The Legitimation Practices of Science and Technology-based Agencies: Isomorphism across EFSA, EMO and EPO?

The Legitimation Practices of Science and Technology-based Agencies: Isomorphism across EFSA, EMO and EPO? Susana Borrás, Copenhagen Business School, Denmark sb.dbp@cbs.dk and Janus Hansen, University of Copenhagen, Denmark JH@soc.ku.dk This version: August 13th, 2015 WORK IN PROGRESS. PLEASE DO NOT CITE OR CIRCULATE WITHOUT PERMISSION ABSTRACT Science and technology-based agencies are increasingly subject to questions of social acceptance. Therefore, they engage in legitimation practices, which we define as the agencies activities seeking to ensure their legitimacy in the society. Sociological institutionalism suggests that as their formal mandate and overall challenges are comparable, we might expect agencies to develop similar practices through isomorphic processes. This paper studies the legitimation practices of three European agencies, EFSA, EMA and EPO, examining variation by means of a two-dimensional matrix distinguishing the agencies response to 1) legitimacy claims (cognitive competence, moral appropriateness, pragmatic problem-solving) and 2) legitimacy sources (input-throughput-output). The findings show limited isomorphism as the agencies legitimation practices are directed at different constituencies and consequently manifest themselves differently across our analytical matrix. This suggests revisiting some of the assumptions of isomorphism theory, taking into consideration the initial conditions and the contextual factors of these agencies. 1

1. Introduction The EU and its member states increasingly rely on independent agencies with high levels of scientific and technical expertise to perform regulatory functions. Such agencies are usually mandated to operate at arm s length from ordinary politics in order to insulate substantive expertise from political pressure and economic interests. This agentification of regulatory functions is however not exempt from discussions about the democratic deficit (Buess 2014). The operations of regulatory agencies have increasingly become the object of attention and sometimes contestation by various stakeholders and publics. Much of the research literature on independent agencies has focused on the properties of such agencies in terms of their mandate and delegated powers (Majone 2001), as well as normative questions regarding how institutional design can ensure adherence to principles of democracy and good governance (Maggetti 2010). Less attention, however, has been devoted to how independent agencies actually operate in practice. Questions about how organizations seek to establish, maintain or regain legitimacy, both in their everyday operations and in times of crisis have not been the main approach in the study of regulatory and independent science-technology agencies (see next section). Whereas some single case studies have recently provided some insights into the organization and operation of particular agencies, there is still limited comparative analysis examining the actual practice across agencies. Starting from the sociological tradition concerned with the actual practice of organisations, this paper examines how selected European agencies develop legitimation practices ; practices deliberately aimed at ensuring the agencies social legitimacy. Questions of legitimacy are particularly intricate for agencies with a strong science and technology component in their remit. Based on science-technical expertise, these agencies operate primarily with reference to non-majoritarian and non-egalitarian principles of scientific knowledge production. Yet, they are often times subject to divergent societal views and expectations, which reach beyond the realm of scientific knowledge production. Such agencies thus serve as interfaces between science, politics and economic interests. They oftentimes need to balance expectations among stakeholders about their political independence, their scientific and technical competences, and their neutrality in the face of value disputes within their areas of responsibility (typically but not always related to risk) (Rothstein, Borraz et al. 2013). This paper seeks to open up the legitimation practices of such agencies to sociological analysis. Empirically, we focus on three European cases: the European Food Safety Authority (EFSA), the European Medicines Agency (EMA) and the European Patent Office (EPO). These agencies have comparable formal mandates and throughout their lifetime they have all been repeatedly exposed to public controversies regarding both their general mode of operations as well as specific regulatory/advisory decision-making. This has likely influenced how they seek to legitimize their operations. We thus want to examine what legitimation practices they have developed through time; how do they operate in practice when seeking to establish, maintain and regain legitimacy among stakeholders and the general public. The theory of sociological institutionalism suggests that, because their mandate is comparable, these agencies are likely to develop similar practices through isomorphic processes. We will examine empirically whether this is in fact the case. Hence, the main research question of this paper is, what specific patterns of legitimation practices have European science-technology-based agencies developed? More concretely, and following from previous 2

