Introduction
The understanding that political legitimacy can spring both, or either, from ‘democracy’ – including some form of consent, collective self-rule, or procedural criteria – and/or from ‘effectiveness’ – including beneficial consequences or utility gains – is as old as the study of political legitimacy. This dichotomy has been characterized as ‘input’ vs. ‘output’ based legitimacy. Fritz Scharpf seminally distinguished between claims to legitimacy based (a) on ‘input authenticity’, that is, some link with the authentic preferences of the members of a community, and those resting (b) on ‘output efficiency’, or the effective promotion of ‘the welfare of the constituency in question’ (Scharpf, Reference Scharpf1999: 6–9). But long before and far beyond the domain of European integration, for which Scharpf developed this classification, conceptions of democratic and procedural input and/or effective performance output as legitimacy sources have marked both normative and empirical accounts of political legitimacy.
Normative political thought on political legitimacy comprises not only social contract and democratic theories, but also utilitarianism and consequentialism (see e.g. Rosanvallon, Reference Rosanvallon2011). Political theory deliberates whether the legitimacy of political institutions and the decisions made within them depends primarily on procedural features – including democracy- and rights-related ones – or whether this rather depends on the quality of the decisions made, and the ‘substantial values’ realized in them. A related point of discussion is whether or not legitimacy demands democracy (see Peter, Reference Peter2014).
On the other hand, empirical social science, and particularly research in the behaviouralist tradition, has also framed legitimacy in terms of the extent to which it is created, maintained, or lost, either on the side of democracy, involving procedural and rights-value, or identity-related considerations, or else on the side of regime or government performance. In contrast to normative theory, this type of work conceives of legitimacy, not as a virtue of political institutions or the decisions made within them, but rather as a belief or attitude on the part of their beholders, expressed in either public opinion or political behaviour, and measurable through large survey or panel data (see Gilley, Reference Gilley2006: 48; Levi et al., Reference Levi, Sacks and Tyler2009). It mostly uses the term ‘legitimacy’, as in social legitimacy, synonymously with ‘regime support’.
This paper explores the nexus between input- and output-based legitimacy (see further Skogstad, Reference Skogstad2003; Torres, Reference Torres2006; Bellamy, Reference Bellamy2010; Heard-Lauréote, Reference Heard-Lauréote2010; Lindgren/Persson, Reference Lindgren and Persson2010; Hobolt, Reference Hobolt2012). The precise research question posed here, queries the nature of the relationship between the two types of legitimacy. Are they mutually dependent, complementary, or do legitimacy gains on one side come, rather, at the price of losses on the other side of the dichotomy? Looking beyond the relative importance of the two types of legitimacy,Footnote 1 this paper challenges the prevalent (and often wrongly cited) topos of a ‘democratic dilemma’ between ‘system effectiveness’ and democratic ‘participation’ (Dahl, Reference Dahl1994). Rather, it makes the case that input and output legitimacy necessarily go together.
The paper is divided into three substantial sections and a conclusion. The first section prepares the ground. It introduces the particular approach to political legitimacy advanced in this paper, and explains how this general approach contributes to the paper’s distinct take on the specific research question it addresses. Further, this section explains the methodological choices that flow from this approach. The second section develops the paper’s argument in engagement with the relevant European Union (EU) Studies literature, situating it in the wider fields of Political Theory and Comparative Politics. It outlines a substantial gap in the literature, namely a certain blind spot regarding possible standards by which to assess output performance as legitimacy enhancing. This section proposes that these measures or standards necessarily link output to input legitimacy. The third section moves on to substantiate the argument empirically. It presents the paper’s case material, a qualitative-interpretive textual re-construction of how the input/output nexus played out in the discursive practices of the European institutions, from the beginning of integration up to the early 2000s. The Conclusion, finally, brings the second and third sections together. It reflects on how the particular and context-dependent discursive history that has been outlined may relate to the nature of the input/output legitimacy nexus more broadly.Footnote 2
The underlying approach to legitimacy and the resulting methodological choices
The paper’s research question regarding the relationship between input and output legitimacy is tackled from a specific angle, which results from the distinctive way of approaching political legitimacy that it seeks to promote (see also Sternberg, Reference Sternberg2013). This approach to legitimacy focuses on historically contingent understandings of what political legitimacy might mean, in a specific context, and as reflected in the discourses of specific, actual actors. It steers a course between two deeply divided camps in the scholarship: normative accounts debating the conditions under which people ought to accept something as legitimate, and empirical research into the extent and causes of them doing so. The conceptual starting point for this approach is the notion that something is legitimate not simply because people believe in its legitimacy, nor only because it meets certain abstract or ideal criteria. In addition to both of these, it is legitimate to the extent that it can be justified in terms of beliefs, narratives, and conceptual languages shared by, and among, dominants and subordinates (see Beetham, Reference Beetham2013: 15–18; see further Habermas’s concept of ‘discursive justifiability’, 1973: 139, 173).
Such belief systems, and what it makes sense to say about an object’s legitimacy, are fundamentally pluralistic. Even the most minimal shared understandings (for instance, of what is relevant to an assessment of legitimacy to begin with) are essentially contested. They are subject to continual reconstruction. Deductive systematic arguments about what hypothetical rational or communicative actors would consider legitimate offer little help in exploring the involved processes of social contestation and construction. Quantitative empirical research, on the other hand, also sheds limited light on contests over what legitimacy means to people. This is because translating explicit and implicit argumentative-narrative logics, and the webs of meaning embedding them, into quantifiable codes, comes at the cost of a losing acuity regarding the very dynamics of meaning-making and contestation under scrutiny in this paper (see Yanow and Schwartz-Shea, Reference Yanow and Schwartz-Shea2006: xii). By contrast to both, this paper therefore investigates, inductively, the structures of arguments used by actual actors.
The paper’s hermeneutic strategy thus emphasises the study of processes of meaning-making and knowledge production. It focuses on such processes as expressed in, or as underlying, the discourses of key actors in key documents. In order to analyse the fine grain of discursive, narrative, and argumentative meaning-making in and across texts, the best method available is close reading, or non-quantitative interpretive textual analysis. Interpretation as a method is concerned, empirically, with how meaning is produced, contested, and reproduced (see Yanow and Schwartz-Shea, Reference Yanow and Schwartz-Shea2006). Textual interpretation, particularly, concentrates on how this happens, both explicitly and implicitly, through language, narratives, imageries, concepts, or discursive logic.
