From Plan B to Plan D : The Challenge of the Constitutional Referendums *

Transcription

1 20 th IPSA World Congress Fukuoka, Japan 10 July 2006 From Plan B to Plan D : The Challenge of the Constitutional Referendums * Min Shu Centre for European Studies & School of Economics Fudan University Min.Shu@fudan.edu.cn Abstract Opinion polls show that most voters who rejected the Constitutional Treaty in the French and Dutch referendums wanted a better constitutional text. The negative result, they thought, would allow an alternative constitution (i.e., Plan B ) to address the problems facing the EU. The European Commission instead launched a Plan D in the aftermath of the double rejections, with the aim of stimulating a frontier-free [EU] democracy by building a European-wide democratic infrastructure. Both turned out to be wishful thinking. The paper argues that the constitutional referendums have imposed serious challenges on EU governance, particularly regarding the separation between the domestic and the European in the EU. While Plan B demands the EU deal with the intra-state economic and political problems, Plan D hopes to build inter-state consensus on domestic issues. Neither is achievable without connecting the domestic and the European and forming a transnational political space in Europe. * This is a preliminary version. Comments welcome, but please do not quote without the permission of the author. 1

2 Introduction From Plan B to Plan D : The Challenge of the Constitutional Referendums The recent referendums on the EU Constitution offered some additional examples of behavioural anomalies in voting campaigns and of the democratic impact of EU referendums. Put simply, there was a shift of attention from what can be called Plan B arguments to Plan D rhetoric. As far as Plan B was concerned, the basic argument maintained that there could be some alternative arrangement that both attain the policy goals of the ballot proposal and meet the special demands of the electorate. In the French referendum, for instance, the debate about an alternative EU Constitution figured prominently in the campaign process. Some EU-critical activists argued that, in order to get a social Constitution for the EU, voters should say no to the proposed Constitutional Treaty. Such arguments were dismissed by pro-integration campaigners as irresponsible and unrealistic, but the yes camp failed to make a convincing case to discredit Plan B arguments. 1 Nevertheless, in despite of the prominence of Plan B in referendum campaigns, Margot Wallström, the Vice President of the European Commission, responded to the French and Dutch rejections of the Constitutional Treaty by claiming that the EU need a Plan D rather than a Plan B. To quote Ms Wallström directly, it is time for a new approach in Europe. Not a Plan B but a Plan D where D stands for democracy and dialogue. The EU must now stimulate a frontier-free democracy by building a European-wide democratic infrastructure (Wallström, 2005, p. 6, emphasis in origin). 2 The campaign debates on a potential Plan B and the European Commission s rhetoric about Plan D raise some important issues. First, how did the no camp exploit the public expectation of a better EU Constitution to its advantage? Second, why did the pro-integration camp fail to address voters concerns over Plan B issues, such as a social Europe? Third, why did the European Commission accept the democratic consequences of referendum crisis? Fourth, from a potential Plan B to a value-laden Plan D, how did issue-oriented campaign mobilisation lead to system-challenging 1 See The future of the European Union if France votes no, an article appeared in the Economist (23/4/2005). 2 Plan D was first used by Ms Margot Wallström in her speech at an extraordinary meeting of the constitutional affairs committee in the European Parliament on 2 June 2004 (EUobserver, 2/6/2005). On 4 June, she published a short article entitled Plan D som i demokrati in the Swedish newspaper Svenska Dagbladet. The text used here is quoted from her speech delivered at 10 th Anniversary Conference of IDEA on 10 June. 2

3 political pressures on the EU? These questions call attention to the behavioural anomalies in referendum campaigns, and highlight the normative impact of referendums on the integration project. The concluding chapter summarises the empirical findings related to the behavioural patterns of political actors in referendum campaigns, and discusses the normative implications of democratic linkages in the EU. In a way the focus of the chapter switches from behavioural analysis to democratic assessment. Attempts are also made to use the analytical arguments developed in previous chapters to explain the roles of Plan B and Plan D in the recent EU referendums. The next section first reviews political mobilisation in EU referendums. Special emphasis is put on the behavioural pattern of political actors. After that, the chapter returns to the research hypotheses and explores the democratic linkages between EU decision-making and the voters. The analysis draws attention to the oft-neglected democratic implications of multi-level governance, and tackles the thorny question of how to democratise the integration project. As a conclusion, the chapter contends that EU referendums have brought to light the structural sources of the democratic deficit in the EU, the solution to which rely on further interaction between democratically separated levels of governance in a transnational European political space. Political Mobilisation in EU Referendums The major part of the thesis has dealt with the behavioural pattern of political actors in EU referendum campaigns. This section gives a summary of the empirical studies in previous chapters. More specifically, it discusses (i) the cognitive structure of mass voters, (ii) campaign mobilisation by domestic parties, and (iii) the difficult positions of transnational actors. While reviewing the theoretical arguments and empirical findings detailed in earlier chapters, the discussion tries to employ these analytical perspectives to explore the special roles of Plan B arguments and Plan D rhetoric in the recent referendums on the EU Constitution. Nonseparable Preferences and Plan B At first glance referendums offer voters two straightforward alternatives: endorsing the ballot proposal or maintaining the status quo. In reality, political manoeuvres in referendum campaigns are more complex. Depending on the discourse of campaign 3

