Introduction
As far as the EU’s executive power is concerned, its institutional development is subject to an increasing diversification. Apart from the empowerment of institutions, in particular the Commission, or bodies closely attached to them, or the empowerment of decentralised European agencies, a third route has emerged: the establishment of independent or otherwise atypical entities which are located within institutions. It is the aim of this blogpost to shed light on the organisational set-up of these latter entities and to explore potential parallels to the governance structure of agencies, notably as regards the engagement of relevant Member State authorities. Taking five examples relating, respectively, to the Commission, the European Parliament (EP) and the European Central Bank (ECB), it will also be addressed which role the ‘agency solution’ – i.e. the political choice in favour of empowering a decentralised agency instead of entrusting an institution – has played during the respective law-making process and, partially, even thereafter. At the end, some of the constitutional questions of EU law this development raises will be pointed out.
The EU’s agencification – an unbroken trend
With a considerable number of new agencies being established or existing ones being reinforced in recent years, the long-term trend of the EU’s agencification has not lost its swing. The past few years have seen the establishment of the European Labour Authority (ELA), the EU Agency for Asylum (EUAA), the Authority for Anti-Money Laundering and Countering the Financing of Terrorism (AMLA) and the EU Customs Authority, as well as the reinforcement of, e.g., the Agency for the Cooperation of Energy Regulators (ACER), the Financial Market Supervisory Authorities (ESAs), and, imminently, the European Chemicals Agency (ECHA). The on-going political discussions on the creation of an EU Centre on Child Sexual Abuse in the form of an agency further testify the unbroken attractiveness of agencies.
Institutional alternatives I: Institutions and their ‘extended arms’
The establishment of a new or the reinforcement of an existing agency is not without alternative, though. When the EU initiates or extends its executive activity in a policy area, it may as well opt for a different institutional design. It is (nearly) always feasible to designate the Commission as the traditional administrative authority of the EU when a new/reformed regime of Union law is set in place. Given the limits to the delegation of powers from the institutions to other bodies the CJEU has drawn in its Meroni doctrine, it may even be legally required to empower an institution like the Commission, either on its own or by allowing it to intervene in decisions principally devised by an agency.
The Commission is regularly vested with new tasks and powers. A recent example for this is the Foreign Subsidies Regulation which provides rules notably with regard to concentrations (M&A) and public procurement procedures. The enforcement of these rules is upon the Commission. As a consequence, the Commission, making use of its right to self-organisation, changed its internal institutional set-up. The Task Force in the International Relations Unit of DG COMP, initially entrusted with the enforcement of the Foreign Subsidies Regulation, in March 2024 was replaced by a newly established Directorate K (Foreign Subsidies) in DG COMP, which is now responsible for addressing foreign subsidies in the context of concentrations and beyond (excluding public procurement procedures). As regards public procurement procedures, since May 2025, a Unit (Foreign Subsidies) in the newly established Directorate G (Compliance and Responsible Business Conduct) in DG GROW has been in charge.
The institutional adaptation to new tasks and powers may – and has done so in the past – as well involve the creation of advisory committees or of executive agencies. This group of entities is, by design, subordinate to or otherwise dependent on the respective institution, normally that is the Commission. Executive agencies, for example, to which the Commission ‘[o]utsourc[ed] certain management tasks’ (Recital 4 of the Executive Agencies Regulation), are heavily influenced by the Commission DG in charge and operate merely as its ‘extended arm’. Thus, they are very different from the (decentralised) agencies referred to before.
Institutional alternatives II: independent/atypical entities within institutions
Apart from these two options – either to empower an institution (or a satellite body) or to put an existing or newly established agency in charge of the implementation of a new/reformed regulatory framework – the EU has repeatedly gone a middle way. By establishing a new independent or otherwise atypical entity within an institution. These entities organisationally belong to the respective institution or agency but remain outside its regular internal structure. Whereas somewhat similar developments have taken place also within the institutional set-up of agencies (note the example of the European Scientific Advisory Board on Climate Change anchored within the European Environmental Agency), the exemplary sample provided here is limited to developments within EU institutions.
