p. 237Introduction

This chapter examines the rise of EU agencies. EU agencies—sometimes also labelled ‘authorities’ or ‘centres’—can be defined as ‘EU level public authorities with a legal personality and a certain degree of organizational and financial autonomy that are created by acts of secondary legislation in order to perform clearly specified tasks’ (Kelemen 2005: 175). Over the past three decades, the EU has established more than forty agencies in a wide variety of policy areas, with headquarters spread out across the member states. With the establishment of these agencies, along with the creation of other de novo bodies as described by Hodson and Peterson in Chapter 1, the EU has quietly managed to expand its regulatory capacity without directly increasing the size or capacity of the EU’s primary executive organ—the European Commission. EU agencies are now well established as an integral part of the EU’s model of governance. This chapter analyses why they have been created and what impact they are having on European governance.

Development

The first EU agencies—the European Centre for the Development of Vocational Training (Cedefop) and the European Foundation for the Improvement of Living and Working Conditions (Eurofound)—were created in the 1970s, but these were operational, rather than regulatory, bodies. The 1990s produced a second wave of agencies, this time dealing with regulatory issues, including the European Environment Agency (EEA) and the European Medicines Agency (EMA). A third wave of agency creation started at the beginning of the twenty-first century with the creation of the European Food Safety Authority (EFSA), the European Maritime Safety Agency (EMSA), the European Aviation Safety Agency (EASA), and the European Railway Agency (ERA). It has continued with the recent creation of EU authorities in the fields of securities, banking, and pension regulation. Most European regulatory-type agencies of the second and third generation advise the Commission on the technical or scientific aspects of regulatory problems, but have not been given the formal authority to take final and binding regulatory decisions.

In addition to the regulatory agencies on which we focus in this chapter, the EU has also established a handful of executive agencies that perform managerial tasks on behalf of the Commission, such as the Education, Audiovisual and Culture Executive Agency (EACEA), the European Research Council Executive Agency (ERC Executive Agency), and the Executive Agency for Small and Medium-sized Enterprises (EASME). It has also set up a number of agencies focusing on foreign and security policy, such as the European Union Institute for Security Studies (EUISS), the p. 238↵European Defence Agency (EDA), and others focused on justice and policing, such as the EU Judicial Cooperation Unit (Eurojust), the European Asylum Support Office (EASO), and the European Agency for the Management of Operational Cooperation at the External Borders (Frontex). While this chapter focuses exclusively on EU agencies with some regulatory dimension (see Table 10.1), the creation of these regulatory agencies should be understood as part of a broader trend in the EU toward delegating authority to autonomous bodies outside the structure of the European Commission.

Why EU agencies? Theories of agency creation

The growing role of agencies

The EEA and the other agencies established in the mid-1990s were ad hoc experiments in institutional innovation. But, over the past decade, EU agencies have become an integral aspect of the EU’s regulatory landscape. Today, proposals to expand EU regulation substantially are regularly accompanied by plans for the establishment of an EU agency, which may take on regulatory, informational, or executive tasks (Rittberger and Wonka 2010). As plans for new agencies proliferated, the Commission called for the establishment of a common framework on which to model them. The 2001 White Paper on European Governance (European Commission 2001b) contained a section on ‘Better application of EU rules through regulatory agencies’, and called for the establishment of a common framework for the creation, operation, and supervision of agencies. The following year, the Commission proposed such a framework for the agencies and, in 2005, a draft inter-institutional agreement on EU agencies was published. However, adoption of the agreement was blocked in the Council. In 2008, the Commission finally withdrew the stalled agreement and presented a new communication on European agencies that promised to relaunch the debate about their place in governance. Representatives of the Commission, EP, and Council subsequently met to discuss issues concerning the structure, supervision, and operations of the agencies—but they did not agree on a uniform framework for European agencies.

While the Council, Commission, and Parliament have been unable to agree on a standard template for agencies, they nonetheless have continued to create new ones. In recent years, agencies have been created in sectors such as energy—the European Agency for the Cooperation of Energy Regulators (ACER), financial market regulation (ESMA, the EBA, and the European Insurance and Occupational Pensions Authority, or EIOPA), and banking (the SRB), in which member states had long resisted the establishment of new agencies (see Table 10.1). In these fields, in p. 247↵which regulatory decisions entail substantial distributional consequences, member states had previously sought to rely on networks of NRAs to provide the necessary coordination without establishing EU bodies that might constrain national regulatory discretion. However, disastrous regulatory failures and a growing acceptance that the network model alone was inadequate led member states to agree to link these networks to powerful EU agencies (Kelemen and Tarrant 2011).

