Conceptualizing throughput legitimacy: Procedural mechanisms of accountability, transparency, inclusiveness and openness in EU governance

Funding information Matthew Wood acknowledges funding from the Economic and Social Research Council Future Leaders Fellowship scheme, Grant/ Award Number: ES/L010925/1; ESRC ‘Health Governance After Brexit: Law, Language and Legitimacy’, Grant/Award Number: ES/S00730X/1. Abstract This symposium demonstrates the potential for throughput legitimacy as a concept for shedding empirical light on the strengths and weaknesses of multi-level governance, as well as challenging the concept theoretically. This article introduces the symposium by conceptualizing throughput legitimacy as an ‘umbrella concept’, encompassing a constellation of normative criteria not necessarily empirically interrelated. It argues that in order to interrogate multi-level governance processes in all their complexity, it makes sense for us to develop normative standards that are not naïve about the empirical realities of how power is exercised within multilevel governance, or how it may interact with legitimacy. We argue that while throughput legitimacy has its normative limits, it can be substantively useful for these purposes. While being no replacement for input and output legitimacy, throughput legitimacy offers distinctive normative criteria— accountability, transparency, inclusiveness and openness— and points towards substantive institutional reforms.

It argues that in order to interrogate multi-level governance processes in all their complexity, it makes sense for us to develop normative standards that are not naïve about the empirical realities of how power is exercised within multilevel governance, or how it may interact with legitimacy.
We argue that while throughput legitimacy has its normative limits, it can be substantively useful for these purposes.
While being no replacement for input and output legitimacy, throughput legitimacy offers distinctive normative criteriaaccountability, transparency, inclusiveness and opennessand points towards substantive institutional reforms.

| INTRODUCTION
In recent years, the procedural aspects of legitimacy at the supranational level have garnered increasing attention. In international relations and regional studies, scholars have been concerned with how to assess the quality of supranational governance (see, e.g., Schmidt 2013; Rittberger and Schroeder 2016;Tallberg and Zürn 2019), while in international organization studies and public administration, scholars have focused on how to evaluate 'good governance' in complex multi-level systems (e.g., Fukayama 2016). In European Union studies as well, scholars whose work is centred on one of the most advanced forms of governance 'beyond the state' also focus on procedural legitimacy (Zürn 2000), but in this case they often use the term 'throughput' to complement the traditional systemic approach to legitimacy in the EU (Zürn 2000;Benz and Papadopoulos 2006;Bekkers and Edwards 2007;Risse and Kleine 2007;Schmidt 2013). Questions regarding how throughput (procedural) legitimacy relates to input (political) legitimacy and/or to output (performance) legitimacy have generated lively debates in recent years in EU studies. These are the focus of this symposium, with the case of the EU used as illustration.
Briefly defined, throughput legitimacy is a procedural criterion concerned with the quality of governance processes, as judged by the accountability of the policy-makers and the transparency, inclusiveness and openness of governance processes (Schmidt 2013). 1 In contrast, input legitimacy is a political criterion focused on citizens' political participation and governments' responsiveness while output legitimacy is a performance criterion encompassing policy effectiveness and outcomes (Scharpf 1999; see also Majone 1998;Bellamy 2010). In the literature, input and output legitimacy are often assumed to allow trade-offs, where good policy results can offset a lack of citizen participation, or vice versa, where bad results matter little if citizens have approved of the policy (Scharpf 1999). In contrast, there are no such trade-offs for throughput legitimacy: however high the quality of the governance processes, throughput is considered no substitute for input or output (Bellamy 2010;Schmidt 2013;Steffek 2019). It cannot make up for bad results or little citizen participation, whereas problematic procedures can throw into question political input and/or policy output (Schmidt 2013). That said, and as we seek to demonstrate in this symposium, throughput is an indispensable component of legitimacy, in particular in multi-level governance arrangements where political input may be far removed from the administrative process, or diffuse, and where policy output is indeterminate, in particular in the short term.
The symposium addresses both theoretical and empirical issues related to throughput legitimacy. This introduction provides an overview of the concept of throughput legitimacy, considers the interrelationship of throughput with input and output, and discusses a range of normative standards by which it can be evaluated. The main articles in the symposium divide between theoretical and empirical contributions. The theoretical contributions discuss the problems with regard to the framework of throughput legitimacy, raising questions about its applicability and normative appropriateness. The empirical contributions, in contrast, demonstrate that the concept of throughput is indispensable, using case studies to illustrate its usefulness in different EU institutional contexts.
