The EP elections and political representation through national parties: the Italian case

The EP elections and political representation through national parties: the Italian case Edoardo Bressanelli* and Daniela Piccio^ * Istituto Universitario Europeo: edoardo.bressanelli@eui.eu ^ Istituto Universitario Europeo/Leiden University: daniela.piccio@eui.eu Paper preparato per la Conferenza annuale della Società Italiana di Scienza Politica, Università degli Studi di Palermo, 8-10 settembre 2011. Panel Rappresentanza e Partiti Politici nell'unione Europea ABSTRACT In this paper, we argue that the 'democratic deficit' of the European Union (EU) might not be as critical as it is often suggested if: (i) a policy congruence between the national political parties and the EP groups to which they associate and (ii) a congruence in the relevant dimensions shaping inter-party competition at the national and the EU levels - exist. When these two conditions are met, citizens preferences would be channelled into the EP groups, through the national parties. Drawing on the data developed by the EU Profiler project, the first Europe-wide Voting Advice Application developed for the 2009 EP elections, this paper empirically measures the degree of congruence between the positions of the Italian parties and the EP groups with respect to the key issues characterizing party competition in Europe. Despite contrasting theoretical expectations, we find that congruence between the Italian and the EP parties and party system exists. On the bases of the evidence for the critical case of Italy, we argue that political representation of the European citizens is taking place nothwithstanding the poor democratic functioning of the representative process in the EU. 1

Introduction In the last few years with the heyday ahead of the June 2009 elections for the European Parliament (EP) the leaders of the Italian political parties have frequently made public remarks about the affiliation of their party with a EU-level organization. For instance, after the split of the Popolo della Libertà (PDL), members of the dissenting wing (Futuro e Libertà) have often accused Berlusconi's party not to fit the underlying ideology and policy programme of the European People's Party (EPP), while the new PDL secretary has claimed that his party is moving towards the constitution of the Italian EPP. On the other side of the policy spectrum, the small Partito Socialista Italiano has frequently emphasized its membership in the Party of the European Socialists vis-a-vis the larger, but more ideologically shaky, Partito Democratico (PD). The issue of transnational affiliation has often been used to play a competitive game with the other parties within the domestic party system, but the real ideological/policy compatibility between the 'new' Italian parties and their EP group has never been systematically measured. In order to fill this gap, this work places itself within the growing field of the academic literature focusing on the potentials of political representation through political parties in the European Union (i.e. Bardi et al. 2010; McElroy and Benoit, 2007; Thomassen, 2009). While the system of political representation linking the citizens and their representatives at the EU level remains structurally underdeveloped, we argue that the presence of a policy congruence between the national political parties and the political groups in the EP, and the similarity in the structure of competition between the party systems at the two levels, would nonetheless make it feasible. Channelled by national parties, the will of the citizens of the Union would find expression, albeit indirectly, in the positions of the party representatives at the European level, as is formally stated by Art. 10.4 of the Lisbon Treaty. 1 On the bases of the data gathered by the EU-Profiler project, the first Europe-wide Voting Advice Application (VAA) designed for the EP elections of June 2009, mapping the policy positions of about 270 national parties in 30 countries on the most salient dimensions of inter-party competition (Trechsel and Mair, 2011), we empirically measure the congruence between the policy positions of the Italian parties 1 We use the label 'EU-level parties' when referring, in general, to the role of parties as representative agents. We use the label 'EP groups' when we refer, specifically, to the party organizations in the EP. 2

and those of the EP groups to which they affiliate. Congruence is observed both in terms of policy agreement between national parties and EP party groups positions as well as in terms of agreement between the main dimensions of party competition. Hence, the specific questions that this work will address are: (i) how closely do the positions of the Italian parties match those of the EP groups to which they are affiliated? (ii) To what extent do the spaces of party competition in Italy and in the EU resemble each other? We will argue that neither policy congruence between the national parties and the EP groups, nor the presence of similar structures of inter-party competition between the two levels can simply be taken as assumptions. This applies even more strongly to the Italian case, where the dramatic events of the early Nineties produced a complete reconfiguration of the individual parties' identities and of the structure of party competition more broadly (Morlino and Tarchi, 2006). Surprisingly enough, our analysis shows that the policy congruence of the Italian parties in their EP groups is strong, when compared with the congruence of the parties of other EU member states. Furthermore, both the Italian parties and the EP groups bind-together the same dimensions of competition, making the Italian and the EP policy space look similar. Although some variation exists, and not all parties fit equally well their EP group, these results show how political parties at the EU level could perform a representative function to a greater extent than it is commonly envisioned. This paper is structured as follows. Section 1 will discuss the democratic deficit in the light of the (failing) process of political representation at the EU level. We argue that when (and if) a strong policy agreement between the national and the EU-level parties exists, the label of democratic deficit might be exaggerated. Section 2 will conceptualize political representation as policy congruence and introduce our case. Section 3 will present the EU-Profiler data and tackle measurement issues. Section 4 will bring forward the empirical analysis. Finally, Section 5 will discuss the results and discuss further implications. 3