This paper’s approach is pragmatic in that it looks at contests over normative beliefs through the lens of the standards to which actors commit themselves, both in their political language and in their attempts to cope with practical problems. Understanding what they take for granted in what they say, and how the EU measures up to the reflected understandings of legitimacy, will ‘help identify [in this case, the EU’s] potential points of vulnerability, and explain any erosion of its ability to secure cooperation […] when under pressure’ (Beetham, Reference Beetham2013: 100). Approaching legitimacy in this way allows for an exploration, for a particular context (and a limited one, as explained shortly), of the discursive production of central justificatory principles and conventions about granting legitimacy, compliance, or consent that underpin the EU’s system of power. The narrow focus of this paper’s specific research question on the input/output nexus, in turn, permits zooming in on how input- and output-related arguments play out in these processes of knowledge production – and how this might reflect on potential general argumentative challenges innate to contests over political legitimacy at large.
This study, then, is about arguments and claims regarding legitimacy, and their theoretical, cognitive, and associative underpinnings. It is not about regime support or the causal factors determining it. In other words, beliefs in legitimacy are not under scrutiny, but rather specific discourses involved in producing meaning and knowledge about legitimacy. The aim here is not to offer a judgement about the degree of the EU’s legitimacy, neither in terms of measurable popular opinion nor in normative terms. Equally, it is not to propose an absolute definition of what legitimacy means in the particular case of the EU, based on some ‘representative sample’ of dominant ideas concerning this question. All of these research aims would impose stringent requirements concerning the avoidance of selection biases and the representativeness of the sources or passages cited in relation to the totality of discourses or beliefs about EU legitimacy.
By comparison, this paper’s ambition is both more limited and more general. The value of its analyses lies in its scrutiny of how, and on what grounds, arguments are made in the particular subset of relevant discourses under study, and to what extent these argumentative structures might tell us something about the relationship between input and output legitimacy more generally. It is, of course, valid to ask whether or not other sets of discourses about EU or political legitimacy would display similar structures and run into similar argumentative challenges. This paper’s two-pronged approach, combining a general discussion of the academic literature (second section) with an analysis of a particular, limited set of EU-official discourses (third section), is an attempt to explore how legitimacy in this discursive context relates to the general nature of political legitimacy.
The source material chosen for third section includes public statements, declarations, treaty preambles, reports, and policy documents by the European institutions and political leaders. The time frame covered reaches from the 1950s to the early 2000s. On some level, even the academic literature discussed in the second section can be seen as another set of discourses, themselves projecting and reflecting specific lines of argument and assumptions. The analysis of these also throws light on the evolution and nature of what could plausibly be said about political legitimacy, and on how input- and output-related considerations come together in creating this plausibility. Excluded from the paper’s corpus are, importantly, other political, intellectual, media, or private discourses in the member-states’ public (and non-public private) spheres. Even so, the discourses discussed in second and third sections arguably did represent important cognitive ‘maps’ available to people for finding their way through the ‘forest’ of the issue of EU legitimacy (see Gamson, Reference Gamson1992: 117). Moreover, they were in dialogue with wider public understandings and hence indirectly reflective of them, even if sometimes inversely, that is, in their reaction to counter-discourses.
Drawn from research presented in a monograph offering a broader discursive history of contests over EU legitimacy in EU-official as well as wider public discourses (Sternberg, Reference Sternberg2013), the sources cited in this paper are ‘representative rather than exhaustive’ (see Mottier, Reference Mottier2005: 258). That is to say, they were selected, in an iterative cycle, to illustrate key discursive positions and patterns, identified on the basis of the more comprehensive corpus used in the book. More particularly, with a view to this paper’s specific focus on the nexus between input- and output-related legitimacy arguments, the selected individual source references exemplify different types of argumentative patterns pertinent to this nexus, as well as particular argumentative challenges in regard to it. On this basis, key discourses discussed in third section include those around integration furthering ‘peace and prosperity’ and a ‘common European good’, those surrounding the advocacy of the 1960s and 1970s of a directly elected and strong European Parliament (EP), those around the cross-institutional ‘People’s Europe’ campaign of the 1980s, and the ‘governance’ discourse of the early 2000s.
Three further methodological criticisms might be anticipated as follows. First, the identification of relevant key discursive positions, patterns, and techniques constitutes the very core of the interpretive process. It cannot therefore be frontloaded, and this is why no rigid scheme of analysis was used. Second, at the centre of this paper are the meanings of legitimacy projected and reflected in the discourses under study, that is, the content and internal logics of those discourses – not the actors advancing or instrumentalizing them (who would be the focus in frame analyses in the Schattschneiderian tradition, see e.g. Fligstein, Reference Fligstein2001; Daviter, Reference Daviter2011). For example, regardless of whether or not one expects the EP to favour input over output legitimacy, what is of interest in this particular paper are the grounds on which it makes whatever claims about EU legitimacy it makes. Finally, the paper does not seek to separate out input- and output-oriented claims about legitimacy from one another; most claims refer to both in some way, as will be argued. It does not measure the relative frequency or legitimacy-enhancing impact, nor possible correlations, of specific motifs, as quantitative discourse, frame, or content analyses endeavour to do (e.g. Bellamy, 2010). This is because this study looks beyond correlations between (perceptions of) overall, input-, and output-based legitimacy (see Lindgren and Persson, Reference Lindgren and Persson2010 for a study of this). Rather, it investigates ways of constructing their relationship, and the argumentative grounds on which they can be founded.
If, say, positive perceptions of input legitimacy do go hand in hand with positive perceptions of output legitimacy, as Lindgren and Persson suggest, how can this link be constituted discursively? On what grounds does it make sense? How do constructions of the nexus change over time – and what might this tell us about the nature of input and output legitimacy, and how they relate to one another? Before turning to these questions in the context of the evolution of the European institutions’ discourses over time, let us first look at the relationship between input and output legitimacy in abstract terms, in reference to the relevant scholarly literature.
The nexus and the literature: can one legitimacy type work without the other? Can it work with the other?
