Holding the European Asylum Support Office Accountable for its role in Asylum Decision-Making: Mission Impossible?

Abstract The Common European Asylum System (CEAS) seeks to harmonize national asylum procedures. The initial implementation design of the CEAS, reflective of the theory of executive federalism, foresaw that national authorities were to conduct asylum processing and implement the harmonized norms. The implementation design of the EU asylum policy has, nevertheless, started to shift. An integrated European administration is emerging. One area this is pronounced in is asylum decision-making, where patterns of joint implementation have surfaced. This term broadly refers to staff and experts deployed by the European Asylum Support Office (EASO), an EU agency, working alongside national administrators, including on the processing of asylum claims. This Article scrutinizes the emergence of joint implementation patterns in EU asylum policy and the resulting accountability challenge, drawing both from legal analysis and political science theories. I also refer to administrative practice as documented in secondary sources. EASO is currently subject to a mosaic of accountability processes. Two main pitfalls emerge: the intricate balance between accountability and independence; and accessibility for the individual. Against this backdrop, I focus on extra-judicial accountability through the European Ombudsman which, combined with the envisaged internal “individual complaints mechanism” within EASO, could go some way in ensuring applicants’ procedural rights.


A. Introduction
The spike in arrivals of individuals seeking asylum in the EU in 2015 highlighted the limitations inherent in the legal design and implementation modes of the EU asylum policy. The initial implementation design foresaw that national executives assume responsibility in the main for the application of European law. Institutionalization of practical cooperation through EU agencies has begun to unsettle this. EU agencies have come at the forefront in the asylum field for two primary reasons: as vessels to overcome the policy implementation gap; as vessels to enhance inter-state solidarity. Initially their mandate was heavily focused on activities such as information exchange, training, and risk analysis. Over time, it has expanded, and so have their human and financial resources.
These developments should be placed at the backdrop of the broader phenomenon of agencification of the European administration. Initially used in 1975 when the first two EU agencies were created, the agency model has been picking up speed since the 1990s. 1 It has been advanced that two types of agencification can be distinguished: quantitative agencification, which refers to the growing number of EU agencies, the increase in staff and total budgets, etc., 2 and qualitative agencification that relates to the growing role of the EU agencies in delivering EU policies and the increasingly important powers conferred on them. 3 Academic commentators and the European institutions explained in several ways the rise of agencification. The European Commission advanced in 2002 that: The independence of their technical and/or scientific assessments is, in fact, their real raison d'être. The main advantage of using the agencies is that their decisions are based on purely technical evaluations of very high quality and are not influenced by political or contingent considerations. 4 The appeal of separating the political from the technical aspects of a decision-making process is obvious. Nonetheless, it has been observed that in reality they are intertwined to such degree that it might be difficult to clearly divide and attribute them respectively to agencies or the Commission. 5 Another set of authors links the need for agencies with "administrative integration," 6 or with the development of a "system of integrated administration" in Europe. 7 The crux of the argument is that agencies coordinate the cooperation of the implementing entities (that is, the Member States), thus contributing to the integration of the EU and national levels when it comes to policy implementation.
Focusing specifically on the de jure and de facto mandate expansion of the European Asylum Support Office (EASO), 8 the operational expansion of its mandate has led to patterns of joint implementation. This term broadly refers to EASO's own staff, and experts deployed by this agency, working alongside national administrators in fields such as information provision to individuals and the processing of asylum claims. 9 The latter has taken place as part of the operationalization of the "hotspot approach" to migration management. 10 These developments point to the gradual emergence of an "integrated European administration" 11 in the asylum area where the EU level is also involved in implementing activities undertaken by Member States authorities. 12 In addition, EASO is increasingly operational in third countries based on bilateral or multilateral agreements, becoming an actor in the EU externalization of migration control. The exercise of executive powers and discretion by EU agency (deployed) staff has direct impact on the fundamental rights of migrants and asylum seekers. In addition, agency deployments in third countries raise further fundamental rights concerns, and the need to coordinate action with international level stakeholders. These developments point to the need for appropriate accountability arrangements so that these implementation activities will not result in violations of fundamental rights, including but not limited to fair procedures.
Some of the most relevant fundamental rights in this context are procedural rights; the prohibition of refoulement; and rights against ill-treatment. Crucially, the right of access to the asylum procedure is a fundamental prerequisite to the identification of persons in need of international protection, and thereby to granting the requisite protection and entitlements; but it can be rendered meaningless if the procedure is not fair and effective. There are numerous issues around the organization of an asylum examination process that affect the quality of decision-making, such as the right to be heard (in a personal interview), the right to legal assistance, and the right to an effective remedy. The latter is a central guarantee as it ensures scrutiny of the quality of the original asylum determination process, thereby preventing prohibited refoulement from occurring. Finally, the effectiveness and fairness of an asylum procedure is also tied up with the conditions that individuals face during this type of administrative processing. Destitution, substandard material conditions, or arbitrary deprivation of liberty of asylum seekers have a direct impact on the quality of the asylum procedure, while at the same time constituting violations of fundamental rights in and of themselves.
Complaints before the European Ombudsman introduced by two civil society organizations, respectively the European Centre for Constitutional and Human Rights (ECCHR), and Advocates International, illustrate these challenges. 13 These complaints, which I analyze in detail in this Article, raise issues of maladministration and procedural failings of EASO for its role in the processing of asylum applications as part of the "hotspot approach to migration management," 14 and also highlight the accountability challenges resulting from joint implementation scenarios.
The "hotspot approach" is part of the EU level responses to the 2015-16 spike in arrivals of third country nationals and stateless persons at the EU's external borders, many of them with international protection needs, for example, fleeing persecution or generalized violence. Notably, Member States detected 1.82 million irregular border crossings in 2015, especially on the so-called Eastern Mediterranean route between Turkey and the Greek islands in the Eastern Aegean Sea (885,386 crossings) 15 and through the so-called Central Mediterranean route to Italy (154,000 crossings). 16 The hotspot approach to migration management essentially concerns interagency collaboration, where deployed national experts under the coordination of a . 15 Frontex Annual Risk Analysis for 2016, at 6 (March 2016), available at https://data.europa.eu/euodp/en/data/dataset/ara-2016. 16 Id. at 6. specific agency-EASO, Frontex, Europol, and Eurojust-operationally assist national administrations in "hotspots" for migrant arrivals. It comprises a variety of administrative tasks, including registration and identification of migrants, and channeling of migrants into further procedures, for example, return or assessment of an international protection claim. 17 This approach, first rolled out in Italy and Greece, was novel: Although the respective agency regulations foresaw deployments, 18 the element of interagency collaboration in what in essence would be a single operational framework was never before so clearly articulated. While joint implementation patterns are not limited to the hotspot approach, the fundamental rights challenges of increased integration between the EU and national levels in assessing asylum claims evidenced in the Ombudsman complaints first emerged in this context. This Article scrutinizes the emergence of patterns of joint implementation in EU asylum policy and the resulting accountability challenge, drawing both from legal analysis and political science theories. I also refer to administrative practice as documented in secondary sources. The analysis proceeds in five steps. I first briefly document the passage from ad hoc projects to institutionalization of practical cooperation in the EU asylum policy (Part B). Next, I develop a conceptual framework for the joint processing of asylum applications. Joint processing is a term which is yet to be clarified and allows for the development of various practices-conceptual clarity is therefore necessary (Part C). Thereafter, I critically assess the implementation of joint processing in Greece, highlighting the constitutional and administrative law challenges it raises (Part D). Next, I explain the definition of accountability I adopt, and I provide a panorama of the existing accountability processes for EASO (Part E), before I analyze more in-depth the operationalization of extra-judicial accountability, that is, the role EU Ombudsman (Part F). I conclude with critical remarks on the existing accountability avenues and their effectiveness on addressing the challenges that mixed proceedings stir up in asylum decision-making (Part G).

