Referendums on European Integration in France: Political Actors Motivations and Voters Attitudes

Referendums on European Integration in France: Political Actors Motivations and Voters Attitudes Geeyoung HONG Yonsei University geeyoung164@yonsei.ac.kr

I Introduction Lindberg and Scheingold (1970) have argued that European citizens have been generally favorable toward European integration but that they have also been indifferent to European project because of the complexity of European issues. This permissive consensus enabled, for many decades, political elites to pursue integration independent of popular preferences. However, more recent studies on public attitudes in Europe suggest that the permissive consensus is now declining. The process of Maastricht Treaty ratification seems to have confirmed this hypothesis. In both the Danish and French referendums, in 1992, popular support for the Maastricht Treaty was lower than expected (Franklin et al., 1994). These referendums showed that there is a deep gap between political elites and public and that public support for further integration could not be taken for granted (Obradovic, 1996). The process for the ratification of Maastricht Treaty posed indeed the question of European Union (EU) legitimacy (Laursen and Vanhoonacker, 1994). Maastricht experience forced political elites to accept that Union should develop policy legitimacy for further integration which depends largely upon the support of its population (Obradovic, 1996). Facing the growing need to develop policy legitimacy, EU member states use increasingly referendums on European integration. Since the first referendum held in 1972, 19 out of the present 25 member states have held EU referendums. Pointing out that most EU referendums have been held since 1992, Taggart (2006) note that referendums have become more popular as the European integration project has become less popular. With the increasing use of referendums on European integration, referendums have recently attracted great attention of many scholars. On the one hand, referendums are perceived as a means of legitimizing the EU policy (Rouke et al., 1992). This view is based on the participative characteristic of referendums: European public participates directly in policy-making through referendums. Indeed, some studies find that states with provisions for referendums adopt policies that are closer to the wishes of the voters 2

(Romer and Rosenthal, 1979; Gerber and Hug, 1999; Hug 2002). In theory, referendums could strengthen democracy in the EU, by giving Europeans a say on the integration and making governments more responsive to public preferences. On the other hand, critics of this approach claim that referendums could not legitimize the European-level policy. As noted by Closa (1994), referendums have only the function of legitimizing domestic regimes. The causes for calling a referendum on European issue are located in the domestic regimes. It is further argued that any European issue, as long as a referendum is held inside the country, ultimately turns into domestic one (Imig and Tarrow, 2001). Past experiences with European elections and referendums have confirmed this domestication of referendums. It became thus important to examine whether or not referendums on European issue are dominated by domestic factors in order to answer the question of whether they can be used as an instrument for securing European Union legitimacy. From the arguments mentioned above, we can draw two conditions which appear to be crucial in order that European Union gains its policy legitimacy through referendums. First, the context in which political actors hold a referendum must not be dominated by domestic factors. Second, voters who are consulted in referendums must express their opinion on the basis of their own decision on European issue. So, it seems imperative to consider both attitudes of political actor and voters behavior in referendums, in order to examine whether or not referendums are framed by domestic context. The aim of the present study is to examine these two factors, that is, political actors motivation in calling an EU referendum and voters attitudes in EU referendum, to know which factors frame primarily the EU referendums. This article attempts to give answer to this question by analyzing the two French referendums held in 1992 and 2005. Analyzing French case seems to be important in two respects. First, with Denmark and Ireland, France has frequently used referendums on European integration. French case could thus provide ample illustrations of referendums. Second, in contrast to Denmark and Ireland in which referendums are constitutionally required, France does not have constitutional requirement. This means that, in France, attitudes of political elites played an important role in choosing to call a referendum, whereas in two other countries, the role of political elites remains limited. 3

The article is organized as follow. I first present some general introduction and theoretical framework on EU referendums. I then examine the two French referendums (1992 and 2005) at two levels: political actors motivation and voters behavior. Ⅱ Referendums on European integration A Referendums in EU Member States There are at least two different models of democratic process which are at play in European project: on one side, representative model of democracy and on the other side, direct model of democracy. Referendum is one of the various forms of direct democracy. And the use of referendums to ratify European treaties, accession and membership constitutes the evident example of the institutionalization of direct democratic practices in EU (Taggart, 2006). It seems necessary, however, to note that referendums are not the main instrument of democracy in EU member states. Most EU member states perceive referendums as supplementary to representative democracy and as extraordinary instrument providing additional legitimacy in pursuing specific policy (Closa, 2007). Several member states such as Belgium, Germany, the Netherlands and the UK do not even have constitutional regulations on referendums. Moreover, according to the British constitutional tradition, referendums are not even compatible with parliamentary sovereignty. Despite the secondary role of referendums in EU member states, the part of referendums seems more and more important in European integration. Since World War Ⅱ, European integration has been the most frequent subject matter for referendums (de Vreese, 2004). Since the first referendum on European integration was held in France in 1972, a total of 44 referendums have been held on issues of European integration. It appears that, in important domains, there is a move towards a model of direct democracy and away from representative democracy (Taggart, 2006). Referendums are indeed being used at crucial moments of European integration. It was by no means a coincidence that 10 EU governments announced their decision to hold 4