The symposium has as its premise that, on thier own, input and output legitimacy are insufficient to evaluate multi-level governance arrangements. Because their normative standards apply either to the political pressures on or the policy outcomes from governance, they cannot provide standards of evaluation for how multi-level governance arrangements work in practice. The procedural quality of multi-level governance demands a different 'made-for-purpose' set of standards, which we group together under the concept of throughput. Electoral input based on citizen participation and representation or policy output leading to effective performance are both certainly necessary for governing activities to be legitimated, but they are not sufficient on their own. They can't account for legitimacy deficits linked to, for example, the potential 'corruption' of citizen input or policy output during the complex processes and procedures of policy deliberation (Kröger 2019), and the 'enclosure' or depoliticization of policy debates during policy deliberation (Coen and Katsaitis 2019). In order to address these and other issues regarding the processes of multi-level governance, in particular when these involve the exercise of power, another kind of legitimizing mechanism is required, throughput, with its own clearly specified normative standards of evaluation.
But the next question is whether throughput legitimacy is itself sufficient on its own to legitimate governance. This is where Parker and Steffek contribute critically to our theoretical understanding of throughput legitimacy, by questioning its normative foundations and emphasis on 'procedure' to the possible exclusion of the other legitimating mechanisms of input and output. The partial answer to such critiques can be found in Kröger and in Coen and Katsaitis. They apply tools and techniques connected with throughput legitimacy to concrete governance processes in ways that show how they can be used critically, and can even provide implicit grounds for substantive reform. In this introduction, we argue that while throughput legitimacy has its normative limits, it can be substantively useful to international organization and public administration scholars, providing critical analytical tools for assessing levels of accountability, transparency, inclusiveness and openness while pointing towards potential institutional reforms.
This introduction proceeds as follows. First, we outline throughput legitimacy and key aspects of the framework, including accountability, transparency, inclusiveness and openness. Second, we detail the charges against throughput legitimacy placed by Parker and Steffek's contributions. Third, we emphasize how Kröger and Coen and Katsaitis use throughput legitimacy concepts to produce substantive contributions to the critical analysis of EU institutions, with potential to inform action in ways promoting their democratization. We conclude by suggesting that, even while largely accepting the theoretical critique of the concept, throughput legitimacy retains significant potential as an analytical concept for critiquing, analysing and reforming multi-level governance.

| KEY ASPECTS OF THROUGHPUT LEGITIMACY: ACCOUNTABILITY, TRANSPARENCY, INCLUSIVENESS AND OPENNESS
The concept of throughput legitimacy has gained salience in multiple disciplinary areas for analysing the legitimacy of complex processes of governance at multiple levels: local, national and global. Scholars have applied it to analyse the legitimacy of local governance in theory (Haus and Heinelt 2005) and in practice, for example in local governance networks in Toronto and Calgary (Doberstein and Millar 2014), urban communities (Eshuis and Edwards 2013) and water governance (Van Buuren et al. 2012) in the Netherlands, wastewater management in Berne and Zurich (Lieberherr 2016), and flood risk management in Hamburg, Helsinki and Rotterdam (Mees et al. 2014). Throughput legitimacy has also been applied at the national level to examine the success or failure of crowdsourcing initiatives for legislation (Christensen et al. 2015) and the governance of parks and forests (Raitio and Harkki 2014) in Finland, oil and gas development in Greenland (Smits et al. 2017), implementation of the United Nations Framework Convention on Climate Change mechanism on reducing emissions from forests in Nepal and Papua New Guinea (Cadman et al. 2016), stakeholder engagement in marine protected areas in Canada (Dehens and Fanning 2018), and participation in onshore wind energy expansion policy-making in Germany (Fraune and Knodt 2017). Scholars of global governance have applied throughput legitimacy, alongside 'input' and 'output' legitimacy, to determine the legitimacy of global internet infrastructure (Take 2012), to assess the application of corporate social responsibility in global supply chains (Mueller et al. 2009), to evaluate the creation of global standards for nanotechnology regulation (Kica and Bowman 2012), and to examine transnational private governance and public-private partnerships (e.g., Wolf 2006;Dingwerth 2007).
Scholars of EU governance have also applied throughput legitimacy to a wide range of policy areas and administrative institutions and processes. Those concerned with governance have examined the throughput legitimacy of the EU's multi-level administration (Benz and Papadopoulos 2006), its democratic processes (Risse and Kleine 2007;Schmidt 2013) and its institutions such as the EC and EP, among others (Schmidt 2016;Iusmen and Boswell 2017;Fromage 2018;Fromage and Van den Brink 2018). In addition, EU scholars have considered throughput legitimacy in terms of policy-making in areas like nature conservation (Engelen et al. 2008) and by EU agencies such as the European Chemicals Agency ECHA (Klika 2015) and the European Food Safety Agency EFSA (Chatzopoulou 2015).