The politics of institutional choice: The birth of the European Environment Agency

In January 1989, Commission President Jacque Delors proposed the establishment of a European Environment Agency, to strengthen the Community’s information-gathering, monitoring, and implementation capacity. Member states, European institutions, and environmentalist groups all voiced support for the proposal, but were deeply divided over specific structural choices—especially those concerning the regulatory powers and effective independence of the new agency. Many in the EP favoured a body with regulatory ‘teeth’. In varying degrees, all member states opposed the idea that the agency could monitor the implementation of European environmental legislation by national regulators, preferring to restrict its task to the collection of environmental information and to networking with national, European, and international research institutions. The position of the Commission was ambivalent. On the one hand, officials in the Directorate-General for the Environment (DG ENV) were concerned about the criticism of industry and some member states that the Commission’s environmental proposals were not grounded in ‘good science’. They were even more concerned by the poor implementation of environmental directives. Hence the idea that the EEA could become a sort of inspectorate of national environmental inspectorates had a number of influential supporters within the Commission. On the other hand, this institution was reluctant to surrender regulatory powers to an agency operating at arm’s length. In a 1989 proposal, the Commission outlined four functions for the new body:

Because of the expected opposition by the member states, no inspection tasks were contemplated.

This proposal was quite distant from the EP’s ‘ideal point’. The fact that Beate Weber, rapporteur of the EP’s Environmental Committee, travelled to Washington p. 248↵DC to gain first-hand knowledge of the United States’ (US) Environmental Protection Agency (EPA) suggests the model of regulatory agency that European parliamentarians had in mind. The EP Environmental Committee maintained that the EEA should be given power to police environmental abuses, to supervise national enforcement of EC environmental regulations, and to carry out environmental impact assessments (EIAs) on certain Community-funded projects. Also, the composition of the management board became a point of contention. According to the EP, environmental groups should be represented on the board, alongside representatives from the member states, the Commission, and the EP itself, and the board should be allowed to take decisions by majority vote.

Comparing the preferences of the main political actors—member states, Commission, EP—with the provisions of the regulation setting up the agency, we see that the member states clearly won the contest over the structure and powers of the agency. The decisive influence of the national governments is revealed by the composition of the management board, which was to be dominated by member state appointees. As mentioned, the main task assigned to the agency is to provide the EU and the member states with environmental information and, in particular, ‘to provide the Commission with the information that it needs to be able to carry out successfully its tasks of identifying, preparing, and evaluating measures and legislation in the field of environment’.⁷ The wording is sufficiently vague, however, to make it unclear whether the agency would be allowed to directly influence policy formulation, for example by evaluating alternative proposals for regulatory measures. Political compromise produced an institutional design characterized by uncertain competences, unresolved conflicts, and failure to deal with the serious implementation problems of EU environmental policy. The compromise over the creation of the EEA established a model that then provided a rough template for subsequent agency proposals. Each agency was created in distinctive political circumstances, and the details of the powers granted to agencies and the oversight structures put in place varied as a result. However, a number of key elements of the design of the EEA have been replicated in other agencies. For example, agency management boards have continued to be dominated by representatives of member state governments and the agencies have served as hubs of regulatory networks that include (rather than replace) NRAs.

From committees to agency: The development of the European Medicines Agency

Our next example provides valuable insights into how obscure comitology committees have been transformed into a structure in which they become the operational arm of European agencies and link with national authorities to form transnational regulatory networks. The first attempt by the EC to regulate the testing and marketing of pharmaceutical products was a directive introduced in 1965 with the dual p. 249↵objective of protecting human health and of eliminating obstacles to intra-Community trade. This directive established only the principle that no medical drug should be placed on the market without prior authorization, and defined the essential criteria of safety and efficacy for drug approval.

