Introduction
Every government needs systems for taking thousands of highly technical decisions to ensure that market transactions are safe, sustainable and compatible with international requirements (Page Reference Page2001). Irrespective of the level of government, choice and evaluation of instruments are challenges for administrations that are becoming increasingly pressing with the mounting number of instruments and the objectives that instruments have to achieve (Peters and Van Nispen Reference Peters and Van Nispen1998; Kassim and Le Galès Reference Kassim and Le Galès2010). Using the EU to discuss the fit of instruments, we can see a multiplication of governance tools including ‘soft’ tools and depoliticised agencies (Treib et al. Reference Treib, Bähr and Falkner2007). This article develops and tests a model to compare the added value of one instrument with others. It does so by studying to what extent the increasing use of agencies is actually an innovation in governance (defined here as ‘norm setting’). Although the EU is the testing ground, ‘agencification’ and the model to asses this development in governance are of general relevance (Christensen and Lægreid Reference Christensen and Lægreid2006).
The EU has relied on comitology for decades. Comitology is hierarchical (top down) rule making in which the European Commission decides on new rules while being controlled by teams of officials from national administrations. The image of comitology as being opaque and sensitive to heavy political lobbying – and hence producing poor measures – was one of the reasons why the EU turned towards agencies (Groenleer Reference Groenleer2009). Compared to the traditional political negotiations, the reference to agencies as ‘non-majoritarian’ suggests a major move towards depoliticisation (Majone Reference Majone1996; Everson et al. Reference Everson, Majone, Metcalfe and Schout1999). The EU agency literature has begun to address the question of whether agencies have indeed developed into the independent regulatory authorities that seemed to be the answer to the shortcomings in the EU's regulatory system (Dehousse Reference Dehousse1997: 246–47). One way to qualify the expectations of change in governance emerged from political scientists and (neo-)institutionalists who started to point to the pressures preventing overhauls. Neo-institutional theory balances ideas about organisational change and centripetal forces associated with traditions, values and interests (Moe Reference Moe1987). Stabilising forces that prevent formal changes in instruments from becoming innovations include existing interests of the major institutional players (Kelemen and Tarrant, forthcoming), lack of resources, and the EU's rule system, including the Meroni doctrine blocking delegation of tasks of the European Commission and constraints imposed by the EU's human resource policy (Schout and Pereyra Reference Schout and Pereyra2011). Hence, Majone's (Reference Majone1996) hope of depoliticisation through agencies can be confronted to the view that agencies result from inter-institutional compromises in which their independence is diluted.
Similar to agencies at the national level, EU agencies held the promise of better management, technical expertise, flexibility, transparency and networking (Majone Reference Majone2002). However, agencies now cannot be compared to the (apparently malfunctioning) instruments they superseded many years ago. The original tools may have developed too. A dynamic framework is needed to compare the relative merit of the agency instrument.
If agencies are different, we need to specify in what ways they actually are innovative. This study recasts the question of newness in asking whether agencies improved the legitimacy of EU policies. The literature on autonomous agencies usually ties the legitimacy question to legality and democratic control (Everson et al. Reference Everson, Majone, Metcalfe and Schout1999). However, ‘legitimacy’ may also help to offer a broader perspective on studying the performance of agencies. It is used here to study the actual innovation in EU governance resulting from the move from more traditional instruments, such as informal networks or comitology, to agencies. In this way we have a tool to compare instruments while allowing for changes in instrument design over time.
The problem of evaluating agencies as instrument has been noted many times before (European Court of Auditors 2008). The European Parliament has struggled with this question in view of its controlling powers. Instead of assessing the effectiveness of agencies, it restricts itself to a rather legalistic discharge of agencies budgets (Jones, forthcoming). The Commission experiences similar difficulties in defining the added value. As a consequence, discussions about abolishing the agency instrument resurface now and then (CEC 2008). EU agencies are generally positively evaluated on the basis of their results – not on the basis of their added value. A ‘meta-evaluation’ of the evaluations of EU agencies typically concludes:
How a policy would work in the absence of an agency is a point which is most often missed. Generally, no alternative to an agency is mentioned in the reviewed documents. …The reviewed material does not say much about the reasons why the agency option has been preferred to others, and evaluative information is even more scarce and superficial as regards the continued rationale of the agencies (Eureval 2008: iv).
The discussion of the added value of agencies compared to other instruments is not just of academic interest. Instruments are not neutral (Lascoumes and Le Galès Reference Lascoumes and Le Galès2007). In the case of the EU, agencies are often regarded as changing the balance of powers between Commission, EP and member states (Everson et al. Reference Everson, Majone, Metcalfe and Schout1999, Kelemen and Tarrant, Reference Kelemen and Tarrantforthcoming). Hence, choice of instruments is as much a question of effectiveness as of altering power relations (Kassim and Le Galès 2010). A methodology to evaluate and compare instruments is therefore important in view of looking beyond political fears and preferences for one type of instruments or another and to actually compare what the effects of changes have been.