B. Institutionalization of Practical Cooperation in the EU Asylum Policy
The initial implementation design of the EU asylum policy foresaw that Member States would realize the Common European Asylum System (CEAS) largely through deploying their own resources. It was therefore broadly underpinned by the theory of executive federalism. 19 This theory can be summed up as follows: Apart from exceptional cases where the EU level directly implements policies (direct implementation), national executives assume responsibility in the main for the application of European law (indirect implementation). 20 Strictly applied, this leads to a neat division of labor for most policies, where legislation is adopted at EU level, and the implementation of EU law is a matter of predominantly national concern. The realization of a common policy is thus secured primarily through legal harmonization, aided by the Court of Justice of the European Union (CJEU) as the authoritative interpreter of EU law. Exceptionally, the Commission is involved in its role of guardian of the treaties, to pursue infringements of EU Lenaerts describes "executive federalism" in the following sense: "[T]he division of powers between the central government and the component entities is not just one defined in terms of areas in which each government holds substantive competence, but relates also to the split between the central government holding the legislative power and the component entities holding the executive power in a given areas"; see Koen Lenaerts, Regulating the Regulatory Process: "Delegation of Powers" in the European Community, 18 EUR. L. REV. 23, 28 (1993). 20 See, e.g., JÜRGEN SCHWARZE, EUROPEAN ADMINISTRATIVE LAW 6-8 (rev. 1st ed. 2006). law. The theory of executive federalism, however, increasingly fails to capture the reality of implementation of EU law, and the intricate links that increasingly exist between the EU and the national levels.
The Common European Asylum System as set down in the Lisbon Treaty was largely conceptualized as a "common system of national variants." 21 What is common is the set of legal rules that Member States are called on to implement, rather than the implementation stage itself. Neither the Treaties nor secondary legislation initially foresaw intense administrative interaction at the implementation stage. On the contrary, each Member State holds primary responsibility for implementation at national level of the harmonized legal rules set at EU level. This can be clearly inferred from Article 78 TFEU, which foresees that "a Member State" is ultimately responsible for the examination of an asylum application. 22 Secondary EU legislation on asylum is also explicit on this point: The norms contained in these instruments impose specific duties on Member States with a view to establishing functioning and well-resourced national asylum systems. 23 In addition, there is no system of mutual recognition of positive asylum decisions on the basis of EU law, nor do recognized beneficiaries of international protection benefit from any type of free movement rights outside the Member State that issues the residence permit on protection grounds. 24 Nonetheless, there have been shifts in the implementation design of the EU asylum policy linking Member States with each other, as well as with the EU institutions. 25 Practical cooperation, which has seen the emergence of joint implementation patterns, is one those shifts.
Practical cooperation between Member States in asylum initially consisted of information exchange through administrative networks and ad hoc projects, 26 and was envisaged to play a marginal support role in the implementation of the policy. These collaborative efforts soon met their limits in boosting Member States' capacity to implement the asylum policy. An implementation gap emerged that was more pronounced in Member States at the external borders of the Union, which had structural deficiencies in their asylum systems and were overburdened by the workings of the so-called Dublin system. 27  For a critical assessment of the shifts in the CEAS implementation design see Tsourdi, supra note 12, at 204-23. 26 Characteristic examples were through EURASIL, an EU network for asylum practitioners chaired by the Commission, and the General Directors of European Immigration Services Conference (GDISC). The latter was established by the 2004 Dutch Presidency. 27 Reference is made implementation challenges led to an institutionalization push, which came to fruition in 2010 through the adoption of the EASO Regulation. 28 EASO was not meant to be involved in asylum decision-making. The EASO Regulation, which remains in force today, explicitly excludes individual decision-making powers: "The Support Office should have no direct or indirect powers in relation to the taking of decisions by Member States' asylum authorities on individual applications for international protection." 29 This formulation is quite far-reaching since it even excludes indirect powers. There is no CJEU case law yet on what a direct or an indirect power is in relation to the taking of a decision on an individual application for international protection. However, to ensure the effet utile of the provision, the term indirect power should be interpreted to refer to the procedural involvement by the agency beyond centralized decision-making by the agency itself. 30 Its mandate also excludes the adoption of legally binding documents that would instruct Member States about the grant or refusal of individual applications. 31 The Regulation envisages three main areas of activities for the agency. First, EASO is to facilitate, coordinate, and strengthen practical cooperation among EU Member States. 32 This includes the gathering and exchange of country of origin information (COI) and the adoption of a common COI methodology, as well as the provision of training for asylum officials based on the European Asylum Curriculum (EAC). The issue of COI, specifically, is intrinsically linked with credibility assessment in the asylum determination process. Here, EASO's mandate includes gathering COI; drafting COI reports; the management and further development of a portal for disseminating COI; as well as the development of a common format and a common methodology for presenting, verifying and using COI. 33 All relevant sources of information should be used, including information gathered through non-governmental organizations. 34 Moreover, the analysis of country of origin information should be transparent and, where appropriate, use the results of meetings of working parties. 35 These latter provisions seek to safeguard the quality of material produced by the agency. As already outlined, the Regulation highlights that the COI analysis should not purport to give instructions to Member States about the grant or refusal of applications for international protection. Thus, Member States are not bound legally by the analysis included in the material EASO produces.
The agency published in 2012 a report stating the methodology and standards it will follow in drafting its COI reports. 36 Although this text is not binding, all Member States were encouraged to use this methodology when producing COI reports. Thereafter, the Management Board relies on a Strategic Network composed of COI Heads of Unit from different Member States and has also set up a series of expert networks on key countries of origin. EASO organizes country specific workshops and jointly produces reports with Member State experts. These reports are public, providing to fair sharing of responsibility between the Member States, since that responsibility will largely depend on the geographical location of each State. See EASO Regulation rec. 14. This is reiterated in EASO Regulation art. 2(6). 30 See also analysis infra Part C. 31 EASO Regulation art. 12(2). On the role of Frontex in the implementation of integrated border management see Melanie Fink, The Action for Damages as a Fundamental Rights Remedy: Holding Frontex Liable, in this issue. 32 See EASO Regulation art. 2(1) and Section I. 33 See EASO Regulation art. 4. 34 See EASO Regulation art. 4(1)(a). 35 See EASO Regulation art. 4(1)(e). 36 See EASO Country of Origin Information Report Methodology (July 10, 2012), https://www.easo.europa.eu/sites/default/ files/public/coireportmethodologyfinallayout_en.pdf. the opportunity to other actors to scrutinize and comment on their content, such as asylum applicants and their advocates at national level. The agency released the first reports in July 2012 focusing on the situation in Afghanistan. 37 Since then, it has been incrementally producing further reports per year, 38 as well as organizing activities that will boost the capacity of Member States to produce and use COI, such as relevant thematic meetings and conferences, or training sessions on how to use and analyze COI.
EASO products should be considered as one of the available tools to national decision-makers in their appreciation of the situation in the country of origin. Despite their non-binding character, EASO COI reports do have the potential to influence national decision-makers, given the authoritative role of the agency. Hence, it is crucial that they integrate information from a variety of actors, that they adhere to the standards of objectivity and impartiality, and that they remain up to date. In this vein, UNHCR published its own perspective on the forced recruitment by the Taliban in Afghanistan shortly after the issuance of one of the agency's first reports in 2012, contesting some of EASO's findings. 39 If the current steering potential of decision-making through COI could be described as indirect, the 2016 Commission proposal for a revamped European Union Agency on Asylum envisages a more robust role for the agency's products and several processes that would grant them a type of "enforceability." 40 However, negotiations on this proposal were still pending in January 2020 and its fate is uncertain, thus it is premature to draw any conclusions as to the future role of the agency.
The second area of activity is support towards EU states under particular pressure, drawing upon all useful resources at EASO's disposal, which may include coordinating resources provided by Member States. 41 This area is intrinsically linked with the element of enhanced solidarity between the Member States. The agency has adopted a flexible definition of what constitutes pressure, and the determination is a relative one, not based on a fixed calculation. 42 Assessed on an absolute scale, the extent of operations is not impressive, but they can greatly enhance the capacity of a smaller Member State. The final area of involvement is the contribution to the development of the CEAS. 43 It includes an annual report on the situation on asylum, and the possibility to adopt guidelines and operating manuals.