1. The Democratic Deficit and Political Representation in the EU The theme of political representation in the EU, concomitantly with the strengthening and the growing institutionalization of the political actors facilitating the representative processes (the EP and the EU level parties) has received increasing attention in the academic literature. The challenge constituted by the shift of sovereignty in a multi-level polity and the formation and institutionalization of a European political space have stimulated numerous theoretical and empirical researches aimed to shed light on the dynamics, peculiarities, limitations, but also potential developments of the processes of political representation at the EU level. From a theoretical perspective, the issues that have mostly been covered relate to the democratic deficit of the EU and to its democratic legitimacy. 2 Scholars did in particular confront each other on how to interpret the concept of political representation at the EU level and on which standards should be adopted in order to make valid inferences on its outcomes and processes. Among the most interesting and relevant debates on this matter is the one centered on the fundamental question on whether the concept of political representation and how it has evolved concomitantly with the evolution of representative democracy at the national level may be applied to the EU multi-level system. It is firstly the EU itself that seems to be aware of the existence of such a democratic challenge. Programmatic statements and institutional reforms aimed at the strengthening of the EU input-democracy (Sharpf, 2003) and at increasing its legitimacy in terms of democratic representativeness have been blossoming over the course of the decades. Since the first direct elections to the EP in 1979, reforms have mainly targeted the EU s elected body the EP and, consequently, the political parties operating internally to it (Kreppel, 2002). Indeed, if the EP has increasingly strengthened its functions and competences, political parties at the European levels moved from performing the roles of passive spectators at the Brussels, Luxembourg and Strasbourg stages (Attinà, 2007, 14) to become central and legally recognized actors in EU politics. 3 The centrality attributed to political parties and to their 2 For a recent review, see Kohler-Koch and Rittberger (eds.), 2007. 3 In legal documents, political parties at the European level were mentioned for the first time in 1992, in art. 138a of the Maastricht Treaty, which later became art. 191 with the Treaty of Nice. Regulation (EU) 2004/2003 set the legal requirement for political parties at the EU level and the conditions concerning their funding. Finally, Regulation 1054/2007 introduced political foudations at the EU level. 4

representative functions result manifestly in the recent disposition entered in the Lisbon Treaty, under which the functioning of the Union shall be founded on representative democracy (Art 10.1), where the actors contributing to forming European political awareness and to expressing the will of citizens of the Union, are the European level political parties (Art. 10.4). It should be noted that scholars have tended to consider the dispositions enshrined in Art. 10 of the Lisbon Treaty as normatively oriented, poorly reflecting the actual functioning of the process of political representation in the EU (Mair and Thomassen, 2010, 22-23). The literature has, in fact, traditionally underlined the weak institutional status of the EP, which not only lacks a number of fundamental powers, such as the right of legislative initiative and instruments to keep the EU executive accountable, but also because of its structural remoteness with respect to the citizens of the Member States that it should represent (Weiler et al., 1995; Follesdal and Hix, 2006). Indeed, EP elections are often conceptualized as second-order elections (Reif and Schmitt 1980), being contested by national political parties, which propose national candidates and structure their electoral competition around national issues and themes. Turnout figures corroborate this definition, being on average about 10-15 per cent lower than those for national elections. Most crucially, the EP elections seem to lack a necessary prerequisite of electoral democracy, characterizing also the most minimal and realist definitions of democracy (Shumpeter 1942, Bobbio 1984), according to which the party system in the electoral arena is reflected by the parliamentary party system, where citizens preferences, aggregated and expressed by party groups, are turned into public policies (Andeweg 1995). However, as Mair and Thomassen (2010) suggest, despite the limitations inherent to the poor democratic functioning of the representative processes in the EU, the instruments of political representation may still function as representative channels. In this light, by focusing on the instruments of political representation, the absence of a direct electoral linkage between the parliamentary and the electoral arenas may not be so disruptive for the EU democracy if one could find some degree of agreement between the national and the EU-level agents of political representation. Indeed, as Thomassen puts it: the effectiveness of a European system of political representation depends on its ability to aggregate and integrate national political agendas and the national cleavage structures at the European level. The major challenge for an effective 5