4 argumentation, secondary concerns and irrelevant issues may also influence voters perceptions of the ballot issue(s). In Chapter 3, this pattern of cognitive process was conceptualised as nonseparable preferences. More precisely, voters hold nonseparable preferences if their evaluation of the ballot depends on issues indirectly related to the voting topic. Due to the impact of nonseparable preferences, mobilised nonseparable issues may substantially transform voting behaviour in referendums (Brams et al., 1997, 1998; Lacy and Niou, 2000). If nonseparable preferences are exogenously predetermined, voting behaviour in EU referendums to a large extent depends on the characteristics of ballot questions. Based on social choice theories (see Muller, 2003), the analysis made a distinction between one-shot and simultaneous policy referendums, and distinguished between multi-issue membership and treaty referendums. The experiences of past EU referendums show that different EU ballots could bring about different problems of nonseparable preferences. Nevertheless, because the majority of the electorate have little knowledge about the EU, nonseparable preferences are more likely to be endogenously mobilised by referendum campaigns. Subject to the intensity of campaign exposures and group mobilisation, voters may subscribe to the discursive campaign linkage between the voting subject and other indirectly related issues. Cognitively, by introducing additional issues into voters evaluations, the mobilisation of nonseparable preferences turns the referendum ballot from a yes-no question into an if-how question. That is, the ballot question is no longer perceived as a simple choice between endorsing a particular mode of European integration and maintaining the status quo. Voters become more concerned with the multiple implications of the ballot proposal. Seemingly irrelevant issues are relevant because voters are uncertain how EU ballots might influence these matters in the long term. Focusing on these mobilised nonseparable issues, it is possible to identify the underlying cognitive structure of mass voters in the electoral process. In some cases, nonseparable issues deal with some unsatisfactory parts of the ballot proposals; in others, they refer to external matters that have not been properly addressed by the ballot proposals. Though nonseparable preferences have brought more complication into the electoral process of EU referendums, the resulting multi-dimensional cognition is normatively appealing. Thanks to the mobilisation of nonseparable preferences, voters are able to deliberate 4

5 on the multiple consequences of the ballots in a way that has not been recognised by existing theories of direct democracy (see Butler and Ranney, 1978, 1994; Budge, 1996; Bowler and Donovan, 1998; Lupia and Matsusaka, 2004). 3 The cognitive feature of Plan B arguments is quite similar to that of nonseparable preferences. Due to simplified options available on the ballot papers, the involvement of the electorate in EU referendums is necessarily limited. As mentioned earlier, they can only choose between approving a particular mode of integration or preserving the status quo. 4 The stark contrast between the two options leaves some leeway for Plan B arguments in EU referendum campaigns. More often than not, Plan B is regarded as a third alternative that the EU would eventually adopt, should the voters reject the ballot proposal. It could be an existing emergency plan behind the scene. It may also be the expected outcome of post-referendum reconciliation. Irrespective of the substantive arguments about Plan B, the core claim is that the rejection of EU ballots would not lead the integration project into stagnation. In many ways the Plan B arguments raised in the recent constitutional referendums have similar functional properties to those of nonseparable preferences. Most Plan B claims raised the unsatisfactory elements in the Constitutional Treaty, and argued that the constitutional project of the EU would be better served if the current document were rejected. Typically in the French referendum campaign, the no camp argued that the Constitutional Treaty had put too much emphasis on the Anglo-Saxon model of economic liberalism. They referred back to the Constitutional Treaty, and pointed out that the word market has been used 78 times and the word competition 27 times in the document. 5 In response, the yes camp found that the word social was mentioned 89 times in the Constitution (the Economist, 26/05/2005). Apart from the internal incoherence, external issues also entered campaign debates about the EU Constitution. 3 Deliberative democracy may provide a useful theoretical approach in this regard. For a recent collection of works on deliberative democracy, see Fishkin and Laslett (2002). 4 Some countries offer a third option in referendums: a blank ballot. For example, voters can cast empty ballots in the referendums conducted in Sweden (Ruin, 1996). This is the reason why the sum of the yes and no votes is always smaller than 100% in Sweden (see Table 3 in Appendices). However, the blank option has not altered the nature of the yes-no alternatives in the referendum. 5 See the commentary article published on the EUobserver website (20/05/2005). The article is written by Susan George, the vice-president of the French Attac organisation. The French Attac is an anti-globalisation group firmed behind the no camp in the French referendum on the EU Constitution. According to the Economist (26/05/2005), the book Cette constitution qui piège l Europe edited by Attac was among the top 20 the non-fiction bestseller list in France immediately before the referendum. 5

