Power structures among political parties and interest groups. The view from Switzerland.

Transcription

1 Power structures among political parties and interest groups. The view from Switzerland. Manuel Fischer Pascal Sciarini Denise Traber Department of Political Science, University of Geneva Paper prepared for the "Jahreskongress der Schweizerischen Vereinigung für Politische Wissenschaft", , St. Gallen Very first draft. Please do not cite without permission. Abstract This paper deals with the power structures among political parties and interest groups in political decision-making processes in Switzerland. In four of the most important decisionmaking processes of the last years, the relations between these two most important groups of intermediary, non-state actors are analyzed along two dimensions: Their share of power and their positions along the major conflict lines. Two potential influence factors are taken into account. On the one hand, as there is not the one and only power structure in one country, we expect differences between policy domains, namely social and financial policy. On the other hand, we expect the degree of Europeanization of an act to influence its power structure. Our empirical results show that interest groups are still very important, even in cases of financial policy, where we expected political parties to be more influential. Interest groups and political parties line up along two major conflict lines: the classical left-right division in domestic cases and the openness-traditions division in Europeanized cases. 1

2 1 Introduction The Swiss direct democratic system grants interest groups with a strong power in the decision-making process. First, interest groups can launch referendums and they can take part in referendum campaigns. Second, and more importantly, direct democracy has favoured the development of a strong pre-parliamentary phase, in which interest groups are invited and where they can voice their policy preferences. The fact that economic interest groups are ex officio represented in a major part of the federal expert commissions underlines their strong standing (Linder 2005: 117). This, together with other historical and institutional reasons (earlier development of interest groups, federalism and weakness of the central state), accounts for the strength of interest groups. On the other hand, political parties are weak in Switzerland, this again due to a mix of historical and institutional reasons such as the bottomup character of the Swiss party system, federalism, direct democracy, the militia system and the resulting scarcity of resources (Kriesi 1998; Ladner 2006). Because of their weak position, political parties are dependent on interest group's financial and expert support. This phenomenon together with personal linkages creates strong ties among the two actors, so that relations between a party and its associated interest group are often closer than its relations with other parties (Kriesi 1980, 1998). In general, there exists a high degree of interpenetration between the system of interest associations and the party system (Kriesi 1982: 155). According to the traditional view, interest groups clearly outweigh political parties in the policy process (e.g. Kriesi 1980, 1998; Linder 2005; Kriesi and Trechsel 2008). Recently however, a debate about the roles and weight of political parties and interest groups in social policy reforms has taken place (see below: Häusermann et al. 2004; Fischer 2005; Kriesi and Trechsel 2008). Even if interest groups and political parties are widely recognized as being among the most important actors in a political system, there exists only a small body of research on their relations. The literature has for the most part treated parties and interest groups separately (Thomas 2001). For instance, the question was largely ignored by neo-corporatists, as they focus on tripartite relations between government, business and labor, but not on political parties. 1 Conversely, the literature on consociationalism focused mainly on political parties and did not take into account the role of interest groups. The absence of clear statements on the party-group relations also holds for general assessments of the political system. Kriesi (1994: 15, 2007: 264) defines political parties and interest groups as part of the intermediary system and specifies the relationships of these actors with the citizens on the one hand and the 1 One important exception is given by Katzenstein (1985: 94ff.) when he discusses the role of political parties in small, corporatist countries. 2

3 state actors (legislative body, administration) on the other hand. However, nothing is said about the relations between political parties and interest groups. Even if there exist some scientific contributions, such as the study of the classic, close ties between socialist parties and unions (Allern et al. 2007), of formal, common institutions (Thomas 2001), or of interpersonal ties between political parties and interest groups (Mcmenamin 2004), the field is largely understudied. Given the importance of these two actors groups in the political process, this is not only surprising, but could have important consequences on our understanding of the political decision-making process. Our paper will contribute to the knowledge about the relations between these two actor types by studying four of the most important decision-making processes in Switzerland in the last few years. On the one side, it compares two cases of social policy, the 11 th pension scheme reform and the extension of the free movement of persons with its related flanking measures, to two cases of financial policy, the program of budget relief 03 and the bilateral agreement on the taxation of savings. On the other side, it compares two cases of domestic policies, the program of budget relief 03 and the 11 th pension scheme reform, to two cases of Europeanized policy, the free movement of persons with its related flanking measures and the bilateral agreement on the taxation of savings. Power structures between political parties and interest groups are analyzed along two dimensions, which are the distribution of power and the major conflict lines. For this purpose, we rely on social network data, which was gathered through interviews. The paper is structured as follows. Chapter two introduces the concept of power structure. On the one hand, we elaborate on the institutional differences between policy domains as well as on their consequences, on the other hand we discuss the influence of Europeanization on power structures. Towards the end of chapter two, we formulate hypotheses. Chapter three presents the data, the operationalization and the method. Analysis follows in chapter four. Chapter five concludes. 2 Theory and hypotheses Dimensions of power structures This paper analyzes and attempts to explain the power structure among political parties and interest groups. Power structures describe the specific system of decision-making and are important because of their consequences on the elaboration of a political decision (see Atkinson and Coleman 1989: 50; Kriesi 1994: 21). Power structures specify the relationships in a policy network. The term of policy network should be considered as a general, 3