Member States contribute to this Pool by proposing experts that correspond to the required profiles. 46 They retain autonomy regarding the selection of the number and profiles of deployed experts, as well as the duration of their deployment. 47 The agency has further refined the types of operational activities to four, without an official basis in the Regulation and offering little conceptual clarity. It distinguishes between "special support," 48 "emergency support," 49 "joint processing activities," 50 and the "hotspot approach." The first such operations were launched shortly after the EASO's establishment, and they gradually grew in number and scope. The initial AST deployments were not operational like the border guard teams deployed by Frontex, which interacted with individual migrants at external borders. 51 Rather, their work consisted of expert advice in ministry departments, or involved training and study visits of members of national administrations. 52 Gradually, EASO tested joint-processing pilots and eventually forms of joint processing emerged. 53 Given the exclusion of even indirect powers in relation to the taking of individual decisions from the agency's mandate, it is important to analyze the emergence of, and EASO's role in "joint processing activities." There is, however, paucity of an analytical framework and absence of a legal definition of the concept of joint processing. Hence, in the next part, I develop my own conceptual framework for the joint processing of asylum applications as a necessary basis to conduct a principled analysis.

C. Establishing a Conceptual Framework for the Joint Processing of Asylum Applications 54
The idea of joint processing of asylum applications has been lingering in the policy agenda, 55 but a first feasibility study for the Commission was only concluded in 2013. 56 The content of the term 46 See EASO Regulation art. 15(2). 47 See EASO Regulation art. 16(1). 48 This, according to EASO, refers to "tailor-made assistance in order to improve the implementation of the CEAS: capacity building, facilitation and coordination of relocation, specific support and special quality control tools." See EASO website, available at https://www.easo.europa.eu/operational-support/types-operations. 49 This refers to "organising solidarity for Member States subject to particular pressures by providing temporary support and assistance to repair or rebuild asylum and reception systems." See id.

50
These refer to "Member States who are in need of external help in the management of their specific case-load can request from EASO the deployment of Joint Processing Support Teams. The joint processing activities are launched by EASO and the respective hosting Member State, after the Terms of reference, including the main criteria of deployment of the joint processing teams, have been agreed upon." See id. joint processing is yet to be clarified. It allows for the development of various practices. For the purpose of the above-mentioned study, the Commission retained a broad definition: An arrangement under which the processing of asylum applications is jointly conducted by two or more Member States, or by the European Asylum Support Office (EASO), with the potential participation of the UNHCR, within the territory of the EU, and which includes the definition of clear responsibilities during the asylum procedure and possibly also for dealing with the person whose application was jointly processed immediately after a decision on his/her case was taken. 57 The think tanks that were responsible for the feasibility study, Ramboll and Eurasylum, then contemplated different options (feasibility assessment technique) ranging from assistance in emergency scenarios through agency deployments to a completely harmonized approach, meaning centralized processing. 58 After developing a conceptual framework for analyzing joint processing (Section I), I retrace the beginnings of EASO's involvement in joint processing of asylum application which took the form of small-scale ad hoc pilot projects (Section II).

I. Unpacking the Concept of Joint Processing
Cognizant of the legal limitation included in the EASO Regulation, 59 I distinguish here between three scenarios: assisted processing; common processing; and EU-level processing. I understand assisted processing to refer to the examination of asylum applications by officials of the competent Member State with the support of officials of one or several other Member States, possibly coordinated through EASO. This would mean in practice either that national officials are active at every procedural stage and are merely assisted by the EU (coordinated) level, or that deployed experts are not independently undertaking actions or adopting decisions that involve executive discretion. By executive discretion here I am not referring to the classical division under national administrative law between discretionary decisions (Ermessensentscheidung or pouvoir discrétionnaire) and those where precise conditions control or "bind" the decision (gebundene Entscheidung or compétence liée). Rather I use this term to denote more broadly administrative actions that have the capacity to affect the legal position of an individual, whether directly or indirectly.
Thereafter, common processing essentially refers to "mixed," or "composite" administrative proceedings. 60 Broadly speaking: [T]hey ensure that input into single administrative procedures can be given from authorities from various jurisdictions. Irrespective of whether a final decision will be taken by a Member State or an EU authority, both levels can thus be directly involved in a single administrative procedure. 61 Their "mixed" or "composite" character refers to the variety of jurisdiction levels involved in a single administrative procedure. In our context, they concern asylum-related decision-making. They occur when the EU (coordinated) level would be exclusively responsible for one or more 57 Id. at 2. 58 Id. at 2-4. 59 See EASO Regulation rec. 14 and art. 2(6). parts of the procedure that involve executive discretion (such as responsibility determination under Dublin, or proposing a decision on the basis of an interview).
The final scenario is then EU-level processing, where the joint elements disappear, as the decision is taken entirely by an EU authority instead of the Member States. This third scenario would require an amendment of the TFEU, which envisages that "a Member State" is ultimately responsible for the examination of an application. 62 The second scenario is beyond the limits of the current mandate of EASO, which excludes the exercise of direct or indirect powers in relation to the taking of decisions by Member States' asylum authorities on individual applications for international protection. 63 Only the first scenario, assisted processing, is within EASO's mandate.

II. Joint Processing: Timid Beginnings
Joint processing of asylum applications emerged first through a number of pilot exercises which took place between June 2014 and June 2015. EASO stated that 22 experts took part in the joint processing pilot projects conducted by EASO in 9 Member States in 2014 and 18 experts from 15 Member States were involved in three EASO pilot projects in 2015. 64 These were one-off, shortterm exercises to test the feasibility (in practice) of joint processing. Very scarce information exists regarding the precise operationalization of these first pilots. Below I undertake an evaluation of their legal nature based on information provided by EASO. From the little information that is available it would appear that the deployed experts undertook activities that fell both under the first (assisted processing) and second (common processing) scenarios contemplated above. Two examples illustrate this point.
A pilot implemented in Poland from January 19 to February 11, 2015, deployed twelve experts from eight countries in three locations. 65 Two teams performed joint registration and identification of applicants together with the Polish Border Guards. 66 This involvement falls under the first scenario, that of assisted processing, since EASO stressed that they were continuously "under the supervision of the Polish Border Guards." Another team jointly processed Dublin cases within the Polish Dublin Unit. Regarding that part of the deployment: Due to the existing high level of harmonization (common templates, DubliNet, English as working language) they were operational within two days and performed their tasks independently. They were able to register and archive their own cases in the Polish national database. 67 In this case, whether we are in the first or second scenario depends on the extent of executive discretion involved in "registering and archiving." If the deployed experts were the ones who made the decision on which Member State was responsible for processing the application as part of their tasks, then this concerned a form of common, rather than assisted processing, as this decision involves elements of executive discretion. 68 If they were merely typing in and archiving the decisions taken by the Polish Authorities, this was assisted processing.
Another "Asylum Determination Pilot," involving Belgian and Swedish officials, was implemented in the Netherlands between February 23 and March 13, 2015. 69  Their task was to perform in-merit personal interviews and based on the results, prepare the draft decisions. The Belgian members of the Processing Support Team performed their tasks in Dutch, while the Swedish expert used the English language (including recording the minutes of the interview). 70 The little information that is available thus seems to be pointing to the fact that experts were not merely assisting the Dutch authorities; "each expert started working independently on their assigned cases within the first days of the exercise." 71 This is arguably the second scenario, common processing, which raises distinct legal questions. These pilots prepared the ground for operations in hotspots. I explore the implementation of joint processing in Greece and the constitutional and administrative challenges this type of processing stirs in the next part.