6 A good example of this is the widespread public concern over the Turkey s accession into the EU. As early as in October 2004, President Chirac made it clear that France would hold referendums on future EU enlargement. The intention was to assure the French people that they would have a say on Turkey s accession, with the aim of separating the issue of Turkish membership from the EU Constitution. However, the attempt was not very successful. On 21 May 2005, one week before the polling day, Philippe de Villiers, a leading no campaigner in France, still maintained that Turkey and the EU Constitution were the same thing. 6 Only by rejecting the Constitution, he claimed, would France save a European Europe rather than create a Euro-Asian Europe. Plan B arguments have met with furious denials from the pro-integration side. In the campaigns leading up to the French referendum, the yes camp initially maintained that no such arrangement as Plan B had ever been conceived for the Constitutional Treaty (EUobserver, 04/03/2005; 01/04/2005). As the no camp went on to argue for the possibility of a new round of constitution-making process, pro-integration campaigners responded that the existing document was the best constitutional text that France could have secured for the EU (EUobserver, 25/05/2005; 27/05/2005). In a word, there was simply no third alternative in the viewpoints of the pro-integration camp. The problem is that such a take-it-or-leave-it argument appeared too arrogant. It tried to centre the campaign debates on the Constitutional Treaty itself by avoiding any other doubt over the process by which the Treaty was made. The voters were not convinced. According to the post-referendum opinion survey conducted by the European Commission, 62% of French voters believed that the no victory would allow the renegotiation to produce a more social Constitution for the EU, while 59% of them thought that the popular veto would lead to the renegotiation in favour of the interests of France (Flash EB171, 2005). A poll conducted in the aftermath of the constitutional referendum in the Netherlands provides similar results: 65% of the Dutch voters agreed that the rejection would allow for the renegotiation of the Constitution to put more emphasis on the social dimension, and 66% believed the no result would serve Dutch interests in the renegotiation process (Flash EB172, 2005). 6 See EUobserver (23/05/2005). The argument of Mr. de Villiers was that Turkey had also signed the EU Constitution. Officially, the Constitutional Treaty was signed by the heads of the state or government of the 25 member states and the 3 candidate countries (i.e., Bulgaria, Romania and Turkey) on 29 October

7 On the whole, Plan B arguments dramatically transformed the campaign debates in recent constitutional referendums. They not only pointed to the internal incoherence of the Constitutional Treaty, but also referred to the indirect implications of the Constitution. In much the same way as nonseparable preferences, Plan B arguments altered the yes-no structure of mass cognition in the constitutional referendums. Theoretically, Plan B issues share many similarities with nonseparable issues. They can be easily integrated into the analytical framework of nonseparable preferences. Empirically, because of the mobilisation of these issues, the electoral process of EU referendums becomes more complicated. However, one can not deny that public deliberation on these important matters may well benefit the integration project in the long run. Instead of blaming voters for irrational voting behaviour, it is perhaps more important to ask why the mainstream yes campaigners fail to claim the ownership of nonseparable issues 7 in referendum campaigns. Two-Dimensional Political Space and Plan B As pointed out in Chapter 2, referendums are contextualised political events. It is difficult to generalise about the content of nonseparable issues, which often vary according to the political contextualisation of referendum campaigns. However, we have identified some consistent behavioural patterns of domestic political parties in response to the challenge of extraordinary campaign linkages. To put it briefly, while the periphery parties on the no side exploit nonseparable issues to their advantage, the mainstream parties on the yes side face a structural dilemma in EU referendums. The reason, as we explained in Chapter 4, lies in the spatial configuration of a two-dimensional political space where domestic political parties compete for the support of individual voters on both domestic and integration issues. A referendum-generated political space can be understood in terms of the two-dimensional spatial model because transnational ballot issues and domestic political concerns are mixed up in EU referendums. The campaigns are dominated by debates about the domestic impacts of European integration as well as the transnational 7 Issue ownership is one of the key concepts developed in electoral studies. It is used to explain how political parties manipulate different policy issues in the electoral campaign, and how voters cast their ballots based on perceived competence of each party on these issues. Classic studies on issue ownership are Budge and Farlie (1983) and Petrocik (1996). Recent works include Petrocik et al. (2003), Holian (2004), and van der Budge (2004). 7

8 implications of domestic voting. As a result, the boundary between domestic and transnational politics is blurred. Domestic political parties are forced to compete on domestic and integration issues at the same time. Depending on the composition of the domestic party system, the structure of the referendum-generated political space can be characterised as orthogonal two-dimension, oblique two-dimension, or linear one-dimension (see Marks and Steenbergen, 2002). Different types of EU ballots may also influence the spatial characteristics of party competition in referendum campaigns. For instance, the conflict between the domestic and the transnational is more pronounced in treaty referendums than in membership referendums. The spatial analogy of a two-dimensional political space thus grasps the special contextualisation of EU referendums. In the two-dimensional political space, the mainstream parties typically suffer from two structural difficulties: intra-party division and inter-party convergence. The problem lies in the fact that, while integration issues are likely to divide individual mainstream parties internally, these issues generally unite the mainstream of the domestic party system. Under these circumstances, referendum campaigns create a structural dilemma for mainstream parties: in order to survive in domestic politics, they should maintain intra-party unity and keep a certain distance amongst themselves; in order to win the integration ballot, they must go beyond partisan antagonism and form cross-party alliances in referendum campaigns. These contradictory motivations show the structural conflict between domestic and integration politics in EU referendums. By contrast, parties on the political periphery are much less constrained by two-dimensionality. Internal cohesion is not an issue for small parties; neither would inter-party alliances blur their distinctive ideological labels. It is therefore much easier for the periphery parties to coordinate on the no side. These parties have taken advantage of their structural positions, and become the main advocates of extraordinary issue linkages in referendum campaigns. Despite the potential challenge of periphery mobilisation, the mainstream parties have not been prepared to mobilise fully in EU referendums. While the official positions of mainstream parties are generally pro-integration, the intensity of their referendum campaigns varies. There is often a noticeable disparity between the incumbent mainstream and the opposition mainstream. The mainstream opposition occupies an 8