Throughput legitimacy is taken in this symposium to be an 'umbrella concept' for evaluating the legitimacy of complex processes and procedures occurring within the 'black box' of multi-level governance, as it processes input demands through governance institutions to produce policy outputs. Umbrella concepts encompass 'a broad constellation of factors, which may or may not be empirically interrelated' (Jackman 1985, p. 169). As such, throughput legitimacy consists of 'the myriad ways in which the policy-making processes work both institutionally and constructively to ensure the efficacy of [multi-level] governance, the accountability of those engaged in making the decisions, the transparency of the information and the inclusiveness and openness to "civil society"' (Schmidt 2013, p. 7). Taking the European Union (EU) as our example, throughput legitimacy can be used to assess its often opaque Community institutions and de novo bodies, including: the European Commission (EC), the EP, the Council of the EU and European Court of Justice, the European Central Bank (ECB) and the financial institutions focused on economic recovery in the euro crisis, along with the EU's multiple decentralized agencies. Institutions set up by the EU to promote interest group involvement with the EU policy process-the 'umbrella' groups-are equally relevant to the analysis of throughput legitimacy (Kröger 2019).
Throughput legitimacy has deep historical roots, as a procedural concept covering a range of normative standards of evaluation, including the principal kinds focused on herein, namely accountability, transparency, inclusiveness and openness. Accountability in the performance of public officials' ruling functions and their giving account and/or being held to account for their actions in public forums are traceable all the way back to Confucius and forward through Hegel's Philosophy of Right and Max Weber's legal-rational authority on to the vast public administration literature focused on these issues. Transparency in the provision of information by public officials was a major concern of Jeremy Bentham and Jean-Jacques Rousseau, while 'transparent management' has been central to good governance, coming into general usage in the 1990s (Hood 2010, p. 990). In contrast, inclusiveness and openness in decisionmaking are often considered in the context of pluralist theories of interest intermediation, and find greatest support from American democratic theorists such as David Truman and Robert Dahl. The theory of 'associative democracy' even sees this as another form of democracy in its own right, as well as a corrective to representative democracy (Cohen and Rogers 1992). Deliberative democracy has also been conceived of as a major procedural remedy to problems related to both input and output legitimacy, which may involve improving the quality of citizen participation via deliberative forums or introducing more deliberation in expert arenas (Bohman 1996;Goodin and Dryzek 2006).
Underpinning these four evaluative criteria of throughput legitimacy are other requirements, including the 'hard' criterion of legality and the 'soft' criterion of trust. Relevant actors generally need to be perceived to act legally within the rules and to inspire trust in those with whom they engage, such that they are believed to act with integrity and without bias so as to ensure equal and open access while meeting expected ethical and moral standards as well as legal ones (Levi 1998, p. 88;Offe 1999). Fairness, meaning enforcing the rules in such a way that they apply equally and appropriately to all, and doing this in a way that can be seen to be fair, is also a key component, and another defining attribute of throughput legitimacy (Franck 1995). As a result, central to throughput legitimacy are expectations about the qualities necessary to policy-makers, which can be summarized by such buzzwords as trustworthiness, integrity, fairness, impartiality and credibility.
Taking account of the quality of the governance processes, and not only the effectiveness of the outcomes and the involvement of the citizenry, is important for evaluating legitimacy at the supranational level. Adding throughput to the systemic mix of criteria for legitimacy is useful because it enables us to separate out issues regarding the procedural legitimacy of the governance processes from the performance legitimacy of the policies they may produce as well as from the political legitimacy of the representative politics by which those policies were generated. At the same time, bringing the different elements involved in assessing procedural quality under the single rubric of throughput legitimacy enables us to theorize their interaction effects even as it allows us to consider how the more general categories of input or output legitimacy relate to throughput legitimacy. Moreover, by shining the spotlight on the procedural aspects of legitimacy, throughput focuses attention not only on the quality of the processes but also on the qualities expected of the agents engaged in policy-making, such as trustworthiness, fairness, integrity, credibility, impartiality, competence and more.
Thus, when we analyse the throughput legitimacy of governance processes we are concerned with how accountable they are, how transparent their practices appear, and how inclusive and open they are to civil society groups in their constructive dialogues over policy decisions. This is the 'broad constellation of factors' we study when analysing 'throughput legitimacy' as an umbrella concept to assess the normative quality of multi-level governance processes.