Scharpf’s basic normative claim was that the key foundation for the legitimacy of European integration and the EU lay not in their ‘input-authenticity’, but in the EU’s policy outputs (1999: 6–9, 283; see also Majone, Reference Majone1996). This constituted a momentous change of perspective from existing accounts that had seen the EU’s legitimacy almost exclusively through the partial lens of its democratic legitimacy, or rather its ‘democratic deficit’. It ushered in a veritable output turn in the study of EU legitimacy (Bellamy, 2010: 2–3). This turn affected both normative accounts engaged in de- and re-constructing standards of legitimate political order in the EU context, and empirical accounts assessing this polity against such standards. In the study of public support for integration, a ‘utilitarian’ cost-benefit consideration came to function as a key explanatory factor (e.g. Gabel, Reference Gabel1998). Beyond the EU case, empirical research into the ‘universal sources’ of legitimacy has also challenged political theory’s common focus on democratic procedures, rights, or moral obligations as foundations of political legitimacy, finding positive correlations between perceptions of performance and of legitimacy (Gilley, Reference Gilley2006; see also Hechter, Reference Hechter2009).
The lacuna of a measure of output legitimacy, and its necessary link to input legitimacy
What precisely counts as legitimacy-enhancing output in accounts emphasizing output legitimacy varies considerably. Definitions range from general problem-solving effectiveness in tackling the complex problems of an internationalizing world (Risse, Reference Risse2006: 191), and effectiveness in fulfilling specific tasks delegated by public actors (Majone, Reference Majone1998), to efficient ‘performance in meeting the needs and values of citizens’ (Lord and Beetham, Reference Lord and Beetham2001: 444). These definitions have in common that they describe abstract categories that need to be filled with substance. Which problems are to be solved, how are tasks to be chosen and how prioritized, how are the burdens and benefits of tackling them divided? Who can legitimately delegate which tasks, according to which procedures, and on what grounds – and how can the accountability of agents be ensured? Finally, how are the relevant citizen needs and values to be determined and weighted against one another? All of these questions effectively raise issues of input-authenticity.
In particular, identifying legitimacy-relevant outputs requires some way of linking outputs to citizen preferences. Not just any output will do. It needs some claim to input authenticity, whether through effective responsiveness or substantive representation, formal representation, authorization, electoral accountability, or other forms of participation. The keystone of all arguments about output legitimacy is some measure of what constitutes legitimacy-enhancing outputs. Most output-focused accounts of legitimacy tend to hinge on an ‘implicit conception’ of the public interest or common welfare, the promotion of which establishes output legitimacy (Moravcsik and Sangiovanni, Reference Moravcsik and Sangiovanni2002: 125–126). In Scharpf’s case, (Reference Scharpf1999: 43–83, 199) this consists of an effective balance between EU-level market liberalization and national social protection, and in Majone’s, the presumed citizen wish to ‘preserve national sovereignty largely intact’ (1998: 5, 7, 14). Both thus link their measure of legitimacy-enhancing output to substantive claims about this being ‘what the citizens want’, and about how effectively EU action responds to it. In Hanna Pitkin’s categorization, they project ‘substantive representation’ (Pitkin, Reference Pitkin1967).
To be sure, the difficulties in determining what the people really want, and hence what kinds of outputs would enhance legitimacy, are considerable. Public opinion research provides useful indications, but reaches its limits when it comes to aggregating preferences between more than two options, or dealing with trade-offs between competing goals, with goals that cannot be ordered on a single (e.g. left-right) dimension (see Moravcsik and Sangiovanni, Reference Moravcsik and Sangiovanni2002: 140–143). In any case, claims to a link between government action and the will of citizens often rely on arguments about the procedural qualities of the processes by which its ends and goals were defined (see above footnote 2). Typically, these take the form, at least partly, of arguments about formal representation through the institutions of majoritarian, competitive party democracy (see Kraus, Reference Kraus2004: 562; Bellamy, 2010). Or they may draw on arguments about representation through non-majoritarian practices of participatory democracy, or through ‘communicatively generated power’ produced through deliberation in independent public spheres (Habermas, Reference Habermas1996: 301–302). All of these arguments are effectively arguments about input legitimacy, in the sense of supporting a claim to a connection between outputs and citizen preferences.
Some scholars further argue, that output legitimacy presupposes yet another type of input authenticity, namely a certain shared belief in a collective identity, or a certain ‘consensus’ over values and principles. Both of these supposedly promote output efficiency by favouring choices in line with a general orientation towards the common good of the political body as a whole (e.g. Scharpf, Reference Scharpf1999: 7–9, 13; Kraus, Reference Kraus2004: 562; but see White, Reference White2010a). Note that Scharpf also posits collective identity to be a necessary condition for input (in addition to output) legitimacy, namely for majority rule to lose its ‘threatening character’, and for engendering ‘trust in the benevolence’ of one’s fellow citizens (1999: 7–9, 13). This consensus or collective identity requirement, of course, ‘misrepresents much mainstream work in democratic theory’ that starts from the premises of pluralistic interests or principled disagreement (Bellamy, 2010: 3). With a view to the relationship between input and output legitimacy, identity and consensus are examples of something held necessary for both, input and output legitimacy thus seem at times to be favoured by the same factors or measures.
To summarize, output-oriented accounts of legitimacy depend on some notion of, or process of defining, the ends and goals of the political order and its actions. These somehow acceptable ends and goals provide the measure of legitimacy-enhancing performance, and necessarily link effective output to some level of input legitimacy. If output legitimacy arguments thus depend, by their own logic, on some claim to input legitimacy, conversely, can input legitimacy work without output legitimacy? Does this dependence run both ways?
Complementarity
There are a number of ways in which input and output legitimacy-related arguments are complementary to one another in building overall legitimacy. This external complementarity is qualitatively different from the internal dependence (by virtue of their own argumentative logic) of output legitimacy arguments on input-related arguments proposed in the previous section. The complementary relationship does seem to work in both directions.
On one hand, input-oriented arguments ‘never carry the full burden of legitimizing the exercise of governing power’, but tend to be supplemented by output-oriented arguments about positive outcomes for the public interest (Scharpf, Reference Scharpf1999: 188, 43–83). Arguments about the legitimacy of democratic systems, for example, involve not only input-related claims about democratic credentials, but often also output-focused democratic growth or democratic peace theses (e.g. Dahl, Reference Dahl1998: 57–59). Even a political system with reasonable claims to input legitimacy (on grounds of electoral accountability, representation, etc.) might be seen as illegitimate if it fails to provide certain outputs, such as securing its citizens’ safety, protecting their property, or providing an environment in which they can secure a livelihood.