An early example is the European Anti-Fraud Office (OLAF). Established as a DG of the Commission, it is independent of the Commission when it comes to OLAF’s investigations (Art 3 of the OLAF Decision). The original plan to establish OLAF as a self-standing entity with legal personality was dropped, inter alia due to Meroni concerns. However, OLAF’s governance has changed over time. Since 2013, it has been adapted so as to include anti-fraud bodies of the Member States and its independence has been strengthened.
In 2010, the European Systemic Risk Board (ESRB) was established as an independent entity without legal personality, seated in Frankfurt/Main. It disposes of close personal and organisational links to the ECB. The ESRB is chaired by the ECB President. The General Board – the ESRB’s decision-making body – inter alia comprises the Governors of the national central banks (as voting members) and high-level representatives of the relevant national (supervisory) authorities (as non-voting members). The ECB ensures a Secretariat of the ESRB and thereby provides analytical, statistical, logistical and administrative support to the ESRB. The human and financial resources for the Secretariat are provided by the ECB as well, and it also appoints, in consultation with the General Board of the ESRB, the head of the Secretariat.
The Authority for European political parties and European political foundations (APPF) was created in 2014 and is in charge of registering, controlling and sanctioning European political parties and European political foundations. The APPF has legal personality and is independent, and it qualifies as ‘body of the Union within the meaning of Article 263 of the TFEU’ (Art 8 para 2 and Recital 10 of the APPF Regulation). The Authority is headed by a Director, taking its decisions. In addition to that, a committee of independent eminent persons advises the APPF on certain matters (Art 16 of the APPF Regulation). The APPF is not only ‘physically located’ in the EP, the latter also provides the ‘necessary offices and administrative support facilities’ (Art 8 para 4 of the APPF Regulation). The appropriations to the APPF emanate from the general budget of the EP.
One of the EU’s institutional responses to the COVID-19 pandemic was the creation of the Health Emergency Preparedness and Response Authority (HERA) in 2021. Originally envisaged as an agency, HERA was eventually situated, qua Commission Decision, within the Commission. Even though it is named ‘Authority’, thereby insinuating a self-standing existence akin to decentralised EU agencies, in fact it is established as a Directorate-General of the Commission, divided in five Units (but no Directorates). HERA’s Head is ‘ranked as’ Director-General, its Deputy Head is ‘ranked as’ Director (Art 4 of the HERA Decision). HERA’s internal structure is very peculiar and different from regular Commission DGs. It disposes of a Coordination Committee, a Board and an Advisory Forum. The Coordination Committee, composed of the Commissioners in charge of health, the internal market, innovation and research and crisis management ‘shall provide political steering on the planning and implementation of HERA’s tasks, without prejudice to the prerogatives of the College of Commissioners’ (Art 5 of the HERA Decision). HERA’s Head and the Director General in charge of health and food safety shall attend the meetings ex officio. Its Board, chaired by HERA’s Head, is, first and foremost, composed of ‘one high-level representative from each Member State’. It shall ‘assist and advise’ the Commission and deliver opinions inter alia on activities HERA may have to implement and new tasks which may be assigned to HERA (Art 6 paras 4 and 5 of the HERA Decision). The Advisory Forum is composed of ‘members from technically competent bodies designated by each Member State’ (Art 7 para 2 of the HERA Decision).
The last example provided here is the institutional set-up of the AI Act adopted in 2024. In anticipation of the adoption of this act, the Commission established the AI Office as a Directorate in the Commission’s DG CONNECT. In addition to that, a European Artificial Intelligence Board was created to ensure the representation of Member States and to give advice on various aspects of the implementation of the AI Act and on shaping AI policy more generally, both at the EU and at the national level.
Institutional diversification and the potency of the ‘agency model’
The examples given above suggest a rather diversified landscape of intra-institutional alternatives to the traditional empowerment of institutions or to the empowerment of agencies. Whereas use cases from the ECB and the EP have been purposely included in the exemplary sample, these independent or otherwise atypical entities seem to be established primarily within the Commission. However, when new powers are to be exercised by the EU’s executive branch – to the extent they are not, qua Meroni, reserved to the institutions –, the ‘agency model’ is still the standard of institutional reform. For one thing, this is proven by the persisting establishment of new, and the reinforcement of existing, agencies up until today (see above). Secondly, the establishment of an agency is regularly on the table in the political discussion, even if eventually an alternative route is chosen. Thirdly, agencies seem to serve as a role model for the governance structure of these specific intra-institutional entities.
