On Infeasibilities of Cosmopolitan Democracy – Lessons from the European Union



Abstract:  This article critically examines institutional versions of cosmopolitan democracy and institutional ideas of global federalism. Assuming global institutions to be as they are designed in conceptions of cosmopolitan democracy, the paper addresses a feasibility question: (To what degree) would such global democratic structures meet normative standards of democratic accountability. As there is no global democracy existing, the argumentation makes a detour, first investigating the experiences of the most advanced project of cosmopolitan democracy – the European Union – and then, second extrapolating from persistent, structural democratic deficits in the EU to the feasibility of global democracy. Compared to national arenas, the EU faces a gradual deficit in democratic capacities and the global reality of cosmopolitan democracy would be –even compared to the EU – a downgraded democracy as well. The main findings are: Vertical accountability is either more ineffective or more inegalitarian or both. Neither unitary nor federal systems should be seen as a plausible solution for the threat of ‘Verselbständigung’ caused by multilevel politics. And nationally segmented public spheres will not promote a similar type of politicised discourses around ‘common’ global issues.


This article critically examines institutional versions of cosmopolitan democracy (e.g. Archibugi 1998, 2004, 2008; Caney 2005, 2006; Held 1995). Defenders of such a position do not just share certain moral or political principles of cosmopolitanism. With respect to certain well defined policy issues (fundamental human rights and policy problems caused by denationalisation), they additionally assume a more or less independent global policy making arena should have normative supremacy in making collectively binding decisions. And this decision-making should and can be democratic.

My concern here addresses a specific variant of the ‘can’-question: I take global institutions as institutional cosmopolitans want them to be, I take the rest of the world to go as it currently does, and I ask: (To what degree) would such global democratic structures meet cosmopolitan democrats’ normative standards of democratic accountability.

Why the feasibility question and why this specific kind of feasibility question? If we solely address the ‘should’-question, normative cosmopolitanism is not that contested in the age of denationalization: Societal denationalisation has led to a relative loss of democratic quality. We now face existing, binding international norms on issues, which most democratic theorists expect to be decided democratically, at least if these norms are decided nationally. The existing international modes of decision-making and policy reform (normally via contracts) are ineffective (substantially under-determined due to unanimity rule) and undemocratic (more or less impossible to change without general consent). Substantially under-determined law leads to strong reliance either on administrative and judicial ‘interpretation’ or second and third order law making which again adds to democratic failure of global politics. Thus, prima facie, institutional versions of cosmopolitan democracy plausibly address and might resolve the problem of democratic fate.

When I assume in this article global institutions to be as they are designed in conceptions of cosmopolitan democracy, I give away a strong and prominent feasibility related criticism towards institutional cosmopolitanism – the realist, agency related charge to be utopian, unrealizable etc. I do this for two reasons: Firstly, as strong as this criticism is empirically (see Zürn 2011), it cannot plausibly challenge our normative attitudes towards institutional cosmopolitanism as a valuable normative ideal. But if the desired global democratic structures would not pay off in adequate realization of democratic principles, then the normative desirability of institutional cosmopolitanism is at stake. Secondly, I expect by following the latter strategy that it is possible to show, that this is precisely the case: global democratic institutions would implement a down-graded democracy at the global level.

Obviously there cannot be a conclusive way to show empirically that non-existing institutions do not work adequately. I suggest to make a detour, first investigating the experiences of the most advanced project of cosmopolitan democracy – the European Union – and then, secondly extrapolating from these results to the feasibility of global democracy. I approach this aim in four main steps: In section 1, the normative standards and institutional visions of cosmopolitan democracy are summarized and the EU is presented as an appropriate test case for such normative aspirations. Then, in section 2 and 3 a demanding, general and norm-driven conception of democracy or rather democratic accountability is introduced and its applicability to the EU and institutional cosmopolitanism defended. In section 4 the deficits of the different channels of European democracy are summarized. The EU democracy is analysed as a sum or synthesis of very different supranational and international institutional settings of opinion and will formation. Compared to national arenas, it can be shown that these EU related settings have to face a gradual deficit in democratic capacities and – equally important – some of these deficits should be seen as persistent, i.e. more or less unchallengeable by institutional reforms. Section 5 finally extends the analysis to the ‘gradual infeasibilities’ of democracy in international arenas. Here theoretically relevant consequences for international democracy and for conceptions of cosmopolitan democracy are discussed.

1. The Case for the EU from Cosmopolitan Democracy Perspective

For the feasibility test cosmopolitan democracy should be taken as if global institutions would be designed just as defenders of cosmopolitan democracy want them to be designed. In other words, normative standards of cosmopolitan democracy are assumed to be desirable and the institutional aspirations are assumed to be implemented in world politics.

As a reference conception of (an institutional version of) cosmopolitan democracy I take Daniele Archibugi’s latest book “Global Commonwealth of Citizens”. The main argument of the book consists of four elements:

  • 1 Democracy is defined as political system implementing three “cardinal principles”: “nonviolence, popular control, and political equality.” (Archibugi 2008: 26)
  • 2 Democracy (in the nation-state) is the best (and fairest) way to resolve societal problems and conflicts (ibid.: 35–51).
  • 3 Due to denationalisation, some important societal problems and conflicts have moved beyond national borders (ibid.: 53–84).
  • 4 Institutional cosmopolitan democracy would be the best (and fairest) way to resolve some global problems and conflicts (ibid.: 85ff.).

None of the four core elements have to be challenged here. What is addressed instead, is the question if and to what extent an implemented ‘institutional cosmopolitan democracy’ would grant democratic principles. The lesser this is the case, the more unlikely it is, that Archibugi’s case for global democracy is sound.

For this purpose Archibugi’s desired global institutional design has to be clarified (see ibid.: 103ff.). He constructs his cosmopolitan democracy model as a third way in between a confederal and a federal model. As in case of the EU national governments and citizens should have their own representative bodies (similar to the EP and the Council of Ministers in the EU). In world parliament citizens should be represented in an egalitarian way (one person, one vote) and in the chamber of states each state has one vote. Electoral and state representation should be supplemented by civil society participation and a global public sphere. In a global constitution the tasks of the global political system are fixed (in the spirit of subsidiarity). And global jurisdiction becomes compulsory and binding.