It is also important to note that the criteria under this 'umbrella' concept can work in contradictory ways, or at crosspurposes. Increasing transparency may uncover a lack of inclusiveness and openness in stakeholder engagement processes, for example. Increasing formal accountability during decision-making may not go hand-in-hand with the inclusion of a wide range of societal stakeholders. Indeed, the opposite may be the case if accountability is viewed as a stronger determinant of 'good governance', while including too many of one kind of stakeholder (business, for example) may be detrimental. We must therefore ensure that throughput legitimacy is not viewed as a 'tick box' set of criteria. Rather, it ought to be viewed holistically as an overall normative good, with four interconnecting processes working to achieve that overall 'good' but not necessarily all in one place or all at the same time or with the same weighting. Beyond this, we should not forget that throughput legitimacy may also serve to enhance or undermine other legitimating criteria, such as citizens' political input or governing bodies' policy output. Finally, we emphasize how ensuring the throughput quality of the governance processes alone is never enough to guarantee an overall sense of legitimacy in citizens, for whom good policy output and/or sufficient political input are equally essential.
The articles in this symposium reflect this nuanced conception of how throughput legitimacy should be evaluated.

| Accountability
First, assessing throughput legitimacy involves analysing accountability mechanisms within governance processes.
Accountability as commonly defined involves the provision of information by an actor or organization, discussion or deliberation on that information, and the (potential) imposition of rewards or sanctions in cases of misconduct (Bovens et al. 2008). Moreover, where accountability is seen as a virtue of actors and not just as an evaluative political or administrative mechanism, it comes with a set of standards of conduct for good governance, including willingness to act in a transparent, fair, compliant and equitable way. Accountability deficits or 'bad' governance instead manifest themselves in 'unresponsive, opaque, irresponsible, and ineffective' behaviour (Bovens et al. 2014). As a procedural criterion of legitimacy, accountability means that technical and political actors can be held to account-by technical and political oversight bodies or other forums-for what they do as they engage in processes of governance. But it also means that actors must give account of their actions in those self-same forums as well as to the public.
In the supranational mode of governance, which is the domain of technical actors in non-majoritarian institutions or public administrations, accountability forums are generally assumed to be specialized, consisting of experts who assess the quality of their (throughput) activities and the effectiveness of their policies' (output) performance. The ECB, for example, seeks to build legitimacy via networks of economists, banking experts, and other central bankers.
But accountability may equally be established by forums made up instead of political actors, such as parliamentary bodies that are by definition input legitimate, which 'take account' not only of the quality of technical actors' (throughput) activities and the effectiveness of their resulting (output) policies but also how such policies resonate with citizen values and the common interest. The EP is the forum in which the ECB is charged to explain its actions and listen to concerns raised by MEPs, although this is a thin kind of accountability since the ECB need not respond to concerns raised by the EP, which cannot impose sanctions.
The EC, instead, is subject to full technical and political accountability in multiple forums. Although initially the Commission was solely accountable to the political accountability forum constituted by the Council, dominated by member state leaders, it increasingly over time became equally accountable to the EP, in particular ever since the appointment of the Commission President through the Spitzenkandidat procedure, as the leader of the winning majority in the 2014 election. At the same time, civil servants in the Commission and de novo bodies are mainly involved in the technical accountability process by consulting with experts and think tanks, producing reports and collecting data on their performance, which also feeds into their political accountability through their responses to the demands of committees of the EP and Council. This is what Bovens et al. (2010) call the 'real world' of accountability in the EU.
The European Council is another matter. As the most powerful EU political actor, the Council is subject to very little scrutiny from any EU forum when in intergovernmental mode. Unlike at the national level, where parliaments largely serve as forums of accountability for national executives, the EP has little de jure authority with regard to the Council, leaving the Council with no formal EU-level forum that can both hold it to account and to which it has to give account. This would not be a problem were accountability predicated solely on the (input-legitimate) linkages between member state leaders and their national parliaments or public opinion more generally. But this logic fails to deal with the fact that in practice very few national parliaments (with the notable exception of some Northern European legislatures) are able to hold their executives to account (Auel 2007;Crum 2017;Kreilinger 2019). But even if national parliaments were able hold their executives to account, this would not be sufficient to guarantee the Council's EU-level accountability because most collective EU-level decisions go beyond the aggregation of member state governments' individual interests in ways that cannot be adequately assessed by any individual national parliament on strictly national accountability grounds (Crum and Curtin 2015, pp. 78-79).
So how to judge the Council's accountability? Only if, following theories of deliberative democracy (e.g., Bohman 1996), especially when applied to the EU (Bickerton et al. 2015), we were to argue that the Council can achieve a kind of mutual accountability as a deliberative body, with its members holding one another accountable for their decisions. But this assumes that deliberation meets certain standards, in particular that it proceed without major inequalities in the exercise of power or voice, or at least that these are balanced out in such a way that member states don't feel unduly disadvantaged. In many domains, this may be the case. It remains in question with regard to Council decision-making during the Eurozone crisis (Schmidt 2019).