The second phase of regulatory developments began in 1975, with another directive setting up the ‘multi-state drug application procedure’ and establishing the Committee for Proprietary Medicinal Products (CPMP). This Committee composed of national experts has played, and continues to play, a key role in the EU’s approach to the regulation of pharmaceuticals. Under the multi-state procedure, a firm that had received a marketing authorization from the regulatory agency of a member state could ask for the recognition of that approval by at least five other member states. The agencies of these countries had to approve or raise objections within 120 days. In the event of objections, the CPMP had to be notified and would express its non-binding opinion within sixty days. The procedure did not work well: the national agencies did not appear to be bound either by the decisions of other regulatory bodies or by the opinion of the CPMP. Subsequent simplifications failed to streamline the approval process, as national regulators continued to raise objections against each other almost routinely. Hence firms generally chose to continue to seek authorization from each national agency separately.

A different approval process was introduced in 1987 for biotechnology and other high-tech products. This new ‘concertation procedure’ required that the application for the authorization be filed both with the national authorities and with the CPMP. The country in which the authorization had been filed acted as rapporteur, but, unlike under the old multi-state procedure, no decision on the application was to be made by any member state before the CPMP had expressed its opinion. The final decision remained with the member states, however. The evaluation of the application, led by the rapporteur country, was carried out at the same time in all of the member states—hence the name ‘concertation procedure’. The new process was an advance with respect to the old practice, but was nevertheless problematic for firms because, as with the previous procedure, there was a tendency for delays in the notification of decisions following the CPMP opinion. Waiting for all countries to notify their decisions following the Committee’s opinion could result in serious delays in a firm’s ability to start marketing a new drug.

In 1995, the problematic multi-state and concertation procedures were replaced by three new approaches and a new agency (see Box 10.1). The multi-state procedure was replaced by a decentralized procedure, which continues and reinforces the principle of mutual recognition introduced in 1975; the concertation procedure was replaced by the centralized procedure set out in the same regulation that also established the EMA (see Box 10.2).⁸

Under the centralized procedure, applications are made directly to the agency, leading to the granting of a European marketing authorization. Use of this procedure is compulsory for products derived from biotechnology and optional for other innovative medicinal products. The EMA is also called on to arbitrate disputes arisingp. 250↵

under the decentralized (mutual recognition) procedure. Opinions adopted by the EMA in either the centralized procedure or following arbitration lead to binding decisions formally adopted by the Commission.

The technical work of the agency is carried out by the Committee for Medicinal Products for Human Use (CHMP, the successor of the old CPMP), by the Committee for Veterinary Medicines (CVMP), and by two smaller and newer bodies: the Committee for Orphan Medicinal Products (COMP), established in 2001 and charged with reviewing designation applications from persons or companies who intend top. 251↵

develop medicines for rare diseases (‘orphan drugs’); and the Committee on Herbal Medicinal Products (HMPC), established in 2004 to provide scientific opinions on traditional herbal medicines. A network of some 3,500 European experts underpins the scientific work of the EMA, and of its committees and working groups.

The CHMP (with similar rules applying to the CVMP) is composed of two members nominated by each member state for a three-year renewable term. These members, in fact, represent the NRAs. Although Commission representatives are entitled to attend the meetings of the Committee, the Commission is no longer represented, no doubt to emphasize the independence of the CHMP.

In keeping with the political compromise common to the establishment of all EU agencies, the new EMA was put under the control of a management board dominated by member state appointees. When the EMA was restructured in 2004, member states acquiesced to the Parliament’s demand that EP and stakeholder representatives p. 252↵be added to the Agency’s management board, although member state representatives still held a majority. The Council also agreed to the Parliament’s demand that the nominee for the position of agency executive director appear before the Parliament prior to her or his formal approval by the Council.

In fact, the Committee has become more important, as well as more independent, since the establishment of the EMA. In the new situation, Committee members have greater incentives to establish the agency’s, and their own, international reputation than to defend national positions. Using Alvin Gouldner’s (1957: 58) terminology, we may say that the agency creates a favourable environment for the transformation of national regulators from ‘locals’ (that is, professionals who have primarily a national orientation) to ‘cosmopolitans’, who are likely to adopt an international reference-group orientation. It does so by providing a stable institutional focus at the European level, and a forum in which different risk philosophies are compared and mutually adjusted, and by establishing strong links to national and extra-European regulatory bodies. Overall, the EMA has been highly successful at least insofar as pharmaceutical manufacturers report satisfaction with the agency’s centralized procedure and many companies seek centralized authorizations (Kelemen and Tarrant 2011). In spite of this success, the EMA will have to move from its home in London once the UK leaves the EU in light of its 2016 referendum vote on membership.