In many respects Archibugi (1998: 219ff.; 2008: 94f.) himself sees the European Union as a – yet imperfect – forerunner to cosmopolitan democracy. At least four reasons make the EU an appropriate test case: Firstly, the desired juridical integration, the combination of citizens and state representation in two different assemblies, but also wide civil society involvement are essential to the EU. This is more than a confederation and less than a federal state. Thus, all relevant features of cosmopolitan democracy exist in various ways at EU level. Secondly, in horizontal perspective across policy fields, the political integration at EU level is strong compared to the rather fragmented governance arrangements at global scale. The EU is the only governance arrangement beyond the nation state which can live up to the systemic expectations of cosmopolitan democracy. Thirdly, even compared to similar institutional arrangements at the fragmented global scale, the democratic performance of the EU is undoubtedly better (see Zweifel 2006). Fourthly, there are not many democratic options to be discussed beyond the institutional modes probed at EU level: i.e. an elected independent parliament, independent administrative bodies (European Commission, regulatory agencies), ‘Second Chambers’ (European Council, Council of the European Union), independent transnational societal bodies (‘Social Dialogue’), various kinds of direct citizens’ participation as well as institutionalized involvement of civil society organisations (policy related consultations, European Citizens Initiative, referendums, elections, deliberative polls etc.) as well as various kinds of cross-organizational bodies (Comitology, expert groups etc.). To cut the long list short, among the various kinds of institutional settings possibly promoting cosmopolitan democracy at global scale, I can see only one which hardly exists at EU level –“multisectoral networks” (see e.g. Benner et al 2005), where various kinds of public and private organizations are merged.2

Hence the democratic performance of the European Union can enlighten our expectations about the possibility of democracy at global scale in many regards. Nevertheless two qualifications must be made. Firstly, the European Union is in a process of ongoing democratization. Many authors rightly claim that the EU has significantly democratized in the previous decades and might do so in future. Thus, a plausible comparison to the EU has to focus not just on actually existing (see below, sections 4.1–4.4) but on persistent or structural European deficits in democratic quality compared to national performances: If we accept ‘ought implies can’ as the ultimate feasibility requirement for normative conceptions, then we have to show, what the best-practice-case EU persistently ‘cannot’ (section 4.5) and that (and to what extent) these restrictions are valid for global democratic institutions (section 5). Secondly, in many regards the societies of the European Union are an ideal case, which is much more convenient for democratic aspirations than the societies of the world. So any deduction from structural democratic deficits in the EU should be seen as still having an optimistic bias towards cosmopolitan democracy.

One might argue that a comparison between national and the European democratic performance is unfair as normative standards should be lower than for nation-states or they should be different. This is not Archibugi’s position for two good reasons. First, if there should be collectively binding decisions on important global issues which deserve to be named democratic, they have to meet democratic standards. Second, Archibugi’s argument heavily rests on the empirical insights into the instrumental values of (national) democracy. If we would modify the essence of democracy for our global aspirations, the empirical connection to the instrumental values of democracy becomes questionable as well.

Unfortunately it is difficult to simply apply Archibugi’s own normative principles to our analysis. They are problematic, too broad and unclear at least in some respects.3 Therefore four normative standards of democratic accountability are outlined in the following section, grasping the normative core of democracy, which are (hopefully) clearer, but none the less consistent with Archibugi’s normative intentions.

2. Conceptualising Democratic Accountability beyond the Nation-State

The core of democracy is what Archibugi calls popular or public control and political equality in the course of political action. Decision-makers and important decisions should be bound to the demos in an egalitarian way. Obviously there are many other pre-conditions and normative aspirations connected to democracy in the nation-state. But it should be undisputed that a variant of popular control and political equality sufficiently indicate democratic quality. In this section I suggest, that these normative aspirations of the democratic core can be appropriately grasped in a conception of democratic accountability.

An elaborate version of such an accountability-centred approach to democracy, which has also been applied to the EU, is presented by Philippe Schmitter (Schmitter 2000, 2007; Schmitter and Karl 1991). For him and Terry Karl effective accountability is a sufficient condition for democracy: “Modern political democracy is a system of governance in which rulers are held accountable for their actions in the public realm by citizens, acting indirectly through competition and cooperation of their elected representatives.” (Schmitter and Karl 1991: 76).

According to this definition mechanisms qualifying as rendering a political system accountable are of two kinds: mechanisms of vertical accountability binding the rulers or rule-making directly to citizens (e.g. elections) and horizontal mechanisms of accountability (or Schmitter’s competition among representatives) binding different kinds of representatives (for the distinction of vertical and horizontal mechanisms, see O’Donnell 1994). I agree with Schmitter in three respects: Firstly, accountability is an essential ingredient of any modern democracy, secondly only a combination of a variety of vertical and horizontal mechanisms can lead to effective democratic accountability, and third that this definition might be a good concept to empirically select more democratic from undemocratic countries in the world. But this minimal definition is not demanding enough for a comparative investigation of the democratic credentials of political systems in the democratic world. Schmitter’s definition explicitly misses any reference to political equality which is a core, if not the primary standard in most other normative conceptions of democracy (Christiano 1996; Dahl 1989; Dworkin 2000; Habermas 1996). Democracy does not materialize if rulers are just held effectively accountable by any citizens, but only if all citizens have an equal right and chance to do so in an egalitarian way. Whereas Schmitter’s conception of democracy can be regarded as the idea of maximizing effective vertical accountability to citizens, I assume the normative aim for a democracy is to optimize effective vertical accountability and political equality or to put it differently, to maximize effective egalitarian vertical accountability.

Other democratic theorists might ask: Why not simply maximize political equality (McGann 2006)? If we only were interested in the egalitarian aspect of citizens’ influence in the political realm, we would more or less automatically favour less demanding procedures of participation over ‘thicker’ participation. For McGann a political system approximating proportional representation in electoral systems and majority rule in the representative assembly would be the best possible democratic system. This is obviously egalitarian, but as Ronald Dworkin has pointed out, also close to minimizing all citizens’ (equal) political influence.4 This would question the value of such a kind of political equality (see Dworkin 2000: 190ff.).

Whereas the weakness of Schmitter’s approach is that his minimal conception of democratic accountability makes oligarchic and other half-way democratic arrangements indistinguishable from democratic arrangements, the pure egalitarian representation approach is concerned with the question whether all votes have been counted and weighted correctly, but not with the question, whether these votes (and the citizens as voters) have a significant influence on the political course of action. It should be clear, that an appropriate idea of democratic accountability has to combine both, the idea of egalitarian and effective impact of the people on public rule.

If maximizing egalitarian vertical accountability is the primary democratic aim, why should we focus on mechanisms of horizontal accountability at all? The institutional and functional differentiation of the political system can both enable as well as constrain vertical accountability. Functioning electoral commissions for example enable and secure the fairness of elections and thereby enable vertical accountability. But the working of other functional agencies might well constrain the options for vertical accountability. Therefore I will have to scrutinize mechanisms of horizontal accountability in order to judge whether or not they add up to make the system democratically accountable.

Here the conceptual reasoning about horizontal accountability will focus on the question whether such mechanisms have the potential to foster vertical accountability or if they are bound to constrain it. Empirical research would have to analyse (1) the quality of vertical accountability mechanisms, (2) the quality of horizontal accountability mechanisms in supporting and enabling vertical accountability and (3) the constraining effects of horizontal mechanisms on vertical accountability. Only when engaging in such a threefold analysis an adequate systematic assessment of democratic accountability is possible.