theoretical perspectives, the paper enquires whether we can identify isomorphic patterns of legitimation practices across the agencies in terms of the legitimacy claims and legitimacy sources that the agencies seek to rely on. In doing so, the paper aims at making three contributions to the literature. Firstly, it combines the organizational sociology approach with the political science and legal studies approach on the legitimacy of regulatory/advisory independent agencies. Secondly, the paper develops the notion legitimation practice which is operationalized in terms of two analytical dimensions (legitimacy claims and legitimacy sources), which is a novel combination of analytical approaches to legitimacy and legitimation from the theoretical approaches mentioned above. Third, and last, this paper aims to contribute to the literature by providing a solid comparative empirical work that substantiates previous normative and sociological approaches into the actual analysis of legitimation practices. These comparative findings cast new light into the agencies responses and strategic decisions regarding how they interact with their complex socio-political environment. The paper proceeds in five steps. First we briefly review the existing literature concerning the legitimacy of science-based independent agencies, with particular focus on European agencies and on the difference between normative and empirical approaches. Here the notions of legitimacy and legitimation as well as theoretical assumptions about isomorphic organizational responses will be carefully spelled out. Based on this, the next section develops an analytical framework to guide our empirical analysis. It is based on to two analytical dimensions, namely, the agencies response to societal legitimacy claims (cognitive, moral, pragmatic) and the legitimacy sources they address (input, throughput, output). This framework is useful to capture complexity and variation in the agencies legitimation practices, and allows us to determine possible isomorphic traits between the agencies. The three sections that follow account one-by-one for our three case studies. The final section compares the three cases, summarizes the findings, answers the main research questions, and concludes with a discussion about the theoretical implications of the comparative findings. On that basis, the paper suggests some lines for further research. 2. On Legitimacy, Legitimation and Isomorphism of Independent Agencies Broadly speaking, legitimacy is at the core of the relationship between a political system and its citizens. It is however a rather elusive phenomenon to study. The social sciences literature has tended to examine legitimacy from two broad angles, a normative-institutional angle and a sociological-behavioural angle (Blatter 2007). Whereas the former is concerned with how far observations of the real world reflect predefined notions from normative theories (of democracy or legal authority), the latter examines the social behaviour regarding the degrees and forms of social support to a political system (Easton 1975). Previous efforts to combine both angles have pointed at their complementarity (Norris 1999); which is particularly obvious in the context of the European Union: The long-standing legitimacy crisis of this supra- and intergovernmental political system is at once linked to issues of normative ideals of democracy (rather than demoi-cracy) (Borrás and Radaelli 2014) as it is to issues of varying social support to the European integration project. The rapid growth of transnational governance (Eberlein, Abbott et al. 2014) and the creation of organisations that are not formally accountable to clearly identified principals, has raised questions about 3

the legitimacy of non-state international and global governance (Bernstein and Cashore 2007) (Bernstein 2011). The role of experts and professions in this transnational governance has also raised the attention of scholars, particularly related to questions of legitimacy (Quack 2010) (Brassett and Tsingou 2011). The organizational focus of these studies has brought forward another important distinction offered by the literature between legitimacy and legitimation. Whereas the former refers to the actual levels of public support at a given point in time, the later refers to actions and processes oriented towards gaining/obtaining/improving those levels of public support (Tallberg and Zürn Forthcoming) (Fransen 2012). This paper is concerned with the study of the organizational practices of these science-technology independent agencies. For this reason we focus on legitimation practices as the set of arrangements and activities that the independent agencies have put in place in order to secure public support of their operations. We base our discussion on the following definition of legitimacy from organizational sociology: Legitimacy is a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed systems of norms, values, beliefs, and definitions. (Suchman 1995) : 574. This definition points to two important features. Firstly, to speak of assumptions and perceptions points to the fact that legitimacy is a relational phenomenon. It is the product of a relationship, not a characteristic of the agency in itself. Secondly, generalized perceptions suggest that not all individual actions or singular decisions of regulatory agencies need to be agreeable to stakeholders for the agency to be considered legitimate. An agency can thus have a certain credibility, which can be maintained in the light of specific disappointments. However, if such disappointments are recurrent among stakeholders, support will eventually be withdrawn and can be withdrawn much faster that it can be built up. Following this we understand legitimation practices as the activities agencies engage in to ensure that their activities are indeed considered legitimate within the social contexts where they operate, understood as the quality of a relationship that must be maintained over time. It has become common in political science, following Easton (1965), to distinguish between input and output legitimacy. Input legitimacy refers to the opportunities of citizens to articulate their demands and requests to the political system (typically through representative mechanisms) and feel they are heard. Output legitimacy refers to the social support granted on the basis of the capacity of the political system to deliver on those demands and requests. This has sometimes been portrayed as an inescapable tension or dilemma between democracy and effectiveness (Scharpf 1999), other times as being mutually dependent and reinforcing through time in the governance of these areas (Borrás 2006). The concept of input legitimacy has been applied in recent studies on EU agencies, primarily to identify tensions between different normative ideals about citizens role in the definition of agencies activities, on the one hand, and between ideals and actual practices on the other. For instance, Klintman and Kronsell (2010) suggest there is a tensions between the normative ideals of inclusiveness and stakeholder participation put forward by deliberative and participatory theories of democracy, on the one hand, and the ideal of the agencies arms length and independence from vested interests that is enshrined in the agencies remit on the other. Their findings suggest that while the EFSA flags a self-proclaimed integrative mode of stakeholder engagement to facilitate genuine deliberation between different interests, the reality is much closer to an aggregative mode, where actors negotiate and bargain across the board. The authors conclude that EFSA often speaks in terms of integrative governance. (h)owever often falls back on 4

