In search of democratic agency in deliberative governance

Introduction

Many scholars and practitioners today agree that regional and global governance institutions suffer from a ‘democratic deficit’ and that prospects for democracy beyond the nation-state must be addressed in this context. If democratic theory took a deliberative turn in the 1990s, deliberative democracy has taken a ‘civil society turn’ to address these shortcomings on the regional and global levels. In light of the present circumstances of world politics, consisting of an increasing asymmetry between rule-makers and rule-takers and inequalities among states, many deliberative democrats investigate the role of transnational non-state actors — ranging from social movements to interest groups and non-governmental organizations (NGOs) — for achieving more transnational or global democracy. Instead of emphasizing juridical aspects, this deliberative civil society or stakeholder view lays stress upon the core democratic qualities or mechanisms of participation, accountability, authorization and deliberation. It is argued that civil society offers a rich soil for reformulating democracy globally since it is inhabited by a growing range of social actors that create new political spaces, which are not delimited by territorial nation-state borders and therefore are more suitable for confronting the globalized political problems that we face today.

This article engages in the debate on democracy as a normative ideal (i.e. ‘the rule by the people’) applied beyond the nation-state, with particular focus on democratic agency. It investigates the notion of democratic agency built into this deliberative civil society view with regard to its democratic qualities. The defended thesis is that the proposed political agent, often conceptualized in terms of ‘stakeholder’, is not equipped to be a democratic agent insofar as the civil society view does not fulfil two basic requirements for an arrangement to qualify as minimally democratic, namely, political equality and political bindingness. It is concluded that insofar as we wish to hold on to a deliberative conception of democracy, something along the lines of Habermas’s two-track view is still our best bet for accommodating these two conditions, even in a transnational context, since it is able to avoid the problems connected with the separability premise. ¹ It is important to note that, since the article examines the question of democratic agency, the primary question of concern is which basic conditions must be fulfilled in order for an arrangement to qualify as minimally democratic, not the question of the justification of democratic borders. While the two questions are intimately related, it is possible to bracket the latter for the argument pursued. However, to analytically separate these two questions is a complex and delicate matter to which we will return.

The argument proceeds in four steps. First, the article sketches the general features of the civil society approach and the presumptions made about democratic agency. Although this approach is rich and embraces numerous different perspectives, the purpose in this section is to illustrate its general characteristics mainly through the work of Terry Macdonald, who offers one of the most theoretically and conceptually sophisticated versions of this view in the present debate, and John Dryzek, who has been a major source of inspiration for the approach as a whole (Section 1). In a second step, I unfold a common feature underlying this view of democracy, what I call the ‘separability premise’. In brief, this premise holds that it is possible to define democracy as consisting of two or more separable democratic qualities or mechanisms — most importantly, inclusive participation, accountability, authorization and deliberation — and that democracy increases the more one or more of these are strengthened. Thereafter, the separability premise is problematized in an effort to draw out the implications for democratic agency. This is done in light of a theoretical framework extracted from the basic requirements of political equality and political bindingness. For political agents to count as democratic agents within this framework, they must be actual agents with equal influence over the decision-making and in the shaping of the common institutions (Section 2). The third section treats a possible objection to this critique, pertaining to an alleged inseparability of the two questions of the justification of democratic borders and the basic requirements of democracy (Section 3). Finally, the article concludes that if we agree with civil society scholars on the strength of a deliberative approach, something along the lines of Habermas’s two-track view of democracy is more appropriate, even in an international context, since it is able to avoid the problems connected with the separability premise. However, it is argued that if we stress democratic law-making in connection with interdependent interests and roughly equal stakes instead of ‘civic solidarity’, as does Habermas, we are able to draw a less pessimistic conclusion about the future of democracy at the global level (Section 4).

Democratic agency in the deliberative civil society approach

In theorizing about how to re-establish a symmetry between rule-makers and rule-takers in global political decision-making, proponents of what I call the deliberative civil society view commonly claim that cosmopolitan theorists keep too much of the Westphalian conception of the state in the translation from nation-state to global democracy. Even if cosmopolitans attempt to rethink sovereignty in functional rather than territorial terms, they still emphasize electoral representation and focus on the juridicalization of international organizations (IOs) through some idea of an overarching cosmopolitan law (Archibugi, 2000, 2002; Held, 2002: 32). Being sceptical of the import of these ‘Westphalian’ features into global politics, the civil society approach wishes instead to enhance democracy in transnational and global decision-making through the increased involvement of transnational non-state actors. Most importantly, such actors are supposed to represent (in a non-electoral way) or speak for marginalized groups (Scholte, 2005). In other words, the representation of political subjects and the expression of stakeholder concerns are made through ‘voice’ rather than vote (Charnovitz, 2006; Keck, 2004; Peruzzotti, 2006; van Rooy, 2004).

Macdonald offers the theoretically most sophisticated and detailed version of this view. She argues that we have to abandon the traditional idea that democracy must take place within a ‘closed’ society if we aim to globalize democracy beyond the nation-state. Under non-ideal conditions, it is suggested instead that a liberal democratic world order ought to be composed of multiple agents of public power held to account by their multiple overlapping stakeholder communities. Even though NGOs are the kind of agent generally discussed by Macdonald, it is argued that the model is applicable to other agents too, such as IOs and transnational corporations (Macdonald and Macdonald, 2010). Against the tendency of using the term ‘stakeholder’ broadly, Macdonald takes a more restricted view in the specification of the stakeholder as a democratic agent. In brief, the individuals who have a relevant interest or ‘stake’ in a decision are those who are subject to problematic impacts on their autonomous capacities. With stakeholders as basic building blocks, she outlines a multi-stakeholder model that in her view has the potential of being applied within a global polity without a need for either formal electoral mechanisms or the establishment of state-like structures of global public power (Macdonald, 2008: 192).