As this study offers a new approach to assess the relative merit of the agency instrument, it needs to be carefully placed among existing studies. Firstly, it departs from mainstream EU agency literature by following the newly emerging sociological policy instrumentation approach as introduced by Lascoumes and Le Galès (Reference Lascoumes and Le Galès2007) and Kassim and Le Galès (2010). The study of (EU) instruments has generally been based on functionalist assumptions (Kassim and Le Galès 2010). At the same time, EU agencies have been studied from a rational power perspective in which actors are assumed to understand the workings of an instrument (e.g. Kelemen and Tarrant, Reference Kelemen and Tarrantforthcoming). The political sociology approach is more dynamic and follows the “career” of instruments by looking at how they are operationalised incrementally (Kassim and Le Galès 2010: 7). It prevents us from taking instruments as “natural”, as being “at our disposal” or as seeing instruments as “offering no alternative” (Lascoumes and Le Galès Reference Lascoumes and Le Galès2007: 2). Instruments are institutions (“organisations”) that structure behaviour and these institutions change through time (Kassim and Le Galès 2010). To understand the added value of the agency instrument we need to explicitly focus on the career of the agency instrument and its alternatives. Although the approach presented below builds on the dynamic view on instruments as presented by Lascoumes and Le Galès (Reference Lascoumes and Le Galès2007), it departs from their work by assuming that even networks can be studied as (developing) organisations (see below).
Secondly, the methodology should add insight into where EU agencies are different. It is by now generally accepted that agencies did not develop into independent authorities but have become a mixture of scientific bodies, traditional EU institutions and networks (Trondal Reference Trondal2010: 129–130). This, however, is too indiscriminate to identify the relative merit of the agency instrument. Does that mean agencies are different, e.g. leaning towards independent expert input, or have they blended in with existing forms of supervision?
Thirdly, as this is a case study about a specific agency, the methodology should inform us about this particular case. The European Aviation Safety Agency (EASA) is used as our N = 1 case to test the comparative framework. All agencies are different, leading Majone (Reference Majone2002: 300) to conclude that ‘agency’ is an “omnibus label”. The primary purpose of an N = 1 study is not to generalise to a population but to generalise to theory (Yin Reference Yin2003). The theoretical question concerns the relevance of the legitimacy model for structuring the debate about the added value. In generalising to theory, Yin (Reference Yin2003: 38) warns not to try to select a representative case. On this basis EASA is relevant, not because it is a typical or an outlier case, but because there are different claims about its contributions to EU governance and these claims can be examined by designing and testing a much-needed research approach. EASA is used to examine the extent to which the model helps to structure the analysis of the relative merit of an agency.
EASA's formal evaluation concluded that it constituted a “fundamental” change and that, contrary to the earlier aviation safety regulation system, it operated effectively and efficiently (Horvath 2007: 8). Similarly, Pierre and Peters (Reference Pierre and Peters2009: 344, 352–3) concluded that EASA changed the aviation safety landscape from an “amazingly loose” club into a transparent bureaucracy which is “better designed to articulate [public interest] than the close-shop nature of” its predecessor. The historical account of EASA by Kassim and Stevens (Reference Kassim and Stevens2010) also underlines the widely felt frustration in safety regulation prior to EASA and the almost surprising shift towards a stronger European regulatory role involving EASA. Yet, they underline that EASA has helped to extend the EU's legal competences despite great resistance but that “it is sometimes less successful in implementing it” (Kassim and Stevens Reference Kassim and Stevens2010: 153). Groenleer et al. (Reference Groenleer, Kaeding and Versluis2010) are also more reserved in their assessment of the added value of EASA. While focusing on inspections and enforcement tasks, they conclude that EASA may “potentially” add value by stimulating learning by distributing and discussing inspection reports within the network of national aviation safety agencies. Valuable as these studies are, they typically offer no insight into the value of the agency instrument as such.
This study focuses on EASA's rule making tasks. It compares EASA to its predecessor (Joint Aviation Authority (JAA)) and to the more traditional comitology structure. Hypothetically, the work of its predecessor could have been taken over by the Commission in combination with a comitology system without creating an agency.
The comparative framework (section 2) follows the obligation arising from the EU's better regulation agenda to carefully argue the choice of instruments (CEC 2009). To focus the discussion, and without aiming at being complete, the emphasis is on a selection of input and output indicators. Using the model, the added value compared to other instruments is discussed in relation to the EU's traditional system for technical decision-making (comitology, section 3), the structure that existed before EASA (JAA, section 4) and EASA (section 5). The conclusions address the question of the relevance of the legitimacy-based approach for analysing the added value of this agency in relation to safety of planes. If the analysis is helpful in structuring the comparison between instruments, we have the basis for studying the added value of agencies more generally through replication.
Legitimacy as organisational challenge
Agencies in the EU have ‘mushroomed’ in number and size, as has the literature on EU agencies. EU agencies have been studied from historical, institutional, management and functional perspectives (Groenleer Reference Groenleer2009; Kelemen and Tarrant, Reference Kelemen and Tarrantforthcoming). The review by Reference Busuioc, Groenleer and TrondalBusuioc et al. (forthcoming) identifies a variety of perspectives that have been taken on EU agencies and refers to literature on agencies as part of the EU's compounded multilevel governance system, to studies on every-day decision-making in EU agencies, and to management and accountability studies. Kelemen and Tarrant (Reference Kelemen and Tarrantforthcoming) emphasize that agencies were mainly created out of rational political motives while others study functional needs (Groenleer et al. Reference Groenleer, Kaeding and Versluis2010). The debate now needs to move beyond the politics and pressures behind agencification and address their added value. A framework is needed to compare the benefits of agencies to those of other instruments. Comparative frameworks should depend on the objective of the comparison. Assessing and comparing organisations involves many potentially relevant variables (Powell Reference Powell1987). External control mechanisms, political salience, internal and external leadership patterns, and personalities are just some of the variables influencing organisational behaviour and hence legitimacy. Therefore, choices are needed to enable a comparison.
