A Road to Agency Accountability? Frontex and Fundamental Rights Fixing the chain of delegation through experimentalist participation

Transcription

1 A Road to Agency Accountability? Frontex and Fundamental Rights Fixing the chain of delegation through experimentalist participation LEILA GIANNETTO University of Trento DRAFT. PLEASE DO NOT CITE OR QUOTE. The recent elections of the European Parliament (EP) are an interesting challenge for the future direction of the European Union. Due to political unrest in Member States (MSs) triggered by the notorious financial, debt and economic crisis, populism and euroscepticism reached high levels of support; particularly, the Front National in France and the UKIP in the UK, but also other nationalist movements across Europe 1, are now the visible expression of the discontent with the management of both European and national politics by Europeans. Moreover, in the last twenty years, ever-decreasing voter turnout at elections both at national and European levels have demonstrated an increasing disaffection of European citizens towards politics and policy-making 2. Sabel (2004), the creator of experimentalist governance would ask: is this be a sign of an evolution of representative democracy? Is it necessary to reconsider the basic criteria of democratic legitimacy of both national and European governance systems? Centuries of philosophical and political studies have not determined where the quest for democracy should end, nor how exactly and incontrovertibly democracy can be evaluated. As a consequence, there has been a continuous redefinition of the concept and the essential features of democracy, especially when it comes to regional polities such as the EU. The implications of this claim are not necessarily negative: democracy is a notion that has been evolving since its inception and the different views that theorists and scholars have embraced over time are all fundamental for understanding this highly complex political and social phenomenon. Recalling Puchala s famous metaphor of the blind men trying to make sense of an elephant from different points of observation (1972) it is clear that the task of describing the elephant democracy in the EU is further complicated by the fact that during the last decade the animal has been moving rapidly, thus leaving the blind men ever more puzzled. 1 Among the most successful: the Danish People s Party, the Austrian Freedom Party and Alternative for Germany. 2 The 2014 European elections, probably thanks to the rise of ant-european parties, have seen this trend halting but not reversing. An analysis of EU voter turnout is presented in Section II of this paper. 1

2 There are a multiplicity of principles that can be argued to be relevant for determining what democracy is and how well it works in a defined polity: delegation (Majone, 1998; Moravcsik, 2002), representation (Piattoni, 2012; Bellamy and Kröeger, 2013), participation (Finke, 2007, Kohler-Koch, 2007), responsiveness, accountability (Curtin, 2006), effectiveness (Majone, 1998) are among the most significant democratic criteria. Bellamy and Castiglione effectively describe the dichotomy between two European integration theories (Piattoni, 2009): the ones focusing on output criteria of democracy, and those focusing on conventional democratic input (2011: 102). Participation, representation and delegation are conventional input democracy criteria due to their interest in the ex-ante legitimacy of policy-making, while responsiveness, accountability and effectiveness pertain to output democracy evaluation. In the constitutional perspective, clearly described by the German Federal Constitutional Court (BVferG) in the Lisbon judgement 3, delegation of power at the national level flows democratically from citizens to parliaments, from parliaments to governments and from governments to the bureaucratic apparatus (Strøm, 2000: 268); in a similar fashion the democratic legitimacy of the European Union derives from delegation of a set of powers from national governments to the European institutions (Bergman, Müller and Strøm, 2000). This view is shared by the liberal intergovernmental school (Moravcsik, 2002, 2004; Keohane, Macedo and Moravcsik, 2009) and challenged by advocates of multi-level governance (Piattoni, 2012); from another perspective, Agné (2007) questions the whole democratic theory of delegation, proposing a new definition of autonomy and of its delegation. Is delegation theory epistemologically able to democratically legitimate the European Union? And what problems of accountability does delegation theory raise? To answer these questions, this paper builds on the literature on delegation as a self-sustaining democratically legitimizing principle at the EU level thus excluding, for space constraints, the parallel evolution of national governance 4 and addresses the case of the growing informal and non-hierarchical governance system. Drawing from recent studies applying the principal-agent model of delegation to EU agencies 5 and to Frontex 6 specifically, democratic legitimation of the EU on the sole basis of 3 BVerfG, 2 BvE 2/08 vom , Absatz-Nr. (1-421). [Available at: ]. 4 For an application of the multi-level governance theory to the nation states, see Hooghe and Marks (2003). 5 EU agencies will be addressed in this paper also as non-majoritarian agencies meaning that they have no electoral basis and derive their powers legally from the Council and the EP even though at the European level the Commission and the ECJ are also non-majoritarian. In the literature independent regulatory agencies (IRAs) are defined as: bod[ies] with [their] own powers and responsibilities given under public law, that [are] organisationally separated from ministries and [are] neither directly elected nor managed by elected officials. (Thatcher 2002). At the 2

3 delegation will be challenged and an alternative solution will be put forward, stressing the need for an experimentalist participation. In order to address these issues the paper is structured in four sections. The first is concerned with an analysis of the literature on democratic legitimacy of the Union, deriving from a principal-agent model of delegation, and its detractors. The second acknowledges the difficulties in having a fully fledged democratic chain of delegation at the EU level and, therefore, an alternative theoretical solution for the accountability problems raised by delegation is proposed based on the experimentalist governance approach integrating the participation of organised civil society in the process. In the third section, an application of democratic accountability will be discussed by employing the principal-agent model and looking at the EU agencies, as the truly ultimate steps of delegation (Gilardi, 2001). In the last section Frontex democratic principals are identified and their accountability relationship with the agency described, in order to explore how the experimentalist participation of civil society can overcome Frontex accountability issues in the field of human rights protection. I. Democratic legitimation through delegation? Since the ratification of the Maastricht Treaty, a sense of entering an era in which MSs were delegating to the EU institutions more competences than they were able to control or more importantly in a democratic delegation perspective to retrieve started to creep, especially among scholars. This led to a widespread debate over the alleged EU democratic deficit (Moravcsik, 2002; Bellamy and Kroger, 2013). This section will enter the debate on the theory of delegation following the various steps of the chain of delegation, by analysing it first in a constitutional perspective, moving on to the liberal intergovernmentalist approach and, lastly, discussing it from a multilevel governance perspective, introducing another input democracy criterion: participation. Finally, the delegation chain will be turned around and the accountability mechanisms that render delegation democratic also from the output side will be described in the following section. European level there are over 40 agencies, divided by the EU in four categories: executive, decentralized, EURATOM agencies and the European Institute of Innovation and Technology. 6 European Agency for the Management of Operational Cooperation at the External Borders of the Member States of the European Union. 3