Further, it is a deliberative model, such that the representatives of multiple stakeholder constituencies are required to deliberate among themselves and reach consensus on a final decision. This deliberative decision procedure is underpinned by a ‘dualist’ conception of equality. In the first instance, stakeholders should be accorded equal opportunities to identify the interests that are supposed to be represented in a deliberative decision process. Second, since these stakeholder interests are not aggregated to reach a decision, as is the case in traditional nation-state models, but rather are advanced by stakeholder representatives, they must be accorded equal consideration by these representatives in the deliberative process (2008: 143). Indeed, Macdonald admits that her deliberative stakeholder model has practical limitations, since it does not incorporate any aggregative procedures for reaching decisions. Therefore, she argues, we most likely would have to employ a hybrid representative model in global politics, in which deliberation among multi-stakeholder representatives is complemented with traditional aggregation among state representatives in cases where consensus through deliberation cannot be reached (2008: 162).

The central question for our purposes is what makes this non-electoral deliberative model democratic and, accordingly, its members (i.e. stakeholders) democratic agents. Through a thorough going analysis of the normative function of elections in traditional models of democracy, Macdonald infers that the main reason why elections have been so attractive is that they can ‘provide stakeholders with a degree of political control over their public political representatives’ and as such function as a mechanism for delivering legitimate representative agency (2008: 170). Nevertheless, she argues, they are not the only effective mechanisms for delivering such control. It is possible to provide alternative non-electoral mechanisms that are able to fulfil equivalent normative functions. In Macdonald’s view, the two mechanisms through which elections deliver political control to stakeholders are authorization and accountability (2008: 171). However, in contrast to much contemporary literature on global governance, which tends to offer non-democratic justifications of these two mechanisms (e.g. Keohane, 1984; Rosenau, 1995, 2000), Macdonald stresses the characteristics that are required of these mechanisms to deliver democratic legitimacy.

Let me briefly illustrate what she has in mind here. Concerning authorization, two distinct elements are required: mechanisms of delegation, for specifying the public political tasks that the representatives are entitled to perform; and mechanisms of empowerment, for according them the appropriate capacity to do so effectively (2008: 180–185). While the latter mechanisms have to do with the NGO gaining trust from other political agents as well as resources, most notably from donors, the former concern delineating appropriate NGO responsibilities, either in terms of broader ‘constitutional’ responsibilities accorded to NGOs through general codes of conduct or in terms of specific policy-relevant mandates, such as ‘stakeholder signalling mechanisms’, through which stakeholders can specify tasks for which they are delegating entitlements to NGOs (2008: 196). Many development NGOs have incorporated processes for consulting stakeholders in the decision-making through the use of participatory methods such as Participatory Rural Appraisal (PRA) and Participatory Learning and Action (PLA). Macdonald illustrates this with Oxfam’s (UK) recent reforms to ensure that its trustees make decisions that are influenced by the views of its stakeholders through a stakeholder ‘assembly’, consisting of around 200 individuals deemed to be representative of the stakeholder community at large, and through regular ‘Stakeholder Surveys’ to identify stakeholder opinions (2008: 196–197).

Similar to authorization, accountability too is constituted by two elements: mechanisms of transparency, for transparently delineating public political roles; and mechanisms of disempowerment, for imposing sanctions that annul certain political resources that enable an actor to perform public political functions (2008: 185–190). While transparency mechanisms in a global political context usually take the form of ‘codes of conduct’, which assist in demarcating transparently the responsibilities of particular NGOs by codifying them within some formalized international charter, mechanisms of disempowerment are similar to mechanisms of empowerment in the sense that social actors are able to disempower NGOs by annulling the resources that enable them to perform public political functions. Stakeholders need not directly participate in this; rather, it can be done ‘by any agents’, as long as it is done in accordance with stakeholder signals (2008: 212–215).

Concerning the relationship between mechanisms of authorization and accountability, Macdonald claims that they are best understood as mutually complementary, since each can ‘operate effectively without the other, conferring democratic legitimacy on public political agents’. Thus, in line with the separability premise discussed below, they can generate some degree of democratic representation and thus legitimation by themselves (2008: 191).

Similar to Macdonald, but with even more emphasis on the deliberative practices of argumentation, Dryzek draws inspiration from the normative force of the all-affected principle for achieving global democracy, that all those relevantly affected by a decision should participate in its making. However, in an article co-written with Simon Niemeyer, he expresses scepticism about the direct application of this principle to the global level, as suggested by cosmopolitans, since the deliberative participation of all affected by a collective decision is infeasible. Since authority is increasingly escaping the boundaries of a well-bounded demos, Dryzek and Niemeyer wish to steer away from the traditional idea of representation defined as the substantive acting for physical others, towards what they call discursive representation, that is, the substantive acting for others’ arguments via the representation of relevant discourses (Dryzek and Niemeyer, 2008: 481). For the present purposes there is no need to go into detail about what is meant by discourse or how these discourses are supposed to be represented in what the authors refer to as the ‘Chamber of Discourses’. It is more important to look at the general democratic features of their proposal, in order to locate the democratic agent.

According to Dryzek and Niemeyer, the priority of discourses has several democratic advantages. To begin with, when representing arguments, proportionality becomes irrelevant. It would even be rational to include a vantage point to which nobody subscribed, since all relevant discourses should be represented for policy rationality. They claim that there are strong moral reasons for discursive representation too. The common liberal argument for representing individuals draws its normative force from the idea of an autonomous subject capable of self-government. But this view is fully compatible with the idea of representing discourses rather than individuals to the extent that we do not assign discourses an independent moral standing but see them as reducible to the individuals who subscribe to them. Dryzek and Niemeyer argue that this in fact has the advantage of taking seriously the multiple characters of persons who engage in numerous different discourses, rather than viewing them as simple unified wholes. To represent an individual as a whole on this account would thus mean to represent all his or her discourses (Dryzek and Niemeyer, 2008: 482–483).

