An accountability deficit? Holding transgovernmental networks to account

ABSTRACT

International cooperation in networks between national civil servants is often judged necessary to address global policy challenges. The civil servants cooperating with each other in these transgovernmental networks (TGNs), are often employed by national agencies that are not part of a ministry. It is therefore assumed in the literature that participating in a TGN renders account giving to a political forum difficult. We offer in this paper a state-centered perspective on the functioning of TGNs to assess whether a trade-off exists between accountability and the creation of technocratic driven networks. We develop a typology of TGNs and accountability, based on the function and the governance style of networks. In three out of four cases studied, variation in the institutional set-up influences which type of accountability is dominant. In one case we expected to find political accountability but observed that elected politicians did not hold civil servants in TGNs to account.

KEYWORDS:

• Accountability
• multilevel governance
• transgovernmental networks

Introduction

Global challenges are often technical challenges. The Dutch city of Amsterdam announced that they would spend 8.2 million euro on the eradication of Japanese knotweed (Meershoek, Het Parool, May 17th 2019). Originally brought to Europe for ornamental garden purposes, this plant adjusted very well to the Dutch climate but has a detrimental effect on biodiversity and can break through tarmac causing problems with sewage (Nobanis, n.d.). The spread is so persistent that the North European and Baltic Network on Invasive Alien Species was created, which prohibited transportation of knotweed (EC, 2019).

The knotwood example is one of many transboundary problems for which a transgovernmental network (TGN) is established (Coen & Thatcher, 2008; Eberlein & Newman, 2008). TGNs bring together national civil servants with specific expertise on a particular policy field, with the aim of forging the convergence of national policies among the participating countries (Slaughter, 2001). TGNs pose a particular challenge to traditional conceptions of accountability. Not only do civil servants working in TGNs convene literally at geographical distance of their national political principals, they are often also employed by independent agencies. Hence, whereas civil servants in a traditional accountability setting have an obligation to explain and justify their conduct to a forum of politicians elected by the population whose interests they need to guard (Bovens, 2007, p. 450; see also Papadopoulos, 2007), within TGNs peer-to-peer interactions are dominant, which creates a form of governance in which the civil servant becomes somewhat of an executive (Raustiala, 2002, p. 25; Slaughter, 2004; Papadopoulos, 2007). Regardless of this organizational distance, participants to these TGNs are civil servants employed by national state authorities. As such they should be held accountable by these authorities, especially when dealing with highly salient policies. However, as Curtin and Egeberg state, it might be necessary to complement it ‘with mechanisms and forums that focus both on the accountability of supranational executive bodies as well as national agencies with dual loyalties’ (Curtin & Egeberg, 2008, p. 640).

Whether a trade-off exists between accountability and the functioning of TGNs, is not yet fully addressed in the literature (LeGrand, 2019, p. 201). We will explore to what extent civil servants in TGNs are held to account for their activities and how this relates to ideal-types of accountability. By combining theories from the disciplines of political science, international relations and public administration, and applying this to a new context, we aim to develop and empirically assess a typology of TGNs and accountability types. Our main findings are that the acclaimed accountability deficit within TGNs is not as imminent as many scholars of international governance are claiming. We find in addition that different types of TGN produce their own types of accountability that yield alternative forms of accountability and in most instances make up for the loss of traditional hierarchical political accountability.

The next sections provide an overview of the state of the art in the literature on TGNs and present a typology of TGNs and accountability type, based on literature on the function of TGNs and the governance style of networks. After describing how we collected and analysed the data for the four TGNs we study, we present the results. In the concluding section we answer the question to what extent TGNs cause accountability deficits at the national level.

Accountability deficits in TGNs

The public policy literature distinguishes three functions of TGNs. First, in an information network, civil servants from different national public administrations come together, discuss their problems, and formulate best practices through exchanging information (Slaughter & Hale, 2013, p. 344, Lavenex, 2008). Second, when a TGN is an enforcement network, the cooperation between civil servants is aimed at improving the enforcement of regulations at the domestic level. This cooperation is mostly on the operational level (Lavenex, 2008, p. 942) and the regulations they aim to enforce may stem from supranational organizations or domestic public administrations (Slaughter, 2004, p. 55). Third, in harmonization networks regulations such as benchmarks and standards in a particular policy area are harmonized between member states of the network, leading to the strongest shift in regulatory capacity from the national state to a TGN (Lavenex, 2008, p. 942).