4 overarching concept describing the relations between political actors in a policy domain (Atkinson and Coleman 1989; Schneider 1992; Kriesi et al. 2006a) or in the political system as a whole (Kriesi 1994, 2007). These relations between political parties and interest groups are generally considered as being complex and fuzzy (Thomas 2001). By using the concept of power structures, we aim at bringing in some order and for this concentrate on two main dimensions of power structures. These dimensions are, on the one hand, the distribution of power, and, on the other hand, the major conflict lines. Together, these two dimensions have proven to be useful for the assessment of power structures (see e.g. Knoke et al. 1996; Fischer et al. 2002; Fischer et al. 2003; Sciarini et al. 2004; Kriesi et al. 2006a). On the one hand, the distribution of power among the different groups of actors gives us information about the influence the different actors have on the process of policy-making. As stated by Wilson (1990: 156), interest groups and political parties are often seen as rivals. It is thus argued that when parties are strong, interest groups are weak, and vice versa. Our assessment of the influence of these two actor groups will shed some light on this assumption. On the other hand, we will look at the most important conflict lines ("opposition structure", Knoke 1990), in the cases under study. This will allow us to reveal patterns of ideological affinity among political parties and interest groups, as well as between them. Let us now turn to two important factors that we suppose to impact on the power structure between political parties and interest groups. Differences across policy domains According to several authors (e.g. Laumann and Knoke 1987; Atkinson and Coleman 1989; Waarden 1992; Knoke et al. 1996; Kriesi et al. 2006a), the analysis of power structures at the level of the political system as a whole is no longer appropriate. A characterization on the national level is too crude to take into account the very important differences between different policy domains and could even lead to wrong conclusions about the political system (see Blom-Hansen 2001). It is thus fair to assume that no single pattern of relations between political parties and interest groups exists within a country (Kriesi 1998: 266; Thomas 2001: 2, 13) and that the strength of interest groups and political parties varies significantly across policy domains (Linder 2005: 117, based on Kriesi 1980). Following a neo-institutionalist view, different institutional settings and competencies in different policy domains may influence the power structures (see e.g. König and Bräuninger 1998). We will discuss differences between two policy domains, namely social policy and financial policy. 4

5 First, we discuss the distribution of power. According to the traditional view, interest groups clearly outweigh political parties in the policy process (e.g. Kriesi 1980, 1998; Linder 2005; Kriesi and Trechsel 2008). Concerning social policy, recent studies challenge this traditional view of the dominance of interest groups and suggest that things might have changed. Since the 1990s, retrenchment pressure, ideological polarization and the emergence of new social demands made it very difficult to reach a compromise in the pre-parliamentary phase, traditionally dominated by interest groups. According to Häusermann et al. (2004), this leads to a reinforcement of the parliament in the decision-making process. They argue that, compared to the corporatist actors, parties were able to encompass wider interests, and therefore gain more influence in the policy process. Similarly, Fischer (2005: 264f.) asserts that with a stronger polarization and conflict between social partners, the compromise is no longer found in the pre-parliamentary phase. Rather, the corporatist actors use the preparliamentary bargaining to state their preferences and signal intransigence. The compromiseseeking and majority-building has been postponed to the parliamentary phase, which gives the parties more weight in the decision-making process. Unlike Häusermann et al. (2004) however, Fischer (2005) does not believe in a diminishing role of the interest groups and points to the fact that they remain powerful in the Swiss decision-making process because they, first, have a strong referendum power, and second, have kept strong ties with their potential coalition partners in the parliamentary arena. Adhering to Fischer's (2005) argument and in line with the traditional view, we expect interest groups to be more powerful than political parties in the domain of social policy. Financial policy, on the other hand, is a domain which usually falls under the control of the legislative body and, therefore, under the control of the political parties represented in parliament. In fact, as stated by Lüthi (2006: 138) and Linder (2005: 210), the parliament has important competencies, and actively makes use of them by modifying the budget proposed by the Federal Council. Additionally, as the annual federal budget as well as resolutions on extraordinary expenditures are made in the form of a simple Federal Decree, they are not subject to an optional referendum. This fact significantly weakens the influence possibilities of interest groups. For these institutional differences, we expect political parties to be more powerful than interest groups in the domain of financial policy. Let us now briefly discuss the second dimension of power structures, the conflict lines. Here, we do not expect major differences between the two policy domains under study. Both social policy and financial policy are expected to be dominated by a classical left-right division. As claimed by Thomas (2001: 16), in policy domains with strong ideological cleavages and a 5