The three dimensions tell us (in a rather abstract way), what we should analyse. But which normative standards should be applied to the accountability mechanisms? Or to put it differently: How do we know if and to what extent such a complex system of accountability makes for democratic accountability? I suggest four normative standards which I consider to be sufficient for indicating the quality of democratic accountability:

  • 1 Maximizing effective egalitarian binding in vertical mechanisms. In a democracy political rules and rulers should be bound to the demos. The more this is the case, the more democratic a political system is. As argued above the binding should be effective and thus the demos should be influential, while respecting political equality. Three clarifications are in place here: Firstly, while the working of effective egalitarian binding is linked to working institutions or mechanisms, the normative standard remains agnostic with respect to concrete procedures. Secondly, the formulation envisages a primarily representative or indirect rule of the people but does not exclude more direct implementations of democracy. And thirdly, within dominantly indirect democratic political rule, the standard of effective egalitarian binding has to be complemented by three secondary normative standards, which are none the less necessary for democratic accountability.
  • 2 Effective guarding of democratic procedures and avoidance of illegitimate mis-use of political power via appropriate horizontal accountability mechanisms. In democracies rulers have a more or less precisely defined mandate and democratic rule has an expiration date. Both have to be safeguarded, that rulers do not mis-use their political rule beyond their mandates and that rulers can be replaced and decisions are alterable over time. Obviously, in modern political systems the democratic nature of the polities has to be safeguarded by distinct agencies, but the normative standard should not be conflated with these institutions.
  • 3 Minimizing ‘horizontally induced’ constraints on vertical accountability. In accordance to the second standard, mechanisms of horizontal accountability are necessary to stabilize the democratic character of political systems. But they also might have negative side-effects on democratic binding in political systems. This potential trade-off between standard 1 and 2 is addressed here. Logically the standard is completely inherent in former standards, but with respect to the analysis of accountability deficits, it is important to have accurate access on processes of horizontally induced ‘Verselbständigung’ (see below). Right now respective criticisms of EU-specific problems of horizontally induced constraints of democratic accountability are beyond the focus of accountability research. And this biases the research in favour of the quality of democratic accountability in the EU. The introduction of the third standard should help to overcome this bias.
  • 4 Maximizing publicity or public discourse on the framing of political processes. At least since Bentham (1816) publicity is seen as a necessary condition for effective accountability: If we do not know who the rascals are we cannot throw them out - or hold them accountable in any demanding way (what is a more precise expression for our normative aspirations). Additionally, actual accountability is and should be for most of political affairs a potential threat, a virtual horizon and not ‘just’ a material practice (see Thompson 1987). It is the overall publicity of political affairs, which decides about the effective potential to hold the political elite accountable in the vertical dimension.

For the purpose of this article I assume these four standards are an appropriate normative yardstick for democratic accountability. But obviously this assumption needs clarification, specification and (not the least) additional justification. The three most important considerations are:

  • 1 All four standards need terminological clarification. There are many potential meanings, which are ‘out there’ in the normative literature on democracy. An important task of investigation is clarifying the precise meaning of ‘effective egalitarian binding’ (Who should bind whom or what and how?), ‘publicity’ and so on (see below, section 4).
  • 2 Depending on the ideological background, liberal and egalitarian conceptions of democracy would ascribe somewhat different functions to democratic accountability, especially to horizontal mechanisms (Bovens 2007: 113ff.). Therefore we have to face a certain degree of normative pluralism also influencing the question of adequate accountability standards: liberals might see a more expanding legitimate role for courts securing individual liberties than more egalitarian democrats and more deliberative conceptions of democratic accountability might emphasize the value of reason-giving and justification, where more aggregative conceptions emphasize the availability of formal sanctions. There are at least three plausible ways to resolve the problem of normative pluralism: constructing overlapping standards (as the four normative principles outlined above), avoidance of contested criteria (see below, section 4), and empirical investigation of plural standards (see Hüller 2010a: 62ff.). I assume (but cannot discuss here) that the problem is best resolved by a combination of all three methods.
  • 3 There is a need for ‘normative ordering’ of the four accountability standards, concerning e.g. the appropriate understanding of optimized political equality and vertical influence (as discussed above). Furthermore in order to gain prescriptive guidance in case of trade-offs between different accountability standards (for expected or at least possible trade-offs, see e.g. Hood 2010; Papadopoulos 2010), a deeper discussion of normative conceptions of accountability is desirable, but beyond the scope of this article.5

To put the previous discussion in a nutshell (or in a formula closest to a definition), the focus has to be on ‘system accountability’: First and foremost, the ‘class’ of rulers should be bound to the demos. And the more effective and egalitarian this binding is, the more democratic a political system is. This effectiveness or success is indicated by the realisation of four (provisionally fixed) normative standards, where egalitarian vertical accountability has normative priority.

3. Analysing ‘Accountability Webs’ in the EU and beyond: Lessons from Comparative Research

Despite many concrete and specific differences of the EU polity in comparison to diverse nation-state polities, there are some structural similarities, with respect to the questions, what and how we have to analyse democratic accountability. So EU research can learn from the diverse national experiences. In this section it should be clarified how comparative research can contribute to the EU analysis.6

Before I outline my approach aimed at comparing diverse national and EU accountability webs against the normative yardstick of the four standards of democratic accountability, I briefly discuss two different comparative approaches: a diachronic intra-EU comparison and an institution-driven comparison of national and EU mechanisms.

Bovens et al (2010: 183ff.) have recently presented a diachronic intra-EU comparison. Their main point: Today the EU is more democratically accountable than in earlier days. This result is (despite serious deficits in operationalization) not surprising, since EU institutions have promoted accountability only a decade ago. But the result is also of limited use value if the aim is to assess the democratic deficit or quality of the EU political system: Even improved accountability performance over time can be significantly below certain threshold levels of adequate accountability (whatever that may be).

A second comparative approach is institution-driven. We know about the performance of certain institutions or institutional arrangements from the comparison of national political systems. One might search for similar institutions at the EU level and expect similar performances as in national arrangements. The main difficulty of this approach is the ‘n = 1’-problem: There simply is no other multi-level political system with such restrictive legislative competences for a parliament at the top level, extensive joint decision-making and such democratically integrated sub-units. Maybe this problem is serious in comparison of national democracies as well, but for our purpose it should be clear, that there is no reference system whose accountability web should be mirrored by the EU.

So how should we compare the European political system? Not just the EU but also national political systems should be analysed as rather unique accountability webs, integrating many different mechanisms in more or less unique ways. And their democratic performances should be analysed against a common normative standard of democratic accountability.

When solely looking at vertical accountability mechanisms (as most comparative research does), elections seem to be the most important, for some authors this is the only instrument worth to be analysed (Przeworski 1999; Hellwig and Samuels 2007; Schmitter and Karl 1991). Nevertheless, a second look reveals that there are multiple possible mechanisms of vertical accountability. Their actual relevance within a political system can only be assessed empirically and might differ from one system to another. The most obvious vertical mechanisms – beyond elections – are instruments of direct democracy (referenda) and different forms of surrogate involvement (citizens’ panels, planning cells, involvement of civil society organisations), but also certain kinds of independent public discourses (Habermas 1996) and contestations (Fraser 2007).

As we know from comparative empirical research, these mechanisms are organized differently in different nation-states, and may foster democratic accountability in different degrees. This issue has been dealt with prominently in the literature on comparative politics focussing on citizens’ influence in different electoral and political systems (for an overview, see Hellwig and Samuels 2007). Pronounced differences can also be expected with respect to divergent kinds of national civil society involvement. More pluralist forms of associational involvement can be expected to have other accountability related credentials than civil society involvement in more corporatist regimes (Lijphart 1999: chap. 9). Furthermore national public spheres are open to very different degrees to (‘critical’) contributions from ‘below’ such as from civil society actors or to investigative journalism (for differences in openness to civil society organization (CSO) contributions in German and American public discourses, see e.g. Marx Ferree et al. 2002). Assessing vertical accountability mechanisms systematically demands research across these very different mechanisms.

