Chapter 6 Democracy, Parties and Elections Democracy: Choosing Parties, Leaders and Policies The 'Democratic Deficit' Debate Parties: Competition and Organization Elections: European Parliament Elections and EU Referendums Conclusion: Towards Democratic Politics in the EU? This chapter looks at how the two central processes of 'democratic politics' -party competition and elections - operate in the EU. At the domestic level in Europe, parties and elections mostly operate hand in hand in a 'competitive party government' model of democracy. There is also an emerging party system at the European level, European-wide elections to the European Parliament are held every five years, competitive and cohesive political parties exist in the European Parliament, and in 2014 there was the first open contest for the Commission Presidency. But genuine democratic politics in the EU remains some way off. Democracy: Choosing Parties, Leaders and Policies Elections are the central mechanism of representative democracy and operate in two interlinked ways (King, 1981), First, elections allow voters to choose who governs. Voters choose between rival candidates for public office, and the winning candidate or party or coalition of parties gets to form the government. Elections consequently allow voters to 'throw the scoundrels out' if they flout their electoral promises, prove incompetent or become less popular than a rival set ol candidates or parties. Second, elections allow voters to choose policies. Political parties and political leaders present citizens with rival policy proposals, and the parties or leaders who emerge as the winners from the democratic process then have a mandate to implement their policies. In this conception, democracy only 146 The Political System of the European Union exists if there is a choice between competing politicians and policies, and if there is a reasonable chance of alternation in government (Schumpeter, 1943; Downs, 1957; Przeworski et a!, 2000), In most democratic systems, competition over policies and for public office is combined in a single model of Competitive party government' (cf. Weber, 1946 [1919]; McDonald and Budge, 2005). In this model the leader of the party or coalition of parties that wins the election becomes the head of the executive (the prime minister), and the party or coalition acts cohesively in the legislative arena to implement the policies presented in the election manifesto in the previous election, or in the post-election coalition agreement. Meanwhile, the losing opposition parties try to demonstrate the failings of the politicians in government and to try to influence policies through the legislative process. In this model voters exercise an indirect influence on policy outcomes. Alternative models allow voters to exercise a direct choice over office holders or the policy agenda. First, in the presidential model, voters directly elect the head of the executive. Parties play an important role in selecting presidential candidates, running their campaigns and forming post-election coalitions to support the legislative agenda of a president, but presidential elections tend to be dominated by the policies and personality of the individual candidates rather than by the manifestos of the parties. In addition, in a presidential system, presidential elections are separate from parliamentary elections - and presidents usually cannot dissolve parliaments - which creates a dispersion of power rather than a concentration of power in the hands of the government. Second, through referendums voters can choose policies directly. Again, parties play a role, advocating one side in a referendum, and the (un)popularity of the parties on each side of the debate (particularly those in government) will affect the way citizens vote in the referendum. So, which model is right for the EU? Following the logic of the competitive party government model, in the past most commentators on the EU's so-called 'democratic deficit' argued that (1) the European Parliament should be directly elected, and (2) it should be given greater powers in the EU legislative process and the selection of the EU executive (i.e. the Commission). 'Direct elections' to the European Parliament were introduced in 1979 and have been held at five-yearly intervals ever since. And, in a series of treaty reforms the European Parliament has been given a greater power in the EU legislative process vis-a-vis the Council and the Commission and also now in the election of the Commission. In purely procedural terms, the election of the European Parliament and the new legislative and executive-appointment powers of the European Parliament suggest that the democratic deficit in the EU has been overcome. In substantive r Democracy, Parties and Elections 147 terms, however, the EU lacks genuine 'democratic polities'. Democratic politics in the EU would require: i a genuine contest between political parties and leaders for control of the policy agenda at the European level and for the main political offices in the EU (such as the president of the Commission) • choices by voters in European Parliament elections on the basis of these rival policy platforms or candidates for political office; and • that the winning electoral choices are translated into legislative and executive action at the European level via cohesive political parties. If this pattern of behaviour does not exist in the EU then the EU cannot be considered to be a democratic polity in the way political science normally understands this concept. The 'Democratic Deficit' Debate Articles on the so-called democratic deficit in the EU started to be published in academic journals in the mid- to late 1980s. More widespread discussion of this issue in the media soon followed, in response to the collapse in support for the EU in the early 1990s (see Chapter 5) There is no single definition of the democratic deficit in the EU. However, Joseph Weiler and his colleagues described what they called a 'standard academics version' of the democratic deficit in 1995: a set of widely used arguments by practitioners and commentators (Weiler et al„ 1995). Adding some elements to Weiler's original definition, an updated 'standard version' of the democratic deficit involves five main sets of claims (cf. Follesdal and Hix, 2006; Hix, 2008; Nanou and Dorussen, 2013; Hobolt and Tilley, 2014; Schneider, 2019): • Increased executive power. EU decisions are made primarily by executive actors: national ministers in the Council and the Commissioners. As powers have shifted to these actors at the European level there has been a reduction of the power of national parliaments, and, in turn, a reduction in the ability of voters to hold their governments to account for actions they take in many major areas of public policy that are now decided at the European level. • The European Parliament is too disconnected from citizens. Increases in the powers of the European Parliament have not sufficiently compensated for the loss of national parliamentary control, since citizens are not as well connected to their MEPs as they are to their national parliamentarians. 