EU agencies’ stakeholder bodies: vehicles of enhanced control, legitimacy or bias?

ABSTRACT

EU agencies are increasingly subject to a flurry of stakeholder bodies. Despite their prevalence, and the considerable variation in structures formally professed to serve the same purpose, we know little about the actor preferences driving the set-up of such structures or the potential implications of specific institutional design choices. We systematically map structural variations across EU agencies and analyse to what extent the establishment and the design of stakeholder bodies is principal-imposed or agency-initiated. Do stakeholder structures enhance political control serving to broadly legitimise agencies, or to the contrary, do they reflect preferences of the bureaucratic actors they are meant to control, and with what implications? We find that, for the most part, weak principal control and steering leaves it to the agencies themselves to design stakeholder bodies as they see fit. This has the potential to introduce unsanctioned biases in favour of specific groups, potentially depleting rather than bolstering legitimacy. A major implication of EU agencies’ stakeholder engagement is that the agency model is currently in flux, moving away from the classic insulated agency towards greater politicization in regulatory policy.

KEYWORDS:

• EU agencies
• stakeholder engagement
• stakeholder bodies
• accountability
• regulation
• politicization

Introduction

EU agencies are increasingly subject to a ‘flurry’ of stakeholder bodies of different shapes and sizes. Despite their prevalence, and the considerable variation in structures formally professed to serve the same purpose (i.e., stakeholder engagement), the actor preferences driving the set-up of such structures and the implications of specific institutional design choices remain uncertain. Prior research suggests two dominant logics at play (Arras & Braun, 2018): Stakeholder bodies may serve as administrative controls formally imposed by the legislator to enfranchise particular interests by design, enabling legislator control (McCubbins et al., 1987; Kelemen, 2002). To the contrary, they can instead reflect bureaucratic preferences for the set-up and the structuring of such bodies – pro-active agency-initiated efforts, where the agency rather the legislator is in the driving seat (Arras & Braun, 2018; Potter, 2019; Bertelli & Busuioc, 2020). We therefore ask: Which actor preferences do EU agencies’ stakeholder bodies reflect? Are stakeholder structures designed by the EU legislator – and as such reflect preferences of a democratic body and contribute to political control – or are they agency-initiated instead i.e., reflect preferences of the very bureaucratic actors they were meant to control?

Why should one care? – the reader might ask. Because procedures matter for substantive outcomes (McCubbins et al., 1987). Stakeholder bodies are a type of procedural requirement for participation and engagement. While the ‘benign’ traditional view of administrative procedures such as these has tended to view them ‘as means of assuring fairness and legitimacy in decisions by administrators’ (McCubbins et al., 1987, p. 244), in a series of seminal articles in the US context, it has been convincingly demonstrated that: ‘this is not all there is to procedures. (…) alterations in procedures will change the expected policy outcomes of administrative agencies by affecting the relative influence of people who are affecting by the policy’ (McCubbins et al., 1987, p. 254 emphasis added).

By empowering and giving access and ‘voice’ in agency processes to specific groups, and not to others, administrative structures can favour particular interests and in doing so, affect substantive regulatory outcomes. When these structures reflect the legitimate intent of the legislator, channelling decision-making towards outcomes preferred by political principals, they serve to effectively allow for legislative control post-delegation. However, when these structures reflect unsanctioned bureaucratic preferences, caution is warranted as through their inclusion – and simultaneous exclusion – function of some interests over others, such structures necessarily introduce unsanctioned biases benefitting specific groups and raising the prospect of coalition-building away from principal preferences and enhanced autonomy, or even capture.

These concerns can become particularly relevant in the case of EU agencies’ stakeholder bodies which have been described as ‘closed instruments’, where ‘access is restricted’ with a ‘limited set of stakeholders’ (Arras & Braun, 2018, p. 1261). Given this power of procedures for substantive outcomes, as well as the qualitative increase in EU agencies’ powers in recent years (Busuioc, 2013a), it is important to study who shapes such procedures and whose preferences they reflect.

In an extensive qualitative study of 34 EU agencies, we systematically map structural variations across agencies and study agency set-up of stakeholder bodies, above and beyond fiat. We analyse to what extent the establishment and the composition of EU agencies’ stakeholder bodies is principal-imposed or agency-initiated. Based on the results of our inquiry, and in light of stakeholder bodies’ prerogatives vis-à-vis the agency, we broadly reflect on the potential implications of these procedural choices: Do stakeholder structures stand to contribute to political control and serve to broadly legitimise EU agencies, or to the contrary, are we witnessing the introduction of principal-unsanctioned structures, potentially stacking the deck and biasing agency-decision-making in unanticipated ways? We focus on durable, institutionalized (as opposed to ad hoc) structures for engaging with stakeholders – stakeholder bodies – and draw on a variety of legal and policy documents such as agency constituent acts, internal rules of procedure, agency websites and annual reports as basis for our empirical investigation.

This is, to our knowledge, the first study to zoom in exclusively on EU agencies’ stakeholder bodies, building upon previous studies of EU agency stakeholder involvement and stakeholder interactions, more broadly (Arras & Braun, 2018; Perez Duran, 2018; Wood, 2018). We go deeper by examining not only the set-up but also the details of stakeholder bodies’ composition (cf. Arras & Braun, 2018) – which we regard as critical to assessing actor preferences – and include in our study a broader array of stakeholder bodies, including those that are voluntarily established by the agency (cf. Perez Duran, 2018). On this basis, and in light of stakeholder bodies’ prerogatives vis-à-vis the agency – i.e., span of access afforded to stakeholder bodies and role (or tasks), we next reflect on implications thereof.

