The binary logic of the vote imposed the formation of two camps in spite of all the .... But, contrary to the American example, it also applies .... reasons that have little to do with their own core beliefs. ..... To identify the coalitions and their possible components on both sides, we have performed a ... A seventh, we thought more.
National Centre of Competence in Research (NCCR) Challenges to Democracy in the 21st Century
Working Paper No. 12
Coalition formation in direct-democratic campaigns: a case study of the vote on the Swiss asylum law
Hanspeter Kriesi, Laurent Bernhard, Regula Hänggli Department of Political Science, University of Zurich
February 2008
1 Abstract
Coalition formation in a direct-democratic campaign follows a logic of its own. In such campaigns, there is little room for bargaining, there are no designated leaders, and the number of potential coalition partners is large. Moreover, coalition formation has a decisive influence on the outcome of the vote. We argue that, in such campaigns, coalition formation is decisively shaped by the institutional setting and the belief systems of the actors involved. The exploratory analysis of the case of the vote on the Swiss asylum law in 2006 confirms the importance of these factors. The binary logic of the vote imposed the formation of two camps in spite of all the substantive distinctions the actors wished to make. In each one of the two camps, however, the heterogeneity of the alliances manifested itself in the formation of component coalitions, each of whom was based on distinct sets of beliefs.
In the literature on direct democracy, hardly any attention is paid to coalition formation, a process that is getting a lot attention in the literature on representative democracy. In the present paper, we propose to take a closer look at the process of coalition formation in the context of a Swiss referendum campaign. Although this case is quite specific, it allows us to explore the more general conditions of the formation of direct-democratic coalitions, and to test some instruments for their systematic analysis. We shall argue that in direct democratic campaigns, the conditions for coalition formation are quite different from those obtaining in representative democracy, with key implications for the way coalitions are formed and should be studied. In representative democracy, coalition formation takes place at different stages of the political process. It may occur before elections in the formation of pre-electoral coalitions (Golder 2006), it occurs in the process of government formation, and it is pervasive in parliamentary logrolling. At each stage, the process is characterized by a certain number of characteristics that distinguish it from the coalition formation in direct-democratic campaigns. The situation is always one of bargaining between a limited number of potential coalition partners – political parties, typically represented by clearly designated leaders, with a zero-sum condition: what the one gains, the other loses. The bargaining process involves multiple issues and, for a given issue, multiple aspects which are packaged into a delicately balanced compromise. Coalition formation at each stage involves promising side payments (attached to holding office or in terms of policy payoffs) to possible partners. In such a situation, size plays a key role. Thus, in the literature on government formation in representative democracy the most cited concept is Riker’s (1962) ‘size principle’, which states that actors will form coalitions just as large as the minimum necessary to obtain control over the government, i.e. they will form ‘minimal winning coalitions’. The situation of a direct-democratic campaign is quite different. In this case, there is little room for bargaining: the proposal submitted to the vote is issue-specific and has been definitely
3 packaged at earlier stages of the process – either by its sponsors or by parliamentary (and preparliamentary) bargaining. The issue-specificity is not an accidental feature, but usually constitutionally prescribed (Clark 1998: 467). As we shall argue below, in their mediated form, direct democratic processes do not preclude logrolling at previous stages of the political process. However, at the stage of the popular vote, there is no room left for bargaining. Moreover, the choice is highly prestructured by the fact that there usually (although not always) are only two options – support or rejection of the proposal. Once the proposal is submitted to the vote, the potential coalition partners have only the choice of taking or leaving it. In addition, each direct-democratic vote is a rather unique event that is hardly connected to other votes at all, which leaves little room for possible side payments among actors. The point is that directdemocratic institutions require a vote by simple majority on a given issue-specific proposal, which strikes at the roots of bargaining (Barry 1975: 485f.). Second, there is no institutionally designated leader who is charged with the task of forming a coalition in favor or against the proposal. Depending on the form of direct democracy, the government may take a leading role in the campaign, but this is not always the case and even when the government leads the campaign of one side, there may be no institutionally designated leader on the other. Third, the number of potential coalition partners is much larger than in representative democracy, since the political parties are not the only relevant actors. This complicates the coalition formation. Fourth, direct-democratic coalitions resemble pre-electoral coalitions to the extent that they have an influence on the outcome of the vote. In fact, the whole point of these coalitions is to control the outcome of the vote. The large number of actors involved increases the uncertainty about the outcome of the vote, and decreases the possibility of control by any single actor. By contrast, compared to pre-electoral coalitions, the issue-specificity of the campaign may, given knowledge about the distribution of the issue-specific popular preferences, reduce the uncertainty of the outcome. Against the background of these specificities of a direct democratic campaign, the problem of size is posed in terms that are rather different from the ones in representative democracy. The outcome of a direct-democratic vote has the characteristics of an uncertain public good (or a
4 public bad, depending on the side you are on). In such a situation, both sides need all the supporters they can get. The larger the coalition, the greater the chance of success at the polls. Accordingly, for potential coalition partners, possible free-riders who benefit from the outcome of the vote without contributing to the campaign effort constitute a key problem. In other words, we do not expect to see ‘minimal-winning coalitions’, but rather large coalitions including actors of different types and political orientations. Since there are no institutional leaders in directdemocratic campaigns, we expect the formation of ad hoc committees on either side, which are constituted for the purpose of coordinating the various actors who are likely to get involved in the campaign. Alternatively, important actors may take it upon themselves to coordinate the alliance partners on their side of the campaign. Finally, increasing the number of coalition partners to get as much support as possible introduces the problem of heterogeneity, or the ‘extension dilemma’ familiar to social movement strategists (Jasper 2006: 127f.): ‘the further you reach out your team or alliance, the more diverse it will be and the less unified’. The more heterogenous you are, the more difficult it is to agree on a common campaign message, and the more difficult it will be to agree on common strategies. The way these coalitions are formed, depends on at least three sets of factors which we would now like to introduce – the institutional setting, the beliefs of the actors, and their considerations for launching a challenge. There are other factors that play a role as well – not the least of which is the structure of the public space. For the present purposes, we shall, however, concentrate our discussion on these three sets. Next, we present the case we shall use to explore the directdemocratic coalition formation empirically. After a presentation of the data and the methods of analysis, we then proceed to a presentation of our results.
5 Determinants of coalition formation in direct-democratic campaigns
The institutional setting Walder (2006: 713) points out that ‘a theory of political choice – or any theory about politics – can be no more valid than its claims about the contexts within which these choices are made’. More specifically, as Sniderman (2000: 69) observes, political institutions are the ‘organizers of political choices’. Institutions define the rules of the game. Direct-democratic institutions come in different guises with different institutional logics attached. One form is the populist, unmediated form to which the American states seem the closest approximation in practice (Cronin 1989: 5057; Smith and Tolbert 2001: 740, 2004: 112ff.; Bowler and Donovan 2006). In the US still today, the initiative process is primarily used by citizens’ and special interest groups to circumvent partisan state legislatures, and the US experience with the use of direct democracy has allowed citizens’ interest groups to flourish. Thus, states with initiatives have developed a larger number of citizens’ interest groups than states without the process. At the same time, having more initiatives on the ballot is not associated with having more economic interest groups (Boehmke 2002; Smith and Tolbert 2004: 96-102). The other form of direct democracy is the more organized and mediated form of most other countries, where popular consultations take place, and where the political parties exert a much closer control over the direct democratic process (Budge 2001: 70)1. In Switzerland, the European country with the greatest experience in direct-democratic procedures and the setting of our case study, a popular vote always intervenes at the end of an extended political process. This is, of course, obvious in the case of the (optional or compulsory) referendums, which concern proposals adopted by Parliament. But, contrary to the American example, it also applies to popular initiatives, which are propositions submitted by a petition signed by some group of
1
Even in the US today, party organizations clearly acknowledge the power of direct democracy, and become involved in initiative contests for various reasons (Smith and Tolbert 2001, 2004: 116f.). If American political parties still rarely sponsor ballot questions, they tend to get more involved in the campaigns before the vote. Although in the European experience, political parties have participated more frequently in direct-democratic procedures, this is not the case in all the countries: for the Irish exception, see Sinnott (2002).
6 voters2. Before being submitted to a popular vote, Swiss initiatives are always debated by the government and by parliament (von Arx 2002). Both decide about whether or not to support the initiatives in the popular vote, and initiatives submitted to a popular vote are always accompanied by their voting recommendations. The parliament even has the possibility to formulate a direct (or indirect) counter-proposal to an initiative which may also be submitted to the popular vote together with the initiative. The extended political process preceding the campaign before the vote serves to prestructure the configuration of the actors involved in the campaign. Although in the final stage of the directdemocratic vote, bargaining is no longer possible, the preceding political process is bound to have included a lot of bargaining, and to have led to the formation of issue-specific parliamentary, and even pre-parliamentary coalitions, which constitute the point of departure for the coalition formation in the direct-democratic campaign. More specifically, the government and the parliamentary majority always are part of one of the two camps facing each other in the campaign. In the case of a referendum, challengers of the proposal adopted by Parliament face the opposition of both the government and the parliamentary majority. In the case of a popular initiative, its sponsors usually have to face the opposition of the two as well, since the government, followed by the parliamentary majority, almost always recommends the rejection of popular initiatives3. Given the key role of the political parties in the political process preceding the vote, the party system constitutes another crucial element of the institutional setting, which serves to 2
The basic Swiss direct democratic institutions can be classified based on the source of the proposition (government or some group of voters), and the initiation of the vote (required by the Constitution or by some group of voters) (Kriesi 2005: 20f.). According to the first criterion, we can distinguish between initiatives and referendums: initiatives are propositions submitted by a petition signed by some group of voters, while referendums are propositions formulated by the government and the majority of Parliament. According to the second criterion, we can distinguish between two versions of the referendum – the one required by the Constitution (the compulsory referendum), and the one required by some group of voters who sign a petition requiring a vote on a parliamentary bill within some specified period after its adoption by Parliament (the optional referendum). 3 For the sponsors of an initiative, but not the challengers of a bill adopted by Parliament in a referendum vote, the institutional logic implies an additional disadvantage – a disadvantage which they share with the sponsors of initiatives in the US: since an initiative asks for a change, its sponsors have to incite the citizens to vote ‘yes’, i.e. they have to lead a ‘status quo modifying campaign’, while the challengers in a referendum vote may lead a ‘status quo preserving campaign’ asking the citizens to vote ‘no’. As is well known, ‘status quo preserving campaigns’ are more easily won than ‘status quo modifying’ campaigns’ (Gerber 1999).
