Representation Deficits and Surpluses in EU Policy-making

Abstract Representation and democracy are not always complementary. Sometimes the one undermines the other. Too much democracy can create a representation deficit, as occurs when majorities oppress or neglect minorities. However, the opposite can also arise. The over representation of different groups can undermine the processes whereby representatives are authorised by and accountable to those they are supposed to serve. The EU offers multiple channels of representation. In some respects, this multiplicity reflects the diversity of the peoples, individuals and interests represented within the EU. Yet in overcoming a potential representation deficit in EU policy-making, this arrangement leads to a representation surplus and creates a democratic deficit.

The European Union (EU) has been diagnosed as suffering from a democratic deficit ever since the ratification of the Maastricht Treaty. Opinions abound as to the origins and severity of this malaise. Some commentators have ascribed this deficit to a failure to institutionalise representative democracy fully at the EU level by rendering the competences of the European Parliament (EP) analogous to those of national parliaments (NPs) (Lodge 1994); others have identified the deficit in the lack of a European demos, which depending on their perspective either could or could not be overcome (Weiler 1995); while still others have located the deficit at the domestic level and bemoaned the incursions of the integration process on the powers of national representative institutions (Neunreither 1994). This introduction adds to this debate by arguing that much of this discussion reflects and interacts with a more specific concern with what we shall call the EU's representation deficits, of which the democratic deficit is both a cause and a symptom. As we shall see, democracy and representation do not easily coexist and indeed can work to the detriment of each other.
Although most contemporary democracies describe themselves as representative democracies, as does the EU in the Lisbon Treaty, the conceptual and empirical compatibility of these two terms is far from selfevident (Bobbio 1987). Historically, thinkers such as Jean-Jacques Rousseau considered representation as intrinsically elitist because it displaces the direct involvement of citizens in collective decision-making (Rousseau 1968, 141). From this perspective, representation creates a democratic deficit by its very nature. The eventual linkage of representation with democracy was a product of contingent developments, most notably the gradual emergence of territorially confined nation states and competitive party systems (Hobson 2008;Rokkan 1974). The French and American Revolutions offered the intellectual and political context in which democratic government came to be envisioned and institutionalised as representative government (Urbinati 2004). What we today call the standard model of representative democracy materialised from these circumstances. According to this model, the democratic credentials of representatives could be assured by having credible electoral mechanisms for their authorisation and accountability. So conceived, representative democracy was viewed not simply as a second best adaption of direct democracy to the size and complexity of modern polities, but also a superior form of democracy whereby popular sovereignty was linked to the benefits of expertise associated with a division of labour between rulers and ruled that allowed for the professionalisation of politicians.
Nevertheless, doubts have arisen about the effectiveness of this linkage from the beginning. On the one hand, the electoral mechanism has been criticised for offering inadequate control over representatives. They have been regarded as forming a class apart, concerned solely with the pursuit of power or the perquisites of office (Mair 2006), and susceptible to capture by special interests with disproportionate influence or lobbying power, such as the financial sector, with whom they often have close personal links (Parry 1969). This first kind of representation deficit we term the elitist deficit. It arises when popular views are under-represented because elites either only represent their own interests or those of small but influential groups, such as bankers. On the other hand, the electoral mechanism has been attacked for exercising too much control over representatives. They become encouraged to strike populist attitudes that pander to popular prejudices and become reluctant to sacrifice short term gains for long term advantages (Brittan 1975). This second kind of representation deficit we term the populist deficit. It arises when unpopular views -of minorities or of expertise -are under-represented.
In different ways, which we explore below, both shortcomings can be regarded as involving a failure of representatives to represent the interests of the political community as a whole, thereby producing a representation deficit. Yet, in the first case it derives from a democratic deficit, in which majority opinion fails to get adequately represented, in the second from what could be called a democratic surplus, in which majority opinion is overly represented. The traditional solutions have been to supplement or constrain the electoral process respectively in ways that depart from the standard model for the democratic authorisation and accountability of representatives. However, as we shall see, each of these solutions threatens the balance between representation and democracy found in the standard model, and in addressing their respective problems risks introducing those associated with the other representation deficit.
Views differ as to how far the EU itself manifests these problems (e.g. Lord 2008;Moravcsik 2008), and the degree to which the integration process is their source (Scharpf 2009) or their solution (Keohane, Macedo and Moravcsik 2009) within the member states (MS). Criticisms of the operation of the standard model of representative democracy have been even harsher with regard to the EU than they are in the case of the MS (Follesdal and Hix 2006). The predominance of domestic issues and parties in European elections, the EU's multilevel nature, the fact that legislative proposals come from the Commission and are subject to co-decision by the Council of Ministers and the EP, among other features, have all resulted in voters often finding it difficult to know what European policies an individual politician or party stands for, let alone to identify which decisions, if any, they can be held responsible for. As a result, a potential representation deficit of the first, elitist, kind has seemed especially compelling given the wide authority enjoyed by national executives and the remote, complex and hard to follow decision-making processes of the EU. One solution would be to move the EU towards the standard model by electing a European government via direct elections to the EP, thereby abolishing the European Council and the Commission in its current form. However, though long advocated by dedicated European Federalists, only a few regard it as likely in the short term (Duff 2012;Habermas 2012), and many doubt its likelihood, workability or desirability in the long term (Mair and Thomassen 2010b).