referendums to ratify the EU Constitution. With the increasing use of referendums, theoretical and empirical research on referendums has proliferated over the last decade. And there are several classifications of the referendums put forward by many scholars. But in order to analyze both the motivations of political actors and the voting behavior in referendums, it is necessary to focus on two ways of classifying the referendums. First, according to constitutional factor, referendums may be either required (mandatory) or non-required (non-mandatory) (Suksi, 1993; Setala, 1999). Required referendums are constitutionally required upon certain issues. Required referendums cannot be avoided, given the constitutional or institutional provisions. Non-required referendums are initiated only at certain political actors demand. Second, referendums may be binding or non-binding, according to legal factors (Suksi, 1993; Hug and Sciarini, 2000). In binding referendums, the government has to implement the policy chosen by the popular vote. In this case, citizens have effectively the last word in a ratification process and rejection by citizens renders the ratification impossible. In nonbinding ( advisory or consultative ) referendums, the parliament or government might not fully respect the outcome of referendums. However, it may be politically dangerous for political actors to ignore the will of voters (Tridimas, 2007). It is important to take into account these two ways of classification, because it has been argued that referendums of different types presented above affect both the motivations of political actors who might call a referendum and the voting behavior of citizens in referendums (Hug and Sciarini, 2000). Among EU member states, only Denmark and Ireland have constitutional obligation for calling a referendum in ratification of EU-level treaties. In ratifying the EU treaties, referendums in other member states are called at the discretion of political actors. B Theoretical Framework for Explaining the Political Actors Motivations As we have seen above, political actors have no choice but to hold a referendum in the case of required referendums. The choice of whether or not to call a referendum 5

occurs only in non-required referendums. That is, only non-required referendums are called at the discretion of political actors. Analyzing the motivations of political actors should be thus limited to the non-required referendums. European integration provides the ample examples of non-required referendums (Tridimas, 2007). Since the first referendum on European integration in 1972, out of a total of 44 EU-related referendums, 23 were not constitutionally required but were called at the discretion of the incumbent government. In 2003, 7 of the 9 referendums held by the new entrants to approve membership were not required. And none of the 4 referendums held in 2005 for ratifying the EU Constitutional Treaty were required (Tridimas, 2007). As we shall examine later, three EU referendums that France has held since 1972 were all initiated by the President. The non-required referendums held frequently on European integration raises the essential question. Why do political actors (government or president) choose to call a referendum rather than relying on parliamentary vote? As has been pointed out, unforeseen outcomes might present obvious political risks for political actor and the efforts required to obtain ratification through referendum might be superior to those for parliamentary ratification. There are two competing approaches for explaining the motivations of political actors who might choose to call a non-required referendum: the first related to the domestic/strategic approach, the second to the European/institutional approach (Closa, 2007). The first one focuses on strategic calculations at domestic level, while the second one emphasizes institutional factors at European level which might lead to the decision of political actors. Domestic/ strategic approach comprises several arguments. First argument suggests that the government use referendums to maximize the outcomes of negotiations (Moravscik, 1993; Schneider and Cederman 1994). Governments with more severe domestic constraints are expected to be more influential in the bargaining process and thus have a greater impact on the negotiation outcomes. Second argument for strategic approach assumes that government holds a referendum to consolidate its own position (Bogdanor, 1994; Hug and Sciarini, 2000; King, 1991; Morel, 2001). In this case, referendums are used as a tactical weapon by reinforcing the electoral position of 6