7 prestructure the direct democratic coalition formation. Switzerland has a multiparty system that, in principle, offers countless possibilities for coalition formation. In practice, however, they are reduced to only a limited number in the context of a direct-democratic campaign. There is a tendency of such a multipolar party system, as Sniderman (2000: 83) correctly guesses, to collapse into a functionally bipolar one. This is not only a result, as he believes, of the irresistible attractions of structural simplicity. Such a reduction is also induced by the structural conditions of a direct-democratic vote, where the voter ultimately only has a binary choice. In spite of its exceptional degree of fragmentation, the Swiss party system essentially has a tripolar structure, which is typical of Western European party systems (Kriesi et al. 2008): it is divided into the left (the Social-Democrats (SP), the Greens, and some minor parties), the moderate right (including the liberal Free Democrats (FDP), the Christian-Democrats (CVP) and a minor Liberal Party (LPS), and the new populist right (the Swiss People’s Party (SVP) and some minor parties). In the peculiar Swiss system of government – a grand coalition composed of the four major parties (SP, FDP, CVP, and SVP), who represent all three poles, even a governing party may oppose the government on specific policies. In some cases, all four parties of the Swiss grand coalition government jointly support the parliamentary proposal, and the challenge is launched by one of the minor parties. Typically, however, the governmental coalition splits, and either the left or the new populist right opposes the governmental coalition (Kriesi 2005: 26-34). In other words, typically, the governmental coalition is either composed of the moderate right and the new populist right (centre-right coalition), or of the moderate right in alliance with the left (centre-left coalition). Less frequently, two of the four major parties join forces in opposing a parliamentary proposal. The widespread notion that the establishment of direct democratic procedures weakens political parties is partly confirmed by the Swiss experience. Among other things, direct democratic decisions induce intra-party conflict (Ladner and Brändli 1999: 286). It is, indeed, not uncommon for cantonal sections of Swiss parties to take a different stand from their national parties, which complicates the coalitional structure of a direct-democratic campaign. Given that Swiss parties traditionally have been rather loose federations of cantonal sections, there is always a possibility
8 that some cantonal sections may deviate from the positions taken by the national parties in a given campaign. However, the number of deviating sections has usually been quite limited and confined to the parties of the moderate right, who are less cohesive than the polar opposites on the left and on the populist right. More generally, the Swiss experience indicates that, while direct democratic procedures may weaken political parties, their impact on parties is not as unambiguous as is often assumed4. While weakening some parties, direct democratic procedures reinforce others, especially the more peripheral and the less established ones among them (Papadopoulos 1991); they help parties maintain a high level of political activity, give them an opportunity to put forward their political ideas, and may contribute to their professionalization and organizational development (Ladner and Brändli 1999). Most importantly, even when organizationally weaker than in other countries, Swiss parties are themselves major sponsors of initiatives and referendums, and their recommendations provide key cues for the voters (Kriesi 2005). If parties are key players in the more mediated form of direct democracy as it is practiced in Switzerland, they are not the only actors involved in direct democratic campaigns. The Swiss experience, indeed, also provides evidence that direct democracy fosters competition between parties and a variety of interest groups. In previous studies, it proved, however, difficult to separate the impact of parties from that of interest groups in direct democratic campaigns (Kriesi 2005: 63; 2006: 607). It seems plausible to assume that the relative importance of parties and interest groups in direct-democratic campaigns varies according to policy- or domainspecific conditions. Policies are made in domain-specific subsystems, which, in addition to political parties, include a large number of actors whose composition varies from one policydomain to the other. From the point of view of direct-democratic campaigns, it is particularly important whether or not a domain concerns broad, diffuse constituencies and/or some narrower interests. Donovan et al. (1998) distinguish between campaigns, where a narrow group challenges another narrow group (‘interest group contests’), broad groups challenge narrow interests (‘entrepreneurial contests’), a narrow group challenges a broad, diffuse group 4
For a more general argument that direct democratic procedures need not necessarily weaken the role of political parties, see Budge (1996: 120ff., 2001)
9 (‘client contests’), and where a broad group challenges another broad group (‘majoritarian contests’). Narrow interests include the key economic interest groups and trade unions, but also a range of more specialized interest groups (such as professional associations), and experts (scientists, policy analysts, consultants, journalists, etc.). Broad, diffuse constituencies, by contrast, are organized by political parties, but also by ‘citizens’ interest groups’ or ‘social movement organizations’. The latter include organizations of the ‘new social movements’ of the seventies and eighties that have become more or less institutionalized interest groups (ecological organizations, human rights and aid organizations, consumers’ organizations, pacifist groups, or women’s associations). In addition, broadly based groups also include more traditional organizations, such as churches, pressure groups defending traditional interests, and, last but not least, offices of the public administration.
Belief systems Within the institutional constraints, we expect the coalition formation of the relevant actors to be primarily based on the actors’ shared beliefs, and less on their short-term strategic considerations. This is in line with the advocacy coalition framework (ACF) (see Sabatier and Weible 2007: 192ff.). The expected importance of beliefs in coalition formation does not imply that the actors are completely disinterested or value-rational in their political orientations. It only means that, when forming coalitions, actors pursue policy-specific goals that are in line with their core interests and/or principles. It is conceivable that actors may join coalitions for strategic reasons that have little to do with their own core beliefs. They may, for example, do so for opportunistic electoral reasons (because they want the electorate to perceive them in a particular way) or for reasons of alliance-building/office-seeking (because they want the other members in the coalition to perceive them as a viable partner). Following the ACF, we propose to conceptualize a three-tiered hierarchical structure of beliefs. At the broadest level are deep core beliefs, which are not policy-specific. They involve fundamental values, general normative assumptions about human nature and about the proper
10 role of government. The traditional left-right scale operates at the deep core level, as do traditional scales of political values. As far as the latter are concerned, we typically find two dimensions in Western Europe – an economic and a cultural one (Kriesi et al. 2008). The economic dimension refers to the classic opposition between state intervention and marketsolutions that has constituted the core of the opposition between left and right. The cultural dimension had its roots in religious confrontations. Under the impact of secularization, the transformation of the class structure and globalization, it has, however, acquired an entirely new meaning. It is now opposing liberal and cosmopolitan values on the one side, and authoritarian and protectionist/traditional values on the other. Moreover, with the opening up of the European nation-states – a result of their increasing integration into the global economy, into global migration flows, and into regional (EU) and global (UN, WTO) political regimes – the cultural dimension has become ever more prominent in the mass publics and in electoral competition. At the next level are policy core beliefs, which are applications of deep core beliefs that span the entire policy subsystem. The final level consists of secondary beliefs. They are relatively narrow in scope and address specific aspects of the policy in question. In particular, they refer to the specific characteristics of the proposal submitted to the vote. Actors of the political elite have well developed political ideologies, which impose constraints on their political belief systems (Converse 1964: 227-230). This means that their secondary beliefs are embedded into their policy cores, which in turn are embedded into their deep core beliefs. But, as we have already pointed out, actors may join a given coalition for different reasons. Simplifying greatly, there are four possible types of alliances on either side (see Table 1). Actors who share both core and secondary beliefs constitute what we propose to call a ‘natural alliance’. Based on the highly constrained character of the political elites’ belief systems, we expect above all to find this kind of coalitions. Second, actors who neither share core, nor secondary beliefs, but still take the same position with regard to the proposal constitute an ‘unnatural alliance’, or what Ossipow (1994: 39) calls an ‘objective coalition’. Such an alliance does not result from any explicit compact, but comes about by the aggregation of actors or component coalitions of actors who campaign independently of each other in pursuit of a
11 common goal. Such actors find themselves on the same side of the campaign, but for entirely different reasons. They do not agree with respect to the specific aspects of the proposal, and these disagreements are rooted in different policy- and/or deep-core beliefs. On the opponents’ side, the proposal may go too far for the ones, not far enough for the others; on the supporters side, actors may support different aspects of the proposal. The third type refers to actors who share secondary beliefs, but for different deep or policy-core reasons. Some group of actors may support a given proposal for cultural reasons, while for another group of actors the same proposal mainly speaks to their economic interests. In this case, the leaders of the two groups may explicitly form partial, or component coalitions that cater to their specific target groups, in order to deal with the problem of heterogeneity implied by the different core beliefs. Contrary to the previous type, the component coalitions are likely to cooperate in this case. They constitute what we might call a ‘pragmatic alliance’. A variant of this type of alliance occurs, when actors join a given component for strategic reasons. The final combination – actors who do not agree on secondary beliefs, but share the same core beliefs – may at first sight be highly unlikely. But it may still happen that, for ideological reasons, actors stick together, although they do not agree on various aspects of the proposal. They constitute what we propose to call an ‘ideological alliance’. Such actors are likely to constitute a common component.
Considerations for launching a challenge On the part of the challengers, the coalition formation process is closely linked to the decision to launch a campaign in the first place. The challengers have a ‘first mover’ advantage. In the case of an optional referendum, which concerns us here, their decision to mobilize corresponds to the decision about whether or not to launch a referendum. This decision depends on a series of considerations. First, the question is whether, from the vantage point of the potential challenger, the reform proposal is sufficiently different from the status quo. The answer to this question is far
12 from straight-forward. Given the complexity of modern pieces of legislation, such proposals tend to be carefully packaged sets of measures. For any given organization, the balance of advantages and disadvantages of a proposal may not be clear-cut. Accordingly, the decision to take up the challenge and to launch a referendum may give rise to serious internal tensions, and to possible intra-organizational fragmentation, as mentioned before. Moreover, different organizations among the set of potential challengers inside and outside of Parliament may arrive at different conclusions about the balance of advantages and disadvantages of the proposal in question, which may undercut the coalition-formation possibilities from the outset. Any actor who comes to the conclusion that the proposal should be challenged is confronted with the question of whether there is a chance for his side to win the popular vote. Provided there is a serious chance of success, and provided the challenger has the required resources (coalition partners) at his disposal, he is likely to decide to launch a referendum. Even if his chances to win the vote are small, he may still be prepared to launch a referendum, provided he has the required resources (coalition partners): whatever his chances for success, he may do so, because he has a reputation to defend, or simply because he cares about principles. The potential challengers may want to signal to their own constituency that they are faithful to their principles, and do not accept the kind of measures adopted by the parliamentary majority. They may also do so, because they can count on sizeable (although not majoritarian) popular support, which would allow them to signal to their opponents that the latters’ manoeuvering space is not illimited in the domain in question. Figure 1 summarizes the reasoning of potential challengers in the form of a decision tree.
