EU Policy Enforcement and Anti-Money Laundering Overhaul in the EU – A New Analytical Approach

EU Policy Enforcement and Anti-Money Laundering Overhaul in the EU – A New Analytical Approach Research Paper No. 2 - 2025 Authors Christy Ann Petit and Nikita Divissenko Dr. Christy Ann Petit is an Assistant Professor at the School of Law and Government of Dublin City University (DCU) and Deputy Director of the DCU Brexit Institute. She is the Director of the Anti-Money Laundering Academy (2025) for the Florence School of Banking and Finance (FBF) at the European University Institute (EUI). Dr. Nikita Divissenko is an Assistant Professor at the International and European Law Department (IER) of the Utrecht University School of Law. His research focuses on the intersection of digital finance, regulation and technological Innovation. [email protected] [email protected] 1 EU policy enforcement and Anti-Money Laundering overhaul in the EU – a new analytical approach Christy Ann Petit and Nikita Divissenko Abstract This paper examines the revised enforcement system for anti-money laundering (AML)in the EU, focusing on the creation of a new EU agency – the Anti-Money Laundering Authority (AMLA). The paper asks how the reformed framework ensures a more consistent and efficient AML enforcement in the EU, compared with prior regimes. The literature primarily examines agencification in the EU through the dimensions of centralisation/decentralization and direct/indirect enforcement powers, however, this analytical lens is incomplete for the revised AML enforcement system. We therefore devise a third dimension for analysing enforcement that relies on operational and collaborative integration throughout the system. This third dimension highlights specific features, namely, how agencification in a common system enables a shared operational responsibility, steered by legal and institutional integration, common methodologies and practices. These features bind the various actors together, moving beyond a network model based on coordination towards a more harmonised and cooperative system of enforcement. The three-dimensional (3D) model serves as an analytical approach for studying the EU enforcement regime for AML, one that could also be used in other areas of EU policy enforcement. Keywords Agencification, anti-money laundering, AML Authority (AMLA), enforcement, supervision. Acknowledgment s The authors wish to thank Larissa Campos for her support in visual designing, as well as Mira Scholten and Lisette Mustert for their thoughtful feedback on an earlier version. All errors remain the authors’ responsibility. I . Introduction The creation of the new European Union (EU) agency – the Authority for Anti-Money Laundering and Countering the Financing of Terrorism (AMLA) – offers a unique opportunity to revisit and examine the approach to EU enforcement, in particular the long-standing 2 phenomenon of agencification. 1 At the first glance the establishment of AMLA appears to constitute (yet another) instance of ‘Europeanisation’ of enforcement in an EU policy area in which the enforcement regime previously relied primarily on national authorities and has proven to be ineffective. In its 2019 report, the European Commission identified several shortcomings and problem drivers stemming from the AML decentralised enforcement mechanism. 2 The report that underpinned the 2021 AML package 3 was based on 10 case studies across eleven Member States between 2012 and 2018. 4 A more intertwined approach between EU and national level had been missing despite its clear necessity and was considered too bold a move only a decade ago when EU legislators undertook the reform from AMLD4 to AMLD5 and in the context of the European Supervisory Authorities (ESAs) Review. 5 We may wonder what is special about that new EU agency set up by the co-legislators, considering the current policymakers’ emphasis on the groundbreaking nature of the creation of AMLA, and compared to the ESAs. 6 The novelty may be linked to the specific sector (AML) itself which requires the continuous sharing of intelligence and information, as well as the interdependence of actors in the system. The powers allocated to AMLA and the supervisory authorities in the system aim at ensuring ‘consistent high-quality application of the AML/CFT supervisory methodology and promot[ing] efficient cooperation between all relevant competent authorities.’ 7 In this regard, the co-legislators agreed, through AMLA Regulation, to constitute an ‘AML/CFT Supervisory System’ 8 designed with both AMLA and the AML/CFT supervisory authorities, which, we argue is contributing to further operational and collaborative integration. 1 Mira Scholten, Marloes van Rijsbergen, (2014) The Limits of Agencification in the European Union, 15(7) German Law Journal,1223-1255; see, also, David Levi-Faur, (2011) Regulatory networks and regulatory agencification: towards a Single European Regulatory Space, 18(6) Journal of European Public Policy, 810-829. 2 European Commission, ‘Report from the Commission to the European Parliament and the Council on the Assessment of Recent Alleged Money Laundering Cases Involving EU Credit Institutions’ (2019) COM(2019) 373 final 23 <https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52019DC0373> accessed 3 June 2025; European Commission, ‘Communication from the Commission to the European Parliament and the Council, Towards Better Implementation of the EU’s Anti-Money Laundering and Countering the Financing of Terrorism Framework (COM(2019) 360 Final); European Commission, ‘Impact Assessment Accompanying the Anti-Money Laundering Package’ [2021] SWD(2021) 190 final 13–19. 3 Regulation (EU) 2024/1620 of the European Parliament and of the Council of 31 May 2024 establishing the AntiMoney Laundering and Counter-Terrorist Financing Authority (AMLA); Regulation (EU) 2024/1624 of the European Parliament and of the Council of 31 May 2024 on the prevention of the use of the financial system for the purposes of money laundering or terrorist financing (AMLR); Directive (EU) 2024/1640 of the European Parliament and of the Council of 31 May 2024 on mechanisms to be implemented by Member States to prevent the use of the financial system for the purposes of money laundering or terrorist financing (AMLD 6). 4 The Commission’s 2019 report on regulatory and supervisory fragmentation in AML/CFT highlighted major divergences in the application of the prior AML/CFT regulatory and supervisor framework raised a structural problem exhibiting the incapacity of the Union to protect the integrity of its financial system. This constituted a strong inadequacy of the AML/CFT enforcement regime, particularly seen in the context of the Single Supervisory Mechanism in place for nearly a decade. See, European Commission, ‘Communication from the Commission to the European Parliament and the Council, Towards Better Implementation of the EU’s Anti-Money Laundering and Countering the Financing of Terrorism Framework (COM(2019) 360 Final)’. 5 The revision of the 4th AML Directive (Directive (EU) 2015/849) was initiated in 2013 with the 5th AML Directive (Directive (EU) 2018/843) adopted in 2018. 