more aggregative forms due to the aggregative logic of the organization (ibid. p. 325). Dreyer and Renn s recent study confirms these previous findings. Their own work shows that EFSA s active policy towards the involvement of stakeholders since its creation has not managed to introduce a truly analytic-deliberative process in the EU s complex food safety governance (Dreyer and Renn 2014). In a recent comparative study of EFSA and EMA Groenleer adds an important temporal dimension to the question of input legitimacy. He examines how agency autonomy the degree to which regulatory agencies are in fact insulated from political processes and pressures depends not only on their formal mandate, but also on their ability to build alliances and legitimacy in the eyes of key stakeholders (Groenleer 2014). Groenleer suggests that the de facto autonomy of EU agencies is likely to be conditional upon acquisition of a level of (organizational) legitimacy. (ibid. p. 261). He thus highlights that input legitimacy among key stakeholders is essential to maintaining regulatory functions satisfactorily. This however, has proved to be difficult in topics of large scientific controversy such as GMOs, or xenotransplantation (Hansen and Allansdottir 2011). To our knowledge there are no studies that actively include output legitimacy performance of specific agencies. This is likely related to the difficulty of studying outputs in these areas, as it involves assessment of the regulatory decisions put forward by the agencies as well as their impact in their socio-political and economic context, which is a daunting task. The interest in input and output legitimacy of independent agencies has recently been supplemented with a growing interest in through-put ; attempts to look inside what is conventionally conceived as the black box of regulatory agencies. This interest comes from the understanding that the input-output model lacks an intermediary notion that captures procedural aspects, which are also very important both for normative theories of democracy and for sociological approaches to legitimacy (Schmidt 2013). The through-put dimensions is approached in different ways in the existing literature. The studies suggest that throughput refers to internal processes of administrative/organizational efficiency, as well as external accountability and transparency of the procedures (Schmidt 2013) (Chatzopoulou 2014). This refers to the social credibility of the internal/ administrative procedures or the agencies, which has been defined as well as procedural credibility (Borrás, Koutalakis et al. 2007). This procedural dimension calls in for a conceptual clarification on the relationship between legitimacy and accoutability. When studying the legitimacy of European agencies, normative-institutional approaches have tended to focus on issues of accountability. Accountability is largely viewed as a specific form of relationship between two (or more) organizations/actors, in which one gives account to the other, who has the formal authority to impose consequences to the former. Normative-institutional theories of democracy argue that accountability secures compliance and control of the agency, as it entails the oversight of the principal (European Parliament and Commission) on the delegated agency. Therefore the functions of democratic oversight and the institutional frameworks for the exercise of accountability are key for the democratic legitimacy of the delegated agencies (Font and Perez 2014) (Koop 2014). However, there is a widespread understanding among scholars that the legitimacy of independent agencies does not only depend on the accountability procedures to their principals. Legitimacy is a wider concept than accountability as it refers to social support rather than parliamentary oversight and principals control. As Black underlines: Although accountability relationships can be critical for legitimacy, legitimacy is not 5