On the other hand, output legitimacy is only ever as stable as performance outputs are. It may depend on input legitimacy to help to ride out periods of performance difficulties (see Easton, Reference Easton1965: 273; Zhao, Reference Zhao2009). This is especially true if the performance of outputs depends on factors beyond the control of the regime in question (Habermas, Reference Habermas1973). The current Eurozone crisis once again underlines the danger of resting the legitimacy of a political order too exclusively on performance outputs, especially under such circumstances.
In sum, neither input authenticity nor output efficiency can durably do the job of building overall legitimacy on their own. Yet, one type can complement and partly, or temporarily, make up for deficiencies of the other. The very backdrop to the output turn in the study of EU legitimacy illustrates this. Generally, legitimacy accounts that put the emphasis on one of either type of legitimacy source are often grounded in a critique of the state of affairs regarding the other type. The academic interest in output-based legitimacy was in tune with a growing disillusionment across liberal democracies with the prospects for input-based legitimacy, and with the traditional institutions of representative democracy (see Norris, Reference Norris1999). In the EU case, these prospects were further constricted by the structural limitations particular to its unique political order. Paradoxically, even a dramatic increase in EP powers had not changed public opinion from seeing the Community as a ‘non-democratic set of institutions’ and the EP as ‘distant, powerless, and poorly representative’. By the mid-1990s, the previously dominant analysis, whereby the ‘obvious’ solution to the EU’s legitimacy deficit was to strengthen the EP, lost its hegemony (Magnette, Reference Magnette2001: 292–293).
The academic ‘output’, ‘participatory’, and ‘governance turns’
One common response was to focus arguments about EU legitimacy on its output-related problem-solving performance rather than on unwinnable claims to its input legitimacy (see Scharpf, Reference Scharpf1999: 187–189, 7–9, 21; Bellamy, 2010: 3). Other parts of the debate, by contrast, took the difficulties of electoral democracy as a starting point for exploring alternative ‘non-majoritarian’ or ‘postparliamentary’ modes of legitimation and governance (Dehousse, Reference Dehousse1995; Majone, Reference Majone1996; Lord and Beetham, Reference Lord and Beetham2001; see Kohler-Koch and Rittberger, Reference Kohler-Koch and Rittberger2007: 27). These modes had a hybrid status uniting output- and input-oriented arguments. They clustered around the ideas of ‘participatory democracy’ and of the EU as a ‘regulatory state’. These will be discussed, in turn, in what follows.
The central idea behind the academic debate’s ‘participatory turn’ was to involve interest groups and civil society organizations (rather than bureaucrats, representatives of individual citizens, or political parties) in policy-making, relying only marginally on legislation (Finke, Reference Finke2007; Greenwood, Reference Greenwood2007: 333; Saurugger, Reference Saurugger2008: 1275). This took inspiration from the models of interest-group pluralism and deliberative democracy. ‘Participatory democracy’ in the EU was advocated partly in appeal to related gains in input legitimacy. More particularly, these appeals concerned equal representation (Saurugger, Reference Saurugger2008: 1276; see Skogstad, Reference Skogstad2003: 322) or equal and fair public deliberation by those concerned by the policy in question. Gains in accountability were also appealed to, if only in the somewhat figurative sense that policy makers could expect to have, in order to justify their decisions to those affected, and would, therefore, feel responsible to them (Risse, Reference Risse2006: 186, 192–193). To be sure, the input legitimacy-related democratic credentials of civil society consultations were controversial. They were especially so on grounds of their elitist and top-down nature, citizens’ uneven access to them, and the insufficiently democratic internal structures of civil society organizations themselves (e.g. Grande, Reference Grande2000: 29–30, 129; Magnette, Reference Magnette2003; Neyer, Reference Neyer2003: 687). Perhaps this is why ‘functional’ or instrumental arguments about increased output effectiveness were also crucial in the advocacy and justification of associative or participatory democracy. Accordingly, involving concerned actors helped overcome implementation problems by mobilizing their willingness, compliance, and expert resources; even purely subjective, informal accountability to stakeholders enhanced the effectiveness of governance arrangements (Hurd, Reference Hurd1999: 387; Neyer, Reference Neyer2003; Risse, Reference Risse2006: 186, 193; Finke, Reference Finke2007: 6–10; Greenwood, Reference Greenwood2007: 340; Saurugger, Reference Saurugger2008: 1276).
Justifications of the EU as a ‘regulatory state’, likewise, included appeals to both input and output legitimacy. At their heart was the delegation of specific tasks to semi-autonomous authorities such as constitutional courts, central banks, or other regulatory and administrative agencies. This was justified as common practice in advanced industrial democracies, and as well suited for the specific policy fields in which the EU was active (Majone, Reference Majone1996; Moravcsik, Reference Moravcsik2002; Lindseth, Reference Lindseth2010: 104). The recipients of delegated regulatory power worked in essentially output-driven, depoliticized and expertise-based modes of ‘administrative governance’. Still, delegation was also an input-driven ‘normative-legal’ principle. It required a ‘lawful legislative enactment’ or ‘loi cadre’, which effectively depended on some input-based claim to majoritarian-democratic legitimacy (Lindseth, Reference Lindseth2010: 2, 104; see Majone, Reference Majone2005: 7). Mostly, however, input-related justifications of delegation were overpowered by output-oriented ones. The ‘loss of democratic control’ following delegation was centrally justified by its consequences, namely that it enhanced ‘the ability of political systems to produce outcomes in the public interest’ (Moravcsik and Sangiovanni, Reference Moravcsik and Sangiovanni2002: 129, see Majone, Reference Majone1998). Is the zero-sum relationship between input and output legitimacy, assumed in this example, to be taken for granted?
Is there an inescapable trade-off?