Once some organizations have decided to go ahead with the referendum, the situation changes for the other challengers, as well as for the parliamentary majority. Now, they all have to position themselves with regard to the new challenge. The other potential challengers are under pressure to join the effort to collect the required number of signatures. They have to decide immediately, since the period for collecting signatures is officially limited in Switzerland to three months. Once it becomes clear that the challengers will succeed in collecting these signatures,
13 the members of the parliamentary majority, and the other supporters of the bill have to position themselves as well: As Sciarini and Trechsel (1996) and Trechsel and Sciarini (1998) point out, the supporters of the bill are likely to consider the electoral consequences of a possible defeat of the bill at the polls, and they may also see the content of the bill in a new light, once it is becoming apparent that it meets with vigorous opposition in the public. In other words, in the shadow of the popular vote, the parliamentary majority may tend to fragment. This applies to the moderate right in particular, who takes a pivotal position between the left and the new populist right, and who either joins forces with the left or the right to constitute centre-left or centre-right coalitions. Fragmentation tendencies are reinforced by the already mentioned sectionalism of the Swiss party system, and by the compeition exerted by interest groups and other types of actors.
The case of the asylum law 2006 The domain of immigration policies which constitutes the object of the present study represents a case of ‘majoritarian contests’. In this domain, one of the directly interested groups – the immigrants – does not constitute a relevant constituency, given that immigrants do not have the right to vote. On the other side, the interests of the indigenous population are mainly defended by new populist parties (Betz 2004, 1993; van der Brug and Fennema (2003), van der Brug et al. (2000, 2005), Kitschelt and McGann (1995), Kriesi et al. (2006, 2008)). In the age of globalization and denationalization, immigration policy has become a key issue for partisan mobilization, which implies that political parties constitute the main protagonists of the broad, diffuse groups confronting each other here. In addition, labour unions and more broadly based citizens’ interest groups in the resident population mobilize on behalf of the migrants’ interests, and are likely to participate in the campaign, too: churches and religious associations, human rights and aid organizations, and some more specialized pro-migrant and anti-racist groups. By contrast, economic interest groups other than labour unions tend to keep a low profile in this domain.
14 Immigration policy constitutes a highly conflictual domain, where consensus formation has proven particularly difficult. The new populist right under the leadership of the SVP has continuously mobilized for more restrictive measures, while the left and its associates kept trying to prevent a modification of the status quo and to defend the immigrants’ rights. The new populist right does not enjoy a policy monopoly in this domain, but it sets the agenda, and with its permanent campaigning has succeeded in keeping immigration issues on the agenda. As a result of these situational factors, there have been repeated modifications of immigration policy in Switzerland, as well as repeated direct-democratic votes related to such issues. Thus, there were no less than four votes on asylum policy over the last twenty years – in 1987, 1999 (involving two proposals), in 2002, and 2006. Three of these votes concerned referendums by organizations from the left against the toughening of asylum law adopted by centre-right coalitions, while the 2002 vote concerned an initiative of the SVP, which attempted to introduce a further turning of the screw into this legislation against the opposition of a centre-left coalition. In addition, over the same period of time, there have been nine more votes dealing with immigration issues, all of which were concerned with the question of the restrictiveness of immigration policy. On September 24, 2006, the Swiss citizens accepted a tough new asylum law with 67.7 per cent in favour to 32.3 per cent against in a referendum vote. As already indicated, the referendum which led to the vote in September 2006 was launched by the left against the tightening of the law adopted by the government and the centre-right Parliamentary majority. The result of the vote in September 2006 was almost identical with that of the two previous votes in April 1987 and June 1999. It was accompanied by a vote on the reform of the Immigration Law, which was also accepted by a similarly impressive majority at the polls. Both reforms had been discussed in Parliament in spring 2004. Compared to the government’s proposal, both bills had been toughened during the parliamentary debates, under the pressure of the populist right (SVP), but with the support of the moderate right. The latter’s support was motivated by the close outcome of the vote on the SVP’s tough asylum initiative in 2002: although the initiative had been rejected, the narrow margin by which it had been turned down
15 made the moderate right receptive for the far-reaching demands of the SVP in this policy domain. Among other elements, the new asylum law stipulated that asylum requests from refugees who were already accepted by another state would not be dealt with. It also included the prohibition of social assistance for refugees whose requests had been legally rejected. The Council of States had gone even one step further and had also abolished the emergency assistance for rejected refugees. But the National Council reintroduced this form of assistance after the Federal Court had declared its suppression unconstitutional. In addition, the new law introduced more restrictive rules for considering the question of refugees without proper identification; it adopted a so called ‘airport procedure’ allowing for rapid decisions at the refugees’ point of entry, and it allowed for the possibility to exchange information with the refugees’ home country. Finally, more drastic coercive measures (various forms of detention of asylum seekers) were adopted, and the duration of the existing measures was extended. Since these coercive measures not only concerned refugees, but all immigrants without a residential permit, they were included in the reform of the immigration law. It is worth mentioning that the UNHCR and the Council of Europe voiced concern about this reform. While the new asylum law was a clear case of retrenchment, the new immigration law had more the character of a compromise solution: in addition to the already mentioned reinforcement of coercive measures, and the introduction of a ‘dual system’ for the labour market that made it much more difficult for foreigners from outside of the EFTA/EU to get a job in Switzerland, the law also introduced some advantages for foreigners, such as facilitated access to residence permits for well integrated foreigners, reintegration aid for foreigners returning to their home country, and residential and occupational changes. The measures facilitating family reunions had, however, been watered down to such an extent – the right to residence permits was restricted to children under age of 12 – that they could no longer be considered a relevant advantage for foreigners. Immediately after the adoption of the two bills in Parliament, the Social-democrats (SP) and the Greens announced that they were going to launch a referendum against the asylum law. The
16 coherence of the left (and as it turned out of the opposing camp as well) was taxed by the fact that the Greens, in collaboration with two specialized aid organizations (FIMM-Forum for the integration of migrants, and Solidarité sans frontières), constituted an ad hoc ‘Committee for a double no’ (K2N) that linked the referendum on the asylum law to a referendum on the new immigration law. While the Greens attacked both laws, a majority within the SP, at first, wanted to concentrate on the asylum law. It was considered an easier target, because it was a pure case of retrenchment. The key arguments (humanitarian tradition, basic human rights, additional burden for the cities) were not as applicable for the immigration law as they were for the asylum law. Together with a policy-specific citizens’ interest group – the Swiss Aid for Refugees (SFHSchweizerische Flüchtlingshilfe), the SP formed an ad hoc committee against the asylum law – the ‘Coalition for a humanitarian Switzerland’ (KHS-Koalition für eine humanitäre Schweiz). Only when Ruth Dreyfuss, a highly respected former social-democratic member of government, decided to head the competing committee that challenged both laws, the SP ended up collecting signatures against both laws as well. Some parts of the left even doubted whether it was opportune to challenge the asylum law: the mobilization against this law would provide the new populist right with yet another chance to mobilize on immigration policy – its preferred issue domain, which might prove counterproductive in light of the upcoming elections in 2007, and in light of the possibility of a poor result at the polls. However, the majority on the left felt that it had no choice but to mobilize against a law that it considered inhuman. On the one hand, the left had already threatened with a referendum during the parliamentary debate, and it had to make sure that its threats were not regarded as idle. On the other hand, its credibility among its own electorate was at stake, given that the new asylum law violated some of its fundamental principles. In addition to this opposition, the governmental centre-right coalition had to face a serious problem of internal cohesion, because a ‘third force’ – the ad hoc committee of the so called ‘Bourgeois coalition’ (BK), composed of actors close to the governmental centre-right coalition, but opposed to the reform proposal – constituted itself and mobilized quite intensely for the referendum against the asylum law (but not for that against the immigration law). Surprisingly,
17 this third force was led by an outsider who did not belong to the political, but to the business elite. It strongly appealed to religious organizations and was supported by both, the Protestant and the Catholic Church leadership, who had already gotten involved in the campaign before. Under the impact of the mobilization by this third force, the pivotal moderate right fragmented to some extent, as expected by our general argument. But in this case, its fragmentation was limited to the French-speaking part of the country. In the German-speaking part, it remained solidly attached to the defense of the two bills. Assuming that the actors involved were aware of the outcome of the earlier votes on immigration and asylum policy, we expect the challengers to have been mainly motivated by reputation or principles, since they must have known that their chances of winning a popular vote were slim in this policy domain. The government’s supporters, by contrast, must have been mainly motivated by the anticipated success at the polls. Eventually, the campaign against the two proposals mobilized a large number of organizations in addition to political parties: these mainly include domain-specific organizations who support refugees and foreigners, as well as religious organizations. The number of organizations participating in the campaign on the government side was more limited, and included the usual allies of the moderate right – the business interest associations, in a minor role, and some domain-specific organizations defending the Swiss national tradition.