Therefore, while the standard model within the EU has been strengthened as the EP has progressively gained more powers, attention among academics and EU policy-makers has turned increasingly to alternative channels and forms of representation that might overcome the distinctive representation deficits of the Council, the Commission and the EP (Krö ger and Friedrich 2013a; Mair and Thomassen 2010a). Three have proved particularly important, addressing each of these bodies respectively. First, NPs have been formally recognised in the Lisbon Treaty as having a role within the representative structure of EU decision-making and have been given new powers to police the integration process. In this channel, the aim has been to strengthen the standard model of representative democracy with regard to one of its historical weaknesses -the discretion of executives with regard to foreign, and specifically EU, affairs. As a result, parliaments have become more proactive in ensuring their governments accurately represent the views of parliaments and citizens when negotiating in Brussels. Second, the Commission has increasingly defended its legitimacy on the grounds that it is precisely the comparative isolation of the EU's decision-making from democratic pressures that helps it overcome a representation deficit of the second, populist, kind by offering a model of expert, efficient, and equitable good governance. Yet, to demonstrate its continued representativeness of the public as a whole it has increasingly consulted with civil society organisations (CSOs) which have a particular interest in the policies that fall within the EU's competence. Though rarely authorised or accountable in the standard sense, CSOs have been seen as offering a more participatory way of gauging the views of citizens on a given issue and even for mobilising a transnational or pan-European constituency around it. Finally, the EU has begun to experiment with more direct forms of citizen involvement, such as the citizens' initiatives and referenda, that address a representation deficit of the first, elitist, kind through bypassing the use of representatives altogether and allowing citizens to represent themselves.
This special issue explores examples of all three of these innovations. While some regard them as a second best to the development of the standard model of representative democracy at the EU level, and even as further intensifying the democratic deficit by muddying the lines of authorisation and accountability (Bellamy 2010;Krö ger and Friedrich 2013b), others view them as part of a further transformation of democracy (Dahl 1989, ch. 6;Bohman 2007). They mirror the use of similar stratagems within the MS to overcome the two representation deficits produced by the standard model by various forms of direct democracy, on the one hand, and non-majoritarian regulators, courts and independent central banks, on the other. They claim these developments may have weakened the influence of the electoral process upon the choice of EU representatives, as per the standard model, but they have done so in ways that have enhanced the representativeness of the system as a whole (Moravcsik 2008). Indeed, they regard the EU's multiple and complex channels of representation as improving the representativeness of not only the EU but also the MS. They help overcome the capture of national governments by powerful domestic groups, enhance the avenues for representation and the resources of expertise available to excluded minorities, and give a voice to the citizens of other countries affected by the domestic decisions of a MS other than their own who would otherwise have no way of influencing them at all (Keohane, Macedo and Moravcsik 2009). In other words, the apparent democratic deficit of the EU is simply a function of its tackling the representation deficits within both the EU and the MS.
As a preliminary to examining the different views sketched above, we shall describe further the two representation deficits below, along with their relationship to persisting tensions between representation and democracy. We shall then apply this perspective to the analysis of the EU's system of representation. As we shall see, the EU provides numerous and potentially conflicting channels of representation, some of which seek to address a democratic deficit and others a democratic surplus -potential or actualwhile attempting to enhance the representative nature of the system as a whole. This evaluation -like that of the articles in this special issue -is necessarily normative as well as empirical, since the various standards of democracy and representation employed reflect normative criteria. Our claim will be that there is often a trade-off between the representativeness of a political system and its democratic responsiveness, and that to some degree the EU is moving towards a representation surplus that needs rebalancing in the direction of democracy.

The Two Representation Deficits of Representative Democracy
As we noted, the tension between representation and democracy has been overcome in standard models of representative democracy via an electoral process for authorising representatives and holding them to account. This process has rested in its turn on certain assumptions about political equality and the nature of the demos. To respect democratic norms, a system of representation needs to offer a form of government that citizens can perceive to be both 'of' and 'for' the people. The process whereby representatives are chosen must be such that citizens can see their interests and views are treated with equal concern and respect, and the decisions their representatives make must promote policies that taken overall can be defended as promoting the equal advancement of those views and interests (Dahl 1989, ch. 3;Christiano 2008). Therefore, as Article 9 of the Lisbon Treaty recognises, political equality provides the meta-norm of representative democracy (Lord and Pollak 2010, 126).
Both ontological and epistemological conditions concerning the nature of the demos ground this emphasis on political equality. Ontologically, the basis for giving everyone an equal say in the collective decision-making process stems from the assumption that all have a roughly equal stake in those decisions, at least as a package if not in each and every one (Christiano 2008, 78-88). Epistemologically, this criterion assumes that each citizen is the best judge of his or her own interests, if not necessarily of how they might be best met then of those potential representatives most likely to find ways to meet them and of their success or failure in doing so (Christiano 2008, 88-100). If both conditions regarding the demos hold, the decision-makers and decisions preferred by the majority of the people on the basis of an equal vote will be those most likely to promote their collective interest (Dahl 1989, chs. 8 and 10;Christiano 2008, ch. 6).