148 The Political System of the European Union • There are no 'European' elections. Citizens are not able to vote on EU policies, except in periodic referendums on EU membership or treaty reforms. National elections are fought on domestic rather than European issues, and parties collude to keep the issue of Europe off the domestic agenda. And, European Parliament elections are not about Europe either, as national parties and the media treat them as mid-term polls on the performance of national governments and parties • The EU is too distant. Citizens cannot understand the EU. The Commission is neither a government nor a bureaucracy, and is appointed through an obscure procedure rather than elected directly by the people or indirectly by a parliament. The Council remains a largely secretive legislature. The European Parliament is impenetrable because of the multilingual nature of the debates. And, the EU policy process is highly technocratic, which prevents actors and citizens from easily identifying political preferences. • The EU locks in unpopular policies. As a result of all these factors, the EU adopts policies that are not supported by a majority of citizens in many (or even most) member states, such as a deregulatory framework for the single market, austerity policies in EMU and liberal migration policies, and these policies constrain the choices national governments and voters can make. However, these arguments are not universally accepted. For example, Giandomenico Majone and Andrew Moravcsik - two of the biggest names in the study of the EU - contested many of these claims. Giandomenico Majone argued that the EU is essentially a 'regulatory state', which should not produce redistributive outcomes, and so does not require democratic legitimation in the same way that domestic 'welfare states' do (Majone, 1993, 1996, 1998, 2002; see also Chapter 8). Regulatory policies in the EU, Majone contends, should mainly aim to correct market failures, and so should be designed to benefit everyone (produce 'pareto-efficienf policy outcomes) rather than to make some people better off at the expense of others (which is the explicit aim of redistributive policies). As a result, Majone argued that EU policy-making, which primarily involves the regulation of goods, services, capital and labour in the single market, should be isolated from the standard processes of majoritarian democratic politics. During the creation of the single market in the 1980s and early 1990s, for example, because the Commission was independent from direct political control, it was able to develop and propose social, economic and environmental standards that were in the interests of all of Europe, rather than in the interests of a particular member state or political majority. If 'normal democratic politics' had set the basic framework for the internal market, Democracy, Parties and Elections 149 instead of an internal market that balanced deregulation of national markets with common environmental and social standards, the EU would have had either an overtly neo-liberal framework, with few social and environmental standards, or an overtly social democratic framework, with high costs for business and consumers. In either of these situations there would have been widespread opposition to the internal market programme from the supporters of the losing side. Majone consequently argued that a more politicized EU during this period would have undermined rather than reinforced the legitimacy of the project. Andrew Moravcsik (2002, 2008) went further, presenting a critique of all the main claims in the standard version of the democratic deficit. Against the argument that power has been centralized in the executive, Moravcsik pointed out that national governments are the most directly accountable politicians in Europe Against the critique that the executives are beyond the control of representative institutions, he argued that the most significant institutional development in the EU has been the increased power of the European Parliament. Moravcsik also argued that EU policy-making was more transparent than most domestic systems of government, that the EU bureaucrats in the Commission and the Council are forced to listen to numerous European and national societal interests, that there is extensive judicial review of EU actions by both the European Court of Justice and national courts, and that the European Parliament and national parliaments have increasing powers of scrutiny that they are not afraid to use. Finally, against the argument that EU policies are far from the views of average voter, Moravcsik argued that the EU's system of checks-and-balances ensures that a broad consensus is required for any policies. Moravcsik's arguments fit with the intergovernmental theory of EU politics (see Chapter 1). In this theory, because the member state governments run the EU and the Commission is simply an agent of these governments, there are no unintended consequences of intergovernmental bargains. As a result, since the governments are accountable to their electorates, and there is no gap between the preferences of the elected governments and final EU policy outcomes, then the EU is not undemocratic because policy outcomes from the EU broadly reflect the preferences of European voters (cf. Crombez, 2003). Regarding Majone's argument, it might have been reasonable to isolate the EU from democratic politics during the market-creating stage of the EU's development, to ensure that the basic political and economic architecture was in the interest of virtually all European citizens. Also, where there were potential redistributive consequences of market integration in Europe, unanimous agreement between the governments ensured that budgetary polices were used as side-payments to compensate potential losers from the process: such as the 150 The Political System of the European Union doubling of EU regional spending in the late 1980s (see Chapter 9). However, most of what the EU decides now has significant distributional consequences, with identifiable winners and losers, and with voters and interest groups and parties with very different preferences regarding what the EU should do. For example, should services markets be opened up to more competition from service providers from other member states? Should taxpayers in northern Europe temporarily transfer wealth to citizens in southern Europe to stabilize the Eurozone? Should the costs of reducing carbon emissions be borne by producers or consumers? Should the EU force countries to accept more refugees as part of a common 'burden sharing' scheme? Should the EU sign a free trade deal with the United States that could allow more lightly regulated goods and services access the single market? And so on. These questions, and many others, are explicitly 'political', with citizens, interest groups, political parties and governments with deeply opposing views on different sides of the debate However, even if EU policies have distributional consequences, and even if these consequences are becoming increasingly salient, the EU might still be 'responsive' to citizens' views on these issues. In a democratic polity, even if citizens are divided on an issue, if the average citizen's opinion changes on a particular issue, then policy outputs on the issue should also change to reflect citizens' preferences. Similarly, if the public wants more (less) action in a particular policy area, then a democratic polity should respond by producing more (less) action in a particular area - this is known as the 'thermostatic model' of policy responsiveness (Wlezien, 1995; Soroka and Wlezien, 2010). In the EU context, even if policy-making power may have shifted to the European level in many areas, as the salience of EU issues has grown in national elections and in domestic party competition, EU policies could still respond to changes in citizens' preferences. The key actors delivering this responsiveness are the member state governments, and there is growing evidence that member state governments do in fact respond to changes in citizens' views on EU issues. Christina Schneider (2019) shows, for example, that member state governments adopt policy positions that are congruent with their voters' preferences, that member state governments defend these positions when bargaining in the Council, and they work hard in the EU institutions to deliver on the policy promises they make to their electorates (cf. Hagemann et al., 2017; Wratil, 2018; Hobolt and Wratil, 2020). So, although there might be executive dominance in EU policy-making, if the national executives reflect domestic voters' preferences or domestic political coalition compromises, then, on average, EU-level policy coalitions should be broadly representative of citizens' political preferences (Franchino and Wratil, 2019). 