Data and method of analysis We identified the relevant organizations on the basis of various sources: the parliamentary debates, the campaign for the collection of signatures, voting recommendations, and the press and websites more generally. We used cross-checks with the persons we interviewed to complete the set of organisations. We ended up with a set of 47 relevant organizations. For all of them, we completed two interviews with their key campaign managers – one at the outset of
18 the campaign, and one after the citizens’ vote. 32 organizations belong to the challengers’ camp, and 15 to the camp of supporters. The difference in size of the two camps is already a first indication of their structural difference. The challenging organizations were not only much more numerous, as we shall see below they also came from more diverse backgrounds. It turned out that two of the organizations – GGK, a PR-firm, and the Swiss Farmers’ Association – were hardly involved at all. They are dropped from the subsequent analysis. Two additional organizations – two very minor parties of the populist right, the Freedom party (FPS) and the Mouvement of Genevan Citizens (MCG), who proved to be completely marginal, i.e. unconnected to the rest of the organizations involved – are also dropped from the analysis. This leaves us with 43 organizations, 31 challengers and 12 supporters of the two proposals. For the subsequent analysis, we treat each organization as a unitary actor, although some of them were internally divided with regard to the campaign. To identify the coalitions and their possible components on both sides, we have performed a network analysis. In our second interviews with the key campaign managers, we presented them with the original list of the 47 organizations. We asked them to mention all the organizations on the list, with whom they had closely collaborated in the course of the campaign. After they had gone through the list, we asked them to indicate the three organizations with whom they had collaborated particularly closely, and finally, we asked for the one organization among the three, with whom they had most closely collaborated. This kind of procedure is very much inspired by earlier work on political elites and their involvement in specific policy domains (e.g. Knoke et al. 1996; Kriesi 1980; Kriesi & Jegen 2001; Laumann & Pappi 1976; Laumann & Knoke 1987). A collaborative relationship is indicated by a ‘1’ in the adjacency matrix, a particularly close relationship by a ‘2’, and the closest collaborative tie by a ‘3’. In other words, we are weighting the closest relationships more heavily. To study coalition structures on the basis of this type of data, we draw on block-model analysis, which allows us to distinguish between structurally equivalent groups of actors (our coalitions) on the basis of an analysis of the cooperative relationships. A block model consists of two elements (Wasserman and Faust 1999:395): (1) a partition of actors in the network into discrete
19 subsets called positions, and (2) for each pair of positions a statement of the presence or absence of a tie within or between the positions. We shall discuss the two, four and eight blockssolutions. The patterns of relationships will be represented by so called “density-matrices”, which summarize the tie-specific densities between blocks (i.e. the share of possible ties between blocks that exist in reality). We also have a measure for the power of an actor, which allows us to specify the influence of different types of actors, and to describe the coalitional structure in more detail. Power is operationalized by a reputational measure and by a measure of structural centrality. The reputational measure is based on a set of questions referring to the list of all organizations involved (Kriesi et al. 2006). In the second interview, the respondents were first asked to name all organizations on this list, which, from their point of view, had been particularly influential during the campaign. Next, they were asked to name the three most influential organizations, and, finally, the most influential one. For each actor, a summary indicator counts the number of times he was mentioned by the other respondents in reaction to these questions, with similar weights attached as for the cooperative ties: mentions as ‘most influential’ are coded as ‘3’, mentions among the ‘three most influential’ as ‘2’, and mentions as ‘influential’ as ‘1’. The values of the indicator range from 0, for an organization that has never been mentioned as influential, to 141 for an organization that would have been considered to be the most influential actor by all 47 respondents. After having identified the component coalitions and their interrelationships, we need to establish to what extent the components share the various types of political beliefs. The deep core beliefs are operationalized in two ways: we measure the self-positioning of the key campaigners on a left-right scale, which gives us a most general indication of their ideological position. In addition, we obtain measures for their political values. For this purpose, we confronted the respondents with seven choices about the Switzerland they desired. Three choices refer to classic economic values (state-intervention in the economy vs. market competition, large income differences vs. small income differences, solidarity vs. self-responsibility), and another set of three choices refers to cultural values (opening up to the world vs. independence, equality of chances for
20 foreigners vs. better chances for the Swiss, modernity vs. tradition). A seventh, we thought more neutral item – reforms vs. status quo – was introduced, as well. A factor analysis of these seven items reveals the expected two dimensions, with the economic dimension clearly constituting the first factor, and the cultural dimension the second one. We note that the item for solidarity/selfresponsibility loads on both dimensions, and that the reform/status quo item is mainly associated with the economic, and not with the cultural dimension. This reflects the notion that nowadays, reformers mainly come from the neoliberal camp, while the left has a tendency to privilege the preservation of the achievements of the past. The economic dimension essentially distinguishes a pro-market from a pro-state position, while the cultural dimension refers to the distinction between a traditional, culturally protectionist position and an open, multicultural position5. Table 1A in Appendix 1 provides the results of these factor analyses. The policy core beliefs are also operationalized in two ways. First, we have a scale measuring the actors’ xenophobia. This scale is composed of a set of five items referring to perceived individual/collective economic, cultural and safety threats due to immigration (see Sniderman et al. 2004). These items form a single factor, which we use as our indicator of xenophobia. Next, we use the key issue-specific arguments that have been formulated in the course of the campaign in favor of, and in opposition to the proposal as indicators of the policy core beliefs. In the case of the asylum law, there are two sets of arguments: four arguments in favor of the proposal (abuse in asylum policy has to be limited; asylum policy needs a more efficient implementation; conditions are too attractive for asylum seekers in Switzerland; there are already too many foreigners in the country), and four opposing it (need to maintain the Swiss humanitarian tradition; need to protect the refugees’ basic human rights; tougher procedures are ineffective for the solution of problems in asylum policy; foreigners contribute to the social and cultural quality of the country). In the case of the immigration law, we have three sets of arguments: three arguments in favor of the proposal (need to fight illegal phenomena such as black labour markets, fake marriages, and intermediairies; need to increase legitimacy of 5
Interestingly enough, an analogous analysis at the level of the citizens based on the same set of indicators results in only one factor, which corresponds to the cultural factor at the elite level. Today, the Swiss voters mainly seem to be divided by deep cultural beliefs, while, contrary to the political elite, economic beliefs do not seem to vary systematically among them anymore (see Table 1A in Appendix 1).
21 immigration policy; restriction of admittance to the highly qualified), three against it (opposition to the dual system, to the absence of a solution for the refugees without proper identification, to the coercive measures), and two (integrative measures, facilitation of family reunions) referring to aspects the opposition, at least in principle, supported, too. Respondents could indicate their opinion about these arguments on a five-point scale ranging from (1) ‘do not agree at all’ to (5) totally agree. We have rescaled these arguments from -2 (position of opposing side) to +2 (position of the supporting side), with 0 indicating a neutral position. These arguments constitute a single factor, which we shall call the ‘arguments’ scales’6. The respondents were also asked – for each law separately – to name the three most important, the most important, and the three most unimportant arguments. Their responses to these questions allow us to construct a weighted version of the arguments’ scale. As is indicated by Table A2 in the appendix, weighting the responses does not result in a substantially different structure of the scale. For the operationalization of the secondary beliefs, we use the issue-specific positioning of the actors. For the asylum law, we asked about the six most contested elements: the third country regulation, the social assistance stop, the airport procedure, the coercive measures and the information exchange. For the immigration law, we also asked about the six most salient elements: the dual system, facilitated access to residence permits, reintegration support, facilitation of residential and occupational change, facilitation of family reunions, and coercive measures. Respondents could again respond on a five-point scale ranging from (1) ‘do not agree at all’ to (5) totally agree. For each law, these issue-specific elements again form a single factor, and again, the weighted version of the corresponding scales essentially replicates the unweighted one (see Table A3 in the Appendix). The strategic considerations have been operationalized in an indirect way: starting from the notion that belief-systems of political elites are typically constrained, we take the degree to 6
For the asylum law, the opinions about the same arguments have also been measured for a panel of voters – at the beginning, in the middle, and at the end of the campaign. At each time point, they also constitute single scales at the level of the voters. As one might have expected, however, these scales show more coherence among the elite than among the voters, but interestingly, the coherence among the voters increases in the course of the campaign This can be seen by comparing the Eigenwerte in Table A2 in the appendix, which provides the results of these factor analyses. This increasing coherence among the voters corresponds the ‘enlightening’ effect of the campaign (Gelman and King 1993).
22 which an actor’s issue-specific positioning cannot be predicted by the deep- and policy core beliefs as an indicator of the extent of his strategic posturing. In other words, the residuals of the regression of the issue-specific position scale on the corresponding core beliefs (left-right scale, the two value scales, the xenophobia and arguments’ scales) will serve as a proxy for strategic considerations. Finally, we made an attempt to indirectly operationalize the motivation for the mobilization of the various actors by asking them in the first interview about their expectations with respect to the outcome of the vote for the one of the two laws, which was most salient to them. An actor who expects to loose the vote and still gets involved in the campaign is, according to the reasoning presented above (see Figure 1) motivated by reputation or principles: i.e. he gets involved because he wishes to signal to the members of his constituency that he defends their cause whatever may happen and/or he wishes to signal to the authorities that he will have to be taken seriously, even if he looses this time. We shall distinguish between three groups of actors in this respect: those who expect to loose, those who expect a close outcome (50 +/-3 percent), i.e. who still think they might have a chance to win, but are not sure, and those who count on winning.
Results The different types of actors involved The political parties have been key actors in this campaign, but together, they make up only one third (34 percent) of the actors involved (see Table 2). In addition to the ‘big five’ – the four members of the grand coalition and the Greens, there are 11 minor parties, including three dissident sections of the two moderate right parties from French-speaking cantons who joined the challengers of the asylum law. Although the parties constitute only a minority of the actors involved, the major parties have been key participants in the campaign (see Table 2), as expected for this type of issue, and the SVP – the party of the new populist right – has even been the most powerful actor in the campaign. In addition to the major parties, the five ad hoc
23 committees and the two branches of government have also been very influential. All five committees have been constituted on the challengers’ side, while the supporters of the law, surprisingly, have not created such a committee to coordinate their campaign efforts. Surprisingly, the Federal Ministry of Justice and Police (EJPD), and the Federal Office of Migration (BFM), who are responsible for the Swiss immigration policy, belong to the key actors in the campaign as well. The Swiss government usually adopts a low key approach in directdemocratic campaigns, but in this case, the Minister of Justice – the leader of the SVP – campaigned rather intensively to obtain support for both laws, and he incited his subordinates to do so as well. A considerable number of citizens’ interest groups (12), and church organizations (4) have also participated in this campaign. However, with the exception of the Swiss Aid for Refugees (SFH), they have been of at best moderate importance. As expected, the economic interest groups (including the unions) have only played a marginal role in this domain, where the broad public interests dominate.
