Resumé

Recent literature suggests that Swedish politics is decorporatizing. This article reports on the results of a survey of Swedish interest groups active in social welfare policy, administered in 1986 and 1994. The results cast considerable doubt on the decorporatization thesis. Little change is discernible between the two surveys on indicators of key corporatist behaviours on the part of interest groups and government. The immediate goal of the article is to provide evidence that the Swedish system is not, in fact, decorporatizing to the extent cited in recent literature. A larger aim, however, is to push the study of corporatism in a more empirical direction. The value of the theory will be enhanced if researchers are more precise about what it means and what behaviours should follow from it.

Introduction

Sweden has long been considered a corporatist country (Grant 1985; Lehmbruch 1984; Schmidt 1982; Schmitter 1981; Strich 1974). Peterson (1977, 59), for example, states: "Corporate structures are particularly well developed in Sweden and their effect on interest group politics is probably more apparent there than in other political systems." Recent work argues, however, that Sweden is becoming less corporatist (Bergström 1991; Cawson 1982; Micheletti 1991; Petersson 1991; Rothstein 1988; Schmitter 1989).

Most authors at present view corporatism as a middle-range theory which is most applicable to economic policy. Some efforts have been made, however, to look at corporatism in other policy areas, including social policy (Cawson 1982; Foster 1983; Harrison 1984; Hoefer 1994; Mishra 1984). While interesting and somewhat useful, many of these efforts have fallen short. As Williamson (1989, 184) states:

. . . the empirical findings, and more importantly, the theoretical and conceptual questions
and leads that these can provide, are relatively thin on the ground.

This article examines whether changes have taken place in the Swedish
interest group system between 1986 and 1994, testing several hypotheses

derived from corporatist theory. The results indicate that, despite the considerable number of articles devoted to the demise of Swedish corporatism, the picture is more complex. Some elements of the Swedish model are becoming weaker, but others are maintaining themselves and a few are even becoming stronger. It is hoped that these findings will add to the thin layer described by Williamson and stimulate further conversation about the subject.

Literature Review

As mentioned above, many recent authors have argued that Sweden is becoming less corporatist than before. Cawson (1982, 85-86) notes that the move towards privatization of social services seen in Sweden is a move away from corporatism. Swedish social policy, traditionally controlled by the Social Democratic thesis that no profit should be made on providing for common human needs, has gradually shifted (Hoefer 1988). At present, private medical and child-care services are permitted. Managers of government-provided services are now in a state of transition, learning to compete for clients.

How far privatization proceeds is partially determined by the ideology of the national government, although local government also has considerable autonomy in this matter. The Conservative Party campaigned in 1991 on a platform of a "New Start for Sweden". Part of that "New Start" was clearly to dismantle major elements of corporatism. With the shift in national government due to the 1994 elections, however, there has been reduced emphasis on these types of possible changes, although movement at local levels of government may still be ongoing.

Rothstein (1988) examined the stated positions of the Swedish Employers' Confederation (SAF) and the votes its representatives made on the budget proposals of many government agencies, including social welfare bureaux. He found that, despite a clear policy on the part of SAF to decrease the size of government programmes, the SAF members on these quasi-governmental panels never once voted against raising agency budgets from one year to the next. Rothstein attributes this result to the power of corporatist arrangements to change the behaviour of powerful interest groups. In this case, SAF helped support more government spending (during both Social Democratic and Conservative coalition administrations). It is thus easy to understand why the Conservatives would like to decrease the level of corporatism in Sweden.

Micheletti (1991) argues that corporatism is a "fair weather phenomenon"
requiring special conditions to exist, and those conditions are disintegrating.
Schmitter (1989) speaks of three major changes which are

occurring in corporatist arrangements in European countries. First is the decreasing importance of national macro-economic decision-making as, for example, the European Community becomes a more important locus of policy determination. Second is the changing relationship between corporatist groups and their members with members demanding more democracy and direct control of their groups. Third is the change in the nature of issues from class-based to a variety of post-modern issues such as quality of life and environmental problems.