From this discursive standpoint, Dryzek has recently proposed three ways of thinking about deliberative democracy and democratic agency in global politics: in terms of a soup, a society or a system. In his view, there are certain deliberative practices that could contribute more democracy on the global level. Apart from representation from NGOs within communities of stakeholders, theorized by Macdonald, Dryzek stresses the importance of practices such as transnational social movement activism, self-appointed unelected popular representation, and deliberation within international negotiations such as the G20 or the World Economic Forum (Dryzek, 2011: 216). According to the soup ‘model’, these practices are ingredients of a ‘democratic soup’ insofar as they are of the right proportion to strengthen democratic norms such as authorization, accountability, deliberation and participation.

Drawing on the concept of the international society, associated with the English School of International Relations, Dryzek further describes the society ‘model’ as showing democratic promise on the global level since it involves norms that regulate the interactions and activities of relevant members. In his view, the constitutional norms and fundamental institutions of international politics evolve not only as a result of strategic action, but also involve communicative action. Similar to recent constructivist proposals in International Relations theory, he points to the global spread of democratic norms of authorization, accountability, deliberation and participation through communicative action, the latter of which both enables and constrains political action. On this society model, international politics is to a large extent a struggle between discourses and what Dryzek calls ‘discursive democracy’ increases to the degree that multiple discourses compete for attention and are subject to critical, inclusive and competent control in transnational public spheres, exercised by actors such as stakeholders, NGOs, social movements and national governments. Since international politics is weak on formal authority structures, Dryzek concludes, such discourses ought to be central to democratization (Dryzek, 2011: 220–222).

The system ‘model’, finally, is conceptualized in terms of a public space, in which free communicative action occurs (not necessarily tied to the exercise of political authority), and an empowered space, in which authoritative collective outcomes are generated (even without clear moments of binding decision). Accountability on this view is understood in terms of the empowered space being answerable to the deliberative public space (Dryzek, 2011: 225–226).

The next section takes a critical look at who is supposed to be the democratic agent in these deliberative civil society proposals, and what makes him or her a democratic agent.

In search of democratic agency

Insofar as one defends democracy as the ultimate foundation of legitimate political authority and thus has democracy as a normative ideal of self-determination in mind, in line with the civil society approach, one has to specify in more detail what makes this self-determination democratic. For only by doing so can we identify what constitutes the basic elements of democratic agency, namely, the specific democratic qualities of those included. In this section, I argue that two conditions are conceptually necessary for democracy and normatively required for an arrangement to qualify as minimally democratic, both of which are fairly uncontroversial and weak enough to be accommodated by a wide range of different normative democratic theories. ²

What seems indisputable is that democracy harbours the basic idea of equal political power, often expressed in terms of an equal distribution of influence over the decision-making. More specifically, at least all persons who are significantly affected by a political decision (or law) should have such an influence. If we unpack this idea, it accommodates two conditions of fundamental importance. The first is political equality. What distinguishes democracy from other forms of government, such as dictatorship, monarchy, or aristocracy, is that it has components that express and secure some form of political equality. While equality plays an important role in democracy in several respects (e.g. in terms of equal respect or equal concern for everyone’s interest), what is of concern here is a specific conception of equality, according to which anyone who is significantly affected by a political decision (or law) has an equal opportunity, secured through an equal right, to participate (directly or indirectly) in the decision-making about it (Christiano, 1996). ³

But apart from this ‘deontological’ dimension of being given an equal opportunity of participation in the decision-making procedure, equal influence also involves a ‘teleological’ dimension, in that people rule over themselves and shape their institutions only if they act politically by ‘exercising’ their political equality. I call this condition ‘political bindingness’.

Let me dwell a little bit more on these two conditions to develop the framework for analysing democratic agency. To begin with, it would be odd to argue that democratic self-determination is premised on a system of elections connected to a unified territory, conceptually speaking, since we could reasonably call a small group of people democratic, a small political organization if you will, which fulfilled political equality and political bindingness. It seems to be primarily an empirical, not a conceptual, question whether democracy is best realized within a territory or not. However, even if the territorial boundary is not the kind of boundary on which democracy relies, democracy is not borderless. It presupposes a particular kind of boundary (or boundaries if we refer to a multilayered system) within which people have the equal opportunity of participating in egalitarian decision-making and bind themselves to political authority as equals. In other words, decision-making or legislation has an inevitable spatial dimension attached to it (Arendt, 2005: 189–190; Lindahl, 2010). ⁴

The reason both conditions must be fulfilled for an arrangement to qualify as minimally democratic, that is, why political equality cannot stand alone, is because it would be peculiar to call a political arrangement democratic in which all those that had the equal opportunity to participate never did. Democratic law- and decision-making draws its legitimating force, not only from equal agency, but also from actual agency, in that democratic legitimacy is premised on the members’ exercise of their political liberties. Only in this way could they indirectly rule over themselves (‘rule by the people’). Indeed, this should not be interpreted in terms of obligations, either as an individual obligation in the Kantian sense, or as a collective obligation to participate in a general or common will in the Rousseauian sense. Further, the condition does not presuppose moralized political action in terms of doing what is right. Rather, bindingness implies that people can rule over themselves only if they have an actual influence in the decision-making, through robust participation in the decision procedures by a major part of the members, or at a minimum by accepting these procedures as valid, without which the right to participate would not have any binding force. ⁵ Built into this condition is also a requirement of ‘positive responsiveness’, presuming that the more people supporting a proposition, the more likely it is to become law (Goodin, 2004; Goodin and List, 2006). So, while the condition of political bindingness itself does not specify a threshold of participation in democratic decision-making — indeed, this obviously varies between different democratic theories — it presupposes that there is such a threshold (Erman, 2011).