necessarily always dependent on accountability relationships. Accountability and legitimacy are usually conflated in debates on regulation and governance, but analytically they are distinct (Black 2008) P. 149. As the three agencies we have selected for examination are comparable in terms of their formal mandates (see next section), their strong reliance on scientific expertise as well as their political environment, we expect that many of their legitimation practices to be somehow similar. This expectation is grounded in sociological institutionalism and the concept of isomorphism. Sociological institutionalism suggests that organisations with comparable mandates operating in the same political environment will develop similar legitimation practices due to isomorphism (DiMaggio and Powell 1983). Isomorphism is a drive towards homogeneity in organizational practices, which can either arise due to similar political framing condition (coercive isomorphism), through imitation of successful practices in other organisations (mimetic isomorphism) or through a logic of appropriateness, that is adhered to shared norms in a given organizational field (normative isomorphism). The concept of isomorphism does not suggest or presuppose that organisations become more effective through homogenization, rather the concept is conceived exactly to explain how and why organisations become more alike, when operating in the same field, due to a pressure to legitimize their operations independently of their actual performance. Institutional isomorphic processes can be expected to proceed in the absence of evidence that it increases internal organization efficiency Similarity can make it easier for organizations to transact with other organizations, to attract career-minded staff, to be acknowledged as legitimate and reputable, and to fit into administrative categories that define eligibility for public and private grants and contracts. (DiMaggio and Powell 1991, p. 73, quoted from Radaelli 2000, p. 28). As we examine independent regulatory agencies, we do not expect coercive isomorphism to be prevalent in regard to legitimation practices. The political principals presumably prefer a hands-off strategy in relation to how the agencies perform their duties, as the operational autonomy is a central raison d etre for such agencies. However, as agencies face significant uncertainties with regard to acceptance of their operations from their environment, we can expect processes of both mimetic and normative isomorphism to operate between them. [W]hen goals are ambiguous, or when the environment creates symbolic uncertainty, organizations may model themselves on other organizations (DiMaggio and Powell 1991, p. 69, quoted from Radaelli 2000, p. 28). Mimetic isomorphism may thus operate when the environment is generally uncertain, or it may be used strategically suggest that organisations are following best practices in a given field. Similarly, the strong reliance on expert knowledge would suggest that some degree of normative isomorphism is present, as the use of expertise has generally has been questioned in the last 2-3 decades, and all expert-based agencies are struggling to maintain their legitimacy. This is manifested, for instance, in growing demands for inclusion, participation and transparency of operations. In his discussion of isomorphism among EU institutions, Radaelli thus suggests that a variant of normative isomorphism can occur in those policy-making processes dominated by cognitive resources, expertise and even technocratic aspects of political life (Radaelli 2000) :29, which fits well with the type of agencies examined here. The concept of isomorphism as suggested by DiMaggio and Powell has been very influential as a research program in sociological institutionalism since it was first introduced in 1983. However, recently Beckert has suggested that the emphasis on isomorphism has been exaggerated in institutional theory (Beckert 2010). He suggests that the organizational mechanisms discussed by DiMaggio and Powell may equally well explain continue divergence among organizations, as isomorphic pressures are contingent on the specific 6

institutional environment in which organisations operate. In this context, we suggest that isomorphism is a productive hypothesis to ground a comparative inquiry on. However, we share some of Beckert s reservations about its uniform effects. In our analysis we thus treat the expectation of isomorphism not as a theoretical assumption, but as an empirical question to be examined. 3. Combining legitimacy claims and legitimacy sources A conceptual framework for analyzing legitimation practices In our empirical examination of legitimation practices we propose to combine an analysis what the agencies perceive to be important expectations/demands/requests by society with an examination of how they go about seeking to fulfill/address those expectations. For this purpose, we combine two sets of conceptual dimensions to form an analytical matrix. For the first dimension, the one that looks into what do the agencies perceive to be important societal demands, we rely on Julia Black (2008). She suggests that legitimacy, understood empirically as social acceptance by individuals, groups or organisations of the operations of a regulatory agency, can be grounded in three analytically distinct sets of societal claims. Firstly, the legitimacy claims by society can be based on cognitive grounds. For science-technology based organisations, cognitively based legitimacy this refers to the social perception of expertise and competence of the agency, and of the agency having sufficient command of the science and technology relevant to its domain of operation. 1 In this regard, the agency legitimation practices that seek to address these societal claims for technical competence and expertise will focus on mobilizing and actively using its technical expertise and competence basis. Secondly, legitimacy claims can be based on ideas and perceptions that the agency operates in ways that are morally appropriate in accordance with prevailing norms about democracy and legality (moral appropriateness legitimacy claim). When agencies are seeking to address those claims, they will put forward legitimation practices that enhance the correctness of the agency operations in terms of general norms like transparency, participation, openness, and the like. Thirdly, following Black, legitimacy can be based on pragmatic claims, which are those associated with the notion that the regulatory agency operates in ways that serves the interests of the individual or social group they were intended to (pragmatic legitimacy claim). In this regard, the agencies will develop legitimation practices that seek to address this claim by underlining and profiling the agency s overall role in the society and economy, and its benefits for specific groups. It is important to note that these three are not mutually exclusive. A particular legitimation practice developed by an agency may seek to address all three claims simultaneously. However, they are nonetheless analytically distinct. For the second dimension, the question how agencies seek to address the expectations/demands from society has to do with the way in which they address different legitimacy sources. For this purpose, we rely on the distinction between input, throughput and output legitimacy suggested by Vivian Schmidt (Schmidt 2012). Applied to our cases, input legitimacy pertains to the way agencies create opportunities for the articulation and uptake of demands and requests by those affected by regulation through various modes of 1 While these different claims to legitimacy are analytical categories, they are paralleled closely by sociological studies of how societal problem-framings direct particular expectations towards the governance of science and technology. 7