A representation deficit arises when either the process fails to offer voters a choice of representatives that reflects equal concern and respect for their various views and interests, or their representatives fail to promote policies that advance them on an equal basis. That may occur due to a lack of responsiveness in the system for selecting representatives, or because either the ontological or the epistemological conditions do not apply. Lack of responsiveness usually involves the tyranny of the minority resulting from the capture of the system by a particularly influential group -either directly, through their presence within the political class, or indirectly, through their power over it. It corresponds to the first, elitist, kind of representative deficit noted above. This deficit is invoked when critics complain -rightly or wrongly -that democratic politicians of all parties form an elite that serves its own interests and those of other members of the elite rather than the people more generally (Parry 1969). By contrast, the absence of the ontological and epistemological conditions is most associated with the tyranny of the majority and populism, and the consequent neglect of either intense or discrete minorities, or the pursuit of short-term benefits that prove costly in the long-term. Thus, the second, populist, kind of representation deficit has two forms. Representation deficit of type 2A results from the absence of the ontological condition. It occurs when citizens have unequal stakes or apply quite different criteria to assessing policies that reflect cultural differences. In such circumstances, the risk that populist majorities may exert majority tyranny over consistent minorities is greatly increased. This concern is often voiced by minority national groups in multinational systems, such as the French Canadians in Quebec or the Scots in the UK. Representation deficit of type 2B results from the absence of the epistemological condition. It occurs when myopia or misinformation or some other cognitive failing means that citizens may press representatives to act against the general long term interest, or be susceptible to their doing so for electoral advantage. A standard example is the temptation for governments to buy voters through a mix of tax cuts and public spending increases that ultimately proves unsustainable (Brittan 1975).
As we observed above, neither of these problems is new, with radical and liberal critics of representative democracy addressing the first and second of these representation deficits respectively from the eighteenth century onwards. By and large, their solutions have involved proposals that seek to enhance the representativeness of the system via alternatives to the democratic processes of authorisation and accountability favoured by the standard model of representative democracy. For example, solutions to the first, elite, deficit have involved introducing forms of direct democracyfrom citizens' juries to referenda. Here the cause of the representation deficit results from the mechanisms of democratic control being insufficient to ensure representatives reflect the views and concerns of the people as opposed to a small section of it or themselves. In the terminology of representation theory, they are insufficiently descriptively representative to stand and act 'as' those they represent. Given this failing results from a deficit in the electoral process, the natural solution is to improve the possibility for citizens to represent themselves directly. By contrast, solutions to the second, populist, deficits of discrete minorities (2A) and myopic, misinformed or self-interested majorities (2B) have involved respectively giving extra weight to minority views by giving them special representation rights, such as extra seats or a guaranteed representation in the executive, (2A) and the removal of certain decisions from democratic control to be decided by non-elected representatives of the public interest, such as courts or experts consulting with relevant groups (2B). In these cases, improving the representativeness of the system involves addressing the relevant ontological and epistemological problems in the composition of the demos by constraining the democratic process in various ways so as to give more than an equal voice to otherwise neglected minority or independent views. Such moves seek to strengthen the capacity of representatives to stand and act 'for' the best interests of the community as a whole.
Worryingly, dealing with the first deficit may exacerbate the second and vice versa. Enhancing direct democracy to overcome the elitism involved in the first kind of representation deficit may give rise to the very populism and prejudice associated with the second kind of representation deficit. Yet, constraining the democratic process and enhancing the authority of representatives risks in its turn the elitism of the first kind of representative deficit in which representatives ignore the interests and views of voters when they should take them into account. Meanwhile, the more the standard model of representative democracy is supplemented or constrained by these two sorts of measures the harder it will be to either authorise representatives or hold them to account. Decisions will emerge from complex, often non-transparent, processes as the product of many hands and different groups of people. The result may be a system that is highly representative in many ways but not particularly democratic in the sense of operating under the equal public control of the people.
The more diverse and complex a society, and the greater its openness to other societies, the more likely these two deficits will occur. A demos assumes a people bound by cooperative relations of mutual advantage that are conducive to political equality (Miller 2009). In these circumstances, the ontological and epistemological conditions for democratic decisionmaking will apply. Though the cross-cutting cleavages of pluralist societies can foster these conditions, the more segmented and differentiated societies become the less that will be the case (Dahl 1989, ch. 5). Interests and the criteria for evaluating them will diverge in ways that prove not only conflicting but incommensurable as well. As a result, pressures will grow to devolve democratic decision-making downwards and often directly to distinct groups to overcome the first, elitist, representation deficit, and to make those decisions that need to be made collectively by a mix of consensual means, in which all groups have an equal say regardless of size, and unelected neutral third parties chosen on the basis of supposedly objective criteria, to address each of the second, populist, representation deficits. Most of the MS have experienced these difficulties and attempted some combination of the remedies noted above to address them -though some, such as Belgium and increasingly the UK, more than others, such as France.
That a political entity as large and diverse as the EU should have been characterised as suffering from both kinds of representation deficit is unsurprising, therefore. On the one hand, it has been viewed as precipitating the elitist kind of deficit. National executives and members of the EP (MEPs) have been charged with a failure to respond to citizens' views, with the Commission -according to such critics -at best an unaccountable bureaucracy concerned to maximise its own power, at worst captured by business and other sinister interests. On the other hand, it has also been regarded as prone to the two populist deficits. Particularly during the Euro crisis, EU decisions have been attacked for reflecting the views of the larger, wealthier states and paying insufficient attention to the social, eco-nomic and cultural diversity of the MS (deficit 2A), or being too abstruse and technical to be responsibly tackled by elected politicians who must pander to the partial and self-interested concerns of domestic electorates (deficit 2B). Consequently, to an even greater extent than in the MS, the EU has evolved a particularly complex system of representation that supplements an incomplete and partial institutionalisation of the standard model of representation within the EP, with numerous other channels and forms of representation. These range from the strengthening of NPs and the introduction of citizen initiatives and referenda, that seek to overcome the first elitist deficit, to the use of consensus or highly qualified majority decision making in the Council and supermajorities in the EP to tackle the 2A version of the second populist deficit, and the Commission's widespread use of non-majoritarian bodies and consultation with CSOs to address the 2B version.
