 31.

Pedersen, M. N. 1977. 'Om det rette brug af historiske materialer i statskunskaben', Festskrift
till Erik Rasmussen. Aarhus: Politica.

Peterson, E. A. 1977. Interest Group Incorporation in Sweden. A Summary of Arguments and
Findings. Working paper presented at the 1977 Annual Meeting of the American Political
Science Association, Washington D.C.

Petri, G. & Vinde, P. 1978. Government Administration in Sweden. Stockholm: Prisma.

Pontusson, J. 1987. 'Swedish Social Democracy, 1967-87: Radicalization and Retreat', New
Left Review 163.

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

Rothstein, B. 1985b. 'Managing the Welfare State: Lessons from Gustav Moller', Scandinavian
Studies 13.

Rothstein, B. 1986. Den socialdemokratiska staten. Reformer och forvaltning inom svensk
arbetsmarknads- och skolpolitik. Lund: Arkiv.

Rothstein, B. 1987a. 'Vålfårdsstat, implementering och legitimitet', Statsvetenskaplig tidskrift
90.

Rothstein, B. 1987b. 'Corporatism and Reformism: The Social Democratic Institutionalization
of Class Conflict', Acta Sociologica 30.

Rothstein, B. forthcoming. 'Aktor - struktur ansatsen: Ett metodiskt dilemma', Staisvetenskaplig

Ruin, O. 1974. 'Towards a Corporate Society? Participatory Democracy and Corporatism:
the Case of Sweden', Scandinavian Political Studies 9.

Scharpf, F. 1987. The Political Calculus of Inflation and Unemployment in Western Europe:
A Game-Theoretical Interpretation. Working paper, Max Planck Institute, Cologne.

Schmitter, P. C. 1979. 'Modes of Interest Intermediation and Models of Social Change in
Western Europe', in Schmitter, P. C. & Lembruch, G., eds., Trends Towards Corporatist
Intermediation. Beverly Hills: Sage.

Schmitter, P. C. 1981. 'Interest Intermediation and Regime Governability in Contemporary
Western Europe and North America', in Berger, S., ed., Organizing Interests in Western
Europe. Cambridge: Cambridge University Press.

Schmitter, P. C. 1985. 'Neo-corporatism and the State', in Grant, W., ed., The Political
Economy of Corporatism. London: Macmillan.

Scholten, I. 1987. 'Introduction: Corporatist and Consociational Arrangements', in Scholten,
1., ed., Political Stability and Neo-Corporatism. London: Sage.

Skocpol, T. 1985. 'Bringing the State Back in', in Evans, P. et al. eds., Bringing the State
Back In. Cambridge: Cambridge University Press.

Statliga myndigheter 1975 and 1986. Stockholm: Statskontoret.

Steeck, W. 1984. Industrial Relations in West Germany. London: Heinemann.

Svensson, L. G. 1987. Higher Education and the State in Swedish History. Stockholm: Almqvist
& Wiksell International.

Therborn, G. 1986a. Why Some People Are More Unemployed Than Others. London: Verso.

Therborn, G. 1986b. 'Class Analysis: History and Defence', in Himmelstrand, U., ed.,
Sociology from Crisis to Science? Vol. 1. London: Sage.

Therborn, G. 'Does Corporatism Matter? The Economic Crises and Issues of Political
Theory', forthcoming in Journal of Public Policy.

Wennerberg, U. 1983. De centrala åmbetsverkens lekmannastyreiser. Research report, Dept.
of Political Science, University of Stockholm.

Wilson, J. Q. 1973. Political Organization. New York: Basic Books.