government or by creating divisions of the opposition. This type of referendum is often called a plebiscite. Third argument suggests that referendums may be held to pass treaties that would not otherwise be ratified (Morel, 2001). Referendums are served as instrument for securing ratification in the absence of parliamentary majorities. Alternative accounts are put forwards by some scholars relying on European/ institutional approach. They emphasize that the decisions of political actors are more influenced by European factors than by domestic factors. First, the force of discourses in favor of referendums which have been developed in European arena and the imitation of path followed in other countries played an important part in the choice of political actors (Closa, 2007). It is also argued that governments call a referendum in order to be more responsive to the popular preference toward European issue. Of course, all these different reasons for holding a referendum play an important role in different contexts. But many scholars are inclined to consider that, in the case of non-required referendums, political strategic calculations are the most important motivation of political actor for calling a referendum (Bogdanor, 1994; Hug and Sciriari, 2000). In contrast, Closa (2007), without denying the importance of strategic factors, focus rather on the importance of the role of institutional factors at European level. C Theoretical Framework for Explaining the Voters Behavior So, how do voters behave in a referendum called by political actors for diverse reasons? On which criterion do they vote in referendum? In recent years, particularly in the wake of Maastricht referendums, much has been studied about this question. There are currently two main approaches for explaining the voting behavior in EU referendums: the second-order election approach and the issue-voting approach. The second-order election is the concept that Reif and Schmitt (1980) employed to describe the European parliamentary elections. In their view, the European elections are dependant on a domestic political context as long as the national political issues remain more important both to parties and voters. This concept can also be applied to the EU referendums: voters behave in EU referendums, not in terms of the European content, 7

but in terms of the domestic context. Franklin, Marsh and McLaren (1994) explained the result of Maastricht Referendums by using this concept. They noted that supporters of government parties voted much more strongly in favor of the Maastricht Treaty than partisans of opposition parties. According to them, national issues tend to dominate the campaigns and voters have thus tendency to consider EU referendums as a means of signaling their attitude toward the government or to follow the recommendations of national parties. This approach is contested by some scholars who focus on voters attitudes and beliefs toward European integration itself. In particular, Siuen, Svensson and Tonsgaard (1994), contesting the argument of Franklin, Marsh and McLaren (1994), claimed that partisanship and government popularity have played no role in the outcomes of Maastricht referendums. The issue-voting approach argues that voting behavior in EU referendums reflect Europeans broad attitudes towards European integration. In other words, referendum outcomes can be seen as a reflection of reasoned decisions taken by voters about the future of European integration (Svensson, 2002). Much interest has been taken on this question of whether Europeans vote really on the issue on which they are asked to vote or do they merely vote on the domestic political context. Hug and Sciarini (2000) attempted to answer this question by taking into account the institutional factors of referendums. They argue that voting behavior in referendums is intimately linked with the types of referendums. According to their account, in a non-required referendum, the government party supporters vote in favor of ratification, because they are well aware of the political calculations behind the government s decision to hold a referendum. If this non-required referendum is binding, they are likely to vote even more strongly in favor of ratification, because of the possible political consequences of their negative vote. In a non-required and nonbinding vote, the partisans of government party appear also to be aware of the intentions of government, but given the fact the outcome of the vote is not binding, they consider the referendum as an opportunity to vote on European issues or to punish the government for diverse reasons. In the view of Hug and Sciarini (2000), both institutional characteristics of referendums (required/non-required, binding/ non-binding) and partisanship (supporters 8

of government party or supporter of opposition parties) affect strongly voters behavior in referendums. Given the fact that it is government who takes the decision to call a referendum, it is no wonder that supporters of government party vote more favorable to ratification than partisans of opposition parties. It seems thus imperative to take into account the institutional factors of referendums in analyzing the voters attitude in referendums. Ⅲ The French Case in Perspective, 1992 and 2005 In the preceding chapter we have examined the classification of referendums and some existing theoretical framework for explaining the political actors motivations and voters behavior in referendums. Now that we saw the intimate correlation between referendum types presented here and attitude of both political actors and voters, we analyze in this chapter the case of two French referendums held in 1992 and 2005. In analyzing the political actors motivations and the voters behavior in French referendums, I rely on the theoretical framework examined above. A Background Since the creation of the Fifth Republic in 1958, France has held in total 10 referendums. Referendums in France are often associated with the plebiscite tradition created by Napoleon the Third and enhanced by Charles De Gaulle (Morel, 2005). De Gaulle, who has not been elected by universal suffrage, did not hesitate to use frequently referendums as an instrument of securing his own legitimacy (Dolez, et al., 2003). Throughout the history of the Fifth Republic, this instrument of direct democracy has contributed to strengthen president s own power by legitimizing a specific policy choice. And French presidents have always understood the power of the referendum (Vassallo, 2007). Out of 10 referendums held by Fifth Republic, three were held on European 9