D. Implementing Joint Processing in Greece: Constitutional and Administrative Challenges
The refugee protection crisis of 2015-16 brought about the passage from short-term, limited scale pilot projects to larger scale implementation of joint processing, initially in the framework of the hotspots. Given the absence of an overarching legal framework regulating the hotspots, there are in fact operational differences in each Member State where this approach is implemented, and the situation on the ground is constantly shifting. Undoubtedly, the operation of hotspots has led to increasing integration between the EU and national levels in implementing EU's asylum and external border control policies. 72 I focus my analysis on the case of Greece, which was the first Member State where this increasing integration led to the emergence of asylum decision-making through mixed administrative proceedings (Section I). An offspring of the hotspot approach to migration management, joint processing has expanded in Greece beyond the implementation of this approach. Thereafter, I critically assess the constitutional and administrative challenges that the implementation of joint processing in Greece raises (Section II).

I. Asylum Decision-Making Through Mixed Administrative Proceedings in Greece
By January 2020 the agency's mandate remained unaltered with no final agreement having been reached on a 2016 Commission proposal which included an expansion of its role in decision-making. 73 In Greece, its role in asylum decision-making is currently regulated exclusively by national administrative law, rendering asylum decision-making a unique case of mixed administrative proceedings established by national law. Therefore, EASO's involvement is de jure only regulated by Greek law, whereas from the perspective of EU law the agency's involvement is a de facto development.
EASO's involvement was first regulated in 2016 with two consecutive amendments of Greek national administrative law. 74 Notably, a law adopted in April 2016 and amended in June 2016, 70 Id.  transposing among other elements the recast Asylum Procedures Directive, 75 establishes an accelerated border asylum procedure addressing also the situation at hotspots. 76 In its 2016 version, the law stated that in case of large numbers of arriving third country nationals or stateless persons who seek asylum at border areas, in transit zones, or in centers of reception and identification (which is the name given under Greek legislation to hotspots), an exceptional procedure applied. 77 As part of this exceptional procedure, the law in question foresaw that "interviews with applicants for international protection may be conducted by personnel made available by EASO." 78 The national law provisions on EASO involvement were amended in June 2016. Notably, the original April 2016 version of Law 4375/2016 stated that interpreters, as well as seconded personnel made available by EASO, "may assist the Greek Asylum Service in recording the claim; the interview; and any other process." Thus, the April 2016 version of the law stated that the Greek Asylum Service can be assisted (μπoρεί να επικoυρείται) by EASO experts and interpreters. However, it already did not reflect the administrative reality on the ground. Hence, the law was amended in June 2016 to state that deployed experts can conduct asylum interviews.
EASO-deployed experts at hotspots in Greece are independently conducting a part of the asylum process that entails discretion. They conduct the asylum admissibility interviews on behalf of the Greek Asylum Service, at least in the majority of cases, and then submit their findings, on the basis of which the Service issues the final admissibility decision. 79 The admissibility phase is geared at weeding out applicants that could be returned to safe third countries. In the case at hand, it is part and parcel of the operationalization of the so-called EU-Turkey deal. 80 This is arguably a legally non-binding document that mentions explicitly commitments of the EU and its Member States towards Turkey as part of a cooperation "bargain" with a view to returning arriving migrants, asylum seekers, and refugees from the EU territory to Turkey. 81 Despite the nonbinding nature of the document, the implementation of these arrangements involves Member State action under the asylum and return acquis, and the consequent applicability of the EU the amendments of June 2016 see ECRE, Greece: Appeal Rules Amended After Rebuttal Of Turkey's Safety (2016), http://www. asylumineurope.org/news/01-07-2016/greece-appeal-rules-amended-after-rebuttal-turkeys-safety. 75 Reference is made to the 2013 Reception Conditions Directive (recast). 76 See Law 4375/2016 art. 60(4). 77 Id. Charter of Fundamental Rights on those Member State actions. 82 Inherent parts of the admissibility process are assessing the credibility of the applicants, detecting vulnerability, and making a finding on the safety of third countries; all of these entail elements of discretion. EASO's implication in processing is numerically significant. For example, EASO conducted 8,958 interviews in the fast-track border procedure during 2018. 83 During the first half of 2019, EASO conducted 2,955 interviews in the fast-track border procedure, mainly covering applicants from Afghanistan, Palestine, Iraq, Syria, and Cameroon. 84 Legal developments at national level did not stop there, though. A 2018 amendment foresaw that Greek speaking EASO staff can be involved in a similar manner in the examination of the merits of asylum claims. 85 Therefore, this no longer concerns operations as part of the hotspot approach to migration management and the admissibility stage, but rather the regular procedure. As of July 2019, EASO experts supporting the regular procedure were deployed in two locations: 21 were deployed with the Regional Asylum Office of Lesbos and nine with that of Thessaloniki. 86 On average, each caseworker conducts four interviews per week. 87 These developments are crystallized in the latest legislative amendments of November 2019 that saw the adoption of a revamped "Greek International Protection Act" bringing together in a single legislative instrument all relevant asylum legislation. 88 EASO is also involved in a support function at second instance decision-making in Greece. It provides so-called rapporteurs to the national Appeals Committees, a function that is expressly stipulated by national law. 89 According to Greek legislation, rapporteurs are tasked with preparing a preliminary report which includes the following elements: "A narrative account and processing of the facts of the case and of the arguments raised by the appellant, as well as their correlation with information on the country of origin that is to be made available to the independent Appeals Committees for their judgment." 90 Recent field research conducted by ECRE confirms that rapporteur tasks are limited to the initial preparation the case file and to conducting COI research upon request by the Committee members. 91 They are therefore not providing members of the Appeals Committees with a concrete legal opinion, even an advisory opinion, regarding the grant of international protection. Hence, stakeholders at national Greek level did not identify conflict of interest issues stemming from the presence of EASO experts in both the first and second instance procedure. 92 The composition of the Appeals Committees has been altered at the end of 2019 through the revamped Greek International Protection Act. Each Appeal Committee now consists of three   Greek International Protection Act art. 95(5) (author's own translation from the original Greek legislation). 91 ECRE, supra 84, at 18 and fns. 110 and 111 (referring respectively to information provided by UNHCR Greece, July 9, 2019, and EASO Operation Greece, July 10, 2019). 92 Id. at 18. administrative law judges, the latest legislative amendment having removed from their composition the UNHCR representative. 93 The previous change had taken place in June 2016, and had altered their composition to two administrative law judges, appointed by the "General Commissioner of the Administrative Courts," and one UNHCR representative. 94 Prior to June 2016, they had comprised one representative of the Ministry of Interior, one UNHCR representative, and one representative appointed from a list of human rights experts compiled by the National Commission on Human Rights. The procedures themselves were already amended in June 2016 to remove a provision that allowed the appellant to request a personal hearing before the Appeals Committees at least two days before the appeal. The Greek Government in 2016 ostensibly undertook this move to address the "disjunction" between the decisions at first instance that authorized return of applicants to Turkey on the basis that it constitutes a safe third country, and the extremely high rate of reversal on appeal. 95

II. Constitutional and Administrative Challenges in the Emerging Integrated European Administration
Asylum decision-making through mixed proceedings stirs up several challenges of a constitutional and administrative nature in EU law. 96 It also constitutes a significant development in the manner of implementation of the EU asylum policy. Regarding the latter point, the de facto workings of the EU asylum agency point towards the emergence of an increasingly integrated European administration. In terms of EU administrative law, there is an emergence of a variant of procedures that could be understood as de facto composite, or mixed, administrative procedures. Namely, although the asylum decision-maker at first instance according to both EU and national law is the Greek Asylum Service, de facto this decision is based on a recommendation from, and facts ascertained during, an interview conducted by experts deployed by an EU agency. While EASO (deployed experts) are not the ones adopting the final decision, their advisory opinions influence the outcome, albeit indirectly. Issues such as fact assessment and credibility assessment are ascertained on the basis of this interview that therefore greatly impacts the final decision. Hence, this is morphing de facto into a composite process.