We find that stakeholder bodies are mostly required by the principal, although voluntary set-ups are also common. Even when mandatory, the composition of stakeholder bodies has been sparsely regulated by the principal. As a result, stakeholder bodies have emerged in a disjointed fashion across EU agencies, for the most part, reflecting a lack of principal steering. As principal preferences remain unarticulated, this leaves room for bureaucratic actors to fill these in, reflecting bureaucratic preferences instead. Moreover, under a combination of legislator-engendered ambiguity and agency-driven initiatives, a whole range of core agency activities and key agency structures have been opened up to societal and interest group input. A major implication of EU agencies’ stakeholder engagement is that the agency model is currently in flux, moving away from the classic insulated agency. Heterogeneous stakeholder bodies entail diverse departures from insulation.

EU agencies and stakeholder involvement: insulation and engagement?

Insulation from politics, more specifically from democratically elected institutions, has been a key rationale for EU agency creation: ‘The independence of their technical and/or scientific assessments is, in fact, their real raison d’être. The main advantage of using agencies is that their decisions are based on purely technical evaluations of very high quality and are not influenced by political or contingent considerations’ (Commission of the European Communities, 2002, p. 5, emphasis added). As seminally noted by Majone (1999, p. 4), ‘democratic politicians have few incentives to develop policies whose success, if at all, will come after the next election’, leading to ‘short-termism’. It is its ‘non-majoritarian’ insulated nature that has been the source of the agency model’s added value in regulation and the source of its credibility and legitimacy: ‘Independent expertise is ‘the be-all-and-end-all’ criterion for legitimacy in the regulatory state’ (Busuioc & Rimkutė, 2020a, p. 1260).

At the same time, the agency model’s very claim to legitimacy simultaneously became the source of pressures in the opposite direction: the need for societal engagement (Busuioc, 2013a). EU agencies’ insulation – the ‘selling point’ of the agency model – created corresponding compensatory pressures for the introduction of mechanisms for enhanced legitimation: ‘the post-delegation efficiency and effectiveness of those agencies is largely contingent upon their capacity to institutionalize participatory regulatory networks in order to secure input legitimacy and generate trust among major stakeholders’ (Borras et al., 2007, p. 584). This became especially pressing in light of the contested nature of many agencies’ roles and responsibilities.

Reflecting such concerns, an enhanced parliamentary role in agency design translated into stronger accountability mechanisms (Font & Perez-Duran, 2016) and devices for societal involvement. There has been a consistent institutional push from the EU legislator for stakeholder involvement (Kelemen, 2002), as evidenced by the variety of stakeholder structures and mechanisms provided by formal design in EU agencies’ mandates and by the importance of stakeholder involvement (re-)iterated in recurrent (inter-)institutional documents. Such mechanisms range from management board presence, stakeholder bodies, to public consultations. At the same time, agencies themselves also have pro-actively initiated a variety of stakeholder engagement and entrepreneurial activities (Wood, 2018), beyond formal requirements, voluntarily wrapping themselves in new engagement ties.

As durable structures of engagement, stakeholder bodies are an especially important manifestation of these patterns. These bodies effectively institutionalize engagement with stakeholders – i.e., non-state organized actors – within agency structures and provide for specific stakeholder groups’ input into agency activities and other output (ranging from annual reports to draft rules in some agencies). Moreover, they facilitate stakeholder access to the agency and its core structures (for instance, to the board or the executive director). The set-up of stakeholder bodies is provided for in the founding regulation of some agencies, while in other cases, stakeholder bodies have been initiated by the agencies themselves, in the absence of such requirements. Moreover, when the structures are mandated by the legislator, their design such as the ‘details’ of their composition can be specified to a greater or lesser extent (Perez Duran, 2018), leaving (more or less) room for the agency to fill in the details. This differentiates them in important ways from permanent structures for involvement such as management boards of EU agencies, whose set-up and composition is principal-stipulated as a matter of constituent act.

While scholars have debated whether functional (Blauberger & Rittberger, 2015, 2017) or political (Kelemen & Tarrant, 2011, 2017) reasons explain the establishment of EU agencies, they converge on the role of actor preferences for explaining variations in agency design. The establishment and composition of a stakeholder body is part of the broader agency design, and we therefore analyse whose preferences are reflected in EU agencies’ stakeholder bodies.

Rationale and implications: a theoretical framing

Seminal literature teaches us, as discussed above, that procedures such as these can be important mechanism for legislative principals to exercise control over bureaucratic bodies, to solve ‘prototypical problems of political control’ (McCubbins et al., 1987, p. 244). In other words, that they reflect a principal control logic. As mechanisms that enfranchise or provide a privileged position to specific groups by design within agency structures, they serve as means for the principal to ‘stack the deck’ in favour of specific (politically-relevant) constituencies: ‘[B]y controlling processes, political leaders assign relative degrees of importance to the constituents whose interests are at stake in an administrative proceeding and thereby channel an agency’s decision toward the substantive outcomes that are most favoured by those who are intended to be benefited by the policy’ (p. 244). Moreover, such mechanisms help the principal oversee the agency. Direct monitoring is costly, and these groups serve to alert the legislator – sound the ‘fire alarm’ – in case of agency non-compliance, allowing for reduced monitoring costs for the principal of agency compliance.

In line with this logic, it is well-recognised within EU regulatory state literature that stakeholder involvement procedures embedded in EU agency structures can reflect parliamentary efforts to exercise control (Kelemen, 2002; Perez Duran, 2018; Arras & Braun, 2018), supplementing existing parliamentary oversight tools vis-à-vis EU agencies such as budget discharge powers or written questions (Busuioc, 2013a; Font & Perez-Duran, 2016). In the words of Kelemen (2002, p. 104):

recognising that it had limited resources to conduct ongoing, direct monitoring, the Parliament pressed for the establishment of formalised, open, transparent administrative procedures that would create opportunities for its interest group allies to engage in indirect, ‘fire-alarm’ oversight and control.