integration, in 1972, 1992 and 2005. As we have noted earlier, there is no legal requirement in the French Constitution to hold a referendum on the EU treaties (Hainsworth, 2006; Hug and Sciarini, 2000). All these three French referendums on European integration were thus non-required and binding, if we classify them. All the EU referendums held in France have attracted considerable attention of French public. In particular, both in 1992 and 2005, French public showed the strong interest to the referendums which could change the future of the EU. The turnout rate in EU referendums confirms the strong interest that French voters displayed to the EU referendums (Table 1). If we compare the EU referendum turnout with other European Election turnout, we can see the difference: the former is higher than the latter. The turnout of around 70% in 1992 and 2005 was largely due to the intensity of the debates in both cases (Marthaler, 2005). Table 1: Evolution in Turnout Rate in European Polls in France, % 23 April 1972 20September 1992 13 June 2004 29 May 2005 EEC Enlargement Maastricht Treaty European Election EU Constitution Turnout Rate 60.24 69.69 42.76 69.34 Source: Flash Eurobarometer 171 The European Constitution: Post-Referendum Survey 2005 On 20 September 1992 the French electorate voted by 51% to 49% to ratify the treaty on enlarging the powers of the European Community agreed by the 12 member states at Maastricht in December 1991. The question asked was Do you approve the draft law put to the French people by the President of the Republic authorizing the ratification of the Treaty on European Union? (Criddle, 1993). On 29 May 2005, the French electorate went once again to the poll to vote on the issue of whether to support the adoption of European Constitution for the EU. The question that was set was Do you approve the proposed law authorizing the ratification of the Treaty establishing a Constitution for Europe? (Hainsworth, 2006). This time, the French rejected the proposal by 55% to 45%. 10

Given the impact of failed result in 2005 referendum, some scholars argue that the 2005 case was totally different in comparison to the successful result in 1992. However, despite the contrasting result in the two French referendums, two cases seem to show more similarities than differences (Sylvain and Tiberj, 2006; Vassallo, 2007). Similarities have been observed between the 1992 and 2005 referendums in many respects: situation, circumstances, political actors decisions and voters attitude. A consideration of these similarities and differences between two French cases would be instructive for understanding the patterns of political actors attitudes and voters behavior. Particularly, in 1992, the left-wing party (Socialist Party, PS) was in office while in 2005, the right-wing president called a referendum. This different but similar situation will enable us to draw more precise conclusion. B Political Actors Motivations in 1992 and 2005 Referendums The three EU-related referendums held in France were all initiated by the President of Republic. What have forced the President to choose to call a referendum whose result was unpredictable, rather than relying on parliamentary ratification procedure in 1992 and 2005? In this section, I will attempt to answer this question by taking a closer look at the context in which the decision to hold a referendums has been taken. Before undertaking this task, it is worth identifying some similarities of background in two cases. First, both François Mitterrand and Jacques Chirac were in their second term of office. In 1988, Mitterand was reelected President by winning a victory over Chirac. In the presidential election of 2002, after seven years in office, Chirac secured a second term by gaining a massive second ballot victory over Jean-Marie Le Pen of the National Front (Front National: FN). The second similarity is related to their low level of popularity. Due to increasing socioeconomic unrest, low economic growth and rising unemployment, Mitterrand and Chirac suffered lack of popularity support at the time of the referendums. Table 2 illustrates high level of unpopularity they faced at that time. The third element of similarity between 1992 and 2005 is that the important elections were to be held in the nearest future: general election due in March 1993 for Mitterrand and presidential election due in 2007 for Chirac. These commonalities mentioned here 11

appear to be crucial in explaining the reasons for which Mitterrand and Chirac have decided to call a non-required referendum. Table 2: Evolution of President Popularity, % 06/1988 06/1989 06/1990 06/1991 12/1991 06/1992 Mitterrand 63 (35) 62 (35 52 (45) 52 (45) 31 (61) 37 (60) 06/2002 06/2003 12/2003 05/2004 07/2004 05/2005 Chirac 50 (48) 49 (49) 40 (58) 32 (65) 34 (64) 32 (66) Source: TNS-Sofre/ Figaro-Magazine. We will now examine more closely the two Presidents choices. We first begin with examining Mitterrand s choice. As we have seen, Mitterrand s popularity levels in 1992 were at its lowest (De la Serre and Lequesne, 1993). Hoping end his presidency on a positive note, Mitterrand was seeking to improve his popularity levels. For him, a successful national referendum on a very important issue was the perfect opportunity to accomplish this purpose (Vassallo, 2007). The second reason concerns his electoral strategy. Mitterand aimed at making significant political gains through referendum, since his governing Socialist Party (Parti socialiste: PS) has experienced bad electoral score since 1988. For example, in regional elections in March 1992, the PS was reduced to 18% of the vote. Mitterrand s strategy was to create divisions among his political opponents (especially right-wing parties) and to attract the centrist voters (Criddle, 1993). For his party, it was imperative to attract centrist voters (in particular centrist voters of the UDF), because the PS required additional votes to win at the second ballot in parliamentary elections. These electoral consideration played great role in Mitterrand s decision to call a referendum. This was also the case for George Pompidou who called a referendum in 1972 to ratify the enlargement of the European Economic Community (EEC) to Britain, Ireland and Denmark. At that occasion, Pompidou intended to consolidate his own power through 12