4 According to the German Federal Constitutional Court whose case law proved to be leading the constitutional discussion over the alleged democratic deficit of the EU since the Solange I case of 1974 the conferral of competences was possible and not contrary to the German Basic Law with Maastricht and it continued to be so also with the Lisbon Treaty. What the BVerfG in its jurisprudence on the matter has stressed is that democracy at the EU level has to be secured by national parliaments and that delegation of core decision-making competences should not be permitted, as listed in paragraph of the Lisbon judgement. Therefore, in a pure constitutional perspective, the democratic deficit arising from the excessive delegation of powers from national governments to the EU institutions is a non issue, as long as the conditions delineated by the BVerfG are met (Tomuschat, 2013). As Curtin describes it, in recent times a legal perspective has been that the Commissions [sic] policy making and executive functions are exercised in a delegated relationship of agency for the politically legitimated Member States. (Curtin, 2006: 88) Climbing up the chain of delegation, the Meroni doctrine of the European Court of Justice, in 1958, has assured compliance of the EU law to the legitimate delegation of powers from a delegating authority, namely EU institutions as the Commission and the Council, to other nonmajoritarian agencies. The same ECJ, in 2005, reaffirmed that the same three principles of the Meroni case have to be respected in case of delegation: first, the powers delegated cannot exceed the powers that have been conferred upon the delegating authority by the Treaties; second, the same conditions of exercise of the delegated power apply both to the delegating authority and the agency; and, third, the power delegated should be very specific and of executive nature. Liberal intergovernmentalism profits from the constitutionalist definition of legitimate delegation and seeks in it the foundations of its aversion towards EU democratic deficit theorists. Moravcsik, in his 2002 paper, argues that the criteria employed by democratic deficit analysts can be reduced to representation issues i.e. European Parliament as the only directly elected entity among un-elected and powerful European institutions (Commission, Council and the ECJ) and output legitimacy issues i.e. informal, opaque and technocratic policy-making. Liberal 7 Paragraph 252 of the Lisbon judgement of 30 June 2009 reads as follows: Particularly sensitive for the ability of a constitutional state to democratically shape itself are decisions on substantive and formal criminal law (1), on the disposition of the monopoly on the use of force by the police within the state and by the military towards the exterior (2), fundamental fiscal decisions on public revenue and public expenditure, the latter being particularly motivated, inter alia, by social policy considerations (3), decisions on the shaping of living conditions in a social state (4) and decisions of particular cultural importance, for example on family law, the school and education system and on dealing with religious communities (5). 4

5 intergovermentalism, instead, assumes that Member States decide to ratify international agreements and to pool their sovereignty in international institutions acting upon their interests formed democratically within the national polity and delegate their powers democratically by virtue of the same democratic chain of delegation that is in place at the national level. From voters to European institutions, the delegation of powers is assumed to be flawless, also and precisely because the delegated powers are mainly relegated to domains that are regulatory in nature (Majone, 1998) or are perceived as non salient for the national voters (Moravcsik, 2002, 2004); moreover, EU delegated powers should not intrude the core policies of Member States, the ones listed in article 79(3) of the German Basic Law (the so-called eternity clause) and the ones that require high levels of tax collection in order to be pursued, namely defence, social welfare 8, education and culture (Moravcsik, 2002:608). It must be acknowledged that delegation is not the only criterion against which democratic legitimacy of the EU is assessed in Moravcsik s perspective. The presence of checks and balances in the EU system of separation of powers, the constraints posed by Member States (fiscally more prepared to implement policies), the direct accountability of the EU through the European Parliament and the indirect accountability of the Council, these are all factors that concur to the overall high level of democratic legitimacy of the Union (Moravcsik, 2002). He goes even further by affirming that the possibility for Member States to act on a broader democratic stage has the effect of even enhancing their domestic democratic life (Keohane, Macedo and Moravcsik, 2009), through the participation in this international organization. However, this set of criteria that Moravcsik puts forward is only meant to address democratic deficit stances, while the substance of his argument, namely that the act of delegation is sufficient to legitimate the European Union and its institutions, is stressed when trying to tackle the assumption that the European Union is not a technocracy and that insulation is not simply an empirical fact; it has normative weight. (Moravcsik, 2002: 613). This implies that also delegation from EU institutions to other non-majoritarian agencies is legitimated by reference to what happens within the Member States: the national central banks are independent from political control and therefore also the European Central Bank is legitimately independent; the same can be applied to regulatory agencies (e.g. food security agencies), even though at the EU level the distance from the citizens the ones from which power flows is much longer than for domestic 8 According to Scharpf and social democratic theorists, it is exactly the absence of a proper welfare that is problematic for the democratic legitimacy of the Union itself (see Scharpf 2009). 5

