Introduction

The rejection of the European Constitution in the French and Dutch referendums of May and June 2005 was the most recent expression of an (often-noted) public disapproval or disengagement with the European Union (EU), which is part of an ongoing problem with EU legitimation (Lord and Beetham, 2001; Eriksen and Fossum, 2004; Mather, 2006).¹ Indeed, since the re-launching of European integration with the Single European Act, Europe's political elite has struggled to convince its populace of both the value and legitimacy of the increasingly empowered supranational tier of European political authority. Besides the French and Dutch no-votes, we have witnessed a series of expressions of popular disapproval, including the French 'petit oui', declining turnout in the elections to the European Parliament, no-votes on European integration in referendums in Denmark, Ireland and Sweden, and an approval rate for the EU that struggles to climb above 50%. Moreover, many of the attempts to respond to these expressions of dissatisfaction have actually exacerbated popular opposition to European integration. Indeed, the European Constitution itself was initially introduced as a means to resolve some of the legitimacy problems faced by the EU (on this tendency for crises of European legitimacy to recur, see Bailey, 2006). Experience therefore suggests the existence of significant obstacles to the realization of a legitimate and popular EU. This article investigates the source of these obstacles, focusing in particular upon the ideas held by the European political elite² as a possible obstacle preventing the legitimation of the EU. It focuses specifically on the interpretation of the French and Dutch referendums on the European Constitutional Treaty.

A Critical Realist Theory of (Elite-Level) Ideas

The formation and effect of ideas has become a major element of study in both contemporary political science (see, for instance, Hall, 1993; Blyth, 2002; Bevir and Rhodes, 2003; Finlayson et al., 2004; Hay, 2004; Quaglia, 2005) and the sub-discipline of EU studies (see, for instance, McNamara, 1998; Kohler-Koch, 2000; Christiansen et al., 2001; Checkel, 2001; Hooghe, 2002). What unites such approaches is their common conviction that ideas play an important role within political processes and in understanding political outcomes. The aim of scholars working within this 'ideational' tradition is to delineate the formation and effect of shared ideas, discourses and ideational frameworks within particular political communities. This is in large part based on the assertion that subjectivity is a socially produced phenomenon, whereby meanings and interpretation are in part formulated within one's social environment. For instance, Wæver argues that, 'it is problematic to ground one's analysis in "given" subjects or objects because both are constituted discursively, and one should therefore study this process of constitution first' (Wæver, 2004, 198). Within this broad ideational perspective, there exists a group of scholars developing a more firmly explanatory method, seeking to assign a causal role to ideas in explaining political outcomes (see, for instance, Blyth, 1997; Berman, 2001; Hay, 2004; Ryner, 2004). From this perspective, it is not simply the formation of common ideational frameworks that need be understood in order to understand political actions and the meanings attached to them. Instead, certain ideas are assigned causal power on the basis of their ability to interact with, and thereby affect, social reality. In the words of Colin Hay (2004), 'we should perhaps be less concerned with (the process of) understanding per se, and more concerned with the role played by particular understandings in motivating the political conduct of particular actors' (p. 147). This approach therefore aims to explain political outcomes in terms of the ideas held by the actors producing them. Thus, while ideas are held to be views about social reality, in addition ideas can also have a causal effect upon the form that reality takes. In this sense, ideas are more than simply the meanings that are attached to certain discourses, but rather form the analytical or 'strategic' basis for the intentional human action that produces the social world as we experience it (for an elaboration of this view, see Hay, 2002, 194–215).

This claim, that ideas have a causal impact upon reality, particularly as presented by Hay, draws largely on the critical realist philosophy of social science (CR) (for applications of this approach within political science see Hay, 1995, 2002; McAnulla, 2005; O'Reilly, 2005; for the key texts outlining CR, see Bhaskar, 1975, 1998; Sayer, 1992). Of importance for this study, CR views social reality as reproduced by the sum activity of socially-embedded individuals on the basis of their ideas about that society. Thus, each action is predicated upon a set of beliefs about the likely effects of that action upon social reality, which forms an integral element in the process through which that reality is itself (re)produced. Social reality is therefore considered 'concept-dependent' because it sustains itself, in part, on the basis of the ideas that its members hold about it. Thus, for Bhaskar, 'social structures exist only in virtue of the activities they govern, they do not exist independently of the conceptions that the agents possess of what they are doing in their activity, that is, of some theory of these activities' (Bhaskar, 1998, 38). On the other hand, social reality is also considered to be knowledge-independent. This means that social entities (including ideas) can be more or less accurately known. Indeed, from a critical realist perspective, it is the knowledge-independent status of reality that provides the basis for scientific activity. The conduct of social scientific investigation presumes that existing knowledge about the social world is incomplete, or even potentially erroneous, and the rationale for that investigation is therefore to attain a more accurate account of the social world (this therefore contrasts starkly with some of the relativist claims implicit within social constructivism and certain discursive approaches, on which see Patomäki and Wight, 2000).

Once we accept that knowledge about reality is able to differ from 'actual' reality, we open the possibility that individuals may conduct certain activities on the basis of erroneous information, or misperception. In this case, we are likely to witness a systematic gap between intended and actual outcomes (or, alternatively, the production of intended outcomes, but for reasons other than those perceived by the agent; Hay, 2004, 147–149). Thus, the attempt by social scientists to improve upon existing knowledge of social reality has the potential to better inform agents and therefore provide the basis for activity that will contribute to an improvement of existing social structures. This is particularly the case if routine misperception can be explained in terms of the social relations within which those holding erroneous beliefs are located (Bhaskar, 1998, 48).