In addition, the different mechanisms are not completely independent from one another: demanding instruments of direct democracy might have a negative effect on electoral participation. Thus in countries applying mechanisms of direct democracy citizens’ vertical influence is more effective (due to referenda) but also less egalitarian, as participation in both instruments is significantly lower than in systems relying only on elections.

Two general lessons can be drawn from these insights about national accountability mechanisms: Firstly, we have to analyse rather unique accountability webs where we have to look at multiple interacting mechanisms of vertical and horizontal accountability. We should not apply a distinct theoretically developed institutional model of democratic accountability to the EU, but rather inductively analyse the existing accountability relations assessing their success in realizing the four normative accountability standards.

Secondly, ideally the comprehensive set of vertical and horizontal mechanisms of accountability (in the broad sense) should be integrated in a complete research program. As we know, however, empirical research normally comes in much smaller portions. Thus we either have to concentrate on investigating only part of the picture or we have to re-assess existing empirical data against the standards provisionally outlined in section 1.

To get a comprehensive view, we have to be aware how levels of analysis further and constrain our insight. On the micro-level different kinds of actors and their accountability relations can be investigated in depth but are hard to generalize; on the meso-level we have the choice between selecting a relevant functional segment of the political system and explore all relevant accountability relations or we can opt for comprehensive investigations on different types of actors and analyze how they perform across policy fields; the macro-level calls for a synthesis of characteristic accountability relations and access them in view with general findings regarding systemic ‘Verselbständigung’ in order to come to a general assessment of democratic accountability.

Systematic incorporation of insights from comparative democratic research has two functions in EU accountability research: First, it gives an overview of potential mechanisms of democratic accountability and thus helps structuring the more inductive approach to gather relevant European procedures. And secondly, nation-state performances in democratic accountability should be the primary arbiter for the European practice. Obviously we could (and we do) apply more ideal normative standards, as the four outlined in section 1. But we know for sure, that no modern political system can perfectly meet such standards in practice. The comparison of European multilevel arrangements with nation-state performances against the normative standards of democratic accountability (and not institutionalized practices) seems to be an appropriate way both to avoid ‘national-container theorizing’ and to investigate the European performance in democratic accountability without over- or underambitious normative expectations.

The standards developed in section 1 are norm-sensitive - they indicate a demanding idea of democratic accountability – and they are methodologically neutral and thus enable an unbiased assessment of the democratic credentials of different institutional sets of political accountability.

4. The Democratic Credentials of the European Accountability Web

European accountability research does not suffer quantitative or qualitative insignificance: Relying on different approaches the recent research has produced quite a number of studies supporting a critical account of EU accountability (Gustavsson et al 2009) and pointing to specific accountability ‘gabs’ in European politics (Harlow 2002; Harlow and Rawlings 2007). Furthermore, as mentioned above, “The Real World of EU Accountability” (Bovens et al 2010) provides valuable insights in individual institutional accountability relations. In addition, successive investigations in parliamentary accountability in the compound system of the EU (Benz 2006, Auel 2007, Papadopoulos 2010) and electoral EU accountability (van der Eijk/Schmitt 2008) has deepened our knowledge concerning the limits of these traditional vertical accountability mechanisms in the EU context. A rather recent claim stipulates a prominent role for civil society actors to put EU policy-making under democratic accountability (Kohler-Koch 2010a; Steffek 2010).

The central aim in the following sub-sections is to re-assess existing empirical insights against the background of the conceptual framework of democratic accountability outlined above. For this purpose I estimate the EU performance in realising the four normative principles introduced in section 2.

4.1 Effective egalitarian vertical accountability

There are six relevant options (plus combinations) for citizens to influence either European policy making or selection of office holders and political agents: EP elections, referenda, the European citizen initiative, participation in European policy deliberation, indirect representation via national parties and governments, indirect representation in policy deliberations via civil society organisations.

None of these six mechanisms can secure vertical accountability to a degree which we find in national political arenas. The main problem of EP elections is neither the relatively low voter turnout (about 50%) nor the paradoxical coincidence of declining turnout and increasing EP competences, but the EP’s missing formal right to autonomously initialize European law making and select governmental office holders. Where majorities in national parliaments can effectively pursue a political program they are elected for, MEPs can only (and not in all cases) veto European policy making, when they expect their constituencies to hold opposing views or when they are negatively affected by certain policy options with respect to proposals made by the Commission. Obviously, this problem of formal competences is remediable.

Another problem with European level elections is: voters’ indication of political preferences to the elected is more or less missing, because neither do candidates offer public European political programs, nor are electors guided by EU related preferences. Neither the specific nor the general direction of future or past European law making is actually the object of the average voters’ decision in EP elections. Thus, even existing formal competences to veto European policy making processes are democratically flawed, because of the deficit in effective symbolic authorization.

Some institutions of direct democracy can be seen as an accountability enhancing supplement to representative democracy in the EU (Trechsel 2010). But as they are not part of the reform agenda of cosmopolitan democracy, I skip the discussion here (but see Hüller 2010a: 100ff.).

Along with the Lisbon treaty the European Citizens Initiative (ECI) has become a new participatory option. A detailed regulation passed by the EP and the Council, however, will come into force only in April 2012. According to this regulation, 1 million European citizens with significant support (i.e. 750 × national seats in EP) from citizens of one quarter of the member states can “invite” the Commission to take a certain policy proposal under consideration, if the Commission expects it to be consistent with European primary law (European Commission 2010). The ECI will hence be a kind of petition with some societal support. It is not possible to assess the political success of the ECI at this state, because this depends on precise implementation and the Commission’s handling with successful initiatives as well. But from democratic perspective no gain in effective egalitarian binding should be expected. The barrier of one million supporters is both too high and too low to secure egalitarianism. It is too low, since many relevant particularistic positions have no problem to overcome the obstacle. Mainly the dominant European societal actors (as the EU funded umbrella organisations) will get another tool to influence European policy making. The main problem will be that the particularistic status of such initiatives remains undiscovered, as long as the more egalitarian second and third steps of direct democratic institutions are missing. In the ECI procedure it will be the Commission’s task to judge the initiatives and the Commission might get trapped in either supporting even particularistic activities or being blamed for undemocratic elitism. Ironically the expected support might also be too high for relevant initiatives, since ordinary citizens and CSOs without Europe-wide structures are expected to regularly fail the test. There is no reason to expect, that relevant initiatives solely come from European organisations or will be supported by them.