6 The Joint Statement and Common Approach on decentralised agencies is still a reference point. See Joint Statement and Common Approach of the European Parliament, the Council of the European Union and the European Commission of 19 July 2012 on decentralised agencies. See Recital 6 of AML Authority Regulation. 7 Recital 16, AMLA Regulation. 8 Section 2, AMLA Regulation. 3 Therefore, we consider that the novelty of the revised AML/CFT enforcement regime and its integrated system go beyond a simple further centralisation of AML/CFT supervision and direct enforcement, transcending a binary approach. Instead, the regime is illustrative of a more significant transformation which combines and integrates – within a 3D enforcement model that we put forward – the ‘binary’ approaches to centralisation and decentralisation, and to direct and indirect vectors of EU enforcement, together with a third dimension of operational and collaborative integration. Thereby, we offer an original way of examining EU enforcement building on scholarship on agencification and EU policy enforcement, and answer the main question: how does the reformed AML framework ensure more consistent and efficient enforcement in the EU? To this end, we examine the powers and tasks as attributed and shared within the AML enforcement system. This third dimension in EU enforcement also offers a useful analytical tool for legal scholars and policy-makers aiming to study EU enforcement regimes considering its potential replicability in other policy areas. The third dimension embodies the essential contributions and roles of joint supervision teams and analyses, peer coordination, and shared intelligence across the AML system. 9 This essential dimension is legally structured and institutionalised under the revised AML regime with a deterrence and punitive function, some discipling device and common methodologies and supervisory approaches for the whole system. We conceptualise this third dimension as a component of a three-dimensional (3D) structure of the AML enforcement system in the EU – besides the (in)direct enforcement and the (de)centralisation dimensions. As such, we argue that this additional dimension shows how agencification in a common system allows moving along from mere coordination to operational and collaborative integration. Such operational integration not only takes place at agency level but also within the system, because all actors are equally involved in common endeavours, with a shared responsibility, and steered by common methodologies. These ‘third dimensional’ features of the revised AML enforcement system create and sustain a connective tissue that ties together the different parts in the enforcement system. In this regard, the new AML integrated system goes way beyond a network-based enforcement model – which would have limited capacity to enforce decisions primarily due to its informality. 10 The third dimension thus constitutes an organisational dimension aimed at fostering a harmonised and cooperative system of enforcement. The paper is structured as follows. First, we offer (II) an overview of the agencification trend and the conceptualisation of EU enforcement in the literature. We then offer (III) a detailed examination of the powers and tasks under the reformed AML enforcement system. We suggest (IV) a new way to conceptualise enforcement in the EU as a 3D structure and highlight 9 Recitals 16 and 27, and Article 7, AMLA Regulation. 10 See discussions in section II. 4 the benefits of this analytical tool to better comprehend EU enforcement in a given EU policy area. The last section (V) concludes. II . Conceptual approach es to EU enforcement Scholarship on enforcement in EU Law has developed various approaches to describe and examine enforcement regimes. This section highlights the main types of EU enforcement models (1.) and discusses key vectors that help to analyse and better comprehend enforcement in the EU (2.). 1. The types of enforcement models in EU Law This subsection gives a brief overview of the more recent literature on EU enforcement models. Scholten, and later Van Kreij, distinguish between the three main modes or regimes of enforcement at the disposal of the EU. 11 The first and the most conventional or default regime, is that where the EU relies on the Member State’s enforcement authorities to implement and enforce EU policies. As such, this regime is decentralised (no single authority but multiple authorities) and is also indirect, meaning that the EU enforcement authorities or institutions do not enforce EU policies directly against citizens or entities. 12 The second type of the enforcement regime is one centred around enforcement networks. 13 These consist of national enforcement agencies, but nodes of these networks may, and often do, comprise the European Commission or specialised EU agencies - although with limited direct enforcement powers. The third type is the regime with enforcement responsibilities vested in an EU agency. This regime usually implies indirect and/or direct powers given to an EU agency (or institution), where the latter oversees and coordinates enforcement by Member State authorities, or directly enforces EU law against citizens or entities. 14 Van Kreij aptly notes, however, that the distinctions between the three enforcement regimes are often ‘blurred in practice’, and often the regimes coexist. 15 For one, these regimes are not always limited to enforcement tasks alone, and include other functions, such as coordination, but also advisory or rulemaking functions. Additionally, in many cases regimes coexist within 11 Miroslava Scholten, (2017) “Mind the trend! Enforcement of EU law has been moving to ‘Brussels’”, 24(9) Journal of European Public Policy , 1348-1366; Laurens Van Kreij, (2019) “Towards a Comprehensive Framework for Understanding EU Enforcement Regimes”, 10(3) European Journal of Risk Regulation , 439-457. 12 Laurens Van Kreij (n 11), 441. 13 Keith G Provan and Robin H Lemaire, ‘Core Concepts and Key Ideas for Understanding Public Sector Organizational Networks: Using Research to Inform Scholarship and Practice’ (2012) 72 Public Administration Review 638; Keith G Provan and Patrick Kenis, ‘Modes of Network Governance: Structure, Management, and Effectiveness’ (2008) 18 Journal of Public Administration Research and Theory 229; Paul Craig, Shared Administration and Networks: Global and EU Perspectives (Hart Publishing 2011) <http://www.bloomsburycollections.com/book/values-in-global-administrative-law> accessed 30 July 2018; Renaud Dehousse, ‘Regulation by Networks in the European Community: The Role of European Agencies’ (1997) 4 Journal of European Public Policy 246; Walter Powell, ‘Neither Market nor Hierarchy: Network Forms of Organization’ (1990) 12 Research in Organizational Behavior 295. 14 Laurens Van Kreij (n 11), 441-442. 15 Ibid, 442. 5 one policy area without fully replacing or eradicating each other. Thirdly, often the EU authority and their tasks (e.g. by the European Commission) transcend multiple enforcement regimes. 16 Scholten and Ottow distinguish between types of enforcement models based on the amount of tasks given from the national to the EU level. The authors distinguish between four models ranging from S (small), M (medium), L (large), and XL (extra-large) enforcement models. 