The aim of the present study is to develop some of the insights gained from CR to identify possible ideational causes for the difficulties experienced by the European political elite in their attempt to resolve the problem of EU legitimation. Ideas are therefore viewed in this study as the strategic analyses informing intentional human agents regarding the likely (social) effect of certain actions. These ideas form the basis for actions, which in turn (re)produce the social structures from which those ideas/actions emerge. From this perspective, we can view the ongoing inability by members of the EU-level political elite to rectify the EU's legitimacy crisis as potentially due to incomplete or incorrect knowledge, or a misperception of the social terrain within which members of the elite seek to act. Indeed, existing research indicates the existence of a gap in attitudes towards the process of European integration between elites and the public, which provides a potential explanation for the inability of the former to produce a polity considered legitimate by the latter. For instance, survey results suggest almost double the level of support for European integration among elites than members of the EU public (Hix, 2005, 165). Similarly, Flockhart (2005) identifies a greater tendency and inclination for members of the Danish political elite to adopt a 'Europeanized' identity than members of the Danish public. Further, Hooghe (2003) identifies a higher level of elite support for EU-level policies, particularly on 'sovereignty' issues as opposed to 'redistributive' policies. Given that the European political elite are more positive about European integration (especially, according to Hooghe (2003), in the case of policies that retain the existing pattern of resource allocation), we might expect the elite to also be more likely than the European public to overlook problems or defects with the current political system. Moreover, given that the present European polity empowers members of the European political elite, we can derive, in accordance with the critical realist method outlined above, an explanation for this oversight in terms of the social relations that comprise the European polity. Thus, the less critical nature of elite-level ideas about the European polity may itself form part of the experience of operating as members of the elite produced by the polity in question. The working hypothesis that forms the basis for this study, therefore, is that members of the European political elite underestimate (or misperceive) the obstacles to the legitimation of the EU polity due to their empowerment by that polity itself, thereby explaining the ongoing problems with EU legitimation. This study attempts to investigate this hypothesis through an examination of the way in which the public rejection of the European Constitution was perceived by members of the European political elite.

Elite Ideas and the Rejection of the European Constitution: Misperceiving Reality?

In seeking to investigate the working hypothesis outlined above, the research presented herein examines the way in which members of Europe's political elite interpreted the French and Dutch no-votes in the 2005 referendums on the European Constitution. This represents an important case study in how members of the European political elite perceive the views and demands of the European citizenry and, thereby, view (obstacles to) the legitimation of the EU. The following section provides a brief introduction to the French and Dutch no-votes, before presenting the results of a survey designed to identify the perceptions held by the European political elite regarding the public motivation for the no-votes.

Background

Following the inability to reach sufficient agreement during both the Amsterdam and Nice Treaty negotiations, the European Convention was given the task of improving the democracy, transparency and efficiency of the EU (European Council, 2001). This was in part a response to a perceived malaise among the European public towards the EU, particularly in the light of the Irish 'no' vote on the Nice Treaty and the resignation of the Santer Commission. Thus, the Laeken European Council noted, Citizens undoubtedly support the Union's broad aims, but they do not always see a connection between those goals and the Union's everyday action. They want the European institutions to be less unwieldy and rigid and, above all, more efficient and open. ... More importantly, however, they feel that deals are all too often cut out of their sight and they want better democratic scrutiny (European Council, 2001).

Nevertheless, the Convention reproduced a number of the problematic practices that were initially intended to be avoided. Most notably, the Convention was a typically elitist and hierarchical organization, comprising (representatives of) national governments, national parliaments, the Commission, MEPs, and led by Valéry Giscard D'Estaing, who 'presided over it in regal splendour drawing on all his attributes and experience ... to overawe the conventioneers' (Dinan, 2004, 28). Moreover, the subsequent intergovernmental wrangling over the terms of the Convention's draft Treaty further increased the elitist nature of the Constitution. Indeed, Magnette and Nicolaïdis (2004) argue that the knowledge that these intergovernmental negotiations would occur actually overshadowed the terms of the deliberations of the Convention itself.

During the campaign leading up to the French referendum, the mainstream political parties in France all supported ratification of the Constitutional Treaty, while the parties on the far right and far left all opposed it (thereby reflecting a general trend among European parties, see Hooghe et al., 2002). As a result, important domestic factors structured the form of the campaign (Taggart, 2006), although controversial factors included EU-level issues such as the unpopularity of the European Commission, the Bolkestein Directive, economic migration by East European workers into France, the principle of free trade and commitment to 'an internal market where competition is free and undistorted' (Art. I: 3-2), and the text of the Constitution itself (Milner, 2006; Hainsworth, 2006; Ivaldi, 2006). The final vote witnessed a 55% no-vote, on a 70% turnout.

In the case of the Dutch referendum, the Constitution again received the widespread support of the Dutch mainstream political elite, with the three governing parties — Christian Democrats, the Liberals and Democrats '66 — and two of the opposition parties — the Labour Party and GreenLeft — each supporting the Treaty (these parties collectively represented 85% of the seats in the Dutch lower house). This reflected a national division between a highly pro-European Dutch political elite and a more hesitant, uninformed and apathetic public (Aarts and van der Kolk, 2006). The 'yes' campaign were relatively inactive in the initial stages of the debate, enabling advocates of a no-vote to argue that European integration was advancing too rapidly, threatening national interests and national identity. The yes campaign was therefore forced to take a reactive stance and concentrate predominantly on refuting the claims that had already been made by the 'no' campaigners. This weak start was further exacerbated by a series of mistakes by the 'yes' campaign, in which claims that a no-vote would threaten post-war European peace, would produce a recession, or would produce a repeat of events similar to those experienced in Nazi Germany or during the Balkans conflict merely produced the appearance of a political elite that was panicked, lacked control of the political situation, and that undervalued the opinions of the electorate (Harmsen, 2005; Qvortrup, 2006, 92–93). The final vote witnessed a 61.5% no-vote, on a 63.3% turnout.

Method

The research presented herein sought to identify the views of the European political elite regarding the reasons for the French and Dutch no-votes, in order that elite-level attitudes about Eurosceptic attitudes could be compared with the reasons identified by Eurosceptic voters themselves, in the case of the Constitutional Treaty referendums. The views of the French and Dutch no-voters were taken from the Eurobarometer surveys conducted immediately after the referendums (Eurobarometer, 2005a, 2005b). In order to access the opinions of the European political elite, a questionnaire was sent to a wide range of the political elite operating at the EU level. The use of these parallel sources of data enabled potential divergences between elite perceptions about public opposition to the EU to be compared with the views of public opponents themselves (in the form of referendum no-voters). This thereby enabled an investigation of the working hypothesis outlined above: that members of the EU-level political elite systematically misperceive the nature of the EU legitimacy question.