‘Legitimacy’ is a relevant starting point because agencies were one of the responses to the EU's legitimacy crisis in the 1990s caused by lacking faith in the quality, transparency and implementation of EU legislation and actions. Agencies were assumed to offer greater transparency, expert authority, efficiency, better informed decisions and better implementation (CEC 2001). The Commission has been discussing how to evaluate agencies but has underlined the difficulties in coming up with an approach (CEC 2008). Acknowledging that experience with agencies in OECD countries qualifies their theoretical advantages (Pollitt et al. Reference Pollitt, Talbot, Smullen and Caulfield2004), the suggestion explored here goes back to one of the basic ideas behind agencies: agencies are an instrument to strengthen the legitimacy of EU rules (Majone Reference Majone1996).
Legitimacy is the belief people and industry have in the rules and actions of government (Weber Reference Weber1968). The public has to trust the policies produced by the systems governments have in place for taking technical decisions. ‘Legitimacy’ is a vast subject and has been used differently over time and between authors. Yet, broad consensus has emerged over the composite nature of legitimacy. The current debate concentrates on input, output and throughput legitimacy (Scharpf Reference Scharpf1999). Input legitimacy addresses democratic control (government by the people) while output legitimacy concerns government for the people (effectiveness or substantial legitimacy). The legitimacy debate has shifted to throughput legitimacy due to limitations of democratic control over – in particular international – governments (Lindseth Reference Lindseth1999). Throughput legitimacy is about how agreements are reached and opens debates about deliberative democracy, transparency and impact assessments (Curtin Reference Curtin2005; Vibert Reference Vibert2006). However, we follow Bekkers and Edwards (Reference Bekkers and Edwards2006, p. 44) who regard throughput legitimacy as part of input legitimacy because it is linked to questions about who has access to decision-making processes and about the – hierarchical and public – control over the procedures.
The distinction between input and output legitimacy helps to operationalise legitimacy in terms of the underlying accountability mechanisms. Curtin defines accountability as “those arrangements made for securing conformity between the values of a delegating body and those to whom powers are delegated” (2005: 87). Hence, accountability mechanisms are the control instruments that make it possible for, among others, politicians to monitor the organisation (e.g. through work planning and evaluations) and that provide the basis for public trust in the actions of public bodies (e.g. through openness of decision-making). Within the constraints of this study, this discussion on the controls to support the legitimacy of organisations can only remain at an exploratory level (see Busuioc Reference Busuioc2009 for a review of different accountability mechanisms). Accountability mechanisms such as transparency rules or input in the form of expert information would require more in-depth analysis and discussion regarding the extent to which they actually translate into legitimacy (Mayntz Reference Mayntz2010).
Accountability mechanisms steer the way in which instruments are used by specifying their objectives, procedures, resources and performance criteria (table 4). As argued by Lascoumes and Le Galès (Reference Lascoumes and Le Galès2007), policy instruments are organisations and can be studied in terms of organisational design. Instruments structure power relations between government and society by shaping, among others, negotiations, agreements and conditions (Kassim and Le Galès 2010). If accountability of instruments is poorly designed it creates principal-agency problems and inconsistent behaviour (Thatcher and Stone Sweet Reference Thatcher and Stone Sweet2002). It can be too elaborate and create opportunities to manipulate performance assessments, for example by emphasising specific objectives while ignoring others. Accountability mechanisms can also cause inefficiencies if procedures are time-consuming or inflexible (Schout and Pereyra Reference Schout and Pereyra2011). They can have a mere symbolic function (DiMaggio and Powell Reference DiMaggio and Powell1983) or lead to fake legitimacy if, for example, evaluations are carried out by stakeholders. Hence, more accountability does not necessarily lead to more legitimacy. If controls are poorly designed, intransparency, inefficiencies and poor policies can undermine the trust in public organisations (e.g. Medina 1997).
The comparative framework is summarised in Table 1. It underlines that governance involves overlapping accountability mechanisms. This study concentrates, firstly, on input legitimacy including hierarchical controls (limited here to political approval of work programmes and budget control) and administrative controls (defined as the rules for decision-making such as work planning, impact assessments, transparency and evaluations). The EU has made great strides in administrative legitimacy. The upswing of better regulation initiatives (CEC 2009) has changed administrative accountability rather drastically to the extent that certain kinds of depoliticisation have been introduced. Moreover, input legitimacy involves legal oversight (access to justice and appeal mechanisms, Majone Reference Majone2002) and functional cooperation (mutual learning between experts, Smismans Reference Smismans2004). Expert communities have played a major role in the legitimacy debate as a way to solve shortcomings in democratic legitimacy. Professional values and the resulting peer pressure are particularly important in discussions of agencies as an instrument to depoliticise decisions (Majone Reference Majone1996).
Framework to assess the legitimacy of EU agencies
With the better regulation agenda, the EU has defined for itself a set of objectives such as transparency, consultation and participation, evidence-based policies, subsidiarity, proportionality, and instruments with the least administrative costs (CEC 2009). Together with evaluation obligations, better regulation includes output and input accountability requirements. All EU decisions, including those originating from EU agencies and comitology, have to respect better regulation principles (CEC 2009).