Now, let us return to the stakeholders of Macdonald’s model and examine their qualities as democratic agents in light of this framework. I cannot see how these conditions are able to make democratic agents out of stakeholders. Concerning political equality, nowhere do these institutional structures secure for stakeholders the equal opportunity to participate in egalitarian decision-making. Recall that this deliberative model lodges a ‘dualist’ conception of equality, according to which stakeholders have equal opportunities to identify the interests that are supposed to be represented by relevant NGOs in the deliberative decision process (for the sake of simplicity, let us call this ‘the equal opportunities condition’), on the one hand, and stakeholder representatives are then required to accord equal consideration to these interests in this process (henceforth ‘the equal consideration condition’), on the other. The problem from the standpoint of democracy is that the authorization of an authority or a political agent is not primarily about the equal possibility of identifying the interest that one wishes to have represented (equal opportunities condition). More importantly, it is to have the equal opportunity to approve of this authority by participating in egalitarian decision-making, thereby accepting its political decisions and laws as binding.

For sure, the equal consideration condition is not able to satisfy this condition either. While the equal opportunities condition concerns stakeholders’ possibility of defining interests to be represented, the equal consideration condition does not involve stakeholders at all, but the equal consideration of their interests by NGO representatives, that is, that decisions are made through procedures that they have ‘equally good reasons to accept’ (Macdonald, 2008: 150). This means that the equal consideration condition is solely tied to some kind of hypothetical consent and as such does not necessitate any political action whatsoever. So, while both the equal opportunities condition and the equal consideration condition might play important normative roles in a democratic theory, they do not connect conceptually and normatively to authorization and cannot therefore replace political equality as specified by the present conceptual framework.

To further unpack the problems of political bindingness and political equality, let us revisit the examples above. Concerning authorization, as we have seen, the most important mechanisms for empowerment are trust from other political actors and donations to have the appropriate resources for political action (2008: 204). But if empowerment is to constitute part of the process of authorization from a democratic point of view, it is the stakeholders who are supposed to empower the NGOs, not other actors. Although Macdonald admits that the connection between the stakeholders and the other actors involved with NGOs ‘is not so straightforward’, it is difficult to see how political equality and bindingness could be fulfilled in this process of empowerment even if there was some kind of ‘indirect mandate’ involved of the kind Macdonald has in mind (2008: 207). Similarly, even if mechanisms of delegation through general codes of conduct had been developed by involved stakeholders (and thus constituted ‘fully democratic mandates’, in Macdonald’s words) rather than by UN agencies, states and NGOs, which is presently the case, what makes them part of an authorization lending NGOs democratic legitimacy? Again, for this act of authorization to be democratic, these stakeholders must not only be involved in developing such codes of conduct, but must have an equal opportunity to participate in the decision-making about them as well as actual influence. Thus, against Macdonald, who argues that these non-electoral mechanisms can serve the same normative purpose as elections by way of providing stakeholders with a degree of political control over their public political representatives, I doubt that this is a political control of the right kind, since it has very little to do with democratic self-determination. ⁶

The same problems arise when we look at accountability. Like empowerment, disempowerment from a democratic point of view does not primarily mean the removal of resources that enable representatives to act politically, as suggested by Macdonald. It means to remove them entirely from that particular political position. What is more, this political act cannot be done by anyone; rather, it comes about through a decision-making process in which stakeholders have the equal opportunity to force the representatives to leave office and replace them in a common act of bindingness. And transparency faces similar problems as delegation, since it is supposed to be established ‘democratically’ by codes of conduct that codify NGO responsibilities within some international charter.

Whilst Dryzek’s deliberative proposal is much less detailed, similar problems appear. First, concerning the idea that discourses rather than individuals should be represented in global governance, it is far from clear how the inclusion of every possible argument makes the political decisions binding to democratic subjects. The moment of factual (not hypothetical) political action, through which a constituency approves of a political authority, is missing. While an important epistemic dimension is expressed by the inclusion of marginalized voices and of all possible arguments, such a dimension cannot alone fill the gap between the citizenry and the political authority in order to generate democratic legitimacy. When we say that democracy is ‘government for the people’, we mean that it exists for the sake of the people and rules in the interest of the governed. But this is only half the story about democracy. Even a compassionate dictator could rule with the interest or ‘voices’ of the governed at heart. Making democracy into a matter of satisfying people’s interests or representing their arguments and voices will not suffice to capture the ‘by’ in ‘the rule by the people’ (Erman, 2010; Rostboll, 2008: 45–77). This ‘by’ cannot be conceptualized without political equality and political bindingness, which turn people into democratic agents and thus into democratic rule-makers.

Second, the three ‘models’ of democracy proposed by Dryzek all rely on the presumption that a few ingredients — consisting of deliberative practices such as representation from NGOs within transnational communities of stakeholders, deliberation within international negotiations, or self-appointed unelected popular representation — lead to increased democracy insofar as they strengthen the democratic norms of authorization, accountability, participation and deliberation. Certainly, Dryzek acknowledges that none of these deliberative practices offers in itself the key to global democracy and that they only strengthen these norms under certain circumstances. Moreover, many of the practices are hard to judge in isolation, because whether their effects are good or bad for democracy depends on other external factors (Dryzek, 2011: 216). Nevertheless, none of the models take into account the conceptual and normative relationship between those very democratic norms. It is true that the system model does stress ‘the need to trace connections’ from the public space (in which deliberation and participation take place) to the empowered space (in which authorization and accountability take place), but it is far from clear what kind of connection would count as sufficient for increased democracy (Dryzek, 2011: 232). Arguably, not all kinds of ‘consequential effects’ would count as desirable from the point of view of democracy. We will have reason to return to this below when I discuss the relationship between the formal and informal ‘tracks’ of the Habermasian deliberative model.