6 major capital-labor division, political parties and their corresponding interest groups defend very close positions and collaborate intensively. We thus expect the left-right conflict line to be the most important one for political parties and interest groups. On the one side, actors on the left such as the Socialist Party or the unions (SGB, Travail.Suisse) are generally in favor of more state intervention and defend the earlier achievements of social security (Armingeon 2006: 688; Bonoli 2006: 801). On the other side, the bourgeois parties (SVP, FDP, CVP) as well as the employers associations (Economiesuisse, SAV, SGV) support fiscal discipline and financial consolidation and therefore reductions of social policy offers (Armingeon 2006: 688; Bonoli 2006: 801f.). Bonoli (2006: 802) argues that there is an additional cleavage on the right side between political parties and employers associations. As they are exposed to political competition, political parties should generally be more reluctant to downsize social security institutions, as these are widely appreciated among the electorate. Interest groups, not exposed to electoral competition, tend to defend more rigid positions. We expect that in both policy domains, the major conflict line is given by the classical left-right division. Additionally, we expect the cleavage between political parties and interest groups, mainly on the right side of the political spectrum, to have certain, however minor, importance. Europeanization As a second factor, we expect the degree of Europeanization to affect the structure of partygroup relations. Despite not being a member of the European Union (EU), Switzerland is strongly influenced by the decisions made at the EU level (Fischer et al. 2002; Mach et al. 2003; Sciarini et al. 2004; Fischer 2005). Sciarini et al. (2004) show that Europeanization does not only affect the substance of public policies, but also the polity (institutions) and politics dimensions (power relations, conflict lines). As for the differences between policy domains, different institutional settings and competencies influence the power structure. However, the Europeanization literature focuses mainly on the power structures between state executive actors on the one hand and society actors on the other hand. Political parties and interest groups are both generally treated as society actors and the relationship between these two actor types is not clearly assessed. Let us first consider the dimension of the distribution of power. In line with the intergovernmentalist argument (Moravcsik 1994), Europeanization leads to the empowerment of state actors because they are directly involved in the international negotiations (Sciarini et al. 2004: 358). At the national level, the strong take-it-or-leave-it character of legislative acts arising from the international arena reduces the room for maneuver of the legislative 6

7 body. Political parties are thus expected to suffer from a loss of power compared to purely domestic decision-making processes (Fischer et al. 2002: 148). In an ordinary decisionmaking process, formal consultation procedures in the pre-parliamentary phase represent a major opportunity for interest groups to voice their preferences and influence the project. Due to the often restricted timetable which governs international negotiations, these formal and very open procedures are often replaced by more informal, selective, consultation mechanisms (Fischer et al. 2002; Sciarini et al. 2002; Sciarini et al. 2004). This enhances the discretionary power of the government, and at the same time the selective nature of these informal consultations favors the few most important interest groups with a high referendum power. Especially business and employers' associations from the export-oriented sector of the economy are expected to gain influence, because their preferences are in line with the liberalization tendencies at the European level and because of their competencies in the international arena (Sciarini et al. 2004: 359). Additionally, the fact that the most important economic interest group of Switzerland, Economiesuisse, is directly represented in Brussels, does clearly underline its international standing and its strong position in Europeanized decision-making processes (Fischer et al. 2002: 148). However, trade unions can exploit the fact that liberalization treaties with the European Union are of vital interest for Swiss business associations. By using the referendum threat, they can enforce some of their preferences, for instance the social protection of workers (Fischer 2002, 2003). Thus, concerning the dimension of power distribution, we expect the most important interest groups from the right as well as from the left (Economiesuisse, SAV, SGB) to be more powerful than political parties in the Europeanized cases. The second dimension of power structures as defined above is the conflict structure. Concerning interest groups, Europeanization is expected to exacerbate the conflict between the export-oriented and the domestic sectors of the economy: Representatives of the internationalized, exposed sectors of the economy have been increasingly reluctant to subsidize the domestic economy and to compensate it for the costs of increased economic openness ("side payments", Sciarini 1994; Bonoli and Mach 2000; Fischer 2005; Mach 2006). Thus, there is an accentuation of the division between the associations representing the interests of the sheltered sector, like the farmers (SBV) and the small businesses (SGV), and those of the exposed sectors (Economiesuisse, SAV, SBVg). One observes a similar conflict within the party system, which lead to an increasing division between the moderate and the nationalist-conservative camps (Kriesi and Trechsel 2008: 112ff.). There is a growing conflict within the party system on the openness-traditions dimension. Issues regarding the opening up 7