9 awkward position in the two-dimensional political space. On the one hand, these parties are in opposition to the government on domestic issues. On the other, they support the government on integration issues. Facing the structural conflict between the domestic and the transnational, the mainstream opposition is very likely to employ the strategy of pro-integration abstention and pursue lacklustre campaigns. The Danish Social Democrats did so in the first Maastricht referendum in 1992 (see Siune, 1993; Schneider and Weitsman, 1996). The Irish Party Fine Gael opted for a lacklustre campaign in the first referendum on the Nice Treaty in 2001 (see Chapter 4). In the recent votes on the Constitutional Treaty, there was again a noticeable gap of yes voters among the incumbent and opposition supporters in Spain, France and the Netherlands. According to the post-referendum surveys conducted by the European Commission, except for the Dutch Green Left, the mainstream opposition all failed to mobilise their supporters on the yes side to the same degree as the incumbent parties (see Table 7.1) Table 7.1 Supporters of the Mainstream Incumbent and Opposition Political Parties in the Referendums on the EU Constitution Political Parties Voted Yes Voted No N/A Spain Mainstream PP Opposition (Popular Party) 72% 19% 9% Mainstream PSOE Incumbent (Socialist Worker s Party) 93% 4% 3% France Mainstream PS 39% 61% -- Opposition (Socialist Party) Mainstream UMP/UDP 75% 25% -- Incumbent PvdA Mainstream (Labour Party) 37% 63% -- Opposition Grown Links (Green Left) 54% 46% -- The CDA Netherlands (Christian Democrats) 53% 47% -- Mainstream VVD Incumbent (Liberals) 49% 51% -- D66 (Democrats 66) 51% 49% -- Source: Flash Eurobarometer 168, 171, 172. Effective pro-integration campaigns require mainstream political parties to transcend domestic political disagreements and form a cross-party campaign alliance in line with their stands on European integration. In doing so, the mainstream parties form a responsive integration cartel to promote the integration project in the domestic arena. 9

10 To explain the cartelisation of integration politics (see Katz and Mair, 1995; Blyth and Katz, 2005), we made an analytical distinction between responsive and irresponsive cartel parties. The argument is that the mainstream convergence on integration issues has more to do with the structural factors of EU decision-making than with inter-party collusion (see Kitchelt, 2000). Because of their structural dominance in domestic politics, the mainstream parties are in a better position to influence the policy-making of the EU. The consensual mode of EU decision-making helps to reconcile the mainstream difference on integration issues, and in turn leads to a mainstream consensus on the overall direction of European integration. Since the support for the integration project has not relied on inter-party collusion, the mainstream parties face the free-rider problem 8 in promoting European integration in the domestic arena. That is, individual parties have an incentive to benefit from the integration project without contributing to integration-related campaigns. The abovementioned pro-integration abstention of the mainstream opposition is a good example of this. As a whole, the mainstream parties tend to avoid integration debates altogether in normal election campaigns, a situation which we consider as denoting an irresponsive integration cartel. Nonetheless, it does happen that the integration cartel coordinates their campaigns and mobilises together in EU referendums. As discussed in Chapter 4, this was the case in the second Danish vote on the Maastricht Treaty in 1993 and in the second Irish vote on the Nice Treaty in On both occasions, the cross-party commitment to the yes campaign was well publicised in order to send credible signals to individual pro-integration parties. In this way inter-party collusion closed the gap between consensual decision-making and collective electoral abstention, and resulted in a responsive integration cartel in EU referendums. A more fundamental problem than campaign mobilisation is that even a responsive integration cartel is unable to deal effectively with the rhetorical challenge of Plan B. Plan B arguments typically feature issues involving both domestic and integration 8 Russell Hardin argues that the free-rider problem, or the logic of collective action, can be strategically considered as n-prisoners dilemma (Hardin, 1971, 1982). This is not entirely correct. A better game-theoretical analogy is provided by the game of stag hunt, where players choose between unilateral or collective action. In this game, collective action yields higher payoff than unilateral action, but the former is risk dominated by the latter (Skyrms, 2003). The stag hunt game differs from the prisoners dilemma in two important respects. First, the stag hunt game explicitly models the conflict between private and public goods in strategic decision-making. Second, collective and unilateral actions are both Nash equilibria in the stag hunt game. The solution to the free-rider problem can thus be regarded as the question of equilibrium selection, and be analysed in the game-theoretical framework. Nevertheless, it is possible to construct a repeated prisoners dilemma and make it analytically equivalent to the stag hunt game (see Skyrms, 2003, pp. 4-6). 10