The pursuit of regulatory credibility: The European Food Safety Authority

The food sector is an area in which EC regulation dates back to the earliest days of the Community. Traditionally, policy on food safety was developed by the Commission, assisted by a large number of comitology and expert committees. Several regulatory failures—of which the bovine spongiform encephalopathy (BSE), or ‘mad cow disease’, epidemic attracted the greatest public attention—revealed the inadequacy of the traditional approach. The BSE outbreak exposed serious shortcomings in the overall coordination of European policies on agriculture, the internal market, and human health. In 1996, the EP set up a temporary committee of inquiry into BSE. The committee concluded that both the Council and the Commission had neglected their duties, and that the UK government had exerted pressures on the Commission’s veterinary services in order to avoid Community inspections and to prevent the extent of the epidemic being made public. The Commission was criticized for having given priority to the management of the beef market rather than to the risks to human health posed by BSE and for having downplayed the problem despite concerns raised by a number of experts. The Committee of Inquiry also noted that there had been severe problems with the workings of the Commission’s Scientific Advisory Committee.p. 253↵

Responding to these and other criticisms, in 1997 the Commission issued a Green Paper on the general principles of food law in the EU. It was followed in 2000 by a White Paper on food safety, proposing the creation of what became the EFSA in the context of a reform of the entire food safety system, ‘from farm to table’. EFSA was to take on responsibilities relating to the risk assessment and risk communication parts of the regulatory system envisaged by the Commission. However, risk management, comprising legislation and control, was not to be transferred to the agency. EFSA was to be guided by the best science, independent of industrial and political interests, open to public scrutiny, scientifically authoritative, and closely linked to national scientific bodies. The reform proposals contained in the White Paper formed the basis of a regulation that laid down the general principles and requirements of food law, established EFSA, and sets out its mission (see Box 10.3).

The organizational design of EFSA is broadly similar to that of the EMA: a management board, an executive director, a scientific committee, and a number of scientific expert panels and their working groups. There are, however, some important differences that are best understood in light of the credibility crisis of EU food safety regulation in the aftermath of BSE and the rising power of the EP. The BSE crisis exposed the potential dangers of interference by national officials in objective risk assessments. Under pressure from the Parliament, the member states agreed to put EFSA under control of a management board that did not guarantee each member government a representative. The board comprises fourteen members appointed by the Council—in consultation with the Parliament, from a list drawn up by the Commission—plus an additional representative of the Commission. Four members must have a background in organizations representing consumer and other interests in the food chain, and no member is an official government representative. The principle of one representative per country has instead been retained in the composition of p. 254↵the advisory forum, which assists the executive director and advises on scientific matters, priorities, and work programmes.

The tension between the desire to enhance regulatory credibility by appealing to independent scientific expertise and the refusal to delegate regulatory powers to EFSA has been temporarily resolved by the doubtful expedient of an organizational separation of risk assessment (the function assigned to the authority) and risk management, which remains the responsibility of the Commission. However, the separation of risk assessment and risk management is problematic, because while the two functions are conceptually distinct—one dealing with scientific issues; the other, with economic, legal, and political issues—they are closely intertwined in practice. The setting of rational regulatory priorities, for example, entails economic, political, and scientific judgements that cannot be easily separated. Again, the determinations of the risk analysts can effectively pre-empt the decisions of the risk managers. Thus it is often impossible to determine with certainty whether a ‘dose–response function’—measuring the probability of an organism’s response to different levels of toxicity—follows a linear or a non-linear model, yet the scientists’ choice of one or the other model is crucially important to the determination of an acceptable level of risk (Majone 2003). Because risk assessment and risk management are so difficult to separate in practice, the refusal to set up a regulatory agency fully responsible for food safety entails a serious accountability deficit without solving the credibility problem.

Independence and accountability

In any democracy, there is a tension between the desire for regulatory bodies to be simultaneously independent and accountable. As we saw at the beginning of this chapter, the desire for policy credibility provides an important functional motivation for agency independence. However, in a democratic polity, when regulatory powers are entrusted to a non-elected body, there is understandably a desire to balance this independence with a healthy dose of accountability. The basic problem is always how ‘to control and validate the exercise of essentially legislative powers by administrative agencies that do not enjoy the formal legitimacy of one-person, one-vote election’ (Stewart 1975: 1688). Law-makers can try to ensure some measure of accountability by giving the agency very specific instructions in the laws that they ask it to implement, putting in place a variety of oversight and judicial review mechanisms, and imposing rigid administrative procedures that force the agency to act in an open and transparent manner (McCubbins et al. 1987).