8 of the country to Europe, to foreign workers or to asylum seekers, as well as cultural conservatism, have pitted the Swiss People's Party (SVP) against the political left and the more moderate right (Brunner and Sciarini 2002; Kriesi et al. 2006b; Kriesi and Trechsel 2008). Thus, we expect the openness-traditions conflict line to be the most important one in Europeanized issues, because these questions concern the opening-up of Switzerland towards the EU. This division should be observable for political parties as well as for interest groups. In non-europeanized projects, the classical left-right division is still expected to be more important. We now briefly summarize our hypotheses developed above. Concerning the first dimension of power structures, which is the distribution of power, we expect the following: 1. In the domain of social policy, interest groups are more powerful than political parties. In the domain of financial policy, political parties are more powerful than interest groups. 2. In Europeanized processes, interest groups are more powerful than political parties. In domestic processes, political parties are more powerful. The second dimension of power structures is defined by the conflict lines. The following hypothesis has been derived from the discussion: 3a. In Europeanized processes, the major conflict line, for both political parties and interest groups, is along the openness-traditions dimension. In domestic processes, the classic left-right dimension is more important. 3b. Differences between policy domains have no influence on the conflict lines. 3 Data and methods Selection of cases Our study relies on a comparison of four policy domains, each represented by one of the most important policy processes in the last years. 2 In order to assess the impact of different institutional settings in different policy domains, we analyze two cases of social policy and two cases of financial policy. To study the second influence factor, we look at one process in domestic politics and one process in foreign (European) policy for each of the two policy domains. The two processes in the field of social policy are the 11 th pension reform (domestic 2 According to an exhaustive expert survey for all processes that passed a parliamentary vote between December 2001 and December

9 case) and the extension of the agreement on the free movement of persons with the EU and related flanking measures against wage dumping (Europeanized case). Our two processes in the field of financial policy are the program of budget relief 2003 (domestic case) and the bilateral agreement with the EU on the taxation of savings (Europeanized case). Table 1 presents the selection of cases. Table 1. Case selection Social policy Domestic policy 11th pension reform Financial policy Program of budget relief 03 Europeanized policy Free movement of persons and flanking measures Taxation of savings The logic of our case selection follows a "most similar systems design" strategy which allows controlling for a certain number of possible independent variables, while isolating the important factor of different institutional rules in different policy domains. As stated by Blom-Hansen (2001: 393), the comparison of different policy domains within one country already has the advantage that some important factors are held constant. By comparing the domain of financial policy to the domain of social policy, we treat two domains which are thematically very close to one another and both belong to the broader field of economic policy. We now briefly present the four cases under study. 11th Pension reform With the 11th pension reform the government s intention was to adapt the Swiss pension system to the changing social and economic circumstances. Its main elements were the financial consolidation of the AHV (first pillar of the Swiss pension system) as well as a social flexibilization of the retirement age. Among other measures, the government proposed an increase in the value added tax, an increase of women s pension age as well as an adjustment of the widow s and the widower s pension level to secure the financial base of the AHV pension fund. Further, the government proposed a social model of flexibilization of the pension age: it offered the possibility of early retirement to those not willing or not able to work until the regular age of 65. Between 1995 and 1997, the administration (federal social insurance office) assigned several expert groups with the task of reviewing the old system and working on new solutions especially regarding the financial security of the system (3-Säulen- 9

10 Bericht, IDA-FiSo1, IDA-FiSo2). Overall, the working groups confirmed the practicability of the three pillars, but at the same time found a need of alternative ways of financing for the first pillar, possibly with an increase in the value added tax. The government s draft was debated in parliament in spring/summer The main controversies were caused by the proposed flexibilization of the retirement age, more precisely; under what circumstances early retirement should be possible. The right proposed linear cuts in pension for those willing to stop working before the ordinary pension age. On the other side, the left claimed better conditions for lower income classes. Further, the left was not willing to accept deterioration in the widow s pension. Soon after the parliament s decision, the social democrats (SP) called for a referendum, together with the green party and the trade unions. A popular vote was held in May 2004, where almost 70% of the Swiss people voted against the bill. Extension of the agreement on the free movement of persons and flanking measures In 2000 the Swiss people had accepted the bilateral treaties with the EU, of which the free movement of persons was one of the most important agreements. In 2002, 10 new states became members of the European Union. The agreement on the free movement of persons between Switzerland and the EU had to be renegotiated, however. The negotiations were closed in May 2004 and the agreement was signed in October of the same year. Earlier in 2004, a tripartite working group with members of the public administration, the unions and business associations had discussed new flanking measures against wage and social welfare dumping. The group s propositions were open to public discussion in the consultation procedure during summer The final draft, published in October after interadministrative revision, included various measures to protect Swiss workers, such as an intensification of controls with the assistance of inspectors, stronger obligations to inform and facilitation for the declaration of collective labor agreements as generally binding. In the lower chamber of the Swiss parliament, the Swiss People's Party (SVP) called for a rejection of the agreement. Apart from the right-wing parties, however, the agreement and its flanking measures were generally supported. Moreover, the lower chamber decided to combine the agreement and the flanking measures in one legislative act. The SVP called for a referendum together with the extreme-right Swiss Democrats (SD). The popular vote showed a relatively strong support (56%) of the Swiss people for the agreement which was taken as an affirmation for the bilateral treaties in general. 3 3 The case of the extension of the free movement of persons and the related flanking measures is conceived as a Europeanized case of social policy. As the flanking measures (social policy) are directly related to and an 10