Some regard this complex system as diminishing the prospect offered by the standard model of knowing what actor is responsible for making which decision and holding them to account for it, both within the MS and the EU. You cannot 'throw the scoundrels out' (Weiler 2012, 829). It becomes harder to judge how representative decisions are -to what degree they do reflect and favour political equality. Another perspective believes either the EU meets -or could meet -the ontological and epistemological conditions of a demos (Hix 2008;Habermas 2012). By contrast, others believe the standard model is inappropriate and unnecessary within the EU. They regard the complexity as an essential new development to meet circumstances where there is no single demos but rather multiple demoiboth national and transnational. Their representation can only be guaranteed through multiple channels of representation that do not employ the formal democratic processes of the standard model (Weiler, Haltern and Meyer 1995;Moravcsik 2008;Fabbrini 2010). It is to the description and evaluation of the EU's system of representation that we now turn, noting along the way the contribution of the various articles in this special issue to this debate.

The EU's System of Representation
Representation is a central concept in the way in which the EU understands its democratic legitimacy. Title II on 'Provisions on Democratic Principles' of the Lisbon Treaty highlights two key principles: 'political equality', in Article 9, and 'representative democracy', in Article 10. These provide the self-proclaimed democratic 'meta-standards' of the EU (Lord and Pollak 2010, 126). As Article 10 spells out: (2) Citizens are directly represented at Union level in the European Parliament. Member States are represented in the European Council by their Heads of State or Government and in the Council by their governments, themselves democratically accountable either to their national Parliaments, or to their citizens.
(3) Every citizen shall have the right to participate in the democratic life of the Union. Decisions shall be taken as openly and as closely as possible to the citizen. (4) Political parties at European level contribute to forming European political awareness and to expressing the will of citizens of the Union.
In addition, Article 11 stipulates that the European institutions shall give citizens and associations 'the opportunity to make known and publicly exchange their views in all areas of Union action' and to maintain a regular and open dialogue with them. It also introduces the European Citizen Initiative whereby a group of at least a million EU citizens may petition the European Commission to further actions that fall within its competences. Therefore, the Lisbon Treaty distinguishes between an electoral, a territorial, a functional, and a direct channel of representation. However, it does not clarify the relationship between them. Moreover, political equality in the Lisbon Treaty refers to two different political subjects: individuals and states (Krö ger and Friedrich 2013b). These two kinds of subjects involve different normative goals and a related distribution of rights and obligations. On the one hand, there are national representative democracies that rest on the political equality of individuals and the EP that seeks, however imperfectly, to transfer this representation of individuals to the EU level. In these cases, political equality refers to the individual's equal right to elect, control and sanction government. On the other hand, states are also the subject of political equality and thus of democratic representation, and it is states which bear certain rights and obligations. In this case, democratic rule is in the hands of elected governments. Both kinds of subjectivity have different reference points: popular sovereignty, in the first, and state sovereignty, in the second.The former points towards an integrated European polity with state-like characteristics, while the latter treats the EU as an advanced intergovernmental organisation. The one is primarily enacted through electoral, functional and potentially direct representation, whereas the other is primarily enacted through territorial representation.
This institutional complexity of different channels of representationand we have not even mentioned the comitology system and all the powers delegated to more or less independent agencies -and as mirrored in the seat and vote distribution in both the EP and the Councils, is precisely an attempt to come to terms with the EU's potential representation deficits. NPs and the EP and the direct democratic channels such as referenda seek to address the first representation deficit by controlling the actions of elites, especially national executives and the Commission. As we noted above, the use of degressive proportionality to allocate seats in the EP and the double majority required for voting in the Council (55 per cent of states representing 65 per cent of the EU population) is designed to ensure that proposals have the backing of small and large states, while the ordinary legislative procedure involving co-decision with the EP effectively means that they also are backed by a broad ideological consensus. As such it overcomes the first version of the second kind of representative deficit (2A) whereby minorities might be overlooked by tyrannous majorities. However, the relative independence of the Commission and the aforementioned use of non-majoritarian regulators and consultation with sectoral groups and CSOs are designed to overcome the second version of this deficit (2B) by ensuring the representation of experts and the general interest in policy-making. The result is a multitude of avenues for representation. Yet, this very abundance arguably creates a representation surplus in many instances, with increasing difficulties for holding the representatives to account.

From Demos to Demoicracy?