Many EU scholars imply, and some directly assert, ‘an inevitable trade-off’ between ‘an emphasis on government for the people and an emphasis on government by the people’ (Katz and Wessels, Reference Katz and Wessels1999: 5; see Torres, Reference Torres2006; Lindgren and Persson, Reference Lindgren and Persson2010: 452–453). The common assumption is that ‘the “effectiveness” of outcomes would be unacceptably harmed if dissenting views were acknowledged and engaged at each stage’ (White, Reference White2010b: 56).
Yet even if input and output legitimacy are indeed antithetical in some of their many respective shapes, or in some contexts, they might not be in others. This has in fact been a central argument in favour of participatory or deliberative (non-majoritarian) democracy. Both are partly advocated as a way of overcoming a supposed input/output trade-off existing for traditional models of representative democracy (see e.g. Skogstad, Reference Skogstad2003: 222–225; see White, Reference White2010b: 59; but Bellamy, 2010: 3). However, non-majoritarian mechanisms of input legitimacy have also been argued to have ‘perverse rather than beneficial effects’ on output efficiency (Bellamy, 2010: 2). One of the challenges of creating and maintaining political legitimacy, then, might be that even if input and output legitimacy do depend on each other in some ways, they might still conflict in others. In other words, they might be both mutually reinforcing and antithetical, either at the same time, or at different points in time.
Winston Churchill, for example, made such a temporal distinction in an interview in January 1939. Asked whether it was ‘possible to combine the reality of democratic freedom with efficient military organisation’, he conceded that it ‘may be that greater efficiency in secret military preparations can be achieved in a country with autocratic institutions than by the democratic system. But,’ he went on, ‘this advantage is not necessarily great, and it is far outweighed by the strength of a democratic country in a long war [in that] the people feel that they are responsible, and […] will hold out much longer than the population of Dictator States’.Footnote 3 In the medium term, Churchill argued, an initially less efficient input-legitimate democratic system would win out in terms of output effectiveness as well.
If, however, input and output legitimacy can indeed be both interdependent or mutually reinforcing and in a trade-off relationship, depending on time and context, then the durable solution cannot be to sacrifice one for the other. Rather, legitimacy claims will have to carefully negotiate their balance, and consider how input and output legitimacy relate to each other, lest overall legitimacy be undermined. Britain’s situation on the eve of World War II powerfully illustrates the urgency of safeguarding democracy as well as efficiency. The next section of this paper explores how the balancing act between input and output legitimacy has played out in another specific context of practical (non-academic) discursive practice.
The nexus in discursive history
This section reconstructs how legitimacy discourses in the environment of the European institutions have negotiated the relationship between input- and output-based legitimacy claims. What references did they make to the respective types of arguments, and to what extent did references to one necessitate an explicit or implicit reference to the other? In other words, to what extent did input and output legitimacy belong together in the logic of these discourses? Does the long-term history of EU legitimacy discourses, too, suggest that the two complemented each other over time? How, in particular, were the legitimacy-enhancing ends and goals of the emerging European polity forged and defined? The underlying aim of this exercise is to investigate what insights such an analysis can offer for the conceptual debate on EU legitimacy.
Forging and furthering a European ‘common good’: pure output legitimacy?
For the early years of integration, claims to legitimacy were to an important degree output-oriented – even if never quite exclusively, as will be argued in the subsequent subsections. Virtually all sources framed the legitimacy of integration at least partly in reference to a promise of making peace and prosperity possible in Europe. On one hand, integration was almost ritually cast against Europe’s long and still fresh history of bloodshed and war – and the continued threat of war ‘if we did nothing’ (Monnet, Reference Monnet1978: 289; see e.g. Preamble ECSC Treaty). More particularly, integration featured as a way to contain Germany (e.g. Schuman, Reference Schuman1950), to discourage the ‘Eastern powers’ from striving for ‘the control of Europe and the continuation of the world revolution’ (Hallstein, Reference Hallstein1955), and to keep totalitarianism at bay at home and abroad (e.g. Preamble EEC Treaty). On the other hand, early discourses legitimating the European Communities routinely highlighted the ‘new prospects of progress and prosperity’ opened up by European integration (CEC, 1958: 9). The promise of a ‘higher standard of living’ or ‘improved living conditions’ (to be achieved through the growth, productivity increases, modernization effects, and economies of scale associated with Europe-wide markets) pervaded pro-Communities discourses throughout the 1950s and 1960s. A common label for this was the emblem of ‘economic and social progress’ (see e.g. Hallstein, Reference Hallstein1951: 3; The governments of the Federal Republic of Germany, Belgium et al. 1955; Preamble ECSC Treaty; Spaak, Reference Spaak1957).
These ‘peace’ and ‘prosperity’ storylines were essentially means of establishing that there was such a thing as a ‘European common good’. The furtherance of this common good formed the basis, and measure, of integration, and of the Communities’ output legitimacy. This storyline was grounded on a related one; that integration was ‘indispensable’, a matter of no alternative, and even of survival (e.g. Marjolin, Reference Marjolin1958; Monnet, Reference Monnet1962: 5). Integration widely featured as indispensable to safeguarding peace in Europe and hence, given the horrors of the alternative, as simply indispensable. Economic reconstruction and better living conditions were a universal, uncontroversial aspiration in war-torn and post-war Europe, and early legitimating discourses framed integration as indispensable to achieving the necessary growth. Moreover, ‘economic and social progress’ was indispensable for making integration sustainable, and thereby peace possible (CEC, 1960: 20). This line of argument transferred the existential indispensability of creating a working peace system to the (far from obvious) choice of economic integration as the way to do it. In the ‘indispensability’ discourse, potentially clashing national, partisan, and individual interests converged in one shared interest, which rested on the existential necessity of peace, economic growth, and hence the European Communities.
The discourse of the ‘common European interest’ was further associated with a discourse about the growing ‘interdependence’ of the West European nation-states. This discourse had the latter as a ‘community of fate’ with a ‘common destiny’, existentially dependent on unity (Monnet, Reference Monnet1962; Preamble ECSC Treaty). The reasons given for this included external factors such as Cold-War international relations (Hallstein, Reference Hallstein1959: 2; Monnet, Reference Monnet1962) and the progress of modern technology and mass production, which had turned national markets into an ‘anachronistic form of economic organization’ (Marjolin, Reference Marjolin1958: 4). In addition, the member-states were deliberately manipulating their opportunity structures in founding the European Communities, indissolubly entangling national interests in a shared European interest in a functionalist fashion (e.g. Schuman, Reference Schuman1950). The Preamble to the ECSC Treaty referred to this strategy as the attempt ‘to substitute for historic rivalries a fusion of their essential interests’.