11 Agreement on the taxation of savings Through the agreement on the taxation of savings income, Switzerland supports the EU system of taxing cross-border payments of interest on savings to natural persons. Swiss paying agents such as banks deduct withholding tax on income from savings in Switzerland on persons liable for taxation in an EU member state. Three quarters of this amount goes as tax revenue to the country of domicile of the income recipient, the other quarter remains in Switzerland. Contrary to most EU-member states, Switzerland has no obligation to exchange information on the savings. Thus, the banking secrecy remains intact. However, the withholding tax can be replaced by voluntary notification of the income to the tax authorities of the country of residence, if demanded by the income recipient. Since the agreement on the taxation of savings is part of the bilateral agreements, the negotiations took place at the same time between 2002 and 2004 and it was treated together with the other agreements during the consultation procedure (2004) and in the government s dispatch. The agreement on the taxation of savings was discussed in parliament in late autumn The main controversies were caused by the discussion around the banking secrecy, which according to the government and the right parties was not be touched by any means. However, apart from this discussion and a controversy on the recipient of the withholding tax (confederation or cantons), the agreement did not cause much stir. In the end, the two chambers decided that the cantons should receive 10 percent of the tax revenue. The agreement was approved by the upper chamber in a unanimous decision, in the lower chamber with 171 against 16 votes. Program of budget relief 2003 Already in 2001, the Swiss people had decided on a program of debt relief, which is a stop on public financing via dept. In 2003 the economic situation had worsened again. The government realized that even the latest budget estimations had to be reconsidered and decided to elaborate a program of budget relief where the principle of dept relief was applied for the first time. It projected saving measures mainly in the fields of social welfare, traffic, national defense, agriculture, education and research and foreign relations (together 85% of the budget cut). Most of the offices were concerned to a greater or lesser extent. After consultation of the parliament s financial delegation and the financial committees of both chambers in winter 2003, the administration organized a very short consultation procedure as well as a consultation conference in June Shortly after, the bill was handed over to imminent consequence of an international treaty, they should be considered as at least partly internationalized (Sciarini et al. 2002: 4). 11

12 special committees of the two chambers established for these purposes. In parliament, which discussed the program of budget relief in early autumn of 2003, the main lines of conflict lay between the advocates of a slim state vs. those (mainly left parties) that feared a retrenchment at the expense of the Swiss people. Luckily, the measures had been elaborated in such a way that almost every area was concerned, which prevented major blockages. Moreover, it was generally accepted that some kind of action was urgently needed, so that the program was adopted in December Data and method Methodologically, our analysis of power structures is based on tools of Social Network Analysis (e.g. Wasserman and Faust 1994). Data was gathered through approximately hundred semi-structured interviews with representatives of organizations involved in the four policy processes under study. Based on positional, decisional, and reputational approaches (see e.g. Knoke 1993: 30), 20 to 30 organizational actors were identified and interviewed. Most of the interviews were conducted between February and July In line with the dominant conception (Laumann and Knoke 1987; Knoke et al. 1996: 7), our study focuses on organizations and the individual interview partners were asked to respond in the name of their organization. The two dimensions of power structures that were presented above are operationalized as follows: The distribution of power is measured by reputational data. Interview partners were asked about all actors that, in their view, have been influential in the specific decision-making process, this from a list comprising all actors that took part in that process. Additionally, interview partners were asked to select the three most influential actors out of the influential ones. This measures what we call strong influence (see table 2). Based on these answers, we first calculate the score of reputational power of each actor, i.e. the percentage of actors which consider a certain actor as being influential. The percentages of all the actors were then summed up to a score for the "overall influence" which is "distributed" to the actors present in the process. From this, we calculated, for each group of actors, their share of this overall influence. The group of political parties is compared to the group of interest groups. Additionally, we make the difference in the groups of interest groups between business interest groups, trade unions, and other interest groups. The conflict lines are measured by data on convergence and divergence of views. Interview partners were asked to mention the actors with whom they had diverging (-1) or converging (1) views on the policy project in question, this again based on a list comprising all major 12