This logic would be expected to have become especially prevalent in recent years as prior literature has linked the introduction of such mechanisms to the EP having co-legislator rights. In recent years, following the entry into force of the Treaty of Lisbon, the EP has gained co-legislator rights vis-à-vis most EU agencies bar three (see Table 1 in Appendix).

Expectation 1: The establishment and composition of EU agencies’ stakeholder bodies reflect a legislator control logic.

To be consistent with this expectation, we would expect to observe empirically not only that the EU legislator – Parliament and Council – provides for the set-up of stakeholder bodies in agencies’ constituent acts (establishment) but also that the legislator specifies and regulates such bodies’ composition as a matter of formal act. In line with the argument of McCubbins et al. (1987), such procedural solutions serve as control mechanisms on the legislator's behalf by allowing the legislator to ‘stack the deck’ in favour of specified groups. This necessarily entails the legislator not only setting up such structures but also closely regulating the details of their composition so as to enfranchise specific groups or constituencies. The composition dimension is thus critical to such bodies serving a control function.

At the same time, it is equally well-established that agencies are not ‘passive actors’ in control processes: ‘they are active participants’ (Black, 2012, p. 13), ‘proactive in managing their own legitimacy’ (Black, 2008, p. 152). Consistently, agencies are found to play an active role in shaping procedural choices such as participation mechanisms. Potter (2019) speaks of ‘procedural politicking by bureaucratic agencies’ – essentially agencies manipulating procedural choices to their advantage: ‘knowledge of procedures – and how to manipulate them strategically – is an essential source of bureaucratic power’ (Potter, 2019, p. 18). In other words, agencies can shape the very rules that are theoretically meant to constrain them: procedures offer principals ‘ample opportunity to intervene in agency decisions, but it is not all driven from the top down. Agencies have considerable powers of their own in the process’ (Potter, 2019, p. 9). This is in line with a broader literature that speaks of the pro-active entrepreneurial role of bureaucratic actors (Carpenter, 2010; Huber, 2007; Moffitt, 2014; Busuioc & Rimkutė, 2020b), actively (co-)shaping their formal constraints and the terms of their formal mandate (Carpenter & Krause, 2015).

Such ‘procedural politicking’ can take unexpected or counter-intuitive forms for instance, an expansion in stakeholder participation opportunities as a way to achieve bureaucratic autonomy through successful coalition-building and bypass principal control. For instance, Potter (2019) finds that bureaucratic agencies in the US context restrict or expand ‘the participation valve’ so as ‘to strategically and systematically insulate their rule-making proposals from scrutiny and interference’ (p. 5, emphasis added). When groups are supportive of an agency’s position, the latter opens up or expands participation opportunities – groups of powerful interests thus serve to bolster the agency’s position towards its principal, raising the principal’s costs of intervention. These findings also reflect Carpenter’s (2001, p. 17) well-established argument that agencies’ successful coalition-building with ally audiences ‘make(s) it costly for politicians to resist them’, leading to enhanced bureaucratic autonomy and principal deference. Thus, EU agencies may voluntarily set up a stakeholder body in accordance with their own bureaucratic preferences.

Expectation 2: The establishment and composition of EU agencies’ stakeholder bodies reflect bureaucratic preferences.

To diagnose alignment with this expectation, we would expect to observe agencies¹pro-actively setting up stakeholder bodies in the absence of formal obligations in this regard, or – when such bodies are provided for by formal design – for agencies to shape these formal constraints. An agency may structure stakeholder composition in line with its preferences when these requirements are not explicitly provided for by formal design. It may structure composition to benefit specific sets of actors, or to the contrary, limit it so as to exclude or disadvantage some interests over others. In doing so, EU agencies may exploit ambiguities in, or go beyond, the legal requirements adopted by the legislator. In other words, ‘bureaucratic preferences’ will be diagnosed as dominant when the agency voluntarily sets up and structures a stakeholder body. It will be similarly diagnosed when, while the stakeholder body is legislator-mandated, the legislator has not carefully tailored composition – it is the agency that shapes composition in the absence of legislator steering. Our argument is essentially that when these choices are not spelled out by the legislator but decided by the agency, they will necessarily reflect bureaucratic preferences rather than those of the legislator – as the legislator has not articulated its own preferences in this respect.

Of course, building autonomy and insulating oneself from control are not the only rationales behind procedural choices made by bureaucratic actors. Perfectly benign and functional (rather than political) rationales such as enhanced expertise or implementation capacity (Arras & Braun, 2018) could be other drivers to do so. Stakeholder bodies might also follow broader trends in EU governance for participation and co-regulation, reflecting isomorphic rather than functional rationales. Acknowledging this does not detract from the earlier point about the significance of administrative procedures for substantive outcomes. By enfranchising particular groups to the detriment of others, stakeholder bodies empower specific groups within agency structures.

This is arguably non-problematic when it reflects legislative intent and a logic of principal control. However, when these choices reflect agency preferences, even entirely justifiable and benign ones, caution must be warranted as to the implications thereof. Such practices effectively amount to the pro-active, legislator-unsanctioned de-insulation of a whole set of agency activities. Moreover, through its inclusion – and simultaneous exclusion function – of some interests over others, such closed structures necessarily introduce biases in favour of specific groups. Unless carefully and cautiously balanced and well-tailored to purpose, putative mechanisms of legitimation can become vehicles of bias. What is more, when the bias is in favour of narrowly defined groups, the specific problem of capture (Stigler, 1971; Carpenter & Moss, 2013) can arise.