Output legitimacy relates to effectiveness. Any performance assessment is difficult in view of competing evaluation criteria (Powell Reference Powell1987). Hence, the Commission has had major difficulties in agreeing on a framework to regularly evaluate agencies (CEC 2008). Following the EU's better regulation agenda (CEC 2009), we can relate to criteria such as effectiveness (do experts agree on the value of the deliveries?), flexibility (responsiveness to new conditions) and subsidiarity (the EU's ambition to leave responsibilities as much as possible at the national level) as emphasised in. The regular reviews of instruments and interviews – including with industry – can be used to uncover the perceptions of performance (see the case of EASA below). Flexibility is included here because of the complaints on inflexible EU standards and the hope that agencies would be more responsive to changes in markets and technologies (Everson et al. Reference Everson, Majone, Metcalfe and Schout1999).
The study is based on two background reports written for the Dutch transport ministry (for details see Schout Reference Schout2008, Reference Schoutforthcoming). Additional interviews took place to monitor the developments in aviation safety regulation until 2010. Interviews were conducted in the process of this study with a number of officials from the EU Commission, the General Secretariat of the Council, industry, two national transport ministries, two national aviation authorities and EASA. Apart from two interviewees, those concerned had experience with EASA and JAA. Most interviewees were followed for a period of 2 years. The structure laid out in Table 1 provided the semi-structured basis for the interviews. The underlying reports and drafts of this paper have been read by representatives from industry, the EU Commission and national civil servants. Their feedback is incorporated.
Comitology
For planes to fly safely and efficiently from one airport to another, thousands of safety standards have to be agreed concerning planes, operators, airports and air traffic management. JAA covered the regulation of planes. An Airbus 380 for instance is composed of 12000 parts, each requiring certificates based on predefined safety standards. In addition, many procedures have to be defined for the operation of a plane, ground handling, noise emission, etc. These standards and procedures, and the resulting certificates, have to be regularly updated due to developments in technologies, new safety principles, etc.
Detailed market regulation is increasingly formulated by international organisations, leaving implementation and application to states. Multilevel international organisations such as the specialised UN agencies dealing with maritime or aviation safety have specific arrangements to pool expertise for taking measures while ensuring transparency and accountability of decisions (Woll Reference Woll2006). The EU has relied mostly on comitology to formulate more routine and technical regulation, but a trend has emerged towards complementing ‘softer’ arrangements with legal approaches. In addition, the EU has used sui generis arrangements particularly in new EU policy fields which mostly involve some sort of ad hoc open coordination of different sorts. Such arrangements come in many shapes and sizes and have developed over time into more permanent structures, including into agencies arrangements, formal networks and comitology (Héritier and Lehmkuhl Reference Héritier and Lehmkuhl2008).
Aviation safety regulation in Europe has traditionally been mostly based on a sui generis arrangement (JAA) outside the EU. Yet, for reasons explained below, there have for a long time been comitology committees. Currently, two comitology committees are listed in the Commission's comitology registry in relation to aviation safety. The first is commonly known as the ‘EASA committee’. It is a more technical committee discussing the implementing rules which the Commission adopts and which are drafted by EASA. The second is the Aviation Safety Committee, which decides on issues where EASA has no formal role and which handles the more political issues. For example, it deals with blacklisting of planes and operators, cooperation with international partners and emission trading schemes for aviation. EASA can be involved in this committee when e.g. technical input is needed in relation to the blacklist. The roots of these committees go back to long before EASA was set up. They have traditionally been staffed by officials from the ministries although it is now common practice, at least in the Netherlands, for experts from the national aviation safety agency to accompany the negotiator. With the extension of EASA's tasks into regulating safety of open skies, EASA is now also involved in the European Single Sky committee.
One way to expand safety regulation in the EU could have been to extend the role – and resources – of the Commission. This would also have required an extension of the work of the aviation safety committees. Table 1 can be used to address the question of whether this would have been a defendable alternative from a legitimacy perspective.
Accountability mechanisms
Comitology has been criticised for making EU decision-making opaque due to the many committees and regimes and because of behind-the-scene influence of Member States and industry on decisions. The weaknesses of comitology became painfully clear in the EP's report on the BSE crisis, which concluded that it was impossible to determine by whom and on what basis decisions had been taken and that experts involved had not been independent (Medina 1997). The poor reputation of comitology was one of the reasons for introducing agencies (Everson et al. Reference Everson, Majone, Metcalfe and Schout1999).
The criticism made reform of comitology unavoidable. In terms of hierarchical control, major changes have included increasing the powers of the EP (Table 2). Until 1987, political accountability through the EP was virtually absent. Hierarchical control at Member State level was also poorly developed because committees were often regarded as technical – not political – which meant that national representatives were hardly coordinated (Brandsma Reference Brandsma2010). The Comitology Decisions of 1987, 1999 and 2006 as well as the Lisbon Treaty constituted major steps towards upgrading transparency and the involvement of the EP (Vos Reference Vos2009). The Comitology Decision of 1999 resulted in more information to the EP on draft measures, voting results and an annual report on the committees. The Comitology Decision of 2006 introduced the ‘regulatory procedure with scrutiny’ (European Council 2006) giving EP powers to block measures with a general scope.