Undoubtedly national governments are strong political actors in European primary and secondary law making via the European Council and the Council of Ministers. But do they contribute in this role to effective egalitarian vertical accountability? They do not. As national governments normally represent marginal absolute majority of national constituencies, majoritarian decisions in the Council might be backed by just about one third of the European citizenry. Only in cases of unanimity in the Council, a very rigid form of aggregative equality is secured. Consequently, the Council’s contribution can either be effective or egalitarian but not both (depending on the Council’s decision rule). Beyond this issue, it remains an open question whether (and to what extent) effective national governments are bound to their demoi. Neither are there effective prior authorisations nor ex post accountability mechanisms to the national demoi. And national elections are no effective means to vote about the general direction of EU politics.

Regarding the contribution of national parliaments, the question is not whether they are ineffective, but where they are ineffective. Either they do not have an effective option to hold their governments accountable with respect to EU issues at the national level, because relevant European decision-making occurs behind closed doors and governmental voting behaviour remains secret. Or there are demanding processes of national parliamentary will formation before European negotiations occur, then these national positions are rendered rigid and ineffective in European policy deliberations, because the national political agents are forced to stick to the deliberately found national positions (see Auel/Benz 2007).

The final option for egalitarian vertical accountability is connected to the work and influence of civil society organisations. Their democratic function is envisaged by the Commission’s famous White Paper on European governance, but up to now there is no effective egalitarian mode of CSO centred accountability. There is pluralist consultation in the EU (Hüller 2010b; Kohler-Koch/Quittkat 2010; Schmidt 2006), but neither do the participating CSOs stand for the European citizenry (Greenwood 2010; Kohler-Koch/Buth 2010) nor are the existing mechanisms of symbolic accountability in European policy deliberations effectively working (Hüller 2010a: 147–215). And due to significant and persistent ‘north-south’ and ‘old-new’ member states divide in CSO participation (Kohler-Koch/Quittkat 2010: 161ff.; Quittkat 2011) as well as due to huge communicative intra-associational gaps (Altides 2010) the egalitarian potential of the CSO input at the European level is significantly lower compared to arrangements at national level (see Hüller 2010a; Kohler-Koch 2010b).

As CSO involvement in EU policy deliberations is purely consultative, it is not enough to focus on CSO contributions (or to put it differently: the demand for symbolic accountability), but also on their effects. Neither the vast literature on influence of CSOs in European policy making nor case studies on the Commission’s symbolic accountability towards CSO contributions support an optimistic view (Hüller 2010a: 200–209; Persson 2009).

To sum up the discussion of the six mechanisms of vertical accountability: For very different reasons, there is no single mechanism securing both effective and egalitarian vertical accountability.

The assessment of ‘compounded’ vertical accountability across mechanisms might show some synergic effects concerning more representative or more pluralist deliberative processes, but a core deficit remains even after combination: there is no effective egalitarian link to the European citizenry. Above this compounded accountability would come with a serious side-effect. A core element of democratic accountability is not, that governments are actually replaced, that policies are vetoed or promoted by a certain majority, but simply the knowledge or belief, that core political actors and policy decisions are bound to the people. This belief or knowledge depends on the visibility of working accountability mechanisms. To put it differently: Accountability must be seen to proceed, and ‘compounded’ vertical accountability has to face the risk of invisibility.

4.2 Effective Horizontal Accountability

There are two main criteria for effective horizontal accountability: securing the effectiveness of existing vertical mechanisms and preventing non- and illegitimate use of power. The criteria should not be designed as a simple matter of fact (Is there corruption, manipulations of electoral outcomes etc.?) but as criteria measuring the institutional provision of these aims: Are there plausible institutional arrangements in the EU to prevent or detect undesired political behaviour? And in contrast to the other three standards, the criteria are not oriented towards approximating a maximum of horizontal accountability, but to secure a degree of horizontal accountability we find on average at national level.

As far as political competences are supranationalised European and national courts, European and national administrations, the EP and national parliaments as well as national governments have various competences to control work and behaviour of political actors and institutions. Roughly three kinds of mechanisms can be distinguished: supranationalised horizontal accountability, where competences to control political actors are located dominantly at European level. The most prominent case is the European Commission, where only the EP can force the demission of the College and only the President of the Commission can recall single members of the Commission.7 At the other end of the continuum horizontal power of supervision is in the political domain of the member states, as e.g. in the case of European elections which are implemented and controlled by national agencies. In between these extremes of national and supranational accountability procedures most European political action is held accountable by institutions from both levels, as in the production of European secondary law or the tripartite supervision in comitology procedures as well as in EU regulatory agencies.

Compared to national settings, European accountability relationships are normally more complex with respect to the institutions involved. No matter how formal competences are shared among control agencies, some deviant behaviour should at least be as visible as in national settings.

There might be two major points of concern with respect to insufficient horizontal accountability in the EU: the dominant role of the European Council as well as the accumulation of competences at the European Commission combined with rather weak obligations to give account. As the causes for such supposed deficits are reversible, depending simply on institutional reform, the specific problems are not relevant for discussing the structural deficits of cosmopolitan democracy.

To sum up, in general horizontal accountability in the EU is at least as strong as in national political systems. By closer inspection we might have to make two qualifications: the legislative powers of the European Council and of the Commission are insufficiently supervised by other institutions of the European political system. But no infeasibility of horizontal accountability in the EU is visible.

4.3 Horizontally Induced Constraints on Vertical Accountability

The third normative standard introduced above is about ‘horizontally induced constraints on vertical accountability’. These constraints can be seen as undesired side-effects (depending on the applied conception of democracy more or less desired) of effective horizontal accountability. Thus, the normative sense of this principle is analogous to the idea of efficiency: upholding a certain degree or level of effective horizontal accountability while minimizing the costs for democratic decision-making or rather the possibility of vertical accountability.

According to section 4.2, horizontal accountability is – with some minor qualifications – well established in the EU, while according to section 4.1, vertical accountability mechanisms are in deficit due to the weak link between the European citizenry and the EU institutions (and channels of political opinion and will formation). Now the general question is, if and to what extent the (weakly) authorized political actors have capacities to govern.

I suggest two criteria: The legislative costs for vetoing juridical and executive activism and the legislative costs for reforming existing law via vertically authorized mechanisms.

As neither the Council, nor the EP, nor any of the other actors discussed in section 1 can initialize secondary law-making, there is no way to reform ‘normal’ European law without the Commission’s explicit approval.

So there are two possible ways (replacing the Commission or replacing primary law via European Council and national ratifications). Both options are extremely costly with respect to the resources time and necessary political support. There is no doubt, even when compared to federal national political systems that it is much more difficult in the EU to reform past decisions and to correct undesired juridical and executive activism in decision-making (see Scharpf 2006, 2009).

4.4 Publicity/Transparency

The fourth and final normative principle of publicity has been introduced as a necessary pre-condition of accountability. There might be other reasons in favour of publicity, which I do not take into consideration here. So what are the appropriate criteria for accountability enhancing publicity in political systems?

To approximate to criteria it is helpful to distinguish between transparency and publicity: “Something would be considered transparent if it was published anywhere and no extraordinary costs and burdens have to be borne by ordinary citizens to get access to the information. But something is public only if, beyond this, almost everybody knows it, and almost everybody knows that everybody knows it” (Hüller 2007b: 566f.).