One recent attempt to justify this representation surplus and come to terms with the apparently insurmountable divide between a stress on the political equality of states, on the one hand, and the political equality of citizens, on the other, has been to think of the EU as a demoicracy (Nikolaïdis 2004). This view stresses 'the horizontal and mutual opening between peoples in a shared polity. It assumes that Europe is not constituted by separate demoi nor demoi-made-into-one demos but by distinct political demoi progressively opening to each other and to each other's democratic systems' (Nikolaïdis 2013, 252). It argues that national representative democracy should remain at the centre of European democracy, but also sees the EU as a way of perfecting rather than dissolving national democracies. It argues in favour of some degree of supranationality and the respective loss of national sovereignty, but also holds that there should be a 'right to exit', which federalists 'decry', thereby presenting European integration as an open-ended process. In sum, in a demoicracy peoples 'govern together but not as one' (Nikolaïdis 2013, 254).
As the demoicratic approach indicates, the question of whether or not citizens or peoples are democratically represented in the EU rests on a satisfactory resolution of the constituency question, namely the question of which body politique should be represented, at which level, and of whether that political community accepts the respective representation as legitimate. If in a globalised world a return to a world of independent sovereign nation-states seems foreclosed, a system of interdependent nation states offers in many respects an appealing vision of the EU. As Richard Bellamy suggests in his contribution to this issue, paradoxical though it may seem supranational integration and cooperation are necessary for nation-states to retain some sovereignty over their territorial affairs and interests. Bellamy proposes that 'EU institutions are best conceived as representing the peoples of Europe'. He shows why democratic legitimacy involves governments being representative of a people and outlines the ontological and epistemological conditions that must hold both domestically and in the supranational environment of the EU in order to establish conditions of civic freedom through collective self-government. Bellamy argues that the EU's system of representation actually facilitates the representation of different European demoi, and suggests that 'moves away from such a union towards greater political unity involve an inevitable loss of representativeness and democratic legitimacy'. From this perspective, the key is to represent citizens in decisions that relate to the ways their political communities interact and interconnect in a manner that shows both citizens and states equal concern and respect. Yet, that solution rests on the EU's competences remaining limited to those issues that facilitate but do not subvert their interdependence.
In their contribution, Christopher Lord and Johannes Pollak offer an alternative account of the representative system of the EU. They argue that it is best understood and evaluated not through the lens of the standard account but via the claims-making approach developed by Michael Saward (2006Saward ( , 2010. Saward argues that representation can be understood as a constantly changing social dialogue in which different actors make claims to audiences which discuss, reject or amend them. His work is grounded on three assumptions. First, representation is not a relationship at precise moments such as elections, but must be thought of as a continuing process that evolves over time. Second, the core of representation is the practice of claims-making to be a representative. Third, constituents and representatives need not be respectively members of electoral districts or elected parliamentarians. Saward insists on the dynamic nature of the relationship and on the performative (rather than the institutional) side of representation, in which both represented and representative play active roles. Such a conception of representation, according to Saward, is well suited to capture power and interests in political representation.
Lord and Pollak argue that the claims-making approach remains theoretically underspecified. They propose a revised version in which claims about rights and obligations are used to specify representative claims, so that it is clear who represents whom and for what purposes. In applying this revised version to the EU, they argue that the representative system outlined above can be seen as a series of claims by the EP, Council, and NPs, among other actors, to represent citizens within the EU arena in which distinctive rights and obligations are attributed to different actors. As such, it offers an alternative sort of representative system to one conceived according to the standard model. Yet, as they concede, worries about authorisation and accountability remain even in their revised model.

The Role of National Parliaments
The traditional mechanism for tackling this issue has been through the parliamentarisation of the EU by strengthening the EP and the European parties and thereby enhancing the link between citizens and their respective MEPs and rendering European elections first-order elections (Follesdal and Hix 2006). Proponents of this position also often suggest the election of the President of the European Commission by the EP (Decker 2002, 261;Hix 2002). They attribute the greatest hindrances to the development of European democracy to the lack of a European-wide party system capable of generating government and opposition groups within the EP (Hix and Follesdal 2006). They contend that by increasing political competition at the EU level by empowering the EP, voters will become more aware of what their MEPs are doing and so be more interested in having a say come the next elections. As a result, the democratic deficit would dissolve.
However, many dispute this analysis. They note that the continuous transfer of competences to the EP has been accompanied by a steady decline in voter turn out (Weiler 2012, 830). They also contend that the standard model would be inappropriate at the EU level because it would produce a representation deficit of minority peoples that the current system serves to protect (Dehousse 1995). Indeed, the current system works well at representing left-right opinion across the EU not despite the secondorder nature of European elections but because of that nature (Mair and Thomassen 2010b). Yet, in avoiding a populist representation deficit of type 2A it creates one of the first elitist kind because it makes it hard to hold representatives to account and to actually influence policy decisions (Mény 2002, 9).
Either in reaction to the prevailing difficulties with democratisation via the EP, or because of different concerns to start with: namely, the potentially damaging effects of European integration on domestic democratic politics and on NPs in particular, increasing attention has been paid to NPs and to strengthening their role in EU policy-making. In the EU context, the domestically rather clear role-distribution between government and parliament gets blurred. Neither the parliament nor the government has direct influence over the European agenda. Governments hardly ever have electoral authorisation for specific Union policies nor can they or NPs initiate legislation. Meanwhile, the powers of accountability of the opposition as well as of parliament more generally are weakened (see Puntscher Riekmann and Wydra in this volume; Bellamy and Krö ger 2012). Not surprisingly then, the 1990s witnessed the emergence of a 'deparliamentarisation' thesis (Raunio and Hix 2000). This thesis seeks to capture the transfer of policy-making powers to the EU and the commensurate strengthening of executives in EU policy-making. The resulting decisional and informational asymmetries between the legislature and the executive produce a loss of power and influence of domestic parliamentsand with them those of the electorate who elect them (Auel and Benz 2005, 373;Raunio and Hix 2000, 145). NPs have no direct control over European policy-making and 'suffer from a lack of authoritative power over transnational policymaking' (Schmidt 1999, 25). Instead, executives have become the 'gatekeepers' in EU policy-making with NPs the main 'losers' of European integration (Maurer and Wessels 2001;Bellamy and Krö ger 2012).