referendum by mobilizing his own camp and by dividing his opponents (Bogdanor, 1994; Criddle, 1993). Moreover, Mitterrand perceived Maastricht Treaty as an instrument both for reasserting France s voice in European institutions and for weakening the dominant position of Germany in European economy. Through the ratification of Maastricht Treaty, he wanted to secure his place in history as a great French statesman (Criddle, 1993). Maastricht referendum was intended to serve these purposes. All this considered, we can assume that in the case of 1992 referendum, the political strategic calculations played the most significant role in the Mitterrand s decision to hold a referendum in 1992. Of course, Mitterrand indicated clearly that the referendum had nothing to do with domestic political considerations (Le Monde, 3.07.92). He indicated that voting for Europe was one thing, voting for him was another. But at the same time, this implies that the possible rejection of Maastricht Treaty by the French should not be linked with his resignation (Le Monde, 3.07.92). This can be interpreted as Mitterrand s strategic consideration of detaching his own position from the possible failure of referendum. The situation of 2005 was not very different from that of 1992. Chirac had been reelected President in 2002 by 85% to 15%. However, the vote in 2002 had been interpreted less as a vote for Chirac than as a vote against the National Front (Front National: FN) s leader, Jean-Marie Le Pen (Hainsworth, 2006). It was thus no surprise that the election results had not been favorable to the UMP (Union for a Popular Movement, Union pour un movement populaire), Chirac s governing party. The right suffered thus several electoral defeats in 2004. For instance, in June 2004, the UMP won only 16.64% of the vote. These outcomes stemmed from the growing unpopularity of both Chirac and Raffarin, which was largely due to unfavorable socioeconomic situation in France (Grunberg, 2005; Hainsworth, 2005). Given this low level of popularity, the political risks involved in calling a referendum were considerable, because the referendum vote could at any time turn into the protest vote (Grunberg, 2005; Marthaler 2005). But the referendum could also be used as an instrument for improving the situation. Chirac seemed, in this sense, to remember Mitterrand s strategy from 1992 (Vassallo, 2007). Chirac intended to use referendum to improve his popularity level. The electoral strategic considerations were also important for understanding 13

Chirac s choice. His decision to hold a referendum was aimed directly at the reelection in 2007 presidential election (Vassallo 2007). For this purpose, Chirac was seeking to hold his centre-right rival, Nicolas Sarkozy in check. Chirac considered also the referendum as a political opportunity for dividing his opponents, in particular the PS (Marthaler 2005; Taggart, 2006). At that time, there was no consensus on ratification of the European Constitution among left-wing parties. Chirac was aware that a referendum would be difficult and divisive for the PS. This led Chirac to hope to gain a major political advantage by aggravating divisions. Other factors than strategic one might also account for Chirac s choice in 2005. Chirac could not ignore the fact that there was an obvious precedent for the referendum (1972 and 1992 EU referendums). Furthermore, in 2005, public and elite opinion was in favor of consulting directly the people about the issue of the European Constitution (Hainsworth, 2006; Marthaler, 2005). The British decision to call a referendum might also have some influence over Chirac s determination (Closa, 2007; Taggart, 2006). Nevertheless, we cannot deny the fact that the considerations and calculations related to domestic politics were the main factor for explaining Chirac s motivations in 2005. The analysis of the two French cases confirm that, in both cases, the political actors decisions for calling a EU referendum were strongly related to political actors domestic tactical considerations. This empirical evidence shows that, out of the two theoretical approaches (strategic or institutional) presented above, strategic approach appears to be more relevant for explaining political actors motivation for calling a non-required referendum. The role of political actors is important, not only in deciding whether or not to call a non-required referendum, but also in deciding when to hold it (Hobolt, 2006). The French case also illustrates that the strategic calculations came into play in setting the timing of both announcement and vote on referendum. The strategy taken by Mitterrand had something to do with the successful outcome of referendum (Vassallo 2007). In 1992, the referendum was announced on 3 June, the day after the narrow rejection of the Maastricht Treaty by the Danish voters and was held on 20 September. Therefore, there was not much time left to the opposition to organize the effective campaign. It was not the case for 2005 referendum. The referendum was announced on 14 July 2004 and held 14