17 The particular characteristics of these models are institutional characteristics of cooperation (e.g. choices between a network and an EU agency) and functional characteristics (the degree of sharing and division of powers). These four sizes also demonstrate various ‘degrees of Europeanisation’, from the smaller to the higher degree. Discrepancies in the practice of enforcement, but also in the enforcement chain and implementation, including sanctioning breaches, have been attributed to the ‘fractured’ system of hierarchical and institutional divisions, and divided competences. 18 In this context, Vervaele discusses ‘shared’ enforcement based on a ‘network-model’, whereby national regulatory and enforcement agencies and the European Commission play the key role in ensuring transnational cooperation and coordination required for the effective enforcement of EU policies. 19 Scholten, Luchtman and Schmidt also assert that no EU enforcement authority can under any legislative or institutional design replace national enforcement authorities in their entirety, thus emphasising the shared nature of EU policy enforcement. 20 Inherent challenges of shared enforcement in the EU include the need for creating common (or shared) terminology, but also shared enforcement practices and strategies. 21 The prism of shared enforcement has thus been used to analyse the challenges stemming from the allocation of powers and tasks between different authorities in EU policy enforcement. 22 Finally, Hofmann and Türk describe the EU law and policy enforcement as an integrated model, which they define as ’the evolution of legal, political and administrative conditions of administering joint policies’. 23 The integrated model of enforcement reflects the functional need for further cooperation, moving away from mere coordination, to secure the uniform 16 Ibid. 17 Miroslava Scholten, Annetje Ottow, (2014) “Institutional Design of Enforcement in the EU: The Case of Financial Markets”, 10(5) Utrecht Law Review , 80-91. 18 John A. E. Vervaele, Shared Governance and Enforcement of European Law: From Comitology to a Multi-level Agency Structure?, in Vervaele, J.A.E. EU Committees: Social Regulation, Law and Politics , Hart Publishing 1999, 130. 19 Ibid, 139; see, also, Laurens van Kreij, (2022) Enforcing EU policies: why do EU legislators prefer new networks of national authorities and not existing EU agencies?, Journal of European Public Policy, 29:10, 1568-1589. 20 Miroslava Scholten, Michiel Luchtman and Elmar Schmidt, “The proliferation of EU enforcement authorities: a new development in law enforcement in the EU”, in Miroslava Scholten and Michiel Luchtman (eds), Law Enforcement by EU Authorities (Edward Elgar, 2017), 3. 21 Miroslava Scholten (n 11), 1360; also, in the context of EU criminal law, see Michiel Luchtman and John Vervaele, (2014) ‘European agencies for criminal Justice and shared enforcement (Eurojust and the European Public Prosecutor’s Office)’, 10(5) Utrecht Law Review, 132–50. 22 See, for instance, Miroslava Scholten, EU (shared) Law Enforcement: Who Does What and How?, in Stefano Montaldo, Francesco Costamagna, Alberto Miglio (eds.), EU Law Enforcement (Routledge 2021), 7 – 23. 23 Herwig Hofmann and Alexander Türk, Introduction: towards a legal framework for Europe’s integrated administration, in Herwig Hofmann and Alexander Türk, Legal Challenges in EU Administrative Law : Towards an Integrated Administration (Edward Elgar 2009), 1-6. 6 implementation of EU legislation. This conception frames EU enforcement not as a conventional multi-level or hierarchical system but rather as ‘a system of integrated levels’. 24 2. Main features of EU enforcement A set of features define enforcement models in the EU as examined by legal scholarship. The inquiry into EU policy enforcement can be framed around two main vectors: first, (a) the degree of centralisation, which depends on the necessity for more harmonisation and consistency of enforcement, better coordination and cooperation between enforcement authorities – which have been primarily conceptualised through ‘Europeanisation’ and ‘agencification’. Second, (b) the existence of powers and tasks exerted by EU agencies, which represents a spectrum from indirect to direct enforcement. a. EU enforcement regime trends: centralisation, Europeanisation and agencification Europeanisation – or the process of shifting national supervisory powers and competences to the EU level – implies “an institutional evolution and functional transformation”, 25 exhibiting a trend across many domains including finance. Scholten and Ottow describe it as a shift from networks of national enforcement agencies to a more centralised structure of agencies and institutions at EU level. 26 This shift in the institutional set-up would often be accompanied by the increase in legally binding and direct powers of EU agencies and EU institutions, including supervisory, regulatory and sanctioning powers. The trend towards Europeanisation of supervision and enforcement, that is so visible in the EU’s financial sector, 27 stems from the need for ever greater coordination and cooperation in enforcement. Van Kreij underscores increased centralisation of EU policy enforcement due to the shift from Member State enforcement to growing reliance and broader responsibilities vested in EU enforcement networks and EU agencies. 28 In our understanding, however, network-based enforcement faces severe limits, considering its informality and functioning based on coordination, primarily, and faithful cooperation among stakeholders. 29 The degree 24 Ibid, 2. 25 Miroslava Scholten, Annetje Ottow (n 17), 86. 26 Ibid. 27 Silvia Allegrezza and Grazia Bruzzese, ‘Supervision and Enforcement of EU’s Anti-Money-Laundering and Countering the Financing of Terrorism’, EU Banking and Capital Markets Regulation (Palgrave Macmillan, Cham 2025) <https://link.springer.com/chapter/10.1007/978-3-031-70529-8_14> accessed 9 September 2025. Sophie Vuarlot-Dignac and Eugenia Siracusa, ‘The European System of Financial Supervision and in Particular the European Securities and Markets Authority’ in Federico Fabbrini and Marco Ventoruzzo (eds), Research Handbook on EU Economic Law (Edward Elgar Publishing 2019); Gianni Lo Schiavo and Alexander Türk, ‘The Institutional Architecture of EU Financial Regulation: The Case of the European Supervisory Authorities in the Aftermath of the European Crisis’ in Leila Simona Talani (ed), Europe in Crisis (Palgrave Macmillan UK 2016) <http://link.springer.com/10.1057/978-1-137-57707-8_5> accessed 2 August 2018; Giulio Vesperini and Edoardo Chiti, The Administrative Architecture of Financial Integration : Institutional Design, Legal Issues, Perspectives (Società editrice Il Mulino 2015). 28 Laurens Van Kreij (2019), 441. 29 L Garicano and RM Lastra, ‘Towards a New Architecture for Financial Stability: Seven Principles’ (2010) 13 Journal of International Economic Law 597, 620. 7 of Europeanisation is most often sector-specific, 30 and depends therefore on the EU policy area, e.g. in competition law, data protection, energy. Despite the need for sector-specific approaches, EU policy enforcement regimes showcase two main trends. First, the push towards further centralisation, including via agencification, addressed an increasing need to foster collaboration and coordination in enforcement, and accompanied a shift away from decentralised enforcement, and network-based enforcement, to EU supervisory authorities or agencies. 