Moreover, as noted above, broad-based coalition-building initiated by agencies, too, can be problematic from a control perspective: agencies’ successful coalition-building with stakeholders can displace the ability of political superiors to enact formal controls by rendering it politically-costly for principals to intervene where an agency enjoys a good standing among stakeholders (Potter, 2019; Bertelli & Busuioc, 2020). In contrast to capture, such reputational coalition-building is broader, rather than focused on narrow interests, but the result can nevertheless be a displacement of principal control (Bertelli & Busuioc, 2020). Our contribution thus also implicitly builds on studies that examine, in the context of EU agencies, how formal configurations are informally shifted to the benefit of different institutional players (Font, 2018).

Categorization, data and method

At the national level, stakeholder bodies have been conceptualized as advisory bodies or committees (Fraussen et al., 2015; Halpin & Fraussen, 2017), or as public committees (Binderkrantz & Christiansen, 2015), composed of non-state organized interests that do not seek public office, e.g., interest groups, companies or other civil society organizations (i.e., stakeholders) (see also Beyers et al., 2008). They are regarded as ‘closed’ instruments for engagement because stakeholders cannot participate unless granted access, in contrast to open forms of stakeholder engagement such as public consultations (Halpin & Fraussen, 2017). At the EU level, stakeholder bodies have been defined as ‘permanent entities within the agency where a substantial number of the members are stakeholders, which meet on a regular basis’ (Arras & Braun 2018, p. 1261) and which ‘enable sustainable interactions with stakeholders’ (p. 1262).

We apply this definition, although we argue that stakeholder bodies are not necessarily permanent, as long as they are durable – i.e., involve sustained interactions. Stakeholder bodies have registered members with repeat access. This means that they consist of a set of stakeholders that are recognized as members thereof (registered or publicly listed as such by the agency). The notion of membership is distinct from access on a single occasion, such as an invitation to a conference or to a closed committee hearing.

Stakeholder bodies are primarily set up for contact with non-state stakeholders rather than governmental organizations or scientists. As such, they are distinct from other internal agency structures such as management boards, which are primarily set up for the representation of national governmental bodies (e.g., domestic agencies) or from scientific committees composed of individuals from universities etc., and which are therefore excluded from the scope of our investigation. Some EU agencies’ management boards may include stakeholders. However, management boards are set up for, and dominated by, representatives of governmental bodies, and stakeholders only make up a smaller fraction of members.

We looked for EU agency structures that predominantly consisted of stakeholders. We only coded an agency structure as a stakeholder body if an absolute majority of member organizations were stakeholders (threshold of more than 50 per cent stakeholders) in order to distinguish stakeholder bodies from agency structures predominantly composed of other groups like governmental bodies. Structures for stakeholder representation that had a minority of non-stakeholders represented (e.g., European Commission representatives) were therefore still categorized as stakeholder bodies.

Our investigation covers all 34 decentralized EU agencies (European Union, 2018, Table 1 in Appendix). In terms of data sources, we draw on all agency founding regulations, on agency documents (management board decisions, internal procedural rules and annual reports) and agency websites. We included stakeholder bodies that were required by EU legislation in force, or bodies that existed in practice in 2020. To investigate our expectations, we map out the set-up of stakeholder bodies across agencies as well as their composition and the extent to which these are mandated within constituent acts (or agency-regulated instead). On this basis, and in light of stakeholder bodies’ prerogatives – tasks (role) vis-à-vis the agency i.e., which agency aspects and activities it provides input on as well as span of access to the agency and its key structures – we reflect on implications thereof.

Stakeholder bodies: a deep dive

In this section, we present empirical data on EU agencies’ stakeholder bodies, and analyse the extent to which their establishment and composition were regulated by the legislator or determined by the agency itself. We then turn to the implications of principal-imposed and agency-initiated stakeholder bodies, where we discuss potentially problematic design features of stakeholder bodies reflecting bureaucratic preferences.

‘What's in a name?’ Of all shapes and sizes

Our examination reveals that stakeholder bodies are a relatively popular stakeholder involvement instrument amongst EU agencies. As seen in Table 1 below, almost half of the EU agencies (16 out of 34 agencies) possess such instruments, while a total of 18 EU agencies have no stakeholder bodies in place – i.e., no provisions for the set-up of such a body are contained in their founding acts and such bodies have not been set up pro-actively by the agency. Overall, stakeholder bodies – regardless of whether voluntarily adopted or legislator mandated – have been on the rise in recent years: Stakeholder bodies are mandated by the legislator within the constituent acts of 11 EU agencies, all of which were established after 2000, during and/or after the third wave of agencification (Table 5 in Appendix). Here we find both agencies that possess decision-making powers (5), and agencies without such powers (6) – see Table 2 in Appendix. Prominent among the former are the most powerful regulatory agencies at the EU level (the European Financial Supervisory Authorities, i.e. the ESAs), while in the latter group we find agencies whose mandates pertain to, or touch upon, fundamental rights issues, and where stakeholder input pertains to these matters (e.g., EASO, FRA, Frontex). Among agencies that can adopt binding acts, roughly half of these (6 out of 11) are not required by the legislator to set up a stakeholder body (see Table 2 in Appendix).

Table 1. Overview of which EU agencies have stakeholder bodies (N = 34).