Accountability analysis of comitology
Administrative accountability was modified through the application of the Commission's better regulation agenda. Comitology is now part of the Commission's work planning system, which includes the obligation to carry out impact assessments. Transparency was improved through the creation of a ‘registry’ – a website listing all committees and providing access to agendas and reports of meetings. However, as interviews with industry showed, the richness of the information in the registry still leaves much to be desired. The transparency of experts involved was also improved by, among others, statements of independence (Vos Reference Vos2000). Legal control over comitology has functioned quite effectively judging by the major cases before the European Court concerning procedures, outcomes, transparency and accountability towards the European Parliament (Bradley Reference Bradley2006). The reinforcement of administrative accountability through legal control underlines that accountability instruments can be mutually reinforcing.
In terms of output legitimacy, comitology has been criticised for being slow and too much concerned with detail. The Commission, among others, has noted that well-argued proposals based on impact assessments are subsequently diluted by committees. Hence, it has considered options for decision-making to become: “simpler, faster and easier to understand … [and to] improve accountability” (CEC 2001: 31). Furthermore, comitology has a reputation of being politicised and leading to suboptimal decisions (Majone Reference Majone1996). On the other hand, it has been discussed favourably in relation to facilitating mutual learning (Joerges and Neyer Reference Joerges and Neyer1997). As suggested by Pollack (Reference Pollack2003: 213), comitology differs in function depending on the extent to which the values of the principles and the agents differ. Delegation can be more elaborate where more agreement exists, whereas control tends to be stronger and more based on deliberative arrangements in contested areas.
These developments show that comitology was not a trustworthy alternative when agencies were created in the 1990s. However, the subsequent reforms have turned comitology into a valued instrument within the EU's set of overlapping and complementary governance tools (Vos Reference Vos2009; CEC 2010). This dynamic perspective is important to assess the innovations introduced by an agency.
EASA's predecessor: JAA
Historically, aviation was a national affair with national carriers, national airports and national aviation industries – combined with bilateral aviation agreements about routes and landing slots. This changed drastically under the influence of liberalisation, open sky agreements with the US and price fighters (Staniland Reference Staniland2008, Kassim and Stevens Reference Kassim and Stevens2010). The ‘normalisation’ of aviation as a market called for a European regulatory framework. The first approach was based on a network of national aviation authorities (NAAs) and was superseded by EASA in 2003.
Aviation regulation can only be understood in its international regulatory context, which started with the International Civil Aviation Authority (ICAO, Kassim and Stevens Reference Kassim and Stevens2010). This UN agency works with panels (comparable to comitology) that report to its ‘councils’. National experts work with ICAO in finding agreement on technical standards. Similar to subsidiarity as the EU's organisational principle, ICAO operates on the basis of international agreed standards which the states have to implement. An increasing number of countries now cooperate in regional groups to pool resources and influence. The pan-European JAA was one of the first regional cooperation projects within ICAO and aimed at countering the US dominance. What started as cooperation between experts became more formalised in 1990 with the Cyprus Agreement whereby the countries agreed to respect the JAA standards and to work towards a high and consistent level of safety. The JAA was a pragmatic cooperative network for collective action in Europe while avoiding handing over powers to ‘Brussels’.
Based on ICAO standards, JAA formulated Joint Aviation Requirements (‘JARs’). The JARs harmonised the certification of planes and thus facilitated the emerging European aircraft industry. In 1987, the work extended to operations and maintenance. This was followed by the licensing of operators, crews and pilots (JAR-ops). This, however, proved difficult. Regulating flying time has a direct impact on social conditions and on the competitiveness of operators (Staniland Reference Staniland2008). Hence, JARs moved much more quickly than the JAR-ops (interviews). A difference that was still relevant at the time of writing (see below).
JAA worked on the basis of minimum standards and mutual recognition. This left ample room for the states to fit the JARs to national situations. In the meantime, the interaction between the experts helped to exchange best practices, created mutual understanding and offered a setting for moving towards common standards (interviews). The result was a gradual equilibration of national standards, at least as far as the standard setting for the technical parts was concerned (JARs – the hardware), but also the ‘Ops’ (the human side) started to progress. In today's terminology, being based on learning and open coordination, JAA would be called ‘new governance’ (Treib et al. Reference Treib, Bähr and Falkner2007).
The EU's competences started to develop around 1990. JAA standards became integrated into the EU through specialised legislation such as BR 3922/1991 (harmonising technical safety requirements and administrative procedures) and BR 2407/1992 (harmonising licensing of air carriers) (Kassim and Stevens Reference Kassim and Stevens2010). This also implied that a comitology committee had to be created operating alongside JAA. Where JAA was staffed with technicians from the NAAs, the comitology committee was chaired by the Commission and staffed with officials from the ministries. The EU, however, had great difficulty in keeping up with the JARs. As explained in the interviews, the national experts cooperating in JAA would adopt standards and requirements much more quickly than the EU could through comitology. The Member States gathered in comitology – represented by officials instead of NAA experts – were biased towards national perspectives and had difficulties accepting binding EU legislation. The negotiations were much more political within the EU context compared to the expert-driven JAA.
This shows that JAA was not much liked politically. To the dismay of the ministries, which wanted more control, experts from national authorities were striking European agreements. The experts were more concerned with safety exclusively and presented civil servants from the ministries as “third rate” experts. The ministries were concerned about the technicians ignoring broader political interests such as social, environmental and regional planning consequences of aviation. What emerged was a sense of urgency among ministries to change JAA.