31 Second, the centralisation trend is supported by an increase in direct enforcement. In this regard, the ‘hub and spoke approach’ 32 is an EU-level enforcement model that relies on an agency and combines coordination and oversight of the work of national authorities, with direct powers to supervise some of the regulated entities, (see further in section IV below). b. From indirect to direct enforcement powers and corresponding tasks Indirect enforcement by an EU institution/agency consists in monitoring the performance of enforcement tasks and exercise of powers by national authorities. Direct enforcement, in contrast, primarily implies that an EU enforcement authority has powers to monitor compliance with legal rules and regulatory requirements by private actors, as well as to investigate and sanction violations of EU law by those private actors. 33 While decentralised enforcement model relies on direct enforcement powers vested in national enforcement authorities, shift of powers to the EU level can either maintain such a decentralised model with EU agencies and EU institutions merely assuming indirect powers, or be accompanied with the attribution of direct enforcement powers to the EU level. Thus, a centralised enforcement model can imply either indirect or direct enforcement powers vested in the EU agency (or institution). In the former situation as said above, the EU agency only oversees and coordinates enforcement that is primarily ensured by national authorities, while in the latter, the agency directly enforces EU law against private. Such a model - considered an ‘EU agency-led’ model - has, in reality, a blend of indirect and indirect powers (see section IV). Direct powers of an EU agency can span from monitoring to investigation and sanctioning. 34 Monitoring powers can thereby amount to the process of ongoing supervision of a specific 30 Madeleine de Cock Buning, Annetje Ottow, John Vervaele, (2014) “Regulation and Enforcement in the EU: Regimes, Strategies and Styles”, 10(5) Utrecht Law Review, 5. 31 See, for instance, Merijn Chamon, (2023). EU Agencies: Shifting Paradigms of EU Administration. EU Law Live , 48; also, David Levi-Faur, (2011). 32 EBA, ‘Response by the EBA to the European Commission’s Public Consultation on an AML/CFT Action Plan and the Establishment of an EU Level AML/CFT Supervisor’ (2020) 13–14. 33 Miroslava Scholten, Michiel Luchtman and Elmar Schmidt (n 23); see also H Hofmann, G Rowe and A Türk, Administrative Law and Policy of the European Union (OUP 2011). 34 The rationales for agencification (and the scope of powers and functions transferred to an EU agency) can be political, institutional and functional rationales, and as such may constrain the types of enforcement regimes. These rationales for giving enforcement tasks to EU agencies (as opposed to Member States’ authorities or EU 8 policy area, regardless of whether there is a suspicion of violation or non-compliance by the actor. Investigation powers, in turn, refer to the action authorities can take in case of suspected infringement of the rules. Finally, sanctioning powers refer to the formalized response by the enforcement authorities to an established violation and can be restorative or punitive in their nature, 35 and may have a deterrence and punitive function (see section III). III . T he reformed AML enforcement system In this section, we examine the reformed enforcement regime in the AML integrated system, with AMLA at its core. To this end, we discuss the reformed AML enforcement regime in light of the objectives of the system (1), then we analyse the two pillars governance set-up and joint arrangements therein (2) and finally study the attribution of tasks (what?) and corresponding powers (how?) of the relevant authorities (3). This legal and institutional analysis supports the third dimension in the 3-D enforcement model put forward in section IV hereinafter. 1. Objectives of the AML integrated system The analysis of shortcomings and problem drivers from the prior AML/CFT framework gives useful insights into the objectives of the reformed AML integrated system. 36 S upervisory divergence was underscored as one core issue, demonstrating inadequate and insufficient supervision at national level, in particular in relation to supervisory authorities, 37 following a threefold dimension. First, the European Commission identified significant differences in the timeliness and effectiveness of supervisory actions at national level. Due to supervisory authorities’ divergent prioritisation and allocation of resources, there was a discrepancy in the intrusiveness, intensity and frequency of AML/CFT supervision, 38 as well as in the methods to identify risks and the risk-based approach. 39 Second, the decentralised AML/CFT supervisory framework showed an overreliance of ‘home’ Member State authorities on the ‘host’ Member State’s. This host country approach 40 meant that the monitoring of entities’ compliance was based on networks) have been extensively discussed in the literature. See Laurens Van Kreij (n 11), 439-457; see, also, Miroslava Scholten (n 11), 1348; Laurens van Kreij, (2022) “Enforcing EU policies: why do EU legislators prefer new networks of national authorities and not existing EU agencies?”, 29(10) Journal of European Public Policy , 15681589; on political rationales, see Laurens van Kreij, (2024) “How Have EU Legislators Established EU Agencies With Enforcement Tasks? Case Studies of the European Aviation Safety Agency and the European Medicines Agency”, 63(2) Journal of Common Market Studies , 592; R. Daniel Kelemen, Andrew D. Tarrant, (2011) “The Political Foundations of the Eurocracy”, 34(5) West European Politics , 922–947. 35 Madeleine de Cock Buning, Annetje Ottow, John Vervaele (n 33), 5. 36 European Commission, ‘Report from the Commission to the European Parliament and the Council on the Assessment of Recent Alleged Money Laundering Cases Involving EU Credit Institutions’ (n 3) 23; European Commission, ‘Communication from the Commission to the European Parliament and the Council, Towards Better Implementation of the EU’s Anti-Money Laundering and Countering the Financing of Terrorism Framework (COM(2019) 360 Final)’ (n 3); European Commission, ‘Impact Assessment Accompanying the Anti-Money Laundering Package’ (n 3) 13–19. 37 There were also specific findings related to credit institutions. 38 European Commission, ‘Report from the Commission to the European Parliament and the Council on the Assessment of Recent Alleged Money Laundering Cases Involving EU Credit Institutions’ (n 3) 21. 39 AML A Regulation proposal, memorandum of understanding 40 European Commission, ‘Report from the Commission to the European Parliament and the Council on the Assessment of Recent Alleged Money Laundering Cases Involving EU Credit Institutions’ (n 3) 22. 15 addressed to them). 86 Thirdly, periodic penalty payments are adopted to compel the obliged entity to put an end to a breach after a failure to comply with prior administrative measures, or to compel natural person to provide further information, data, procedures or other material in the context of an invitation. 