In our discussions of political actors such as the Council or the EP, it is also important to underline the distinction between procedural (throughput) accountability and political (input) accountability. With the concept of accountability for elected officials, we need to be careful to differentiate between political actors being 'held accountable' by citizens through elections (input legitimacy) and by their reason-giving in public forums (throughput legitimacy) (Borowiak 2011;Crum and Curtin 2015;Wood 2015). Crum and Curtin (2015) explain the difference as one in which accountability in public forums is 'the ex post complement to the ex ante mechanisms of democratic election or authorization through which executive actors are initially appointed'. In other words, accountability in the throughput sense comes after input-legitimate elections, and applies to the ways in which elected officials are expected to justify their exercise of power in public forums, most often parliamentary but also in the forum of public opinion, with the understanding that they will be judged on the basis of whether that exercise serves the popular constituency (Crum and Curtin 2015, pp. 64-66).
Accountability is not just about rendering account to specialized forums, though. It is also about making them public. In the EU, the increasing recourse to hearings held by the EP with the other institutional actors represents a kind of accountability forum. But this alone is not enough to ensure accountability to the public. The paradox is that the Commission's increasing attention to internal accountability has not solved 'the problems of rendering accounts externally' (Wille 2010), or its invisibility, remoteness and seeming unaccountability to the public (see also Schmidt 2013). But this is also where transparency comes in.

| Transparency
Transparency is closely related to accountability, but is not coterminous. It has long been seen as a key accompaniment to accountability in EU governance, as matching parts producing good governance only when in combination, or even as 'Siamese twins' (Hood 2010). Transparency means that citizens and political representatives have access to information about governance processes and that the processes along with the resulting decisions are public (Héritier 2003;Novak 2013). Transparency serves to ensure that the public knows that political and technical actors not only follow ethical standards but also that they are doing what they say they are doing, such that their discourse can be measured against their actions, and that citizens can hold them accountable for what they say and what they do. In transnational democracy, moreover, transparency can equally be seen as a component of citizen empowerment (Smith 2012). That said, it is important to remember that transparency is no substitute for accountability, since although account may be given and discussion may follow, nothing may subsequently happen (Papadopoulos 2010(Papadopoulos , p. 1034).
In the EU, transparency refers to when EU institutions make available information about their internal processes (as opposed to providing it to the Commission or Parliament as part of formal accountability processes). Such practices can also fit within an output legitimacy approach, focused on the production of documents concerning effectiveness, which enable citizens, the media and various 'sleaze-busters' to evaluate the quality of policy implementation. This focus, however, does not capture the extent to which such documentation is often required throughout the process of policy-making, for example in documentation about the criteria civil servants use to structure their decision-making procedures. This latter focus is where throughput legitimacy is centred.
The problems with regard to transparency in the EU are many. Often, the increased access to information results in less rather than more transparency, because citizens find it difficult to navigate the mountains of data available via the internet (Héritier 2003). Moreover, agencies are often reluctant to release information, sometimes for privacy reasons, other times for fear of negative spillover effects. For example, the European Stability Mechanism is deliberately lacking in transparency, as Ban and Seabrooke (2017) illustrate, mainly because of fears that providing information on their actions would tip off the markets, thereby hurting the very countries they are trying to help through their transactions. In so doing, their legitimacy concerns are more focused on output legitimacy, to produce good results for the countries concerned, as well as on throughput accountability to their 'principals' in the Eurogroup.

| Inclusiveness and openness
Inclusiveness and openness are found in 'the intermediation processes through which citizens organized in interest groups have a direct influence on policy making' (Schmidt 2013). Openness means that members of the public, whether as individuals or in organized groups, have access to policy-makers regarding the policies in which they are most interested. It can be defined as the extent of opportunities for non-state organizations and individuals to become involved in EU consultations and civil society umbrella groups. Inclusiveness means that policy-makers are open to all such groups, and bring them in in such a way as to ensure balance in their representation. Inclusiveness can therefore be defined as the breadth of non-state organizations and individuals involved in EU consultations and civil society public and private interest groups. This is an area that can be seen to overlap with input legitimacy because interest groups are taken as representing their constituents' interests. But such 'pluralist' inclusiveness and openness is not the same as political representation. Interest groups are engaged in processes of representation when they organize grassroots letter-writing campaigns, bring busloads of farmers to national capitals or Brussels to protest against a new bill, organize demonstrations of mothers for peace, and represent 'civil society', meaning citizens, in public interest-related activities meant to influence elected officials. They are involved in procedural inclusiveness and openness when they become part of an elaborate interest intermediation process focused on policy-making, in which they lobby political and technical officials, testify in committee meetings and parliamentary hearings, provide informational evidence to administrative functionaries, and serve as counterweights to one another as they seek to influence policy formulation and implementation. Ensuring such inclusiveness and openness is not always easy. There is a vast literature on the problems faced by administrative bodies and regulatory agencies with regard to interest intermediation, in particular because of the dangers of agency 'capture' by special interests and of client politics (e.g., Bianculli et al. 2014).