Against this trend, NPs have become actors in their own right in EU policy-making. The Lisbon Treaty mentions NPs in the main text of the Treaty for the first time. Article 12 details the basic rights and functions of NPs in EU matters and introduces an 'Early Warning Mechanism' (EWM) that assigns national legislatures the right to scrutinise proposed EU decisions and initiatives for compliance with the principles of subsidiarity and proportionality. Furthermore, NPs can have a collective legislative influence in that a majority of them may force, by way of a so-called 'orange card', an early vote on an EU legislative proposal in the Council and the EP.
In his contribution to this issue, Ian Cooper argues that this development suggests a 'tricameral' model in which 'NPs constitute the third chamber in the EU's representative system'. Cooper defends this kind of reconfiguration of representation and argues that it 'moves the EU away from traditional models of representative democracy and more towards a complex "demoi-cracy"'. However, as Carina Sprungk notes in her contribution, the role of NPs is changing in the process. Her main contention is that 'rather than simply re-enhancing traditional powers of NPs, these reforms imply ideas of a new type of parliamentary democracy in Europe', leading to a 'sustainable transformation of national democracy'. More specifically, Sprungk shows that under the Lisbon Treaty, and as a reaction to being the 'main losers' of European integration, NPs need to adapt new roles in order to fight their own disempowerment, which however 'significantly deviate from or even conflict with their traditional roles'. These new roles, which the empirical analysis explores in the French, German and Polish parliaments, are gatekeepers, networkers, and unitary scrutinizers, all of which require NPs to become more cooperative actors while neglecting their traditional representing and controlling functions, findings that are echoed in the analysis by Puntscher Riekmann and Wydra.
Indeed, Sonja Puntscher Riekmann's and Doris Wydra's contribution suggests an even more pessimistic analysis. They note the difficult balance within the EU between the representation of citizens and that of demoi, or, as they put it the 'representation of the whole versus the representation of the parts'. As they show, in EU policy-making achieving this balance is particularly difficult for parliaments that are asked to 'endorse the European decisions of their governments and simultaneously to sell the sacrifices to their constituencies' given that the preferences of governments resulting from international obligations and those of domestic constituencies can be quite different. The empirical evidence presented shows that from 2010-2012 in Italy, Germany, and Austria legislation relating to the handling of the Euro-crisis was passed in NPs by majorities that were time and again formed by government and opposition parties. This suggests that many MPs entrapped in the European rescue discourse preferred to represent 'the whole' rather than its 'part', privileging the supposed European demos over their own constituency. That could be regarded as a responsible response to the crisis that avoids a populist representation deficit of type 2B. Yet, it potentially involves representation deficits of the first elitist kind and possibly of type 2A as well (Mair 2006). It suggests that opposition parties in particular have neither controlled the 'part' that their national government represents nor 'the whole' that the Council represents, allowing an elite solution to be imposed without popular support. That proves an especially deleterious desertion of their duty to oppose given the bailout and austerity packages have been largely set outside the normal EU decision-making process, thereby allowing the more powerful credit states to set terms for the weaker and smaller debtor states -the resort to an extra-Union Treaty being 'but the poignant legal manifesta-tion of this political reality' (Weiler 2012, 831). As a result, Puntscher Riekmann and Wydra contend that the demoicratic approach of Bellamy fails, and that there may be no alternative to strengthening the standard model of representation within the EP and creating a European demos.

The Role of Non-State Actors
Disillusionment with parliamentary institutions, both the EP and the NPs, has led some to question the continued viability of the standard model of representation that they embody altogether. They note that while the traditional institutions of representative democracies are facing falling voter turnout, a massive drop in party membership and declining overall trust, other forms of political action coming from civil society, both unorganised and organised, have firmly established themselves in the political arena. They contend that the modern territorial state, and with it the link between democracy and representation, is challenged through a variety of diversification processes, including those of supranational (European) integration, of competences, actors, and arenas, which have contributed to the increasing dilution of traditional representative politics (Warren and Castiglione 2004). Given the transfer of competences to the EU, its fragmentation and its densely structured multi-level politics, all of which weaken the ability of national democracies to keep decision-making authority in their hands, the relationship between representation, democracy and the nation-state in the EU is particularly strained. What is more, the nation-state's apparently clear demarcation of the demos that can or should be represented is increasingly dissolving and there is no clear institutional centre of authority anymore given the delegation of competences to a multitude of non-state or semi-state actors. In the light of these developments, it has subsequently been argued that representation can no longer be restricted to electoral representation or to representation in the nation-state (Rehfeld 2006;Lord and Pollak 2010).