Scholars from various disciplines such as international relations, public administration and political science, raised concerns on accountability deficits caused by the relatively large degree of independence of TGNs (Busuioc, 2010, p. 3; see also Raustiala, 2002, p. 418; Slaughter, 2002, p. 1042). Principals lack oversight over the actions of civil servants in the network, while peer-to-peer accountability within TGNs replaces the political accountability relationships running from voter to elected politicians (Papadopoulos, 2007; Sabel & Zeitlin, 2010, p. 12; Papadopoulos, 2014, p. 283).

Concerns about accountability deficits in democracies are not limited to TGNs (Majone, 1999; Schillemans, 2008; Van Beek, 2018; Zürn, 2004), but are also related with the rise of policy networks in general (Harlow & Rawlings, 2007; Sørensen & Torfing, 2009, p. 234; Vantaggiato et al., 2020). In networks, civil servants are subjected to multiple accountability relationships (Magetti, 2012; Scott, 2000), which complicates the accountability between civil servants and their political principals (Curtin & Egeberg, 2008, p. 655; Messner, 2009, p. 919). In order to understand the varieties of accountability relationships within network settings we adopt in this study Bovens’ relational conception of accountability:

Accountability is a relationship between an actor and a forum, in which the actor has an obligation to explain and to justify his or her conduct, the forum can pose questions and pass judgement, and the actor may face consequences. (Bovens, 2007, p. 450)

Theoretical framework

It is claimed in this paper that prevailing accountability relationships in TGNs are a function of the function and governance style of TGNs (Vantaggiato et al., 2020). Provan and Kenis (2008) identify three distinct styles of network governance. A participant-governed network is based on the collaboration of the members themselves, without a clear entity within the structural make-up of the network that coordinates or supports the collaboration. This type of network is dependent on the involvement of the members as they are solely responsible for the work of the network. The second type is the lead-organization governed network. This is a more centralized governance structure in the sense that ‘all major network-level activities and key decisions are coordinated through and by a single participating member, acting as a lead organization’ (Provan & Kenis, 2008, p. 235). Often a lead organization offers secretarial support for the activities of the network. The third type is the network administrative organization, with a secretariat to help steer, govern and coordinate the network in a centralized manner. In this network the members are no longer in the lead and the secretariat is the network broker (Provan & Kenis, 2008).

Combining TGNs’ network governance style and function yields four distinct varieties of institutional set-ups that correspond, both logically as well as normatively, with one of Romzek and Dubnick’s (1987) four accountability types (see Table 1). In bureaucratic accountability the relationship between actor and forum is based on close supervision, the codification of rules is strong and the subordinate faces internal controls (Romzek, 2000). In a legal accountability type, the civil servants have slightly more autonomy compared to the bureaucratic type because the source of control is external to the TGN the civil servant is part of and ‘the accountability standard focuses on whether the agent has complied with the principal’s expectations’ (Romzek, 2000, p. 25). The third type of accountability is professional accountability, consisting of peer-to-peer accountability based on relatively loosely formulated professional norms and standards, with high discretion for the civil servant vis-à-vis the home organization (Romzek, 2000, p. 26). The fourth type of accountability is political accountability; the civil servant in a TGN has considerable discretion to perform tasks but has to give account to a political forum external to the TGN, such as a parliament or an elected politicians heading a ministry.

Table 1. Four types of accountability. Adapted from Romzek and Dubnick (1987).