Petersson (1991) and Micheletti (1991), focusing on the Swedish situation, agree that the established way of doing things in Sweden has changed. Petersson, one of the principals on the Swedish Study of Power and Democracy, indicates that there have been cutbacks in the system of government-appointed commissions, leaving fewer opportunities for interest groups to exert influence. Further, more interest groups now exist, new power centres have emerged and the power structure has become more fragmented.

Micheletti contends that post-modern politics is making the system of corporatism untenable. The three changes taking place, post-modernization, post-industrialism and internationalism, are leading to a situation where "compromise and consensus in political decision-making is emphasized less today than in the past, and Sweden is gradually becoming decorporatized" (Micheletti 1991, 154).

The implications of this shift are profound, and several authors have stated that the "Swedish model", as traditionally understood, is unlikely to survive intact. Bergström (1991, 29) discusses how and why Sweden is drifting from its unique characteristics to become a "more ordinary Western European country". Hoefer (1988) found that post-materialism was a primary cause of the move towards privatization of government services. An increased amount of decentralization in decision-making is noted by Eländer (1991). Micheletti (1991, 163) states that:

The Swedish model of corporatism is an example of how the balance between organized
interests and the state was struck in the past. Its character was technocratic and elite
democratic. Now it is obvious that Sweden must strike a new balance. . . .

The data collected in this research project cast doubt on the conclusion that corporatism is dying, at least in the area of social welfare policy. At best, there are a few minor changes taking place. The reason for other authors' incorrect analysis of the situation is related to the lack of behavioural and empirical testing of corporatist theory. If Sweden is becoming less corporatist, one should expect to see changes in certain behaviours exhibited by government and interest groups. These changes are not happening to the degree one would expect if the literature on the corporatist system in Sweden was completely correct.

Research Methods and Hypotheses

The information for this analysis comes from a survey administered to Swedish interest groups active in social welfare policy in 1986 and 1994 (before the election). The list of organizations sent a survey was created in 1986 by examining all legislative proposals emanating from the Department of Social Services during the period 1982-86. In these government documents, all interest groups which had responded to a request for comments to Royal Commission proposals (used the remiss process, in other words) are listed. Addresses were collected and surveys were sent to these organizations. The 1994 list of groups included all of the 1986 groups for which a current address could be found and a number of other groups listed in the Stockholm "Yellow Pages". The 1986 survey was sent to 92 groups with a response rate of 73 percent. The 1994 survey was sent to 102 groups, with a 60 percent response rate.

Before continuing, it is important to note the difficulty of providing a single definition of the term corporatism. To date, there is no clear consensus regarding what corporatism is, how to measure it, whether it is simply a variant of pluralism, or on many other important areas (see Martin 1983; Panitch 1980; Williamson 1989; among others). According to Micheletti (1991, 148-149):

A minimal definition of corporatism is an integral relationship for decision-making and implementation between the state and encompassing interest organizations. Encompassing interest organizations represent the greater part of the collectivity which they can organize as members. For Sweden, the distinguishing characteristic of corporatism is the incorporation of antagonistic, encompassing interests in state policymaking.

Corporatism is an often used term, but there have been few attempts to tease out testable hypotheses concerning the theory (one recent exception is Hoefer (1994)). Using Micheletti's definition, the data from these surveys allow us to test the degree of corporatism found in Sweden, based on hypotheses derived from theory and recent literature. We can thus note in some detail what changes, if any, have occurred in the past few years.

Assuming that reports of change in the Swedish political system are accurate, we should be able to see differences in the relationship of the central government to interest groups in a few key areas. Micheletti (1991, 154), as noted earlier, argues that a decline in the use of "compromise and consensus" in policy-making has occurred. If so, we should see this reflected in several behaviours measured on the survey. Second, the role of "encompassing organizations" vital to corporatism should be diminished if the system is "decorporatizing."