87 If compared with pecuniary sanctions and administrative measures in the legal framework 88 there is no express dissuasive component as those periodic penalty payments are already having a punitive function on the obliged entities. However, the peculiarity of those penalties is their potential retroactive effect ‘up to the date of application of the administrative measure’. 89 This option, left open to AMLA, 90 may be consequential for the obliged entity concerned by those penalties and actually adds, by this legal feature, a deterrence function next to the punitive effect of those measures. Furthermore, penalties are also envisaged in the case of breaches of beneficial ownership transparency requirements, 91 with an obligation for Member States to adopt the rules applicable to such breaches and ensure their implementation. 92 In AMLA’s interactions with the financial supervisors, its power to issue requests to act and instructions 93 to ensure AML/CFT regulatory compliance, and to request information, 94 is wideranging and constitutes an important disciplining device within the system. Furthermore, in its indirect oversight, AMLA has a power to request supervisors and supervisory authorities to act and can instruct them to take measures in relation to non-selected obliged entities. 95 In cases where the requests to act are not complied with or relevant information about the same are not provided within 10 working days, AMLA may ‘take over’ the supervision of those nonselected obliged entities – after obtaining the permission from the European Commission. 96 Such change in the tasks and powers assumed by AMLA, in lieu of the financial supervisory authority, will be temporary. Furthermore, AMLA will investigate systematic supervisory failures, non-application or incorrect application of EU Law. 97 The systematic failures may lead to Commission’s infringement procedures, and issuance of an opinion to compel the supervisor to comply with AMLA’s recommendation. 98 The powers attached to the investigation in case of systematic 86 Articles 22 and 6(1) AMLA Regulation. 87 See Article 23(1) AMLA Regulation. 88 See Articles 21, 22 AMLA Regulation. 89 Article 23(5) AMLA Regulation. 90 Indeed, AMLA’s decision to impose a periodic penalty payment ‘ may be taken at a later stage with retroactive effect…’ (emphasis added), as per Article 23(5) AMLA Regulation. 91 Set out in Chapter IV, AML Regulation. 92 Article 68(1) AML Regulation. 93 Art.6(2)(c) AMLA Regulation. 94 E.g. very detailed provision on the request for statistical information and internal processes or arrangements of authorities, through structured questionnaires and online/offline tools, see Article 6(2)(a) AMLA Regulation. 95 Article 6(2)(c) AMLA Regulation. 96 Article 32(5) AMLA Regulation. Regarding the process, see paragraphs (6) to (9). 97 Article 34, see also Recitals 9 and 35 AMLA Regulation. 98 Article 34(6) and (7) AMLA Regulation. 16 failures of supervision 99 are essential to guarantee the effectiveness of the AML/CFT supervisory system. In the interactions with the non-financial supervisors third, 100 AMLA have some powers to address the failures committed by authorities breaching Union law. Following the investigation of authorities’ potential breaches or non-application of Union law, AMLA can issue some recommendations with remedial actions, 101 and some warnings, in case of non-compliance with the prior recommendations including the measures that must be implemented to mitigate the effects of the breaches. 102 Actually, another forceful disciplining device can be found in the publication of peer reviews reports, both at EU and international level, unveiling the shortcomings along good practices. 103 As AMLA is tasked with the coordination of peer reviews of supervisory standards and practices in AML/CFT in the non-financial sector, 104 this peer pressure will begin from the start of AMLA’s operations, and contributes to supervisory convergence (see section d.). All these instruments strengthen the deterrence and punitive function of the EU AML/CFT enforcement regime, enhancing its effect across Member States and supporting supervisory convergence. Overall, the publication of pecuniary sanctions, the application of administrative measures in relation to breaches, 105 and periodic penalty payments 106 contributes to the deterrence function considering the anticipation of its potential impact on the reputation and credibility of the obliged entities. Moreover, the systematic publication of national and supranational risk assessments 107 exerts a reputational form of deterrence, which can prompt both public authorities and private actors to proactively address and effectively mitigate ML/TF risks. Lastly, an additional layer of deterrence in the AML enforcement architecture resides in the potential infringement proceedings by the European Commission. 108 This supranational oversight – upholding the integrity of the AML regime – will contribute to a more uniform application and enforcement of EU AML/CFT rules. d. Legal and policy tools to support efficiency in enforcement The framework provides for several legal and policy tools to support efficiency in AML/CFT enforcement: starting from data sharing enhanced arrangements, including unified methodologies and common approaches, as well as a duty of mutual assistance. 99 See paragraphs 2 to 4 of Article 34 AMLA Regulation. 100 Here this includes both non-financial supervisors and the public authorities that oversee self-regulatory bodies. 101 It is also within six months from initiating the investigation, see Article 37(3) AMLA Regulation. 102 Article 5(4)(c) and Article 37(4) AMLA Regulation. 103 Recital 39, 104 Article 5(4)(b) AMLA Regulation. 105 It is important to signal the co-legislators’ approach to the ‘punitive and educational goal of publications’ where preventive administrative measures (before any breaches) are not published, and therefore, only the decisions addressing AML/CFT requirements breaches with administrative measures. See Recital 105 AMLD6. 106 Article 58 AMLD6, Article 25 AMLA Regulation. 107 See Recitals 14 and 19 and Articles 7 and 8 AMLD6. 108 Article 78 AMLD6 provides a general transposition date by 10 July 2027, with a few derogations for dedicated provisions with two earlier date and one later. See the overview of national transposition on Eur-Lex https://eurlex.europa.eu/legal-content/EN/NIM/?uri=CELEX:32024L1640 as of 23 June 2025. 17 The data and information sharing arrangements are fundamental to enable AMLA, supervisory authorities and FIUs to carry out their responsibilities efficiently, contributing to better and faster data analysis. 