		1. Source of control over actor
		Internal	External
2. Degree of control over actions actor	Low	Professional accountability	Political accountability
	High	Bureaucratic accountability	Legal accountability

Table 2 visualizes, then, our typology of accountability forms. We claim that the type of accountability present in a TGN is causally related to the function of a TGN and the governance style of the network. The arrows stipulate that, in reality, TGNs may have mixed functions and styles and, hence, display a mix of accountability types. In line with the typology we expect, first, that when a TGN is an information network and participant governed, the TGN incorporates professional accountability in its day-to-day functioning. In an information network, civil servants come together on a voluntary basis, discuss their problems, and formulate best practices. When the governance style of a network is participant-governed the members themselves collaborate, without a secretariat coordinating the collaboration. Given the horizontal actor-forum relationship in participant-governed information networks, the actors within the TGNs are peers. These institutional characteristics increase the likelihood that this type of TGN incorporates professional accountability. This type of accountability consists of peer-to-peer accountability based on relatively loosely formulated professional norms and standards, with high discretion for the civil servant.

Table 2. A typology of TGNs and accountability.

Second, when an information network is network administrative governed, the TGN incorporates bureaucratic accountability in its day-to-day functioning. When an information network is network administrative governed, its main function is exchanging information between participants. A secretariat is set up within the network to help steer, govern and coordinate the network in a centralized manner. We expect bureaucratic accountability to be dominant here. In this type, the relationship between actor and forum is based on close supervision. The civil servant participating in the TGN faces on a regular basis internal controls through occasional participation in the TGN by the supervisor from the home organization. There is a strong vertical dimension, with codification of rules. The impact of an information network is not perceived to be great by the home organization, but at the same time, the network administrative part leads to a formalized component in the TGN, which institutionalizes the regular involvement of supervisors from the home organization to a higher degree than if the network was only participant governed.

Third, when a harmonization network is participant governed, the TGN incorporates political accountability in its day-to-day functioning. In a harmonization network, regulations such as benchmarks and standards in a particular policy area are harmonized between the participating countries through the activities of the TGN. When the governance style of a network is participant-governed the members themselves collaborate without a clear organizational entity within the structural make-up of the network that coordinates or supports the collaboration. The accountability type fitting this institutional set-up of a TGN best is political accountability, in which an actor needs to give account to a domestic political forum such as a national parliament or a minister. Under political accountability, the civil servants have considerable discretion to perform tasks, which is key to ensure expertise of the civil servants in the TGN can be used to the full to fulfil the important task of harmonizing benchmarks and standards across member states of the network. Given the importance of this type of TGN for the member states because of its high impact on national policies, and the need for a high level of discretion for the expert, the civil servants in the TGN needs to be responsive to the supervisor in the home organization. At the same time, the supervisor needs to have some discretion to slightly diverge from legal standards because of the importance to use its expertise in a flexible manner.

Finally, when a harmonization network is network administrative governed, the TGN incorporates legal accountability in its day-to-day functioning. When the main function of the network is harmonizing benchmarks and standards between participating countries and a secretariat within the network is set up to help steer, govern and coordinate the network in a centralized manner, the impact of the network on the state is likely to be considerable. At the same time, the activities of the civil servants in the TGN are subject to rules externally set at the global level which are subsequently monitored by audit teams working independently from the secretariat of the TGN. This fits the type of legal accountability because here emphasis is put on compliance with set rules and procedures.

The typology helps us to explore the claim that the challenge to traditional conceptions of accountability is not as problematic, from a normative perspective, for all types of TGNs. First, in information networks, where bureaucratic or professional accountability prevails, civil servants can be held accountable easier because accountability norms are less stringent than in more formalized networks. Second, when information is shared in an information network and professional or bureaucratic accountability is absent, political principals fail to hold account civil servants for sharing policy information with civil servants from other states in the TGN. Although this situation is undesirable, the consequences of sheer information-sharing for domestic policy practices are negligible, and therefore the accountability deficit less problematic. By contrast, harmonization TGNs do influence domestic policies and when these networks do not meet political or legal accountability standards, policies at the national level are influenced but not accounted for by either elected politicians at the domestic level or other civil servants in the home organizations. Hence, when in harmonization networks the accountability standards of political or legal accountability are not met, we can speak of major accountability deficits arising because of the domestic policy consequences of harmonization TGNs. Finally, when a network deals with salient topics, meeting the accountability standard fitting with the institutional set-up of the TGN becomes even more necessary in harmonization networks to prevent an accountability deficit from arising.