1 Democracy, Parties and Elections 151 Nevertheless, there are problems with relying on policy responsiveness of national governments as the main mechanism for ensuring a democratically accountable EU. For example, as Kleine and Minaudier (2019) argue, as governments become more responsive during national elections, this can lead to policy gridlock at the EU level, particularly during elections in the larger member states. Above all, though, member state governments are only concerned with their own voters, and aggregating the median voter or the domestic political coalition in each member state is not necessarily the same as representing the interests of EU citizens as a whole. Yes, the EU may be accountable to national electorates, but that is not the same as European voters collectively making a choice about who should govern them at the European level, or about the direction of the EU policy agenda. National governments may have mandates from their separate electorates, but in the absence of genuinely European-wide electoral contest, there is no sense of a collective European-wide mandate. What political scientists call 'losers' consent'- where citizens accept that they did not 'win' a particular election or on a particular policy issue - is critical for the survival of political systems (Anderson et al., 2005). Relying on national governments as the main mechanism for securing political and policy responsiveness is a weak foundation on which to build losers' consent in the EU. Imagine telling someone who is angry that an EU decision or set of policies went against their strongly held views that they should be happy that their national government effectively represented them in EU decision-making, when their government was either outvoted, or was forced to compromise, or traded-off their policy position on the issue they care about in return for securing an outcome they wanted on another political issue. In other words, even if one accepts most of the arguments of Majone and Moravcsik (for example about the need to isolate some regulatory policies from democratic competition, or the openness of the EU policy process) and most of the evidence of Schneider and others (about the responsiveness of national governments when acting at the EU level), the EU polity cannot be considered to be 'democratic' unless there is some form of European-wide democratic political contestation over who governs at the European level and over the direction of the EU policy agenda. Hence, the rest of this chapter focuses on the electoral and political processes in EU politics, and whether they come anywhere close to this expectation. Parties: Competition and Organization EU politics is party politics. This may not seem obvious to a casual observer of the EU. But, on closer inspection, party organizations, labels, ideologies, policies, coalitions and interests take centre stage. All politicians at the domestic and 152 The Political System of the European Union European levels are party politicians who owe their current positions and future careers to the electoral success and policy positions of 'their' parties. Parties are the main actors in domestic elections, European Parliament elections and referendums. They are the main organs connecting governments to parliaments and parliaments to voters. As a result, they provide vital links between the national and EU arenas and between the EU institutions themselves. To understand how EU politics works, then, we need to understand how parties compete and organize in the EU. Party Positions in EU Politics No single party family dominates EU politics. The social democrats are the main force on the centre-left in all member states, and are members of the Socialists and Democrats (S&D) group in the European Parliament. The main force on the centre-right is the European People's Party (EPP), which is composed of Christian democratic parties as well as the more pro-European conservatives. Between these forces are the centrist and liberal parties in Renew Europe, which used to be called the Alliance of Liberals and Democrats for Europe. Outside these three main political forces are several forces to the left of the social democrats: the greens and left-wing regionalist parties, who sit together in the Greens/European Free Alliance (G/EFA), as well as radical left parties who sit in the European United Left/Nordic Green Left (EUL/NGL) group. Finally, there are several forces to the right of the EPP: a range of nationalist/Eurosceptic conservative parties (such as the Polish Law and Justice Party) who sit in the European Conservatives and Reformists (ECR) group; right-wing populists parties (such as the French National Front and Dutch Party for Freedom) who sit in the Identity and Democracy (ID) group; plus a number of extreme right parties (such as JOBBIK in Hungary) who are not members of any group in the European Parliament (and hence sit as 'non-attached' members). Figure 6.1 shows the approximate location of the national parties and these political forces on the two main dimensions of EU politics: the left-right, and the EU integration dimension (cf. Hix and Lord, 1997; Ray, 1999; Bakker et al„ 2020). The locations suggest several natural alliances in EU politics. The clearest natural allies are the EPP and Renew, who are both moderately centre-right and pro-European. The S&D also have a natural ally on the centre-left in G/EFA, who have become more pro-European over the past two decades. On issues related to the speed and depth of European integration, however, radical left and radical right parties find themselves closer to each other, in terms of their 1 Democracy, Parties and Elections u w " u uGus ! .1 $ U EUL/NGL U SG L E CCL S > E L G C U U u -, c e c XECR I c A 1 c C a a A' /C i AA Left-Right Figure 6.1 National party policy positions in EU politics Note: This figure plots the positions of the national member parties of the political groups in the European Parliament in 2019 on the two main dimensions in EU politics, using expert surveys of where national parties are located on the Left-Right and Anti-/Pro-Europe dimensions. The average positions of the national parties in each political group are marked by an X. Key: ECR European Conservatives and Reformists (C) EPP European People's Party (E) EUL/NGL European United Left/Nordic Green Left (U) G/EFA Greens/European Free Alliance (G) ID Identity and Democracy (A) NA non-attached (independents) (I) Renew Renew Europe (L) S&D Socialist arid Democrats (S) Source: Calculated from data in Bakker et al. (2020). shared anti-European positions, than they are to the mainstream political forces in Europe, who are all pro-European. The locations also show that some forces are more naturally 'cohesive' in policy terms than others. While most party families are relatively homogenous on the left-right dimension, they are more internally divided on Europe integration. Party families are historically defined in relation to the left-right dimension and not on the question of European integration. As a result, parties in the same political 154 The Political System of the European Union family have similar policy preferences on left-right issues but often have different national-based preferences towards the EU. This is because in some countries EU policy outcomes tend to change domestic policies in a leftwards direction, by introducing new social regulations (as in the UK. for example), whereas in other countries EU policies tend to move policies rightwards, by liberalization domestic markets (as in France, for example). Also, because EU policy outcomes tend to be rather centrist, as a result of the checks-and-balances in the system, the further a party is from the centre, either on the radical left or the radical right, the more likely a party is to be anti-European (cf. Aspinwall, 2002; Marks, Wilson and Ray, 2002; McElroy and Benoit, 2012; Bakker et al„ 2020). The dominance of the centrist pro-EU alliance, which has often prevented the mainstream parties from competing on European issues in domestic electoral campaigns, means that anti-European positions amongst the electorate have not been articulated by parties in many national party systems (Mair, 2000; Gabel and Hix. 2003; cf. De Vries, 2007). On average, there is a high 'congruence' between the preferences of voters on European integration and the EU policy positions of the national parties they vote for (Carrubba, 2001; van der Eijk et al., 2001). However, in most national electorates there is an increasing number of anti-European voters, who choose to support parties on the extremes who have more anti-European positions, and particularly in electoral arenas - such as European Parliament elections and referendums on European issues - which provide opportunities for voters to protest against the centrist pro-European 'cartel'. And, as voters have become more Eurosceptic, the congruence between voters' and mainstream parties' positions on Europe has declined (cf. Lefkofridi and Katsanidou, 2013; Wiiliams and Spoon, 2015; Dalton, 2017), particularly for lower income and lower educated voters (Walczak and van der Brug, 2013). Parties in the EU institutions There are two party organizational structures at the European level. The most prominent of these are the political groups in the European Parliament. These groups were first formed in the Assembly of the European Coal and Steel Community in 1953; the precursor of the modern European Parliament (Hix et al., 2003). Since then, the political groups in the European Parliament have evolved into highly developed organizations, with their own budgets, leadership structures, administrative staff, rules of procedure, committees and working groups. The second organizational structure consists of the transnational party federations outside the European Parliament, which were formed in the run-up to the first direct elections in the mid-1970s. The Confederation of Socialist Parties of I Democracy, Parties and Elections I55 the EC was the first to be established, in April 1974, followed by the Federation of Liberal and Democratic Parties of the EC, in March 1976, and the European People's Party of Christian democratic parties in April 1976. Despite their names, these were rather loose organizations, did not have highly sophisticated organizations at the European level, and did not have a clear and coherent policy orientation, despite biannual European party conferences. Nevertheless, at the instigation of the three secretaries-general of the party federations introduced a new 'party article' in the Maastricht Treaty, which stated that: Political parties at the European level are important as a factor for integration within the Union. They contribute to forming a European awareness and to expressing the will of the citizens of the Union. Following this article, the party federations established new and more coherent organizations. The Party of European Socialists was launched in November 1992, the EPP adopted a set of new statutes in November 1992. a new European Federation of Green Parties was set up in June 1993 and the European Liberal, Democratic and Reform Party was established in December 1993. These party organizations strengthened their links with the political groups in the European Parliament, and also with the representatives of these party federations in the Commission, Council and European Council, mainly via quarterly party federation leaders' summits. Instead of being umbrella organizations for fighting European Parliament elections, these federatidns have gradually developed into genuine 'Euro-parties', albeit with limited power over their constituent national member parties or the political groups in the European Parliament. Figure 6.2 shows the party political make-up of the EU institutions in January 2010 and January 2020 (following the UK's departure from the EU). At the end of the 1990s, the centre-left had dominated EU politics, with the socialist group the largest group in the European Parliament (until 1999) and centre-left parties running thirteen of the then fifteen member state governments. By 2010 there had been a considerable shift rightwards, with the EPP dominating the Commission, the Council and the European Council, although reliant on building coalitions in the European Parliament. At the start of 2020. the EPP was still the largest political force: with 175 of the 705 MEPs, 11 of the 27 Commissioners in the von der Leyen Commission, 11 of the EU heads of government in the European Council, and 34 per cent of the votes (by population) in the EU Council. Nevertheless, the European Parliament had become more fragmented, with the rise of the populist right, and liberal/centrist politicians were more represented in all the EU institutions. In other words, whereas the EPP had dominated in 2010, the political make-up of the EU institutions was more balanced in 2020, with the liberals playing a pivotal role. The Political System of the European Union European Parliament European Council (no. of Heads of Government) ^^^^^^^^^^^^^^ European Commission EU Council (voting/population weights) Key: ■ Radical Lett (EUI./NGL) ■ GreenaCG/EFA) ■ Social Democrats (SAD) ■ Liberals (ALUE/ Renew) ■ Centre Right (EPP) Conservatives (ECRi Populist Riyhl (EFD/ID) Non-attached Figure 6.2 Party make-up of the EU institutions in 2010 and 2020 Note: The inner pie-charls show the make-ups in January 2010 and the outer pie-charts show the make-ups ,n January 2020 (foitowing the UK's exit). The EU Council figure shows the po cent of votes under the QMV system in 2010 and the per cent of EU populaL in 2020 If democratic politics works in the EU, this centre-right dominance of the EU's legislative and executive institutions should produce a concomitant shift in the EU