109 For instance, Article 75 AML Regulation is particularly relevant as it concerns the exchange of information in the framework of partnerships for information sharing. Overall, the rationale behind this provision is to break silos, and to avoid having intelligence ‘trapped’ in between actors. The legal arrangements in this provision should foster data and information sharing in practice so that AML/CFT stakeholders conduct intelligence sharing both at operational and strategic level. Moreover, this information flowing among partners facilitates the use of shared risk analysis, which also contributes to breaking silos. Furthermore, AMLA is responsible, in collaboration with national authorities, to ensure the convergence of supervisory practices and the promotion of high supervisory standards, 110 which is primarily done through the adoption of system-wide methodologies and common approaches. AMLA will be responsible for developing a consistent, system-wide supervisory methodology, in cooperation with supervisory authorities. In particular, the whole AML system will maintain an ’up-to-date and harmonised AML/CFT supervisory methodology’, 111 which will detail the risk-based approach to the supervision of obliged entities. The legal tools supporting the common methodology correspond to the now standard approach to the Single Rulebook legal instruments, that is a combination of level 2 and level 3 acts (e.g., regulatory and implementing technical standards. guidelines, recommendations, opinions and other measures and instruments as appropriate). 112 However, the level of details as to the elements which must be contained in the supervisory methodology, and the means to reach that outcome, are quite unprecedented within a level 1 act in the Single Rulebook, i.e. here the AMLA Regulation. 113 This prescription from the legislators is an effective support for the whole AML/CFT enforcement system, considering its legally binding nature. Some elements of this supervisory methodology are already being crafted before AMLA’s full operation in 2028. 114 In this regard, the co-legislators envisaged that the system relies on some work carried out by the EBA, as outlined in the introduction, including for the development of a ‘refined and detailed harmonised methodology’ 115 that will allow to assess residual risk comparably (‘residual risk methodology’). Such a methodology is essential in the identification 109 All subject to the provisions of AMLD6 and AML R related to data protection, see in particular Article 70 AMLD6, and Article 75 AML Regulation. 110 Article 5(3)(e), AMLA Regulation. See also Article 5(4)(e) as regards tasks with respect to non-financial supervisors. 111 Article 8(1), AMLA Regulation. 112 Ibid. See also Recital 10, AMLA Regulation. 113 Article 8(2) letters (a) to (d) list the elements to be included within the risk-based supervisory methodology. Article 6(3) gives an overview of some practical tools for data collection and analysis. 114 Common approaches started to be discussed, during the transition period before AMLA creation, through regular meetings of national authorities in a Forum of EU AML/CFT supervisors, every 6 weeks as of June 2024, which is being phased out as AMLA takes over with its own governance set up. See EBA, ’Getting ready for AMLA : the EBA’s contribution to the new AML/CFT regime’, 26 June 2024. https://www.eba.europa.eu/sites/default/files/2024-06/9dc69374-bd6b-4d2d-8f588f36c8e6c6e7/Getting%20ready%20for%20AMLA%20factsheet%20FINAL.pdf 115 Recital 22, AMLA Regulation. 18 of the selected obliged entities and should strive to ‘harmonise’ the different national approaches. 116 The two EBA draft Regulatory Technical Standards related to methodology 117 will have an impact on the AML system by determining the scope of AMLA’s direct supervision and the work of supervisory authorities towards obliged entities in assessing and classifying their inherent and residual ML/TF risk profiles. 118 Finally, the authorities have a duty of mutual assistance. 119 The supervisory cooperation in the system mandates national supervisory authorities to provide wide-ranging information to AMLA on the obliged entities that they directly supervise and are under a duty of assistance in situation involving the application of national laws. 120 This duty relies primarily on thematic reviews, 121 training, staff exchanges and secondments of personnel, sharing best practices, the use of IT tools (e.g. ‘Next Generation’ FIU.net, IT and AI services). 122 In the context of FIUs, there is a specific objective of capacity building, knowledge-sharing and good practices, due to the softer nature of the support and coordination mechanism provided by AMLA. Interestingly, for both supervisory authorities and FIUs, they may introduce requests to ‘trigger’ such mutual assistance (e.g. for practical tools and methods, 123 or more generally, among FIUs or FIU-AML 124 ). On the other side of the coin, they are bound by obligations of information and data sharing. The co-legislators also resorted to a voluntary dimension in such mutual assistance, in which AMLA may ensure that other supervisory authorities mobilise its resources or FIUs provide assistance on a voluntary basis. 125 This soft element must be read together with the principle of sincere cooperation and mutual trust. IV . The 3 - D enforcement model and agencification in the new AML integrated system We argue that the novelty of the revised AML/CFT enforcement regime and its integrated system go beyond a simple further centralisation of AML/CFT supervision and enforcement (1). Instead, the regime is illustrative of a more significant transformation which combines and integrates – within a 3D enforcement model – the ‘binary’ approaches to centralisation/decentralisation, and to direct/indirect EU enforcement powers, together with 116 Recital 22, AMLA Regulation. However, the EBA will cease to have tasks and powers related to AML/CFT once the AMLA Regulation is fully applicable, including with a transfer of resources, see Recital 84. Furthermore, AMLA will also be able to rely on the EuReCa database (AML/CFT database) operated by the EBA until no later than 30 June 2027. See Article 106 AMLA Regulation. 117 Placeholder for the EBA Draft RTS to be published in October 2025. 118 Public consultation was open until 6 June 2025, https://www.eba.europa.eu/publications-and-media/pressreleases/eba-consults-new-rules-related-anti-money-laundering-and-countering-financing-terrorism-package 119 Article 10, Article 45 and Recitals 16, 51 AMLA Regulation. The Commission can also provide technical support to Member States. 120 Article 7(3) and (4), AMLA Regulation; see, also, Andrea Minto and Thomaz De Arruda, ‘The New EU Anti-Money Laundering Authority (AMLA) and the Standard for the Application of National Laws - “What’s Past Is Prologue”?’ Forthcoming EBOR. 121 Article 9(5) AMLA Regulation. All supervisory authorities are required to provide upfront information on which thematic annual reviews they envisage to conduct during the following year or supervisory term, on a yearly basis. 122 Recital 9, AMLA Regulation. 123 Respectively, Art. 10(1)(b), 10(3) AMLA Regulation. 124 Art. 45(2) AMLA Regulation. 125 Art, 10(5) and Art. 45(3) AMLA Regulation 19 a third dimension of operational and collaborative integration moving away from mere coordination (2). 