on 29 May 2005. Because of the very long time between its official announcement and the actual vote, the No camp was able to reverse the outcome of referendum (Vassallo, 2007). C French Voters Behavior in 1992 and 2005 Referendums How did the French voters behave in two EU referendums called by President as a result of strategic calculations? Did they take into account the political actors strategies in voting in referendums, as has been argued by Hug and Sciarini (2000)? It seems reasonable to think that voters who are close to opposition parties are generally indifferent to the government s propositions. I will thus consider first the attitudes of supporters of government party in order to answer this question. So, what was the voting behavior of supporter of government party in 1992 and 2005? Table 3 shows the evolution of the percent of three main parties supporters voting Yes in two referendums in France. We can note here an important and decisive change between 1992 and 2005. According to exit polls, the PS partisans who supported largely the ratification in 1992 (84%) became much less favorable to the ratification in 2005 (39%). At the same time, the RPR (UMP in 2005) which persuaded only 37% of its voters to support the ratification in 1992 (37%) succeeded in mobilizing the massive yes vote in 2005 (75%). It is also noteworthy that in 1992 the strongly pro-european UDF mobilized less vote favorable than the PS. Table 3: Evolution of the yes vote percent in 1992 and 2005 referendums, % 1992 (n= 2731) 2005(n=2015) PS 84 39 UDF 68 75 RPR(1992)/ UMP(2005) 37 75 Source: 1992: SOFRES Exit poll (reported in Franklin et al., 1993); 2005:Flash Eurobarometer 171(2005). 15

All this data indicates that the partisanship have been the main factor conditioning voters attitudes in EU referendums. That is to say, supporters of government party voted strongly in favor of the issue on which they are asked to vote, whereas these same supporters tended to vote largely against when this same party was in opposition. So what led the French electorate to vote differently according to the position of the party they support within the party system in their country? But before drawing conclusion from what we have seen, one possibility must be examined: possibility that the the UMP s supporters might vote massively in favor of the ratification in 2005 because they supported largely the idea of European integration as a whole. Table 4 provides some data which reveals the development of public opinion over European integration for the period when the European Constitution issue was not yet debated (2000-2002). If we consider the massive support for EU Constitution in 2005 referendum, we can expect the high level of support for European integration for partisans of the UMP. But, according to opinion polls data, partisans of the UMP were not so favorable to the European integration. This does not sufficiently explain the massive support that voters of the UMP gave for ratification in 2005 referendum. We can also note that voters of the PS were more favorable to Europe than those of the UMP. Table 4: Public Support for European Integration, % 12/2000 12/2002 PS supporter UDF supporters Favorable : 56 Against: 43 Favorable : 61 Against: 38 Favorable : 56 Against: 43 Favorable : 68 Against: 32 RPR(2000)/UMP(2002) supporters Favorable : 46 Against: 53 Favorable : 55 Against: 41 Source: CSA Poll (2000), CSA Poll (2002) 16

Furthermore, it had been admitted, through several historical evidences, that support for the Treaty is in general lower than that for the European integration as a whole (Franklin, et al., 1994). But this was not the case for the UMP supporters in 2005 French referendum. From this, we can assume that the strong support that the UMP voters gave in 2005 referendum can be explained by the position of the UMP within the party system (government party). In other words, the UMP partisans were supportive to the decision taken strategically by President Chirac. It seems clear that they were aware of strategic considerations of Chirac and voted on the basis of their partisanship in 2005 referendum. This logic can be also applied to the 1992 case: the PS supporters voted massively in line with their presidential wishes in 1992. Voting behavior of supporters of opposition parties seems to be more complex. Unlike supporters of government party, partisans of opposition parties were more at liberty to vote in referendums because they are less concerned about government s will: they vote either on issue itself, or on their partisanship. Their voting behavior is thus more unpredictable than that of supporters of government party. Nevertheless, in France, in both 1992 and 2005, supporters of the RPR (1992) and those of the PS (2005) voted massively No in EU referendums. This is all the more incomprehensible because the RPR and the PS adopted officially pro-ratification stance in 1992 and 2005 respectively. This means that the majority of supporters of these parties did not follow the party line when their party was in opposition. This leads us naturally to the question of why supporters of the RPR and the PS defected massively from their party line. Why they voted strongly against the ratification? We can note here some similar phenomenon occurred to opposition parties, that is, the RPR in 1992 and the PS in 2005: the intra-party dissent. Both the RPR in 1992 and the PS in 2005 were not united on the referendum issue and divided seriously. This intra-party dissent on the referendum issue has not occurred in government party, that is, the PS in 1992 and the UMP in 2005. In 1992, the two sides of the issue consisted of parts of the RPR: the top leadership of the RPR including Chirac sided beside the Yes camp, while two former ministers in Chirac s 1986-88 government, Charles Pasqua and Philippe Séguin, both of the RPR, 17