In the EU, efforts to ensure the openness and inclusiveness of its various institutional bodies have grown over the years, in particular with regard to 'civil society' (a term deliberately used in order to suggest input responsiveness) (Smismans 2003). Legitimacy here has generally been tied to questions regarding the balance in access and influence among organized interests representing business versus those representing unions or public interest organizations (Greenwood 2007), although it can also be related to the kind of trust-based relationship demanded by the Commission of all participants in ongoing consultation processes (Coen and Richardson 2009). But whatever the good intentions, EU access and inclusiveness remain limited (Kröger 2008), especially because of the difficulties of transnational mobilization (Della Porta 2009).

| CRITIQUES OF THROUGHPUT LEGITIMACY
Whatever the benefits we have just outlined with regard to defining throughput as a separate procedural mechanism for assessing legitimacy, it is both necessary and appropriate to hold it up to rigorous scrutiny. Moreover, with the rise of deliberative and participatory theories of democracy that are often used within the throughput legitimacy framework outlined above (Pateman 2012), it is useful to question the key assumptions and effects of these theories, in particular with regard to their usage in international organization and public administration. We therefore welcome the important theoretical challenges by Steffek and Parker to throughput legitimacy in this symposium, and respond to them here.

| Throughput legitimacy: past, present and future
Parker charts a genealogy of deliberative democracy's popularity by reflecting on its origins within the EC's Forward Studies Unit (FSU), perhaps the most influential in-house think tank created in the 1990s to develop a more deliberative and participatory approach to EU multi-level governance. Parker's key point is that the FSU started out making extensive and radical proposals that were significantly watered down in practice with the introduction of the Open Method of Coordination (OMC) and the 2001 White Paper on Governance. Using the tools available from a Foucauldian account of power and ideas, Parker shows convincingly that, when translated into the hard political realities of multi-level governance, 'What can be said in these ostensibly deliberative forums is linked, of course, to who is able to partake in them' (p. 8). Taking the OMC as key example, Parker observes that 'in practice, inclusion via the OMC does not equate with deliberative governance in the inclusive way in which it is conceptualized by the FSU; peerreview processes may seem to include an array of actors, but power is very much in the hands of member state governments' (p. 8). Parker's argument is that deliberative governance is inherently exclusionary in the sense that the image of an 'array' of stakeholders or publics will necessarily be limited in practice due to the requirement for consensual decision-making: the ontological centrality of conditional consensus to even an abstract and theoretically reflexive discourse of deliberative governance such as that proffered by the FSU necessarily implies some form of exclusion. (p. 9, italics added) Attempts at instituting deliberative governance that satisfy throughput legitimacy criteria are thus, in many respects, bound to fail on Parker's terms, especially when instituted at the transnational level. This does not mean that the normative ideals of deliberative democracy are undesirable, but rather proponents of concepts linked to itinclusiveness and openness in particular-should be wary that attempts to translate this institutionally are prone to violate the very criteria they espouse.