Several other hypotheses based on corporatist theory are also tested in
this article. First, a corporatist system should have some interest groups
which have a close relationship with the government. Usually, one would

expect these groups to be among the largest peak organizations in the country (Micheletti's "encompassing organizations"), because they have the ability to represent and "control" the largest number of persons. In Sweden, because of its allegedly relatively open style of corporatism (Peterson 1977), the number may be larger or the type more diverse than in other corporatist countries. This close group-government relationship might be seen in several observable ways: (1) by being consulted regularly by the government; (2) by having a steady relationship with the government, no matter which party or party coalition is in power; (3) by a lack of partisanship in group behaviour, since the group must work with whoever is in power; and (4) by groups receiving relatively large amounts of funding from the government because these groups are functioning in a quasigovernmental way.

In addition, under corporatism, we should expect that groups are not particularly competitive with each other. Each has a specific niche, either created by government or social forces, and competition for members or other resources should be subdued. If, as Micheletti suggests, compromise and consensus (the hallmarks of corporatism) are declining in Sweden, we should see changes in many of these characteristics.

Finally, if corporatism is in fact less important than it used to be, we should see that groups' ratings of different ways of influencing policy should reflect a decrease in the importance of corporatist ways of policy-making such as Royal Commissions and the remiss system.

Results

This section examines the areas noted in the previous section as being areas where differences in behaviour should be evident if Sweden is indeed becoming less corporatist. The first area is in compromise and consensus, the second is the role of encompassing organizations, the third is groupgovernment relations, the fourth is the level of competition groups feel towards each other and the fifth is the ratings of where in the political process groups' influence can best be used.

Compromise and Consensus

The survey provides us with several questions related to the use of compromise and consensus by interest groups in the Swedish social welfare policy arena. Two of these are essentially mirror images of each other. The first asks to what degree the respondent feels that the social policy area is characterized by intense conflict; the second asks to what degree the respondent feels that the social policy area is characterized by consensus.

In Table 1 it is shown that the tendency between 1986 and 1994 is for
groups to feel less conflict and more consensus, although these changes
were not statistically significant.

Two tactics for influencing policy which could lead to understanding whether compromise and consensus were becoming less important are the use of public protests and demonstrations, which would give the impression that less consensus existed, and trying to develop consensus among experts before moving forward with policy proposals. Table 2 reports the percentages of groups that feel that these tactics were important or very important in achieving their policy goals, as well as whether the group uses that tactic at all. Once again, the results show no statistically significant changes between 1986 and 1994, thus indicating that Micheletti's comments may not be fully accurate. In short, there does not seem to be less cooperation and consensus in social welfare policy now compared to the mid-1980s. The trend, though not statistically significant, as shown in Table 2, is in the opposite direction. More groups "never use" the conflictual approach of protests and demonstrations in 1994 than in 1986, while the percentage who consider it important or very important has fallen by a small amount. The number of groups who never try to develop consensus among experts has decreased slightly (that is, more groups use this tactic in 1994 than they did in 1986) and the percentage of groups who believe it is an important or very important tactic has increased by one-half, from 20 percent to 30 percent.

The Role of Encompassing Organizations

Micheletti uses the term "encompassing organization" without any definition being offered. This concept is operationalized here as being an organization whose members are other organizations. Thus, LO or SAF are considered encompassing organizations, whereas the members of these federations probably are not. This idea is presumably at least close to what

one would mean by the term "encompassing" organization (such groups
are also sometimes known as "peak" organizations).

The overall message of this section is that the role of encompassing organizations is not diminishing. The differences found between 1986 and 1994 are not statistically significant and, in fact, the numbers are often not even moving in the direction expected if decorporatization is happening. In 1986, for example, 88 percent of the 24 peak organizations that responded indicated that they were regularly consulted by government; in 1994 the percentage was 94 percent of the 17 respondents (chi-square = 0.494, p = 0.482).