democratic political system at the European level is to overcome the traditional dividing lines in Europe, the national borders (2009, 215) If the national parties aggregating in the same political party at the EU level share similar policy platforms, and competition among parties is not nationally delimited and defined, then the EU party system has a strong representative potential. European voters might not be yet aware of it, and the EU-level parties might not yet directly compete on different policy platforms, but the conditions for a representative system are in nuce present. Hence, we argue here that the EP groups would function as effective instruments of representation in the EU, even in the absence of a direct electoral connection between them and the European citizens, if two conditions are met: (i) There is policy congruence between the national political parties and the EP groups to which the former belong to; (ii) There is congruence in the relevant dimensions shaping inter-party competition at the national and the EU levels. 2. Political Representation and Congruence: the Case of Italy Since Miller and Stokes seminal article (1963), political representation has often been conceptualized in the literature in terms of policy agreement, correspondence or congruence between citizens preferences and representatives positions. Even though further components in the represented-representative linkage have then been proposed (Eulau and Karps, 1977) and despite a number of epistemological limitations 4 in this approach, testing the presence of a policy congruence between the citizens and the political elites remains an appealing way of operationalizing the concept of political representation. This applies for scholars looking at both the domestic and the EU levels. 5 On the latter, they have normally placed their focus on the comparison between the preferences of the MEPs and those of the European citizens, and a 4 Not least, the relative ignorance argument, claiming that even the more educated and informed among citizens are neither aware on where their interests lie nor are able to formulate specific demands that governments should put into practice (Diamond and Morlino, 2005). 5 On measurements of issue congruence at the domestic level, see inter alia Dalton et al. 2010; on issue congruence at EU level, see Bardi, 1989 and Thomassen and Schmitt 1999. 6

reasonable degree of policy congruence between the party groups in the EP and the national electorates has been found. This proved to be the case, in particular, for those policy issues related to the left-right dimension (Thomassen and Schmitt 1999). In our paper, we follow the literature in operationalizing political representation in terms of policy congruence, but we shift the analytical focus from the traditional nexus between citizens and representatives' preferences to that between the national parties and the EP party groups to which the former are affiliated. The traditional definition of issue congruence the similarity of opinions between party voters and party elites (Dalton, 2008, 227) is here adapted as to capture 'the similarity of opinions between national parties and their EP group'. Our choice to adapt the traditional definition of congruence stems directly from the complexity of the EU political system, where voters' preferences are not directly collected by the EP party groups, but are aggregated at the EU-level only through national parties. Of course, a complete story about the EU system of political representation can only be told taking into account all the three levels, as (i) voters' demands are collected by (ii) national parties which, in turn, coalesce into the (iii) EP party groups. In this paper, we opted to place our focus selectively, and originally, on the congruence among national and party groups in the EP. Keeping citizens' demands aside for the moment, we argue that it is still crucial for political representation in the EU to observe congruence among the national and the EP parties. The literature provides us with a number of reasons casting a bleak light on the feasibility of a EU system of political representation (Schmitt and Thomassen, 1999) because of which it might be untenable to assume that the EP groups are coherent parties of like-minded members. As Fabbrini strongly puts it: differences within each of the groups are so impressive that the label of 'party' is, in most cases, wholly instrumental (2003, 9-10). With the waning of class distinctions and the fading away of the traditional belief-systems, the same meaning of the concept of party family which the EP groups were originally based upon might have lost some of its original appeal (Mair and Mudde, 1997). Furthermore, the appearance of new parties which do not belong to the traditional families, together with the EU expansion to countries with young and less developed party systems as it is the case of Eastern Europe might make the EP groups less cohesive internally, and less different from each other. Hence, if the policy agendas and the patterns of party competition are determined 7

domestically, the EP groups cannot be anything but broad agglomerations of national parties, that is heterogeneous coalitions incapable of expressing clear and coherent positions. Another factor has to do with the institutional incentives for the national parties to join the EP groups, however strong their ideological differences. As it has been argued elsewhere (Bressanelli, 2012), there are several rules in the EP Rules of Procedure (RoP) which make membership in a party group convenient. The most coveted committee chairmanships and rapporteurships, the allocation of financial resources and staff, issues of legitimacy and prestige push national parties to affiliate with the established EP groups and, in particular, drive large parties to seek membership in the large EP groups. Ideological and political compatibility between the national parties and the EP groups shall not be taken for granted on the basis of the extremely loose procedural grounds: according to the RoP (art. 29), political affinities among national members only need to be not explicitly denied in act of constitution of a political group. Indeed, drawing from their first exploration on the policy congruence between national parties and the EP groups, McElroy and Benoit (2007, 20) note that, even if the congruence on a left-right scale between the EP groups and the national parties seem to be strong, there are a number of parties in each group especially, in the People's Party and the conservative UEN whose ideologically fit is very loose. They further recognize that, including a larger number of policy dimensions, a greater divergence between the national and the European levels may result. Besides congruence at the level of parties, we also investigate congruence at the party-system level. If parties at the national and at the EU-level bind together different policy issues if, in other words, it is not possible to reduce policy competition among parties to the same underlying axis, such as the left-right 'super'-dimension then the representative channel is flawed. The existence of a similar dimension of party competition in the EU and in its member states allows voters to recognize the EU politics as simply domestic politics by other means. Most crucially, given that the empirical literature has demonstrated that a bidimensional space of political contestation, defined by a left-right and an integrationsovereignty dimension, is the most effective way to describe the EU policy space (Hix and Lord, 1997; Hix et al. 2007; McElroy, Benoit, 2007), what is really required is that national parties and the EP groups similarly associate these two dimensions 8