The sample was made up of 198 MEPs, 204 Commission fonctionnaires, 25 members of COREPER, 48 members of the Commissioners' cabinets³ and every member (105) of the European Convention. This gave a total sample of 580. The composition of the sample was chosen in order to reflect both the balance of the actors involved in EU-level politics (hence the majority presence of the Commission and European Parliament and a smaller sample from the Council) and also to access political actors with direct relevance to the Constitution itself (hence every member of the European Convention was included). It was also decided to leave out the most high-profile political actors, such as national ministers and members of the College of Commissioners, due to the strong likelihood that the response rate among this group would be very low. The selected sample was initially contacted by email, which minimized the effort involved in responding. In order to ensure an adequate response rate, respondents were guaranteed anonymity and reminders were sent out three times (creating a total of four contacts with the sample), both of which measures commonly increase response rates (see Czaja and Blair, 2005, 209, 213). A number of problems are involved in email surveys, including difficulties in formatting the questionnaire in a way that is compatible with a variety of email systems, the possibility of using inaccurate email addresses, and the difficulties involved in ensuring anonymity for email responses (Czaja and Blair, 2005, 227–228). In order to avoid any problems associated with this method, therefore, two (of the three) reminders were sent by post. Further, in order to increase opportunities to respond, responses were collected over a 12-month period (from September 2005 to September 2006). The response rate was reasonable, with 121 responses (21%), particularly considering that there exists a widely noted problem with response rates to elite-level surveys, in large part due to the experience of both survey fatigue among politicians and officials (Smith and Hay, forthcoming) and also due to a reluctance to provide written answers to potentially controversial questions (Hooghe, 2002). We should also consider that, in absolute terms, 121 responses, by EU actors from a broad range of positions, nationalities and party affiliations, provide an important insight into the opinions of a large number of elite-level actors (see Bochel and Defty (2007) for a similar argument). Further, the responses were reasonably representative (see Table 1). Indeed, the response rates show a significant⁴ response bias in only two cases. There is an over-representation, by 22%, of responses from the Party of European Socialists (PES) group, and an under-representation, by 12%, of the European People's Party and European Democrats group (EPP). However, it should also be noted that 47% of respondents declared no party affiliation at all, so distortions in party representation should not be viewed as reflective of the sample as a whole. Further, in order to guard against potential distortions in the findings as a result of this partisan response bias, all significant claims made in the paper are tested for significant associations with respondents' membership of either the PES or EPP.

Table 1 - Response rates.

Full table

Regarding the questions, those surveyed were requested to answer two questions in which it was asked: 'What ONE factor do you think played the largest role in the [French/Dutch] 'NO' vote to the European Constitution?' The participants were then provided with a list of key factors drawn from the actual responses provided to the Eurobarometer (2005a, 2005b) surveys of French and Dutch citizens (see Appendix A for the list of factors). The questionnaire also asked respondents to respond to the question, 'Should there be a new attempt to adopt a revised European Constitution?' The aim of this question was to identify the extent to which members of the European political elite felt the problems surrounding the adoption of the ECT were surmountable. Finally, those surveyed were asked to identify 'what ONE measure would most improve the public's support for European integration?' and again presented with a list of possible choices (see Appendix B). The aim here was to identify the factors the European political elite focused on in seeking to improve EU legitimacy.

The decision to compare the results of the elite survey with the Eurobarometer post-referendum surveys was not straightforward, particularly given the existence of alternative polls conducted within France and the Netherlands following the referendum results.⁵ Indeed, in the case of France, alternative polls include the exit-polls conducted by IPSOS (2005) and the post-referendum survey conducted by TNS SOFRES (2005) (see Ricard-Nihoul (2005) for an interesting comparison of the main findings of the various post-referendum surveys). Similarly, following the Dutch referendum, an exit-poll was conducted by, among others, Interview-NSS⁶ (see Harmsen (2005) for an overview). However, the Eurobarometer survey was used in the present study for a number of reasons. Firstly, it provided the most detailed insights into the motives expressed by no-voters in both countries. For instance, whereas the IPSOS survey identified eight categories of motives for no-votes, and the Interview-NSS identified the top 10 expressed motives, the Eurobarometer polls identified 18 different motivations in both cases. Secondly, the Eurobarometer surveys had significantly lower levels of multiple responses for individual participants than the alternative surveys in both countries, thereby providing significantly more focused access to the priorities attached to the different motives of no-voters. For instance, whereas the accumulated IPSOS survey responses amounted to a total response level of 282% (indicating an average identification of nearly three (of eight) motives for each respondent), the French Eurobarometer survey totalled only 186%. Similarly, in the Netherlands, the Interview-NSS poll identified the top 10 motives expressed by no-voters, which totalled 455%, while the Eurobarometer survey amounted to 157%. The Eurobarometer polls therefore provide greater insight into the priorities attached to specific motives as respondents identified a smaller number (i.e. the most highly prioritized) of motives. Thirdly, the Eurobarometer survey represents the 'official' survey into the motivations of the no-voters in both France and the Netherlands. Indeed, as it was conducted by the European Commission itself, and as the results had been published before the elite-level survey was conducted, we can assume that EU elite-level actors had access to this information and were relatively well-disposed towards the credibility of its results. An instance of divergence between the Eurobarometer survey results, and the views of the members of the EU political elite regarding the reasons for no-votes, would therefore suggest considerable conviction among members of the elite with regard to the reasons for the no-vote identified by themselves. Finally, the official nature of the Eurobarometer surveys, and the fact that they were conducted by the same organization, ensure that a uniformity of method is employed for the two (non-elite) surveys, thereby making them particularly suitable for comparison. Despite these advantages with the Eurobarometer data, all significant findings identified below were cross-checked, when possible, against the results of either the IPSOS (France) or Interview-NSS (Netherlands) polls in order to confirm the validity of the findings. In each case, the alternative data were used to identify new chi-square values, and failed to produce P-values greater than 0.05 in all but one instance (see below), thereby further confirming the claims made herein.