Network of actors in the campaign As is shown by Table 3a, the block-model after the first split neatly divides our set of actors into the two camps that confronted each other in the campaign. As expected, the opposing camps are internally connected by multiple ties, but hardly at all linked to each other. It may surprise that there are any collaborative ties between the two at all. Table 3b presents the block-model for the four-block solution, where each camp is again divided into two component coalitions. We find the three components of the tripolar party system, plus the ‘third force’ that resulted from the split of the moderate right. On the challengers’ side, the coalitions of the left is joined by the ‘third force’, while the two components of the political right – the moderate and the new populist right – make up the supporters’ side. The density matrix of the four-block solution reveals three important points: 1) each block constitutes a more or less cohesive coalition – the densities are highest in the diagonal, indicating intense cooperative ties among the members of each one of
24 the four blocks. Compared to the other three blocks, the cooperative relations appear somewhat less dense on the left, which is related to the fact that the left constitutes the largest block. 2) As already observed, there are virtually no relationships between the blocks from opposite camps, as is indicated by the corresponding zero-blocks. Only the moderate right is tenuously connected to the ‘third force’. As it turns out, these ties typically involve actors from the moderate right on both sides: (dissident) moderate right parties (LPS, FDP-GE) and a business association (SGV) on the pro-side maintain ties to four key actors from the ‘third force’. 3) The two components on each side are somewhat connected to each other, but these connections are not very intense. On the challengers’ side in particular, the two coalitions seem to have led their campaigns quite independently. On the supporters’ side, the populist right claims to have been in close cooperation with the moderate right, but the latter did not reciprocate in kind7. In fact, in this particular campaign, the major parties of the moderate right (FDP and CVP) chose not to cooperate with the populist right. In case of a centre-right coalition, they usually form a joint committee with the SVP, with one of the them adopting the leading role, but in the present case, each party led its own campaign. The FDP attempted to distance itself from the SVP by focusing on its own liberal migration concept, while the CVP adopted a low key approach, since the proposal not only gave rise to internal tensions, but also caused conflicts with the Catholic church. The SVP, by contrast, was very active and, in fact, dominated the campaign. Its professionally organized campaigning team leads a permanent electoral campaign, of which direct-democratic campaigns on issues related to immigration are an integral part (Kriesi 2005: 268).
The eight-block solution presented in Table 3c adds some additional detail about the composition of each block and the relationships obtaining between them. Appendix 2 provides an overview over the composition of the eight blocks, including the power and centrality of the various organizations. Basically, each one of the four previous blocks is divided into a core and 7
One organization of the populist right (the EDU, see below) claimed a close collaborative tie to the Swiss Red Cross. Since this single tie led to a less than satisfactory overall solution, it was dropped from the analysis presented in the text.
25 a satellite, where the satellite is internally less connected than it is to the core. This is most obvious on the supporters’ side, while, the satellites on the challengers’ side are somewhat more internally cohesive. On the latter’s side, the core of the left coalition constitutes the largest block. Its satellite is mainly composed of organizations from the French-speaking part of the country. The core is dominated by the most influential organizations of the left campaign – the coordinating committee K2N, and the Social-democratic party. Although the Social-democrats had not mobilized against the immigration law at first, and although they did not consider the vote on this law as very important, they nevertheless ended up in the core of the coalition of those who opposed both laws – quite simply because of their political weight on the left. Relatively unconnected to the orther organizations, the ad hoc committee constituted by cultural professionals and artists represents the third most influential organization in the left core component, followed by the Greens. The remaining, less influential members of the left core include the major unions (SGB, Unia, Travail Suisse), and some issue-specific organizations (sosf, sanspapier, Politakt, FIMM). Although part of the coalition, the unions did not consider this campaign as very important and kept a rather low profile. The existence of a French-speaking satellite points to the relevance of the linguistically segmented public sphere for the coalition formation – an aspect we are not treating here in detail, but which is of considerable importance for the Swiss case. It is coordinated by the CCEX, an actor explicitly set up by the left for the coordination of the campaign in the Frenchspeaking part of the country. It includes several relatively marginal organizations who are all attached to the CCEX, but have hardly any additional ties to other parts of the left campaign structure: two radical left parties who are exclusively present in the French-speaking part of the country – the Communist party (PdA) and solidaritéS (an alliance of radical left mouvements), two dissident sections of the Christian-Democrats (CVP-VD and CVP-GE), a union (Comedia), as well as an issue-specific citizens’ interest group from Geneva (agora) 8.
8
The EVP – the Evangelical People’s Party, a minor party who is exclusively based in the Germanspeaking part of the country – is also part of this satellit component. It was also a member of the K2N, but not otherwise connected to any of the organizations in this component. It ended up in the satellite component by default.
26 The ‘third force’ is also divided into a core component and a satellite. Its core includes three highly influential organizations: the coordinating committee of the ‘third force’ (BK) as well as the Swiss aid for refugees (SFH) and the committee (KHS), which the SFH had created together with the Social-democrats to collect signatures against the asylum law. These are the key organizations of the campaign exclusively focused on the asylum law. Amnesty international, the aid-organization of the protestant church (HEKS), both modestly influential and well connected to the other organizations, and a Jewish aid organization (VSJF) complete this block. The second component of the ‘third force’ corresponds to a church block, which is internally more coherent than the other satellites and includes the moderately influential peak organizations of the Catholic (SBK) and the Protestant church (SEK) of Switzerland, as well as the Swiss Red Cross (SRK), who also proves to have been quite influential. A religious organization (CJLA) and a regional branch of the Protestant church (ERLZH) complete this block. On the supporters side, the core block of the moderate right is composed of the corresponding two major parties (CVP and FDP), and of the two branches of government who were in charge of the issue in question. It is interesting to note that the two influential branches of the federal government (EJPD and BFM) are rather closely related to the moderate right parties, although the actively campaigning minister in charge of the EJPD was the leader of the SVP at the time. The satellite of this very cohesive block of the moderate right is composed of two peak business interest associations – the employers’ association (SAV) and the association of small businesses and trade (SGV), as well as of the small Liberal party (LPS), and of the Genevan branch of the Radical party (FDP-GE). These organizations had a rather marginal status in the campaign on the asylum law. They cared much more about the immigration law. The block of the new populist right is the smallest one. It only includes four actors. Key among them is the SVP, the major promoter of a tough immigration legislation in Switzerland, and, as we already mentioned, the most influential actor in this campaign. It is accompanied by the AUNS (Association for an independent and neutral Switzerland), an advocacy group closely associated with the SVP that has been founded in 1986 and that mobilizes against the integration of Switzerland into the international political community, as exemplified by the UN
27 and the EU. Surprisingly, the AUNS did not consider the two laws involved here as very important. The satellite of this block includes the youth organization of the AUNS, and a small party of the evangelical fundamentalists in German-speaking Switzerland (EDU), two very marginal and powerless organizations.
The beliefs of the actors in the campaign The values and beliefs of the key actors involved in the campaign, indeed, form a highly integrated set of ideas, as is indicated by the correlation matrix in Table 4. There are, however, two exceptions to the generally high correlations (generally>.70): on the one hand, the economic values are considerably less connected to the overall set of beliefs, which supports the expectation that asylum and immigration policies are mainly driven by cultural concerns, and not by economic considerations. On the other hand, the issue-specific positions with regard to the immigration law are less closely linked to the deep core and policy core (xenophobia) beliefs than those with regard to the asylum law. This suggests that the asylum law was the more critical issue of the two.
The belief structures of the component coalitions allow us to characterize them in terms of the distinctions made in Table 1. Figure 2 provides the mean factor values of the four component coalitions’ beliefs at each one of the three levels. As expected, the two camps confronting each other in the campaign are clearly distinguished by their beliefs at all three levels. In fact, on the basis of economic values and issue-specific secondary beliefs alone, it is possible to perfectly predict, which side of the campaign an actor was on with respect to both laws. However, while the two component coalitions on either side largely share the same secondary beliefs, they are distinct with regard to the core beliefs. This is particularly striking on the side of the laws’ supporters: they have different values at the level of the deep core, and they also hold quite distinct policy-core beliefs. Thus, the populist right takes an extremely protectionist and traditional position on the cultural value dimension. It is also extremely xenophobe and strongly
28 believes in tougher measures to render Switzerland less attractive for refugees. The moderate right, by contrast, is culturally moderate at the level of the deep core, not at all xenophobe, and its support for the arguments in favor of the two laws is also quite moderate. In fact, considering its cultural beliefs, the moderate right is located more closely to the challengers of the law than to its allies from the populist right. It is its decidedly pro-market position, which distinguishes it from the challengers, and from its populist allies. In sum, the two components on the supporters’ side constitute a clear example of a ‘pragmatic coalition’.
On the challengers’ side, the two components not only share the secondary beliefs, but also the policy-core beliefs. Surprisingly, there are little differences in this respect between the two laws, although the third force only mobilized against the asylum law. Where the challengers’ components differ to some extent is at the level of the deep core beliefs, with respect to the economic value dimension. The third force did not quite share the left’s position in traditional economic terms. Therefore, although the two component coalitions on the challengers’ side share the policy-core beliefs, they were not able to organize a joint campaign, and their combined effort resembles a ‘pragmatic alliance’ once again. This becomes apparent, too, when we consider the actors’ self-placement on the left-right scale in Table 5, which constitutes a kind of summary indicator for their deep core beliefs. The two camps are quite distinct in this respect, as are the two component coalitions on either side – the left is distinctly more to the left than the third force, and both of them are more to the left than the moderate right, which in turn is clearly more moderate than the new populist right. Both the third force and the moderate right are located close to the centre (=5) of the scale, the former to the centre-left (4.3), the latter to the centre-right (6.3), while the left (2.7) and especially the populist right (8.3) take quite extreme positions. Except for the moderate right, the satellites always resemble their core on this dimension.
29 A more detailed analysis of the individual arguments about the asylum law reveals that, surprisingly, the key campaign arguments of both sides are not the most divisive ones (Figure 3a): the supporters of the law obviously brand the abuse of the existing legislation by ‘false refugees’, but the third force is also in favor of fighting against such abuses, and even the left is not clearly opposed to such measures. The challengers are, however, less convinced that the previous legislation has been ineffective in fighting such abuses, and they are very much opposed to the belief that tougher measures would provide effective solutions to the problems in asylum policy. Moreover, they do not believe at all that Switzerland is too attractive for asylum seekers. It is this belief that tougher measures provide effective solutions to the current problems, which most clearly divide the two camps. By contrast, the actors from both camps, including the populist right, are convinced that the Swiss humanitarian tradition needs to be preserved. As it turns out, this is the only argument that neither discriminates between the two camps, nor between the four component blocks in any significant way (see Appendix 3a). The moderate right fully supports the application of this general principle to the basic human rights of the refugees, too, while the populist right still pays some lip service to the preservation of their rights. As far as the moderate right is concerned, the support of the humanitarian tradition reflects its deep core cultural beliefs. In the case of the populist right, however, this support is much less credible. It not only contrasts with its deep core cultural beliefs, but also with its extreme xenophobia, and its policy-core beliefs in the effectiveness of tougher measures to render Switzerland less attractive for refugees. A more detailed analysis of the secondary beliefs shows that all six issues give rise to similar contrasts with regard to the asylum law.