6 regulatory agencies; moreover, what is the legitimacy of EU agencies (e.g. Frontex, Europol 9, EASO 10 ) to whom powers cannot be said to have been delegated but rather created (Curtin, 2006; Pollak and Slominski, 2009)? At the other end of the debate, multi-level governance (MLG) departs from a different ontological assumption: in the EU hierarchical (i.e. vertical) relationships between delegating authorities and delegated entities are not so clearly defined and the governance system can be better described as multi-layered (Hooghe and Marks, 2003). This means that also sub-national and non-governmental entities shape the EU governance system differently from intergovernmental liberalism, in which only Member States and the EU institutions are recognized as levels of political mobilization and these various levels have overlapping competences, thus creating ad-hoc policy networks (Piattoni, 2009). Also this governance framework, as the one described by Moravcsik, affects the EU and its Member States alike. Particularly regarding the overlapping competences, Majone would argue that this could happen only in some specific policy fields, the ones that are not decided through the Community Method in which a hierarchical a division of powers applies and only clearly identified EU institutions can take part to policy-making but through the so called New Forms of Governance (NFG) (Majone, 1998; Bellamy and Castiglione, 2011). The NFGs were firstly accounted for by the Commission in its 2001 White Paper on Governance but they had been operating for quite some time before being legally recognised and given a general framework; these NFG acknowledged the participation of governmental and non-governmental actors in the EU policy-making fields concerned with environmental and social policies, mainly, and could be described as horizontal relationships based on plurality of actors and networks (Gilardi, 2001:2). Therefore, challenging delegation as the ultimate legitimizing principle of the European Union from a MLG perspective, Piattoni argues that this intergovernmentalist assumption derives from a strong matching between delegation and representation at the national level and, following her argument: representation as delegation [...] cannot serve as a basis for EU democracy (Piattoni, 2012: 2). Moravcsik and Majone, indeed, see delegation as entrusting a trustee, an expert, who does not need to consult with the delegating authority to take decisions on its behalf and thus enjoys an ample room for manoeuvre; at the same time, in order to control the trustee and ensure that he/she complies with national preferences, liberal 9 European Union s law enforcement agency. 10 European Asylum Support Office. 6

7 intergovernmentalists assume that accountability mechanisms are available to national governmental institutions (Piattoni, 2012). These two considerations ample room for manoeuvre for the trustee and strong accountability mechanisms seem to go inevitably in opposite directions. The solution proposed by Piattoni implies a re-definition of the concept of representation according to which national governmental representatives should explain and justify what they do in Brussels (Urbinati 2006) by delivering (two-way) deliberative accountability (Piattoni, 2012: 16). The perils of international delegation are presented also by Agné in his 2007 study on the myth of international delegation. He argues that the action of delegation per se does not imply that the actions of the delegated agency will be democratic. Following Agné argument, delegation to the European Union can be defined democratic in theory inasmuch as the powers that have been delegated can be restored to the democratic hands of national constituencies 11. However, the full retrievability of delegated powers is nowadays virtually impossible, as has been discussed regarding the possible exit of Greece from the Euro Area due to the crisis 12. Moreover, for what concerns delegation of powers to European agencies, the delegation mechanism and the accountability chain(s) are very difficult to define clearly and therefore their democratic legitimacy seems difficult to establish in this perspective. Agné s proposed solution is to reconsider the democratic theory of delegation (Agné, 2007: 20) by departing from the idea that power is delegated or rather alienated as it can never be fully retrieved at any level of detail by the delegating authorities and defining the delegation of autonomy ; in his conception delegation to EU agencies is democratically legitimate if [they] succeed in enlarging people s autonomy, even in cases where authority cannot be recovered by any democratic body (Agné, 2007: 39). The participation criterion is embedded in his theoretical approach, in contrast with intergovernmentalists, but it is also limited by another criterion, that is essential for EU independent or semi-independent agencies: efficiency. The problems of delegation to EU institutions evidenced so far will be discussed further in the next sections, by introducing and applying the principal-agent model of delegation to the ultimate step of delegation (Gilardi, 2001): EU non-majoritarian agencies. This rational choice 11 Art. 49 of the Lisbon Treaty introduces the possibility for Member States to exit the EU. 12 For an interesting informal debate on the issue see the Economist s debate Should Greece leave the Eurozone and return to drachma? available at: (accessed November 2012) 7

8 model is discussed as being the one that seem to give greater plausibility to the assumption that delegation could serve as the foundation of a solid democratic legitimation for the Union. II. Participation, not delegation But what if, as I and many other assume, there are no principals in civil society not even the political parties that connect it to the agents in public administration with the robust and panoramic knowledge needed for this directive role? (Sabel 2004: 3) According to Sabel, the party systems are facing an era of extreme weakness. Piattoni points out some other major problems in our representative democracies, that seem to be disregarded by the theory of delegation: first, a declining party membership, secondly a declining partisan identification of the electorate and, third, the escalating power of the executives (Piattoni, 2012: 2). And there is another feature of our representative democracies that cannot be disregarded; Moravcsik is right in affirming that the national arena is similar to the European arena: voter turnout at the EP elections has been steadily decreasing in the last twenty years (see Graph 1) and the same can be said for Member States national elections (see Graph 2). These issues, taken together, should alert the analysts of democratic legitimacy that the system is changing and, therefore, that there is a need to adapt to the new conditions in order to keep democratic legitimacy in place. As Bellamy and Kröeger point out the EU is a problem for democratic theory insofar as it cannot be democratic according to modern accounts of democracy (2013: 493). Clear signs of discontent and disaffection of citizens towards the existing multi-layered governance system 13 are relevant issues for the development of democratic theory, particularly when the current crisis has made evident that the clear disconnection between European citizens opinion on the EU before the crisis and now is a result of the lack of responsiveness of the institutions. In the first draft of a paper presented in a Conference in Trento in 2013, Moravcsik describes the European voters in a clear way: Europeans [...] are functional pragmatists. They do not care much about the level of European demos, we-feeling, or community, which has not changed much in decades. [...] In order to support EU action, Europeans require a concrete stake in its issues (Moravcsik, 2013: 18). The aim of introducing an experimentalist participation in the EU governance system is to highlight that citizens already 13 Before the crisis, the Eurobarometer recorded high consensus towards the European Union institutions, compared to national governments. 8