1. The revised AML/CFT enforcement regime : a third dimension The new AMLA agency-centred and integrated system embodies a qualitatively new dimension to the understanding of EU policy enforcement, which is one that transcends the traditional narrative of a major shift of direct enforcement powers from national to EU level and further centralisation. We argue that the centrepiece of the revised AML/CFT regime is the dimension of operational and collaborative integration – moving beyond mere coordination (figure 1 below). This dimension is, for the first time, legally structured to this extent in EU secondary law and institutionalised under the AML/CFT enforcement regime. It embodies the essential contributions and roles of joint supervision and analyses, peer coordination and pressure, shared information and intelligence across authorities and actors across the AML integrated system, as examined above. First, the trend towards greater Europeanisation of enforcement in the EU is confirmed with AMLA adding a further degree of centralisation and assuming greater direct enforcement powers at EU level – although it remains a ‘decentralised’ agency (see subsection 2 below). The direct supervision of selected obliged entities and powers vis-à-vis national competent authorities (pillar 1) constitute the main illustrations of this trend. 126 Second, and more intriguingly, the revamped AML multi-actor enforcement system exhibits a third dimension in the features shaping enforcement, which builds on the two traditional vectors of centralisation/decentralisation and direct/indirect powers. Notably, we observe a more integrated enforcement system fostered by a number of elements examined above: AMLA’s tasks and powers, all AML/CFT system actors’ duties in information and knowledge sharing aimed at breaking existing silos, the new mechanisms for mutual assistance and capacity building, as well as AMLA’s powers to exercise peer pressure and mediate disputes. 127 Moreover, AMLA supervisory oversight under the new system rests on a deterrence and punitive function, including important disciplining device. 128 As such, we argue that this additional dimension or ‘third dimension’ shows how agencification in a common system allows moving along from mere coordination to operational and collaborative integration. Such operational integration not only takes place at agency level but also within the system, because all actors are equally involved in common endeavours, with a shared responsibility, and steered by common methodologies (see section III above). These ‘third dimensional’ features of the revised AML enforcement system create and sustain a connective tissue that ties together the different parts in the enforcement system. Indeed, the 126 See, Section III [a]. See, Section III [a] [b] and [d]. 128 See, Section III [c]. 20 new AML integrated system also goes beyond a network-based enforcement model – which would have limited capacity to enforce decisions due to primarily informality. 129 The third dimension thus constitutes an organisational dimension aimed at fostering a harmonised and cooperative system of enforcement. It is also in this third dimension that we find main novelties aimed at addressing the AML system’s earlier shortcomings, including ‘home’-‘host’ coordination issues, and inefficiencies in cooperation and information exchange. They appear to rest on the principles of sincere cooperation and mutual trust. 2. The 3D enforcement model Our three-dimensional approach to EU enforcement aims at representing better the two traditional vectors of centralisation/decentralisation and direct/indirect enforcement powers, and their interplay with operational and collaborative integration. We examine this in the context of the revised AML/CFT enforcement regime. We propose a conceptual framework to visualise the revised AML enforcement system in the EU through a three-dimensional model (Figure 1 below). This model transcends a binary approach and illustrates the interdependence of actors in the EU’s AML/CFT enforcement system. This 3D structure, represented in the shape of a cube, captures the dynamic interaction between three vectors: Decentralisation – Centralisation ; Indirect – Direct Enforcement ; Coordination – Operational and Collaborative Integration (numbered 1 to 3 below). The three-axis diagram includes the degree of centralisation versus decentralisation (X-axis), enforcement powers ranging from indirect to direct (Y-axis), and the extent of coordination, operational and collaborative integration among actors (Z-axis). Those dimensions are interrelated and embedded in the legal and institutional framework above examined and will be further shaped by practice within the system. The first axis (X) captures the distribution of supervisory powers across the AML/CFT system, from decentralised national actors to centralised EU authority (agency or institution). On the left end of the spectrum (A) lie national authorities, including those overseeing self-regulated entities such as Bar Associations in some Member States. At the opposite end stands a hypothetical fully centralised entity at EU level in a situation in which national authorities would no longer exist. In the revised regime, the EU’s Authority, AMLA, should be positioned tilted towards centralisation 130 considering its direct supervisory responsibilities over selected obliged entities, that coexist with some shared responsibilities with the national authorities responsible for the lions’ share of non-selected obliged entities, and FIUs. AMLA is also responsible for the oversight over national authorities, and in this context, it can adopt binding instructions, launch inquiries, and take over the supervision of non-selected obliged entities. Yet, the national authorities and FIUs still have an essential role. Ultimately, the adopted AMLA 129 See discussions in section II. 130 We inserted a vertical mark on the X-axis, in a qualitative way, therefore, there is no proportional representation to observe here. 21 legal and institutional framework contains direct and indirect supervisory powers combined in a decentralised EU agency. 131 The second Y-axis reflects the enforcement powers along a spectrum from indirect to direct enforcement powers. At the bottom of this axis lie indirect enforcement tools, such as nonbinding guidance and informal coordination mechanisms, which represent earlier stages of EU AML policy (before the EBA having more powers through the 2019 ESAs Review). If we go up the axis, we find shared instruments, such as the development of common methodologies and the conduct of peer reviews, which blur the distinction between soft coordination and more formal oversight. In this regard, AMLA has been given by the EU legislators the responsibility for common methodologies and quite intrusive tools in its oversight over the system (e.g. instructions to national supervisors). Moreover, intrusive tools exist among supervisors (e.g. inquiries to each other) which demonstrates a somehow shared coercive dimension, which can exist thanks to the deterrence and punitive function in the overall system. Towards the top, the model captures direct enforcement powers, such as AMLA’s ability to impose binding supervisory decisions, or issue sanctions against selected obliged entities. When these tools reach the highest point in the cube, they reflect a qualitative difference: the imposition of legal obligations and a coercive effect that builds upon the deterrence and punitive function (see section III). Importantly, the AML/CFT enforcement