When asked how, if at all, their efforts to influence policy had changed in the last five years, most groups in both years indicated that there had been no change (56 percent in 1986 vs. 64 percent in 1994). The 1986 figures show that 44 percent increased the number of influence attempts compared with 36 percent who had increased their efforts in 1994. In neither year did any peak organization say that it had decreased its influence efforts (chisquare = 2.866, p = 0.413).

Peak groups are in contact with about the same number of government agencies in 1994 compared with 1986, with the vast majority at both times interacting with more than five agencies (85 percent in 1986 and 88 percent in 1994). The frequency of interaction seems to have decreased, however, with the agency with which they had most contact. In 1986, 80 percent said they interacted "frequently" compared with only 71 percent who said that in 1994. All other groups indicated that they interacted "occasionally" with the government agency (chi-square = 0.442, p = 0.506).

Group-Government Relations

This section looks at all organizations, not just the encompassing organizations

of the previous section. Changes in the way that groups relate to
government are examined. We find that there are no significant changes
between 1986 and 1994.

Similar to the high numbers of peak organizations consulted by government, the percentage of all responding groups which are regularly consulted by government is very high. In 1986, 90 percent of organizations were consulted with and this increases in 1994 to 95 percent. These numbers indicate that whatever exclusivity might once have existed in governmentgroup consultations is now gone.

An important element in corporatism is the extent of cooperation between important interest groups and government. Cooperation between these actors should remain fairly stable no matter which party controls the government, according to corporatist theory. Table 3 provides information regarding how much change there was in the relationship of groups to government after an election where the ruling party changed. The 1982 election returned the Social Democrats to the leading role after six years of right-wing government. The 1991 election put the Conservative Party and its allies in control again.

In both cases, most groups report no change in their relations with the government, despite a seemingly significant change in the policies that the government might be expected to pursue. Those groups that do report a change after an election generally seem to find better relations. Increased conflict is indicated, however, by the larger percentage of groups which perceive worse relations after the 1994 election than after the 1982 election.

Another question addresses this same issue: 28 percent of Swedish groups in 1986 reported that "Some elected officials oppose my group's goals" vs. 39 percent in 1994 (chi-square of 3.99, p = 0.551). This indicates increased disagreement and more contention between groups and government and is consistent with the information in the previous paragraph.

Assuming that many interest groups active in the social welfare policy

area would be pushing for more liberal benefits or programmes, a Social Democratic government could signal better chances for such changes than a right-wing government. Although these non-statistically significant results are in the direction of decreasing cooperation between government and groups and thus can be interpreted as evidence of decorporatization, I believe another explanation is more likely. This set of interest groups has more in common with the Social Democratic Party than the Conservative Party, so relations have suffered. At any rate, the changes measured are not great, so it is difficult to conclude that decorporatization is occurring.

If government-group relations are worse, as the decorporatization thesis argues, one would expect cuts in government funding of groups. Table 4 indicates the levels of funding provided by the government in the two time periods. The government is providing considerably more of the budget for many groups in 1994 than in 1986. The percentage of groups receiving more than half of their budget from government sources tripled between 1986 and 1994 (from 10 percent to 30 percent of all respondents) (chi-square of 33.038, p = 0.06). If one looks at the mean percent of government funding instead of using these categories, the results are even stronger. In 1986, the mean level of government support was 16 percent (s.d. = 23.8) while in 1994, the mean was 29 percent (s.d. = 29.9) (p = 0.01).

The clear conclusion from this section is that group-government relations are not significantly worse than before, and much of the change in their relations may be due to the nature of the party in power, not changes in the level of corporatism.

Level of Competition Between Groups

One of the outcomes of decorporatization would be increased conflict and competition between interest groups. Contrary to this expectation, the information in this section shows that there is not more competition between groups. Table 5 shows that there was no change on this dimension between 1986 and 1994 (chi-square of 0.248, p = 0.88). Over half of the Swedish

groups indicate that there is "little" competition. Less than one-quarter of the groups state that there is "much" competition, while one-quarter say they experience some competition. These numbers have remained impressively constant.