11 implications. In the two-dimensional political space, these issues reshape the balance of power between the pro-integration cartel and the anti-integration camp. Firstly, the domestic implications of Plan B issues draw internal dissidents from the mainstream parties into the EU-critical camp. As a result, the campaign persuasion of intra-divided integration cartel is weakened. Secondly, the integration dimension of Plan B issues requests the integration cartel to respond to demands beyond their control. EU decision-making is a collective process, in which the domestic integration cartel only comprises one of the twenty-five. Without the consent of the other member states, it is simply impossible for the integration cartel to give credible answers to unsettled integration issues. In the recent referendums on the EU Constitution, the debate on social Europe loomed large in the French and Dutch campaigns. The arguments highlighted the socio-welfare impacts of the Constitution in the domestic arena, and the social dimension of the Constitutional project in Europe. Not surprisingly, the left-wing mainstream parties in both France (i.e., the Socialist Party) and the Netherlands (i.e., the Labour Party) were seriously divided on the domestic implications of social Europe. As it turned out, the majority of their supporters failed to follow official party lines, and voted against the Constitution in the referendums (see Table 7.1). In the meantime, social Europe as a constitutional principle for the EU received almost no effective response from the pro-integration camp except for some abrupt denials. 9 Another example of Plan B issues dealt with the possible renegotiation of the Constitutional Treaty. Renegotiation is both a domestic and an integration issue because it involves two interrelated aspects: who should renegotiate and how to renegotiate. The first is a domestic issue because it challenges the competence of the incumbent government in representing the country in EU negotiations. The second is an integration matter, as it queries whether there could be a better outcome from the EU negotiations. In the French and Dutch referendums, the renegotiation argument nearly paralysed the pro-integration alliances. At the domestic level, the yes camp failed to take a consistent stance on renegotiation. 10 Prominent pro-integration figures, such as Jacques Delors, made public statements 9 In a televised debate on the Constitution held on 14 April 2005, the French president Jacques Chirac argued that the Constitution was necessary to fight against ultra-liberalism, and to make Europe stronger against big powers (EUobserver, 15/04/2005). 10 It would be quite interesting to see whether the renegotiation argument invited more furious denies from the incumbent than from the opposition. 11

12 admitting the possibility of renegotiation in the campaign process (EUobserver, 17/05/2005). At the transnational level, the renegotiation arguments weakened the pro-integration camps across the member states. As the prospect of the Constitutional Treaty looked dimmer after the French rejection, 11 the renegotiation arguments became even more plausible. Indeed, as showed in the Eurobarometer polls quoted earlier, there were even more Dutch voters who agreed with the renegotiation arguments than French. In a referendum-generated two-dimensional political space, domestic political parties face considerable political uncertainties. To reconcile domestic and integration politics in EU referendums, individual mainstream parties, especially those in the opposition, often strategically withdraw from the campaigns. As a whole, the integration cartel is in need of some credible assurance to stage responsive campaign mobilisation. Nonetheless, even the responsive integration cartel has trouble dealing with Plan B issues. The reason for this lies in the fact that the two-dimensional political space has constrained the effective response of the pro-integration cartel to issues having both domestic and integration implications. In fact, not only did mainstream parties find it difficult to address Plan B arguments, but also transnational actors shied away from ideological mobilisation in EU referendums. The domestic barriers to transnational mobilisation make more salient the communication gap between the domestic and the transnational. As a result, the debate about Plan B shifts to the discussion of Plan D at the transnational level. Transnational Actors and Plan D The transnational implications of EU referendums vary according to the underlying issues in EU ballots. Membership referendums decide the geographical size of the EU; policy referendums influence the policy scale of the integration project; treaty referendums determine the general direction of European integration. On the whole, these popular votes all exert certain transnational impacts on the integration process. It is reasonable to expect that transnational interests mobilise in EU referendums. To find out whether this is true, Chapter 5 examined how transnational actors mobilise in EU referendum campaigns. The discussion concentrated on three sets of transnational 11 See, for example, a series of articles published in the Economist after the French and Dutch referendums. Most tellingly, one article is entitled Dead, but not yet buried. 12