Indeed, there has been a systematic attempt by the European Commission to promote and develop 'societal representation' at the EU level and include non-state actors, and in particular CSOs, in EU politics (Bellamy and Castiglione 2010;Kohler-Koch and Finke 2007;Saurugger 2010). In the much debated White Paper on Governance (2001), the Commission even went so far as to claim 'its legitimacy today depends on involvement and participation' (European Commission 2001, 11). Since the early 1990s, it has massively invested in the creation and survival of some large, non-profit seeking European umbrella organizations, thereby seeking to increase the legitimacy of its proposals and to contribute to the construction of a transnational demos. The EU more generally has also accorded CSOs an important role in its policy-making, culminating in the Lisbon Treaty which establishes a legal duty to consult with them.
Many scholars of the role of CSOs in the EU have been interested in whether -or indeed hoped that -CSOs could act as a means of bridging the gap between EU citizens and Brussels, a functional equivalent to (the lack of) parties at the EU level, a force of Europeanisation and democrati-sation at the same time, constructing a European demos and thereby rendering the EU more legitimate overall. The hope is that CSOs and their interactions and input into policy-making could provide the mutual trust that is needed under circumstances of mutual (inter-)dependence as we find them in the EU as well as for more legitimate governance of the EU itself as it expands its competences. As is well known, the EU lacks the fundamental features of nation-states which enable citizens to collectively govern themselves: a demos, a public sphere for debate, and therefore an authoritative channel of representation. Seeking to enforce further political integration in the EU without the necessary democratic substructure may indeed be dangerous both for the EU and the national democracies that compose it (Schmitter 2000, 115), and CSOs are imagined to bridge that gap. More normatively, some authors argue that party politics is in some structural sense exclusive and that therefore additional forms of representation are required (Mansbridge 1999;Young 2000). From this perspective, functional representation, according to some authors, could contribute to the realisation of political equality if it brings weak interests into the political process (Young 2000).
Two contributions to this issue address related questions empirically. Sandra Krö ger explores whether European umbrellas of CSOs are actually representative of the constituencies they claim to represent, arguing that no matter whether these organisations represent professional interests, a cause or weak interests, societal involvement in their representation is required for it to be legitimate. Her findings show that umbrellas active in the field of agricultural policy are highly representative of their constituencies, while less representativeness is found for environmental groups and still less for anti-poverty organisations. In the two latter cases, EU policymaking seems to still be a matter of a few Europeanised policy officers, who additionally often use other lobbying strategies than the investigated European umbrellas. The related CSOs, therefore, do not seem to mediate between the EU and its citizens as hoped by some normative theorists or Commission officials -not via the investigated EU umbrellas anyway. Instead, the involvement of the European umbrellas contributes to the creation of a representation surplus.
In her contribution, Elizabeth Monaghan looks at CSOs not so much as interest aggregators, but as dynamic claims-makers who dispose of multiple ways to make issues present (see the contribution by Lord and Pollak to this issue). Addressing the role CSOs have played in the context of EU climate change policy, she makes three points. First, CSOs are autonomous actors of representation which do not necessarily represent pre-existing constituencies and interests, but instead shape an issue and the representation of that issue themselves. They represent an issue rather than a people. Second, Monaghan notes 'the independent role of ideas in shaping actors' behaviour alongside an assessment of which course of action promotes their material benefit'. Third, in the context of EU democracy, CSOs are associated with a normative project which is 'about democracy as building a political community', the desirability of which as a top-down project Monaghan questions.

A More Direct Say for Citizens
The last channel of representation is the direct channel. This is the most recent channel, which was introduced with the Treaty of Lisbon in the form of the European Citizen's Initiative (ECI) and only came into force in April 2012. Its legal basis is set out in Article 11, para. 4 TEU and in Article 24, para. 1 TFEU. By way of an ECI, European citizens can invite the Commission to propose legislation on matters where the EU has competence to legislate. The ECI has to be backed by at least one million EU citizens, coming from at least seven out of the 28 MS. Once officially registered, the organisers of the ECI have one year to collect signatures. After that, the Commission examines the initiative and decides whether to take action or not. If it decides to put forward a legislative proposal, the normal legislative procedure is in place, and, if adopted, becomes law.
Another The ECI harks back to a tradition of directly involving citizens that long pre-dates this innovation -namely, the use of referenda in nation-states. The Commission and pro-Europeans more generally have often been fearful of such exercises in direct democracy. Following the 'petit oui' of the French Maastricht referendum, and the rejection of the Constitutional Treaty by France and the Netherlands and, in the first referendum, of the Lisbon Treaty by Ireland, they have understandably worried that such moves favour those opposed to European integration. In their study, Richard Rose and Gabriela Borz present evidence that challenges their assumption. Multi-level statistical analysis shows that citizens dissatisfied with government performance are more likely to want referendums to check their governors and that national context matters. However, they are in a minority. Most of those who endorse EU referendums favour the EU as it is: what raises doubts are steps toward further integration. As a result, the risk of defeat if the EU consulted its citizens in a pan-European referendum is less than might have been supposed. As such, it offers a way of offering popular endorsement for the EU that may overcome the first, elitist, representation deficit and some of the democratic deficit. The present practice of regarding referendums as subsidiary choices that are left to each MS to decide creates inequalities among EU citizens. This has meant that from 72 to 99 per cent of citizens have had no chance to vote on a treaty. A pan-European referendum would give all citizens a chance to vote. The result would need a federal type decision rule for constitutional amendment, for example, a super-majority of voters and of MS, and provisions for national majorities opposed to further integration to opt out rather than veto, as at present. Otherwise, this proposal would run the risk of both forms of populist representation deficit in that if, as they suggest, national preoccupations remain salient, then certain national groups might override the views of other less populous nations and, though they believe it less likely, there is also the danger of a rejection of long term European for short-term national advantages. In other words, even more than with the EP, a pan-European referendum requires that the ontological and epistemological conditions for a demos are met.