policy agenda towards more market liberalization and deregulation (ct. Manow et al., 2008; Warntjen et al., 2008). However, this translation from party strengths to policy outputs requires party actors in the same political family to co-operate within and across the three institutions, which is not always the case in EU politics. There is some evidence that Commissioners and governments are influenced by their party affiliations as well as their nationality when acting in the EU policy process (Hagemann and Hoyland, 2008: Tallberg and Johansson, 2008; Wonka, 2008). However, the transnational party linkages are still weak in the Commission and the Council. In the European Parliament, in contrast, political behaviour is dominated by supranational party organization and competition. Democracy, Parties and Elections Research on rollcall (recorded) votes in the European Parliament has revealed that voting along supranational party lines has increased while voting along national lines has decreased (Attina, 1990; Hixet al., 2005, 2007; Hixand Noury, 2009). This growing voting cohesion of the political groups cannot be explained by increasing internal ideological homogeneity of the groups. The national member parties of the main political groups have quite diverse policy positions, and the internal heterogeneity of the main groups has grown as the groups have expanded their membership as a result of EU enlargement. For example, the British and Czech conservatives left the EPP in 2009, as they were critical of the pro-EU integration position of the EPP, and formed ECR with several other conservative and populist right parties. But, despite these departures, the EPP remains a broad coalition of relatively market interventionist parties from the Benelux, France, Germany and Austria, more free market and socially liberal parties from Scandinavia, and some socially conservative parties from Eastern Europe and Iberia. The growing party-based voting in the European Parliament is more a result of the increased powers of the European Parliament than the internal ideological coherence of the groups (esp. Raunio. 1997; Kreppel, 2002a; Hix et al., 2007). As the European Parliament has gained more influence over policy outcomes from the EU, the stakes have raised for the MEPs. More at stake means more incentives to strengthen the division of labour inside the political groups, to try to win votes and shift outcomes in a particular policy direction. The political groups have consequently strengthened their leaderships and established mechanism for rewarding and punishing 'backbenchers' who toe the group line: via the system of party whips (who monitor group voting) and committee assignments (party co-ordinators in committees and the allocation of the rapporteurships). However, the political groups are rarely able to prevent particular national delegations of MEPs from defecting in key votes (e.g. Gabel and Hix, 2002; Hix, 2002b; Ringe, 2005; Lindberg, 2008). For example, between 1999 and 2009, the British conservative MEPs, who belonged to the EPP group in that period, voted differently from the majority of their European political group about 30 per cent of the time. The political groups may control rewards inside the parliament, but national parties still control the selection of candidates for European Parliament elections and hence whether MEPs will be reselected. As a result, if MEPs have conflicting voting instructions from their political group and their national party, they will usually follow the instructions of their national party (Ringe, 2009). MEPs from member states with candidate-centred electoral systems (such as open-list proportional representation or single transferable vote) tend to be freer from their national parties (Hix, 2004). Nevertheless, the growing cohesion of the political The Political System of the European Union groups, despite the power of national parties over their MEPs, suggests that national parties rarely instruct their MEPs to vote differently from their groups, either because they share the same preferences as their group on most issues on the EU agenda, or because issues in the European Parliament are not salient enough. Also, repeated interactions in the European Parliament provide incentives for national parties to support their political group, since voting against the group rarely changes the policy outcome yet risks the withdrawal of important positions for national parties, such as a committee chair or a key rapporteurship (esp. Hix et al, 2009). Competition between the European political groups has grown (Hix et al., 2003; Hix et al., 2005). Unlike in national parliaments in Europe, there is no governing majority in the European Parliament. The separation of executive and legislative powers in the EU means that coalitions in the European Parliament are built issue by issue - rather like in the US Congress (cf. Hix et al,, 2006). On most policy issues this means that the groups in the European Parliament form coalitions along the left-right dimension: with a centre-left majority, of the S&D, ALDE, G/EFA and EUL/NGL, forming a winning majority on environmental regulation, gender equality, and civil liberties and justice and home affairs; and a centre-right majority, of EPP, ALDE and ECR forming a winning majority on single market liberalization and global trade agreements. As a result of these shifting coalitions, the European Parliament, and hence the EU more broadly, tends to adopt policies that are 'liberal' in terms of both liberalization of the single market and liberalization of the free movement of people and gender, racial and sexual equality. Not surprisingly, citizens who agree with these policies (with centrist political preferences) tend to support the EU, while citizens who disagree with these policies (either on the radical left, who oppose the EU's free market economic policies, or on the radical right, who oppose the EU's liberal social policies) tend to oppose the EU and the European Parliament. Interestingly, though, during the financial crisis, when the EU adopted a package of legislation on the regulation of banks as well as rules restricting national macro-economic polices, the standard structure of coalition formation and left-right political competition in the European Parliament shifted (Otjes and van der Veer 2016; Blumenau and Lauderdale, 2018). On these critical issues, the political battle in the European Parliament shifted from the left-right dimension to a pro-/ anti-Europe dimension: with a 'pro-European' bloc of EPP most of the national parties in S&D, and ALDE (who favoured moves to promote deeper political and economic integration within the Eurozone) lining up against a 'Eurosceptic bloc of ECR, G/EFA, EUL/NGL. EFD, who opposed the EU imposing fiscal austerity and any new restrictions on national macro-economic autonomy. Democracy, Parties and Elections Left-Right Figure 6.3 Three votes in the 2014-19 European Parliament Wore: The figure shows the location of the 'cutting line' in three votes in the 2014-19 European Parliament. The two-dimensional locations of the individual MEPs are calculated using Optimal Classification. See the note to Figure 3.2 for the political group labels. Source: Rollcall voting data from VoteWatch.eu. Top: Fossil Fuel Subsidies, 17 January 2018 Middle: Tackling the abuse of fixed-term contracts, 31 May 2018. Bottom: Asylum-seeker burden-sharing, 17 September 2015 160 