9 have this concrete stake and to uncover new opportunities for civil society participation in a democratic polity. It is hardly tenable that participation of citizens to political life is a negotiable issue in a functioning representative democracy. Dahl is probably the most prominent author to discuss the idea of democratically acceptable delegation within the nation state; in order to be considered democratically legitimate, in Dahl s view, any political procedure should fulfil five criteria: political equality, immediately followed by effective citizen participation ( input participation), enlightened understanding, final control over the agenda-setting and, finally, universal suffrage. If it is true that the requirements of democratic legitimacy cannot be equally applied to national states and the European Union, due to the different nature of the two political entities (Majone, 1998; Moravcsik, 2002) and this view is supported in this paper, it is nonetheless necessary to agree on some basic principles that can sustain and improve the democratic legitimacy of the Union, which is lately highly questioned by nationalist and populist parties within the Member States. All the more so because the national and the European arenas are undergoing similar changes and, as a consequence, delegation seems no longer the democratic principle that can primarily fulfil this task. Voters are more difficult to please than ever. In Hirschman s (1970) terms, they increasingly opt for exit strategies. Not only is loyalty to political parties in retreat, but the declining number of party members... indicates that the same applies to voice. In other words, voters increasingly behave like consumers shopping around for the best short-term deal. This may not in the long run produce the best politicians. (Bergman, Müller and Strøm, 2003: 745) Even though voters apathy and shopping for votes, described by Bergman, Müller and Strøm, along with non-salience of some policies 14 are relevant issues concerning the decision of the voters to participate in elections (Moravcsik, 2002: 615), these two characteristics usually affect input participation but they might be reversed when policy-making is actually implemented at the local level and accountability mechanisms become relevant. Implementation might uncover unintended drifts of delegated agencies or bureaucracy from the preferences of their principals that are, ultimately, the citizens themselves ; these drifts affect directly the 14 Populist anti-eu parties and austerity have created new saliency for EU policies, especially in the fields of immigration and economic steering (redistributive policies); moreover, voter turnout remained stable from previous EP elections to the latest, in spite of the steady declining trend (since 1979) shown in Graph 1. 9

10 citizenry, thus activating civil society 15 that is therefore interested in organising itself and in acquiring expertise in order to enter the debate over the substance of policy-making, on a voluntary basis. Also, in this framework, citizens output participation (Finke, 2007) is no more only episodic (Urbinati, 2006: 2 cited in Piattoni, 2012: 6; Sabel, 2004), meaning citizens can directly participate in the EU political life only every five years when elections of the European Parliament require them to and reluctantly so. This form of output participation can be called experimentalist, drawing form Sabel and Zeitlin s experimentalist governance or Directly Deliberative Poliarchy (DDP). According to them, delegating authorities at the top of a completely blurred chain of delegation 16 in order to control the implementation of their preferences can only decide to set framework goals and benchmarks, according to which the actors at the bottom of the chain can be evaluated. Continuous monitoring, at different levels of this blurred chain, detects errors and breakdown, uses these findings to trigger searches for the root causes of design or other flaws that escaped earlier examination. The goal of such root cause analysis is to trace disruption back to its original source, which may not be palpably linked to the proximate cause of the breakdown. (Sabel, 2004: 11). This governance system implies that delegation chains are not relevant, while, at the same time, accountability gains a paramount role. In this framework, accountability makes policyformation more democratic not only by enhancing control but also by promoting trust through more responsive institutions (Fisher, 2004). Moreover, as Sabel explains, non-compliance to the rules of delegation, in a principal-agent perspective 17 would trigger sanctioning while in an experimentalist governance perspective it would trigger mobilization of networks, learning and creation of solutions for the specific case. Civil society organizations role in this framework would therefore be to help revise means and ends (Sabel and Zeitlin, 2008) of European policy-making, by experimenting and learning, in a bottom-up perspective; the consequence would be an increase in the overall legitimacy of the governance system by enhancing citizens awareness of and inclusion in European policy-making (Smismans, 2006: 307). Drawing from the solutions proposed by Agné and Piattoni, the recognition of the role of civil society in experimentalist participation, would meet some of the requirements proposed by 15 The definition of civil society that will be implied in this paper is the one given by Finke: civil society is composed of voluntary groups, organizations and associations which articulate the variety of societal voices in the EU multi-level system. (Finke 2007: 21). 16 Sabel and Zeitlin build their new architecture (Sabel and Zeitlin 2008) on the multi-level definition of governance and more specifically on the II ideal-typical model of governance, in which hierarchical relationships are difficult to define and governmental and non-governmental actors are networked. 17 See next section of the paper. 10

11 the two scholars: on the one hand, a larger number of people can be involved in making policy without endangering either their collective or individual autonomy (Agné, 2007: 40) without hindering effectiveness; on the other hand, policy-makers would be required to meet higher standards of accountability and reason giving due to the expertise and the interest of the accountors. Civil society organizations are particularly suited to participate in the accountability mechanisms of EU governance due to the introduction of article 11 TEU with the Lisbon Treaty, which states: The institutions shall, by appropriate means, give citizens and representative associations the opportunity to make known and publicly exchange their views in all areas of Union action.. The need for more participation at the EU level was already acknowledged in the 2001 Commission s White Paper on Governance, with the recognition of the development of new modes of governance such as the Open Method of Coordination. These new forms of governance involve civil society in the policy-making process from the input side (Smismans, 2006) but only in some specific fields and only with a consultative role. On the other hand, even though article 11 TEU does not empower directly civil society organizations to hold to account European decisionmakers, it gives CSOs a strong leverage to ask for accountability (Kohler-Koch and Kotzian, 2012: 7) and to make EU institutions more responsive to European citizens. Also, the act of holding someone accountable for something is strongly linked with the interest and the capacity or expertise of the accountor and these characteristics are both present when considering civil society organizations vis-à-vis European institutions: CSOs act in the interest of their members and/or have a mission they want to pursue. (Kohler-Koch and Kotzian, 2012: 9). Nevertheless, there are a number of downsides of civil society involvement in the multilayered and non-hierarchical EU governance. Kohler-Koch has long been researching the pros and cons of organised civil society participation through accountability mechanisms and her studies describe a situation in which, in spite of the high expectations on the democratic improvement that civil society participation would have brought about, civil society scored poorly in affecting policy outcomes due to various reasons (e.g. lack of resources, shifting interest of the media, etc.) but in particular to the impossibility to sanction international organizations. Other recent empirical studies have uncovered some problems both in the legitimating role of organized civil society and in the normative strength of this claim. Finke describes two theoretical approaches to the participation of civil society into governance systems: calls for deliberative democracy, and for a democratizing role for civil society through socialization within the European arena (Finke, 2007); she underlines that there is this unresolved tension among scholars but also within the Commission between involvement of 11