For the immigration law (Figure 3b), it comes as less of a surprise that two of the arguments in favor of the law – support for the integration of foreigners, and for the facilitation of family reunion – are also backed by its challengers. Together with the third force and the moderate right (but against the populist right), the left strongly supported integration measures and family reunions, and it was not really opposed to fighting illegal practices. It is at the level of the concrete implementation of these principles, where the left had reservations about the new law.
30 While it supported the facilitation of occupational and residential changes as well as the facilitation of a return to the home countries, with regard to family reunions, together with the third force, it parted company with the moderate right, who had opted for a solution that, as already mentioned, was not really a step in the direction of liberalization. Moreover, it was strictly opposed to the coercive measures, to the dual system and some other aspects of the new law. To summarize this discussion, we can demonstrate the embeddedness of the argument-related beliefs and the issue-positions in the deep core beliefs by straight-forward OLS-regression analyses. The results are presented in Table 6. They allow us to make four points: first, the crucial importance of the cultural values in this policy domain is confirmed: for both laws, cultural values prove to be more decisive determinants of policy-core beliefs than economic values (columns ‘1’ in Table 6). Note that self-positioning on the left-right scale does not contribute to the argument-related beliefs, because it is very closely associated with the two value dimensions9. Second, economic values are only relevant for the moderate right, especially for its issue-specific positioning (columns ‘2’, which introduce an interaction term between economic values and the moderate right). Contrary to the other participants in the campaign, the moderate right seems to have been considerably motivated by economic reasons. Third, xenophobia, which is closely associated with the cultural, but not with the economic value dimension (Table 4), proves to have been particularly relevant for the determination of argument-related beliefs in the case of the asylum law. It was less relevant for argument-related beliefs in the case of the immigration law, and irrelevant for the issue-specific positioning with respect to both laws (also columns 2). More than the immigration law, the asylum law apparently touched upon the policycore beliefs. However, even in this case, the actors’ issue-specific positioning is not related to xenophobia. Finally, the argument-related beliefs and the issue-positions are strongly determined by the (deep) core beliefs, as is indicated by the very sizeable values of R2. This indicates that the actors mainly pursued policy-specific goals in this particular campaign and that, for most of them, strategic considerations unrelated to this specific vote, were at best of secondary importance. 9
Jointly, they explain 80% of the variance of the self-positioning on the left-right scale.
31
Motivation for mobilization and strategic considerations Table 7 presents the expected outcome of the vote for the four component coalitions. As is perfectly clear from this table, the uncertainty in this particular case was not overwhelming. Given the outcome of the previous votes, i.e. given the known long-term issue-specific distribution of the population’s preferences, the pro-side generally expected to win. On the supporters’ side, no actor expected a clear defeat and only two organizations expected a close race. Surprisingly, the organizations who were uncertain about the outcome of the vote on the supporters’ side include the most influential actor in this campaign – the SVP, and the most important party of the moderate right – the FDP. The SVP hoped for a success of the two laws at the polls, since they had been forged under the pressure it had exerted in Parliament. Its goal was to harvest the fruits of its long-term campaign for a tougher immigration policy, and to confirm its victorious image in the light of the upcoming elections in Fall 2007. But it was not quite sure about its chances at the polls. Similarly, the FDP was not entirely sure about whether the laws would be adopted. Moreover, for the FDP a victorious outcome was not yet sufficient enough: the victory at the polls would also have to be attributed to its own labours in order to be of any value at all. The two laws that had finally been adopted by Parliament owed a lot to the efforts of several members of the FDP parliamentary group, but the FDP was rightly preoccupied by the possibility that a victory at the polls would essentially be attributed to its populist competitor, the SVP. On the challengers side, one mostly expected that a majority of the citizens would vote in favor of the laws as well. Only minor actors on the left, actors with little experience in federal votes hoped for a close race or even a success of the opposition. On the left, the general feeling was that any result that was better than the outcome of the previous votes on immigration policies would be a success. One hoped for as good a result as possible without expecting to win. The unions hoped to signal to their members (among whom are many foreigners) that they mobilized
32 on their behalf, they hoped to be present in the media. For the Social-democrats, too, it was important to signal to their electorate that they were committed to the defense of the humanitarian tradition of Switzerland and that they did not accept fundamental attacks on that tradition as exemplified by the new asylum law. In addition, they hoped to make some inroads into the electorate of the moderate right, anything beyond the electorate of the left (roughly 30 percent) for them would have been a success. A union organizer hoped that a relatively large number of ‘no’-votes would strengthen the position of the challengers in the later stages of the implementation of the two laws, which still left a lot of details to be specified. Some organizations on the left considered it already a success that they had been able to collect the required number of signatures for the vote to take place. Others thought it was already a success that their point of view could be heard in public. Some hoped for a defeat of the law at least in Geneva, or at least in the French-speaking part of the country, and in the larger cities. Some took a more long-term view and considered this campaign in the context of a long-term mobilization in support of the immigrants. Expectations were somewhat higher among the organizations of the third force, half of whom expected a close race. Among those hopefuls, we find the outsider who had created the ‘Bourgeois coalition’ against the asylum law. He was the CEO of a major Swiss company, but politically not active. In the run-up to the campaign, he had written a letter to the editor, which got him a lot of media attention, and he was contacted for support by the Swiss Aid for Refugees. In reaction, he took it upon himself to create the ‘Bourgeois coalition’, very much in line with the notion that ‘if I don’t do it, nobody else will’, which reflects the pessimism of activists, who are willing to absorb the costs of mobilizing on behalf of a common public good (Oliver 1984). The other influential members of the third force did not want to specify their expectations, but they were also quite hopeful with respect to the final score for the asylum law. The christian members of this coalition wanted to mobilize the churches for a ‘no’-vote and to prevent any further turning of the screw with as a high a share of ‘no’-votes as possible. Overall, however, it is fair to say that the challengers were mainly motivated by reputation and principle, and to only a limited extent by the expectation that they could win the campaign.
33 Adding the arguments’ scale to the second model for the actors’ issue-position on the asylum law in Table 6, we have calculated the corresponding residuals as an indicator for the actors’ strategic considerations. Given that the issue-specific positions are closely embedded into the core beliefs, such considerations did not generally play an important role in this particular vote. However, for a limited number of actors, the residuals turned out to be sizeable, i.e. considerably larger than half a standard deviation. There are five actors concerned, among whom four of the 12 members of the moderate right. The Federal Migration Office (BFM) and the Christian-democrats (CVP) are characterized by the largest residuals, which are both positive, indicating that the issue-positions of these two actors were more to the right than we would have expected on the basis of their core-beliefs. As far as the Christian-democrats are concerned, they, indeed, took a surprisingly tough position in this specific policy-proposal, apparently with an eye to the upcoming federal elections: for electoral reasons, they did not want to leave the field of immigration policy over to the SVP, even though they were considerably more moderate in their overall outlook than they chose to position themselves in this particular case. As for the Federal Migration Office, we can only speculate that many staff members of this office, who had been recruited by the previous minister (who had also belonged to the CVP), were much less radical than their current minister and his policy priorities, but had to remain loyal to the latter’s proposal in public (and in our interviews). The only actor on the challengers’ side who appears to have been strongly strategically oriented is, surprisingly, the committee K2N. It turns out to have been more critical of the proposal than one would have expected on the basis of its core beliefs. This is likely to be a result of the fact that the person responsible for the campaign of the K2N, whom we had interviewed on its behalf, was a professional campaigner who had been employed especially for this specific campaign. His core beliefs may not have reflected the core beliefs of the committee’s regular members. While this specific case raises the issue of the adequacy of our selection of the interview partners, it constitutes the exception that serves to confirm the overall result of the limited importance of strategic considerations in this particular campaign.
34 Finally, we have made an attempt to predict the block-membership on the basis of the issuepositioning, the motivation for mobilization (the expected closeness of the outcome), and the magnitude of strategic considerations (independent of sign). This analysis is restricted to the first three blocks. The populist right not only is too small (only four cases), but it is also clearly distinct from the other three blocks by its extreme core beliefs. For the other three blocks, the analysis is possible and makes a lot of sense: on the one hand, the question is whether the members of the moderate right are really distinct from the third force with regard to their issuespecific positioning, or whether the moderate right joined forces with the populist right for purely opportunistic reasons. On the other hand, the question is whether the third force was more opportunistic than the left, not only believing in the possibility of a victory, but also positioning itself strategically in view of a possible victory. Table 7 reports the results. As it turns out, the differences between the moderate right and the third force are not significant. However, given that the number of actors involved is rather limited, we should note that, even if not significant, the differences go in the expected directions: the moderate right is positioned closer to the right, it is more strategically-minded, but not at all different from the third force as far as the expected closeness of the vote is concerned. By contrast, the left differs significantly in all three respects from the third force: it is more critical of the law, it was without illusions about the outcome of the vote, and it was more strategically minded than the third force. From this comparison, the latter appears to have been somewhat naïve, getting involved for a cause that it believed could be won, but free of any strategic consideration.