Results

The responses to the first set of questions can be seen in Table 2. This shows that the factors identified by members of Europe's political elite differed considerably between the two referendums. In the case of the French referendum, the fear of economic liberalization (33%), a protest vote against the national government (27%) and protest against national economic conditions (14%) were the factors most commonly identified in explaining the no-vote. In the case of the Dutch referendum, an anti-EU/nationalist sentiment (16%), opposition to deeper integration (14%), opposition to Turkey's application to join the EU (13%) and a protest against Europe's political system (12%) were the factors most frequently identified. In order to compare these responses with the actual responses given by members of the French and Dutch public in the Eurobarometer survey, the various responses were grouped together into six different categories: opposition to European integration and/or support for national sovereignty; opposition to the EU (neo-liberal) economic model; opposition to the form of the European Constitution; opposition to enlargement; a failure to understand the Constitution; and a protest against national factors. Those responses that were given as 'other' were either re-coded in terms of the existing codes (where possible) or discarded (as they were too varied to provide any useful insight into a general view held within the European political elite) (see Appendix C). The purpose of grouping these variables together was to facilitate comparison of broad trends of opinions within both the European political elite and the French and Dutch citizenry, thereby enabling a comparison in general terms of the factors that both the public voted against and the EU political elite perceived them to be voting against. This aggregation of the responses facilitated comparison between the surveys by grouping together similar, but sometimes differently worded, motives. It also increased the likelihood of being able to make valid statistical comparisons between the responses, as the number of cases within each category was obviously increased (thereby reducing the possibility of cells with expected counts less than 5 for the purposes of measuring valid chi-square values). While this may have blunted slightly the detailed insight that the studies provide into the motives for no-voters and the elite perceptions of those motives, nevertheless it does provide a useful view of the broad area of focus that both elites and the public look to with regard to EU legitimation. Moreover, the results in Table 2 and in the Eurobarometer (2005a, 2005b) surveys can be used to derive more detailed comparisons.

Table 2 - Factors identified by the European political elite in explaining the French and Dutch no-votes (percentages).

Full table

Figures 1 and 2 compare the results of the elite-level survey and the actual public responses provided in the Eurobarometer survey. These figures show a clear discrepancy in both cases between the factors identified by members of the political elite in explaining the French and Dutch no-votes, and the actual reasons identified by the no-voters themselves. Indeed, these differences can be seen more clearly in Tables 3 and 4, which show the percentage difference between the frequency with which a factor was identified by members of the political elite in explaining the rejection of the European Constitution and the frequency with which a factor was identified by the French and Dutch no-voters themselves, along with a chi-squared test of association between whether a respondent is a member of the political elite or not, and the factor identified in explaining the no-vote.⁷ For those chi-square tests that produced a P-value less than 0.01 we can assume a significant association between the variable measuring whether or not a respondent is a member of the European political elite and his/her identification of a particular causal factor in explaining the reason for the no-vote. The factors for which responses significantly differed according to membership, or not, of the political elite are shown in bold in Tables 3 and 4.⁸ Further, for each of the responses for which a significant association is identified, an additional chi-square test was performed to measure the significance of party affiliation in order to identify possible effects of the response bias of PES and EPP members. In none of these cases is there a significant association between affiliation with the PES or the EPP and the response given, therefore suggesting that the significant partisan response bias noted above has not skewed the results of the research.

Figure 1.

Reasons identified for no-vote, by elite and non-elite, in French referendum.

Full figure and legend (21K)

Figure 2.

Reasons identified for no-vote, by elite and non-elite, in Dutch referendum.

Full figure and legend (21K)

Table 3 - Comparing elite ideas and public reponses — France.

Full table

Table 4 - Comparing elite ideas and public reponses — Netherlands.

Full table

In the case of the French referendum, in Table 3 we witness a significant under-estimation by Europe's political elite of the opposition by the members of the French public to the form taken by the European Constitution. Thus, 6% of the public interviewed in the Eurobarometer survey identified their opposition to the form of the Constitution (in particular, its undemocratic and insubstantive nature) as their reason for voting 'no', while no respondents to the elite survey identified this as a factor. We also witness a significant over-estimation of the extent to which the no-vote was a protest against national factors. Thus, 41% of the elite survey felt the no-vote was a protest against national factors (economic or political), while only 27% of the French public claimed to hold such motives.

In the case of the Dutch referendum, there was also a significant misperception of factors producing the no-vote. Thus, members of the European political elite most significantly over-estimated the strength of the anti-enlargement vote against Dutch no-voters. Twenty-five per cent of the respondents to the elite survey viewed the no-vote in the Netherlands as a vote against enlargement, whereas in fact only 5% of the public declared that to be the case. Further, elite respondents significantly underestimated the anti-EU (neoliberal) economic model sentiment in the no-vote (identified by 1% of the elite and 8% of the public), the extent to which it reflected an opposition to the form of the Constitution (identified by no members of the elite and 13% of the public),⁹ and the failure of the population to understand the Constitution (identified by 9% of the elite but 22% of the public).¹⁰

In response to the question regarding whether there should be an attempt to adopt a revised European Constitution, the results show a very clear preference among Europe's political elite for a continued attempt to adopt a European Constitution, with 76% of respondents favouring the adoption of a Constitution (see Figure 3). This is actually quite close to the response identified among the European public in the latest Eurobarometer survey, in which 61% of respondents claimed they were in favour of a constitution for the EU (Eurobarometer, 2006). Indeed, if we exclude the 'don't know/no answer' responses from the surveys there are no significant differences between the responses provided in the elite and Eurobarometer surveys. Thus, the only response that does indicate a significant difference is for the 'don't know/no answer' response, with 2.5% of elite responses in this category and 17% of Eurobarometer responses (²=18.1; P<0.01).

Figure 3.

Views on the need to adopt a revised European Constitution.