The Political System of the European Union Some of these voting patterns are illustrated in Figure 6.3, which shows the splits in the votes on three policy issues in the 2014-19 parliament (EP8). In the key vote on whether the EU should seek to end fossil fuel subsidies, the Parliament divided along left-right lines, with the left supporting ending support for the coal industry on full pay, and the right opposing, and with AIDE voting with the left. In the key vote on the regulation of fixed-term contracts, the Parliament again divided along left-right lines, with the right supporting a more free market approach to the use of these contracts and the left supporting more restrictive rules. This time, ALDE voted with the right. In contrast, in the key vote on asylum burden-sharing, the Parliament split along more pro-/anti-European lines, with centrist parties, as well as the two more leftist groups (G/EFA and EUL/NGL) supporting a common EU burden-sharing framework and the more Eurosceptic groups supporting national sovereignty to decide who to accept as asylum seekers. In sum, parties at the European level are underdeveloped compared to national parties, and national parties remain the key actors in the European-level party organizations. However the political groups in the European Parliament do behave like national parliamentary parties: MEPs vote along political lines, and coalition formation in the European Parliament is mainly driven by left-right policy positions Also, the transnational party federations are beginning to serve as arenas for linking key party actors at the European level: the national party leaders, the European political group leaders and the European Commissioners. But, do voters choose between rival policy agendas for European action in European electoral contests? Elections: European Parliament Elections and EU Referendums There are two types of EU-related electoral contest: European Parliament elections, and referendums on EU treaty reforms or other major changes to the EU. Whereas European Parliament elections are held throughout the EU according to a fixed schedule (every five years), referendums on EU-related issues have only been held sporadically in some member states European Parliament Elections: National or European Contests? Direct elections to the European Parliament were first held in June 1979. and since then have been held every five years. In the run-up to the first elections many scholars expected that elections to the European Parliament would provide a new legitimacy for the EU (e.g. Marquand, 1978; Pridham and Pridham, 1979), r Democracy, Parties and Elections 161 According to Walter Hallstein (1972: 74), a former president of the European Commission: Such a campaign would force those entitled to vote to look at and examine the questions and the various options on which the European Parliament would have to decide in the monlhs and years ahead. It would give candidates who emerged victorious from such a campaign a truly European mandate from their electors; and it would encourage the emergence of truly European political parties. After nine sets of European Parliament elections it is clear that the reality is still some way short of this optimistic prediction. This is because European Parliament elections are in practice 'second-order national contests' rather than 'European contests'. Karlheinz Reif and Hermann Schmitt (1980) came up with this concept having observed that first elections in 1979 were mainly about national political issues, national parties and national government office. The main goal of national parties throughout is to win and retain national government office. Elections that decide who holds national executive office are therefore 'first-order contests', and parties consequently treat all other elections - including European Parliament elections, regional and local elections, second chamber elections and elections to choose a ceremonial head of state -as beauty contests fought in the shadow of the ongoing first-order election contest. The second-order and national character of European Parliament elections has two effects. First, because second-order elections are less important than first-order elections, there is less incentive for people to vote in European Parliament elections, and hence there is a lower turnout in these elections than in national elections. Turnout in European Parliament elections has always been approximately 20 per cent lower than in national parliamentary elections. Participation in European Parliament elections has also declined over time: from 63 per cent in 1979 to 43 per cent in 2014, although in 2019 turnout rise to 51 per cent. Mark Franklin (2001) demonstrates, however, that most of the decline in turnout in European Parliament elections can be explained by falling turnout in national elections, the fact that voting is compulsory in fewer member states, that national elections are held on the same day as European Parliament elections in a declining proportion of member states, and that the EU has enlarged to countries that have generally had lower levels of turnout in national elections. Once these factors have been controlled for, turnout in European Parliament elections has in fact been relatively stable. Second, because European Parliament elections are mainly about the performance of national governments, many people vote differently in a European Parliament election than they would if it were a national election. Some citizens use 162 The Political System of the European Union the elections to vote sincerely rather than strategically, by voting for a (small) party that is closest to their preferences and which has a chance of winning a seat under the electoral system used in a European Parliament election. Meanwhile, other people use the elections to express their dissatisfaction with the party or parties in government, by voting for an opposition party. Either way, the consequence of such 'vote switching' is that large governing parties lose votes in European Parliament elections, while opposition parties and small parties gain votes (Reif, 1984; van der Eijk and Franklin, 1996: Marsh, 1998; Hix and Marsh, 2007). The second-order national elections model consequently suggests that lower turnout in European Parliament elections and the outcome of these elections have nothing to do with 'Europe'. This is not entirely true. Focusing on the motivations for voting, Blondel etal. (1997,1998) and Stockemer (2011) find evidence that voters' attitudes towards Europe are relevant in explaining participation in European Parliament elections. At the individual level, citizens who are more anti-European are less likely to vote in European Parliament elections, and lower levels of support at the aggregate level are also correlated with lower turnout at the national level, both across country and across time. There is also some evidence that voters' attitudes towards European integration influence which parties they support in European Parliament elections (Weber. 