together. According to the Hix-Lord model, the left-right and the integration dimension are not linearly related. Their association is, rather, summarized by an inverted U-curve: extremism on left-right corresponds to Euroskepticism, while parties at the system 'core' all support further integration. When a party system does not fit this structure and the UK is probably the foremost example tensions are bound to follow. In this paper, congruence between the national and the European levels are investigated with exclusive reference to Italy. The Italian case appears to be particularly appealing for several reasons. First, the party system has been subject to a continuous process of change as it is indicated by the appearance of new parties, the changes in party labels and their frequent mergers and splits that have been taking place in the two latest decades. In the EU arena, the fluidity of the Italian parties implied frequent switches between the different EP party groups: the LN has seated among the Non-Attached, the Liberals, the Euroskeptics (IND/DEM) and the Conservatives (UEN); Berlusconi's party has been among the Non-Attached, formed a group on its own (Forza Europa) and finally joined the EPP after its affiliation being vetoed for several years by other EPP's member parties 6 ; members of the then PD have been members of the Communist group, the Socialists, the Liberals and the People's Party. In occasion of the 2009 elections party membership in the EP became a hotly debated issue, and in particular for the newly formed and heterogeneous-based PD and SeL. The PD, funded in 2007 as a merger of the left-wing Christian-Democrats (the 'Daisy') and the Democratic Party of the Left (PDS), was split among two different EP groups in the 2004-2009 Parliament (the previous Christian-Democrats seated in the ALDE, whereas the previous members of the PDS seated among the Socialists). The question of transnational affiliation in the EP became very divisive within the PD and it became increasingly more salient as the 2009 elections approached. 7 Similarly complex was the issue of the transnational affiliation of SeL, as clearly manifested by its party symbol for the 2009 EP elections that presented at is bottom the symbols of 6 Not only the Italian People's Party vetoed FI's affiliation to the EPP but other Christian Democratic parties were overtly skeptical about the party's affiliation to the group. See for instance Martens, 2009, 141-147. 7 It should be remarked how the debate within the PD, which ultimately became member of the S&D group, is still ongoing. For instance, one MEP recently contested the choice to affiliate with the S&D and decided to leave the party. 9

three EP groups: the PES, the Greens and the GUE/NGL. 8 In such a new and fluid political landscape, formed by parties whose constitution (or re-constitution) is relatively recent, with the old parties based on the traditional party families of Western Europe being replaced by parties of a more uncertain ideological outlook, policy congruence with the party groups in the EP was difficult to be expected. Overall, it can be argued that Italy is a critical case to test for the policy congruence among the national parties and the EP groups. Moreover, it shed some light on the fit of 'new' parties into the transnational groups, providing interesting insights on the consequences of enlarging the Union beyond the safe boundaries of Western Europe. Finally, it is an interesting case-study per se indicating to what extent the process of representation works in one of the largest members of the EU-27. 3. Data: the EU Profiler The data we use to measure the congruence between the Italian parties and the EP groups derive from the EU Profiler project (www.euprofiler.eu). The EU Profiler developed by an international team of researchers led by the European University Institute 9 is the first Europe-wide Voting Advice Application (VAA) ever launched. Developed for the 2009 elections to the EP, it is essentially a means for voters to gain a view about the European political landscape and discover, through means of an online survey, which party best matches their political preference. Hence, the EU Profiler produces two different data: it assembles information on voters' preferences, including background information; and it gives information on political parties' positions on a variety of key thematic issues. In our paper, we only use the data on the latter. Experts were gathered for each country to collect information of the individual policy statements into more synthetic dimensions is made deductively. In other words, the EU Profiler team has, on the bases of a priori, theoretical considerations, aggregated the policy statements into a smaller number of policy dimensions. The aggregation of preferences per dimensions has been made in two different ways, which will also be used in the empirical analysis that follows. First, 7 thematic dimensions on economic liberalism, restrictive financial policy, law & order, 8 SeL run the 2009 EP elections as an alliance between the Federation of Greens, the Italian Socialist Party, the Democratic Left and Movement for the Left. In late 2009, the Federation of the Greens and the Italian Socialist Party then abandoned the project of SeL. 9 The Italian expert team was composed by Edoardo Bressanelli, Andrea Calderaro, Daniela Piccio and Fulvio Stamati. For an overview, see Gagatek (ed.), 2010. 10