Full figure and legend (21K)

Finally, in response to the question asking which factor would improve public support for the EU, the results show that responses were fairly evenly distributed over a number of factors (see Table 5). Interestingly, some of the most frequently identified factors were in areas related to the EU's 'democratic deficit', including the need for increased democratic accountability and more public participation, which arguably contrasts with the elitist manner in which the Constitution itself was drafted. This illustrates to a certain extent the dilemmatic nature of the EU's legitimation problem, in that the actions required to legitimate the EU are not always those that EU-level actors actually wish (or are able) to carry out in practice (Bailey, 2006). Also of interest, the results show a marked distinction between the responses given by members of the two main parties (see Figure 4). Thus, most notably, we witness an increase in social policy as the factor identified most frequently by PES members (33%). In contrast, no EPP members identified social policy as a means of improving EU popularity. Moreover, 12.5% of EPP members identified a reduction in regulation as one of the key factors that could improve EU popularity, compared to only 3.1% of PES members.

Figure 4.

Factors to improve public support for EU, by party.

Full figure and legend (38K)

Table 5 - Elite views on factors that will improve public support for the EU.

Full table

Discussion

There clearly exists a divergence between the ideas held by members of the European political elite and public about EU legitimation, at least with regard to the rejection of the European Constitution. A number of interesting observations can be made. Firstly, and perhaps most strikingly, in both cases elite respondents failed to identify an underlying opposition to the form of the Constitution in their explanation for the rejection of it. This is of interest both because it illustrates a distinct misperception within the European elite regarding the target of popular disaffection with the EU, and because it can, more critically, be interpreted as an explicit redirection of blame by members of the European political elite for the failure of a major element of the European integration project. Indeed, an interpretation of the no-votes as a direct rejection of the actual Constitution can be seen as most damaging for the European political elite (as it represents both a criticism of the general programme of European integration and a more direct criticism of a specific political initiative developed by members of the European political elite themselves). The considerable downplaying of the importance of the form of the Constitution itself in understanding the no-votes in each country, therefore, can be viewed as an attempt to reinterpret popular disapproval in a less negative light. This view is confirmed by much of the post-referendum debate, in which we have witnessed an attempt to depict the no-votes as something other than a rejection of the European Constitution itself. For instance, during the plenary European Parliament debate that followed the referendums, Commission President José Manuel Barroso stated: We must all respect the expression of the citizens' will, but we must honestly recognise that this debate was often mixed up with national issues and that, sometimes, the arguments had little to do with the European Constitution (speech to the European Parliament, 8 June 2005).

Similarly, in his book on the European Constitution, Andrew Duff MEP (ALDE), a keen supporter of the Constitution, notes that, 'the undeniable virtues of the EU's Constitution offered only limited insurance against a vote of dissent against something entirely domestic' (Duff, 2005, 187). It is also confirmed by the findings of Brouard and Tiberj (2006), who present statistical evidence that rejects what they call the 'political distrust' explanation for the no-vote (the view that it was 'a vote of distrust connected to national issues directly rather than to the European issue itself'), despite the importance placed on this factor by the European political elite. Our first observation, therefore, is that members of the European political elite seemingly under-estimate the extent to which the no-vote on the European Constitution is a direct criticism of the Constitution itself, arguably in an attempt to divert blame away from themselves in the light of political failure. This is a particularly interesting development, given the more often-noted occurrence of blame-shifting towards the EU-level by national-level actors (see, for instance, Koenig-Archibugi, 2004), suggesting a two-way 'multi-level' process of blame diversion.

Secondly, if we accept this argument regarding blame diversion, the factors that elite survey respondents have over-estimated, or shifted the blame towards, are also of interest. Indeed, in the case of both France and the Netherlands, there is a clear factor (although it is different in each case) that has been overestimated by elite respondents. In the case of France, the respondents to the elite survey significantly over-estimated the extent to which the no-vote was a protest against national (economic or political) factors. This suggests that members of the European political elite were particularly keen to explain the no-vote as a protest against national political and economic conditions (i.e. factors that did not directly result from the policies and practices they had enacted). In the case of the Netherlands, in contrast, the elite respondents most significantly over-estimated the extent to which the no-vote was a protest against the enlargement of the EU. This also represents a possible diversion of blame towards external factors, in this case problematic states that are (or were) outside of the process of European integration. This interpretation is somewhat supported by a comparison of the distribution of elite responses according to whether the respondents originate from the 2004 accession countries or the pre-2004 EU15 (see Figure 5). Respondents were 10% less likely to identify opposition to enlargement as the reason for the Dutch no-vote if they were from one of the applicant states during the 2004 accession.¹¹

Figure 5.

Difference between new member state respondents and EU15 respondents on the role of anti-enlargement in explaining the Dutch no-vote.

Full figure and legend (22K)

Thirdly, on the results to the elite survey question on whether a revised European Constitution should be adopted, as noted, the most significant finding is that members of the European citizenry responding to the Eurobarometer were less likely to provide a firm, 'yes' or 'no', response than those responding to the elite surveys from the Eurobarometer survey. This can perhaps be taken to indicate a (not particularly surprising) greater level of engagement among the European political elite on this issue than among the European citizenry as a whole.

Finally, the response to the question on factors that will improve popular support for the EU indicates that the way in which elite-level actors view problems with, and potential solutions to, EU legitimacy are in large part coloured by their ideological predisposition. As EU-level social policy is a key demand of the PES, and an opposition to regulation is associated with the position of the EPP (these positions have become increasingly demarcated, at the EU-level, since the mid-1990s, see Gabel and Hix, 2002), the results indicate a tendency for partisan politicians to identify key elements within their own party political programme as important means to improve the popularity of the EU. While this may be unsurprising on one level, it is also true that such divergent opinions cannot simultaneously be correct. Partisan politicians therefore appear to view the problems and potential solutions to European legitimacy in terms that confirm (rather than challenge) their initial viewpoints. This therefore suggests a potential source of political inertia, in that expressions of popular disaffection are interpreted by the European political elite through conceptual lenses that confirm, rather than challenge, existing viewpoints with regard to the EU.