Conclusion Our exploratory analysis of the coalition formation in the direct-democratic campaigns for the asylum law and the closely related immigration law has confirmed the usefulness of the basic theoretical ideas and the empirical tools we introduced for such an analysis. As we pointed out in the beginning, the institutional setting, the belief systems and considerations for launching a challenge prove to be crucial for the configuration of actors that takes shape in the run-up to such a campaign. With regard to the institutional setting, the binary institutional logic of the vote
35 imposed itself in the present case, although the situation was complicated by the presence of two related, but not equally challenged pieces of legislation. While the opposition initially tried to make a distinction between the two laws, the binary logic of the vote imposed the formation of two camps in spite of all the substantive considerations that the actors involved tried to make. In each one of the camps, however, the heterogeneity of the alliances required to win a vote manifested itself in the formation of component coalitions, each of whom, in turn, was based on some distinct sets of beliefs. Political parties are generally very important actors in the coalition formation in Swiss direct democratic campaigns. In the policy domain of immigration, where the broad-based interests involved are mainly defended by political parties, and where narrower based economic interest groups are less concerned, the importance of parties for coalition formation can be assumed to be even greater than usual. The parliamentary alliances between the parties are typically reproduced in the course of the campaign. But, as our case study very clearly demonstrated, even in a policy domain that is very favorable to partisan mobilization, the parties made up only a minority of the actors involved. The coalitions on both sides of the campaign included a variety of actors, especially a sizeable number of citizens’ interest groups and churches defending the immigrants’ cause on the side of the challengers. Most surprisingly, the supporters were joined by two branches of the federal administration who played a key role in the campaign. As expected, the campaign effort was coordinated by several ad hoc committees on the challengers’ side. On the supporters’ side, however, no such committee was formed because of the reticence of the two major parties of the moderate right. As expected, the heterogeneity of the participants led to the constitution of component coalitions on either side – component coalitions that formed what we have proposed to call ‘pragmatic alliances’. The pro-side was composed of the two components of the political right – the moderate and the populist right. These two components differed considerably in terms of their cultural deep-core and policy-core beliefs. However, they shared the concrete secondary beliefs which have been implemented in the two laws. Pressured by the populist right, the moderate right accepted the details of the new laws for pragmatic, strategic, and, not in the least,
36 economic reasons, although, in cultural terms, it proved to be closer to the challengers of the two laws. In principle, this made the moderate right particularly vulnerable to the challenge, as is illustrated by the constitution of a ‘third force’, and by the dissidence of some of its sections in the French-speaking part of the country. While the third force shared the policy-core and the deep core cultural beliefs of the left, it differed from the latter in economic terms, as well as in terms of its more general political background. This is why it could not openly cooperate with the left, and the two components of the opposition, too, mobilized ‘apart together’. At the level of the concrete issue-positions concerning both laws, however, the two components showed surprisingly little differences, although the one only mobilized against the asylum law, while the other opposed both laws. Since the challengers have a first mover advantage in direct-democratic campaigns, their motivation to mobilize is particularly important. In the present case of a highly conflictual, mainly culturally determined policy, the challengers were not so much motivated by the chances of success – in their overwhelming majority they did not expect to win the vote, but by questions related to their reputation (credibility) and fundamental principles. Although many in the third force believed they had a chance to win the vote, their involvement was not in any way strategically oriented, but rather motivated by a combination of a somewhat naïve belief in the possibility of success and a sincere belief in the core principles of the Swiss humanitarian tradition.
37 Bibliography Arx, von Nicolas 2002. Ähnlich, aber anders. Die Volksinitiative in Kalifornien und in der Schweiz. Basel: Helbing und Lichtenhahn. Barry, Brian 1975. ‚Review Article: Political Accomodation and Consociational Democracy’, British Jorunal of Political Science 5: 477-505. Betz, Hans-Georg 2004. La droite populiste en Europe: extrème et démocrate, Paris: Autrement. Betz, Hans-Georg 1993. The New Politics of Resentment: Radical Right-Wing Populist Parties in Western Europe. Comparative Politics 25, 4: 413-427. Boehmke, Fredrick 2002. ‘The Effect of Direct Democracy on the Size and Diversity of State Interest Group Populations’, Journal of Politics 64: 827-44. Bowler, Shaun, and Todd Donovan 2006. ‘Direct Democracy and Politial Parties in America’, Party Politics 12, 5: 649-670. Budge, Ian 2001. ‘Political Parties in Direct Democracy’, pp.67-87 in Referendum Democracy. Citizens, Elites and Deliberation in Referendum Campaigns, edited by Mathes Mendelsohn and Andrew Parkin, Basingstoke: Palgrave. Budge, Ian 1996. The New Challenge of Direct Democracy. Cambridge: Polity Press. Clark, Sherman J. 1998. ‘A Populist Critique of Direct Democracy’, Harvard Law Review 112: 434-482. Converse, Philip E. 1964. ‘The Nature of Belief Systems in Mass Publics’, pp. 206-61 in Ideology and Discontent, edited by David Apter. New York: The Free Press. Cronin, Thomas 1989. Direct Democracy: The Politics of Initiative, Referendum, and Recall. Cambridge: Harvard University Press. Donovan, Todd, Shaun Bowler, David McCuan, and Ken Fernandez 1998. ‘Contending Players and Strategies: Opposition Advantages in Initiative Elections’, pp. 80-104 in Citizens as Legislators. Direct Democracy in the United States, edited by Shaun Bowler, Todd Donovan and Caroline J. Tolbert. Columbus: Ohio State University Press. Golder, Sona Nadenichek 2006. ‘Pre-Electoral Coalition Formation in Parliamentary Democracies’, British Journal of Political Science 36: 193-212. Jasper, James M. 2006. Getting Your Way. Strategic Dilemmas in the Real World. Chicago: The University of Chicago Press. Kitschelt, Herbert in collaboration with Anthony J. McGann 1995. The Radical Right in Western Europe. Ann Arbor: The University of Michigan Press. Knoke, David (1990). Political Networks: The Structural Perspective. Cambridge: Cambridge U Press. Knoke, David, Franz U. Pappi, Jeff Broadbent and Y. Tsujinaka 1996. Comparing Policy Networks: Labor Politics in the U.S., Germany, and Japan. Cambridge U Press. Kriesi, Hanspeter 2006.’The Role of the Political Elite in Swiss Direct-Democratic Votes’, Party Politics 12,5: 599-622. Kriesi, Hanspeter 2005. Direct Democratic Choice. The Swiss Experience. Lanham, Md: Lexington. Kriesi, Hanspeter 2005a. ‚Zusammenfassung und Schlussfolgerung’, pp. 256-70 in: Der Aufstieg der SVP. Acht Kantone im Vergleich, edited by Hanspeter Kriesi, Romain Lachat, Peter Selb, Simon Bornschier, Marc Helbling. Zürich: NZZ-Verlag. Kriesi, Hanspeter 1980. Entscheidungsstrukturen und Entscheidungsprozesse in der Schweizer Politik. Frankfurt: Campus.
38 Kriesi, Hanspeter, Edgar Grande, Romain Lachat, Martin Dolezal, Simon Bornschier, Tim Frey 2008. West European Politics in the Age of Globalization. Cambridge University Press (forthcoming). Kriesi, Hanspeter, Silke Adam and Margit Jochum 2006. “Comparative analysis of policy networks in Western Europe”, Journal of European Public Policy 13, 3: 341-361. Kriesi, Hanspeter and Maya Jegen (2001). „The Swiss energy policy elite: The actor constellation of a policy domain in transition“, European Journal of Political Research 39: 251-287. Ladner, Andreas and Michael Brändle 1999. ‘Does Direct Democracy Matter for Political Parties? An Empirical test in the Swiss Cantons’, Party Politics 5, 3: 283-302. Laumann, Edward O. and Franz U. Pappi 1976. Networks of Collective Action: a Perspective on Community Influence Systems. New York: Academic Press. Laumann, Edward O. and David Knoke 1987. The Organizational State. Social Choice in National Policy Domains. Madison: The University of Wisconsin Press. Neidhart, Leonhard 1970. Plebiszit und pluralitäre Demokratie. Bern: Francke. Oliver, Pamela 1984. ‘’If you don’t do it, nobody else will’: active and token contributors to local collective action’, American Sociological Review 49: 601-610. Ossipow, William 1994. ‘Le système politique suisse ou l’art de la compensation’, pp. 9-56 in Elites politiques et peuple en Suisse. Analyse des votations fédérales: 1970-1987, edited by Yannis Papadopoulos. Lausanne: réalités sociales. Papadopoulos Yannis, 1991, “Quel rôle pour les petits partis dans la démocratie directe?”, Annuaire suisse de science politique 31, 131-50. Riker, William H. 1962. The Theory of Political Coalitions. New Haven: Yale University Press. Sabatier, Paul A. and Christopher M. Weible 2007. ‘The Advocacy Coalition Framework: Innovations and Clarifications’, pp. 189-220 in Theories of the Policy Process, 2nd edition, edited by Paul A. Sabatier. Boulder, Col: Westview. Sciarini, Pascal and Alexander H. Trechsel 1996. ‘La démocratie directe en Suisse: L’élite politique victime des droits populaires?’, Revue suisse de science politique 2, 3: 201-232. Sinnott, Richard 2002. ‚Cleavages, Parties and Referendums: Relationships between Representative and Direct Democracy in the Republic of Ireland’, European Journal of Political Research 41: 811-826. Smith, Daniel A. and Caroline J. Tolbert 2004. Educated by Initiaive. The Effects of Direct Democracy on Citizens and Political Organizations in the American States. Ann Arbor: The University of Michigan Press. Smith, Daniel A. and Caroline J. Tolbert 2001. ‘The Initiative to Party. Partisanship and Ballot Initiatives in California’, Party Politics 7, 6: 739-757. Sniderman, Paul A. 2000. ‘Taking Sides : A Fixed Choice Theory of Political Reasoning », pp. 67-84 in Elements of Reason: Cognition, Choice, and the Bounds of Rationality, edited by Arthur Lupia, Mathew D. McCubbins, and Samuel P. Popkin. Cambridge, UK: Cambridge University Press. Sniderman, Paul M., Louk Hagendoorn, and Markus Prior 2004. ‘Predisposing Factors and Situational Triggers: Exclusionary Reactions to Immigrant Minorities’, American Political Science Review 98, 1: 35-50. Trechsel, Alexander H. and Pascal Sciarini 1998. ‘Direct Democracy in Switzerland: Do Elites Matter?’, European Journal of Political Research 33: 99-124. Van der Brug, Wouter and Meindert Fennema (2003). ‚Protest or mainstream? How the European anti-immigrant parties developed into two separate groups by 1999’, European Journal of Political Research 42, 1: 55-76.
39 Van der Brug, Wouter, Meindert Fennema, Jean Tillie (2005). ‚Why some anti-immigrant parties fail and others succeed: A two-step model of aggregate electoral support’, Comparative political studies 38, 5: 537-73. Van der Brug, Wouter, Meindert Fennema and Jean Tillie (2000). ‚Anti-immigrant parties in Europe : Ideological or protest vote ?’ European Journal of Political Research 37 : 77-102. Walder, Andrew G. 2006. ‘Amiguity and Choice in Political Movements: The Origins of Beijing Red Guard Factionalism’, American Journal of Sociology 112, 3: 710-50. Wassermann, Stanley and Katherine Faust 1994. Social Network Analysis. Methods and Applications. Cambridge University Press.