| The turn to proceduralism
While Parker restates some of the theoretical limits of deliberative democracy 'classically' conceived, Steffek's argument is grounded in a normative interrogation of what he calls 'the turn to proceduralism' more generally: In the attempt to establish throughput as a separate and independent type of legitimacy we thus can see the broader trend to proceduralism at work-a remarkable transformation of the concepts and imagery of modern politics and administration. (p. 3) Steffek argues that proceduralism does not add anything substantive to existing normative theories derived from input and output legitimacy criteria. Indeed, put in its strongest terms, his critique is that 'Throughput legitimacy … does not prompt us to formulate critical questions of [a] quite fundamental sort but requires political institutions to comply with procedural standards of good governance' (p. 11). Moreover, he 'wonders' whether 'popular disaffection with European governance is really rooted in concerns about procedure' (p. 11). While Parker maintains that a deliberative approach can provide desirable normative standards, Steffek's charge is that throughput legitimacy looks in the wrong place for improving democratic legitimacy. He sees it as a hybrid formation of input and output legitimacy that does not have the normative force of either, and projects a way of analysing governing institutions that is at once oblique and overwrought. In short, focusing on improving procedure will not suffice in an era of multiple clashing crises. Steffek hence provides a powerful challenge to research in this area, when he states that: European governance fails to efficiently address current political challenges, such as mass migration and the persistent economic stagnation of large parts of the continent. This brings about a fundamental legitimation crisis of the European integration project as populist movements in virtually all regions of the Union, from Finland to Greece, refer to cultural and religious values, to national identity but also to material justice and fair distribution. Responses to such challenges are unlikely to be found in the procedural fine-tuning of the European political machinery. (p. 11)

| RESTATING THE VALUE OF THROUGHPUT LEGITIMACY
These challenges force us to reflect on why throughput legitimacy is a valuable concept. Our argument in response is that both Parker and Steffek place too heavy a burden on throughput legitimacy to provide foundational criteria by which legitimate governance can be built, or grounds for robust normative critique. As this introduction argues, throughput legitimacy is not meant to stand alone. High-quality throughput processes can only serve as a complement to, not a substitute for, good policy output and adequate political input in any democratic system. Moreover, as the empirical articles in this symposium show, throughput legitimacy can indeed be put to work as a criterion for legitimate governance in a way that shows the limitations, and potential, of maintaining high procedural standards of governance in existing supranational institutions. While these standards are never perfect, the value in taking account of them lies in allowing us to better explore all aspects of democratic legitimacy within the messy institutional processes and procedures of European multi-level governance. As such, the various criteria encompassed by our umbrella concept of throughput legitimacy are both normatively potent and empirically flexible, and a necessary criterion for studying such complexity, we suggest.
First, Sandra Kröger's article provides a compelling rationale for why we need throughput legitimacy. She argues that 'Assessing the quality of EU governance processes needs to include an assessment of whether the actors present in those processes have a reliable link to their domestic constituencies, without which they are likely to feel less bound to the related inputs, making it more likely that the input will not come out as "uncorrupted output"' (p. 2).
Assessing the representativeness of actors at the supranational level is thus crucial: 'throughput legitimacy needs to be conceptualized more in terms of a two-level mechanism than has hitherto been the case' (p. 2). If interest groups claiming to be representative of publics at the national level become influential in consultation processes without a direct electoral link, then legitimacy may be said to be compromised during the process of consultation, thus damaging Schmidt's criterion of 'inclusion'. Kröger's survey and interview data, which show 'a degree of structural remoteness', particularly for anti-poverty and environmental groups, has a normative force to it and suggests the need for institutional change in the form of improving representation between supranational interest groups and those they represent at the national level. Kröger's article thus shows that to be fully throughput legitimate, interest groups need not just have barrier-less access as part of pluralist processes of interest intermediation but that they also need to remain closely linked to their membership in order to ensure that they are accountable to their members, by representing them to their members' satisfaction.
Coen and Katsaitis capture a similar feature of the processes and procedures of European governance in their analysis of those who give evidence to select committees in the EP. Their rationale is again focused on delving into the intricacies of these committees to reveal how inclusive they really are, against the idealized image painted by the EP. They aim to: Assess which interest groups access the EP's deliberative processes, focusing on its hearings as an understudied area of European governance. Committee hearings are one of the few forums where the Parliament invites organizations to provide information and deliberate on policy issues. Two connected questions emerge: How open are hearings to interest groups? Do they include some interest groups more than others? (p. 2, italics added) Their analysis shows that think tanks are overwhelmingly the most common actors involved in committee deliberations, and that this tends to have a simultaneous effect of both establishing an image of inclusiveness, while also having a depoliticizing effect. Hence, they find both an effort at deliberation and also a process of 'enclosure'. As deliberations become more specific, actors without the requisite 'expertise' are crowded out of committee hearings and the deliberations become 'depoliticized'. Coen and Katsaitis reflect that the concept of throughput legitimacy helps to disentangle and account for the complex dynamic of 'enclosure' they reveal: Polities such as the EU generate hybrid forms of deliberation, seeking to tick the boxes of different criteria within a specific procedure at the same time. Conceptually, this demands frameworks that can encompass potentially competing schools of thought in the same space. The legitimacy of these procedures cannot be understood via dichotomizing mechanisms, but through an analytic lens that can evaluate different legitimacy objectives. (p. 14) The implication of their argument is that committees ought to be careful when creating procedural constraints on debate that set prerequisites on knowledge or expertise, and instead favour an inclusive style of determining what expertise means, and who possesses it, particularly in the context of contentious policy deliberations. Empirically, then, Coen and Katsaitis show that a deeply anti-democratic favouring technocratic governance in the EU is manifested empirically within the committee hearings of the EP. Moreover, they show where reforms might be introduced to ensure that, practically speaking, a genuinely more inclusive deliberative approach might emerge. While Parker (2019) rightly points out that there will necessarily be limitations in such formal arenas, the analysis they provide offers an important starting point for reform.