Possible objections

In asking for those surveyed to identify the factor that most explains the no-votes in France and the Netherlands, it is possible that the methodology may have skewed the responses towards the most important factors and away from the less commonly cited factors, in which case the results noted above may reflect the research methodology adopted rather than a divergence between the views of the European political elite and public. While this is certainly a possibility, nevertheless, this criticism is not sustained by the empirical results. Indeed, the results show that in fact the over-estimation and under-estimation of factors is not limited to the most important and least important factors, respectively, thereby enabling a rejection of the claim that distortions reflect the wording of the question rather than an actual discrepancy in elite-level and public views. For instance, in the case of France the most over-estimated factor (national protest) was the second-most cited factor in the Eurobarometer survey, while the most cited factor (anti-economic model) was actually under-estimated in the elite survey. Further, in the case of the Netherlands, the most over-estimated factor in the elite survey was actually the least cited factor in the Eurobarometer survey (anti-enlargement), while the most under-estimated factor was mid-ranking in the Eurobarometer survey (opposition to the form of the Constitution). The discrepancies between elite opinions about public opinion and public opinions themselves, therefore, should not be interpreted as a result of the wording of the questionnaire.

A second possible objection is that there is a disparity in the methodology of the two surveys, in particular between the instruction to identify only one factor in the elite survey, compared with unlimited responses in the Eurobarometer survey. However, it should be noted that in the Eurobarometer survey respondents on average identified less than two factors, therefore significantly mitigating the possibility that this disparity in methods could produce divergent outcomes. This possibility is mitigated further still by the decision to record all responses provided in the elite survey in the case of respondents identifying more than one factor. In this case of multiple responses, responses were weighted so that one respondent would only be allocated the weight of one response overall. The decision to (seek to) limit elite-survey responses to one factor was taken to avoid respondents choosing a large number of responses, thereby preventing the emphasis which respondents placed upon the various factors from being identified.

Conclusion

The research presented herein has focused upon the divergence in ideas between the elite and the European citizenry regarding European legitimation. This has been conducted through a case study of the no-votes on the European Constitution in the French and Dutch referendums of 2005. The survey identified a tendency to re-direct blame for political failure by political actors operating at the supranational level, either towards political actors operating at the national level or to problems existing outside of the EU (in the form of anti-enlargement). The survey also found that members of the political elite were significantly more engaged with the debate on the future of the European Constitution than were members of the European public. Finally, we have witnessed a significant partisan difference between factors identified that have the potential to improve the popularity of the EU, suggesting that elite-level political actors interpret experiences of political failure and political unpopularity in terms that reinforce, rather than challenge, ideological predispositions. The findings therefore suggest that the problems experienced by the European political elite in its attempts to legitimate the EU in part reflects (and can therefore be explained, in part, in terms of) an inclination to (re-)interpret experiences of political failure in a light that diverts attention away from direct criticism of EU-level initiatives and/or that confirms existing ideological predispositions.

Notes

¹ I am very grateful to the anonymous respondents who completed this survey. A coded SPSS file containing the results of the survey is available upon request. I acknowledge the very helpful comments made by the participants of the Mapping Elite Attitudes ECPR workshop 2006, convened by Andrew Gamble and Bruno Cautres, and in particular the comments of Ben Clift. I am also grateful for comments and advice provided by Saori Shibata, Hugh Bochel, Sara Motta, the three anonymous referees, and for the editorial guidance of Ben Rosamond. This research was in part conducted at the Department of Politics and Public Administration, University of Limerick, and at the School of Languages and Social Science, Aston University. All errors are my own.

² For the purposes of this article, members of the European political elite are defined as the major participants in EU-level political decision-making, including representatives of member states, Commission officials and MEPs (for a similar view, see Hooghe, 2003).

³ These were all selected using a systematic random sample.

⁴ A significant bias in response rates is taken here to be anything over 10%.

⁵ I am grateful to one of the anonymous referees for bringing these alternative sources of data to my attention.

⁶ A summary of this exit poll can be found at: http://www.nos.nl/nosjournaal/dossiers/europesegrondwet
/020605_tabel_redenen.html.

⁷ In all the chi-squared results reported in Tables 3 and 4 there were no expected values less than 5 in any of the cells.

⁸ The extent to which the vote was a general statement against European integration and/or for the protection of national sovereignty was overestimated by respondents to the survey in the case of the Netherlands. However, as this was the most commonly identified factor by the public it is possible that the high incidence of elite respondents identifying this factor reflects the wording of the elite survey (which asks respondents to identify one factor which they felt played the largest role in the no-vote) (see Discussion below). For this reason, this result is not considered significant and so is not in bold in the table.

⁹ There was no response in the Interview-NSS survey that could be placed within this category (opposition to the form of the Constitution), thereby preventing comparison in this case. This in part reflects the generality of the terms used, including 'increasing bureaucracy' and 'the EU has more disadvantages than advantages', both of which might have fallen within this category under a more detailed survey.

¹⁰ This divergence was not confirmed by the Interview-NSS survey, where roughly the same proportion of respondents identified 'The provision of information was poor' as the reason for not voting, as did members of the elite survey identify a failure to understand the Constitution.

¹¹ It should be noted that this did not produce a significant chi-square test of association (P=0.347).