The appropriate criteria depend on whether we are concerned either with vertical accountability, where the European citizens are the ultimate ‘principal’ or if we are concerned with mechanisms of horizontal accountability, where political actors hold accountable other political actors. If professional political actors are expected to be the forum (in Bovens’ words), then transparency would suffice, as we expect professional actors to be aware of relevant and transparent political documents and processes. But if the egalitarian demos is expected to be the relevant forum, then actual publicity is a necessary condition for adequate or functioning accountability. Thus, if we are concerned with vertical accountability mechanisms, the appropriate criterion is ‘publicity’.

Transparency and publicity of what? The objects of transparency, relevant for vertical as well as horizontal processes of accountability are: relevant documents, policy proposals, as well as the general political agenda of relevant political institutions, and policy deliberations concerning relevant decisions.8

Beyond that, criteria for publicity enhancing vertical accountability have to vary from mechanism to mechanism. The criteria concern actual citizens’ knowledge of EU specific electoral alternatives, plausible understandings of choices in referenda as well as of successful ECI initiatives. In cases of direct citizen participation in European policy deliberation, we have to expect wide public knowledge about issues which are on the political agenda as well as egalitarian participation. In cases of indirect mechanisms national and CSO actors have to be consciously authorized or hold accountable for their positions in European politics.

The more these criteria are met, the more accountability enhancing transparency and publicity is in place. Thus, the general aim is to maximize these kinds of transparency/publicity. Nonetheless, as far as we have similar mechanisms on national level, we should ‘realistically’ only expect similar performances on European level.

In the previous decade the EU has made impressive progress in making the vertical mechanisms discussed above transparent. Elections, referenda and the ECI are perfectly transparent as are similar processes in the nation-state, and due to the reform of the consultation system ‘normal’ European policy making is visible to all interested parties. Even the Commission opened their documents and their policy related communications to external scrutiny (Hüller 2007b: 564ff.). The most important exemption is the European Council. But secret voting in the Council is obviously no structural necessity of European politics.

Almost perfect transparency does not pay off in publicity similar to national publicity of accountability related political affairs. There is a ‘structural’ (mechanisms transcending) deficit of publicity (see for EP elections: de Vreese 2006; for referenda: Hüller 2007a; for direct citizen participation Hüller 2010b; for CSO representation: Altides 2010; Hüller 2010a: part III).

Why do vertical accountability mechanisms have to face such a ‘structural’ publicity deficit in the EU? First and foremost it is plausible to assume a positive correlation between (expected) influence and public attention (the weaker influence via vertical mechanisms of accountability is, the lower the public attention for these institutions and processes is) and most (but not all) mechanisms are rather weak (compared to national mechanisms). Secondly the attention of national news media for EU issues is still in deficit. Even similarly important and transparent EU policy deliberations and conflicts find less and deficient attention in dominantly national media than national issues. Many important and controversial policy deliberations in the EU do not find significant attention in national newspapers at all (e.g. the deliberations on labour law reform in 2006) (see Altides 2010). Some issues find national media attraction but for usual with a remarkable delay. In Germany the anti-discrimination directive from 1999 was only debated hotly after 2004 at the dawn of national implementation. Even public discussion of the much criticised service directive started roughly a year after Frits Bolkestein, then the EU commissioner responsible for internal market regulations, presented his famous proposal in January 2004. In the Frankfurter Allgemeine Zeitung, a major German newspaper, 388 articles have dealt with the service directive since then, 288 of them were published in 2005 and 2006, but only 11 in the entire year of 2004. And if European issues advance to the public agenda (as the service directive did), the public communication remains nationally segmented with respect to the speakers, issues and transported political positions (Wessler et al. 2008). Third, the European consensus-style of communication as well as the compounded political arrangements negatively affects the simplicity of political alternatives, which in turn negatively affects the news value and in consequence visibility.

4.5 Persistent European deficits of democracy

In the comparison of the realization of these standards in the European and national political systems it has been shown, that except for horizontal accountability there is some empirical evidence for a European deficit in democratic accountability.

But should these empirical findings inform our normative reasoning about the desirability of EU or even global democracy? A direct deduction from empirical inadequacy would obviously lead to a naturalist fallacy. The ‘is’ would determine the ‘ought’. This kind of argument would not take seriously the prima-facie rationale for global democracy outlined in the introduction. But the relevance of the empirical knowledge would be of quite another kind if it can be shown that empirical deficits will safely survive over time, if we cannot have EU democracy in a quality as we can have democracy in the nation state. Thus, the question turns up, if these deficits are accidental, context-specific and alterable or if they are backed by more or less persistent causes.

As no relevant deficits have been detected for horizontal accountability, I only discuss the remaining three standards briefly: Firstly, there are many mechanisms of vertical accountability but none leads to effective egalitarian binding. It would be unfair to draw the conclusion of a structural down-graded political participation by taking first and foremost low electoral participation as a sufficient proof. Since the EP competences within the European political system are significantly below the competences within the EU member states it might be a sign of prudence not to vote. But there are participatory opportunities which obviously outweigh similar national procedures, as in case of the online consultations. There are no similarly inclusive measures giving a say for all citizens in normal lawmaking in none of the EU member states. No matter which issues are at stake in the consultations, there is a clear pattern of in-egalitarian participation within these consultations: there is more or less non-participation from new Eastern member states’ citizens and CSOs, low participation from Southern Europe, and more or less plural participation from Northern and Western Europe. There is no evidence for hoping that things will improve over time. There is a combination of endogenous variables which have a strong effect on transnational participation. Neither the types of participatory procedures nor specific national affectedness can explain the continuous result of in-egalitarian participation in the EU. Two possible explanations draw on the influence of national democratic cultures and on national economic prosperity. The influence of both factors is well documented in democratic theory.

Secondly, there is effective horizontal accountability in the EU, which is precisely the reason for the high degree of political ‘Verselbständigung’ of EU politics. In turn this constrains effective vertical accountability, and thus establishes a trade-off between functioning vertical or horizontal mechanisms. Parts of the legislative costs to veto juridical and executive activism as well as to democratically reform existing law certainly have EU specific causes, most important from a cosmopolitan perspective are the accumulation of competences in the European Commission and the exercise of unanimity rule in primary law making. Nevertheless some democratic defects are common to all systems with strong federal elements: The more levels have to be integrated in a political polity, the more effective egalitarian vertical accountability is at stake - ‘trapped’ by the requirement of multi-level approval and of a combined support by supranational and national majorities. This in fact leads to super-majoritarian rule and thus undermines effective equality in vertical accountability (see McGann 2006).

Third and finally the comparatively poor publicity of vertical mechanisms adds to the rather ‘bad’ accountability performance of the EU. Weak opportunities of political influence, a consensus-style within the European political system, and nationally segmented media combines for weak publicity of European politics. How persistent should we expect these causes to be? By institutional reform towards a more unitary and politicised European political system (Hix 2008), we might expect a significant increase of public attention to follow. There can be public discourses about EU issues similar to discourses about national issues and also the timing of public communication might improve. But in more then half of a century of European integration not just one relevant European newspaper or TV station came into being and there is no reason to assume the nationally segmented media organisation to transcend beyond the national and linguistic boundaries. The relevant public communicative processes for political opinion and will formation remain nationally segmented for the time being.