13 actors: cross-border EU-critical movements, transnational pro-integration organisations and the EU institutions. Of the three, cross-border EU-critical movements are the least organised and institutionalised. However, they tend to be the most influential in EU referendum campaigns. In contrast to the domestic grouping of Eurosceptic political parties, cross-border EU-critical movements do not take ideological stands in their platforms. Intra-party Eurosceptic groups and extremist political parties are important political forces behind domestic anti-integration campaigns, but their ideological roots in domestic politics prevent them from mobilising beyond the national context. By contrast, non-partisan Eurosceptic groups and single-issue Eurosceptic parties do not subscribe to explicit ideological positions. Because of this, they are in a better position to mobilise in the transnational arena. These groups typically form the core of cross-border EU-critical movements. For example, TEAM, a transnational EU-critical network, has adopted a strict organisational principle with regard to the ideological balance of its members. As detailed in Chapter 5, most of its members are non-partisan groups and single-issue Eurosceptic parties. The principle of ideological neutrality not only makes it easier for these groups to coordinate activities against European integration, but also weakens the ideological criticisms levelled at such coordination. Free from ideological constraints, cross-border EU-critical movements have played an important part in providing information, financial and rhetorical support for domestic anti-integration campaigns in EU referendums. In the Irish referendums on the Nice Treaty, for instance, TEAM directly coordinated the transfer of campaign donations from Danish Eurosceptic groups to the no campaigners in Ireland. In recent votes on the EU Constitution, cross-border EU-critical movements again staged successful transnational campaign coordination. It was reported that Nigel Farage, a British MEP affiliated to UKIP, financially supported Philippe de Villiers campaign against the EU Constitution in France (EUobserver, 23/05/2005). Moreover, Mr. Farage even participated in the French campaign by himself. In a no camp rally held in Paris one week ahead of the polling day, he gave a public speech saying that, while he might have different reasons for against the Constitution, the French non would be the first step towards a Europe of nations, which all of us want to achieve (EUobserver, 23/05/2005). 13

14 Compared to cross-border EU-critical movements, transnational pro-integration organisations usually have a more organised institutional structure. Paradoxically, however, their presence in domestic EU referendum campaigns is less visible. Based on the evidence arising from the elite interviews, we have identified two possible reasons for this resource-competence paradox. First, transnational pro-integration organisations tend to have a top-down organisational structure. That is, the transnational presence of pro-integration organisations mainly originates from the exogenous necessity of political mobilisation rather than the endogenous demands of the rank and file. Because of this, the institutional structure of these organisations does not match its actual mobilisation power. Second, within transnational pro-integration organisations, there are considerable internal disagreements regarding the preferred speed and direction of European integration. It is often up to the domestic offices to decide how to conduct their own voting campaigns. As a consequence, instead of transnational pro-integration organisations, the domestic offices tend to stand out in the campaign process of EU referendums. Recently, to support the Constitutional Treaty, some transnational pro-integration organisations (including the EMI) have made a significant breakthrough by launching a pan-european YES campaign for the first time. A number of campaign activities were conducted in different countries. 12 However, it is worth asking whether these activities were organised by the transnational office in Brussels or domestic offices based in individual countries. Without directly addressing the organisational weakness mentioned above, transnational pro-integration organisations will probably remain the emperor s invisible clothes. The EU institutions are the most institutionalised transnational actors of the three. At the same time, they are the least rhetorical campaigners in EU referendums. The institutional actors of the EU, especially the European Commission, prefer to be seen as neutral information providers in the campaign process. However, information provision is an institutional solution rather than a deliberate choice. On the one hand, using the EU budget to finance rhetorical campaigns is quite controversial. It is much 12 For further details, see (accessed on 5 July 2005). 14

15 less problematic to work on neutral information provision. On the other, the institutional actors of the EU face the most stringent domestic barriers to referendum campaigns. Information provision is again a safer option in this respect. However, without the rhetorical power of campaign persuasion, it is debatable as to what extent the information provision of EU institutions may achieve its initial goals. 13 The rejection of the EU Constitution in the French and Dutch referendums has brought further into question the communication efforts of the EU institutions. On the whole, the discrepancy between transnational actors organisational strength and rhetorical power is astonishing. At the transnational level, the organisational strengths apparently fail to deliver the rhetorical powers that one would expect in normal political settings. This is due to the fact that domestic ideological labels are no longer effective benchmarks, based on which political actors may form alliances and mobilise supporters. Meanwhile, the integration process has not yet created a coherent line of political disagreement, except for some crude divisions between more or less integration. Under these circumstances, the organisational labels of transnational actors contain few political connotations. The organisational strengths thus can not easily translate into rhetorical power in the campaign process. In the case of EU institutions, institutional strengths have even turned into a burden in campaign mobilisation. From this perspective, it may be easier to understand why Plan B issues have lost their appeal in the transnational campaign rhetoric; and why Plan D becomes a prominent matter at the transnational level. Plan B arguments rely heavily on the connection between the domestic and the transnational. They deal with the domestic implications of transnational ballots and the transnational impacts of domestic voting. The non-ideological platform of transnational mobilisation makes it very difficult for transnational actors to pursue campaign rhetoric on Plan B issues. After all, different countries may support or oppose certain EU ballot for different reasons. Transnational actors are not equipped to campaign on issues specific to the domestic political context. By contrast, Plan D highlights the existing democratic gap between the EU and its 13 The European Commission was reported to have admitted the deficiency of neural information provision. Louis Michel, the Belgian commissioner, said, Communication is more than information and it is difficult to see this process as a value-free or neutral exercise. It is a true political exercise. (EUobserver, 28/04/2005). 15