It is possible to get a more systematic grasp of the problems that have been identified in the deliberative civil society approach pertaining to democratic agency, and perhaps even one explanation for why they occur, if we take a closer look at the underlying separability premise. Recall that this premise accommodates the presumption that democracy consists of two or more separable core democratic qualities, norms or mechanisms, and that democracy increases the more one or more of these are strengthened. The term ‘separability’ refers to the idea that they are conceptually and normatively detachable in two important ways: first, they potentially involve and bind different subjects; and, second, it is possible to determine the increase of one democratic quality, norm or mechanism independent of the others. The key mechanisms of Macdonald’s model are accountability and authorization, which supposedly are equipped to do the normative work of conferring democratic legitimacy effectively independently of each other (Erman, 2008, 2010). Apart from accountability and authorization, Dryzek also theorizes participation and deliberative practices among actors such as stakeholders, NGOs, protesters and governments. The problem with the separability premise, or so I argue, is that if we detach accountability, authorization, deliberation and participation in this way, we cannot account for ‘the rule by the people’. For in order to do so, these democratic mechanisms or norms must be knit to each other and to the same actors through political equality and political bindingness.

Indeed, Macdonald admits that accountability and authorization constitute different stages of the representative process, where authorization precedes and initiates representation, while accountability follows and terminates it. However, they not only constitute two temporal stages but are conceptually and normatively related in relevant ways. A citizenry, the group of relevant stakeholders, binds itself to an authority (or agent) through an act of authorization via egalitarian decision procedures. The demand for accountability emerges through this very act of legitimation. If the authority is successful in being accountable, this accountability is directed back at the very same subjects (even if others would benefit from it as well). ⁷ Similarly, if it is unsuccessful, this gives the subjects good reasons to de-authorize it and re-authorize another. If we hold on to the separability premise and conceptualize authorization and accountability as separate mechanisms, however, this essential normative two-way relation would get lost.

Dryzek is even further from capturing this two-way relation, since he does not offer any account (not even a temporal one) of the relationship between the democratic norms of authorization, accountability, deliberation and participation. But as I argued above, we cannot draw any conclusions about increased democracy solely on the basis of the strengthening of each of these norms in isolation. While I agree that communicative action both enables and constrains political action and that constitutional norms that regulate international interaction evolve partly as a result of communicative action, this in itself does not tell us very much about increased democracy. It is unclear what reasons there are for thinking that central to democratization is an increasingly inclusive representation of multiple discourses exercised by stakeholders, together with NGOs, social movements, governments, and so on. For this seems to entail an arbitrary and unequal influence over the decision-making. By contrast, this is prohibited by the requirements of political equality and political bindingness, through which relevant actors have equal influence in the empowered space, and deliberation and participation in the public space have an impact on the very same decision-making processes in the empowered space. In fact, only when the actors involved have equal influence in this respect are they able to authorize the empowered space, in turn making this space accountable, not only to the public space in the abstract but to the very democratic subjects located in it.

Once we remove the separability premise, we see that accountability and authorization, as much as deliberation and participation, must be understood as parts of a conceptual and normative package. Thus, we cannot draw any conclusions about increased democracy through increased accountability independent of authorization, because they hang together and involve the same subjects (Erman, 2010; see also Follesdal, 2011). There are of course numerous ways for authorities and agents to be accountable in politics, but increased accountability without any authorization would not be democratic accountability (Erman, 2006; see also Buchanan and Keohane, 2006; Grant and Keohane, 2005).

A possible objection

Now, before concluding with some general remarks on deliberative democracy and democratic agency in a transnational context, let me respond to a possible objection to the proposed critique of the civil society approach. It concerns the methodology applied, where I separate analytically the question of the justification of democratic borders (henceforth ‘the boundary question’) from the question of which conditions must be fulfilled in order for an arrangement to qualify as minimally democratic (henceforth ‘the basic requirements question’), bracketing the former while directing focus on the latter. Indeed, in a sense they bring to the fore different aspects of agency: the boundary question has to do with who should be included in a democratic arrangement and the basic requirements question has to do with which conditions must be fulfilled for it to be a democratic arrangement in which those included become democratic agents.

Macdonald might reply that these questions on her account are inseparable, and that by bracketing the boundary question one simply cannot understand her answer to the basic requirements question. In other words, only by taking into consideration her particular version of the all-affected principle (i.e. those who have a relevant interest or ‘stake’ in a decision), which is the criterion used to demarcate stakeholder constituencies, is it possible to make sense of her defence of the non-electoral mechanisms of authorization and accountability, which if fulfilled are supposed to make these arrangements democratic and turn the stakeholders into democratic agents. For Macdonald, the boundary question is not only the first and most basic question of democratic theory, it is also the first question of democracy in the world that we confront under existing non-ideal political circumstances, in which the boundaries of democratic decision-making processes will be pluralist in structure.