To sum up, the EU has been discussed as an ideal case of a forerunner of cosmopolitan democracy. But we have to acknowledge, that democratic quality in the EU is in deficit and this deficit is to some extent alterable but there are also persistent causes for a down-graded democratic quality in the EU.

5. EU Lessons for Cosmopolitan Democracy

This section extends the EU analysis to the ‘gradual infeasibilities’ of global democracy. Before doing this, I give three reasons why global institutions are empirically less democratic than the EU and why cosmopolitan democracy might be seen as unrealistic: First, almost all policy issues, which are dealt with at the European level are on the global political agenda as well. But when the same political actors deal with these issues in the EU in few organisational settings which are rather strongly integrated, the institutional frameworks of global politics show much more variety, combining more or less independent and policy specific inter-national organisations with various trans-national arrangements. Thus, the institutional anchors of ‘systemic’ democracy, as desired by cosmopolitan democrats, are more pronounced at the EU level. Such a systemic institutionalisation is a pre-condition for functioning representative democracy opposing the ‘natural’ political power game among states, political parties or interest groups. Cosmopolitan democrats tend to have a master plan for global democracy in hand, which does not go smoothly with incremental and fragmented activities of global reform.9

Second, the ‘democratic opportunity structures’ in the EU are empirically more advanced compared to the global level: there are European elections and there is a state-independent parliament, there are – despite all the empirical deficits – many channels for direct citizen participation (see above, section 4.1), which are completely inexistent at the global level. The systems of interest group involvement are comparatively more advanced at the European level as well. Public communication about EU issues goes well beyond event-driven and protest communication about global affairs. European issues of policy reform are significantly more visible in national media than similar global issues are (see Wessler et al. 2008). By contrast, on the global level we can presently only choose between minority government (when decisions are made by majority rule among nation-state representatives, who represent marginal national majorities) and politically ineffective unanimity rule (representing a marginal, but rigid absolute majority).

Third, the political course of action makes it much more ‘realistic’ to expect democratic politicisation at the European level (see e.g. Hix 2008) than the rise of global parties, the supply of egalitarian structures of global political will formation and so on or any kind of functioning institutional substitutes for them.

Nevertheless my main point is not that cosmopolitan democracy is unrealistic, utopian or infeasible to create. My main point is that such institutions at the global level would and can not live up to the normative principles of cosmopolitan democracy: If we would have EU-like institutions at global level, they would be less democratic than similar national arrangements and even than EU democracy.

5.1 Restrictions to effective egalitarian vertical accountability

We should expect more problems in effective egalitarian vertical accountability, even if there would be similar institutions as in the EU. First, we would have to ask who would speak for the citizenry of authoritarian regimes? Neither the citizens themselves, nor hardly existing civil society organisations nor the governments can democratically represent these citizens. As long as we live in a world of democratic and non-democratic societies, participatory structures at global level would fleece citizens from non-democratic countries.

Second, in a more ideal world (not coming into existence for the time being) we might expect global democracy to consist purely of democratic nation-states – as it is the case in the EU. Since the national histories of democratic experiences, national economic performances as well as the distribution of individual capabilities to function as democratic citizens are much more inegalitarian across the (democratic) world, we should also expect more inegalitarian global participation. To be clear, it is not just the case, that we have inegalitarian citizens and CSO participation in global politics (see e.g. Scholte 2011). The causes of the EU participation disease would be even more effective among the nation-states around the world. With respect to different variables of the Human Development Index we can see relevant differences among the EU societies, most are ranked ‘very high human development’ and few ‘high human development’. So even if the differences in societal endowments with politically relevant capabilities are moderate among EU member states (compared to the huge differences between the most distant societies in the world), political participation is rather unequal. Anyway if the differences are actually the cause of unequal participation in the EU, as long as most fundamental political capabilities are not assured broadly in a significant number of societies (as e.g. access to education, to pluralist information, to internet etc.) political participation at global level would be much more unequal than in the EU. The same is true for national participatory cultures. In consequence, similar democratic opportunity structures would create citizens and CSO participation which would be even more an elite phenomenon than it already is in the EU (see section 4.1 and 4.5).

So it should be a fair expectation, that the more extreme social inequalities in the world will survive institutional reforms towards cosmopolitan democracy and vertical mechanisms of accountability might be effective (in theory) but participation would be unequal (in practice).10

5.2 Horizontally induced constraints on vertical accountability

It has been argued above (section 4.5), that only more unitary and more politicised courses of European politics might overcome the defects of juridical and administrative capture as well as of non-solutions at the expense of existing majorities. The same problem is even more effective at the global level, as the ‘democratic’ decision-making costs in an extended multi level governance arrangement would increase.

If we now assume (for the sake of argument) a more unitary democratic arrangement at the global level, then the democratic weight lies predominantly on effective egalitarian vertical accountability. As just pointed out, vertical accountability is and will be more inegalitarian at the global level than in national and European politics. Additionally the idea of a unitary world state is also under normative attack far beyond the realists’ camp: Even if we would subscribe to the general ideal of global human rights and certain common global problems, it is quite a different thing to get consensus in many concrete cases, if and to what extent whose human rights have been violated (and by whom). A unitary global body would more or less automatically try to extent the boundaries of common global affairs, just as the European Commission, the ECJ and the EP are rightly ‘accused’ to promote European integration. And without the national and regional safeguards of federalism an expanding global polity might rapidly be experienced as ‘tyrannical’ (e.g. Rawls 1999).

5.3 More invisible global politics

Concerning transparency and publicity the persistent visibility problems of global politics are to a great extent similar to the EU. Even if the issues of public discourse are (or: would be) supranationalised, the speakers, the media and the political positions remain national ones. In consequence, relevant parts of national communications about EU issues are invisible for nationally segmented audiences outside these countries. If we now assume the political cleavages across the world to be more diverse than among EU societies and if we assume similar kinds of nationally segmented public discourses, the part of the relevant public communication about global affairs, which remains invisible to national audiences would increase.

To give an example: public communication (arguments, positions etc.) about the legitimacy of military interventions (in Iraq, Kosovo etc.) differ significantly between nationally segmented public discourses in countries as Germany, France and the UK (see Wessler et al 2008: 107ff.). It should be clear that the ideological distances between national public opinions would increase, if we would add Arab countries to the sample. Thus, the invisibility of relevant positions and arguments would increase in nationally segmented communications about global affairs.

National segmentation of public discourses should also be expected to be a restriction for the acceptance of global policy making for at least two reasons: In national discourses among national speakers about European politics, ‘blaming Brussels’ is a prudent and actually existing strategy to attribute responsibility, when unpopular policies get under public scrutiny (see Gerhards et al 2009). The same mechanism should be expected to occur in cases of contested global policies. In consequence, more public communication about EU and global affairs in nationally segmented public spheres might constrain instead of foster cosmopolitan ‘informed loyalty’.

For more representative modes of political will formation, transparency should be seen as a sufficient normative aim. The democratic functioning of ‘mere’ transparency depends on attentive ‘representative’ actors as CSOs, political parties and media securing that relevant information meet the eye of the public or are treated in inclusive policy deliberations. This democratic functioning of transparency now depends on extensive attention of these representatives. The expectable North-South divide in attending transparent policy arenas (see above, 5.1) would have consequences for the appropriateness of transparency: full transparency of documents, political processes etc. would enable in-equal effective chances to influence global policy making.