Many early legitimating discourses, furthermore, placed great emphasis on hope, agency, and the determination to bring about a better future through political action (e.g. Hallstein, Reference Hallstein1951: 14; Martino, Reference Martino1957; Mansholt, Reference Mansholt1958). European integration was commonly celebrated as the ‘greatest voluntary and purposeful transformation in the history of Europe’ (Spaak, Reference Spaak1957; see Marjolin, Reference Marjolin1958: 1). It embodied a new type of politics: no longer ‘the art of the possible’, but ‘the art of the maximum possible’ (Hallstein, Reference Hallstein1959: 1). This way of seeing integration was embedded in a general confidence, in post-war Europe, in social engineering and progress on the basis of expert rationalities, planning, and impartial technocracy (see Walters and Haahr, Reference Walters and Haahr2005: 21–41). According to this mindset, good and legitimate government was government that was efficient, impartial, predictable, and helpful in solving concrete problems. This was often opposed to the ‘excited demands’, passions, and nationalist impulses associated with ‘politics’ (e.g. Haas, Reference Haas1963: 159).
In light of the period’s recent history, moreover, unobstructed mass politics and majoritarian democracy were potentially suspicious, as well as possibly harmful to performance efficiency. Democratic inputs, on this understanding, could be antithetical to output efficiency, which was the overpowering aspiration. Arguably ‘democracy’ emerged only gradually, over the course of the 15 years following Second World War II, as the key element of political legitimacy discourses in the member-states (Conway and Depkat, Reference Conway and Depkat2010). Of course, a contemporaneous counter-discourse did insist on democracy, including on supranational elements of it, as a condition for the emerging Communities’ legitimacy (see e.g. Sternberg, Reference Sternberg2013: 46–61). In addition, even output-biased claims to their legitimacy never really left questions of democratic input legitimacy out of the equation.
The need for input legitimacy, and the ‘eminently political’ nature of integration
Even where early legitimation discourses gave a central place to the delivery of efficient problem solving, they nevertheless often combined their claims to output legitimacy with some – possibly implicit – reference to input legitimacy as well. Efficient outputs such as ‘economic and social progress’ featured not least as means of ‘prov[ing] to its constituent peoples the advantages of integration’ (CEC, 1960: 20). This framed citizen approval and some level of input authenticity as necessary, if only for making integration sustainable, and the efficient delivery of other outputs, specifically durable peace, possible.
What is more, even the most heavily output-biased legitimation discourses depended on some claim to input legitimacy. The voluntary-transformation discourse, for one, implicitly depended on input legitimacy by way of the normative-legal principle of delegation discussed in second section. Even if eventually, once the new institutions were in place, ‘the moment of the technicians arrive[d]’ (Monnet, Reference Monnet1978: 321), the legitimacy of the institutions and their subsequent actions did rely on an input-legitimate initial act of authorization. By signing and ratifying the ‘outline’ treaties (Hallstein, Reference Hallstein1965), legitimate national democratic representatives had fixed the ‘general objectives’ for Community action (CEC, 1960: 19), and the ‘normative frameworks’ and ‘standards […] within which the Community [could] act’ (CEC, 1972: 17, see also CEC, 1958: 15).
More fundamentally, the very aim of the ‘common European good’ discourse was to make it plausible that European integration delivered essentially what the European citizens wanted. Much like Scharpf and Majone’s accounts, discussed above, it linked up with a particular conception of what the Europeans wanted, or should want; namely, peace and prosperity. Seemingly purely output-focused, the discourses establishing the European Communities as the best available way to achieving peace, prosperity, and generally efficient problem-solving, also laid claim to the input-authenticity or substantive representativeness of these outputs, that is, to their accordance with citizen preferences.
Indeed, early legitimation discourses tended to project integration as responding to a public consensus about its purpose, ends, and goals. This consensus supposedly arose from an insight into integration’s existential indispensability (e.g. Haas, Reference Haas1968: 456; CEC, 1976: 11). The dependency of output legitimacy on some claim to the input authenticity of the outputs at hand came to light particularly compellingly wherever such a consensus could not be taken for granted. Early integration history was full of illustrations that not everyone agreed about what the Communities should be doing and how (e.g. the failure of the European Defence Community, or the struggles of the 1960s over how supranational the Communities should be). Discursive projections of consensus were up against powerful counter-discourses that emphasized, instead, disagreement, pluralism of preferences, and clashing interests. These counter-discourses attacked representations of integration as a non-contentious win-win enterprise in everyone’s interest, where short-term ‘sacrifices’ would be ‘compensated’ by the commonly enjoyed peace and ‘the shared prosperity of tomorrow’ (Martino, Reference Martino1957; see CEC, 1972: 34).Footnote 4
Even if it could have been claimed that there had been at one time a consensus about the overarching objectives of integration, or if the treaties had somehow legitimately fixed them beyond the realm of contestation – there still remained ‘fundamental choices’ to be made about the ‘ways and means’ of pursuing them (CEC, 1960: 19; EPA, 1960b: 17). These choices were ‘eminently political act[s]’. Not only did their consequences benefit and tax some more than others (Hallstein, Reference Hallstein1965), they were also simply too important and ‘far-reaching’, and comprised too existential a ‘gamble on the future’ embracing ‘the whole economic life of our six countries’, to be left to ‘a handful of good experts’ who would ‘settle all problems to general satisfaction’ (EPA, 1960b: 17). Whereas technocratic discourses portrayed the nature and goals of the integration project as too important to allow the people to interfere with and possibly obstruct them, this counter-discourse held them as too important not to involve the people in their pursuit (EP, 1963b: 2). In brief, the Communities’ legitimacy could not rest on outputs alone.
Supranational parliamentary democracy: the solution to input-legitimacy deficits only?