To begin with, I see no reason to demand that democratic theory must always answer both questions in order to do some proper normative work. Rather, it seems to depend on the normative problem at hand and whether one sees the boundary question as ‘endogenous’ or ‘exogenous’ to democratic theory (Hurley, 1999: 126–129). On a narrow exogenous view, democratic theory is used to answer the basic requirements question while leaving the boundary question aside, say, when members of a club wish to organize themselves democratically and seek to secure the basic conditions for doing so. Or take Habermas’s deliberative theory, in which the principle of democracy is premised on a legal community already being in place (Habermas, 1996a). On a broader endogenous view, which indeed seems more appropriate when theorizing democracy in a transnational and global context, any fully fledged conception of democracy would have to address both questions. To this extent, I find Macdonald’s ambitions appropriate for the problem she addresses. ⁸

It is almost a truism among political philosophers that the boundary question is prior to the basic requirements question in democratic theory. The controversies have revolved more around the conceptual and normative significance of this fact and its implications. For example, while Robert Goodin sees the boundary question as the first challenge for democratic theory (Goodin, 2007: 40–41), Robert Dahl infers that since democracy logically presupposes a ‘demos’, its very boundaries are not a question for democratic theory (Dahl, 1989: 119–131). However, a crucial aspect that is overlooked is what exactly ‘prior’ alludes to in this context. Most importantly, from the fact that the boundary question is causally and empirically prior, it does not follow that it is also normatively prior. In fact, it would be peculiar to argue that it is normatively prior, since there are a lot of boundary problems ‘out there’ in normative space, concerned with the who question from various normative ideals (e.g. human rights or global justice), such as who should be included and excluded from having group rights or certain primary goods. The particular boundary question of relevance here concerns the ideal of democracy. For this reason, we need a pair of ‘democratic glasses’ to approach the boundary question, specifying the basic requirements for an arrangement to qualify as minimally democratic, in order to know where to look. Indeed, this is why all or most proposed criteria for justified inclusion, including Macdonald’s, are expressed relatively similarly in terms of some version of the all-affected principle, focusing either on affected interests or more narrowly on those subjected to laws and regulations (e.g. Abizadeh, 2008; Archibugi, 1998; Benhabib, 2004; Goodin, 2007; Gould, 2004; Habermas, 1996a; Held, 1995; Shapiro, 1999; Whelan, 1983).

Of course, this does not mean that there is not a lot of conceptual and normative work left for a fully fledged democratic theory when we have moved from the basic requirements question to the boundary question and answered both in a compatible way, not least pertaining to the institutional and practical dimensions of democracy, such as the contextual specification of these requirements and further conditions following from them. However, it means that for the present purposes it is by no means implausible to bracket the boundary question in addressing Macdonald’s answer to the basic requirements question. For it is argued that the non-electoral mechanisms of authorization and accountability in this model are not able to accommodate even the minimal requirements of political equality and political bindingness and are therefore incompatible with any version of the all-affected principle (the all-subjected principle included).

Winding up: Transnational deliberative democracy and democratic agency

Let me conclude by taking a look at deliberative democracy and democratic agency in a transnational or global context. As is evident from the broad range of existing democratic models, it is possible to satisfy the conditions of political equality and political bindingness in numerous ways. For example, some contemporary scholars have proposed a conception of global democracy constituted by a world government bound to a global demos through a ‘formal track’ in the form of aggregative procedures among representatives of the world’s populations (Cabrera, 2004; Tännsjö, 2008). However, civil society scholars have directed forceful criticism towards such formalistic approaches to a world state or an overarching global political order. They argue that unsupported by an ‘informal track’ — consisting of civil society engagement, public debate and deliberation — a democratic global structure would uphold unjustified power structures by being incapable of identifying problems hidden from formal political decision-making and thus unable to place them on the agenda (Dryzek, 2006; Macdonald, 2008; Macdonald and Macdonald, 2010; Price, 2003; Scholte, 2005). Further, as Macdonald points out, the liberal individualist model of representation (Dahl, 1967; see also Pitkin, 1972: 192–196) fails to make theoretical provision for the establishment of legitimate deliberative structures that underpin it (Dryzek and List, 2003; Macdonald, 2008: 137; Miller, 1992).

To sidestep this criticism but simultaneously avoid the problems arising from the separability premise, I think that something along the lines of Habermas’s two-track view of deliberative democracy looks more promising to meet the conditions of political equality and political bindingness. On a Habermasian view, democratic legitimacy is generated through two practices: informal processes of opinion-formation and formal institutionalized deliberative and aggregative decision-making procedures (Habermas, 1996a: 486–487). These two tracks are complementary in the sense that informal public discussions can identify social problems that lie outside the agenda of formal politics and bring them into political decision-making as well as critically examine political rule-makers and require accountability (Habermas, 1996a: 365). But the tracks are also interdependent, since legally facilitated informal opinion-formation in the public sphere is channelled into legally institutionalized will-formation in terms of deliberations and bargainings, the outcomes of which are combined with legally binding decision procedures, among which the most important is majority rule (Habermas, 1996b: 1494).

In line with the two-way relation discussed earlier, and against the separability premise, the normative relationship between governing authorities and those subjected to their decisions or laws consists of two interdependent tracks, which bind the same subjects through a common legal and institutional framework. In other words, citizens who engage in civil society activities to push an authority (e.g. a global governance institution) for increased accountability should also possess a formal equal status as participants in its egalitarian decision-making. Compare this with Dryzek’s deliberative system model, consisting of a public space which is not necessarily tied to the exercise of political power, and an empowered space which is not necessarily tied to any ‘moments of binding collective decision’ (Dryzek, 2011: 226).