Data and methods

To assess how TGNs incorporate accountability in their day-to-day functioning, we first, demarcated the universe to which we aim to generalize. Second, by means of a typology of nine varieties of TGNs we indicated along which dimensions the cases within this universe vary. Third, we used the typology as a sampling frame by selecting a small number of variances from it. Furthermore, the cases selected are likely to be the clearest examples of where, because of particular combinations of function and governing style, a single accountability type is likely to be dominant over others. To assess which accountability type is present in practice, we examine how accountability is given, on what and to whom. Moreover, we also assess the extent to which sanctions are possible in a TGN, the degree of autonomy, and the source of control. See the online appendix for more information on the operationalization of accountability types.

We study Dutch representatives in TGNs because the Netherlands is represented in a large number of TGNs over a long period of time. Next to the Dutch context, we also kept constant the international embeddedness of TGNs. All four TGNs have close links with the EU, e.g., with the European Commission playing a role in the establishment or financing of TGNs or TGNs performing a task set out by the EU legislation (see also Curtin & Egeberg, 2008).

The participant governed information network we selected is the European Union Network for the Implementation and Enforcement of Environmental Law (IMPEL). As an example of a network administrative information network, we selected the Association of European Vehicle and Driver Registration Authorities (EReg). The Western European Nuclear Regulators Association (WENRA) is the case study for the participant governed harmonization network. The fourth case is a study of a network administrative governed harmonization network, i.e., the European Cooperation for Accreditation (EA). All TGNs deal with complex technical issues for which participating civil servants need a high level of expertise. However, two TGNs deal with salient issues, i.e., nuclear energy (WENRA) and mutual recognition in the EU internal market (EA).

The governance style can be determined based on inaugural documents¹ as these often list the components of the networks and the decision-making process. Next, we have looked at newspaper articles² on the activities of the networks. Reports by TGNs have been analysed with reference to the definition of accountability by Bovens (2007) and the defining elements of the types. Reports of the home organization³ have been cross checked for references to the TGN. European Commission documents mentioning the TGNs were also analysed.⁴ Second, the method of participant observation was used to understand the context of the TGN and the home organization in which civil servants participate and are held to account. In this research observations of two networks were undertaken, comprising a total of 11 full working days. These observations were of working group meetings and one annual meeting.⁵ In the case of WENRA and EA observations were not allowed because both TGNS deal with issues that are sensitive in nature, respectively nuclear safety and the accreditation of organizations.

In addition to the participant observations, 26 civil servants in total were interviewed for the four TGNs.⁶ The semi-structured interviews played a crucial role in the micro-level approach this research adopts by offering for all four cases insights with regard to the perspective of those in charge of constructing and interpreting the relational concept of accountability. The results from the interviews conducted with civil servants involved in the activities of IMPEL and EReg confirmed the insights obtained from the participant observations. Hence, we assume that the interviews conducted with civil servants involved in the activities of WENRA and EA have compensated for the lack of participant observations in these cases. Furthermore, a higher number of documents was studied for the WENRA and EA cases compared to the IMPEL and EReg cases.

Results

IMPEL: a participant governed information network

We expect to find in the participant governed information IMPEL network, professional accountability to be incorporated in its day-to-day functioning. The activities of IMPEL range from wildlife protection against illegal hunting to waste shipment management. Given that the capacity to implement environmental policies in Europe lies with the EU member states, EU environmental legislation comes in the form of directives and member states only exchange information in the context of IMPEL on domestic implementation experiences. The civil servants in IMPEL are typically from environmental protection agencies, inspectorates, and ministries. The General Assembly is the main decision-making body, with representatives of all the member authorities present, convening at least once a year and presided over by the national coordinator of the host country to the Assembly and by the president of the board. The network activities consist mainly of exchanges of information and experiences on ‘the practicability and enforceability of European environmental legislation’ (IMPEL, n.d.). These activities are organized in clusters of experts of the different member countries that set out strategies in particular fields of interest. In addition to these projects, IMPEL organizes three-day conferences to discuss the implementation and enforcement of EU environmental law. From these examples it shows that IMPEL is an information network (IMPEL respondents 4, 12). However, we see some evidence for a harmonization function as well. IMPEL is mentioned in EU legislation as a partner for further cooperation regarding the harmonization of the implementation of environmental laws at the member state level. Respondents (IMPEL respondents 12, 4) considered this could potentially be a first step towards harmonization but currently the network lacks organizational capacity to perform such a role.