inclusion and consultation. Throughput pertains to the way the agencies organize their work internally. Regulatory agencies can no longer afford to behave as the black boxes originally entailed in the systems theoretical distinction between input and output legitimacy. For purposes of legitimacy, it is increasingly important that also internal operations are transparent and credible to the outside world. Finally, output pertains to the outcome of regulatory agencies activities, that they in fact produce the desired regulatory results (e.g. that food is safe for consumption, medicines effective and free of side effects, protection of patents robust etc). Outputs do not necessarily speak for themselves, and agencies may engage in practices to ensure that their performance is in fact considered satisfactory by the relevant social environment and recipients of their operations. For our particular purpose we can thus operationalize input as the channels of engagement with various types of stakeholders, throughput with the efforts to render internal operations in the agencies acceptable to observers in their environment, and output with the efforts to signal that their outcomes are satisfactory to their relevant social environment. When we cross-tabulate the above analytically distinct dimensions, we get the analytical matrix shown in Table 1. Table 1: Two Dimensions of the Agencies Legitimation Practices What -Practices addressing different social legitimacy claims How - Practices addressing different legitimacy sources Input legitimacy sources Throughput legitimacy sources Output legitimacy sources Cognitive competence claims Moral appropriateness claims Pragmatic goal achievement claims In the next sessions we report case studies on the three agencies structured along the dimensions outlined in Table 1. The case studies consist of qualitative mappings of the content and the form of the legitimation practices of these three agencies. The data is largely based on the publicly available and accessible information on the official webpages of the agencies. The data has been based on a systematic and careful examination of the content of the respective websites, identifying the legitimation practices according to the definition provided in section 3 of this paper. The legitimation practices of each agency are then mapped onto the cells of Table 1, which subsequently allows for an encompassing and organized comparison of the legitimation practices of the three agencies across each of the two analytical dimensions. The European Union has more than 40 agencies. Most of them have been created since the mid-1990s, when the EU integration process extended its functional operations into a large number of specific policy areas. There are different types of agencies in the European context. Some of them are executive agencies, 8

as they are in charge of implementing specific policy areas delegated directly from the executive powers of the European Commission. Another set of agencies have a diffusion, communication and monitoring purpose (like promoting and monitoring gender equality, fundamental rights, drug addiction etc). A third set of agencies are mainly dedicated to the exchange of operational information across member states in specific areas of collaboration (i.e. vocational training, disease prevention and control, etc). Other agencies are also engaged in mobilizing national and EU-level resources into direct activities at European level, for example on regulatory issues of the single market (i.e. energy regulators, electronic communication), or on important cross-european security issues like cross-border organized crime. Last but not least, other specific agencies like Euratom or EIT agencies, have a particular historical background and enjoy significant budgetary freedom. This paper studies the legitimization practices in three specific agencies: the European Food Safety Agency (EFSA), the European Medicine Agency (EMA) and the European Patent Office (EPO). The first two ones are EU agencies dedicated to provide regulatory advice (EFSA) and decisions (EMA) on single market issues of product authorization and safety. Formally speaking the EPO is an international agency, hence it does not belong to the list of formal EU agencies. However, the forthcoming creation of a Unitary Patent and the European Patent Court (approved in February 2013 and pending of ratification at the time of writing), will create a single patent system in the EU and will de facto make the EPO an EU agency. The selection of these three agencies as case studies is based on two key criteria. Firstly, these three agencies provide (quasiregulatory) advice (EFSA) and regulatory decisions (EMA and EPO) on the basis of highly specialized and cutting-edge scientific and technical knowledge. This feature makes them excellent cases for analyzing the legitimation practices, as it is particularly relevant to see the extent to which these agencies tend to focus or not their legitimation practices on this scientific-technical knowledge basis. Secondly, these three agencies have strong industry-user constituencies, as they all serve to authorize products (safety)/new knowledge (patent protection) into the single market. As such, these agencies are at the forefront of tensions between market development (industry interests) and broader societal issues (patient/consumer safety; or morality and limits of patent subject-matter). Last but not least, these three agencies have been hitting regularly the media headlines, with specific controversial issues like GMO approvals (EFSA), medicines/medical devices unexpected secondary effects/ defects (EMA), and issues of morality/ethical concerns on specific patents (i.e. patent on diagnostic test for a breast-cancer gene) and the associated public health problems of excessively costly treatments. For all these three reasons (scientific-technical knowledge, tension between different interests, and regular controversies since their creation), these three agencies constitute particularly good cases for studying the way in which they have addressed legitimacy issues, by way of putting up specific legitimation practices. As mentioned above, given these common features, we expect to find a certain level of isomorphism across those practices. 9