16 citizens. It raises doubts about how effective the communication capacity of transnational actors really is in domestic referendum campaigns. For transnational pro-integration organisations and EU institutions, this is the real problem they are facing in the campaign process. Not surprisingly, as the referendum in Luxemburg endorsed the Constitution on 10 July 2005, José Maria Gil-Robles, the President of the EMI and the former President of the EP, still insisted that the need for better communication persisted. 14 Ten days later, the European Commission approved a new Action Plan to Improve Communicating Europe. 15 Nonetheless, the demand for Plan D in the EU has its roots deeper than the communication deficit in transnational campaigns. As we examine political mobilisation as a whole, there emerges a gap between the verticalisation of campaign mobilisation in EU referendums and the structural change of multi-level governance in the EU, a more serious problem to which the following sections now turn. Political Verticalisation and Multi-Level Governance In the beginning of this thesis, we hypothesised that as European integration deepens and widens, political mobilisation in EU referendums becomes increasingly verticalised. The analysis of campaign mobilisation at three political levels (i.e., mass, domestic and transnational) lent support to the hypothesis about the verticalisation of campaign mobilisation. Firstly, voters refer to diversified issues directly or indirectly related to the voting topics. These issues are related to both the domestic and the transnational impact of EU ballots. Secondly, instead of the political mainstream, domestic political parties on the periphery are likely to play a key role in mobilising public opinion in EU referendums. The domestic party systems are under increasing pressures to adapt to the political changes resulting from the integration process. Thirdly, transnational actors are taking initial steps to influence domestic referendum campaigns. Yet, they still lack the rhetorical power to mobilise the public with respect to the transnational dimension of EU ballots. These findings basically confirmed H4. However, it is necessary to point out that such a verticalised constellation of campaign mobilisation is different from what we had expected in the beginning. Most notably, 14 For details, see (accessed on 15 July 2005). 15 See (accessed on 15 August 2005). 16

17 the declining role of domestic mainstream parties has not been matched by transnational mobilisation in referendum campaigns. The domestic periphery parties instead gain more prominence in the campaign process. Meanwhile, voters care about both the domestic and transnational impacts of EU ballots, but they still seek primary references in the domestic political context. Thus, while campaign mobilisation has been verticalised, the national level remains the focal point of political verticalisation in EU referendums. The question arises as to what extent the verticalisation of campaign mobilisation in EU referendums has reflected the multi-level political governance in the EU. In the introductory chapter, we argued that the verticalisation of integration politics was different from the multi-level governance perspective (Scharpf, 1994; Marks et al., 1996; Hooghe and Marks, 2001; Bache and Flinder, 2004) in that it highlights the potential conflict between different political levels in the dynamic transformation of European political governance. The theoretical focus of multi-level governance draws attention to two important aspects of the political dynamics of European integration. First, in terms of multi-level, decision-making in the EU has dispersed across multiple territorial levels (Hooghe and Marks, 2001, xi; see also Kohler-Koch, 2003). Second, as far as governance is concerned, EU policy-making is considered as a problem-solving process, 16 which replaces governments with policy networks and prioritises policy coordination over political controls (Bache and Flinder, 2004, pp ; also see Rhodes, 1997; Scharpf, 1997b). Thus, on the vertical side, multi-level governance stresses the declining role of national governments in EU decision-making; on the horizontal side, it emphasises the role of policy networks as a novel mode of political governance. In many ways, political verticalisation in EU referendums has reflected the overall trend of multi-level governance in Europe. For instance, domestic mainstream parties become less capable of mobilising public opinion in their favour. Campaign alliances in EU referendums also tend to depoliticise the mainstream disagreements on European integration. Nevertheless, in contrast to the benign picture of multi-level governance, political verticalisation has brought to light the structural tension between the transnational and the national in EU politics. In EU referendums, 16 Hooghe and Marks (2003) make a distinction between two types of multi-level governance: One is oriented to general purpose, nonintersecting and durable jurisdictions ; the other focuses on task-specific, intersecting, and flexible jurisdictions (p. 233). 17