References

1. Aarts, K. and van der Kolk, H. (2006) 'Understanding the Dutch "No": the Euro, the East, and the Elite', PS: Political Science and Politics 39(2): 243–246.
2. Bailey, D.J. (2006) 'Governance or the crisis of Governmentality? Applying critical state theory at the European level', Journal of European Public Policy 13(1): 16–33. | Article |
3. Berman, S. (2001) 'Ideas, norms, and culture in political science', Comparative Politics 33(2): 231–250. | Article |
4. Bevir, M. and Rhodes, R.A.W. (2003) Interpreting British Governance, London: Routledge.
5. Bhaskar, R. (1975 [1997]) A Realist Theory of Science (2nd edn), London: Verso.
6. Bhaskar, R. (1998) The Possibility of Naturalism: A Philosophical Critique of the Contemporary Human Sciences, (3rd edn), London: Routledge.
7. Blyth, M. (1997) 'Any more bright ideas?', Comparative Politics 29(2): 229–250. | Article | ISI |
8. Blyth, M. (2002) Great Transformations: Economic Ideas and Institutional Change in the Twentieth Century, Cambridge: Cambridge University Press.
9. Bochel, H. and Defty, A. (2007) 'MPs' attitudes to welfare: a new consensus?' Journal of Social Policy 36(1): 1–17. | Article |
10. Brouard, S. and Tiberj, V. (2006) 'The French referendum: the not so simple act of saying nay', PS: Political Science and Politics 39(2): 261–268.
11. Checkel, J.T. (2001) 'Why comply? Social learning and European identity change', International Organization 55(3): 553–588. | Article |
12. Christiansen, T., Jørgensen, K.E. and Wiener, A. (eds.) (2001) The Social Construction of Europe, London: Sage.
13. Czaja, R. and Blair, J. (2005) Designing Surveys: A Guide to Decisions and Procedures (2nd edn), London: Pine Forge Press.
14. Dinan, D. (2004) 'Governance and institutions: the convention and the intergovernmental conference', Journal of Common Market Studies 42(Annual Review): 27–42. | Article |
15. Duff, A. (2005) The Struggle for Europe's Constitution, London: Federal Trust.
16. Eriksen, E.O. and Fossum, J.E. (2004) 'Europe in search of legitimacy: strategies of legitimation assessed', International Political Science Review 25(4): 435–459. | Article |
17. Eurobarometer (2005a) The European Constitution: Post-referendum survey in France, Brussels: European Commission, Available at http://www.europa.eu.int/comm/public_opinion/flash/fl171_en.pdf.
18. Eurobarometer (2005b) The European Constitution: Post-referendum survey in the Netherlands, Brussels: European Commission, Available at http://www.europa.eu.int/comm/public_opinion/flash/fl172_en.pdf.
19. Eurobarometer (2006) Eurobarometer 65: Public Opinion in the European Union, Brussels: European Commission, Available at http://ec.europa.eu/public_opinion/archives/eb/eb65/eb65_first_en.pdf.
20. European Council (2001) 'The future of the European union — Laeken declaration', Available at: http://europa.eu.int/constitution/futurum/documents/offtext/doc151201_en.htm.
21. Finlayson, A., Bevir, M., Rhodes, R.A.W., Dowding, K. and Hay, C. (2004) 'The interpretive approach in political science: a symposium', British Journal of Politics and International Relations 6(2): 129–164. | Article |
22. Flockhart, T. (2005) 'Critical junctures and social identity theory: explaining the gap between Danish mass and elite attitudes to Europeanization', Journal of Common Market Studies 43(2): 251–271. | Article |
23. Gabel, M. and Hix, S. (2002) 'Defining the EU political space: an empirical study of the European elections manifestos, 1979–1999', Comparative Political Studies 35(8): 934–964. | Article |
24. Hainsworth, P. (2006) 'France says no: the 29 May 2005 referendum on the European constitution', Parliamentary Affairs 59(1): 98–117. | Article |
25. Hall, P.A. (1993) 'Policy paradigms, social learning and the state: the case of economic policy-making in Britain', Comparative Politics 25: 175–196. | Article |
26. Harmsen, R. (2005) 'The Dutch referendum on the ratification of the European constitutional treaty 1 June 2005', European Parties Elections and Referendums Network: Referendum Briefing Paper No. 13, Available at http://www.sussex.ac.uk/sei/1-4-2-9.html.
27. Hay, C. (1995) 'Structure and Agency', in D. Marsh and G. Stoker (eds.) Theory and Methods in Political Science, Basingstoke: Macmillan, pp. 189–206.
28. Hay, C. (2002) Political Analysis: A Critical Introduction, Basingstoke: Palgrave.
29. Hay, C. (2004) 'Taking ideas seriously in explanatory political analysis', British Journal of Politics and International Relations 6(2): 142–149.
30. Hix, S. (2005) The Political System of the European Union: Second Edition, Basingstoke: Palgrave.
31. Hooghe, L. (2002) The European Commission and the Integration of Europe: Images of Governance, Cambridge: Cambridge University Press.
32. Hooghe, L. (2003) 'Europe divided? Elites vs public opinion on European integration', European Union Politics 4(3): 281–304. | Article |
33. Hooghe, L., Marks, G. and Wilson, C.J. (2002) 'Does left/right structure party positions on European integration?' Comparative Political Studies 35(8): 965–989. | Article |
34. IPSOS (2005) La Banque des sondages: Referendum 29 Mai 2005: Le sondage sorti des urnes, Available at: http://www.ipsos.fr/CanalIpsos/poll/8074.asp.
35. Ivaldi, G. (2006) 'Beyond France's 2005 referendum on the European constitutional treaty: second-order model, anti-establishment attitudes and the end of the alternative European utopia', West European Politics 29(1): 47–69. | Article | ISI |
36. Kohler-Koch, B. (2000) 'Framing: the bottleneck of constructing legitimate institutions', Journal of European Public Policy 7(4): 513–531. | Article |
37. Koenig-Archibugi, M. (2004) 'International governance as new raison d'Etat? The case of the EU common foreign and security policy', European Journal of International Relations 10(2): 147–188. | Article |
38. Lord, C. and Beetham, D. (2001) 'Legitimizing the EU: is there a "Post-parliamentary Basis" for its legitimation?' Journal of Common Market Studies 39(3): 443–462. | Article |
39. McAnulla, S. (2005) 'Making hay with actualism? The need for a realist concept of structure', Politics 25(1): 31–38. | Article |
40. McNamara, K.R. (1998) The Currency of Ideas: Monetary Politics in the European Union, Ithaca: Cornell University Press.
41. Magnette, P. and Nicolaïdis, K. (2004) 'The European convention: bargaining in the shadow of rhetoric', West European Politics 27(3): 381–404. | Article |
42. Mather, J. (2006) Legitimating the European Union: Aspirations, Inputs and Performance, Basingstoke: Palgrave.
43. Milner, H. (2006) '"YES to the Europe I want; NO to this one." Some reflections on France's rejection of the EU Constitution', PS: Political Science and Politics 39(2): 257–260.
44. O'Reilly, D. (2005) 'Social inclusion: a Philosophical Anthropology', Politics 25(2): 80–88. | Article |
45. Patomäki, H. and Wight, C. (2000) 'After postpositivism? The promises of critical realism', International Studies Quarterly 44: 213–237. | Article |
46. Quaglia, L. (2005) 'Civil servants, economic ideas, and economic policies: lessons from Italy', Governance 18(4): 545–566. | Article |
47. Qvortrup, M. (2006) 'The three referendums on the European constitution treaty in 2005', Political Quarterly 77(1): 89–97. | Article |
48. Ricard-Nihoul, G. (2005) 'The French "no" vote on 29 May 2005: understanding and action', Notre Europe, Studies and Research No. 44, Available at http://www.notre-europe.eu/uploads/tx_publication/Etud44-en.pdf.
49. Ryner, J.M. (2004) 'Neo-liberalization of social democracy: the Swedish case', Comparative European Politics 2(1): 97–119. | Article |
50. Sayer, A. (1992) Method in Social Science: A realist approach (2nd edn), London: Routledge.
51. Smith, N. and Hay, C. (forthcoming) 'Mapping the political discourse of globalisation and European integration in the UK and Ireland empirically', European Journal of Political Research.
52. Taggart, P. (2006) 'Keynote article: 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 44(Annual Review): 7–25. | Article |
53. TNS Sofres (2005) 'Le referendum du 29 Mai', conducted for TF1, RTL and Le Monde on 29 May, 2005.
54. Wæver, O. (2004) 'Discursive Approaches', in A. Wiener and T. Diez (eds.) European Integration Theory, Oxford: Oxford University Press, pp. 197–215.