restrictive immigration policy, environmental protection, expanded welfare state and liberal society are visualized in a 'spider'. The dimensions are created by summing up individual policy items and are constructed from a priori considerations on issuebinding by the parties. Second, party positions have been displayed in a 'compass'. This is a simple bidimensional space, where the horizontal axis is a socio-economic left-right scale and the vertical axis represents opposition-support for EU integration. The Italian political parties select ed for the EU Profiler project are the following: Unione Democratici di Centro (UDC), Popolo delle Liberta (PDL), Partito Democratico (PD), Italia dei Valori (IdV), Sinistra e Liberta (SeL), the left wing alliance Partito dei Comunisti Italiani and Rifondazione Comunista (PdCI/RC), the Lega Nord (LN) and the right wing alliance Fiamma Tricolore La Destra MPA. 10 These specific parties (or alliances) were select ed based on their likelihood, according to opinion polls and previous election results, of meeting the electoral threshold and to gain representation in the EP. 11 Based on the positions of the Italian parties and on those of the party groups to which they affiliate in the EP, in the next section we take as empirical measure of congruence the distance between the national member party and its EP group on a given policy dimension. Clearly, the more distant the two parties, the least congruent. Euclidean distances are calculated as: p i p j where p is the position of parties i and j on a given dimension, or in a bidimensional space as: [( p i p j ) 2 + ( q i q j ) 2 ] While we derived the positions of the Italian parties directly from the EU Profiler data, for determining the positions of the EP groups we employed a different procedure. As comprehensive policy documents (such as manifestos) that can be conveniently used for coding the EP party groups positioning are barely present for 10 This electoral alliance has not been subsequently included in the empirical analysis because its members have traditionally seated among the Non-Attached. 11 It is useful to remind that the electoral law was changed before the 2009 EP elections most importantly, a four-percent threshold was added, leading to a strong reduction in the number of Italian parties represented in the EP. 11

both the smaller and younger party groups as well as for the more consolidated groups in the EP (the short, ten-points Euromanifesto presented by the Liberals for the 2009 EP elections well illustrates the problem), 12 the position of the EP groups is computed by aggregating (averaging) the positions of the national member parties. This measure is further weighted per number of seats, in order to get a better approximation of the 'real' position of the EP groups. 13 4. Empirical Analysis 4.1 Congruence at the party level We begin our analysis by comparing the policy positions of the Italian parties and the EP groups on the seven policy dimensions previously discussed: economic liberalism, restrictive financial policy, law & order, restrictive immigration policy, environmental protection, expanded welfare state and liberal society. Data are visualized in spider graphs, with each spike representing a policy dimension and the area that connects the partisan positions on each dimension representing the party location in the policy space. Each dimension ranges from 0 to 100, where 100 means 'complete agreeement' and 0 'complete disagreement' with the issue. While the EU Profiler originally used the spiders to display the match between voters and parties, we use them to visualize the congruence between the national that is: Italian parties and the EP groups. We regard the spiders as a very effective way to visually capture policy congruence. In Figure 1, each spider displays both the position of the EP group and that of its Italian member party. The overlap between the two areas indicates the policy congruence of a national party in its group. Consequently, the more the areas overlap, the stronger the congruence among the two parties. With such analysis, we can both underline areas of tension and disagreement between members and EP groups and attempt a cross-party comparison, by emphasizing which Italian parties are more (or less) congruent with their group. The policy congruence between the Italian parties and the EP groups spans from 7.3 for the UDC to about 20 for both the IDV and the LN. There is a strong variation in 12 The coding of the EU-level parties is due to Edoardo Bressanelli and Wojciech Gagatek. 13 Indeed, both in the extra-parliamentary parties and in the political groups in the parliament the strength of the national member parties in the political organs corresponds to their size. In other words, large parties have more votes and count more. 12

the congruence between the national parties and the groups, which is worth exploring in more depth. To begin with the two largest parties the centre-right PDL and the centre-left PD Table 1 shows that they are respectively 14.1 and 12.4 points distant from their group (the EPP and the S&D). As for Berlusconi's party, the strongest differences with the transnational group are in the areas of immigration policies, environment, law & order and liberal society, while the Democrats diverge most on immigration, environment and welfare state expansion. If overall there seems to be good policy congruence between the Italian parties and the EP groups to which they affiliate, for two parties, the IDV and LN, outcomes seems instead to diverge. Indeed, by comparing the policy positions of the IDV and the ALDE reveals quite strong differences in the areas of environmental protection and welfare state expansion, where Di Pietro s party takes a more left-wing position as compared to the ALDE group. Another clear distinction is on policies to liberalize the economy, with the IDV less supportive than its political group. In general, it can be observed than the ALDE takes a remarkably centre position on all the policy dimensions, while the area occupied by the IDV stretches towards the top-right corner. A poor congruence is also evident between the LN and the EFD. In particular, the LN is tougher than its EP group on immigration policies and on law & order issues, being considerably more conservative on post-materialist issues ('liberal society'). However, the poor fit of the LN should not come as a surprise. The EFD is notoriously very heterogeneous in terms of policy, with its member parties covering a broad spectrum of positions. The broad composition of the group is formally recognized in its Statutes, where it is stated that only administrative questions are decided by majority vote, while political decisions need unanimous agreement of the member parties (art. 2). Finally, moving to the left of the Italian party system, it is worth noticing that the positions of PdCI/RC and the younger SeL are very close, with both parties being about 16 points distant from their EP group. On environmental protection and liberal society, both Italian parties are more supportive than their EP groups, while on the other policy scales with the exception of welfare state expansion where the policy match is great they stand against. [ Figure 1 and Table 1 about here ] 13