occupied No side (Criddile, 1993). Chirac, who aimed at electoral success, had no choice but to take the position on Europe compatible with the centrist voters. But Charles Pasqua and Philippe Séguin who focused on the technocratic implications of the Treaty and its threat to sovereignty, whilst maintaining their commitment to the Treaty of Rome and the Single European Act, took the anti-maastricht Treaty stance (Criddle, 1993). The intra-party dissent was more notable in 2005. The official party line of the PS was determined by an internal poll in December 2004 when 59% of the militants of the PS opted for the Yes vote. But a minority of leaders, especially Laurent Fabius, the deputy leader of the PS and a former prime minister, expressed their disappointment with the European Constitutional Treaty and campaigned actively for the No vote (Brouard and Tiberj, 2006; Grunberg, 2006; Hainsworth, 2006; Marthaler, 2005). Fabius opposition to his party line was interpreted widely as a strategic move to challenge François Hollande, the First Secretary of the PS, for the leadership of the party and to become the main candidate of the PS in 2007 presidential election (Hainsworth, 2006). Crum(2007) who examined the behavior of political parties in referendums on the European Constitutional Treaty in 2005, put forward some conditions in which parties divide on European issue. According to him, oppositions parties that side with the Yes camp are more likely to divide on European issue, because they are torn between pro- European ideological inclinations and strategic considerations as an opposition party (Crum, 2007). This intra-party dissent influences strongly attitudes of the partisans: Rather than taking cues from their preferred party, the supporters of opposition party in which elite opinion diverges are prone to differentiate among the cues and to adopt positions that are inconsistent with the official position of the party (Gabel and Scheve, 2007). Indeed, the two French referendums in 1992 and 2005 confirm these arguments. In two referendums, supporters of the parties where internal divisions have occurred were strongly influenced by this dissent and did not consequently follow the party line. In particular, the 2005 French referendum case supports notably this interpretation. In 2004, when apparent divisions within the party did not yet occur, the supporters of the PS showed the high level of support for European integration in general as well as for the 18

idea of European Constitution. All this considered, we can assume that they voted massively against in referendum vote because they were influenced by this intra-party dissent. The comparison of voters behavior in two French cases showed how supporters of government party as well as of opposition parties behave in EU referendums. This comparison has led to the following observations. Concerning supporters of government party, they perceive referendums as an occasion for expressing their approval for the incumbent government and vote consequently massively in favor of the ratification. This means that the argument of Hug and Sciriani (2000) is valid at least for the French case, according to which in a non-required and binding referendum, the government supporters vote strongly in favor of ratification, because they are aware of the political calculations of the government in place and the political consequences of their vote. In contrast to partisans of the government party, supporters of the opposition parties, although they are more at liberty to vote, are inclined to vote against the ratification. This negative vote can be explained either by the partisanship for the supporters of opposition parties adopting anti-ratification stance, or by the intra-party dissent for the supporters of oppositions parties adopting pro-ratification position. To sum up, voting behavior is largely framed by domestic political context and thus far from the issue voting behavior. Ⅳ Conclusion The aim of this article was to examine whether the EU referendums used increasingly by the member states are dominated by domestic factors. We have attempted to answer this question by analyzing the French case at two levels: political actors motivations to call a referendum and voters behavior in referendums. The analysis of two EU referendums which were held in France in 1992 and 2005, led us to the following two observations. First, in both 1992 and 2005, the political actors decisions for calling an EU referendum were strongly related to political actors domestic strategic considerations. 19

In two cases, political actors used EU referendums mainly to strengthen their own position or to gain electoral advantages. The strategic calculations came also into play in setting the timing of both announcement and vote of referendum. Second, in nonrequired referendums called strategically by political actors, voters of government party support strongly the ratification whatever their attitude toward the issue whereas those of opposition parties are prone to oppose the ratification. In contrast to partisans of government party, supporters of opposition parties, although they are more at liberty to vote, are inclined to vote against the ratification either by the partisanship for the supporters of opposition parties adopting anti-ratification stance, or by the intra-party dissent for the supporters of oppositions parties adopting pro-ratification position This was evidently the case for both 1992 and 2005 referendums. All these findings indicate clearly that referendums on European integration are, throughout all the process, intimately bound up with domestic politics. Given the fact that EU referendums are strongly framed by domestic political context, we can assume that EU-referendums cannot be seen as an instrument for securing the EU policy legitimacy. This conclusion is particularly valid for non-required and binding referendums. But for required referendums, some different conclusion could be drawn. Since the present study is limited to a non-required referendum case, it cannot, unfortunately, provide satisfactory answer to the question concerning the required referendum case. Further research is thus needed to examine whether required referendums can sufficiently legitimize the EU policy. 20