IMPEL is a network that is highly reliant on its participants for functioning smoothly – a typical characteristic of a participant governed network. Although IMPEL has a secretariat, it is very small with only 1.5 fte, has a marginal facilitating role and is not present in the project groups (IMPEL respondents 1, 2, 4). Hence, the governance style is highly decentralized. We expect that the function and governance style of the TGN lead to incorporation of professional accountability in the functioning of IMPEL, with a dominance of peer-to-peer accountability based on relatively loosely formulated professional norms and standards.

The empirical material points to the incorporation of professional accountability in this participant governed information network. The network is organized to allow the expert to do his/her work. The General Assembly is the highest decision-making body but it operates on input that has been pre-discussed in the expert setting of the project teams. The account giving in the organization of IMPEL relies heavily on how the expert team leader decides to communicate reports. The sharing of results of the network are mostly done in the expert team leader meeting and are shared on a general strategic level with the General Assembly. Respondents indicate that they disseminate the information from the network meetings they attend to their direct expert colleagues at home, sometimes in the form of newsletters with summaries of IMPEL meetings (IMPEL respondents 3, 2, 8).

For the Dutch participants in this TGN, IMPEL is not a political salient network, with as a result the dominant perception that politicians or higher-level civil servants do not need to be involved regularly. As one of the respondents stated on the involvement of the ministry:

I tell them and the others what the Dutch approach will be in the support group. I tell them that if I do not hear back from them, I can assume that this is the way we will address all the issues. In all those years I have seldom heard that it should be done differently. (IMPEL respondent 1)

Clearly, there is a horizontal relationship between the civil servants in the TGN and the home organization, with a high level of discretion for the civil servant.

EReg: a network administrative governed information network

Due to the development of the internal market, the need emerged to cross-border monitor vehicles and drivers of EU member states and EReg was set up. EReg is organized into four different components: topic groups, secretariat and chair, a general assembly, and an advisory board. The technical discussions and technical transfer of knowledge are done in specialized topic groups. The secretariat functions as a facilitator and hub to the network, which counts 29 member authorities responsible for vehicle and driver registration from both inside and outside the EU (EReg, 2019). The secretariat is responsible for the day-to-day operations of the network and operates from the headquarters of one of the member organizations, the RDW in the Netherlands. The topics are filtered out by means of surveys conducted by the secretariat, and the working programme is prepared for and written by the secretariat and the chair of the network (EReg respondents 1, 4). The role of the secretariat is also to monitor the progress of topic groups. This progress is reported to the general assembly, which convenes annually. The general assembly is the highest decision-making unit of the network, but the day-to-day operations are run by the chairman together with the secretariat with the consent of the advisory board. In sum, the secretariat operates as an independent entity within the network and can therefore best be characterized as a network administrative network.

The work of EReg is primarily aimed at the sharing of knowledge and experiences to enhance the capacity of their own national organizations. For instance, the network organizes one-day conferences called EReg Academy. These are sessions in which policy makers and/or academic professionals come together to discuss a subject more in depth (EReg a, n.d.). The knowledge experts in EReg possess is communicated to policy makers in a way allowing them to improve policy without direct interference of experts in the political process of policy making (EReg respondent 1; EReg, 2019; EReg 2016; EReg 2019; EReg a, n.d.). Clearly, EReg is an information network.