12 citizens organizations in policy-making and active citizenship (Finke, 2007: 21). Among other studies it is interesting to mention: Warleigh in 2001 concluded that: NGOs are currently unsuited to the task of Europeanizing civil society thanks to their inability to promote the political socialization of their supporters (2001); Saurugger, more recently, focused on the analysis of the professionalization of civil society organizations, thus describing the emergence of possible issues of legitimacy regarding civil society representation (Smismans, 2006: ). However, the experimentalist participation proposed here does not require a Europeanized civil society that would give substance to the democratic legitimacy of the Union addressing the no demos argument (Weiler, 1995) in order to function; the idea is that through active citizenship and mechanisms of accountability that provide for the possibility of revising means and ends of policy-making, democratic legitimacy is ensured and could be strengthened by managing diversity while implementing European policies: [t]he only way citizens and states are going to see the Union as legitimate is if it generally does what they want it to do, and does not oblige them to adopt policies (Warleigh in Smismans, 2006: 75). It is important to stress, at this point of the discussion, that there is an issue regarding democratic legitimacy that must be taken into account, which is the so-called Monotonicity Mistake (Føllesdal 2012: 4; Smismans 2006: ). Føllesdal describes it as the risk of strengthening one single component of democracy in order to increase the overall democratic legitimacy and achieve mixed results at best (Bellamy and Kroeger, 2013): increasing transparency of EU agencies, for example, might undermine their effectiveness. In this sense, this paper suggests that strengthening experimentalist participation by involving civil society both in the implementation/revision of policies and in the chain of accountability, does not disregard the need to strengthen e.g. judicial accountability of agencies operating in the internal and external defence/security field, or to keep pushing for a revision of the principle of representation in a globalized and media-based democracy 18 ; this is in line with Smismans proposal of adopting a reflexive deliberative poliarchy (Smismans 2006: 310). In the words of Bergman, Müller and Strøm, democracy means effective but limited government containing mechanisms of delegation as well as accountability (2003: 4). 18 The concept of e-democracy is evolving rapidly; all the major scholars debating democratic theories acknowledge the evolution that technological advancement is imposing on policy-makers and on the whole governance system, first and foremost in terms of transparency and quantity of data available to the internet/media user, but also highly affecting the campaigning of politicians and the way in which bureaucracies dialogue with citizens (for the latest development in the field see the ejournal of edemocracy and Open Government, available at 12

13 The empirical validity of experimentalist participation through accountability mechanisms, which sustain the legitimacy of delegation, is discussed in the next section. III. The chain of delegation and EU agencies Inasmuch as EU non-majoritarian agencies are the very last-step of delegation in the chain (Gilardi, 2001) and affect European citizens everyday lives, they seem to be the perfect environment to test this paper s assumptions, because accountability is important when public scrutiny is at its lowest, namely in the day-to-day affairs of [European] governance. (Kohler-Koch and Kotzian, 2012: 4). Moreover, the peculiar form of participatory democracy that is presented in this paper has its main rationale in the principle of good governance, and, as Fukuyama affirms: if we care about the health of democracies around the world, we must also care about the performance of their governments that is, the quality of their state bureaucracies. (Fukuyama, 2013: 6). In order to address these issues, this last section first gives a brief presentation of the principal-agent model and then it analyses what are its shortcomings when applied to EU nonmajoritarian agencies and, specifically, to Frontex; finally, it proposes the application of the experimentalist participation in order to enhance accountability of the European Border Management Agency. American rational-choice institutionalists created a very useful tool to analyse the delegation of power from the national to the international level: the principal-agent theory. In the EU the literature on principal-agent models of delegations started to gain consensus only in the last fifteen to twenty years, giving some very useful insights not only on the reasons why Member States delegate competences to EU institutions in some specific policy-fields and the mechanisms that they enact to avoid that the EU institutions shirk from Member States preferences, but also on why and how EU institutions (mainly the Commission and the Council) delegate some of their competences to other EU non-majoritarian agencies regulatory and coordinating agencies and with what consequences for accountability. The literature review on the principal-agent model and its accountability implications presented here will follow the path traced mainly by Gilardi and Bovens. Principal-agent is a mid-range theory (Ruse and Ruse 2013: 4), thus a useful tool to describe the reasons of delegation, the mechanisms of control enacted by the principals to avoid 13

14 shirking and slippage by the agents, and the effectiveness of said mechanisms. Its application was launched with a Special Issue of the European Journal of Political Research of 2000, whose introduction was titled Parliamentary Democracy and the Chain of Delegation which evolved into a book published in 2003 by the same authors (Bergman, Müller and Strøm, 2000, 2003). Gilardi defines the principal-agent relationship as a social transaction, or interaction, in which one actor, the agent, carries out actions that are intended to fulfil the interest of another actor, the principal (2001: 3); the act of delegation is, therefore, the basis of this relationship. The relationship is defined through a contract between the two actors, in which the principal establishes the limits of action of the agent and tries to set up mechanisms or an incentive structure to avoid that the agent shirks, meaning avoids pursuing principal s interests to maximize its own power/utility, and slips when the structure of delegation the contract itself leaves room for manoeuvre to the agent in a way that might run counter to the principal preferences (Kassim and Menon, 2003): these two problems are called agency losses. Control on the agent is exerted by the principal in seven different ways, enumerated by Gilardi: police patrols oversight, costly and constant monitoring of agent s behaviour by the principal; and fire alarm oversight, in which a third party monitors the conduct of the agent and report it to the principal 19 ; but also control through administrative procedures, through appointments of head positions of the agent, through the budget 20, through the amendment of relevant legislation and, ultimately, through institutional checks (e.g. the Fundamental Rights Agency supervise the conduct of Europol and Frontex for what concerns the correct application of the European Charter of Fundamental Rights). The principles that inform this principal-agent relationship are the same of rational-choice theory: 1) the actors try to maximize their utility; 2) their interests will eventually conflict; 3) the agent is more specialized than the principal, which is one of the most compelling reasons for delegation to an EU agency, thus leading to an asymmetry of information between the actors. Other reasons of delegation have been described as functional or institutional or intergovernmental according to the theory that employed principal-agent models to explain the reasons for integration and the way in which integration moved forward. These rationales for delegation include the need of the principal to gain credibility, to overcome problems of collective action (Kassim and Menon, 2003: 123), to solve problems of incomplete contracting 19 A role that civil society organizations already have in the international arena but that does not have binding nature. 20 The European Parliament has a direct role in its control for EU agencies and can sanction, through budget constraints a misbehaviour of the agency. 14