The time gap between JARs and EU-jars virtually incapacitated the Commission. As appeared from interviews, while JAA was developing JARs 12 – version 12 – the EU would still use JARs 3. The Commission was unlikely to go to the European Court of Justice with outdated EU legislation (compared to the JARs). Apart from the legal limitations, JAA was also criticised for lacking a legal appeal mechanism for industry. However, as discussed below, this proved not much of a problem.
The development of the Airbus 380 had cast its shadow. Those interviewed took different views on its certification (Schout Reference Schout2008). Some emphasised the costs of getting all the parts through 27 different procedures and considered a centralised system (EASA) unavoidable. Others downplayed this by pointing to other planes that had been certified by JAA (including the Airbus 320, 330 and 340) and argued that the costs should not be overstated. Certification was handled by teams and, once approved, Member States would acknowledge them by way of routine. Despite demands for a ‘one-stop authority’ for certificates, little in terms of analysis actually showed that JAA was underperforming. As emerged from the interviews, the intellectual framing of problems and solutions following liberalisation of aviation was rather limited (Schout, forthcoming). Yet, the initial preference in Member States and the Commission tended heavily towards centralisation.
Accountability mechanisms
Starting with administrative control mechanisms (Table 3), JAA was managed by the network of directors from the national administrations (European Civil Aviation Conference). This network set the strategic agenda and discussed major differences and problems. The elaboration of the work programme was in the hands of teams of experts from national authorities. Once the programme was announced – and, in later days, put on the JAA website – the experts would discuss the new standards. They would also involve the USA Federal Aviation Authority in their discussions in order to ensure the international context of legislation. Industry was also consulted in the elaboration of new regulations. JAA consisted of a core of 20 national senior experts who developed the work programme in their areas (related to structural design, safety, use of materials and other key issues in the design of planes). They assigned pairs of experts from the same field but from different states to lead the work on certification requirements in smaller teams. The full core group monitored the development of these teams.
Accountability analysis (JAA)
Accountability analysis (EASA)
In the latter days, influenced by better regulation, impact assessments were applied to new standards. Agreements on standards were published, followed by a consultation phase of usually 3 months to allow national administrations and industry to examine the proposed measure. The feedback would be reviewed by the experts before finalising the proposal. If a measure involved changes in legislation, it would go through the national parliaments. Codes or guidelines would be implemented through the normal ministerial arrangements (hierarchical legitimacy).
The JAA standards would have to be implemented by the NAAs and related bodies (e.g. maintenance organisations). Satisfactory implementation would grant organisations mutual recognition status (legal legitimacy). In case of major deficiencies, Member States had to elaborate follow-up measures and were monitored by the JAA teams. Organisations would lose the mutual recognition in case of repeated failure. One of JAA's alleged weaknesses was the absence of binding legal implementation and of appeal mechanisms. However, there were economic consequences (withdrawing mutual recognition), and JAA's successor (the Commission supported by EASA) is also keen to avoid legal actions and focuses on learning instead (Groenleer et al. Reference Groenleer, Kaeding and Versluis2010). Moreover, the absence of a formal appeal mechanism was compensated for by the consultations in which industry was closely involved. This situation has been reversed by EASA which has an appeal mechanism but less intense – but more general – consultations with industry (see below).
In terms of functional legitimacy, the experts developed common objectives throughout the years, gravitating towards high and uniform levels of safety. They were truly concerned with safety and in interviews prided themselves with “having grease on their hands” and “putting safety above everything else”. They were usually located at airports operating at arms’ length from the ministries. The standards were formulated on the basis of problem solving and best practices and not on the basis of negotiations. This was, as stressed in interviews, not a legal process. These discussions helped to create common understandings of intentions while leaving leeway to implement the standards in the national contexts and allowing for gradual convergence (cf. Joerges and Neyer Reference Joerges and Neyer1997). Moreover, the experts interviewed underlined the flexibility in setting priorities and in responding to technological changes.
In terms of output legitimacy, JAA helped to harmonise aviation regulation by providing a basis of thousands of standards and procedures for products and services. The JARs were more or less literally taken over by EASA and turned into EU standards and administrative procedures. The effectiveness of the consultations with industry and with the extensive pan-European network is also demonstrated by the fact that few letters of complaints followed when the measures were issued. In interviews, experts estimated that, on average, there would be about 10–20 complaints. The profound expert discussions apparently resulted in broad support. JAA is usually associated with the limitations of mutual recognition and voluntary cooperation, but it can equally be regarded as a case of successful open coordination. As such, this network followed the classical pattern of an OMC preparing the grounds for EU legislation (cf. Lascoumes and Le Galès Reference Lascoumes and Le Galès2007). The positive assessment of JAA's input and output legitimacy implies that the criticism regarding JAA was largely unjustified. It seems there was an element of groupthink involved among officials and mistaken hopes of ‘getting governments out’ among industry.
EASA
As discussed by Kassim and Stevens (Reference Kassim and Stevens2010 chapter 7), the choice for an agency remained far from self-evident for quite some time. Some within the Commission held strong reservations towards agencies. However, interviewees explained that the Director-General for Transport, Lamoureux, was a supporter of the idea of independent authorities. His preference was explained in an interview as a recognition that expansion of DG Transport itself was not an option.