Group leaders were also asked if they believed that other groups opposed their group's policy goals. There was a very small increase in the number of groups saying that this was a good or very good description of their situation (32 percent in 1986 compared with 35 percent in 1994, chisquare = 4.151, p = 0.528).

Where in Political Process Influence Can Best Be Used

The final section of results looks at where in the political process group leaders believe they can best use their influence. The Swedish system can be broken into five stages. First, Royal Commissions are appointed to conduct research and make recommendations for laws. Second, the remiss process allows all interested parties to make comments on the Royal Commission's recommendations. Third, the affected Ministry and the Government prepares and makes final changes to the proposed law, taking into account the comments forwarded to them during the remiss process. Fourth, the Parliament votes on the Government proposal. Finally, the law is implemented by the Bureaucracy. (For more detail on the role of interest groups in the Swedish political system, see, for example, Pestoff (1983)).

Respondents were asked to rank these parts of the policy-making process in terms of where they were able to exert influence. If corporatism is breaking down, then the traditional corporatist parts of the political process such as Royal Commissions and the remiss system should be seen as less important than before. Table 6 shows that this is not the case, in absolute or relative terms. There are no statistically significant differences between the 1986 and the 1994 surveys for any of the variables.

Royal commissions are considered the place where groups can most use
their influence, followed by use of the remiss process, in both 1986 and

1994. These two elements of the Swedish law-making system are most closely linked to corporatism. There is one shift in relative ranking between the two years. The fourth and fifth ranked stages (Parliament and Bureaucracy) change order.

Another set of questions presents seven different activities that groups can engage in to achieve their policy goals. Respondents are asked to rate the importance of each activity. These results also show no statistically significant changes between 1986 and 1994 (see Table 7).

While there is a little discrepancy between Tables 6 and 7 regarding the importance of commissions vis-å-vis remiss submissions, the important point for this article is that there are none of the changes one might expect if corporatism were fading in Sweden.

Conclusion

This article looks empirically at the supposed demise of corporatism in

Sweden and finds that there are few, if any, grounds to believe that this process is occurring. In area after area, with variable after variable, there are no changes in behaviour that should see change, assuming decorporatization is happening.

What possible explanations could account for these results? First, one could say that the variables looked at are not the important ones to examine. While this may be true for some of the variables, it is difficult to believe this is the primary reason because the variables chosen are tied into corporatist theory and such a large number of them are used.

Another explanation may be that 1986 is too late for the first measurement. Many of the articles describing the end of corporatism in Sweden were already written by the mid-1980s. And with the typical time-lag between writing and publishing, it is clear that many of the elements that are destabilizing corporatism already had a lot of time to operate before 1986. This explanation is probably partially correct but does not explain the entire situation either. First, Sweden appears to be still a very corporatist country. Second, if decorporatization is a process, and the underlying elements causing it have not disappeared, the process should be continuing. Thus, change should still be happening, yet the measurements presented in this article show very little sign of it.

One other possible explanation is that the later time-point in this timeseries is too premature. Change is happening to the Swedish system as a whole but it has not yet penetrated throughout the system. A few glimpses have been spied by perceptive observers but the empirical measures used in this article are too crude to pick up these elements. This reasoning may also be partially correct but misses the crux of the argument. We should not be so eager to rush into making judgements without adequate empirical evidence.

A final possible reason for the findings might be that social welfare policy is different from other types of policy and is not losing its corporatist character as quickly as other types of Swedish policy, especially economic policy. As pointed out by Hoefer (1994), social welfare policy in Sweden is marked by elements of both corporatism and pluralism. This argument is also perhaps at least partially true. Nonetheless, it is not a very telling argument because the articles discussed in the literature review do not make any claim that their observations are true for only one segment of the polity. At the very least, the data in this article show that more precision needs to be employed when arguing that the Swedish political system is changing.