15 (Pollak, 1997: 103) in the case in which interests are fragmented and bargaining is long and complex, the agent is better placed to mediate among them, to shift blame away from the principal on politically costly decisions and, finally, to escape from political uncertainty. Gilardi maintains that when analysing independent regulatory agencies established within the Member States that according to Thatcher represent [for principal-agent models] a prime example of delegation (2002: 129) the only reason that can justify the principal s decision to delegate powers and competences to an independent agency working at arm s length from the principal is exactly avoiding political uncertainty. This is the only transaction cost, provided by the principal-agent theory that can explain the real independence of national regulatory agencies in Western Europe. However, Gilardi s discussion over independent regulatory agencies, stops at the national level and does not address the democratic legitimacy of delegation to these agencies, but only the reasons why delegation is possible and is permitted at the national level. From a European perspective, EU agencies can surely shift blame away from national government representatives and thus reduce the costs of bargaining, overcoming problems of collective action and providing a higher level of expertise; conversely, expertise can produce information asymmetries, thus it can be accounted for as a potential agency loss. However, agency losses shirking and slippage and controlling and sanctioning EU agencies directly are too costly to bear for any principal for two reasons: firstly, because it is not clear who the ultimate democratic principal is, but many have control of different aspects of the life of the agency (Busuioc, 2010), and, secondly, because the coordinating, managing and regulating powers that these EU agencies are entrusted are new powers that no EU democratic principal (nor the European Parliament, nor the Council) could exercise by itself e.g. the management of operations of Member States border guards (Curtin, 2007; Pollak and Slominski, 2009). As a matter of fact, information asymmetries in this framework increase exponentially, due to the inability of each democratic principal to understand how the agency operates at any level of detail (Agné, 2007). Moreover, at the EU level, democratic principals are not driven by the need to avoid political uncertainty when they establish EU autonomous or quasi-autonomous agencies, as was the case for national independent regulatory agencies. EU agencies started to boom during 2000s, a time in which integration of the Union was perceived as having a steady pace, so much that the idea of a Constitutional process was launched (Laeken Process). These aspects of delegation, described employing the principal-agent model, seem to hinder the democratic legitimacy, at least, of such agencies. 15

16 Nonetheless, in a recent judgement (22 January 2014), the Court of Justice of the European Union decided on the ESMA (European Securities and Markets Authority) case 21, in which the UK sought to limit the powers of the EU legislative branch to delegate competences to EU discretionary regulatory agencies in the field of financial assets and security markets; the Advocate General, in its Opinion, evidenced the possibility for the Court to balance the functional benefits and independence of agencies against the possibility of them becoming uncontrollable centres of arbitrary power (Para. 19) 22. The Court, instead, took an opposite position on the matter and affirmed that the Treaties (Articles 290 and 291) do not shape a closed system of delegation, thus allowing for delegation as long as there is possibility for judicial review of the acts of the delegated EU agency and the Meroni criteria are met. Even though constitutionally this is a strong argument legitimating delegation to EU independent and non-majoritarian agencies, the democratic legitimacy of these agencies might still be perceived as a problem, as the action for annulment that led to this judgement shows. The problem seems all the more real when considering that the agencification phenomenon in the EU has gained momentum in the last twenty years and is shaping European governance (Papadopulos, 2007; Curtin, 2006, 2007; Shapiro, 1997); the picture is complicated further by the existence of European agencies that operate in fields such as internal and external security or taxation, which are core issues for national sovereignty and mentioned by the BVerfG in paragraph 252 of the Lisbon judgement. According to Bovens, the accountability arrangements that can be deemed to democratically legitimize government action are the ones that enable democratically legitimised bodies to monitor and evaluate executive behaviour and to induce executive actors to modify that behaviour in accordance with their preferences (Bovens, 2007: 465). This is the same reasoning that lies behind a well-functioning and hierarchically defined chain of delegation, counterbalanced by a straight and short chain of accountability. As described in this section, the European governance system is not shaped in this way and thus an alternative accountability mechanism that might democratically justify the already legally legitimised delegation to EU agencies, is needed. The social or horizontal type of accountabilities described by Bovens might be helpful in this regard because they allow organised civil society participation in accountability mechanisms. 21 Court of Justice of the European Union, Case C-270/12 UK v Parliament and Council, 22 January More information about the case can be accessed at: 16

17 Thanks to these mechanisms, agency 23 is obliged to explain and justify its conduct vis-à-vis the civil society organizations (the forum) that have an interest to pose questions and have the expertise to pass judgement, according to which the agency may face consequences, including both formal (e.g. Regulation revision) and informal (e.g. media exposure for wrongful conduct) sanctioning. IV. Frontex The European Agency for the Management of the External Borders of the EU is a rather complex compromise between the need of Member States to de-politicise the issue of border management and the will of the Commission not to lose grip upon the European dimension of the issue. At the same time it is a compromise between the intergovenamental method and the supranational method, but it can also be said it to be a compromise between the need for security of European citizens and the duty to protect the rights of third country nationals. Finally, even concerning the principle of solidarity Frontex can be seen as a compromise: between the voluntarism of MSs which may decide whether to take part to Frontex operations or not (now slightly diluted with Regulation 1168/2011) and the burden-sharing idea without which integration in the Union area could not be possible. This last section, describes the accountability regime the coherent complex of accountability relationships described by Bovens (2007a) to which Frontex is subject when human rights are called into question. In order analyse Frontex accountability regime the principal-agent tool is employed, but first it is fundamental to assess who is/are the principal/s in a relationship in which the agent (Frontex) is the only specified actor. Principals in this hierarchical relationship to be considered as Boven s democratically legitimised bodies (2007: 465) are first and foremost elected bodies 24, i.e. the EP and the national parliaments; other EU institutions, among which indirectly elected bodies 25, the Commission, and the European judicial bodies i.e., the Court of Justice of the EU and the European Ombudsman are also in a relationship of accountability with the agency when fundamental rights are at stake and are, therefore, addressed as principals in this section. 23 The actor in Boven s definition of accountability and the agent in principal-agent theory. 24 Lord reports that the EP agrees on this definition of political principals: all administrative bodies should be part of a single hierarchy of political control that leads back to the public via the power of an elected Parliament to sanction a politically appointed leadership of an integrated executive (2011: 912). 25 As evidenced by Busuioc, the political accountability relationship with the Council would be particularly relevant for the study of Council agencies (2010, p. 118), not for Frontex. 17