The 2016 Commission proposal to expand the powers of EASO and transform it into an EU Asylum Agency confirms these trends. While its fate is unclear, that proposal introduced elements of not only assisted but also common processing in the agency's mandate. The envisaged measures as part of operational support, are variegated. They include preparatory acts of the asylum procedure that do not entail executive discretion, such as assistance with the identification and registration of third country nationals, and assistance with the provision of information 93 Greek International Protection Act art. 116(2). 94 This modification took place as part of the same June 2016 package that clarified that EASO deployed experts can conduct interviews at hotspots. See Law 4375/2016 art. 5(2). The modification seems to have been supported by the European Commission. A trace of this can be discerned in the Commission's second report on the progress made in implementing the EU-Turkey statement. After noting that the pace of returns had been slower than expected, and that the great majority of initial inadmissibility decisions had been overturned by the Appeals Committees, the Commission notes the legal steps that Greece and Turkey have undertaken to achieve further progress, stating: "To ensure full respect of EU and international law, on the international protection procedure. 97 The proposed regulation also includes a form of common processing: Migration management teams deployed in areas under pressure could potentially be tasked with the "examination of claims," 98 even though the final decision on protected refugee status would remain the competence of Member States. 99 These trends were further enhanced in the Commission's amended proposal, released in September 2018, which in many respects complements, rather than repeals, the earlier proposal. 100 Nonetheless, at the time of writing, these legal developments remained prospective.
In the meantime, de facto administrative integration raises challenges of both a constitutional and administrative nature. In order to ascertain the constitutional level challenges, I shall begin by categorizing the national legislative texts and the type of joint processing activities that are taking place in practice under the conceptual framework that I previously developed.
The April 2016 version of the Greek law stated that the Greek Asylum Service can be assisted by EASO experts and interpreters. This would seem to point to actions under assisted processing, that is, would include EASO experts independently undertaking actions that entail executive discretion. That version was compatible with the limitations upon EASO according to its current mandate, notably that it "shall have no powers in relation to the taking of decisions by Member States' asylum authorities on individual applications for international protection." 101 Nevertheless, the law was swiftly amended in June 2016 to state that deployed experts can conduct asylum interviews so as to reflect actual practice on the ground. As previously analyzed, conducting an admissibility interview independently entails undertaking actions and decisions that involve executive discretion. The administrative reality is that this moves clearly beyond assisted processing, to the realm of what I defined above as common processing. The June 2016 Greek legislation and its operationalization arguably give "powers in relation to the taking of decisions on individual applications" to the agency, in the very least indirect powers. 102 In this sense they exceed the legal limits under the EASO regulation. There is no CJEU case law on what a direct or an indirect power is in relation to taking a decision on an individual application for international protection. However, emitting an opinion, even a non-binding one, on an individual case, on the basis of an independently conducted interview arguably qualifies at least as an "indirect power." 103 The same holds true for the May 2018 amendments to the Greek law and subsequent practice regarding interviews on the merits of asylum claims. These interviews on the merits that are conducted independently entail the exercise of executive discretion. The Greek-speaking staff that EASO hires issue a non-binding opinion on the international protection needs of individual asylum seekers. These actions as well arguably qualify at least as an "indirect power in relation to the taking of decisions on individual applications." Therefore, EASO is acting outside the scope of its current mandate and thereby in violation of its Regulation. EASO's action is constrained by its founding Regulation, which constitutes secondary EU law, and the agency is not in a legal position to unilaterally alter its mandate. Only the co-legislators (the Council and the European Parliament) can do so through the adoption of subsequent EU legislation.
Turning to the challenges on a constitutional level, it is necessary to examine the compatibility of these practices with primary EU law. In relation to the specific provisions on the CEAS, EASO's involvement in Greece is compatible with Article 78(2)(e) TFEU, which foresees that "a Member State" is to be responsible for the examination of an application. The deployed experts are only formulating an opinion, which is not binding on the Greek Asylum Service according to law. It is  . 12, 2018). 101 See EASO Regulation art. 2(6). 102 See EASO Regulation rec. 14 and art. 2(6). 103 I first developed this argument in Tsourdi, supra note 12, at 1024. the Greek Asylum Service that formally adopts the admissibility decision, and it has the power to adopt a decision that goes against the proposal of the deployed experts. The Greek Asylum Service does not merely rubberstamp the non-binding EASO advisory opinions. For example, ECRE reports that: [A]s far as non-Syrian applicants undergoing admissibility procedures are concerned, EASO systematically recommends inadmissibility decisions based on the "safe third country" concept, subject to a few exceptions. The Asylum Service overturns these opinions as a matter of policy. 104 This means that existing practice in Greece is far from what I defined above as EU-level processing and therefore respects the limits posed by EU primary law relating to asylum processing.
Another question is the general compatibility of these practices with the case law of the CJEU, and notably the criteria developed in the Meroni 105 and Short Selling 106 case law on the constitutional limits to the delegation of powers of EU agencies. Analyzing the aforementioned and subsequent case-law, Chamon concludes that what is prohibited is essentially a form of discretion that allows an EU agency to develop policy on its own, that is, being empowered to make political, economical, or social policy choices. 107 The type of discretion that I have been referring to, i.e., executive discretion to decide, for example, whether an individual fulfils the criteria of the legal definition of a refugee, does not amount to the prohibited discretion of formulating policy as defined above.
Apart from constitutional level challenges relating mainly to the separation of powers between Members States and the EU in implementing policies, another set of challenges is of an administrative nature. Here, I am referring to the compatibility of these practices with procedural standards and guarantees enshrined in primary and secondary EU law. Relevant primary EU law is for example the EU Charter on Fundamental Rights (CFR) which enshrines procedural rights such as Article 41 CFR (right to good administration) and Article 47 CFR (right to an effective remedy). Relevant secondary EU law includes, for example, instruments of the EU acquis that establish procedural rights for asylum seekers, such as the 2013 Asylum Procedures Directive. The processes described are part of the asylum procedure, and applicants should enjoy the full array of rights foreseen by EU law no matter who is conducting the interview; the fact that the EU level is operational cannot lead to a diminution of procedural rights. On the ground, however, there is uncertainty as to the procedural rights available, with civil society organizations reporting shortcomings relating for example: to the manner of assessing vulnerability; the manner of conducting admissibility interviews; or the fact that advisory opinions on admissibility are issued in English and in practice not translated in Greek, and the fact that interviews are conducted in English, undermining the quality of legal representation by Greek lawyers. 108 104 The constitutional challenges, resulting from the overstepping by the agency of its mandate as established by its regulation, which excludes even indirect powers in relation to individual decision-making, can only be addressed through the adoption of an amended agency regulation addressing these points. Moreover, administrative integration requires a rethink of accountability processes so that it does not lead to a de facto watering down of procedural guarantees. In the next part, I provide my understanding on the concept of accountability while also sketching out the different accountability processes the current EASO Regulation foresees, and commenting on their effectiveness.
E. Holding EASO Accountable: An Ineffective Overload? Accountability has been characterized as "a chameleon-type term : : : that crops up everywhere performing all manner of analytical and rhetorical tasks and carrying most of the major burdens of democratic 'governance.'" 109 The EASO Regulation foresees several potential accountability procedures: judicial, financial, political, extra-judicial, and social. While the Regulation's provisions make explicit reference to the term accountability, no concrete definition is offered, hindering clarity. Thus, it is necessary first to resort to understandings of the term beyond the legal text. Thereafter, I provide a panoramic view of the potential accountability avenues that exist under EASO's founding Regulation (Section I). The reason I refer to them as "potential" accountability avenues is that in order to ascertain if a specific process is indeed an accountability process the study of its legal design is sometimes not conclusive and needs to be complemented by empirical data. A good example to illustrate this point is the case of social accountability processes. Thereafter, I critically assess the current arrangements, highlighting pitfalls and limitations. I base my remarks solely on the legal design of these processes (Section II). My study scrutinizes the operationalization of a specific accountability process, namely extra-judicial accountability, in the following part (Part F).