Conclusion: The Conflicts and Complementarities between Representation and Democracy
European integration has brought about a transformation of political authority by which the state loses its monopoly on collectively-binding decision-making. The result is a transformation of national sovereignty which is occurring empirically, and which scholars are struggling to come to terms with conceptually. Since it took on its modern meaning, national sovereignty provided an account of legitimate rule, within a confined territory and associated with defined functions, in which political authority was 'singular and supreme' (Goodhart 2007, 573). What contemporary scholars are struggling with, then, is to detach the notion of sovereigntyand thereby representation -from boundaries that bind together territory, demos and the idea of the rightful rule of that demos within those defined boundaries, leading some to criticise what they call 'methodological nationalism' (Zü rn 2000) which does not question the Westphalian state as its point of departure when thinking of normative democratic standards today.
Indeed, the EU is a problem for democratic theory insofar as it cannot be democratic according to modern accounts of democracy. On the one hand, why should we expect institutions and norms of democracy which were developed for and in the national context to retain their meaning if translated to a fundamentally different, supranational context? Given the historical contingency of territorial boundaries and the decreasing congruence between the people affected by political decisions and nation-state boundaries, democratic theory is not well advised to treat the identity of a people as given. On the other hand, it would be unacceptable to not be guided by normative standards when evaluating the democratic quality of the EU, and citizens are likely to be inspired in their assessment by what they consider to be democratically legitimate in their respective national contexts.
The problem is epitomised by the way representation and democracy are coming apart within the EU. In the context of a sovereign state, where representatives are responsible for the main policies affecting citizens and these citizens form a people that meet the ontological and epistemological conditions specified above, minimal tension need exist between democracy and representation. Popular sovereignty can be exercised to authorise and hold representatives to account without the danger of any of the representation deficits. Yet, as states become more interconnected, their populations more diverse, and decision-making more complex, the more likely representation deficits become. To the extent the EU involves demoi rather than a demos, the different peoples need to be represented to prevent any becoming a consistent minority -the representation deficit identified above as type 2B. That requirement may now apply to the 12 million EU citizens who live in a MS other than that of their nationality -roughly 24 times the populations of Luxembourg or Malta. Yet, that group, forming only 2 per cent of the EU population, does not itself constitute a European demos. However, representing such diversity makes it hard to broker decisions that are in the general interest. Veto players and joint decision traps abound. As a result, a deficit of type 2A may arise in which each people seeks its own advantage at the expense of other peoples. For example, an adequate solution to the Euro-crisis is hampered by the fact that national politicians in both debtor and credit states have had to balance immediate domestic electoral demands against the long term need for a fundamental restructuring to render the Euro zone an optimal currency area. The difficulties of doing so has led some to argue that the key problem in the EU is not a democratic deficit but a democratic surplus provoking a 'populist' representation deficit, whereby local priorities and preferences prevent governments from committing to the long term changes the future of the euro requires (Moravcsik 2012, 66-7). Yet, the alternatives of a settlement imposed from above by technocrats or the more powerful credit states return us to a representative deficit of the elitist type, that has spawned increasingly vocal forms of populist anti-Europeanism (Leconte 2010).
In this special issue we have explored a number of attempts to overcome these problems and rebalance representation and democracy: through empowering NPs, through enhancing consultation with non-state actors and through the direct participation of citizens. They seek respectively to revive the standard model of representation, offer a non-standard form of representation and do away with representation altogether. Each was found wanting in themselves. The NPs were thought to strengthen the standard model of representation with regard to national executives negotiating in the Council, thereby tackling an aspect of the first, elitist, representation deficit within the EU. Yet, not only was its effectiveness doubted, but also a risk exists that it might strengthen the second, populist, representation deficits. National politicians may be less inclined to make national sacrifices for the greater European good, thereby risking both the more powerful states overruling the weaker (type 2A) and producing suboptimal policies as a result (type 2B). If this leads to a strengthening of technocracy and elite rule, legitimacy might be obtained through consultation with CSOs as representatives of an emerging transnational European civil society. However, their representative nature was disputed, as was their Europeanised character. As a result, employing this measure to address the second, populist, deficits would appear, as predicted, to have reintroduced the first, elitist, deficit. Finally, we explored direct democracy in the form of referenda as a solution to this latter problem. While the concern that it would in its turn promote one form of the second, populist, deficit (2B) was shown to be less than feared, referenda can themselves be mechanisms of elite manipulation given they are usually in charge of posing the question. In the absence of a demos it may also still lead to a deficit of type 2A, in which a majority tyrannises a minority.
The balance between representation and democracy in the EU proves hard to achieve, therefore. Overcoming the democratic deficit risks producing a representation deficit and vice versa. The challenge and direction for future research must be to look at both the separate channels of representation and the ways they interact. Only such a holistic approach is likely to produce a view of the EU's political system that can combine both representation and democracy in ways that are complementary rather than conflicting.