Do we need the concept of throughput legitimacy for Kröger and Coen and Katsaitis to make their contributions?
If we return to Steffek's key critique that procedural and processual inequities are neither the cause of Europe's widespread malady of democratic disaffection, nor is a focus on those processes and procedures the primary solution, we might argue in return that neither empirical study in this symposium provides, or can claim to provide, such a broad analysis. What they do provide is a careful empirical assessment of the problems of throughput legitimacy in European-level public administration along with ideas for potential reform. These may not be prescriptions capable of curing widespread democratic disaffection, but is this really the bar we want to set ourselves for judging the relevance of these studies?

| THE STAKES ARE HIGH
The question this symposium addresses is whether an input-output dichotomy is sufficient to capture and normatively evaluate complex supranational administrative and representative institutions. Certainly, the input-output dichotomy is valuable and straightforward, as Steffek argues. He suggests that a turn to 'proceduralism' obscures the crucial importance of these criteria, in particular today. However, we might ask in response, is it sufficient to evaluate systematic efforts at coordinating complex supranational policy deliberations and administration through such powerful, yet crude, indicators? If we were to simply use input legitimacy as a standard for EU governance, for example, it is almost destined to fail because of the very nature of legitimacy in the EU, split between 'input' that is most robust at the national level, where citizens elect their representatives, and 'output' performance that is generated via throughput processes at the EU level (Schmidt 2019). Moreover, how could we use input legitimacy as the normative standard by which to evaluate how centuries-old ethno-nationalist cleavages diminished enough to produce shared standards for market supervision, trade and human rights? We cannot assess this without turning to the different criteria of throughput legitimacy, to evaluate the supranational public administration that consists of the 'Bureaucrats in Brussels'. But of course, we cannot do so without an understanding of the forces to which Brussels needs to respond, in which national politics has turned increasingly populist. As Jennings and Lodge (2019) argue, the recent Brexit vote and the rise of ethno-nationalist populism across Europe leave scholars of public administration with difficult choices and responsibilities that we cannot wish away.
The broader relevance of this introductory article (and the symposium more generally), beyond the EU, is to explain why using the concept of throughput legitimacy can be useful for international organization and public administration scholars. This article delineated and defended four criteria that can be studied under the broad rubric of throughput legitimacy, viewed as an 'umbrella concept' that constitutes an important topic of study distinct from input and output legitimacy. These criteria are diverse enough to provide room for a range of scholars interested in the quality of procedures and processes of multi-level governance to identify with; but they are also specific enough to connect with bodies of substantive theoretical work as well as empirical research with rigorous methodological frameworks. This symposium as a whole provides a potential organizing rationale for scholars working on accountability, transparency, inclusiveness and openness in multiple domains to see and justify their work as contributing to the broader aims of evaluating the quality of governance processes according to normative standards alongside 'traditional' input and output legitimacy standards. This can help, for example, to address concerns expressed in recent literature on accountability about the normative orientations and practical applicability of accountability research (see for example, Schillemans 2016). Moreover, studies of stakeholder engagement and transparency, which often contain implicit appeals to normative values, can make these evaluative-normative claims clearer by linking to throughput legitimacy.
Ultimately, in our view, the concept of throughput legitimacy is valuable because it allows us to create clear criteria for legitimacy that simultaneously reflect and account for the necessary complexity involved in (attempting to) construct high standards between and across multiple levels of governance. This is a project about which, we believe, scholars of public administration ought to be sympathetic. In Sandra Kröger's words, our task in studying multi-level governance is to make sure that inputs to supranational institutions do not come out as 'corrupted outputs'. Moreover, in Coen and Katsaitis' words, with throughput and its attendant criteria of inclusiveness, openness, accountability and transparency we are able to create 'an analytic lens that can evaluate different legitimacy objectives'. In this regard, it could be argued that the concept of throughput might be better seen as an 'umbrella concept' accompanying the other two such umbrella concepts of input and output legitimacy, as criteria for dynamic and 'multi-centric' processes of policy formulation and implementation (Cairney et al. 2019). As such, normative disputes about justifiability might be replaced with debates about empirical foci. Where and how do we apply normative criteria for evaluating complex governance processes in light of their (relatively newly acknowledged) complexity and multidimensionality? While we should be aware of the limits Steffek and Parker note, Kröger and Coen and Katsaitis provide reasoned and empirically rigorous studies that show the way towards-admittedly incremental and partial-reform.
They also acknowledge the useful comments of Martin Lodge and the anonymous referees at Public Administration.
Matthew Wood acknowledges funding from the Economic and Social Research Council Future Leaders Fellowship scheme and Governance After Brexit grant 'Health Governance After Brexit: Law, Language and Legitimacy'.