The immediate goal of this article is to provide evidence that the Swedish system is not, in fact, decorporatizing as is argued in recent literature. A more wide-ranging aim, however, is to push the study of corporatism in a more empirical direction. The debate about corporatism is no longer on

the cutting edge of political science, at least in part because so little theorytesting empirical work has ever been conducted on the subject. If there is any value to the theory, and I believe there is, it is important to be more precise about what it means and what behaviours should follow from it. Only then can the discussion about the strengths and weaknesses of the theory resume with any meaning.

ACKNOWLEDGEMENTS

The author expresses his thanks to the School of Social Work at the University of Texas at Arlington's Research Enhancement Fund for funding the 1994 survey and to Erik Amnå, Novemus School of Public Affairs, Örebro, Sweden for his comments on an earlier draft of the article. Suzie Haslinger, 8.5.W., handled the details of the 1994 survey with aplomb. The American-Scandinavian Foundation and the Fulbright Commission funded the 1986 survey.

REFERENCES

Bergström, H. 1991. "Sweden's Politics and Party System at the Crossroads", West European
Politics 14, 8-30.

Cawson, A. 1982. Corporatism and Welfare: Social Policy and State Intervention in Britain.
London: Heinemann Educational Books.

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

Eländer, I. 1991. "Analyzing Central-Local Government Relations in Different Systems: A
Conceptual Framework and Some Empirical Illustrations", Scandinavian Political Studies
14, 31-58.

Foster, P. 1983. Access to Welfare: An Introduction to Welfare Rationing. London: MacMillan.

Grant, W. 1985. "Introduction", in Grant, W., ed., The Political Economy of Corporatism.
London: MacMillan.

Harrison, M. 1984. "Corporatism, Incorporation and the Welfare State", in Harrison, M.,
ed., Corporatism and the Welfare State. Aldershot: Gower.

Hoefer, R. 1994. "Corporatism, Pluralism and Swedish Interest Group Influence in Social
Welfare Policymaking", Governance 7, 165-181.

Hoefer, R. 1988. "Post-Materialism at Work in Social Welfare Policy: The Swedish Case",
Social Service Review 62, 383-395.

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

Martin, R. M. 1983. "Pluralism and the New Corporatism", Political Studies 31, 86-102.

Micheletti, M. 1991. "Swedish Corporatism at a Crossroads: The Impact of New Politics and
New Social Movements", West European Politics 14, 144-165.

Micheletti, M. 1990. "Toward Interest Inarticulation: A Major Consequence of Corporatism
for Interest Organizations", Scandinavian Political Studies 13, 255-276.

Mishra, R. 1984. The Welfare State in Crisis. Brighton: Wheatsheaf.

Panitch, L. 1980. "Recent Theorizations of Corporatism: Reflections on a Growth Industry",
British Journal of Sociology 31, 159-187.

Pestoff, V. 1983. The Swedish Organizational Community and its Participation in Public Policy-Making: An Introductory Overview. University of Stockholm, Department of Political Science, Research Report No. 6, The Associative Action of Swedish Business Interests, October.

Peterson, E. 1977. "Interest Group Incorporation in Sweden: A Summary of Arguments
and Findings", paper, Annual Meeting of the American Political Science Association,
Washington, DC.

Petersson, O. 1991. "Democracy and Power in Sweden", Scandinavian Political Studies 14,
173-191.

Rothstein, B. 1988. "State and Capital in Sweden: The Importance of Corporatist Arrangements",
Scandinavian Political Studies 11, 235-260.

Schmidt, M. 1982. "Does Corporatism Matter? Economic Crisis, Politics and Rates of
Unemployment in Capitalist Democracies in the 19705", in Lehmbruch, G. & Schmitter,
P., eds., Patterns of Corporatist Policy-Making. Beverly Hills, CA: Sage.

Schmitter, P. 1989. "Corporatism is Dead! Long Live Corporatism!", Government and
Opposition 24, 54-73.

Schmitter, P. 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.

Strich, T. 1974. "Introduction", in Pike F., & Strich, T., eds., The New Corporatism. London:
University of Notre Dame.

Williamson, P. 1989. Corporatism in Perspective: An Introductory Guide to Corporatist
Theory. Newbury Park: Sage.