I. A Mosaic of Accountability Processes
Given the broad array of accountability processes envisaged in the Regulation, both judicial and non-judicial (i.e., political, financial, extra-judicial, and social) accountability processes are pertinent. Adopting the concept of accountability developed by Bovens, Goodin, and Schillemans, accountability may be conceptualized as "an institutional relation or arrangement in which an agent can be held to account by another agent or institution." 110 It consists of three elements or stages: i) The actor should be obliged to inform the forum about his or her conduct, by providing various sorts of information about the performance of tasks, about outcomes, or about procedures; ii) there needs to be a possibility for the forum to interrogate the actor and to question the adequacy of the explanation or the legitimacy of the conduct; and finally, iii) the forum may pass judgment on the conduct of the actor. 111 The mention of a "judgment" in this context should not be equated with a legally binding final pronouncement by a court or tribunal. Rather what is meant is the possibility of concrete consequences following the information provision and debate stages. Although it might seem minimalist, this understanding is broad enough to accommodate processes such as those envisaged in the EASO Regulation. Accountability in this sense operates ex post, as distinct from control which will often operate ex ante; this sense of accountability is also distinct from issues such as responsiveness and participation. 112 The EASO Regulation foresees, first, several political accountability processes: before the agency's own Management Board; 113 before the Council; 114 before the European Commission; 115 and finally before the European Parliament. 116 Next, the Regulation envisages judicial accountability through the Court of Justice, for example, an action for annulment against agency decisions relating to access to documents, 117 jurisdiction over instances of contractual liability, 118 as well as for instances of non-contractual liability, for example, an action for damages in relation to "any damage caused by its departments or by its staff in the performance of their duties." 119 While the case of EASO's involvement in asylum processing is not explicitly mentioned as it is arguably beyond its current mandate, it also falls within the remit of the CJEU. An intricate tango involving the Commission, the Council, the European Court of Auditors, and the European Parliament ensures financial accountability. 120 When it comes to social accountability processes, these are most arguably lacking. The agency's Consultative Forum is not conceptualized as an accountability forum; thus, the agency does not have to report, or explain its actions, to representatives of civil society. The main aim is one of dialogue, exchange of information, and pooling of knowledge. 121 Finally, the EASO Regulation contains a general provision on extra-judicial accountability stating that "[t]he activities of the Support Office shall be subject to the controls of the Ombudsman in accordance with Article 228 of the TFEU." 122 It also contains an additional reference to the European Ombudsman's (EO) mandate in relation to the specific issue of access to documents. 123 The EO has already been called to investigate complaints of maladministration in relation to EASO's involvement in asylum processing at the hotspots. The fact that this mechanism has been the main avenue for holding EASO accountable, despite the disappointing level of scrutiny it offered (Part F), reflects the main pitfalls of the current agency accountability approach.

II. EASO's Accountability Matrix: A Critical Analysis
When it comes to agency accountability, scholars have contested the value of the principal-agent framework of analysis, a predominant model in public management. 124 Deirdre Curtin aptly observed three concrete problems around this specific model: [F]irst, powers are not necessarily being delegated by the EU legislative power; second, there are (very) often multiple principals involved in setting up such agencies (including, in some instances, the European Parliament). Third, the tasks being "delegated" may be those of the Member States, not of the formal principals. 125 113 See, e.g., EASO Regulation art. 29. 114 See, e.g., EASO Regulation art. 31(3). 115 See, e.g., EASO Regulation arts. 7(1), 12(2). 116 See, e.g., EASO Regulation art. 30(1). 117 See EASO Regulation art. 42(3) and TFEU art. 263. 118 See EASO Regulation art. 45(2). 119 See EASO Regulation art. 45(3) and TFEU art. 340(2). 120 See EASO Regulation arts. 34 and 36. 121 See EASO Regulation art. 51(2). 122 See EASO Regulation art. 47. 123 See EASO Regulation art. 42(3). 124 Gailmard provides a succinct understanding of principal-agent models: "in principal-agent models, some actor (or group of actors) called an agent undertakes an action on behalf of another actor (or group of actors) called a principal. The principal, for its part, can make decisions that affect the incentives of the agent to take any of its various possible actions. Madalina Busuioc added further dimensions on agency accountability integrating empirical findings. She mapped out the abundance of accountability arrangements, creating an intertwined system. 126 While some of the ensuing dynamics were positive, with accountability processes reinforcing each other, some of the dynamics were negative, leading to accountability overload of relatively small agencies from the sheer volume of accountability obligations. 127 Apart from accountability overload, another complexity is that the imperatives for accountability and independence can sometimes conflict. 128 Accountability for fundamental rights violations raises specific challenges, bringing to sharp relief the position of the individual. In the specific case of EASO and its role in the processing of asylum claims, I identify two main pitfalls in the current accountability arrangements: The first is the conflict between accountability and independence, and the second is the accessibility for the individual.
The conflict between accountability and independence is most vividly portrayed through the accountability of the agency towards its Management Board. The EASO Regulation places great importance on the independence of the Director. It foresees that she should be independent in the performance of her duties and that she should "neither seek nor take instructions from any government or from any other body." 129 Therefore, independence is sought vis-à-vis both national governments and the EU institutions or bodies. However, the latter statement is conditioned on the premise that it is "[w]ithout prejudice to the powers of the Commission, the Management Board : : : " 130 Given the predominantly Member State composition of some of the accountability fora, such as the Management Board, 131 there is a delicate balance between "not taking instructions from any government" and respecting the prerogatives of the Management Board as an accountability holder.
It has been observed that "having all Member States represented at agency boards is in line with the conceptual understanding of the EU executive as an integrated administration and is an expression of the composite or shared character of the EU executive." 132 When the European level, through an EU agency, starts to be more implicated in policy implementation, including through the deployment of statutory staff and experts on the ground, Member States are understandably keen to have a strong say. The operational tasks undertaken are intrinsically linked with the implementation of the EU asylum policy and this legally rests with the Member States. Therefore, it cannot be concluded that the national level is seeking to "reappropriate powers" through the back door. At the same time, the independence challenge posed should not be underestimated.
The expansion of EASO's operational capacities, and its increasing role in implementing the EU asylum policy affecting the rights of individuals in the process, makes safeguarding of independence a more pressing issue. Research has illustrated the importance of administrative insulation for refugee status determination outcomes, and how the politics of deterrence can affect refugee status determination and strip away administrative insulation. 133 The supranational character of EASO means that there is also another danger. Given the distribution of power and political stakes in the field of asylum, EASO risks being captured by strong regulators and used as "proxy" to control weaker ones. 134 Indeed, understanding "national interest" in these fields as one-dimensional does not do justice to the divergence of interests between Member States, nor their power differential.
Accessibility for the individual is the second pitfall. It is clear from the governance of some of the accountability processes, for example, vis-à-vis the EU institutions, that there is only a very limited and indirect role for the individual to play. Judicial accountability avenues are, in principle, more accessible to the individual. However, under EU law the accessibility of judicial accountability is constrained by the strict rules on standing or criteria for ascertaining liability. For example, in order to ascertain noncontractual liability of EASO, an applicant would have to substantiate a sufficiently serious breach of a rule intended to confer rights on individuals; damage; and a causal relationship between the unlawful conduct and the victim's damage. 135 These elements would have to be substantiated in the context of a multi-actor situation, taking into account the non-binding nature of EASO's opinions. 136 Contesting the illegality of EASO's advisory opinions through an action for annulment before the CJEU is also not without its difficulties. 137 Before getting to the substantive arguments of the claimant, the conditions of admissibility for such type of action would need to be fulfilled. 138 Apart from the compliance with a specific time limit, these include the reviewability of the act and the standing of the applicant, where different rules apply according to whether an applicant is "privileged," "semi-privileged," or "non-privileged." The main contentious element would be whether the acts by EASO-deployed experts (i.e., the non-binding advisory opinions) fulfil the definition of a "reviewable act" for the purposes of an action for annulment, that is, whether they are "intended to have legal effects." 139 In my analysis above, I have posited that it is arguable that the advisory opinions produce such legal effects (hence, I found them to fall under what I termed acts that entail executive discretion). The opposite has also been argued, namely that "EASO's activities will not be considered a reviewable act before the EU courts, because they are not capable of affecting an individual's legal sphere, but constitute mere preparatory actions for the purposes of taking the final decisions on the admissibility of an asylum claim." 140 Given the shifting practices on the ground, and the de facto character of these developments from the perspective of EU law, it seems that this is a question that necessitates further empirical research in order to be answered fully.