Although the 7 spider-dimensions provide a fine-grained picture about the most relevant dimensions of party competition in Europe, often the literature simply describes parties in terms of their positions on two main axes: left-right and anti-pro integration (Hix and Lord, 1997; see also below). In what follows, the congruence between the Italian parties and the EP groups is then measured by their distance in this bidimensional space (Table 2). For the sake of interpretation, it is important to stress that each scale might range from -2 (most left-wing; outright opposition to integration) to 2 (most right-wing; full support to integration). In terms of congruence, using the integration dimension and a socio-economic leftright scale 14 modifies, to some extent, the picture we have drawn before. Thus, the largest distance that is: the least congruence is now between SeL and the GUE/NGL. This is largely due to their different preferences on EU integration: while the former is broadly positive, the latter has a critical attitude towards the EU. This is well illustrated by comparing two key party documents. In the 2009 EP election manifesto of the European Left the transnational organization to which most of the GUE/NGL member parties belong the neoliberal tendencies of the EU were bitterly criticized and a sound 'no' to the Lisbon Treaty was reaffirmed. By contrast, in the 2008 programmatic manifesto of the Democratic Left a founding member of SeL the EU was considered to be 'the main political arena for the Italian Left'. Moreover, the other evident difference is in the (good) congruence between the IDV and the ALDE, whose positions are close in the policy space. Although the IDV is coded as a left-wing party, while the ALDE lies on the right side of the spectrum, the socioeconomic distance between the two is limited. Additionally, both parties are strong supporters of integration. As for the other parties, the analysis confirms the strong congruence of the Italian parties in the EP groups, with the exception of the LN only. Further grist to the mill of this evaluation comes from Table 3.1, where we measured the congruence of the parties in the three largest member countries of the EU Germany, France and the UK. The three cases constitute important benchmarks for congruence evaluation: the EP party system is often said to be modeled on the German one; French parties are in a state of flux and can be roughly compared to Italian ones; and the UK party system is generally considered at odds with 'continental' alignments. As the Table shows, on average, the Italian parties are more congruent with their EP groups than the German, French and British ones. As 14 For details on the construction of the policy scales, see www.euprofiler.eu. 14

expectable, the UK has the worst programmatic fit, averaging 1.54 in terms of distance. France and Germany have lower average distances scoring, respectively, 0.73 and 0.61. Among the individual parties, the best congruence is between the S&D and the German SPD, which are only 0.16 distant, while the worst score is for the EFD and the UKIP (3.03). Hence, with an average distance of only 0.55, Italian parties are more congruent with their EP groups than the German ones, and even the parties showing the least policy congruence with their EP groups (SeL and the LN) seem to be not too worrisome, in comparative perspective. It is remarkable to notice that none of the Italian parties come even close to the Tories-EPP distance (1.71) a notorious case of poor congruence between a national party and the EPP group due to its critical attitude towards EU integration as the least congruent party (SeL) is only 0.84 distant from its transnational group. [ Table 2 about here ] 4.2 Congruence at the party system level Having measured the congruence of the Italian parties with their EP group the first of our conditions we move on to assess the similarity in the structure of inter-party competition the second of our conditions (see Section 1). In what follows, we look first at the correlation matrices 15 among the seven dimensions previously visualized in the spider-graphs; second, we observe the structure of party competition in the bidimensional space defined by left-right and support for integration. By looking at the correlations among the seven dimensions (Table 3), we clearly see how the same policy dimensions are interconnected in both the Italian parties and the EP groups: parties supporting a liberal society (pro-euthanasia, abortion, civil rights) are also supporting stronger green policies and the expansion of welfare provisions. In contrast, liberal society is inversely related to economic liberalization, restrictive financial policy, law & order and restrictive immigration policy. More generally, and both in Italy and in the EU at large, policy dimensions are clustering in two groups: a first one with pro-environment, welfare and social liberal positions, a second with pro-market economy and conservative positions on civil liberties. In order to provide a more stringent test about the underlying structure behind the 15 See Benoit and Laver, 2006, 102-103, for a comparative approach to policy spaces. 15