The term ‘regulatory’ EU agency is often used even where regulatory powers are weak (Groenleer Reference Groenleer2009). EASA does have some regulatory powers. The director can take “completely independent” decisions (Regulation 216/2008 Art.38). Yet, the agency has been in a position of considerable uncertainty about the extent to which it can perform its tasks by itself. Plans to reconsider the involvement of the Commission are more or less constantly on the table due to the political nature of aviation regulation and because industry lobbies the Commission on political issues (interviews). The initial plans followed a rather centralised approach and put Member States at a distance by also avoiding comitology (CEC 2000). According to interviewees, a “fossilisation” of national influence on aviation had to be prevented. The hope to keep governments at a distance pushed towards an agency solution rather than giving the Commission new powers or creating an intergovernmental body like Eurocontrol. According to interviews, national governments should not be replaced by a “European government”. The agency seemed to offer a new approach based on expertise instead of on politics.
Member States quickly started to doubt their support for a centralised approached and consistently aimed at new involvements in aviation regulation (Schout Reference Schout2008). Board decisions also tied the hands of EASA considerably. For example, the board decided that EASA should hire experts from the NAAs instead of commercial experts. This decision caused substantial losses given that the board also agreed high fees for experts to support income for the NAAs but kept prices for certificates low for industrial policy reasons. Moreover, with its expanding tasks, accumulating losses and inability to hire sufficient experts, EASA became seriously overloaded. As a result, it remained in many ways a subsidiarity-based network organisation with close ties to the NAAs.
In terms of rule making, EASA has to support the development of a high and uniform level of safety and of environmental standards in aviation. These tasks have been elaborated in several phases, developing from managing the certification of design, production and maintenance of planes, to air traffic management and the operations of airports (EC1108/2009). EASA mainly ‘assists’ the Commission in the preparation of measures (Art. 19). The requirements normally go, via the Commission, to the EASA committee. The Commission cannot change the opinions of EASA without prior coordination with EASA (Art. 17.1.b). This is a modification of comitology because it imposes restrictions on the Commission. Respecting the Meroni principle, standards and recommended practices are presented as “semi law making powers” or “soft law” without general applicability. Yet, owners of certificates have to meet the standards. Although the agency formally ‘assists’ the Commission, standards and recommended practices are set by the executive director and apply to all owners of a specific type of plane. The withdrawal of a certificate can mean that all owners of that plane are hit. Hence, EASA's soft-law can be “quite hard” (as expressed in interviews) and stretches the limits of ‘Meroni’.
Accountability mechanisms
The background of EASA's activities is its work plan. As EASA is part of the Commission system, the planning of budgets and programmes is formalised also for agencies and reflects the detailed administrative controls including consultation (CEC 2009). The programme is developed in consultation with advisory bodies (the Advisory Group of National Authorities and a committee which includes industry). The draft programme is sent to the board, EP, Council, Commission and the Member States. Art 56 (BR 216/2008) specifies that it has to be presented according to the Commission's activity-based management style to ensure a feasible workload in relation to resources and to link EASA's activities to the Commission's priorities. The fact that the Commission's comments will be attached to the programme underlines the initiating powers of EASA. The board adopts the programme and uses it to evaluate the performance of EASA in its annual report.
Once the programme is agreed, rule making starts with a preliminary impact assessment examining the need and options for new measures. A team headed by an EASA expert, and usually including national experts, will produce a ‘notice of proposed amendment’ (the draft measure). These can be accompanied by an impact assessment in accordance with the Commission's guidelines, and this can be checked by the Commission's Impact Assessment Board. Subsequently, a consultation period of 3 months allows persons and organisations to give their comments. In case industry has major difficulties with the measures, they can turn to EASA's board of appeal. So far this has never happened (early 2010). As regards implementation, Member States, contrary to JAA, are now forced by law to adhere to the EU standards and can be taken to Court; however, the Commission is very reluctant to do so (Groenleer et al. Reference Groenleer, Kaeding and Versluis2010).
In terms of peer group involvement, EASA has developed into a network. From a subsidiarity perspective and in view of the fact that major airports need aviation authorities too, it is doubtful whether EASA could work without the NAAs (Schout Reference Schout2008). Airports and Member States need expertise for their regulatory and inspection tasks. Moreover, EASA relies on the expertise of the NAAs (as did its predecessor).
As regards output legitimacy, EASA's position lacked clarity at the start, it had major financial difficulties, and there were clashes between Member States and EASA about the position of EASA. Moreover, EASA suffered from overambitious objectives, and staff proved difficult to hire because of salary limitations and obligations to work with limited contracts (Schout and Pereyra Reference Schout and Pereyra2011). The criticism regarding EASA was summarised in, among others, the Report from the House of Commons (2006) which concluded that EASA threatened aviation safety. To force some realism into its planning, the Commission stated that EASA had to become more professional and only accept tasks that it could deliver (COM 2009: 322). Given these start-up problems combined with EASA's ambitions to grow, there has been increasing annoyance over the behaviour of EASA bordering on arrogance and leading to an atmosphere of ‘EASA against the Member States’ (Groenleer et al. Reference Groenleer, Kaeding and Versluis2010).
This underlines that, despite the enthusiasm to move away from JAA, it was unclear what EASA should actually look like. The interviews indicated that it was unclear from the regulation who precisely was responsible for what. Failing to consider subsidiarity, the first EASA evaluation (Horvath 2007: 38) noted that the distribution of tasks between the Commission, EASA and the NAAs “creates complexity” and needed to be “centralised”. Given the distribution of expertise, the Commission is responsible for regulatory tasks while many responsibilities are in fact to a large extent carried out by EASA. Moreover, in its operations, EASA has to work with NAAs due to shortages in staff and expertise. Hence, instead of clarity, EASA caused confusion as regards responsibilities.