Another difference is that the defended two-track view rests on the principle that democratically legitimate laws and political decisions must not win assent by those affected by them, as stressed by Macdonald and Dryzek, but only by those subjected to them, that is, those whose actions are governed by them (Dahl, 1989; Habermas, 1996a: 110; Lopez-Guerra, 2005). The reliance on a narrower ‘subjected view’ rather than an ‘affected view’ has several advantages for deliberative democracy. As we have seen, even if the all-affected principle is not directly applied by the deliberative civil society approach (cf. Archibugi, 1998, 2000, 2002; Held, 1995, 2002), it is used as a fundamental device for theorizing democracy. On Macdonald’s view, those who are affected in the sense of having a relevant interest or ‘stake’ in a decision should be able to participate in the decision-making about it (Macdonald, 2008: 192). On Dryzek’s view, democratic legitimacy is generated to the extent that those affected by collective decisions are able to participate in consequential deliberation about them and their related discourses (Dryzek, 2011: 212, 222). But as I have tried to show, their respective use of this principle is unsuccessful in accommodating political equality and political bindingness. However, there seem to be more profound problems involved.

Democratic theories built up around the all-affected principle tend to replace political equality with some kind of moral equality, for example, anchored to an egalitarian idea of the equal worth or dignity of all human beings. But it is far from clear how such a moral equality is supposed to translate into political equality, taking us from counting all relevant interests equally to equal political power (Erman, 2011, 2012). ⁹ For political equality is not only premised on the idea that members are morally equal and as such have a right to participate in decision-making to the extent that they are affected, but also that they have equal influence in the decision-making. This is precisely what the conditions of political equality and political bindingness are able to capture. Further, the all-affected principle allows for a proportional view of affectedness, according to which those who are more affected by a decision should have more influence than those who are less affected (see Gould, 2004; Rawls, 1971). Indeed, in line with Macdonald, supporting proportional influence seems sensible since it is affectedness that motivates a right to participate in the decision-making in the first place (Bergström, 2007 ; Macdonald, 2008). The problem from the point of view of democracy, however, is that members are supposed to rule over themselves through a political arrangement which takes numerous decisions on a wide range of social, political, legal and economic issues. ¹⁰

The question is what the defended two-track view could reasonably look like beyond the nation-state under the pluralist conditions that Macdonald and Dryzek explore. Habermas is reluctant to stretch the two tracks transnationally and globally and instead recurs to a three-tiered system on the supranational, transnational and nation-state levels. He argues that supranational coordination and transnational negotiation replace democratic legitimation at the former two levels because the form of civic solidarity that the latter requires cannot be extended beyond states or continental regimes such as the EU (Habermas, 2006: 138–139). However, while I find it analytically useful to construe the global political system in terms of three systems of organization and find Habermas’s diagnosis of the present state of affairs to be largely correct, taking our starting-point in the two conditions of political equality and political bindingness makes it possible to disconnect democracy conceptually and normatively from territorial units as well as from the nation-state as the sole democratic organizational form. This, I claim, allows for a less pessimistic view of the future of transnational and global democracy than Habermas’s. It is suggested below that a shift of focus from civic solidarity to democratic law-making connected to interdependent interests and roughly equal stakes opens the possibility for theorizing the expansion of a deliberative two-track view beyond the state on the basis of what Thomas Christiano calls ‘a common world’ (Christiano, 2006). Rather than waiting for civic solidarity to emerge on transnational and supranational levels before calling for a democratization of global governance structures, democratically institutionalized ‘common worlds’ might be precisely what is needed in order for such a solidarity to arise on these levels.

What characterizes a modern political society, in contrast to associations such as clubs, commercial ventures and economic enterprises, is that members’ interests in it are not partial. Instead, to a significant degree they share a common world. Such a world is a set of circumstances among a group of people, in which the realization of all or nearly all of the fundamental interests of each member is connected with the realization of all or nearly all of the fundamental interests of every other member. Thus, modern political societies consist of comprehensive systems of provision and regulation of the basic needs of all members, providing for public goods such as education, health care, redistribution of income and wealth, and criminal law, often through a system of rights. In contrast to David Miller and others who connect the question of the emergence of democratic polities to national and cultural identities, to say that people share a common world in the legal-political and interest-based sense stressed here does not conceptually require a shared culture or national identity, since multinational and multi-ethnic states could inhabit such a world (Miller, 2009, 2011).

Undeniably, we live in a world in which our interests are affected by what people do around the globe. As argued by Goodin, virtually everybody in one way or the other seems affected by everybody else (Goodin, 2007). However, the common world condition is distinct in that citizens’ interests are deeply interdependent and as a result are connected in multiple ways, both legal-politically and institutionally (Christiano, 2006: 97). In contrast to arrangements such as organizations and associations in which people have very different stakes and are differently affected, in a modern democratic society, individuals have roughly equal stakes in the world in which they live. Even if they do not have equal stakes in each and every political decision, they do have over time (Christiano, 2006: 98). Indeed, if a political system only used democracy for deciding upon one issue and the rest was decided by, say, an elite or a dictator, the majority voting on this one and only issue would needless to say be undemocratic, as some would have a much higher stake in the decision than others, which would generate clear winners and losers.

The common world condition is here to be understood in descriptive terms such that it does not say anything about whether it is desirable to strive towards creating common world contexts. Furthermore, in contrast to comprehensive moral ideals, it is argued that democracy is desirable only in common world contexts and as such is defended as a partial ideal (see Robeyns, 2008: 344). In other words, democracy is appropriate for the political decision-making once it is an empirical fact that people have interdependent interests and roughly equal stakes (Christiano, 2006: 98). To be sure, the common worlds that we know of are products of highly morally arbitrary and highly contingent causes. But arbitrariness of origins does not take away the fact that a common world is a morally relevant fact now (Christiano, 2006: 85–87). ¹¹