seven policy dimensions, we run a confirmatory factor analysis. 16 Among the EP groups, a single factor explains a large share of the variance among the seven dimension (72 %). As for Italy, the common variance accounted for by the first extracted factor is even larger (89 %). In both cases, the extracted factor loads negatively with liberal society, environmental protection and welfare state expansion, whereas it loads positively with the other policy dimensions. [ Table 3 about here ] As for any analysis involving data reduction, the key is to provide a meaningful interpretation of the extracted factor(s). What does this one factor providing a parsimonious and efficient description of party competition both in Italy and the EU indicate? By confronting state intervention to laisser-faire in the economy and libertarian to traditional values, our factor seems to capture a general left-right overarching dimension. Indeed, looking back at the multidimensional spiders (Figure 1), while all the Italian parties self-placing themselves in the centre-left (PdCI/RC; SeL; IDV; PD) occupy the upper-left corner, all parties normally placed at the right (UDC; PDL; LN) occupy the area located towards the bottom-right angle. In extreme synthesis, the left-right dimension is able to capture the most important divide between parties in Italy as well as in the EP. In Graph 2, the left-right dimension together with the integration dimension defines the policy space where both the EP groups and the Italian parties are located. The neat correspondence between the Italian and the EP party system finds further support. In the EP party system, the fringe groups (GUE/NGL and EFD) are the most Euroskeptic, while groups located towards the centre of the EP party system are prointegration. When we shift our focus to the Italian party system, the inverted-u curve is still the most effective way to describe party positions. 17 On the extreme left, the PdCI/RC has a very cold attitude towards integration; on the right, the LN takes a Euroskeptic position. In between these two 'extremes', all other parties are Europhiles. It is, indeed, remarkable to observe how similar the inverted U-curve looks like in the EP and the Italian party system. By choosing a party in their domestic party system, 16 A principal factor analysis with varimax (orthogonal) rotation has been run separately for the EP and Italy. 17 The left-right dimension and the integration dimension are orthogonal to each other. The correlations among the two dimensions are: -0.043 (EP); 0.060 (Italy), both not statistically significant. 16

therefore, the Italian voters knowledgeable about it or, most likely, not are de facto choosing for EU alternatives. Although parties fit their EP group less well than others, the patterns of competition among the Italian parties clearly mimic the party system they are embedded in at the EU level. [ Graph 2 about here ] 5. Conclusions How much similarity exists between the political profile of the national parties and that of the EP groups to which they are affiliate? And how much similarity there is between the national and the EU polity levels in terms of the relevant dimensions of inter-party competition? We considered the analysis of those two conditions as crucial for shedding light on the processes of political representation in the EU and for testing whether the EP groups may indeed function as effective instruments of representation in the EU, as formally stated by the Lisbon Treaty. Drawing on the original dataset of the EU Profiler project, in this paper we specifically focus on the case of Italy, by empirically measuring the degree of congruence between the Italian political parties and the EP groups. Results showed that the Italian and the EP groups positions with respect to the key policy dimensions characterizing the European political space were overall quite close to each other. With a few exceptions, most noticeably SeL and the LN, we found policy congruence between the party actors at the two polity levels. As a means of comparison, we showed how congruence between the Italian parties and the EP groups is greater than that in other member countries such as France, Germany and the UK. Furthermore, a similar pattern of congruence was observed for the space of inter-party competition at the national and the EU levels. Such a result, we argued, was far from expectable. Scholars frequently emphasize the mismatch existing between the national member parties and the EP groups. This applies particularly for the case of Italy: in the light of the country s recent reconfiguration of the party system, still considered as being in flux because of its recurrent party splits and mergers, and because of the creation or re-orientation of political parties characterized by loose ideological underpinnings which escape the traditional party families classification that instead define (in principle, at least) the 17

EP groups, policy congruence could hardly be assumed. Having observed congruence at the party and party system levels, we do not intend to make the claim that no democratic deficit exists in the EU. On the contrary, we argued that the processes of political representation at the EU level lack a number of fundamental preconditions for being assimilated to a democratic polity. Most importantly, we stressed upon the absence of a direct electoral linkage. Nonetheless, what we do argue is that voters may, although indirectly or even inadvertently, find their national party preferences being reflected in the EP groups, when the positions of the national parties that they vote for and those of the EP groups are similar one to the other, and when the same policy issues are bound together at the two polity levels. In fact, one could argue that political representation of the European citizens may take place even despite the poor democratic functioning of the representative processes in the EU (see also Mair, Thomassen, 2010). The question is, of course, whether this same match would hold both across time and across a wider number of cases. The analysis presented in this paper faces at least one important limitation. Namely, that political representation, as a process, necessarily involves considering the input side i.e.: the citizens demands. As we argued before, a complete story about the EU system of political representation can only be told taking into account all the three levels, starting from the voters' demands, which in turn are collected by national parties, which in turn coalesce into the EP groups. By indicating congruence between the positions of the institutional actors, it is one side of the coin only. This aspect is even more crucial considering the poor performance of political parties as representative agents at the very national level (see, for all, Mair 1998), and is corroborated by opinion polls systematically pointing to the (national) political parties inability as adequate channels of political representation. Integrating the demand side citizens preferences and testing the degree of congruence at the three levels, would therefore undoubtedly enrich the analysis of political representation at the EU level. This said, there is a growing area of research focusing on differences in policy competition at the national and the EU levels. As McElroy and Benoit argued, more in depth analyses into the two overlapping arenas of European party competition is one of the most promising areas for emerging research (2007, 18-19). In the context of the growing attention by the EU itself on the representative actors and processes, such research would indeed be important to verify the normative oriented dispositions 18

brought forward in the light of the actual representative dynamics taking place between the national and the EU levels. 19