18 this tension takes the form of structural conflicts between transnational ballot issues and domestic electoral mobilisation. Because of this tension, mass voters invoke nonseparable preferences (see Chapter 3); domestic political parties are constrained by a two-dimensional political space (see Chapter 4); transnational actors are confronted with the legal and political barriers to domestic campaigns (see Chapter 5); and there is a dual constitutionalisation process embedded in these popular votes (see Chapter 6). Taken together, the political verticalisation of campaign mobilisation not only reflects the political structure, but also highlights the structural politics, of multi-level governance. From this perspective, the analysis of campaign mobilisation in EU referendums has unintentionally uncovered an important gap between the structural transformation of European governance and the verticalisation of integration politics. On the one hand, the decision-making process of the EU has led to various structural connections amongst the political actors at the regional, national and supranational levels. Because of these connections, national governments are gradually losing their monopoly over EU policy-making and implementation (Hooghe and Marks, 2001). On the other hand, the integration project still relies mainly on the national electoral process to obtain direct and indirect political legitimacy (Moravcsik, 2002). Voters tend to put primary emphasis on political accountability at the national level. They often connect EU ballots with various domestic political concerns, despite the efforts of national politicians to prioritise the transnational dimension of integration-related ballot issues. 17 Transnational actors stand a better chance of informing the public in this regard, but there are substantial domestic barriers to transnational mobilisation. Thus, the verticalisation of integration politics has significantly lagged behind the structural transformation of multi-level governance in Europe. In fact, the structural gap between entrenched democratic legitimacy at national level and the political linkages between different levels of European governance has characterised various forms of democratic deficits in the EU. It underlies oft-quoted problems such as diluted national parliamentary controls, second-order EP elections, and the distance between Brussels and EU citizens (see Hix, 2005, pp ). Given 17 In his TV debate on the EU Constitution, President Jacques Chirac commented that France has its problems, and has to resolve them. It is not the Constitution that will resolve them for us, he added (EUobserver, 15/04/2005). 18

19 the significance of the problem, a Plan D is indeed necessary in order to deal with the democratic challenges of EU referendums. The Democratic Challenges of EU Referendums Different from many other studies of EU referendums (e.g., Jenssen et al., 1998; Hug, 2002; Szczerbiak and Taggart, 2005), this thesis has examined the mobilisation patterns of political actors rather than the reasons behind the voting results. The analytical focus on political mobilisation was adopted for two main reasons. First, a comparative study of EU referendums demands a generalised analytical perspective. Such a perspective must go beyond the political context of voting countries. Second, the research question of the thesis focuses on the political interaction between referendums and European integration. The mobilisation perspective avoids value-laden arguments such as a yes result contributes to the integration process and a no vote blocks further integration. Instead, attention is paid to the way in which particular mobilisation patterns have both reflected and influenced the integration process. It is from such a mobilisation perspective that we are able to make an assessment of the democratic challenges imposed by EU referendums. In the past decade a few prominent democratic and integration theorists have argued that European integration necessarily induces the problem of democratic deficit. According to such views, the integration process is vulnerable to the democratic deficit precisely because of the inherent trade-offs between system effectiveness and citizen participation (Dahl, 1994); between political interdependence and democratic legitimation (Scharpf, 2000); and between non-majoritarian institutions and electoral accountability (Majone, 2005, Ch. 2). In the early 1990s, the integration project made some significant changes in the Maastricht Treaty. The aims were to improve the system effectiveness of the decision-making process (by introducing more QMV), to recognise the structural interdependence across policy areas and between levels of policy-making (by creating a three-pillared policy-making structure and setting up the Committee of Regions), and to establish non-majoritarian institutions (such as the ECB). It is not surprising that the democratic problem of the EU was subject to intensive public debate during the ratification process of the Maastricht Treaty. Further integration in these directions only intensified the democratic challenges of the integration project in the post-maastricht era. 19

20 A problem with these arguments, however, is their prediction about the inevitability of democratic deficits in the EU. That is, further efforts to build effective, interdependent and non-majoritarian decision-making mechanisms in the EU cannot go hand in hand with preserving democratic legitimacy in a normative manner. Such a gloomy picture has led to some radical assessments of the democratic status of the EU. On the one hand, Dahl (1999) contends that it is impossible to build democratic organisations and institutions at the international level. In his views, this is due to the huge scale of international governance and the deep-rooted heterogeneity of the electorate. Dahl s proposal is to treat the governing organs of international organisations as bureaucratic bargaining systems (p. 33, emphasis in origin). On the other hand, Moravcsik (2002, 2004, 2005) denies the applicability of ideal democratic standards to the EU. He argues that the democratic evaluation of the EU must take into account the limitations of real-world democracy. Based on the criteria which translate the ideal to the real, Moravcsik claims that the EU is democratically legitimate. 18 Instead of considering the democratic problems of the EU either inevitable or non-existent, we try to understand the democratic challenges of EU referendums from a mobilisation perspective. Arguably, EU referendums have offered the best way for ordinary voters to make a judgement about the democratic nature of the integration project. The democratic deficit of the EU is not only one of the most debated topics in referendum campaigns, but also figures prominently in post-referendum political reconciliation. However, it would be a mistake to dichotomise the assessment in terms of a simple yes-no answer according to the voting results. Indeed, democracy has more to do with the decision-making process than the decision itself. The democratic challenges of EU referendums should lie more importantly in the specific mode of campaign mobilisation than in the voting results. For instance, in recent referendums on the EU Constitution, while Spanish voters endorsed the Constitution with a quite high approval rate of 76.73%, the French rejected the Constitution with a negative majority of 54.68%. Yet, campaign mobilisation around the EU Constitution was much more intensive in the French referendum than in the Spanish vote. In this sense, compared with the Spanish referendum, the French vote have made a more significant 18 Zweifel (2002) comes to a similar argument by comparing the democratic status of the EU with Switzerland and the US. 20