With EReg being a network administrative information network, we expect bureaucratic accountability to be incorporated. The empirical findings indicate that account giving in the structure of EReg is focused on two tiers of the organization, the annual meeting, and the topic groups. The topic groups report back after each session to the chair and the secretariat, and they subsequently use a standardized format to present the result to the General Assembly (EReg respondents 2, 4, 5). The members of the topic groups are mostly from the operational level, whereas the attendees at the General Assembly are from the executive level, such as directors of the national public authorities participating in the TGN. It only very rarely occurs that the ministry at the national level plays a role in holding civil servants in topic groups to account. In almost all cases this task is taken up by the director of the national public authority participating in the TGN (EReg respondents 7, 8). In sum, as expected EReg incorporates bureaucratic accountability in its day-to-day functioning. The manner of monitoring the progress and activities is increasingly embedded in the network as is the pivotal role of the secretariat in professionalizing the network further. This monitoring is coordinated by the secretariat through structured reporting and runs from the activities of civil servants in the EReg topic groups to the directors of the national public authorities organized in the General Assembly. Hence, the bulk of the accountability line remains in the internal management structure of the network and the vertical line between a subordinate and a supervisor is present and well-integrated in the day-to-day functioning of the TGN. The subordinate is the civil servant participating in the topic group, who is highly aware of the need to provide information on the activities of the network to the supervisor. The supervisor is the director of the national public authority participating in the TGN and organized in the General Assembly of EReg. In most cases there is a rather passive attitude stemming from directors towards EReg activities (EReg respondents 1, 8).

WENRA: a participant governed harmonization network

The raison d’être of WENRA is to protect people and the environment from nuclear waste and radiation levels. WENRA tries to achieve this aim by harmonizing the implementation of safety reference levels (SRLs) among all participating organizations.

The WENRA network is structured as a participant governed network, with the networks’ activities carried out by the participants of the network with limited secretarial support. Regulatory bodies of European Union member states and third countries partake in the work of the network and the heads of the regulatory bodies convene twice a year for the General Assembly. In the structure of the working groups, the participant governed aspect of WENRA is most clear, with a focus on the individual working group member and the collaborative nature of the working groups. When a working group comes to a conclusion, the work will be presented to the plenary meeting of the national nuclear safety authorities.

The function of WENRA is harmonization, manifested in the creation of benchmarks or standards combined with a concern for their implementation at the domestic level (WENRA, 2019). The harmonization focus is clearly visible in the working groups operating within the network structure. Triggered by the Fukushima accident, the working group Reactor Harmonization Working Group (RHWG) reviewed the implementation of the SRLs set by WENRA in the member countries. RHWG produced revised standards in 2014 and has since conducted a pilot study to see whether two specific SRLs were implemented as intended (WENRA RHWG, 2019). The Fukushima disaster led the Council to decide on stress tests to be carried out by WENRA and ENSREG in all member states and to determine a uniform assessment.

After publication of the SLRs, the process of peer review starts to benchmark whether national regulators implemented the SRL in their legal framework. In the Dutch context, the national regulator is the ANVS, which prepares each WENRA meeting based on the agenda of the meeting which is send in advance (WENRA respondents 1, 3). When a working group discusses the setting up of draft SRLs, consultation takes place with the in-house experts of ANVS and the directors of nuclear installations (WENRA respondent 2). Each plenary meeting is prepared for by the most senior management level delegate (WENRA respondents 1, 2, 3). Next to the working group this delegate also attends the plenary as a second to the director of ANVS. The director is informed about these activities, but this seldom leads to discussions in the ANVS management (WENRA respondent 2). When a plenary meeting concludes with agreement on changes on for instance the SRLs, the ANVS then starts a process of benchmarking to make sure the new SRLs are implemented in the Netherlands. Dutch ministries are not involved in the work of WENRA because ANVS has its own mandate to develop rules regarding nuclear safety.

In a participant governed harmonization network like WENRA, we expected political accountability to be incorporated. However, the empirical findings do not provide sufficient ground to identify a single dominant accountability type. We can identify several accountability lines. First, the delegates participating in WENRA are aware that their knowledge is limited and often consult experts from within their home organization (ANVS). Due to the high level of expertise needed to understand the discussions within WENRA, there is no involvement of political principles at the domestic level in these consultations. The second accountability line consists of internal monitoring in the network between the national representative in the working group and the national representative in the General Assembly. This comes closest to bureaucratic accountability. The third line is account giving directed to external stakeholders such as the directors of nuclear installations whom are asked for feedback on SRLs (WENRA respondents 2, 3).