Reference Bogdanor, Vernon. 1994. Western Europe. In Referendums around the world, eds. David Butler and Austin Ranney. Carruba, Clifford J. 2001. The Electoral Connection in European Union Politics. Journal of Politics, Vol.63, No.1 Closa, Carlos. 1994. Referendum as a Legitimatory Instrument in the Integration Process. Paper presented at the ECPR Workshop on Democratic Representation and the Legitimacy of Government in the EC of the 22 nd Annual Joint Sessions of the European Consortium for Political Research, Madrid, 17-22 April. Closa, Carlos. 2007. Why Convene Referendums? Explaining Choices in EU Constitutional Politics. Journal of European Public Policy, Vol. 14, No. 8. Criddle, Byron. 1993. The French Referendum on the Maastricht Treaty September 1992. Parliamentary Affairs, Vol. 46, No. 2, pp. 228-238. Crum, Ben. 2007. Party Stances in the Referendums on the EU Constitution: Causes and Consequences of Competition and Collusion. European Union Politics, Vol.8, No.1, pp.61-82. Dolez, Bernard, Annie Lauren, and Laurence Morel. 2003. Les référendums en France sous la Ve République. Revue Internationale de Politiaue Comparée, Vol. 10, No. 1, pp. 111-127. Franklin, Mark, Michael Marsh, and Lauren McLaren. 1994. Uncorking the Bottle: Popular Opposition to European Unification in the Wake of Maastricht. Journal of Common Market Studies, Vol. 32, No. 4, pp. 455-472. Gabel, Matthew, and Kenneth Scheve. 2007. Party Dissent and Mass Opinion on European Integration. European Union Politics, Vol. 8, No. 1, pp.37-59. Grunberg, Gérard. 2005. Le référendum Français de ratification du Traité constitutionnel européen du 29 mai 2005. French Politics, Culture & Society, Vol. 23, No. 3, Winter 2005, pp. 128-144. Hainsworth, Paul. 2006. France Says No : The 29 May 2005 Referendum on the European Constitution. Parliamentary Affairs, Vol. 59, No. 1, pp. 98-117. Hug, Simon, and Pascal Sciarini. 2000. Referendums on European Integration: Do Institutions 21

Matter in the Voter s Decision?. Comparative Political Studies, Vol. 33, No.3, pp. 3-36. Imig, Doug, and Sidney Tarrow. 2001. Contentious Europeans. Protest and Politics in an Emerging Polity. Lanham: Rowman & Littlefield Publishers, Inc. Laursen, F., and S. Vanhoonacker. 1994. The Ratification of Maastricht Treaty: Issues, Debates and Future Implications. Maastricht: IEPA. Lindberg, Leon, and Stuart Scheingold. 1970. Europe s would-be Polity. Englewood Cliffs: Prentice-Hall. Marthaler, Sally. 2005 The French Referendum on Ratification of the EU Constitutional Treaty, 29 May 2005. Representation, Vol. 41, No. 3, pp. 230-239. Morel, Laurence. 2001. The rise of government-initiated referendums in consolidated democracy. In Referendum Democracy: Citizens, Elites and Deliberation in Referendum Campaigns, eds. M. Mendelsohn and A.Parkin. New York: Palgrave. Obradovic, Daniela. 1996. Policy Legitimacy and the European Union. Journal of Common Market Studies, Vol. 34, No.2. Reif, K. and H. Schmitt. 1980. Nine second-order national elections: a conceptual framework for the analysis of European election results. European Journal of Political Research, Vol. 8, No. 1, pp.3-44. Rouke, J., R., Hiskes, and C. Zirakzadeh. 1992. Direct Democracy and International Politics: Deciding International Issues Through Referendums. Boulder: Lynne Rienner. Setala, M. 1999. Referendums in Western Europe- A Wave of Direct Democracy. Scandinavian Political Studies, Vol. 22, No, 4, pp.327-340. Slater, Martin. 1982. Popular Indifference, Political Elites and Community Building. Journal of Common Market Studies, Vol. 21, No.1. Suksi, M. 1993. Bringing in the People. Dordrecht: Martinus Nijhof Publisher. Taggart, Paul. 2006. Questions of Europe- The Domestic Politics of the 2005 French and Dutch Referendums and their Challenge for the Study of European Integration. Journal of Common Market Studies, Vol. 44, Annual Review, pp. 7-25. Vassallo, Francesca. 2007. The Failed EU Constitution Referendums: The French Case in Perspective, 1992 and 2005. Paper prepared for the First Annual Research Conference of the EU Centre of Excellence(EUCE) at Dalhousie University, Halifax, Canada, May 21-23, 2007. 22