13 actors participating in that process. The resulting data matrix was then analyzed with blockmodeling approaches in the program UCINET (Borgatti et al. 2002). The program first calculates similarities in the convergence/divergence profile of the actors and then clusters them, beginning with the most similar ones. 4 The last remaining division between two clusters of actors thus represents the most important conflict line in the case. The substantial content of the conflict line has to be interpreted in function of the actors which compose the two clusters. 4 Results and discussion Distribution of power We now present the results for the four cases under study. Based on this, we will then discuss and compare the four cases regarding the two influence factors. First, table 2 shows the results for our first dimension of power structures, which is the distribution of power. The numbers in the table are percentages and indicate the distribution of influence among the different actor groups. The two indicators for influence (influence and strong influence) are considered together in order to assess whether political parties or interest groups are more powerful. For reasons of completeness, the figures for the other groups of actors, not in the focus of this paper, are also included in the tables. For cases of social policy (eleventh pension reform and free movement of persons), we expected interest groups to be more powerful than political parties. This is clearly the case for the eleventh pension reform. While interest groups have 34% of the total influence (resp. 36% of the total strong influence) in the case, political parties only have 26% (resp. 23%). Looking more precisely at the separate, most important actors, this finding must however be slightly nuanced. 5 First, the most influential actor, together with the Federal Department responsible for the project (EDI), is the Socialist Party. It was considered as being influential by all of the interviewed actors. This, together with the finding that the Swiss Federation of Trade Unions (SGB) was the most influential interest group, expresses the fact that these actors won the referendum against the bill. In addition to these two actors from the left side of the political spectrum, the three bourgeois parties and the three most important business interest groups are very influential as well. 4 This indirect, two-step method of blockmodeling has recently been criticized. Instead, a direct method was proposed (see Doreian et al. 2005: 168ff.). However, the analysis of our data with this direct method, called "generalized blockmodeling", in the program PAJEK (Batagelj and Mrvar 1996), gave us very similar results to the ones presented in this paper, calculated with the indirect method. 5 Tables with the ranking for specific, not aggregated actors can be found in Appendix 1. A list with abbreviations of actor names can be found in Appendix 2. 13

14 Table 2. Influence of actor groups Eleventh pension reform Program of budget relief 03 Free movement of persons Taxation of savings N Influence Strong influence N Influence Strong influence N Influence Strong influence N Influence Political parties Strong influence Interest groups Business Unions Other Federal Council Public Administration Cantons, etc Experts and other

15 Thus, among the most important political parties and interest groups, the power distribution is balanced. The fact that some smaller interest groups such as the union Travail.Suisse, the Swiss Farmer's Association (SBV) or organizations of the elderly and women have clearly more influence than the smaller parties explains the overall stronger positions of interest groups. Looking at the indicator for strong influence, however, the quite clear advantage for interest groups is again confirmed. For the case of the free movement of persons, even if political parties and interest groups are equal in terms of influence, the indicator for strong influence shows a clear domination of interest groups (42% against 14%). Looking more precisely at the different actors, one can detect clearly that actors defending the open economy (Economiesuisse, FDP) as well as actors in favour of the working population (SGB, SP) dominate the decision-making process. The first group was clearly in favour of the extension of the free movement of persons for the benefit of the domestic economy and has therefore invested quite an amount of resources in the decision-making process. The second group was the one that influenced the process by requiring an extension of the flanking measures, without what they would have rejected the extension (see Fischer 2003). Thus, in terms of simple influence, the most important actors are again interest groups and political parties. It is, as in the case of the pension reform, the indicator for strong influence that points heavily in the direction of a dominance of interest groups. Thus, in summary and for both cases of social policy, interest groups seem to be more powerful. This is in accordance with our expectations. Nevertheless, to confirm our hypothesis 1, we have to consider whether there is variation concerning influence, thus whether things are different in cases of financial policy. For the cases of financial policy, we expected the opposite situation, thus political parties being more influential than interest groups. If we consider both indicators for influence, this expectation can not be confirmed. In the case of the program of budget relief 03, political parties and interest groups are equally influential (21% resp. 18%). On the side of interest groups, Economiesuisse is by far the most influential actor. This position is confirmed by the measure of strong influence, where this economic interest group ranks far above all the other society actors. This fact underlines the interest of Economiesuisse in a well-balanced state budget and tax lowering. The four biggest political parties as well as the other important interest groups did have some limited influence in this process. Note also that state actors like the Department and Administration of Finance or the Canton s Finance Directors are clearly more powerful in this issue than political parties and interest groups. The same is true for the other case of financial policy, the bilateral agreement on the taxation of savings. Actors from 15