The discourse of the ‘eminently political’ nature of Community politics marked much of the advocacy for European elections and EP powers of the 1960s and 1970s (as well as earlier). Often articulated with a critique of the technocratic approach, this advocacy linked up with the argument that legitimately taking the essentially political choices involved required some element of democratic input, and more specifically, an element of supranational parliamentary democracy. Versions of this argument sometimes appealed to entrenched beliefs in the legitimating power of representative democracy in the national context, but added that ‘the Community need[ed] to find its own democratic legitimation beyond that which can be transmitted to it by the governments responsible’ or the national parliaments (CEC, 1972: 12, 32; see e.g. EPA, 1960b; EP, 1963a, 1969). These discourses formed a prototype (Sternberg, Reference Sternberg2013: 57–58) of the later academic ‘classic democratic deficit theory’ (Dehousse, Reference Dehousse1995: 125). They constituted counter-discourses to the above output-focused legitimacy narratives from the very beginning of integration.
Interestingly, advocates of a strong and directly elected EP often demanded these supranational electoral-representative elements of input-legitimation at least partly on output-oriented grounds. To be sure, one role of the EP was to make executive policy making responsive to citizen needs and expectations. It was to keep the Communities ‘in close and permanent touch with political and human realities’ (CEC, 1961: 19). In this discourse, the EP featured as the ‘sounding board’ as well as the ‘stimulator of this public opinion’ (CEC, 1972: 34). European elections were attributed an almost magical power to mobilize popular support for integration and ‘directly to associate the peoples to the building of Europe’ (EPA, 1960a: 834, see EPA, 1960b: 16–17, 1963b: 25; CEC, 1961: 19). These discourses partly equated the Communities’ democratic legitimacy and input authenticity with popular support. Yet they represented popular support, in turn, as necessary not least for output-oriented reasons: the ‘active support of public opinion’ and the pressure of public opinion on political leaders and policy makers were needed for ‘sustaining’ or ‘advancing Europe’, and specifically for overcoming collective action problems and the ‘divergences and particularisms of the moment’ (CEC, 1958: 13–14; EPA, 1960b: 16–17). On this understanding, input legitimacy was required for making efficient outputs possible, by giving room to the ‘political will’ that provided ‘the only way out of dead ends’ once ‘the experts’ resources’ were ‘exhausted’ (EPA, 1960b: 16–17).
Finally, an elected EP would improve the Communities’ output efficiency by allowing it to capitalize on its ‘new political authority’, flowing from its democratic input legitimacy (CEC, 1976: 29). The EP was celebrated as a ‘motor’ of European integration (EP, 1963b: 2–4), or a counterweight to the ‘centrifugal tendencies’ plaguing the Community construction (EPA, 1960b: 16; see van der Stoel, Reference van der Stoel1976). A strong and elected EP could help to generate a ‘supranational will’ (EP, 1963b: 1). ‘Supranational’ was used here in the sense not only of standing above ‘national thinking’ (e.g. Hallstein, Reference Hallstein1955, Reference Hallstein1959: 2; Mansholt, Reference Mansholt1958), but also in that of wanting more supranationalism – which, in these discourses, would make the Communities more output efficient. Both discourses, of the EP as a stimulator of public support and a motor of integration, framed input authenticity as necessary for optimising the Communities’ problem-solving and policy output, making them sustainable, and helping them to ‘do their job well’.
In sum, input authenticity and output legitimacy were intertwined in discourses advocating a strong supranational parliamentary element for the Communities’ emerging political order. These discourses counter-balanced the output-focused ‘peace and prosperity’ and ‘European common good’ narratives from the very outset of integration.
The People’s Europe and governance discourses: overcoming the trade-off?
As with the more recent Eurozone crisis, the financial and economic crises of the 1970s dealt a blow to legitimacy claims founded on the provision of prosperity and peace in Europe. In the face of recession, many questioned whether economic integration actually was effective, and the only available choice, in seeking to safeguard peace. A concurrent debate on the ‘legitimation crisis’ of the capitalist welfare state underlined the danger of resting legitimacy exclusively on performance outputs (e.g. Habermas, Reference Habermas1973). The star of social engineering was fading, which further rendered the foundation of the Communities’ legitimacy on efficient, technocratic output performance a liability. In fact, claims to integration’s legitimacy were now increasingly threatened by the association of the Communities, and especially of its supranational elements, with ‘technocracy’ and legitimacy only through output performance. This was true especially when ‘politics’ and input legitimacy through representative democracy were associated with the nation-state (e.g. de Gaulle, Reference de Gaulle1965).
As a result of all this, the storyline of integration as furthering an uncontroversial European common good was crumbling. Integration had lost its ‘guiding light, namely the political consensus […] on our reasons for undertaking this joint task’ (CEC, 1976: 11). The Communities’ output legitimacy increasingly seemed to depend on input legitimacy as well, both as a way of riding out these times of performance difficulties and for endowing its outputs with some claim to input authenticity. Otherwise these outputs’ legitimating power would be lost, and the de-legitimating power of performance difficulties heightened. What the Communities did, and how they evolved, needed to rest on a plausible claim to be reflecting the needs, desires, and interests of the European citizens, if it were to pass as legitimate itself and to convey overall legitimacy on the Communities.
The ‘need to redefine the objectives of European integration’ in line with what would make its subjects endorse the process, became a frequent motif in discourses on the Community’s legitimacy in the late 1970s and the 1980s (here EP, 1984; see similarly Council, 1984; CEC, 1985, 1988a: 4). This motif formed part of a radical change of perspective in official legitimation rhetoric, which increasingly centred on the point of view of the European citizens: ‘We must listen to our people. What do the Europeans want? What do they expect from a united Europe?’ (CEC, 1976: 11). In the 1980s, this understanding found an emblem in the discourse and policies around the ‘People’s Europe’ campaign. The latter’s defining target stated that ‘the Community should respond to the expectations of the people of Europe’ (Council, 1984; see CEC, 1976: 13; 1985: 5).
On the face of it, this change of perspective constituted a turn towards input-based foundations of legitimacy, towards concern for input authenticity and substantive representation. Yet the ‘citizen expectations’ at the heart of this discourse played an ambiguous role. They featured as an independent source of legitimacy, but at the same time also as an object of top-down manipulation. Citizen expectations were to be moulded, for instance, through intensified, carefully tailored communication and information policies, the quantification of the ‘costs of non-Europe’, or cultural and identity-building policies (e.g. CEC, 1985: 10–11, 20, 1988b). On balance, legitimacy in the People’s Europe discourse still very much flowed from this Europe’s performing specific outputs. But the objectives of these outputs were now increasingly framed in terms of what the citizens wanted, of their ‘most immediate concerns’ or ‘deepest aspirations’ (CEC, 1973: I; Santer, Reference Santer1985; see Council, 1984: 11; 1983: 24). A certain misgiving about giving citizens too much agency – which they may well put to the use of obstruction – betrayed a persisting understanding that input authenticity and unmediated participation or even representation could be antithetical to output efficiency.