References Andeweg R. B. (1995) The Reshaping of National Party Systems. West European Politics, 18, 3, pp. 58-78. Attinà, F. (2007) EU democratization. From patchy democracy to party democracy, Paper presented at the ECSA-USA 5 th Biennial International Conference, Seattle. Bardi, L., (1989) Il Parlamento della Comunità Europea: legittimità e riforma, Il Mulino: Bologna. Bardi, L., E. Bressanelli, E. Calossi, W. Gagatek, P. Mair and E. Pizzimenti (2010) How to Create a Transnational Party System, Florence: EUDO Report, 2. Bressanelli, E., (2012) 'National Parties and Group Membership in the European Parliament. Ideology or Pragmatism?', Journal of European Public Policy, forthcoming. Bobbio, N. (1984) The Future Of Democracy: A Defence Of The Rules Of The Game. Polity Press: Cambridge. Cedroni L. and D. Garzia (eds.) (2010) Voting advice applications in Europe: the state of the art, ScriptaWeb: Naples. Dalton R. J. (2008) Citizens politics: public opinion and political parties in advanced industrial democracies, Washington D.C: CQ Press. Dalton R. J., D. M. Farrell and I. McAllister (2011) 'The dynamics of political representation', in M. Rosema, B. Denters and K. Aarts (eds.), How Democracy Works, Amsterdam: Pallas Publications Amsterdam University Press, pp. 21-38. Diamond, L. and L. Morlino (2005) Assessing the Quality of Democracy. Baltimore: The Johns Hopkins University Press. Eulau H. and P. D. Karps (1977) The Puzzle of Representation: Specifying Components of Responsiveness, Legislative Studies Quarterly, 2, pp. 233-54. EU Profiler (2009) General Description and Methodology'. Follesdal A. and S. Hix (2006) Why there is a Democratic Deficit in the EU, Journal of Common Market Studies, 44., 3, pp. 533 62. Gagatek, W., (ed.) (2010) The 2009 elections to the European Parliament. Country Results, Florence: European University Institute. Hix, S. and C. Lord (1997) Political Parties in the European Union. London: Macmillan. Hix S., A. G. Noury, G. Roland (2007) Democratic Politics in the European Parliament, Cambridge, UK, New York: Cambridge University Press. 20

Kohler-Koch, B. and B. Rittberger (2007) Charting Crowded Territory: Debating the Democratic Legitimacy of the European Union, in B. Kohler-Koch and B. Rittberger (eds.), Debating the democratic legitimacy of the European Union. Lanham, MD: Rowman & Littlefield, pp. 1-29. Kreppel, A. (2002) The European Parliament and Supranational Party System: a study in institutional development, New York: Cambridge University Press. Mair, P. (1998) Representation and Participation in the Changing World of Party Politics, European Review, 6, 2, pp. 161-74. Mair P. and C. Mudde (1998) 'The party family and its study', Annual Review of Political Science, 1, pp. 211-229. Mair P. and J. J. Thomassen (2010) Political representation and government in the European Union, Journal of European Public Policy, 17, 1, pp. 20-35. Martens, W. (2009) Europe: I struggle, I overcome, Brussels: Springer. McElroy, G. and K. Benoit (2007) Party groups and policy positions in the European Parliament. Party Politics, 13, 1, 5, pp. 5-28. Miller, W. E. and D. E. Stokes (1963) Constituency Influence in Congress, The American Political Science Review, Vol. 57, No. 1, pp. 45-56. Morlino L. and M. Tarchi (2006) Introduzione, in L. Morlino and M. Tarchi (eds.), Partiti e caso italiano, pp. 7-15. Bologna: Il Mulino. Reif K. H. and H. Schmitt (1980) 'Nine-second order national elections a conceptual framework for the analysis of European election results', European Journal of Political Research, 8, 1, pp. 3-44. Sharpf, F. W. (2003) Problem-Solving Effectiveness and Democratic Accountability in the EU, MPIfG Working Paper 03/01. Schmitt, H. and J. Thomassen (2009) The EU party system after European enlargement, Journal of European integration, 31, 5, pp. 569-587. Schmitt, H., and J. J. Thomassen eds. (1999) Political Representation and Legitimacy in the European Union. Oxford: Oxford University Press. Schumpeter, J.A. [1942] (1962) Capitalism, Socialism and Democracy. London: George Allen & Unwin. Thomassen, J. J. (2009) 'Introduction', in J. J. Thomassen (ed.), The Legitimacy of the European Union after Enlargement. Oxford: Oxford University Press. Thomassen, J. J. and H. Schmitt (1999) 'Issue Congruence', in Schmitt, H., and J. J. Thomassen (eds.) Political Representation and Legitimacy in the European Union. Oxford: Oxford University Press, pp. 186-208. 21