From the outset, EU agencies have been subject to a variety of control mechanisms designed to ensure that they cannot deviate too far from the will of their political masters. This oversight has long been evident in the composition of agency management boards, whereby member states ensured that their representatives could keep a watchful eye over agency operations. As the power of the EP has grown over the p. 255↵past two decades, it too has asserted itself as a political master of the EU agencies. This assertion has been evident both in the EP’s scrutiny of agencies’ budgets and in the Parliament’s demands that the agencies adopt more formal, transparent, and judicially enforceable administrative procedures. The impact of these demands is evident in the regulations founding a number of the recently created EU agencies. For example, the EASA and the ECA, both of which have been given authority to make decisions concerning the safety of products, are required to follow detailed, transparent procedures, to give reasons for their decisions, and to provide applicants access to a board of appeal to contest decisions. Substantial accountability mechanisms have been put in place, and certainly the EU agencies operate in a far more transparent and accountable manner than the comitology committees that preceded them.

The network model

It is clearly impossible to transpose to the EU the American model of federal agencies operating independently from the regulatory authorities of the states. Regardless of what one thinks of the alleged legal obstacles to the adoption of such a model, it is certain that the member states would reject it. However, again and again, member states have proven themselves willing to construct systems in which the national regulators become components of EU-wide networks, coordinated by European agencies.

The model in which an EU agency serves as the coordinating hub of a network of NRAs should not be confused with a model of governance that relies exclusively on a network of NRAs in the absence of a European agency. Some scholars (Eberlein and Grande 2005) and many practitioners have argued that EU-wide networks of NRAs provide an effective alternative to establishing more formal, centralized EU agencies. In some sectors, such as energy and financial market regulation, many national governments long resisted calls for the establishment of EU agencies by insisting that networks of NRAs could achieve the needed cooperation and harmonization. However, the actual experience with such networks—in fields from pharmaceuticals, to telecoms, to energy, to financial market regulation—has been unimpressive. EU networks regularly fail to deliver regulatory harmonization because pursuit of national self-interests regularly trumps the impact of ‘professionalization’ and other normative pressures that network governance theorists claim should facilitate effective cooperation. Indeed, it is for this very reason that some of these loose, ineffectual networks were eventually subsumed into more centralized structures under the leadership of a European agency (as in the fields of medicines, energy, and financial services regulation).

The network model has been crucial to the proliferation of EU agencies. By relying on networks of NRAs, EU agencies render themselves less threatening to those who oppose centralization of power. Nevertheless, by creating and coordinating networks of NRAs, the EU agencies can encourage the spread of common regulatory norms p. 256↵and practices across the member states. Ultimately, the agencies manage to harness the capacities of existing NRAs to serve European ends.

Conclusions

As shown by most chapters in this volume, the rate of institutional innovation in the EU, after six decades of integration, is still remarkable. This is certainly true in the case of European regulatory agencies. The quantitative growth of EU agencies since the 1990s and, with it, expansion of the range of policy areas in which agencies play a role in governance is striking. More recently, we have also seen an increase in the substantive powers of at least some EU agencies, with doctrinal change at the Court of Justice opening the way for a greater delegation of regulatory powers to these agencies.

Inter-institutional politics continues to play a crucial role in the design of EU agencies. In most areas of regulation, design of new regulatory agencies today requires the agreement of the Council, the EP, and the Commission, each of which can simply refuse to propose reforms that would undermine its own authority. The precise terms of the compromise between these institutions varies in the case of each agency and they have failed thus far to agree on a standard template. Nevertheless, the basic outlines of a model of EU agencies has emerged and is now a vital feature of the Union’s regulatory landscape. EU agencies rely heavily on NRAs, but the agencies provide central leadership and harness these authorities into pan-European networks. The two most dramatic crises faced by the EU in recent years—the euro crisis and the refugee crisis—have each revealed the need for greater centralization of regulatory authority in their respective fields, and these crises have prompted the establishment of new agencies (such as ESMA, the EBA, EIOPA, the SRB) or the dramatic scaling up of existing ones (Frontex and EASO). It seems highly likely that the number and authority of EU agencies will continue to grow in the decade to come.