The discourse of ‘governance’, in turn, was championed by the Commission’s 2001 White Paper, in close dialogue with the academic paradigm of participatory democracy discussed above. It was one of the responses by the European institutions to the EU’s ongoing and much-deplored ‘legitimacy crisis’, ushered in not least by the difficult ratification of Maastricht.Footnote 5 The governance discourse had strong input legitimacy-related elements: ‘When we speak of “governance” we are, in fact, discussing democracy’ (Prodi, Reference Prodi2001; see similarly CEC, 2001: 32). Yet it aspired to a new type of democracy, superior to the classic majoritarian, electoral, and particularly, the parliamentary models of democracy, all of which suffered from a ‘growing crisis of faith’ and ‘disenchantment’ (Prodi, Reference Prodi2001; see Vignon, Reference Vignon2000: 4), and had led to the citizens’ ‘alienation from politics’ (CEC, 2001: 32). Governance by contrast embodied the ‘kind of democracy our fellow-citizens want’ (Prodi, Reference Prodi2001): a more “genuine” type of democracy that was ‘much more participatory, “hands-on”’ (Prodi, Reference Prodi2000). Like the academic ‘participatory turn’ discussed above, the governance paradigm rested centrally on the involvement and consultation of civil society organizations (CEC, 2001: 11–18). While the motivation for this consisted partly in ‘giving voice to the concerns of citizens’ and hence in increasing input authenticity, it also lay in more successfully ‘delivering services that meet people’s needs’. Explicitly: ‘Participation is not about institutionalising protest. It is about more effective policy shaping’ (CEC, 2001: 14–15).
In sum, both the People’s Europe and the governance discourses were in part attempts to explore and institutionalize new kinds of input legitimation; ones that would not endanger policy outputs and the progress of integration as a whole. Yet both discourses continued to subscribe to the output-oriented principle that ‘Effective action by European institutions [was] the greatest source of their legitimacy’ (Prodi, Reference Prodi2000; see e.g. CEC, 1995: 2, 5). The novelty was that they made more of a discursive effort to represent this action in reference to what the citizens wanted and needed.
Outlook and conclusion
The recent Eurozone crisis has once again placed output-based claims to the EU’s legitimacy under fire. In addition, fundamental disagreements and conflicts about how to address the crisis have raised questions regarding the input-legitimacy of any course of action, including its authorization, its accountability, or link to the will of European citizens. Eurosceptic parties are on the rise in many member-states. Even before the current crisis, gaining support for the draft constitutional and Lisbon treaties had proved extremely difficult. Debates in (and possibly among) member-state publics about these reform attempts, as well as about the current crisis, have turned on conflicting visions for their countries’ and Europe’s political, economic, and social futures. In short, during the past decade or so, EU politics and European integration have been subject to intense public and political politicization.
This suggests that interlocking input and output legitimacy might not so much require a simple match between outputs and citizen preferences, to which the People’s Europe discourse aspired – and which has shown itself downright impossible given the heterogeneity and fundamental clash of citizen preferences and interests. Rather, such an interlocking seems to require opportunities and structures appraising and channelling this contestation. The absence of a consensus on what kinds of policies the EU should deliver has become impossible to deny. The politicization of EU politics of the past few years can be read, moreover, as a forceful expression of people’s will to influence decisions about their countries’ and Europe’s future, and not only through the Habermasian ‘communicatively generated power’ of public deliberation, but also directly through the classic mechanisms of electoral party democracy (see Bellamy, 2010). Ironically, the polarization of public debate and popular opinion makes majoritarian democracy, which concentrates democratic input on elections, actually seem to imply less of a trade-off between input and output legitimacy than post-majoritarian modes of democratic legitimation, which had precisely been promoted as ways of overcoming the democracy/efficiency trade-off. Furthermore, the ongoing politicization of the EU’s policy direction, specifically, casts a critical light on non-majoritarian claims to input legitimacy that are based on the involvement of organized stakeholders. Rather, politicization may reassert the need for more universal or inclusive modes of representation. In any case, the legitimacy of any kind of action by the EU needs a strong foundation in concurrent claims to its input legitimacy.
In conclusion, both this paper’s conceptual discussion and its reconstruction of practical legitimacy discourses in the context of the EU institutions have suggested that input and output legitimacy necessarily belong together. Most arguments about the EU’s legitimacy, whether academic or practical, can be shown to involve both input- and output-related claims, whether explicit or implicit. Virtually all claims to one legitimacy type effectively require some claim or reference to the other type. On one hand, this can be a requirement in terms of the very logic of the claims about legitimacy they advance, in that one type of argument does not work without the other. This type of internal dependency works especially in one direction. Output legitimacy, in particular, requires some reference to the input authenticity of the specific ends and goals of performance outputs, be it through substantive or formal, majoritarian or non-majoritarian, representation. On the other hand, input and output legitimacy seem mutually dependent in that they complement one another. They each make up for weaknesses at the other end of the spectrum, especially during periods of difficulty. This is why most legitimacy discourses, both academic and practical, effectively cover both bases and lay some claim to both input and output legitimacy. The discursive practice of the European institutions, but also of academics and other discursive actors, notably in the national public spheres, continuously reconstructed what the legitimacy of integration and the EU might mean and to what extent it depended on input- vs. output-related factors, as well as how the two types of legitimacy related to each other. What emerges as continuous in the discursive history recounted in this paper is that input and output legitimacy essentially depend on each other, whether internally or with a view to the overall goal of laying plausible claim to reasonable legitimacy.
Acknowledgements
This paper, in various forms and iterations, has benefited from helpful comments by François Foret, Peter A. Kraus, Brigitte Leucht, Juan Diez Medrano, Kalypso Nicolaïdis, William Paterson, Janie Pélabay, David Robertson, Martin A. Schain, and Jonathan White, as well as the anonymous reviewers. Many thanks.