What follows from this concerning deliberative democratic agency is this: first, political actors ought to become democratic agents if and only if they have interdependent interests and roughly equal stakes; and, second, in order to turn into democratic agents they must be part of a legal-political structure which fulfills the conditions of political equality and political bindingness. If we take a look at the transnational level in light of these considerations, the state will continue to be the political community of primary importance in the foreseeable future on the two-track view, not only from a pragmatic but also from a normative point of view. At the same time, this does not mean that we should not opt for expanding our polities beyond the nation-state, for example, as is presently being done in Europe. Quite the contrary. In the examination of alternative democratic units to the state, Christiano neglects important qualitative differences between international institutions such as the WTO, the IMF and the World Bank, on the one hand, and regional bodies such as the EU, on the other (Christiano, 2006: 94–96). Members of the EU increasingly have at a minimum similar stakes in the mutual effects of their interactions, at least in some areas, the more their interests have become connected in multiple ways through the development of common legal-political and institutional structures. At the same time, the problems of the democratic deficit of the EU, together with the recent crises where members do not seem to accept that their interests are in fact interdependent, pose great challenges for the EU as a deliberative two-track project. However, this is a topic for another article.

A move from the regional to the global level forces us to yet again acknowledge the differences between a body such as the EU and international institutions such as the WTO and IMF. Concerning global governance institutions, the prospects for democracy look much bleaker, at least under the present circumstances. However, as our fundamental interests become increasingly interdependent in the moral and political global world order through human rights, international law and common global governance structures, the more we have roughly equal stakes at least within some domains, such as important environmental issues, peace and security. If we become sufficiently and significantly interdependent in the mutual effects of our interactions in these domains — legally, politically and institutionally — democracy becomes an attractive ideal. And if this is the case, an advantage of a two-track view anchored to the conditions of interdependent interests and roughly equal stakes — rather than to civic solidarity, which is too entangled with a statist framework — is that institution building plays a much more important part, opening the door for institutional innovation as well as the emergence of civic solidarity. In this endeavour, the empirical conditions of interdependent interests and roughly equal stakes, together with the conceptual and normative conditions of political equality and political bindingness, form a fruitful basis for the shaping of common institutions, which are not necessarily limited to or even tied to territorial borders.

Indeed, analysing the prospects and challenges facing democracy at the global level is a task for non-ideal theory and thus not a task for the present article. That said, it is worth mentioning that the newly emerging field of global administrative law (GAL) could be a useful tool in this enterprise, constituting a regulatory ‘space’ that transcends inter-state relations (Kingsbury et al., 2005: 26). A basic presumption shared by scholars of GAL is that we can no longer hold that adequate accountability for global regulatory governance can be achieved through the application of domestic administrative law requirements to domestic regulatory decisions alone in light of emerging patterns of global governance structures (Boisson de Chazournes et al., 2009; Cassese, 2005; Dyzenhaus, 2005; Kingsbury et al., 2005: 26). According to this view, understanding global governance as administration makes it possible to find innovative ways of theorizing legitimacy of international institutions. In contrast to cosmopolitan scholars who devote much attention to constitutional aspects of law — which are supposed to transform cosmopolitan rights into human rights — GAL shifts the attention to administrative aspects of law-making. Such a shift highlights the extent to which mechanisms of participation and review are missing on the global level as well as inviting development of institutional principles and procedures for improving the prerequisites for global democratization and the democratization of international law (Boisson de Chazournes et al., 2009: 315–317; Kingsbury et al., 2005: 27).

Undoubtedly, much of the normative work on GAL has bought the separability premise, focusing on the operation of rules, procedures and mechanisms relating to separate democratic norms such as accountability, transparency and participation. This is not the best way forward from the standpoint of the suggested ideal-theoretical framework. Rather, the challenge for non-ideal theory would be to study in much more detail how global administrative regulations, procedures and mechanisms could and should be knit together with (rather than separated from) democratic procedures such as elections and parliamentary control. For it is one thing to assure the legality in global governance; quite another to assure its democratic legitimacy. To contribute to the latter, GAL could for example assist in transforming moral equality into political equality by protecting political rights. In administrative law conceptions, rights are broadly understood, usually regarded as held by direct subjects of regulation, including not only individuals, but also state and non-state actors (Kingsbury et al., 2005: 45). Moreover, while GAL is not sufficient to secure the condition of political bindingness, it looks promising for improving the prerequisites of an ongoing democratic practice by contributing to the institutionalization of the multifaceted deliberative arenas, of both formal and informal kinds, on which ‘bindingness’ depends.

Conclusion

Winding up, in this article I have examined the notion of democratic agency in the deliberative civil society approach with regard to its democratic qualities, illustrated by the work of Macdonald and Dryzek. It has been argued that their respective conceptions of who is supposed to be a democratic agent do not fulfil the minimal requirements for democratic agency because they are not able to accommodate political equality and political bindingness. In light of the tendency in the contemporary debate to define transnational political activities of all sorts in terms of democratic agency, this critique calls for a more careful conceptualization in this regard, since the democratic agent is a specific kind of political agent. Moreover, I concluded with a brief sketch of the contours of a Habermasian two-track view, which refutes the separability premise and sees democratic qualities such as accountability and authorization as parts of one conceptual and normative package.

This certainly does not suggest that civil society actors cannot contribute better global governance in important ways. For example, existing political practices show that NGOs are contributing to a more just world order by being ‘agents of justice’ and ‘agents of democracy’. As the latter, they can push international organizations towards increased transparency and accountability and as such improve the empirical prerequisites for global democracy. In fact, they have had a large impact in this role — undertheorized by Habermas — putting important issues on the agenda as well as influencing policy outcomes in a variety of policy fields (Erman and Follesdal, 2011; Erman and Higgott, 2010). The point made here is merely that they cannot be democratic agents the way they are conceptualized by the deliberative civil society approach, because in order to transform from agents of democracy to democratic agents, political equality and political bindingness must be secured.