40 Table 1: Coalition types, based on shared beliefs Secondary beliefs
shared Not shared
Core beliefs shared
Not shared
natural alliance
pragmatic alliance
ideological alliance
unnatural alliance
41 Figure 1: Decision-tree for possible challengers (optional referendum) Bill adopted by Parliament: Is it sufficiently different from the status quo (SQ)?
No: Does it block further Reform (too small a step)? No: Stop
Yes: Yes: Is there a chance to win the vote?
No: Is there a reputation to defend?
No: stop
Yes: Is it worse than the SQ? (too radical)
Yes: Are the required Resources/partners available?
Yes: No: are the reqiured stop resources/partners available?
No: Stop
No: stop
yes: launch referendum
Yes: Launch Referendum
42 Table 2: Distribution of actor types, and their influence Actor type big 5 parties other parties ad hoc committees economic interest groups citizens' interest groups public administration churches+ total
frequency abs % 5 10.6 11 23.4 5 10.6
mean 31.4 4.1 39.2
s.d. 26.7 2.5 12.0
7 12 2 5 47
3.0 11.8 51.5 12.6 15.4
3.4 12.4 9.2 9.5 18.1
14.9 25.5 4.3 10.6 100.0%
power
43 Table 3: Density matrices1) a) 2-block solution Origin of tie contra pro
Target of tie contra 0.33 0.03
pro 0.01 0.52
b) 4-block solution target of tie origin of tie left Third force moderate right populist right
1) These densities are calculated for the weighted cooperative ties, i.e. they do not represent the share of possible ties which are present, but a relative density measure that may take on values greater than 1.
44 Table 4: Correlations between the various beliefs (n=42)
left left economic values cultural values xenophobia arguments-asylum argumentsimmigration position-asylum position-immigration
economic cult Xenovalues values phobia
ArguArguments ments asylum imm
. 0.75 0.65 0.68 0.80
. 0.18 0.38 0.57
. 0.81 0.75
. 0.84
.
0.81 0.79 0.72
0.62 0.62 0.63
0.71 0.70 0.54
0.76 0.71 0.46
0.90 0.87 0.74
. 0.92 0.81
position asylum
. 0.86
45 Figure 2: Beliefs of the four component coalitions: mean factor scores a) deep core: political values
mean factor scores
2 1.5 1
left third force
0.5
mod right
0
populist right
-0.5 -1 economic values
cultural values factors
b) policy core: arguments and xenophobia
mean factor score
2.5 2 1.5
left
1
third force moderate right
0.5
populist right
0 -0.5 -1 arguments asylum
arguments immigration
xenophobia
factors
c) secondary beliefs: issue positions on the two laws 2
mean score
1.5 1
left third force moderate right populist right
0.5 0 -0.5 -1 asylum
immigration factors
46 Table 5: Self-positioning on the left-right scale, means (on a scale ranging from 0 to 10) Camp block2 block4 block8 against 3.0 left 2.3 satellite 2.8 core left 2.0 third force 4.3 core 4.5 churches 4.0 pro 7.0 mod right 6.4 satellite 7.0 core 5.8 populist right 8.3 satellite 7.5 core 9.0
47 Figure 3: Policy-core beliefs – position of component coalitions on individual arguments: mean scores and overall standard deviations (in parentheses) a) asylum 2.0 1.5 mean score
1.0 left third force moderate right populist right
0.5 0.0 -0.5 -1.0 -1.5
ab us e
(1 .2 ef 8) fe ct iv e to (1 o .4 at 9) tra ct iv e (1 in .5 ef 9) fe ct iv e (1 .5 to 9) o m an y (1 .2 3) qu al ity (.9 ba 6) si c rig ht s (.9 3) tra di tio n (.9 0)
-2.0
arguments
b) immigration
2.0 1.0 left third force moderate right populist right
0.5 0.0 -0.5 -1.0 -1.5 -2.0 ille ga l( 1. le 30 gi tim ) ac y (1 .5 de 3) te nt io n du (1 .6 al 1) sy st em sa (1 ns .4 0) pa pi er hi s gh (1 ly .4 qu 4) al ifi ed (1 in .3 te 6) gr at io n fa (1 m .1 ily 6) re un io n (.8 9)
mean score
1.5
argument
48 Table 6: Determinants of positions on argument scales: OLS-regressions, unstandardized regression coefficients, standard errors (in italics), levels of significance 1) Determinants
49 Table 7: Expected outcome of the vote by component coalition: Percentages expecting a ‘yes’vote Expected ‘yes’vote of less than 47% 47 to 53% more than 53% total n
left 5.0 20.0 75.0 100.0% 20
third force 0.0 50.0 50.0 100.0% 8
moderate right 0.0 14.3 85.7 100.0% 7
populist right 0.0 25.0 75.0 100.0% 4
all 2.6 25.6 71.8 100.0% 39
Table 8: Determinants of component coalition membership: multinomial regressions, unstandardized regression coefficients, standard errors (in italics), levels of significance1) Determinant constant issue position (asylum)
50 Appendix 1: Factor analyses Table A1: Factor analysis for deep core beliefs, for campaign elite and voters: factor loadings of original items1) Item
elites factor 1 0.88 0.71
factor 2 . .
voters factor 1 . .
state/market income equality Solidarity/selfresponsibility 0.66 0.52 0.46 Reform/status quo 0.66 . . Opening/independence . 0.76 0.65 Equal/unequal chances . 0.70 0.54 Modernity/tradition . 0.50 0.46 Eigenvalue 2.92 1.16 1.23 1) only loadings greater than .4 are presented in the table Table A2: Factor analysis for policy core beliefs (arguments), for campaign elite and voters: factor loadings of original items1) a) asylum elites
voters start of Midunweighted weighted campaign campaign abuse 0.57 0.87 0.59 0.63 effective 0.66 0.75 0.51 0.54 too attractive 0.88 0.78 0.72 0.75 too many 0.85 0.56 0.72 0.69 ineffective -0.68 -0.58 . . quality -0.78 . -0.53 -0.57 basic rights -0.77 -0.75 -0.45 -0.45 tradition . . . . Eigenvalue 3.95 3.25 2.19 2.38 1) only loadings greater than .4 are presented in the table b) immigration unweighted weighted 0.46 0.84 0.40 0.53
illegal law highlyqualified only 0.64 0.58 Coercive measures -0.90 -0.71 Dual system -0.92 -0.77 sanspapier -0.93 -0.82 integration -0.63 . children -0.53 . Eigenvalue 3.97 3.19 1) only loadings greater than .4 are presented in the table
End of campaign 0.69 0.60 0.77 0.71 . -0.55 -0.41 . 2.59
51 Table A3: Factor analysis for secondary beliefs (issue positions), for campaign elite: factor loadings of original items1) asylum Third country
unweighted weighted 0.83 0.55
immigration Dual system Permanent residence
social assistance 0.91 0.78 Airport procedure 0.71 . return aid id papers 0.93 0.86 occ-resid change detention 0.95 0.86 Family reunion data 0.91 0.64 Coercive measures Eigenwert 4.6 2.95 Eigenwert 1) only loadings greater than .4 are presented in the table
unweighted weighted 0.83 0.78 0.85
0.63
0.43 0.44 0.85 0.77 3.12
. 0.44 0.80 0.78 2.58
52 Appendix 2: Overview over the organizations involved in the campaign Name Block1: satellite of French-speaking left Christlichdemokratische Volkspartei des Kantons Waadt Partei der Arbeit Comedia Aumônerie Genevoise œcuménique auprès des Requérants d' Asile Evangelische Volkspartei solidaritéS Christlichdemokratische Volkspartei des Kantons Genf Coordination contre l'exclusion et la xénophobie Block2: left core Politakt Travail.Suisse JungsozialistInnen Schweiz Schweizerischer Gewerkschaftsbund Forum für die Integration von Migrantinnen und Migranten (FIMM) Sans-papiers Kollektiv Unia Solidarité sans frontières Grüne Partei der Schweiz Kulturschaffende gegen das Asylgesetz Sozialdemokratische Partei der Schweiz Komitee 2x Nein Block3: core of third force Verband Schweizerischer Jüdischer Fürsorgen Amnesty International Hilfswerk der Evangelischen Kirchen der Schweiz Koalition für eine humanitäre Schweiz Schweizerische Flüchtlingshilfe Bürgerliches Komitee gegen das Asylgesetz Block4: churches Evangelisch-Reformierte Landeskirche des Kantons Zürich Chrétiens et juifs pour la liberté d'aider Schweizerischer Evangelischer Kirchenbund Schweizerisches Rotes Kreuz Schweizerische Bischofskonferenz Block5: satellite of moderate right Schweizerischer Arbeitgeberverband Schweizerischer Gewerbeverband Freisinnig-Demokratische Partei des Kantons Genf Liberale Partei der Schweiz Block6: core moderate right Freisinnig-Demokratische Partei Christlichdemokratische Volkspartei Bundesamt für Migration Eidgenössisches Justiz- und Polizeidepartement Block7: satellite of new populist right Junge für Freiheit, Unabhängigkeit und Neutralität Schweiz Eidgenössisch-Demokratische Union Block8: core new populist right
Abbreviation
Influence
Indegree
CVPVD PdA Comedia
3 4 4
2 4 5
Agora EVP solida CVPGE CCEX
4 5 6 6 24
1 1 4 3 13
Politakt TS JUSO SGB
0 1 4 7
0 1 2 5
FIMM Sanspap Unia sosf GPS Kultur SPS K2N
8 8 8 11 18 29 33 51
9 6 12 13 11 5 11 19
VSJF ai HEKS KHS SFH BK
5 18 18 44 46 48
4 10 8 15 14 11
ERLZH CJLA SEK SRK SBK
3 6 16 16 25
0 4 8 4 8
SAV SGV FDPGE LPS
0 1 5 9
2 0 2 2
FDP CVP BFM EJPD
11 18 45 58
5 3 9 4
Y4fun.ch EDU
1 1
1 0
53 Aktion für eine unabhängige und neutrale Schweiz AUNS 16 2 Schweizerische Volkspartei SVP 77 6 Appendix 3: Results of t- and F-tests for the differences between the positioning of the blocks on the weighted and unweighted argument scales: t- and F-values, and levels of significance1) a) aylum argument