Set-up Provided by Mandate (11)	Voluntarily Adopted (5)	No Stakeholder Body (18)
Composition stipulated (5): EBA, EIOPA, ESMA, EFCA, ELA	ACER, ECHA, EFSA, EUIPO, EU-OSHA	BEREC, CdT, Cedefop, Cepol, CPVO, ECDC, EEA, EIGE, EMA, EMCDDA, EMSA, ETF, EU-LISA, Eurofound, Eurojust, Europol, GSA, SRB
Partially stipulated (5): EASO, ENISA, ERA, FRA, Frontex
No specific composition (1): EASA

A total of 5 EU agencies have set up stakeholder bodies pro-actively, in the absence of a formal obligations to do so (see Table 1 above; for additional details, see Table 4 in Appendix). For instance, EFSA has set up a permanent stakeholder platform (EFSA’s Stakeholder Forum), with its own smaller-scale Stakeholder Bureau. ECHA too, has put a process in place for accrediting stakeholders and carries out a yearly Stakeholder Workshop with its accredited stakeholders aimed at providing input on agency work programme and priorities.² ACER also has multiple specialized stakeholder bodies, with three ‘European Stakeholder Committees’ for different electricity issues and a ‘Gas Network Codes Functionality Platform’. Again, here too, we find agencies with the power to adopt binding decisions with voluntarily-established stakeholder bodies in place (ACER, ECHA, EUIPO) as well as without such voluntary structures (CPVO, EMA, SRB) – see Table 2, Appendix.

Upon closer examination, it is directly striking that stakeholder bodies vary considerably in their size, membership, and indeed, the names of different structures – ranging from ‘platform’, ‘consultative forum’, ‘advisory body’, ‘advisory forum’ ‘stakeholder group’ etc. (see Tables 3 and 4, Appendix). This observed variation is consistent with findings of earlier studies with respect to societal and stakeholder input, more broadly (Borras et al., 2007; Perez Duran, 2018).

The size of stakeholder bodies can vary widely from one agency to the next (see Tables 3 and 4 in Appendix). Some stakeholder bodies are comparatively limited in size e.g., EBA, EIOPA and ESMA (30); Frontex (14); ENISA (32); EFCA (11). For instance, EBA’s ‘Banking Stakeholder Group’ has a total of 30 members, and the legislator has provided specific requirements as to its composition with the aim for a balanced representation of specific groups. Similar rules are in place for EIOPA and ESMA. By contrast, other stakeholder bodies are unwieldy: the EASO ‘Consultative Forum’ includes members from 230 organizations; FRA boasts more than 700 organizations, while EFSA’s Stakeholder Forum list comprises 120 registered organizations. Such instances are a far-cry from the portrayal of stakeholder bodies as exclusive clubs: ‘closed’, ‘limited access’ instruments. In some cases, we see the set-up of a two-tiered system, where a smaller-scale body or bureau is put in place (e.g., FRA, EFSA) to help co-ordinate large and unwieldy stakeholder bodies. Beyond this, most agencies have a single stakeholder body – although we also find agencies with multiple specialized stakeholder bodies for specific issues (e.g., EIOPA, ACER).

Composition-wise we see a lot of heterogeneity across stakeholder bodies (see Tables 3 and 4 in Appendix). Some agencies’ structures aim for a balanced composition among different stakeholder types/categories (e.g., EBA, EIOPA, ESMA), while others are open to all (e.g., EASO, ECHA, FRA), albeit loose eligibility criteria usually apply, such as being active within Europe and working within the relevant policy area. Others yet are dominated by one type of stakeholder (e.g., ACER, EASA, ENISA – industry). This variation across agencies is to no small degree the product of the fact that composition is only to a limited degree provided by constituent act. While the EU legislator has placed specific requirements as to stakeholder bodies’ composition and membership breakdown (distribution of different types of stakeholders) in some agencies’ founding acts, for the majority of EU agencies, this is either not regulated at all, or provided for only in partial terms (see Table 1). For instance, the composition of the stakeholder bodies of the European Financial Supervisory Authorities is structured in detail by the legislator, including the number of seats assigned to different stakeholder types (see Table 3 in Appendix). Such agencies represent a minority of cases, however, which has perpetuated heterogeneity.

Procedural choices: agency or principal-structured?

We now turn to ascertaining to what extent stakeholder bodies reflect legislator or bureaucratic preferences. In terms of establishment, as we have seen, most existing stakeholder bodies are mandated by the legislator (11), while some have been voluntarily set up by the agency (5). A first cursory view would then seem to suggest a mixed logic, with a principal control as the dominant logic at play, given that most stakeholder bodies in existence are required by founding acts. A closer look at formal requirements and agency practices, however, reveals otherwise.

It is interesting to note that even when stakeholder bodies are mandated by the EU legislator, in the majority of cases, the legislator does not appear to be in the driving seat as to their composition. While as noted above, the stakeholder bodies of 5 agencies (among them, the 3 ESAs) are carefully tailored by the legislator, most other agencies’ composition details are only partially – or not at all – stipulated. For instance, the composition of the stakeholder bodies of Frontex and EASO is only to a limited degree spelled out by the legislator. Frontex is to include three specified organizational actors: EASO, FRA and the United Nations High Commissioner for Refugees (UNHCR), leaving it for the agency to decide composition beyond these three actors. Similarly, EASO’s is to include UNCHR as a member (ex officio), leaving it for the agency to decide on the remaining composition. In practice, these two agencies have opted for very different compositions in giving shape to their respective stakeholder bodies: while Frontex opted for 14 members, including the 3 permanent members provided for by constituent act, EASO included over 230 organisations. Such divergences become difficult to account for in structures purported to serve the same purpose (stakeholder engagement), and is even more striking as these two agencies operate in the same field (Justice and Home Affairs). This is illustrative of the considerable variation in how agencies have implemented broad legislative provisions as to stakeholder composition, and of the limited degree of legislator control exercised over the structural set-up of such bodies.