2007; Tilley et al., 2008; Hobolt et a)., 2009). For example, De Vries and Hobolt (2012) and Hong (2015) find that citizens with anti-European attitudes are more likely to switch from supporting a governing party in a national election to a smaller party in a European Parliament election. This consequently affects which parties win and lose. Some political families have systematically done better than others in these elections, irrespective of their size or governing status, which cannot be explained by the standard second-order national elections model (Hix and Marsh, 2011). In particular anti-European parties did particularly well in Denmark in 1979,1994 and 2014, in France in 1999 and 2004, and in the UK in 2004, 2009 and 2014. Green parties also performed well in the 1989 and 2019 elections. And, socialist parties, both in government and opposition, did particularly badly in the 2009 and 2014 elections compared to the other main political families. In addition, the standard second-order model does not fit the post-Communist states in Central and Eastern Europe as well as the Old15 member states in Western Europe: in that there has not been such a general anti-government swing in European Parliament elections in these new member states (Koepke and Ringe, 2006). Also, Ferrara and Weishaupt (2004) demonstrate that while the general position of parties on the issue of Europe may not have an effect on whether they gain or lose votes in European Parliament elections, whether a party is divided on the question of Europe does seem to matter, in that parties whose Democracy, Parties and Elections 163 elites are divided over European integration have performed worse than parties who have a united pro- or anti-European position, other things being equal. At the level of individual M EPs and candidates in European Parliament elections, the elections work differently in different member states. For a start, there is not a uniform electoral system in European Parliament elections, as Table 6.1 shows (esp. Farrell and Scully. 2005). Since the 1999 elections, when the UK replaced its single-member constituencies with regional list-proportional representation (PR), all member states have used some form of PR. However, there are two major differences in the type of PR that are used. First, many member states have some form of preferential voting - either 'open' list-PR or single-transferable-vote - which allows citizens to choose between politicians from the same political party, rather than simply to choose between political parties. Second, while most member states elect their MEPs in one single national constituency, several member states elect their MEPs in several sub-national constituencies. Comparative political science research has found that preferential voting and the number of politicians elected in each constituency have a significant effect on how candidates campaign in elections and the relationship between politicians and their parties (e.g. Carey and Shugart, 1995; Shugart et al., 2005). In particular, preferential voting systems encourage politicians to campaign directly to voters, to persuade voters to support them rather than other politicians from the same party. Also, larger constituencies under preferential voting systems produce more competition between candidates and hence more direct campaigning and candidate recognition by the citizens. There is some evidence that there are similar effects in European Parliament elections (Farrell and Scully, 2007). For example, in the 2009 European Elections Study, citizens in countries that had small multi-member constituencies and preferential voting systems - as in Estonia and Ireland - were more likely to be contacted by parties or candidates during the election campaign than citizens in countries where non-preferential voting systems are used - such as Poland, Greece, Portugal or Germany (Hix and Hagemann, 2009). Similarly, in the 2014 elections, candidates in member states with preferential voting systems or small constituencies were more active on social media than candidates in member states with party-list systems or larger constituencies (Obholzer and Daniel, 2016). Finally, there was a new aspect to the European Parliament election campaign in 2014. The Lisbon Treaty, which entered into force in December 2009, changed Article 17, on the appointment of the Commission President. Paragraph 7 of the new article said: Taking into account the elections to the European Parliament and after having held the appropriate consultations, the European Council, acting by a qualified Table 6.1 Electoral systems used in the 2019 European Parliament elections Member state No. of Electoral Preferential No. oi Average Formal Voting days MEPs system type voting constituencies constituency size electoral threshold Austria 18 List PR Yes 1 18 4% Sunday 26 May Belgium 21 List PR Yes 3 7 Sunday 26 May Bulgaria 17 List PR Yes 1 17 5.7% Sunday 26 May Croatia ' List PR Yes 1 11 5% Sunday 26 May Cyprus 6 List PR Yes I 5 1.8% Sunday 26 May Czech Republic 21 List PR Yes 21 5% Friday 24 May & Sat. 25 May Denmark 13 List PR Yes 1 13 - Sunday 26 May Estonia 6 Lisl PR Yes 1 6 Sunday 26 May Finland 13 List PR Yes 13 Sunday 26 May France 74 List PR No 1 74 5% Sunday 26 May Germany 96 List PR No (16) 96 Sunday 26 May Greece ?! List PR Yes 1 21 3% Sunday 26 May Hungary 21 List PR No 1 21 5% Sunday 26 May Ireland 11 STV Yes 4 Friday 24 May Italy 73 isl PR Yes (5) 73 4% Sunday 26 May Latvia i> Lisl PR Yes 1 8 5% Saturday 25 May Lithuania 11 List PR Yes 1 11 5% Sunday 26 May Luxembourg 6 List PR Yes 1 6 - Sunday 26 May Malta 6 STV Yes I 6 - Saturday 25 May Netherlands 26 List PR Yes 1 26 3.8% Thursday 23 May Poland 51 List PR Yes (13) 51 5% Sunday 26 May Portugal 21 List PR No I 21 • Sunday 26 May Romania 32 List PR No 1 32 5% Sunday 26 May Slovakia 13 List PR Yes 1 13 5% Saturday 25 May Slovenia 8 List PR Yes 1 8 4% Sunday 26 May Spain 5- List PR No 1 54 - Sunday 26 May Sweden 20 List PR Yes I 20 4% Sunday 26 May UK-Great Britain 70 List PR No ll C - Thursday 23 May UK-N.Ireland 3 STV Yes I 3 - Thursday 23 May Total 751 43 17 Note: List PR = party-list proportional representation; STV = single-transferablo-vote. In Germany, Italy and Poland candidates are presented in constituencies but votes are counted and seats allocated in a single national constituency. 166 The Political System of the European Union majority, shall propose to the European Parliament a candidate for President of the Commission. This candidate shall be elected by the European Parliament by a majority of its component members. If he does not obtain the required majority, the European Council, acting by a qualified majority, shall within one month propose a new candidate who shall be elected by the European Parliament following the same procedure. The two key new phrases, indicated in italics, meant that the outcome of the European Parliament elections would now have a direct impact on the choice of the Commission President. In response to this treaty change, ahead of the 2014 elections the main European political parties put forward their candidates for this position, who became known as Spitzenkandldaten (lead candidates): Jean-Claude Juncker for EPP, Martin Schulz for S&D, Guy Verhofstadt for ALDE, Alexis Tsipras for EUL/NGL, and Ska Keller and Jose Bove for the Greens (Hobolt, 2014). The candidates held several live debates, which were broadcast on television and on the internet throughout the EU, and the two main lead candidates' (Juncker and Schulz) went head to head in two separate debates on French and Austrian/German TV. Did this innovation create a link between voters and executive power at the European level for the first time? On the one hand, Schmitt et al. (2015) found that there was a connection between the campaign visits of the candidates to member states and