18 There are three ways in which the EP can control 26 agencies and Frontex in particular and make them accountable. First of all, by making use of its co-legislator role in the drafting of Regulations establishing the agencies: Frontex Regulation 2007/2004 was passed as a Council Regulation and the EP was confined to a consultative role, while the 1168/2011 Regulation was approved with the ordinary legislative procedure. Secondly, through budgetarial control: agencies are financed mainly through the Union s budget which has to be approved by the EP and their financing is often subordinated to an annual budgetary discharge procedure; as a consequence, the EP not only can control agencies agendas, by deciding whether to sign off their annual accounts, but also it can summon executive directors to question them and require satisfactory answers on budgetary issues; however, there is no provision that formally requires the EP to produce an evaluation of the agency activities. The third form of EP oversight is the possibility to always require agencies directors to report on agencies activities - provided by Article 25(2) of Regulation 1168/2011 and to inquire them, even though this is not as binding as a tool as the legislative and budgetary powers 27. In conclusion, the combination of these tools Regulation revision, budgetary control, hearings could describe an effective accountability relationship, provided the EP or, more specifically, the LIBE Commission 28 has the sufficient know-how and expertise to pose the right questions and therefore to evaluate Frontex activity. National Parliaments role in the AFSJ, instead, is delineated by Article 12 TEU. They should always be informed regarding legislative proceedings and they are the depositors of the control on the application of the principle of subsidiarity. Moreover, point (c) of Art. 12 provides a specific evaluation power for national parliaments in the AFSJ especially concerning the implementation of Union policies in that area (see also Art. 70 TFEU), and by being directly involved in the monitoring of the activities of two agencies operating in the Area of Freedom Security and Justice (AFSJ): Europol and Eurojust (Art TFEU). No mention is made of Frontex in the Treaty; this is so because Frontex has no executive power whatsoever as Illka Laitinen affirmed in the LIBE Interparliamentary Committee meeting of 2011, before the representatives of 18 national parliaments of the Union. 26 Control and accountability are two different concepts. Busuioc (2010) argues that accountability is a mechanism of control (even if of indirect control) but the viceversa is not true, meaning that there are forms of direct control that cannot be considered as accountability relations. 27 Debating is probably the most essential requirement for accountability to work; M. Busuioc cites R. Mulgan in this respect, who writes that forcing people to explain what they have done is perhaps the essential component of making them accountable. In this sense, the core of accountability becomes a dialogue between accountors and account-holder. (Busuioc, 2010, p. 34). 28 Civil Liberties, Justice and Home Affairs Committe of the European Parliament. 18

19 Nonetheless, Frontex, as any other EU agency, since 2009 is subject to the jurisdiction of the Court of Justice of the EU, according to Art. 263 TFEU as amended by the Lisbon Treaty. Any act or measure adopted by Frontex can be reviewed by the Court and according to the Charter of Fundamental Rights (Rijpma, 2009). It must be acknowledged that it is difficult to assess how the Court can enter in a judicial accountability relationship with Frontex because still no cases have been brought before it. However, the first quasi-judicial proceeding against Frontex was initiated by the European Ombudsman, Nikiforos Diamandouros on 6 March Even though this owninitiative inquiry by the Ombudsman might or might not become the subject of a judicial action 29, it is the first time Frontex undergoes legal scrutiny from a recognised European quasi-judicial body on the subject of the implementation by Frontex of its fundamental rights obligations (Diamandouros, 2012). Frontex has a duty to answer 30 the information request; as a consequence, this leads to debate and scrutiny, not only between the Ombudsman 31 and the agency, but also among stakeholders willing to give their contribution (i.e., organised civil society, institutional bodies and individuals), so that the debate phase is particularly rich and potentially fruitful. For what concerns sanctioning, the Ombudsman after collecting Frontex and all the stakeholders opinions required Frontex to establish a mechanism for dealing with complaints about infringements of fundamental rights in all Frontex-labelled joint operations (European Ombudsman, 2013). In conclusion, the European Ombudsman, notwithstanding its limited powers, has the great advantage to work as a trait d union among the EP, the Court and the independent agencies and, therefore, it help[s] to reconcile delegation with parliamentary democracy (Magnette, 2003, p. 678). For what concerns Frontex relationship with the European Commission, the main role of the Commission as Frontex principal is to approve the budgetary planning and to control its implementation. Frontex Management Board should also inform the Commission and discuss about the signature of working arrangements with third countries and the deployment of Frontex officials in these non-european countries; the Commission has the possibility to give an opinion on these issues and on Agency s programme of work for the coming year, before it is approved by the 29 The European Ombudsman, however, cannot adopt legally binding decisions but can try to remedy the conflicts through negotiation and, in particularly difficult cases, the European Parliament can decide to bring the case before the Court. The threshold of access to the Ombudsman is very low (anyone can apply for his services), while to access the Court the requirements are very strict and the costs difficult to meet. 30 Which it did on 17 May 2012, with a very detailed Opinion (Frontex, 2012). 31 According to C. Harlow & R. Rawlings, in their Promoting Accountability in Multilevel Governance: A Network Approach (2006), the Ombudsman also engages in standard setting for accountability structures and processes. 19