What is more, in one case a stakeholder body was mandated without any requirements from the legislator on composition (i.e., EASA), leaving complete free rein for the agency to fill in these details and shape stakeholder structures. Here, the set-up of an advisory body ‘representing the full range of interested parties affected by the work of the Agency’ is provided for in its founding regulation (Art. 98(4)). In practice, the actual composition of this stakeholder body has been specified by the agency (through a decision of its management board) and predominantly includes industry actors (see Table 3 in Appendix).

These patterns are at odds with a principal control logic. Such a logic would lead us to expect that the legislator structures procedures, stacking the deck in favour of specified constituencies, which can channel decision-making, and serve as ‘fire alarms’ for the legislator. This expectation does not hold across-the-board as the composition of many stakeholder bodies is either not specified, or only very generally specified, by formal design. In other words, it appears that in many cases the legislator found it important that such stakeholder mechanisms are nominally in place, but not necessarily how these bodies are structured, who actually is involved, and to what end. The legislator did not attempt to control or regulate these processes equally for all agencies. This seems to point rather at a logic of legislator efforts at signalling stakeholder involvement as way to tap into the legitimacy benefits thereof, effectively setting up stakeholder bodies as ‘empty shells’ which agencies fill with content.

One notable exception emerges, however, as noted above: the stakeholder bodies of the financial supervisory authorities (EBA, EIOPA and ESMA), where the legislator has spelled out in detail composition details. These agencies possess far-reaching supervisory competences as well as formidable (quasi-)rule-making powers that ‘break the mould (…) in terms of their unprecedentedly wide-ranging powers compared with earlier agencies’ (Busuioc, 2013b, p. 112). The financial agencies can draft regulatory technical standards, which are endorsed by the Commission as delegated acts under Article 290 TFEU or implementing technical standards, adopted by the Commission as implementing acts under Article 291 TFEU. The founding regulations of the financial supervisory authorities provide that their respective Stakeholder Groups need to be consulted, among others, on draft rule-making, on actions concerning regulatory technical standards and on implementing technical standards (as well as guidelines and recommendations).

Here a principal logic control clearly emerges. Article 290 and 291 TFEU have been the ‘battle lines’ over non-legislative rule-making between the Commission and the EP, with well-documented and recurrent tensions between the Commission and the EP specifically with respect to the process of adoption of ESAs technical standards and its implications for institutional prerogatives (Busuioc, 2013b, pp. 115–117). In this context, that these stakeholder bodies – which are consulted on draft regulatory and implementing technical standards, prior to the Commission’s endorsement – should be structured in great detail by the legislator is no surprise. The EP has a range of checks in place on the Commission’s role when adopting delegated and implementing acts but the Stakeholder Groups can provide an early ‘fire-alarm’ opportunity for the EP, acting as a check both vis-à-vis the agency as well as the Commission. Thus, this suggest that the EP does carefully tailor such bodies where it matters for its prerogatives. These instances are consistent with a legislator control logic, as theorized in the literature, and indicate – perhaps not surprisingly – that the EP yields this in a strategic, sporadic manner, rather than by default.

To sum up, for most EU agencies, stakeholder bodies in place do not seem to reflect a legislator control logic. Out of the 11 agencies where the set-up of a stakeholder body is legislator mandated, only for 5 agencies are the full details of their composition specified by the legislator. Little legislative input beyond underspecified demands for stakeholder bodies is in place for the remainder (6), leaving significant leeway for agencies to shape their design. We find examples of agencies with significant mandates where stakeholder bodies remain weakly regulated (e.g., EASA, ERA) or neither required nor regulated by the legislator at all (e.g., ACER). A number of EU agencies (5) have also set up stakeholder bodies on a voluntary basis, in the absence of any formal requirements in this respect. Observed patterns are thus predominantly consistent with a bureaucratic preferences logic rather than indicative of attempts at principal control. Nevertheless, there are limited but important instances of principal control where the EP has carefully structured stakeholder involvement when it mattered for its prerogatives.

Implications: still insulated?

Are these patterns potentially disconcerting? Is the diagnosed ambivalence of the legislator in this respect – on the one hand, demanding the set-up of such structures but on the other sparsely regulating their composition – potentially problematic from a legitimacy perspective? A cursory view of the prerogatives of stakeholder bodies (and their members) vis-à-vis the agency – both in terms access to agency structures and input into agency activities – certainly raises some points for concern in this context.

In terms of span of access, stakeholder bodies afford member stakeholders opportunities to provide input and/or interact with the core agency structures and decision-makers such as the executive director, the management board and even scientific committee members. Multiple stakeholder bodies are granted access to the agency head: For instance, the FRA’s Fundamental Rights Platform is coordinated ’under the authority of the Director’ (founding regulation Art. 10(5)), while ENISA’s Advisory Group, EFCA’s Advisory Board and EFSA’s (voluntarily established) Stakeholder Forum are all chaired by the agency’s executive director.³ Similarly, ECHA’s Stakeholder Workshop with accredited stakeholders is attended by several of the ECHA’s directors, including the executive director.⁴

Access is often not limited to the agency head. For instance, in the case of EFSA, in addition to the executive director, representatives from EFSA’s management board, advisory forum, and Scientific Committee are also invited to attend Stakeholder Forum meetings.⁵ The Boards of Supervisors of EBA, EIOPA and EMSA – the EU financial authorities’ main rule-making and decision-making bodies – convene meetings with their respective Stakeholder Group ‘regularly, at least twice a year’.⁶ ENISA’s stakeholder body can take part in the meetings of the management board at the invitation of the Chairperson,⁷ while individual stakeholder body members of several agencies can attend management board meetings as observers (EASA, EFCA).⁸ Agency stakeholder bodies also provide input and recommendations on specific matters to management boards (e.g., EASA, EASO, ENISA, FRA, Frontex), agency head (e.g., EFCA, ENISA, FRA, Frontex) and/or agency’s Scientific Committee (e.g., FRA).