A final avenue would be having indirect access to the CJEU, through a reference for a preliminary ruling by a national court. 141  and that the national court would refer a question on the validity of EASO's act. 144 The CJEU can review the validity of non-binding acts such as recommendations. 145 However, this indirect route means that the national courts need to make such a reference, which in practice is far from certain. It has also been argued that "because of the close cooperation of the national authorities and the EU agency, and the operation of EASO being based on purely de facto arrangements, it seems next to impossible to identify, pinpoint, and, consequently, subject to judicial control, EASO's contribution to the decision-making process." 146 This argument raises valid points, such as the de facto nature of EASO's involvement under EU law. However, as the complaints before the European Ombudsman reveal, it seems that EASO's contribution can be discerned from that of the Greek authorities, and in fact as will be analyzed in the next part of this Article, the Ombudsman has begun to adduce concrete procedural obligations for the agency in its cooperation with national authorities. These practical difficulties could well explain why to date the main means of holding EASO accountable for its role in asylum processing has been through an extra-judicial mechanism, that is, complaints before the European Ombudsman analyzed below.
F. Extra-Judicial Accountability in Practice: Fit for the Task?
The role of the European Ombudsman (EO) is complementary to that of the CJEU. The European Ombudsman's mandate concerns "instances of maladministration in the activities of the Union institutions, bodies, offices or agencies, with the exception of the Court of Justice of the European Union acting in its judicial role." 147 Broadly speaking, while the judiciary mainly deals with the legality of administrative decisions, the Ombudsman additionally assesses the conduct of administration in the broadest possible way. 148 This "life beyond legality" entails that the EO: is not merely advancing claims on behalf of the individual, but also aiming towards more accountable and responsive administrative procedures in a broader sense : : : [and] to encourage both parties to adjust their practices so that similar disputes do not arise in the future (and can leave the process with mutually beneficial outcomes). 149 Busuioc has eloquently summarized the potential benefits of this accountability avenue for agency action specifically: flexibility, accessibility, and a broad scope of review. 150 Namely, the Ombudsman model is not premised on a strictly adversarial logic, the rules of standing are not as strict as those applicable for the CJEU, and maladministration includes non-legal rules and principles that restrain administrative behavior even though they lack binding force. 151 144 In his in-depth study on the office of the EO, Vogiatzis has assessed that as an accountability mechanism the EO is related and has contributed to the EU's democratization. 152 He links this with the notions of openness, transparency, participation and efficiency. 153 While the language employed is not one of fundamental rights, this does not mean that the EO does not play a role in upholding the Union's fundamental rights obligations. In its 2018 annual report, the EO selfidentified complaints linked with issues of equality, non-discrimination, and the right to strike as directly concerning fundamental rights. 154 Nevertheless, fundamental rights feature also in other areas of activity of the EO, such as access to documents, which constitutes the bread and butter of her Office. In the same way, fundamental rights feature prominently in the complaints analyzed below which concern the role of EASO in national decision-making processes. The role of the EO is even broader, however, since maladministration is broader than illegality. In essence, "the Ombudsman expects of the EU institutions to abide by the principles of good administration even when acting lawfully. Thus, the Ombudsman's role goes beyond legality and therefore makes a substantial contribution to good governance, too." 155 In this sense, the EO also scrutinizes the agency's adherence to soft law, for example, such as its own internal guidelines.
While the EO's decisions are not legally binding, they are impactful. As Harlow and Rawlings observe, in some ways national ombudsmen's influence is more direct compared to courts and this is linked to their investigative method, access to files, and power to institute own initiative procedures. 156 All this holds true also for the European Ombudsman. The EU Ombudsman undoubtedly constitutes an accountability forum based on the broad definition of accountability adopted above. The agency (and any other Union institution, body, or office) needs to provide the EO with information and access to files, 157 the possibility for dialogue and questioning is foreseen (e.g., provision of reply by the relevant institution or body), 158 and the EO's recommendations and potentially reports to the European Parliament 159 fulfil the requirement of concrete consequences following the information-provision and debate stages, even if they lack legally binding force.
A parallel evolution is that of the emergence of internal extra-judicial accountability mechanisms that has gradually taken place in Frontex. Namely, consecutive amendments to the Frontex Regulation have led to the development of novel fundamental rights oversight mechanisms, such as an independent Fundamental Rights Officer that is responsible for the running of ombudsman-type processes taking the form of an "individual complaints mechanism." 160 The mandate of the Fundamental Rights Office was further strengthened in the recent November 2019 amendments to the founding Regulation of that agency, thanks to the enhancement of its capacities and the creation of the function of fundamental rights monitors. 161 While the effectiveness of these mechanisms will be tested in the years to come, they hold the promise of meaningfully complementing as well as interacting with judicial accountability processes at both national and EU levels.

I. EASO's Involvement in Asylum Decision-Making Before the European Ombudsman
As mentioned at the outset of this contribution, two civil society organizations have activated extra-judicial accountability: the ECCHR and Advocates International. I critically assess the European Ombudsman's decision in these two instances.
The ECCHR, a German civil society organization, had raised both constitutional and administrative aspects in their complaint before the EU Ombudsman. Factually the complaint targeted EASO's role in asylum decision-making at the Greek hotspots. 162 They argued, firstly, that EASO acts outside of its mandate under EU law through its participation in asylum processing in Greek hotspots. Next, the organization argued that when conducting such interviews, EASO fails to comply with the provisions on "the right to be heard" in the Charter of Fundamental Rights (Article 41), as well as EASO's own guidelines. The Ombudsman's reasoning on both limbs of the complaint leaves a lot to be desired.
To substantiate the constitutional level of its complaint, the ECCHR referred both to the limitations of EASO's mandate as set out in its founding Regulation, 163 and in the practice on the ground to argue that EASO's involvement amounted at least to an indirect decision-making power on the admissibility of claims and on the vulnerability of individual applicants. While it noted that "the ultimate responsibility for decisions on asylum applications rests with the Greek authorities," 164 the EU Ombudsman nevertheless found that "EASO is being encouraged politically to act in a way which is, arguably, not in line with its existing statutory role." 165 Therefore, the EU Ombudsman essentially held that EASO through its involvement in asylum processing in Greek hotspots is acting ultra vires. Despite this position, the Ombudsman swiftly brushed aside the incompatibility of EASO's actions with its mandate. She next opined that "it is likely that EASO's founding Regulation will be amended in the near future to provide explicitly for the type of activity in which EASO is currently engaged, thus resolving the issue of EASO possibly operating outside of its statutory brief." 166 This is an unsatisfactory reasoning in the sense that it falls short of upholding the requirements of the legality of action of EU institutions and bodies, and the principle of the rule of law. It is tenuous for the Ombudsman to discontinue inquiries based on a future legislative amendment to redress the illegality of action of an EU body. Firstly, the timing of the adoption of a given legislative instrument is unpredictable. In the case at hand, the Ombudsman rendered her decision in July 2018 and one and a half years later, in December 2019, there is still no firm political agreement between the European Parliament and the Council on the adoption of the amended EASO Regulation that has been negotiated since May 2016. Secondly, apart from the timing, also the final content of a legislative instrument is unpredictable. The fact that the Commission proposed amendments to the agency Regulation moving in this direction does not guarantee that they will survive the negotiation procedures between the two co-legislators. Betting on the future redress of the situation through envisaged amendments in ongoing negotiations in a politically sensitive area such as asylum is a rather risky gamble.