Appendices

Appendix A

List of factors from which participants in the survey were asked to identify the most important causal factor in explaining the French/Dutch no-votes:

1. Anti-EU sentiment/Euroskepticism/Nationalism
2. Opposition to 'deeper' European integration
3. Fear of economic liberalization/Constitution too 'neo-liberal'
4. Not enough 'Social Europe' in Constitution
5. EU is too powerful
6. Protest against Europe's political system
7. Protest against national government
8. Protest against national economic conditions
9. Opposition to 2004 enlargement
10. Opposition to future enlargement
11. Opposition to Turkey's application to join the EU
12. Undemocratic Constitution
13. No substance to Constitution
14. Failure to understand Constitution
15. Other (please specify)

Appendix B

List of factors that respondents were asked to choose from in identifying which factors would most improve the public's level of support for European integration.

1. More social policy
2. More employment policy
3. Less regulation
4. More regulation
5. More economic liberalization
6. Better publicity
7. Greater transparency in decision-making
8. Less EU-level expenditure
9. More EU-level public policies
10. Increased democratic accountability
11. More public participation
12. End to enlargement
13. Other (please specify)

Appendix C

Categories of responses

1. Anti-European integration/pro-national sovereignty

Elite Survey

 Anti-EU sentiment/euroskepticism/nationalism

 Opposition to 'deeper' European integration

 EU is too powerful

 Protest against Europe's political system

Eurobarometer (France)

 Loss of national sovereignty

 I am against Europe//European construction

 The draft goes too far/advances too quickly

 I do not want a European political union/a European federal state/USE

Eurobarometer (Netherlands)

 Loss of national sovereignty

 I am against Europe/European integration/European construction

 The draft goes too far/advances too quickly

 I do not want a European political union/a European federal state/USE

 Europe is evolving too fast

 Europe is too expensive

 Loss of the Dutch Identity

 I do not trust Brussels

2. Anti- EU (neoliberal) economic model

Elite Survey

 Fear of economic liberalization/Constitution too neo-liberal

 Not enough Social Europe in Constitution

Eurobarometer (France)

 Economically speaking, the draft is too liberal

 It will have negative effects upon the employment situation in France

 I am against the Bolkestein Directive

 Not enough Social Europe

Eurobarometer (Netherlands)

 It will have negative effects on the employment situation in the Netherlands/relocation of Dutch enterprises/loss of jobs

 Economically speaking, the draft is too liberal

 Not enough social Europe

3. Anti-form of EU constitution

Elite Survey

 Undemocratic constitution

 No substance to constitution

Eurobarometer (France)

 The draft does not go far enough

 Not democratic enough

 I do not see what is positive in the text

 Too technocratic/juridical/too much regulation

Eurobarometer (Netherlands)

 I do not see what is positive in the text

 Not democratic enough

 This constitution is imposed on us

 Too technocratic/juridical/too much regulation

4. Anti-enlargement

Elite Survey

 Opposition to 2004 enlargement

 Opposition to future enlargement

 Opposition to Turkey's application to join the EU

Eurobarometer (France)

 Opposition to future enlargement

 Does not want Turkey in the EU

Eurobarometer (Netherlands)

 Opposition to further enlargement

 Does not want Turkey in the EU

5. Failure to understand constitution

Elite Survey

 Failure to understand constitution

Eurobarometer (France)

 Lack of information

 Too complex

Eurobarometer (Netherlands)

 Lack of information

 Too complex

6. Protest against national factors

Elite Survey

Protest against national government

 Protest against national economic conditions

Eurobarometer (France)

 Opposes the president of the Republic/the national government/certain political parties

 The economic situation in France is too weak/there is too much unemployment in France

Eurobarometer (Netherlands)

 Opposes the national government/certain political parties

 The economic situation in the Netherlands is too weak/there is too much unemployment in the Netherlands

 The Netherlands must first settle its own problems

7. Other

Elite Survey

 Other

Eurobarometer (France)

 Other/don't know

Eurobarometer (Netherlands)

 Other/don't know

 The yes campaign was not convincing enough

 There is nothing on human rights or animal rights

 Influenced by the 'No campaign'

About the author

David J. Bailey is Lecturer in the Department of Political Science and International Studies, University of Birmingham. He completed his Ph.D. at the London School of Economics, where he focused on the relationship between social democratic parties and their position on European integration. He has published articles in the Journal of Common Market Studies, Journal of European Public Policy, and Capital and Class. He is currently working on a manuscript — The Political Economy of European Social Democracy: A Critical Realist Perspective — to be published with Routledge in 2008.