the level of turnout in these member states, as well as a connection at the individual level between the awareness of the candidates and a propensity to vote in the elections. However, on the other hand, the Spitzenkandldaten had no clearly identifiable effect on the overall 2014 election results, as the second-order effect remained strong: large parties lost votes while small parties gained votes, opposition parties did better than parties in government, and anti-EU parties did well, particularly in the UK, France and Denmark. The European parties again chose Spitzenkandldaten ahead of the 2019 elections. This time there was a more open contest within the EPP, with Manfred Weber (EPP leader in the European Parliament) eventually defeating Alexander Stubb (former Finnish prime minister) at the EPP Congress in November 2018. The S&D chose Frans Timmermans, the Greens picked Ska Keller again, ECR chose Jan Zahradil and the European Left chose Nico Cue. French President Macron expressed his scepticism towards the process, on the grounds that the EPP would almost certainly win, which prevented ALDE from choosing a single candidate. Instead, they backed a 'team' for the European Parliament elections, although Margrethe Vestager emerged in the TV debates as ALDE's de facto lead candidate. The TV debates seemed to be more widely covered than in 2014, Democracy, Parties and Elections 167 and several lead candidates featured prominently in national campaigns, such as Weber in Germany and Timmermans in the Netherlands. And, in the election itself, the Dutch Labour Party (PvdA) received a boost as a result of Timmermans' candidacy: almost doubling its vote-share between 2014 and 2019. Nevertheless, the 2019 Spitzenkandldaten process ended in failure. The EPP emerged as the largest political group again, and backed Weber. However, S&D and Renew refused to support him because of his hesitancy to eject Orban's Fidesz from the group after the democratic backsliding of Hungary. In return, the EPP refused to support Timmermans or Vestager. Macron also refused to support Weber, as the EPP had rejected Macron's proposal for 'transnational lists', which he had put forward as part of his plan to reinvigorate EU democracy. Macron and Merkel were willing to support Timmermans, but he was blocked by the 'Visegrád four', who opposed his support for refugee burden-sharing, and EPP prime ministers refused to support Timmermans or Vestager as neither vjas from the 'winning group' in the election. In the end, a compromise was then struck between the Heads of Government, for Ursula von der Leyen (a German Christian democrat/EPP) to be put forward by the Heads of Government as the (first woman) Commission President, with Timmermans and Vestager as key vice presidents in the new Commission. The deal commanded a large majority in the European Parliament, supported by EPP, S&D and Renew. Although von der Leyen was an EPP politician, and so reflected in spirit the outcome of the election, it was widely suggested that 'the Spitzenkandldaten process is dead and the intergovernmental counter revolution has succeeded' (de Wilde, 2020: 16). The key test, though, is what will happen in the next elections in 2024. If the European parties pick more prominent politicians with widespread political support ahead of the elections, it might be more difficult for the Heads of Government to reject the process again. On the other hand, prominent politicians may be reluctant to put their names forward if there is an expectation that the candidates will be ignored in the post-election horse-trading. In short, despite the increased powers of the European Parliament in both the legislative arena (vis-á-vis the Council) and the executive arena (vis-á-vis the selection and accountability of the Commission), European Parliament elections are still not mainly about the performance and the policy agendas of parties or politicians at the European level, which would need to be the case for the EU to be seen by many citizens as truly 'democratic'. The Spitzenkandldaten process was an attempt to connect the election campaigns, and voters' choices to EU executive power, but has so far hardly made much of a difference. Other reforms are being discussed, such as electing a certain number of MEPs on transnational lists, as Macron and others have proposed. 168 The Political System of the European Union But, again, most member states opposed these measures, and there does not seem to be much support amongst European citizens for any further reforms in this direction Referendums on EU Membership and Treaty Reforms Do referendums on European issues do better than European Parliament elections in terms of enabling citizens to express their preferences on European issues? Figure 6.4 shows the fifty-six referendums that have taken place on a variety of EU-related issues in twenty-five different countries or territories. Almost two-thirds of these referendums have been in just three states: Denmark (where eight have been held), Ireland (with nine) and Switzerland (with ten). While the pro-EU side has won in most of these referendums, it has not done so well recently. For example, of the seventeen referendums that were won by the anti-EU side, two were before 1990, four were in the 1990s and eleven have been since 2000. In fact, if one only counts referendums amongst EU member states (either on treaty reforms, membership of the Euro/EMU, or adopting a particular EU initiative), then since 2000 the pro-EU side has only won in a minority of cases (five out of fourteen), and two of these victories were re-run referendums in Ireland, after Irish voters had initially rejected a treaty reform and were then asked to vote again on a slightly amended version. In fact, the pro-EU side has lost four out of five referendums since 2012: one in Switzerland in 2014 on restricting free movement of people from the EU (which was then over-ridden by a referendum in 2020 accepting the agreement with the EU); in Denmark in 2015 on whether Denmark should fully join the justice and home affairs pillar of the Treaties; one in the Netherlands on the EU-Ukraine Association Agreement; and, most devastatingly of all from the EU's perspective, one in the UK in 2016 on whether the UK should remain or leave the EU. But, are these referendums really about the EU? Mark Franklin and his co-authors argue that most EU referendums work much like European Parliament elections: in that the European issues in an EU referendum are of such low salience that the popularity of the government of the day dominates in the campaign and the outcome (Franklin, Marsh and McLaren, 1994; Franklin, Marsh and Wlezien, 1994: Franklin et al., I995; Franklin, 2002). For example, looking at the Danish, Irish and French referendums on the Maastricht Treaty in 1992 and 1993, the main determinant of individual votes was support for or opposition to the party or parties in government at the time of the referendums. In the case of the 1992 French referendum, polls showed overwhelming support for the EU: but voters used the referendum to punish the unpopular Mitterrand presidency. In Democracy, Parties and Elections a o OCH FH , Gin: ml' C LTO n FR*SE O Agreement „Ith EU LT#S1 o •ft OCH • mi001 •n SB -l—I—I—I—r~ i s S Figure 6.4 Referendums on European integration. 1972-2020 Sources: Compiled from data in Hug (2002), European Elections Database (httPY/'www.nsd.uib. no/european_election_database.