The lack of active involvement of an elected political principal – against our expectations – is something which is also noted by one of the respondents:

The ministry was not involved. Rather our expert advice was simply followed. Meaning that in parliamentary debates someone from ANVS sat directly next to the minister telling her what to do. (WENRA respondent 4)

In 2020 preliminary steps have been taken to rearrange the accountability structure to externalize the source of control by giving a limited degree of control to political principals in the home organization (Van Veldhoven – Van der Meer, January 17th 2020). It is still early days to determine whether this will result in practice in a dominance of political accountability over other accountability types

The EA: a network administrative governed harmonization network

The EA has 50 members, all of which are national accreditation bodies for the certification, verification, and testing of standards. Examples of fields in which conformity assessments and accreditation takes place are food safety, laboratories, mechanical testing, fire safety, and the railway system. The work of the EA is considered vital to create a level playing field, ensuring the smooth working of the EU internal market and the placing of products on the market.

The EA is structured as a network administrative governed organization. This means that an external administrative entity is set up to steer and organize the network. The EA has a secretariat comprised of eight full-time employees, with an independent status from the member organizations in the TGN. The tasks of the secretariat are, among others, the management of projects and contracts with the European Commission/EFTA, the contact point for stakeholders, management of the peer-evaluation programme, providing support to the executive and technical committees, and implementation of the EA communications and marketing strategy (EA, n.d.-a).

The EA also has other organizational components next to the secretariat. The member organizations convene in the General Assembly each year, in which peer reviews of accreditation bodies are discussed. The assembly is presided over by the chair of the executive committee. This executive committee is tasked with the implementation of the resolutions of the General Assembly and the general oversight of the network. The budget is decided upon in the General Assembly and stakeholders are represented through an advisory board.

The function of the EA is harmonization through the creation of new rules and procedures which are subsequently implemented by the national accreditation bodies (EA, n.d.-b). An example of how harmonization takes place in the EA is the publication of guidance documents, describing how a national accreditation body should interpret a standard set by EA (EA, 2018, October 19). Preceding the drawing up of these documents, rigorous peer reviews are held to establish the correct way of interpreting the standards by exchanging information on best practices. The execution of the interpretation set is also part of the harmonization, with strict sanctions attached to the harmonized rules. The ultimate consequence for not adhering to the accreditation guidance documents results in a permanent or temporarily expulsion of the national accreditation body from EA (EA respondents 1, 2, 3).

The type of accountability we expected to be incorporated in the functioning of a network administrative harmonization network is legal accountability. The empirical findings give strong support for the dominance of this type of accountability, which is based on an assessment of whether the agent has complied with the formalized principal’s rules. The audits taking place in the peer reviews are the most important technique deployed in the account giving structure (EA respondents 1, 2, 3).

Each national accreditation body is bound by the rules of the EA and is audited on whether the national authorities are upholding the standards of accreditation (EA respondent 1). The audits are conducted every four years by a team of national accreditation body delegates from different countries. The audit makes use of peer reviews with audit teams independent from national accreditation bodies. This indicates the prominence of legal accountability in the procedures, i.e., the source of control is external to the TGN and the degree of control over the civil servants in the TGN is high, possibly even leading to suspension of TGN member states in the case of non-compliance.

Conclusion

Is there an accountability deficit regarding TGNs? In the literature on TGNs, it is often assumed that TGNs should be judged along the lines of the political accountability type, with an actor giving account to a political forum such as a national parliament or a minister (see, for example, Papadopoulos, 2007). We showed by drawing on insights from contributions to public administration, IR and the political science literature that in practice civil servants in TGNs are held to account by a forum in various ways, dependent on the institutional set-up of TGNs. We find that in three out of four cases TGNs incorporated an accountability type which is most fitting to the institutional set-up of the TGN, leading to no or very limited accountability deficits. In these cases, a relationship existed between the actor and the forum, in which the actor had an obligation to explain and justify his or her conduct, the forum could pose questions and pass judgement, and the actor faces consequences. However, small accountability deficits arise in TGNs where the sanctions for non-compliance are limited, such as in the case of EReg and IMPEL. In WENRA we expected to find political accountability but we could not identify on the basis of our empirical findings a dominant type of accountability. This confirms the difficulties reported in the literature of attracting attention by elected politicians for the activities of the network (Busuioc, 2010; Papadopoulos, 2007, 2014; Raustiala, 2000; Sabel & Zeitlin, 2010; Slaughter, 2002). However, we also found that the challenge of TGNs to comply with traditional conceptions of accountability is from a normative perspective not problematic for all types of TGNs. TGNs with an information function such as EReg and IMPEL are aimed at sharing information between participants and in practice these TGNs comply with an accountability type expected to be dominant in these TGNs based on the function and governing style of the network. The absence of sanctions does not lead to a severe accountability deficit, given that the networks are organized around only sharing of information with peers. However, we also found that sanctions are absent in the case of WENRA and the most logical accountability type (i.e., political accountability) is missing in practice in this harmonization participant governed network. Hence, in the case of WENRA – a TGN with a high policy impact because of the harmonization function of this network and touching upon a salient policy field (i.e., nuclear energy) – it was even more important in order to prevent a severe accountability deficit that WENRA complied with the accountability standard of political accountability. Clearly, this was not the case. It should not come as a surprise that civil servants consider that in order to reduce the accountability deficit, there is a need for reforming the accountability relationships within WENRA towards political accountability.