4. EFSA legitimation practices The European Food Safety Agency was created in 2002 in the context of a general revision of the European Union policy on food safety following the food safety scandals and heated debates in the late 1990s (Dioxine, BSE, Genetically Modified Organisms moratorium, etc). The General Food Law of 2002 defines the role of EFSA in terms of providing independent scientific advice regarding risks and scientific opinion on matters of food safety, animal feed safety, nutrition, animal health and welfare, plant protection and plant health. This means that EFSA s tasks are of risk assessment, issuing scientific opinions on these matters; whereas risk management decisions are taken by the Members States and the European Commission through the comitology procedures. EFSA was somehow born under fire, and this has naturally had an important effect on the way in which this agency has developed its remit and modes of operation and thus also its legitimation practices, which have been central to EFSA from the formation of the agency. In regard to input legitimation practices EFSA has developed a primary focus on two types of social legitimacy claims: cognitive competence and moral appropriateness. EFSA run a number of forums with a technical focus concerning the cognitive dimensions of its practices. EFSA has created a remarkable number of input legitimation practices with focus on scientific dimension. In these practices EFSA invites stakeholders to give their opinions and consultations on specific technical matters. These take different forms: from regular meetings of stakeholders with plenary scientific panels and EFSA working groups, open Scientific Colloquium series, to specifically focused consultation meetings EFSA meets stakeholders. Recently, this has expanded into forward-looking practices dealing with issues such as emerging risks and food chemical occurrence data. Particularly relevant, though are the scientific public consultations that EFSA regularly calls upon on scientific subjects specific to its remit. In these public consultations, interested parties can submit written scientific comments on specific substances and their active components. This is an important input legitimation practice based on cognitive competence. Other types of input legitimation practices are more oriented towards claims of moral appropriateness. EFSA has established several permanent consultative forums, the most relevant of which is the Stakeholder Consultative Platform, created in 2005 and formed by a wide range of stakeholders from 24 different organisations operating at different stations of the food chain. National sources of input legitimacy are important in a supra- and inter-national agency like EFSA, for this reason the Advisory Forum, formed by national risk assessment representatives plays an important role. The Forum has recently put forward a strategy for networking among EFSA and member states for sharing scientific information, pooling resources and coordinating work programs. As far as we have been able to establish, EFSA does not have any particular consultative forums for the business community. 10

Because EFSA was created under particular circumstances (food scandals and controversies in the late 1990s early 2000s) and its rulings and recommendations are regularly publicly contested the agency has put a lot of emphasis on its throughput legitimation, and it has done that in different ways Focusing on cognitive competence dimensions, EFSA has created a series of throughput legitimation practices on scientific aspects. The most relevant practice here is the EFSA Scientific Committee publication of a guidance document regarding procedural issues to ensure transparency of EFSA scientific outputs. This guidance deals with key scientific procedural issues, like the identification of data sources, the criteria for inclusion/exclusion of data, the confidentiality of data, and how to address and make explicit the scientific assumptions and uncertainties. Another scientific/procedural practice is the Guidance on Expert Knowledge Elicitation, which defines the method for selecting experts (from industry, national agencies, scientific community, others), and for obtaining and analyze their knowledge in a reliable, unbiased and reproducible way. From the perspective of throughput legitimation, EFSA has defined a set of Openness, transparency and confidentiality principles, a Code of good administrative behavior and a Decision on access to documents. These are initiatives and principles focusing on the moral appropriateness of administrative and organizational practices. That the internal procedures conform to codes of good governance and are transparently shown to do so, does not guarantee the strong regulatory performance that produce output legitimation. Consequently, EFSA also engages in intensive outreach activities to document its substantive performance. This is seen in a series of practices oriented towards cognitive and scientific dimensions. In this regard EFSA publishes a wide range of scientific publications on specific scientific issues, for example the publication of a large collection of data and reports on a specific scientific area in its remit; reports explaining the safety of specific substances, or the more encompassing scientific EFSA journal, or the more dissemination-oriented In Focus publication. Likewise, other periodic outreach publications and activities have a broader audience, targeting non-scientific stakeholders. This include activities and publications in the form of EFSA Press center, and stakeholder news, as well as reports about processes on science communication. Last but not least, output-oriented legitimation practices focusing on achievement of goals include the EFSA annual report, as well as publications and outreach activities on quality management goals and action plans, EFSA budget and financial statements, as well as international cooperation goals and action plans. The mapping of EFSA legitimation practices above suggests that EFSA seeks to achieve legitimacy by ensuring that stakeholders can get involved in some activities of the agency, and that they can understand its throughput mechanisms, whether these are related to technical assessments or consultations with various societal actors. EFSA obviously devotes significant resources to solicit scientific knowledge and expert assessment on food-safety, but it cannot rely on its recommendations and rulings being accepted simply because they are from EFSA. Therefore, the agency puts a lot of emphasis on documenting the principles underlying the composition of its scientific advisory boards, its selection of scientific experts and their declaration of conflicting interests. One important milestone in the legitimation practices of the agency is the recent initiative transformation to an Open EFSA. With the recent publication in March 2015 of an implementation plan, the agency has identified a series of measures to be put in place or reinforced in order to strengthen EFSAs transparency and openness. The action plan 2 is a well-structured approach of the agency s own activities, and a way of defining a strategic line of action to improve it. 2 EFSA: Preliminary Implementation Plan Transformation to an Open EFSA March 27th, 2015. 11