regime represents a layering of several instruments across the system, with direct and indirect enforcement tools within a common institutional and legal framework. This will lead to implement AML/CFT policy, through intrusive tools when critical (for both selected and nonselected obliged entities) while preserving supportive or coordinating functions in the system, including in the coordination and support mechanism for FIUs. The third Z-axis introduces the novel third dimension, directed into the depth of the cube, which reflects the degree of coordination, and further operational and collaborative integration among AML actors in the system. At the front of the cube (point C), the system exhibits low level of integration, with minimal and informal coordination between national authorities and the EU, e.g. through networks, where enforcement largely occurs in parallel, with inefficiencies and loopholes that we observed. As we move deeper into the cube, the model represents the increasing level of integration which we discussed above regarding joint supervisory teams and joint analysis teams, shared statistical platforms, centralised databases, and harmonised risk assessments, among others. These developments represent operational and collaborative integration (see point I) not through hierarchical relationship in a simplified binary fashion, but through blending hard and soft enforcement tools, which are, once again embedded legally and institutionally in the revised AML/CFT enforcement regime. At the top-right corner (point F) lies a hypothetical situation of a fully centralised and operationally integrated regime — not (yet) realised. No distinction would remain between EU131 See impact assessment, pp. 33 and 45. 22 level and national enforcement actors; the EU agency would have strategic oversight and entire operational execution. This third axis thus completes the 3D model by highlighting how operational and collaborative integration dynamics reinforce the formal allocation of powers and attribution of competences in the EU AML enforcement regime. Figure 1: The 3D Structure of the AML Enforcement System in the EU Source: Authors’ elaboration This model, on one hand, illustrates more precisely and adequately a deepening of agencification as a mode of governance in financial regulation. On the other, it showcases how the co-legislators’ choice, on a spectrum between mere coordination and further operational and collaborative integration, accompanies the general trend for further direct enforcement powers. This policy choice does not necessarily mean that there is full centralisation. On the contrary, this shows the importance of the whole system tied together to implement the EU policy area for which the co-legislators chose a common system of enforcement (where uniformity, consistency and convergence are fundamental). V . Conclusion The main contribution of this paper, based on the analysis of the powers and tasks under the revised AML enforcement system, is a new analytical framework of enforcement structure in 23 the EU, that rests on a three-dimensional (3D) model (see figure 1 above). The model, applied in the case of the revised AML/CFT enforcement regime, combines several ‘soft’ instruments and tools for operational and collaborative integration, 132 and underscores the deterrent and punitive function 133 of the system with binding tools and powers, disciplining device, 134 combined with a shared expertise and common methodologies and approaches. This contributes to reducing fragmentation and to ensuring greater consistency and convergence of supervisory practices and the promotion of supervisory standards across the system. In other words, the revised enforcement system, with AMLA at its core with national authorities and FIUs, showed that the prism of ‘Europeanisation’ and that of ‘centralisation’ of enforcement, or the focus on ‘direct’ enforcement powers, do not give a sufficiently robust analysis of powers, tasks, and key mechanisms underpinning the revised regime. The third dimension is complementary to the more traditional and binary approach to ‘Europeanisation’ and ‘agencification’ of EU enforcement – i.e., the decentralization/ centralisation and direct/indirect enforcement vectors. As such, the 3D model enables a more robust analysis, one that illuminates the elements of coordination (often characterized by informality and lack of intrusiveness as in a network-based enforcement) and of deeper operational and collaborative integration, including through the joint conduct, reporting and transmission of analyses, as well as capacity building initiatives, or a harmonised AML/CFT supervisory methodology for the whole system. Moreover, it can prove useful in identifying and studying the challenges and potential weaknesses of an integrated, agency-centred system, such as the new AML/CFT enforcement regime. The paper builds on the literature on EU enforcement and the analysis of the new AML enforcement system. Overall, the revised AML/CFT enforcement regime can be characterised as an extra-large enforcement model and exhibits a deep integration of EU law and policy 135 through the latest review of its legal and institutional framework. The EBA held a public consultation regarding four draft Regulatory Technical Standards (RTS) in spring 2025, following the European Commission’s call for advice. While the results are yet to be published, 136 these RTS, as level 2 instruments in the AML Single Rulebook, form part of the AML/CFT enforcement regime. Their impact will be wide-ranging as they focus on the sanction regime, customer due diligence, the methodology to assess and classify obliged entities’ inherent and residual risks, and the methodology to determine the selected obliged entities based on risk assessment. Furthermore, AMLA will issue an opinion addressed to the 132 E.g. the voluntary dimension in mutual assistance, in which AMLA may ensure that other supervisory authorities mobilise its resources or FIUs provide assistance on a voluntary basis. 133 See above section III.3.c. 134 In that respect, any failure to comply could lead to AMLA’s recommendations and warnings, or even, to the ‘takeover’ of (the direct supervision of) non-selected obliged entities. See Article 32(5), AMLA Regulation. 135 See section II. In particular: Miroslava Scholten, Annetje Ottow, (2014) “Institutional Design of Enforcement in the EU: The Case of Financial Markets”, 10(5) Utrecht Law Review , 80-91. Herwig Hofmann and Alexander Türk, Introduction: towards a legal framework for Europe’s integrated administration, in Herwig Hofmann and Alexander Türk, Legal Challenges in EU Administrative Law : Towards an Integrated Administration (Edward Elgar 2009), 1-6. 136 Expected 31st October 2025. 24 European Commission on the ML/TF risks affecting the Union every two years from 2030 onwards. 137 The findings of this paper, in particular the delineation of the third dimension in operational and collaborative integration also points at new directions for further research that could address the challenges faced by enforcement systems, or explore new avenues for reviewing EU enforcement regimes in other policy areas. 137 Article 7(5) AMLD6. AMLA can decide to do it more frequently.