Moreover, and while this varies from one agency to the next, in terms of their tasks (role), stakeholder bodies are involved in providing input on varied, and in some cases, a large array of agency activities. Some of this input pertains to core agency tasks such as rule-making or enforcement roles. Thus, while some stakeholder bodies have a role with respect to planning, agency priorities and work programme (e.g., EFSA’s Stakeholder Forum, FRA’s Platform, EASO’s Consultative Forum, ENISA’s Advisory Group), others have been granted broad mandates with respect to providing input on a broad range of aspects pertaining to agency’s functioning and core tasks, including in relation to key functions such as rule-making tasks (e.g., EBA, EIOPA and ESMA, as discussed above).

In some cases, such extensive mandates are granted to stakeholder bodies whose composition has not been spelled out by the legislator. For instance, the functions of EASA’s stakeholder advisory body (SAB) – whose composition is stipulated by an agency board decision and is predominantly comprised of industry actors – are to provide recommendations, among others, on: ‘all aspects related to the programming and rulemaking activities of the Agency’.⁹ Other tasks include: to support the Agency in the ex post evaluation of rules, to support rule-making groups, provide advice on ‘the content, priorities and its execution of its safety programmes’, provide advice ‘on the policy of acceptance of industry standards’ etc.¹⁰ These tasks, specified by the agency’s management board, are a substantial expansion to the functions explicitly provided for in the founding regulation: i.e., primarily pertaining to consultation on the work programme and some budget aspects (such as fees and charges).¹¹ By extension, the agency has granted the stakeholder body explicit remit with respect to the agency’s rule-making activities.

Such broad mandates are also in place for tasks related to enforcement and supervision. Frontex’ Consultative Forum not only assists the executive director and management board on fundamental rights matters, but also provides input on the border guard core curricula and gives ‘strategic advice to Frontex operational activities’.¹² The Forum can also conduct on-spot visits of joint operations, rapid border interventions, hotspot areas, return operations and interventions (founding regulation, Art. 108(5)). The financial supervisory authorities’ Stakeholder Groups too, have a role in enforcement and supervisory matters – they can for instance, submit a request to the agency to investigate the alleged breach or non-application of Union law.¹³

In search for involvement, agencies’ stakeholder bodies seems to have become in many cases, ‘jacks of all trades’ that give input on a considerable array agency activities of consequence ranging from agency priorities to rule-making (e.g., EBA, EIOPA, ESMA; EASA) and enforcement (e.g., EBA, EIOPA, ESMA, Frontex). The clear implication that emerges from our investigation is that the agency model is being incrementally transformed: the model that is emerging is a far-cry from the classic model at the heart of regulation – the insulated agency model. A whole range of core agency activities have been opened up to societal and interest group input.

Given for the most part a lack of consistent principal steering, such procedures have emerged in a disjointed fashion across agencies. Only a minority of agencies seem to have stakeholder bodies that are exclusive clubs whose members represent diverse interests in balanced proportions (i.e., balanced representation), under explicit and specific guidance by the legislator. To the contrary, several agencies have adopted broad, unwieldy structures that are effectively ‘open to all’ organizations that sign up for membership (see Tables 3 and 4, Appendix). Other agencies yet, as discussed above, have adopted narrow structures that mainly reflect one stakeholder category (i.e., industry). Both patterns are potentially problematic: While a narrow representation (i.e., dominance by predominantly one type of interest) increases the prospect of capture dynamics emerging with a move away from the public interest (Carpenter & Moss, 2013), agency-initiated broad-based engagement too, can blunt principal control through successful coalition-building with a variety of regulatory audiences (Bertelli & Busuioc, 2020). Ultimately, an agency could bring supporters on board – in a stakeholder body – to strategically strengthen its position vis-à-vis the principal:

[I]n circumstances where groups are likely to bolster the agency’s position, the agency may find itself increasing participation opportunities. That is, if groups support the agency’s position and principals do not, having groups weigh in on the agency’s proposal may (…) serve to convince overseers that the agency proposal has merit – or at least that intervening in the rulemaking may raise issues with a groups of powerful interests. (Potter, 2019, p. 78, emphasis added)

Given the high stakes, it is paramount that future studies investigate to which extent such dynamics are materializing in practice as a result of the structural choices EU agencies have made in implementing loose legislator requirements for stakeholder engagement.

Conclusion: politics in the regulatory state

We set out to investigate the actor preferences shaping structural choices involved in EU agencies’ stakeholder bodies and reflect on implications thereof. Our study diagnosed a move away from regulatory insulation among a broad array of EU agencies: Nearly half of EU agencies have internalized stakeholder engagement within their structures, opening up a broad span of agency tasks to stakeholder input. We regard this as symptomatic of a broader move towards politicization among key actors of the EU regulatory state. Through opening up an array of core activities to societal inputs, a broad array of EU agencies are no longer strictly legitimized by their technical outputs, but are increasingly making efforts to expand the basis of their claim to legitimation on demonstrating that such outputs reflect – and are informed by – societal considerations i.e., stakeholder concerns, preferences and priorities.