Although it seems to have found a productive modus operandi in monitoring implementation (Groenleer et al. Reference Groenleer, Kaeding and Versluis2010), EASA's role in rule making has remained contested. Compared to the JAA system, the greater independence of EASA in formulating measures leads to much more feedback letters (criticism) from its clients, sometimes in the thousands (according to interviewees). The interviews have also shown that in the technical complex aviation sector with several backup mechanisms, the agency tends towards formalisation at the expense of the flexibility that the sector needs. Interviewees and Horvath (2007, p. 42) speak of “over-engineering” rules and complain about EASA's risk-avoiding and legalistic attitude (“losing sight of the practicalities of flying”). The increase in feedback on proposed measures shows at least that its output legitimacy is not beyond doubt and implementation problems have remained. Similarly, Kassim and Stevens (Reference Kassim and Stevens2010: 153) are not so sure whether EASA has overcome the difficulties as they existed.
Conclusions
Agencies have mushroomed in national and EU administrative systems. One of the pending issues of agencification is the question concerning their added value. The legitimacy framework (table 1) makes it possible to define and assess the contributions of an agency and, hence, to compare it to alternative policy instruments. This study offers an analysis of the merit of an agency (EASA) compared to its preceding network (JAA) and to comitology. The literature on EASA is positive about its contributions but pays little attention to other instruments. The legitimacy-based model allows a dynamic comparison of the career of EASA and its hypothetical alternative.
This approach indicates that the perception of JAA as underperformer is unjustified. Allegedly, the network was unsuitable for the liberalising EU market. The agency instrument was supposed to change decision-making from an opaque and ineffective club-type OMC (JAA) towards expert decisions with a one-stop shop for certificates without having to rely on the much-criticised comitology. Our analysis indicates however that output legitimacy was at a higher level than perceived. It was efficient and explicitly subsidiarity-based. JAA was to some extent victim of civil servants in the ministries who wanted more control over national experts and of impressions among industry that governments could be put at a distance by creating an agency. As regards input control, JAA seemed to be well equipped in terms of administrative and hierarchical control.
The second conclusion is that comitology has moved from closed negotiations towards more transparency and stronger democratic control. Moreover, comitology has proven to be effective in terms of output (including observance of subsidiarity). It is not just a form of hierarchical rule making but also of deliberative governance. From a legitimacy perspective, the move away from comitology towards agencies was as far as EASA is concerned less necessary due to these comitology reforms. Importantly, the analysis shows that any discussion on advantages of an instrument has to take into account the dynamics in alternative instruments. As a corollary, expanding the use of comitology could have been explored more seriously as an alternative to the agency route.
Thirdly, the application of the model suggests that the agency instrument has not produced the major change in regulatory governance that was hoped for around 2000. It has not resulted in less involvement of national governments because the networks have been continued and Member States are still prominently involved through the board. As regards its regulatory role, the output legitimacy of the agency is to a large extent based on the successes of the JAA network. In terms of input legitimacy, there are signs of over-control, such as micromanagement by the agency board concerning fees and the obligation to use experts from national authorities. Hence, as regards the initial hope that agencies would increase expert input in decisions, this has only happened to a limited extent. EASA has decision-making powers in some areas (where it replaces comitology), it binds the hands of the European Commission because the latter cannot simply overrule EASA's advice, and it shifted the right of initiative away from the Commission (by proposing opinions). This limits the extent to which the creation of the agency implied an institutional innovation. It mainly assists the Commission and its comitology system. Paradoxically, the conclusion is that the agency in fact shifted regulation towards a more legalistic and risk-avoiding attitude whereas the JAA network was more oriented towards depoliticised decisions and expert deliberations. Hence, EASA is quite the opposite of what proponents of EU agency instrument hoped for.
Given that this is an N = 1 study, the generalisation concerns primarily the use of the legitimacy-based approach. Importantly, the analysis shows that the framework can be used to compare and assess instruments in terms of administrative accountability, performance, and hierarchical and legal control. It offers the much-searched-for comparison between instruments and allows for a dynamic perspective. The application of the model leads to a different take on EASA compared to the more descriptive assessments (Horvath 2007; Pierre and Peters Reference Pierre and Peters2009). Judged in isolation, evaluators see progress and attribute this to the agency. However, the comparative approach indicates that progress in this field could also have been achieved through other instruments.
EASA is not an ideal type agency and the realities of comitology committees is more differentiated than could be discussed here (Brandsma Reference Brandsma2010). Yet, this comparison of JAA and EASA and a hypothetical alternative instrument (comitology), shows that each of these instruments has been upgraded overtime under the influence of better regulation agendas. They have become more alike in terms of the legitimacy framework.
Following Dukes (Reference Dukes1965), N = 1 studies can also be used to generalize beyond the specific case. This analysis underlines that one has to be careful claiming that one instrument is better than others. The better regulation influence suggests that EU instruments have been improved generally and, hence, that the agency instrument has been less of an innovation than sometimes claimed. The analysis also shows that policy makers were not aware of the strengths and weaknesses of different instruments nor of reforms of instruments. As the case shows, assuming rational behaviour of decision makers and functional necessity seems dangerous for understanding the development of an instrument.