6. Conclusion

If global political institutions would be designed the way cosmopolitan democrats want them to be, we would have an EU-like institutional design at global level (without its alterable democratic flaws). Such fictitious global democratic procedures would – for persistent empirical reasons – be significantly less democratic than similar processes within the nation-state: vertical accountability is either more ineffective or more inegalitarian or both as compared mechanisms at nation-state level. Neither unitary nor federal systems should be seen as a plausible solution for the threat of ‘Verselbständigung’ caused by multi-level politics. And nationally segmented public spheres will not promote a similar type of politicised discourse around ‘common’ global issues. Normal global policy making (just as EU policy making) would have a higher chance not to be publicly discussed at all and if discussion occurs, then in nationally segmented ways. To sum up: The global reality of cosmopolitan democracy would be a downgraded institutionalization of the normative standards of democratic accountability.

This criticism goes as far as it does. In a comparative perspective the critique is strong: if and to the extent global institutions imitate European democratic settings, just like cosmopolitan democrats have proposed (see section 1), these global settings would be significantly less democratic than similar settings at European level, which are significantly less democratic than national ones.

But to do justice to the Cosmopolitan’s claim, two qualifications have to be made: Firstly, the argument obviously rests on my ‘reverse-Sinatra-assumption’ outlined and defended in section 1: If cosmopolitan institutions can’t make it in the EU, cosmopolitan institutions can’t make it anywhere (beyond the nation state). Secondly, it is not enough to prove comparative deficits: if the initial rationale to advance cosmopolitan democracy is the growing political incongruence caused by denationalisation, then the fact of a downgraded global democracy does not automatically blow away the normative ideal of cosmopolitan democracy. Thus, my criticism is less straight forward: There is obviously a trade-off in democratic quality by ‘scaling up’ democracy, between the aims of ‘catching’ affected parties in global political decision-making (fostering the cosmopolitan ideal) and maintaining the effectiveness of accustomed democratic standards (challenging the institutional vision of the cosmopolitan ideal). To put it differently: Cosmopolitan democracy is not a perfect solution for growing political incongruence due to societal denationalisation.

Therefore we have to comparatively scrutinize different possible second-best solutions: The ultimate normative concern shared among normative cosmopolitans is to establish institutional settings within and beyond the nation state which can be expected to maximize democracy in a multi-level perspective. In accordance with Zürn (2011), there are at least four kinds of normative institutional visions in the field of international political theory. But there is no thorough (counterfactual) empirical assessment yet which of these different institutional designs would maximize democratic quality in the face of persistent circumstances of global politics. The article just proofed the imperfection of cosmopolitan democracy, but did not assess which institutional design would make for a second-best ideal. Because the other models do not aspire the institutional ideal of EU democracy, it does not make sense to assess their democratic potential against the background of the EU’s best possible practice.


  • 1

     A first version of the paper was presented at the Kick-Off Workshop of the Research Program ‘Global Democratic Governance’“The Causes, Consequences, and Democratic Legitimacy of International Institutions” at University of St. Gallen, June 9–10, 2011. I am deeply grateful to the participants for a stimulating discussion. For very helpful written comments on earlier versions of the article I would like to thank Andreas von Staden and the anonymous reviewers of SPSR.

  • 2

     Compared to international negotiations, multisectoral networks might actually be more inclusive and effective in certain policy fields at global level. But they are not part of the cosmopolitan democrats’ desired institutional design, because these networks can only meet standards of plurality but not of democracy.

  • 3

     Is ‘nonviolence’ really a normative principle which indicates the democratic quality of a political system? A principle of nonviolence risks to level potential differences between all forms of (legitimate and non-legitimate) violence. This is far from being a consented position in democratic theory. Above this, the focus on nonviolence might also exceed the core of democracy. But also the ‘operationalization’ of the two undisputed principles is far from being without ambiguity: How do we know it is ‘popular control’ or ‘political equality’ when we see them? There is intense debate on the many options of the precise meanings of both principles. But in this point Archibugi’s conception is short of accuracy.

  • 4

     To participate or not to participate in political affairs is to some extent a question of personal ‘taste’. It is plausible to assume a positive correlation between participation and political influence. The idea of maximizing equality would constraint per se types of political participation, which are not conducted in an egalitarian way. As we know from empirical research, the more demanding political participation is, the more unequal are the patterns of participation. In the consequence the idea of maximizing political equality (alone) would put a relative premium on the attitudes of non-political actors.

  • 5

     The methodological considerations on doing this kind of normative theorizing adequately are skipped here, but briefly outlined in Hüller (2010a: chap. 3). The main idea is, normative theorizing in the face of reasonable normative pluralism should aim at avoiding or minimizing reasonable rejection by an open normative definition as well as by restricting the empirical analysis to unproblematic cases. If (or to the extent) both is unattainable, the empirical research has to investigate their data against plural standards.

  • 6

     Maybe somewhat controversial is the background assumption that the European Union has to comply with similar normative standards of democracy as the nation-state. I have discussed the ‘ought’ and ‘can’ of EU democracy elsewhere (Hüller 2010a: chap. 2).

  • 7

     The exemption from supranationality in this case is the European Council’s option to recall the ‘High Representative of the Union for Foreign Affairs and Security Policy’, and the elections of the commissioners.

  • 8

     A contested issue here is the transparency of representatives’ voting behaviour. On the one hand secret voting is dysfunctional for accountability, on the other hand might secrecy be a pre-condition for desired independence of EU actors from national constituencies as well as a supportive for Europeanized considerations in political negotiations.

  • 9

     From a more realist perspective, we would have to analyse more closely the democratic potentials to reform existing international organisations as well as how to overcome specific fragmentations of world politics. Nevertheless discussing such realist options would not do justice to the integrative expectations of cosmopolitan democracy.

  • 10

     It might be added, that programs of affirmative action to increase egalitarian participation have structural defects. Paying for participation by the ‘state’ creates fears of capture, so that funded CSO do not speak for the people but for the funding global institutions. And such funding always creates an incentive to invent ‘concernments’. And constructed modes of egalitarian participation (like in deliberative polls) beyond the local level are normally hidden in transparency and politically ineffective (for the EU, see Hüller 2010b).

Thorsten Hüller is research fellow at the Collaborative Research Centre 597 ‘Transformations of the State’ at the University of Bremen. His research is mainly on normative and empirical questions of democracy and on the European Union. His most recent book is Demokratie und Sozialregulierung in Europa (Campus 2010). His articles have appeared in Acta Politica, Journal of European Public Policy, Swiss Political Science Review and Zeitschrift für Politikwissenschaft. Address for correspondence: Collaborative Research Centre 597 “Transformations of the State”, University of Bremen, P.O.Box 330 440, D-28334 Bremen, Phone: +49 (0)421 218 – 56680/Fax: +49 (0)421 218 – 87 21, Email: thorsten.hueller@sfb597.uni-bremen.de.