20 Management Board. More importantly, the Commission has the role of proposing the Executive Director of the agency and it has two representatives sitting in the Management Board of the agency. According to Egeberg, Trondal and Vestlund, Frontex and the Commission cooperate very closely: Frontex and the other agencies of the AFSJ are subject to the Commission s supervision and involvement in their work programmes. They take part in regular meetings with the Commission, and the term parent DG is used across agencies (2014: 15). In sum, the Commission does not have a clear role in holding the agency accountable concerning fundamental rights, but as a close working partner, could have the potential to do so. From this brief review of Frontex accountability relationships, it seems that there is a problem of effectiveness of accountability, particularly vis-à-vis the EP and the national parliaments, when principals either lack the incentives or lack the knowledge and expertise (i.e., information asymmetry) to enter this relationship and engage in the debate over fundamental rights. As Bergman, Müller and Strøm note, Indeed, one might ask whether the main problem in democratic delegation is that the principals, and not the agents, fail to respect their contracts. (Bergman, Müller and Strøm, 2003: 745). This is true particularly for the EP and the national parliaments, except for crisis situations or where serious problems involving a particular agency come to the public eye (Busuioc, 2010: 131). Participation of civil society organizations does not have the two problems faced by principals placed in a hierarchical relationship with Frontex. A clear example is provided by the 2011 revision of Frontex founding Regulation; the revision was triggered by the constant monitoring of the actions of the agency by organized civil society that has questioned repeatedly the conduct of the agency, mobilizing media on the issue (Pollak and Slominski, 2009). The revision of the Regulation can be numbered among the success stories of civil society holding a EU agency accountable on the issue of fundamental rights: the Regulation 1168/2011, indeed, included the provision of the establishment of a consultative body, within the agency, composed of civil society organizations; these civil society organizations are the most interested in monitoring Frontex activities and have the highest expertise to hold the agency accountable. Interestingly, this model of monitoring and consulting has been extended also to the new agency European Asylum Support Office (EASO). Other important amendments of the Regulation repeatedly requested from civil 20

21 society organizations since Frontex inception (e.g., ECRE, Human Rigts Watch, Amnesty International, Statewatch, Migreurop, etc.) are the introduction of a Fundamental Rights Officer, the draw up of a Code of Conduct for Frontex officials and the provision of the suspension or termination of an operation by the Executive Director in case of a serious violation of fundamental rights or of the rights deriving from international protection. In sum, fundamental rights have been recognized as a crucial issue to be addressed by Frontex, thanks to the experimentalist participation of civil society organizations, pressuring Frontex itself, the Commission, the EP and the European Ombudsman to take action. In this system of governance in which autonomous agencies regulate (e.g., ESMA), manage (e.g. Frontex, Europol) or coordinate (e.g., EASO) aspects of EU policy making, and networks are established in order to implement EU policies taking into account the diversity of each Member States, regional and local values and institutional set-up (e.g. the European Patrols Network is the network supporting Frontex work while ESMA's Standing Committees and groups constitute ESMA s network), the vertical, principal-agent delegation does not help explain and enhance the democratic legitimacy of the Union. According to Curtin, the reasons for the demise of delegation as the ultimate democratically legitimizing principle is that the chain of delegation that links the voters to the European autonomous and non-majoritarian agencies (and their networks) is broken: there are multiple principals the problem of many eyes presented by Bovens (2007) and multiple agents the problem of many hands (Bovens, 2007). EU agencies respond to multiple democratic principals; in the case of Frontex the principals are at the same time: Member States that are present in the Management Board, the European Parliament that controls the budget and the Council which, together with the EP has set up the agency; moreover, networks dilute responsibility among a large number of actors (Papadopulos, 2007: 473). Conclusion Social processes, actors interests, electoral systems and decision making patterns are in continuous evolution. What remains constant is the need of citizens to have governments that are the expression of their will and to face a bureaucratic body that is efficient, responsive and responsible for its conduct. The trend of the past twenty years, indicating that voter turnout is constantly diminishing, does not necessary confirm the reticence of European citizens to participate to the political life of the EU. 21

22 Intergovernmental liberalism supports the idea that delegation is key to determining the democratic legitimacy of the European Union; Member States delegating a number of competences 32 to the Commission and the Council of the European Union are the bulwarks of democracy in so far as they can fully retrieve these powers, when their interests are no longer met at the EU level. This view is shared by the German Constitutional Court, which clearly stated it in the famous Lisbon judgement of 30 June However, Agné argues (2007) that the European Union has reached a level of integration (or deepening ) that makes it practically impossible to completely retrieve the delegated powers. Even more so, when looking at powers that have been delegated to or created for EU agencies. As demonstrated in this paper, principal-agent delegation cannot stand as the ultimate foundation for democratic legitimation, particularly not at the EU level. Delegation is a procedural requirement, that has to follow specific criteria in order to be legitimate criteria described by the constitutional courts and the Meroni doctrine, in a legal perspective, and by the principal-agent theory, in a rational-choice perspective. However, in order to be perceived by European citizens as democratically legitimate, a more substantial principle might be of support: experimentalist participation. It is exactly because there is a broader trend toward professional administration, judicial enforcement of rights and strong executive leadership (Moravcsik, 2002: 607) in both the national and the international arenas that it is fundamental to re-introduce, in this new system, citizens voice through an enhanced participation in fora other than the ballot box. In conclusion, this paper has sought to develop the current literature on delegation and accountability mechanisms by proposing a change in perspective regarding civil society participation in the EU governance system. In his 2007 article on the role of horizontal forms of accountability, Bovens (2007b) describes the involvement of civil society organisations and the scrutiny by the media as a supplementary means for holding the multiple levels of governance accountable, together with more traditional vertical and hierarchical forms of accountability. The impossibility of substituting vertical accountability, described by the principal-agent model of delegation, with scrutiny by civil society organisations and the public in general is recognised in this paper; however, building on Bovens analysis, the argument presented here has gone further by describing a situation in which the crisis of representation is evident both at the national and at the European levels, thus placing high expectations on the promise of a larger involvement of interested civil society organizations into the EU governance system. It will be interesting to 32 Principle of conferral, article 114 TFEU. 22

23 analyse, with further research, how much European agencies actually learn from the exchanges with civil society and how much civil society organizations become institutionalized, thus focusing on whether or not they tend to lose their link with the reality they were trying to promote and protect. 23

24 Graph 1 Voter turnout, European Parliament elections ( ) Source: TNS/Scytl in cooperation with the European Parliament. Available at: [accessed 30/05/2014] Graph 2 Source: IDEA Voter Turnout Database, Eurostat [accessed January 2014] 24