The prominence of the ‘agency solution’ is most visible in the case of HERA. The reasons why HERA was not set up as an agency first and foremost were of a temporal nature. In the midst of the COVID-19 pandemic, it was the political aim to establish a ‘flexible structure’ (Commission Communication on HERA) which can start operations and contribute to fighting the pandemic as soon as possible. A legislative procedure required for the establishment of an agency and the subsequent build-up phase of the agency arguably would have taken too much time. HERA’s establishment as part of the Commission, on the contrary, allowed it to benefit from the existing infrastructure and expertise.
The idea to establish HERA as an agency, however, was given up only preliminarily. The Council requested a review of the idea to establish HERA as a ‘distinct entity’, by 2024 at the latest (Art 16 of the Public Health Emergency Regulation). As part of this review, the pros and cons of the transformation of HERA into an agency were discussed in a differentiated manner. The final report does not recommend this transformation. Accordingly, such transformation would require political balancing, as both alternatives – leaving HERA as a special DG within the Commission or establishing it as an agency – come with advantages and disadvantages.
But also other examples give proof of the relevance of agencies. OLAF’s internal set-up has been amended so that today its internal structure is much more akin to that of an agency than when OLAF was created in 1999. And for the institutional set-up relating to the AI Act, an ‘agency solution’ still remains on the table. In spite of the institutional alternative the legislator has preliminarily opted for, it requested the Commission, by August 2029 and every four years thereafter, to draw up a report, including ‘an assessment with regard to […] the possible need for a Union agency to resolve any identified shortcomings’ (Art 112 para 3 of the AI Act).
The ESRB resembles an agency in that its Governing Board comprises, apart from representatives of various EU institutions and bodies, representatives of national authorities (for the Supervisory Board and its partial similarity to agencies see here). This decentralising element of representation of Member States has been part of the EU’s institutional DNA ever since, it is true, but the representation of specific (independent) national authorities has been particularly prominent within agencies.
The APPF seems to be an outlier here. The Commission proposal initiating the legislative proceedings eventually resulting in the original APPF Regulation did not make provision for the establishment of an independent authority within the EP, but saw the EP itself at the centre of implementation of the legal framework. It was the EP which made the suggestion to establish the APPF in its position at first reading, and several trilogue meetings were required to agree on this. Apparently, the establishment of an agency or the introduction of an agency-like governance structure has not been discussed. The EP’s institutional purpose, i.e. to represent the citizens of the EU (not its Member States), may have been the main reason for that. Furthermore, the small scale at which the APPF with currently only 14 employees operates (see Annual Activity Report 2025) does not speak in favour of a more encompassing governance structure.
The organisational set-up of most of the intra-institutional models presented here has a decentralising effect, even though the empowerment of EU institutions prima facie appears as an instance of centralisation. This broader institutional development raises numerous questions of EU constitutional law. While the Treaties allow for the institutions’ internal self-organisation, there are limits to deviating from the top-down decision-making process within these institutions. In an intra-institutional set-up, the delegation of powers to an independent body may be less visible, but the underlying problems are the same. As regards, for example, the Commission, this development affects its independence from the Member States, but also the principle of collegiality which – at least in its political dimension – is under pressure anyway by the more President-focused leadership in recent years. As regards the relationship between institutions, the rise of intra-institutional bodies, which are at the same time granted (some) independence from their respective host institution, challenges the EU’s inter-institutional balance, as struck by primary law, in a way similar to decentralised agencies. However, tackling these questions would go beyond the scope of this blogpost, whose purpose is instead to raise awareness of the different routes of institutional empowerment which coexist in the shadow of the still dominant ‘agency solution’.
Paul Weismann is Associate Professor of EU law at the Faculty of Law and Economics of the University of Salzburg and at the Salzburg Centre of European Union Studies (SCEUS). He studied law and history at the Universities of Vienna and Salzburg and at the London School of Economics and Political Science. He has taught and researched at various universities and research institutes and has published widely on EU law, notably on institutional issues (including European agencies), EU soft law and EU economic law (in particular Banking Union, EMU and State aid law).