Table 2: EFSA Legitimation Practices What - Social legitimacy claims Cognitive competence claims Moral appropriateness claims Pragmatic goal achievement claims How - Legitimacy sources Input: Stakeholder meetings with plenary scientific panels and EFSA working groups EFSA Scientific Colloquium Series Specifically focused consultation meetings EFSA meets stakeholders Scientific Public Consultations Advisory Forum and Scientific cooperation at national-efsa level Stakeholder Consultative Group on Emerging Risks Stakeholder Consultative Group on Food Chemical Occurrence Data Stakeholder Consultative Platform Regular publication Moving Together Stakeholder Consultative Group on process transparency and information EFSA Stakeholder conference on transparency Throughput: Guidance of the Scienfic Committeee on Transparency in the Scientfic Aspects of Risk Assessment Guidance on Expert Knowledge Elicitation Rules defining the workflow of Scientific Opinions Openness, Transparency and confidentiality principles Code of good administrative behavior Decision on access to documents Transparency and confidentiality report Output: EFSA Scientific Journal Corporate publications on specific scientific issues: collection of data and reports on specific scientific area; explanation of safety of specific substances, etc. In Focus: Animals In Focus: Food In Focus: Plants Corporate publications on risk communication, discussions about Open EFSA EFSA Press Center Stakeholder News Annual report Corporate Publications on quality management plans, budgets, cooperation goals. 5. EMA legitimation practices The European Medicines Agency started its operations in 1995, when an internal market for medicines and medical devices was created in the European Union. The creation of the internal market was accompanied 12

by a regulatory reform, which introduced a centralized procedure for the authorization for pharmaceutical products in the European Union. This is the so-called centralized procedure, managed by EMA, which is compulsory for some products; and is different from the decentralized procedure of mutual recognition by which nationally approved products are accepted by other national agencies. Among the chief aims of EMA is to enhance the scientific capacity underpinning pharmaceutical authorizations and prevent regulatory agency-shopping in the European single market. Similar to EFSA, following this reform EMA is responsible of the scientific assessment, and the European Commission and Member States are responsible of the actual approval and management aspects of medicine marketing authorization. Along with providing scientific expertise for authorizations, EMA is thus also charged with the task of balancing rigorous oversight of pharmaceuticals with the need to ensure a market-supporting regulatory system, which ensures that effective medicines are made available to patients without unnecessary delays. In terms of input legitimation, EMA is rather explicit about the different stakeholder involvement in its operations. Firstly, it maintains a close dialogue with the relevant scientific communities through its six Human Scientific Committees (CAT, CHMP, COMP, HMCP, PDCO and PRAC). These committees hold regular meetings with active researchers within their domains, as well as creating research-oriented networks in specific scientific areas, for example the European Network of Pediatric Research at the European Medicines Agency (Enpr-EMA), with focus on input from the interaction with the pediatrics research community. Furthermore, the agency launches regular public consultations on specific medicines or active components, to gather external advice and scientific inputs. Secondly, EMA has established several forums to solicit inputs to its operations from civil society actors, most noticeably patient organisations in the Patients and Consumers Working Party (PCWP) as well as the Working Party with Healthcare Professionals Organisations, which were both created in 2013. Thirdly, EMA has been more careful regarding the role of industry, as industry representatives are not members of scientific committees or other structures within the agency. However, EMA has organized or participate in several forums for dialogue with industry actors on a wider basis. Likewise, EMA has created the European Network of Pediatric Research at the European Medicines Agency (Enpr-EMA). Taken all these altogether, EMA has become increasingly focused on creating channels for input-legitimacy on different grounds and claims (cognitive, normative, pragmatic), thus making itself sensitive to demands and wishes from different kinds of stakeholders, putting emphasis on the scientific community, as well as patients and consumers. Most recently (December 2014) EMA has revised its overall framework on the interaction between the European Medicines Agency and patients and consumers, defining more clearly the role of patients in scientific committees. In terms of throughput legitimation, EMA has also gradually adjusted its internal working procedures to be more transparent. A key challenge for EMA is rooted in the cross-pressure between promoting transparency in its assessments and recommendations to maintain public confidence that it is unbiased, on the one hand, and the pharmaceutical industry s interests in non-disclosure of confidential material included in the applications. At times this balance has been quite difficult to strike. At its formation in 1995 one of the first principled legitimation practices was the creation of regular European public assessment report (EPAR), which describes the scientific basis for the opinions on medicines. However, as public debates about transparency grew during the late 1990s and early 2000s, these reports were increasingly considered too generic. Following the general policy of the EU to become more transparent, EMA management thus passed the internal Rules for the implementation of Regulation (EC) no 1049/2001 on 13