Our investigation suggests that the stakeholder mechanisms in place seem, for the most part, to reflect bureaucratic preferences for engagement rather than principal control. For the majority of agencies, the legislator has not mandated the set-up of a stakeholder body at all, or has stopped short of actually regulating the composition of such structures in a way that would allow for meaningful oversight. Agencies have stepped in to fill the gaps and even to create such bodies from scratch as they see fit. In a departure from the above, the legislator has closely regulated such bodies for a handful of agencies – notably, the EP has been diligent in spelling out the details of stakeholder bodies’ composition, size and stakeholder distribution with respect to the European Financial Supervisory Authorities, the most powerful EU agencies in existence (Busuioc, 2013a). This is not coincidental. It is an area –non-legislative rule-making in the financial sector – where the EP has been concerned with its own prerogatives and one of protracted turf tensions with the Commission over institutional prerogatives (Busuioc, 2013b). Such instances are consistent with a legislator control logic, yet our findings suggest that the EP deploys this sporadically, rather than across-the-board. For the most part, beyond such notable instances, composition requirements are provided for by the legislator only in a broad sense, leaving the ‘details’ of structural choice and implementation to agencies. EU agencies enjoy leeway to shape and structure their stakeholder bodies in significant ways.

These developments warrant caution. While often associated with democratic desiderata of participation and open government, stakeholder engagement is not an unmitigated good. Structures are determinant of policy outcomes. When stakeholder bodies are not carefully balanced, when engagement becomes an end in itself without much consideration to the details of such structures, such bodies can become vehicles for bias. This becomes especially relevant as our investigation reveals that key agency structures and an ever broader array of agency activities are opened up to stakeholder input, ranging from agency priorities to their (draft) rule-making as well as enforcement and supervisory tasks.

The instruments discussed here are only one among the multiple stakeholder involvement strategies EU agencies are pursuing (Perez Duran, 2018; Arras & Braun, 2018), the engagement picture is more extensive still. In response to pressures for engagement, and given the contested nature of regulatory policy, EU agencies are increasingly turning towards building support and allies from their environment. The agency model is slowly and incrementally being morphed away from its core feature – insulation. While not politicized in a majoritarian sense, the growing role and span of involvement of interested parties in agencies’ work, certainly points to creeping, and growing, politicization in regulatory policy.

Acknowledgements

We would like to thank Thijs de Boer for research assistance provided for this study in early 2018. We are also very grateful for helpful comments received from Åse Gornitzka, Karl Bjurstrøm, Sigurd Lindstad, Thijs de Boer, Vilde Hernes, Leif Erlend Kårtvedt as well as from the three anonymous reviewers.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This article is part of a project that has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme [grant agreement number 716439].

Notes on contributors

Madalina Busuioc

Madalina Busuioc is Associate Professor at the Institute of Public Administration, Leiden University, where she leads a prestigious European Research Council (ERC) grant investigating public sector reputation and its effects within the European regulatory state. Her recent work draws on reputational approaches to study and theorize about their implications for our understanding of public accountability and theories of political control.

Torbjørg Jevnaker

Torbjørg Jevnaker is a Doctoral Research Fellow at the Department of Political Science, University of Oslo. With a research interest in the EU and public administration, she works on EU agencies, regulatory issues and member state implementation. She has also published about EU governance of, and policy-making on, energy and climate issues.

Notes

1 We acknowledge that agencies are not unitary actors, however, the adoption of a stakeholder body requires coherent internal action manifest in the adoption of internal rules of procedure regulating set-up and composition (where such bodies are not legislator-mandated).

1 We acknowledge that agencies are not unitary actors, however, the adoption of a stakeholder body requires coherent internal action manifest in the adoption of internal rules of procedure regulating set-up and composition (where such bodies are not legislator-mandated).

2 ECHA ‘Cooperation with accredited stakeholders’, available online: https://echa.europa.eu/about-us/partners-and-networks/stakeholders/cooperation-with-accredited-stakeholder-organisations [accessed 13 May 2020].

3 ENISA founding regulation, Art. 21(3); EFCA founding regulation, Art. 40(4). EFSA: Art. 4(2): EFSA (2018) Decision of the Management Board of the European Food Safety Authority on the criteria for Establishing a List of Stakeholders and the Establishment of the Stakeholder Forum and the Stakeholder Bureau, 9 October, available online: https://www.efsa.europa.eu/sites/default/files/Document18992.pdf [accessed 1 May 2020].

4 ECHA ‘Accredited Stakeholder Workshop’, available online: https://echa.europa.eu/about-us/partners-and-networks/stakeholders/cooperation-with-accredited-stakeholders/accredited-stakeholder-workshop [accessed 13 May 2020].

5 Art. 4(8) (EFSA, 2018; see endnote 2).

6 Art. 40(2) in all three agencies’ founding regulations.

7 Art. 17(4) of ENISA founding regulation.

8 EASA founding regulation, Art. 99(5); EFCA founding regulation, Art. 40(2).

9 Article 5(1): EASA (2015) Management Board Decision 20-2015 replacing Decision 3/2002 of the Management Board establishing the Advisory Body of Interested Parties, 15 December, available online: https://www.easa.europa.eu/sites/default/files/dfu/EASA%20MB%20Decision%2020-2015%20Stakeholder%20Advisory%20Board%20(SAB).pdf [accessed 12 May 2020].

10 Article 3 (EASA, 2015; see endnote 8).

11 The founding regulation leaves it open to the agency should it choose to consult on other issues as well. See Art 98(4) of Regulation (EU) No. 2018/1139 OJ L 212, 22.8.2018, p.1-120.

12 FRONTEX (2020) ‘Consultative Forum’, January 2020, available online: https://frontex.europa.eu/fundamental-rights/consultative-forum/general/ [accessed 30 April 2020].

13 Art. 17(2) of the ESMA, EBA and EIOPA founding regulations.