There are some characteristics of the Dutch case which limit the extent to which the findings can be generalized to other contexts. First, the Netherlands can be seen as a pace setter of Europeanization in networks because of its international outlook and tradition of strong regulators. Second, compared to the unitary state of the Netherlands, federal states might have more difficulty satisfying the accountability standards due to veto players at the federal and sub-national level. Even if the Netherlands adopts the appropriate mechanisms for establishing accountability, this does not automatically affect accountability standards in other national contexts. Further research is necessary to shed light on the implications of participating countries in TGNs not evenly addressing accountability deficits at the national and sub-national level. Such follow-up research could also explore whether there are differences between countries with regard to how the choice is made for a particular accountability type and who is involved in these decision-making processes.

An important feature of this study is the novel perspective with which it contributes to the study of European regulatory networks (ERNs). The majority of studies of ERNs have been top-down and network-centered. Especially after the establishment and spread of European agencies modes of networked governance have come to be seen as the European Commission’s way to create an European administrative space and to address its administrative deficit. ERNs were the extended arms of the Commission administration. Next to these top-down approaches ERNs have been studied as sui generis type of networks. Numerous studies pertain to the emergence, structures, tasks, and functioning of these networks. What our study brings to this field is a bottom-up or state-centered perspective on the functioning and organization of ERNs and, more broadly, TGNs. What has often been overlooked is the fact that ERNs are extended arms of national administrations, too. As Egeberg (2006) once succinctly argued national administrations have become ‘double hatted’. The process of European integration ties together national administrative systems more closely to each other. At the domestic level, the increasing integration of national administrative agencies into ERNs affects the structure and organization of national political-administrative systems at their core. One of the most important implications of this integration is addressed in this study: accountability.

Acknowledgements

We would like to thank the three anonymous reviewers for their comments and suggestions.

Disclosure statement

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

Additional information

Notes on contributors

Daniëlle van Osch

Daniëlle van Osch is a PhD candidate at Leiden University, Institute of Public Administration, The Netherlands.

Rik de Ruiter

Rik de Ruiter is Associate Professor at Leiden University, Institute of Public Administration, The Netherlands.

Kutsal Yesilkagit

Kutsal Yesilkagit is Professor of International Governance at Leiden University, Institute of Public Administration, The Netherlands.

Notes

1 The number of documents studied authored by the network is: 67 for IMPEL, 31 for EReg, 119 for WENRA, and 113 for EA.

2 The number of news articles studied: 90 for IMPEL, 33 for EReg, 147 for WENRA, and 15 for EA.

3 The number of documents studied from the home organization: 18 for IMPEL, 14 for EReg, 4 for WENRA and 10 for EA.

4 The number of documents studied from the European Commission: 7 for WENRA, 8 for EReg, 34 for WENRA, and 7 for EA.

5 Observation 1: 18–21 September 2017; observation 2: 18–20 January 2017; observation 3: 20 February 2018; observation 4: 2–4 May 2018.

6 The number of interviews conducted is: 11 for IMPEL, 8 for EReg, 4 for WENRA, and 3 for EA.