The second limb of the complaint related to a violation of applicants' procedural rights, and namely those guaranteed under Article 41 of the EU Charter (right to good administration). The complainant referred to non-compatibility with the principle of fairness and lack of transparency drawing from practice and invoking also EASO's own guidelines on the use of interview techniques, especially in what concerns the identification of vulnerability. The Ombudsman found that "EASO has made considerable efforts to improve its practices in those areas highlighted 162 While EASO also undertakes interviews on the merits of asylum claims after the legislative amendment of national Greek administrative law in May 2018, this was a later development that did not form part of the complaint to the EU Ombudsman. 163 See my analysis, supra Part D, Section II, as well as my previous analysis in Tsourdi, supra note 12, at 1024. 164 European Ombudsman, Case 735/2017/MDC, supra note 13, para. 32. 165 Id. at para. 33. 166 Id. at para. 34. by the complainant." 167 Still, she accepted that "there are genuine concerns about the quality of the admissibility interviews as well as about the procedural fairness of how they are conducted." 168 Nevertheless, she placed great emphasis on the fact that EASO's opinions are advisory and that the Greek Asylum Service can disagree with EASO's expert opinions. 169 In addition, the Ombudsman referred to the possibility of an appeal before the Greek Appeals Committees to redress procedural faults. 170 It is undeniable that EASO has taken steps to improve the quality of its processing, such as a multi-layered quality assurance system made up of a team leader and a helpdesk and containing training and case sampling components. 171 Moreover, it is true that the Ombudsman is best placed to identify areas of systemic improvement rather than act as a de facto appeals body for shortcoming in each individual case. Nevertheless, the Ombudsman's reasoning fails to take into account the capacity challenges that Greek national authorities face, which prompted the EASO deployments in the first place. Of course, it remains open to the Greek Asylum Service to overturn the opinions of EASO experts. The Greek Asylum Service has done so as a matter of policy where the agency experts propose findings of inadmissibility (that would subsequently lead to returns to Turkey) for non-Syrian nationals that will not find adequate protection in that country. Here the indicator of the incompatibility of EASO's findings with international and regional human rights law is easy to set out (nationality of the individual asylum seeker). However, it remains doubtful whether the overloaded Greek administrative authorities have the capacity to thoroughly scrutinize each interview. After all, it was systemic issues of diminished capacities in relation to the numbers of existing asylum seekers that led to the deployments of EASO experts in the first place, whose presence will have little impact if they end up generating more work for the Greek authorities than assistance. Thus, the quality of decision-making that EASO offers is crucial. In addition, the Ombudsman seems not to fully acknowledge the complexities of mixed proceedings and the practical difficulties for the individual asylum seeker to contest the quality of preparatory acts by an EU body before national appeals committees. EASO's role in fact calls for a broader rethink of EU procedural asylum law to fully account for this new administrative landscape.
The second complaint concerned the same factual background, that is, EASO's involvement in the admissibility interviews. Nonetheless, the complainant did not raise arguments relating to the constitutional aspect of EASO's involvement in asylum processing but rather raised violations of procedural rights and errors during the interview phase with EASO of an asylum seeker that was consequently deported based on a decision of the Greek Asylum Service and feared dead. As mentioned above, procedural rights are intrinsically linked with the absolute prohibition of refoulement, as the rejection of (or refusal to process) an application could lead to the return of the individual to third countries or the country of origin. The important issue was that during an internal screening as part of its own quality procedures, EASO had uncovered the errors of judgment during the interview and the problems with the interpreter. 172 It had not, however, alerted the Greek authorities about these procedural irregularities. EASO argued that its stance was part of its effort not to overstep its mandate and not to influence the final decision of the Greek Asylum Service. 173 Thus, the aim of this internal review was exclusively geared to improve the quality of decision-making and not to intervene in "open cases." 174 167 Id. at para. 45. 168 Id. at para. 46. 169 Id.
EASO's reasoning is not convincing. Were the agency adamant about not overstepping its legal mandate, it should refuse to be involved even indirectly in individual decision-making. As it has chosen to participate in decision-making by independently conducting actions that entail executive discretion, it cannot argue that its functions continue to remain purely supportive. As the Ombudsman correctly noted in her observations, despite their non-binding nature, these opinions can help inform the final decision of the Greek authorities, and the errors identified by EASO may have contributed to the fact that Mr. X was deported. 175 Where errors are identified, EASO should do its utmost to remedy them. Even if EASO does not have the power to ask the national authorities to take specific action, it has according to the European Ombudsman an obligation to notify the national authorities of these errors; thereafter national authorities can assess the gravity of the errors and decide whether any remedial action needs to be taken. 176 Therefore, in this case, the Ombudsman found that the serious errors in the case of the asylum seeker who was later deported constituted maladministration.
In this second case, the Ombudsman seems to have become more aware of the intricacies that mixed administrative proceedings raise. While she reiterated her point that individual shortcomings can better be addressed in appeals, 177 she sought to establish specific procedural guarantees, such as the obligation to notify national authorities of identified errors, or the creation of an individual complaints mechanism to ensure procedural fundamental rights in the context of mixed proceedings. 178 Such an individual complaints mechanism would in fact be a novel internal ombuds-type procedure along the lines of the mechanism that already exists as part of the functioning of Frontex. It was added at the insistence of the European Parliament as part of a revamped EU Agency on Asylum; 179 nonetheless the fate of these negotiations is uncertain, and for now the agency lacks the concrete legal basis, as well as the dedicated resources to establish such an internal extra-judicial accountability mechanism.
If it is to be established, it will be important that the agency will be well resourced to run this mechanism (in particular, with ear-marked funding for that purpose). In addition, complaints should lead to concrete consequences, including review of practice by the agency and suspension or potential removal from the pool of experts or agency pay roll of specific individuals where they are found to be systematically underperforming or conducting procedural violations. More important even will be the combination with the obligation of EASO to alert national authorities of procedural errors it has identified.

G. Conclusion
The implementation modes of the EU asylum policy have significantly shifted since its inception. Institutionalization of practical cooperation through EASO, an EU agency, has incrementally led to the emergence of patterns of joint implementation through the joint processing of asylum applications. Joint implementation patterns and the augmentation of the financial and human resources available to EU agencies could act as precursors to deeper forms of integration, eventually leading to a full "Europeanization" of this policy's implementation modes. This should not be specifically linked with political aspirations of an increasingly "federalized" EU, but rather could be viewed as a pragmatic approach to effectively implementing a policy that arguably leads to the provision of regional public goods, here, asylum. 175 Id. at paras. 13-14. 176 Id. at para. 16. 177 Id. at para. 28. 178 Id. at para. 29. 179 See, e.g., Proposal Council of The European Union for a Regulation of the European Parliament and of the Council on the European Union Agency for Asylum and Repealing Regulation (EU) 439/2010 on State of Play and Guidance for Further Work, amendments 30, 92 (June 27, 2017), https://data.consilium.europa.eu/doc/document/ST-9563-2017-COR-1/en/pdf. The first experiences with joint processing in Greece illustrate that enhanced administrative integration should not be celebrated as something that is inherently positive. Administrative integration brings its own challenges, of both a constitutional and practical nature, and requires a rethink of accountability processes so that it does not lead to a de facto watering down of procedural guarantees. EASO is currently subject to a mosaic of accountability processes. Two main pitfalls emerge through the current arrangements. First, the intricate balance between accountability and independence: Current accountability processes could render EASO hostage to Member State political influence, mainly through the workings of its Member State dominated Management Board. The second pitfall is accessibility: Judicial accountability at EU level remains largely inaccessible to individual asylum seekers due to the strict rules on admissibility, standing, and establishment of liability. Other processes fall short of being genuine accountability processes, as they remain at the level of information exchange.
Extra-judicial accountability emerges as a promising avenue to ensure procedural fundamental rights compliance. The European Ombudsman seems to be increasingly grappling with the complexities that mixed processing raises. Her scrutiny of the first such complaint was arguably superficial, brushing aside the constitutional-level challenges based on uncertain future legislative amendments and placing the entire burden of safeguarding procedural standards on the under-resourced Greek administrative authorities that EASO experts have been deployed to assist. In the treatment of the second complaint though, the Ombudsman recommended the adoption of concrete procedural standards, such as notification obligations to national authorities of errors identified by the agency, and the establishment by EASO of an internal individual complaints mechanism. In this sense, extra-judicial accountability through the EO, combined with the envisaged internal "individual complaints mechanism" within EASO, could go some way in ensuring the respect of applicants' procedural rights in this brave new administrative landscape.