16 the state administration are much more influential than political parties or interest groups, with one important exception, however. The Swiss Banker's Association (SBVg), being directly concerned by this issue, was the most influential actor in this process. Together with other associations defending the interests of the Swiss financial sector (Economiesuisse, Association of Private Bankers, Swissholdings), the SBVg is responsible for the clear dominance of interest groups over political parties in this case. Widely considered as a very technical issue requiring sophisticated and detailed knowledge, the bilateral agreement on the taxation of savings was probably very difficult to influence for political parties with their limited resources. Note also that trade unions had no stake in this issue and did not even try to influence the process. Summarizing the findings for the power distribution in our cases of financial policy, we can conclude that interest groups are generally more influential than political parties. This counters our hypothesis 1. In sum, all of the four cases under study display the same pattern, namely the clear domination of interest groups over political parties. While this was expected for cases of social policy, the cases of financial policy were expected to display a slightly different pattern. Consequently, we can only partly confirm our hypothesis 1. This implies that also our second hypothesis, according to which we expected differences between domestic and Europeanized cases, can only partly be confirmed. While in Europeanized cases, interest groups are generally more influential than political parties; the same is also true for cases of domestic policies. This, however, does not correspond to our theoretical reasoning. Conflict lines Let us now consider the second dimension of power structures. Tables 3-6 show the major opposing positions for the four cases under study. The figures in the intersection of two positions represent the average relation between the actors composing that cluster. Below, the actors sharing a position are listed. In function of the actors composing the different clusters, substantial meaning has to be attributed to the conflict lines separating them. While we expected a classical left-right conflict line for cases of domestic policies, the actors in the Europeanized cases should mainly line up along the openness-traditions conflict line. We did not expect the differences between the two policy domains to impact upon the conflict line. 16

17 Table 3: Conflict lines in the 11 th pension reform Position 1 Position 2 Position 3 Position Position Position Position 1: Position 2: BR, EDI, BSV SP, SGB, Travail.Suisse, Unia Position 3: Average conflict level: CVP, FDP, SVP, EDU, LPS, SAV/Economiesuisse, SGV, Alliance F, EFD, EFV, ESTV, BfS, Seco, SODK, FDK, Experten Table 4: Conflict lines in the program of budget relief 03 Position 1 Position 2 Position 3 Position Position Position Position 1: Position 2: Position 3: Average conflict level: SVP, FDP, CVP, Economiesuisse, SAV, SGV, BR, EFD, EFV SP, GPS, SGB, Travail.Suisse, VGB, transfair EDU, SBV, FDK, KdK, SSV, EPA, DEZA, Seco, BSV, BFE, BABS Looking at the three main clusters for the domestic cases (Eleventh pension reform and program of budget relief 03), there is a clear left-right conflict line emerging. 6 In the case of the pension reform, the main left party (SP) shares a position with the most important and powerful trade unions. These actors traditionally tend to defend the welfare state and try to expand social security systems. In the present case, these actors launched a referendum against the reform and did eventually win it, preventing the pension reform coming into force. These left actors are quite strongly opposed to the administrative leaders of the reform, which share position 1, as well as to position 3. In position 3, the three right parties are grouped 6 We present three clusters of actors in these tables instead of two. Note, however, that the most important conflict line is given by the opposition between the actors of the political left and the actors of the political right. An additional and substantial, though less important, conflict line separates a group of state actors from the right and the left cluster, respectively. Because this additional separation has substantial meaning in the two cases, we decided to present them in the tables. 17

18 together with the most important economic interest groups. These actors supported the 11 th pension scheme reform. In the case of the program of budget relief, one can observe a very similar pattern. While the most important actors on the left side of the political spectrum (SP, unions) share a position, the actors on the right side of the political spectrum share another position with the administrative units responsible for the program. These two clusters of actors are very strongly opposed to one another, as can be detected by looking at the average conflict level between their positions. The bourgeois parties, together with the main economic interest groups, are traditionally in favor of a balanced budget and thus supported the program of budget relief. The actors on the left traditionally defend a strong state administration and were consequently opposed to budget restrictions. A third position is formed by actors from the federal administration which potentially suffered from the retrenchment measures. Interestingly, an important interest group, the Swiss Farmer's Association (SBV), shares its position with these administrative units. This can probably be explained by the fact that the sector of agriculture, and thus agrarian subsidies, formed part of the planned cuts. Thus, position 3 seems to be shared by the potential direct victims of the program of budget relief 03. Summarizing, one can observe a clear left-right division which forms the main conflict line in the two domestic cases. This assertion gives support to the second part of hypothesis 3a. Table 5: Conflict lines in the case of the free movement of persons Position 1 Position 2 Position Position Position 1: SVP, SD, Lega, FPS, PdA, SolidaritéS/Alliance de gauche, AUNS Position 2: CVP, FDP, SP, GPS, EVP, Economiesuisse, SAV, SBV, SGV, SGB, Travail.Suisse, BR, EDA, IB, Mission, BJ, BFM, EVD/Seco, KdK, VSAA Average conflict level:

19 Table 6: Conflict lines in the case of the taxation of savings Position 1 Position 2 Position Position Position 1: SVP Position 2: FDP, CVP, SP, GPS, EVP, EDU, Economiesuisse, SGV, Travail.Suisse, SBVg, Swissholdings, Treuhandkammer, BR, EDA, IB, EFD, EFV, ESTV, BJ, KdK, FDK Average conflict level: In order to analyze the first part of hypothesis 3a, we have to look at the conflict lines in the two Europeanized cases. For the case of the free movement of persons and the related flanking measures as well as for the bilateral agreement on the taxation of savings, the main conflict line is given by the openness-traditions division, as can be seen in tables 5 and 6. Concerning the free movement of persons, the Swiss People's Party (SVP), together with small extreme-right and extreme-left parties, is opposed to the other important parties and to all of the most important interest groups. These two clusters have a high level of divergence between them and a high level of internal convergence. The SVP and the small extreme parties all opposed the extension of the free movement of persons to the ten new EU member states by referendum. They clearly defend a policy of traditions vis-à-vis the European Union and their citizens. While the two centre-right parties (FDP, CVP) and the economic interest groups were generally in favor of the treaty because of its benefits for the Swiss economy, the Socialist Party and the trade unions supported the agreement only under the condition of a reinforcement of the flanking measures. These were designed to protect Swiss workers against wage dumping. In the second Europeanized case, the openness-traditions conflict is far less pronounced, but still clearly observable. As for the former case, the SVP is opposed to the other main political parties as well as to the most important interest groups. This big majority of the political actors agreed upon the necessity to negotiate with the EU on the problem of the taxation of savings and was thus in favor of the treaty. The SVP on the contrary estimated that the solution found was a clear attack on the Swiss banking secrecy and that Switzerland made too many concessions to the EU. Again, we can observe that the main conflict line is given by the openness-traditions division. However, we need to specify this finding. According to the literature, we expected the openness-traditions conflict line to be important for political parties and interest groups. While we observed the cleavage for political parties, interest groups seem to be less 19

20 concerned. Looking at the clustering of actors, it appears that representatives of the business interest groups focused on world markets and those focused on the internal market do not have major conflicts. Economiesuisse and the SAV on the one side and the SGV and the SBV on the other side do in fact have very similar convergence/divergence profiles in the case of the free movement of persons. In the case of taxation of savings, the SGV is separated from the Economiesuisse only by the fourth most important conflict line. 7 Thus, we found no support for the widespread idea that there is an accentuation of the external-internal division for economic interest groups. Note also that we found no support for Bonoli's (2006) hypothesis which argued that there exists an additional cleavage between right-wing political parties and employers organizations. In fact, these two groups of actors do have very similar profiles of divergence/convergence and are not divided by a major conflict line. We can thus support our hypotheses 3a and 3b on the conflict lines: While the institutional differences between the domains of social and financial policy have no influence on the conflict line, Europeanization does so. In Europeanized cases, the openness-traditions division is more salient than the classical left-right division. This, however, is only true for political parties, but not for interest groups. 5 Conclusions The purpose of this paper was to contribute to the discussion about the relations of political parties and interest groups. Starting from the debate about the influence of interest groups in social policy issues in Switzerland, we tried to shed some light on the understudied issue of party-group relations. We hypothesized that institutional differences between domestic and Europeanized cases on the one hand, and between cases of social policy and cases of financial policy on the other hand influence the power structure among political parties and interest groups. On the one hand, we looked at the distribution of power among the different actor groups; on the other hand we identified the most important conflict line within the different cases. Social network data, gathered through interviews with representatives of the most important political actors, provided us with high-quality and dense information about the structures in the different processes. With respect to the distribution of power, we expected interest groups to be more influential than political parties in the domain of social policy, while parties were expected to play a dominant role in the domain of financial policy. In the Europeanized cases, we hypothesized interest groups to be more influential than political parties. The opposite was expected for the 7 This information is not visible in the data presented in the tables, but only in the data matrices. 20