access to EMA documents in 2006 as well as standard operation produces (SOPs) and work instructions (WINs) on how to handle requests to access documents. This had been followed in 2013 by Principles for publication of agendas and minutes of EMA scientific committees, which means that all minutes are made publicly available. In terms of output legitimation, the primary purpose of EMA is to ensure that pharmaceuticals that are marketed are both effective and free of unacceptable side-effects. EMA has launched a number of initiatives to signal this message to the surrounding society. This includes, for instance, the outreach activities directed at patient groups and health care professionals. EMA wishes to communicate that they do in fact have robust risk assessment procedures and are able to monitor pharmaceuticals in the postapproval phase, when products are released on the market. Particularly relevant in this regard is EMA s creation in 2006 of the European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP) for expert collaboration in post-authorization monitoring and post-authorisation studies on drug safety. Likewise, EMA participates actively in the European Technology Platform for Global Animal Health, an industry-led network promoting exchange of data and initiatives in the field of animal health. These two networks are important output oriented legitimation practices in relation to cognitive-based legitimation practices. Furthermore, EMA organizes or participates in multiple research conferences and scientific workshops to disseminate its results, and to obtain new knowledge. The training of patients and consumers is a relevant, and unusual practice of EMA, half way between input legitimation (empowering patient organisations in scientific advice processes within EMA), and output legitimation (focusing on the results of EMA regulatory decisions and its scientific basis). Last but not least, EMAs outreach activities through different publications and events to the general public are also important and visible forms of addressing claims for general moral appropriateness of the agency s activities. Table 3 EMA Legitimation Practices What - Social legitimacy claims Cognitive competence claims Moral appropriateness claims Pragmatic goal achievement claims How - Legitimacy sources Input: Scientific communities interaction with scientific committees Scientific public consultations European Network of Paediatric Research at the European Medicines Agency (Enpr-EMA) Patient and Consumers Scientific-advice meetings (SA Meetings) Patient and Consumers Scientific-advisory-group meetings (SAG Meetings) Patients' and Consumers' Working Party (PCWP) Revised framework for interaction between the European Medicines Agency and patients and consumers Enpr-EMA Involvement of young people and families in the Enpr-EMA networks Healthcare Professionals (HP) involvement Interaction with European HTA bodies Enpr-EMA Collaboration with micro, small and medium-sized enterprises (SMEs) 14

Throughput: European public assessment report (EPAR) Rules for the implementation of regulation 1049/2001 on access to EMA documents Standard Operation Procedures (SOPs) and Work Instructions (WINs) on how to handle requests to access documents Principles for publication of agendas and minutes of EMA scientific committees Output: European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP) European Technology Platform for Global Animal Health Research conferences and scientific workshops with experts and stakeholders Patients and Consumers Training and support Newsletters and other forms of broader societal dissemination of EMA s outputs 6. EPO legitimation practices The European Patent Office was established in 1973 by the European Patent Convention, of international nature. EPO grants (or not) patent rights to applicants (typically firms) after a process of examination conducted by in-house patent examiners. Technically speaking the EPO is the implementing agency of the European patent organization, currently with 40 Member States, which is not part of the European Union. However, the recent agreement on December 2012 to create a EU-based Unitary patent for 25 of the EU Member States and which will be granted by EPO, might turn EPO into a de facto regulatory agency of the European Union. Given its historical background and its international (rather than supranational) origin, the EPO has developed substantially less legitimation practices than the other two agencies examined above. In terms of input legitimation, the EPO considers itself a much less subject to address general legitimacy claims from the public or societal stakeholders that the two agencies examined above. There are two primary constituencies the EPO systematically consults with, patent professionals and the businesses it serves. This is reflected in the composition of The Standing Advisory Committee before the European Patent Office (SACEPO), which is formed by 10 representatives from industry (which are nominated by BUSINESS-EUROPE), 10 representatives of the patent profession (which are nominated by EPI), and by 10 legal experts in industrial property law appointed ad-personnam. The EPI is the Institute of professional representatives before the European Patent Office which was created in 1977 by the EPO itself, and it is ruled by EPO own regulation. This focus on industry and legal experts/ practitioners was expanded when in 2011 EPO created the Economic and Scientific Advisory Board (ESAB) formed by 11 members, some of which are scientists, and some of which are experts from industry and the patent profession. Last but not least, given its international (rather than supra-national) nature EPO is largely dependent on its member 15