Priority Setting as the blind spot of administrative law enforcement: Theoretical, Conceptual, and Empirical Study of Competition Authorities in Europe

Priority setting by independent regulatory agencies (IRAs) is an invisible, yet essential component of regulatory enforcement. The selection of which cases to enforce and which to disregard is vital given IRAs’ finite resources, and due to its function of concretising open-ended administrative norms. Clear enforcement priorities allow IRAs to focus on matters of genuine economic and doctrinal importance, solve complex socio-economic problems and build credible, independent, and accountable authorities. However, as a blindspot of administrative discretion, to date neither a normative framework was developed to assess IRAs’ priority setting rules and practices, nor a shared terminology exists to evaluate its different features. This article fills this gap by developing a novel typology and normative framework to guide IRAs’ priority setting, based on a historical, conceptual, and empirical study. It combines insights from top-down analysis of administrative and criminal law enforcement with bottom-up empirical research and engagement with IRAs using EU competition law enforcement as a case study.


INTRODUCTION Priority setting: between expertise and the rule of law
Setting priorities by independent regulatory agencies ('IRAs') is a crucial component of effective expert-driven enforcement, free from politics.As IRAs are constrained by scarce financial and human resources, it is neither possible, nor desirable, that they enforce every possible law infringement.The power to choose which cases to pursue and which to disregard is a precondition for preserving society's resources to tackle the most harmful infringements.Such power affords authorities the autonomy to focus on matters of genuine economic and doctrinal importance, and hence, can contribute to credible enforcement priorities.
When setting priorities, IRAs exercise administrative discretion, that is, the power left to decision-makers to choose 'between different alternatives when concretizing legal norms with a circumstances, how can one differentiate between prioritisation decisions being made in the public interest to those that advance private interests, such as increasing the reputation of the IRA or its officials (as opposed to the public interest as articulated by elected politicians) or self-enrichment? 9w can one safeguard technical expertise and avoid unjustified prioritisation practices based on cheery-picking, regulatory capture, 10 revolving doors 11 , or populist initiatives? 12me of the conceptual challenges common to priority setting by IRAs -such as the lack of clear legal standards, the democratic deficit of their operation, scarce resources, lack of information on non-compliance, and reputational effects -have already been explored.Scholarship on regulatory enforcement adapts enforcement strategies to specific cases, such as the 'regulatory pyramid' of Responsive Regulation, 13 or Smart-and Risk-based Regulation. 14In the fields of taxation and environmental protection, the concept of 'enforcement elasticity' was developed to inform a cost-benefit based selection of enforcement targets. 15Research on administrative discretion identified it as 'space within and between rules in which legal actors exercise choice', 16 and discussed how discretion can be delineated, structured and controlled. 17Political scientists examined the emergence and operation of IRAs as a process of delegation within principle-agent context. 18However, none of these strands of literature focused on IRAs' enforcement priorities.Despite these scholarly efforts, the role of priority setting is largely overlooked not only by 'ordinary citizens', but also by scholars, policymakers, and courts.IRAs were developed incrementally, as a 'historical accident'. 19Lacking a clear regulatory philosophy, many were created to respond to specific political challenges.As elaborated below, the scope and nature of IRAs' priority setting powers were typically defined implicitly, corresponding to their national administrative, constitutional, and criminal law traditions.Moreover, priority setting rules and practices are also influenced by a complex matrix of non-legal factors, such as broader political and economic circumstances, bureaucratic and organisational norms, personal experiences, the decision makers' perceptions and attitudes, and moral and social norms. 20At the same time, courts exercised only limited review over the exercise of priority-setting by IRAs, mostly ensuring that IRAs did not overstep their legislative boundaries. 21They typically focus on cases IRAs select to pursue, not on the process and the impact of case selection.Focusing on formal actions by IRAs, scholars, policymakers, and courts tend to overlook instances of inaction or informal action, even if the latter is estimated to characterise the vast majority (90%) of the IRAs' efforts. 22cordingly, to date, no normative framework guides the assessment of IRAs' priority setting rules and practices and no shared terminology explains its different aspects.There are neither 'bestpractices' setting and controlling enforcement priorities, nor benchmarks for measuring their effects.

Aims, methodology, and approach
This article aims to fill this gap by developing a novel typology and normative framework to guide IRAs' priority setting, based on a historical, theoretical, and empirical study.In a bid to capture the legal, institutional, and practical contexts of priority setting, it combines insights from topdown analysis of administrative and criminal law enforcement with bottom-up empirical research and engagement with IRAs.The empirical analysis is based on a systematic and comprehensive mapping ('coding') of the procedural and substantive rules and practices that define the way competition authorities ('CAs') of 27 EU Member States, the United Kingdom (UK), and the EU Commission ('Commission') set their priorities. 23The data was collected through desk research of the publicly available legislation, case law and policy documents in each jurisdiction combined with written questionnaires and interviews with officials of the CAs. 24The cut-off date for the data collection was December 2020. 25r typology and normative framework were presented in a Policy Report, 26 summarising the empirical findings and formulating policy implications for competition authorities.This article places the empirical findings and subsequent feedback to the Report from interviewees and workshops with enforcers (including the Commission, and European and international CAs) in the context of existing scholarship of regulatory enforcement and administrative discretion.The article further develops our findings, framework, and typology for the operation of various IRAs beyond competition authorities.
The remainder of this article is organised as follows.The first section traces the historical origins of priority setting powers as they developed in the shadow of the modern regulatory state.
It points to the diversity of IRAs' structures and powers across national legal orders, in particular between Anglo-Saxon and European jurisdictions.The second section draws insights from the rich theoretical account of administrative discretion, to propose normative benchmarks to guide the assessment of priority setting rules and practices.Departing from the narrow judicial review centred approach to control discretion (ultra vires principle), it advocates a broad public interestbased approach and identifies five good governance benchmarks: effectiveness, efficiency, independence, transparency, and accountability.
The third section explores IRAs' priority setting by using EU competition law as a case study.
It discusses the emergence of CAs in Europe and the evolution of their prioritisation powers following the 'modernisation' of the enforcement in 2004.EU competition law offers important insights into priority setting as CAs are the second most common IRAs globally, 27 and were among the first IRAs in the US and Europe, serving as a blueprint for other IRAs. 28The fourth section 24 The empirical study was undertaken pursuant to the approval of the Ethics Committee of the University of Amsterdam.Furthermore, training seminars with European and international CAs enriched the insights gained. 25This date also represents the end of the implementation period of Directive 2019/1 of the European Parliament and of the Council of 11 December 2018 to empower the competition authorities of the Member States to be more effective enforcers and to ensure the proper functioning of the Internal Market [2019] OJ L 11/3 (the 'ECN+ Directive'), which will be discussed below.Since this date, some Member States reformed their priority setting rules and practices, which are not reflected in this study. 26Or Brook and Kati Cseres, 'Policy Report: Priority Setting in EU and National Competition Law Enforcement' Available at https://ssrn.com/abstract=3930189or http://dx.doi.org/10.2139/ssrn.3930189. 27Mattia Guidi, 'Competition Authorities' in Martino Maggetti, Fabrizio Di Mascio, and Alessandro Natalini (eds.),Handbook of Regulatory Authorities (Edward Elgar Publishing, 2022), 113-114. 27Ibid, 115. 28Many IRAs have the power to regulate the behaviour of market parties ex ante, a power that other agencies usually do not have Even though IRAs sometimes have a statutory duty to perform certain tasks (ie.requirement for market assessments in the telecommunications sector) unlike CAs, still, agencies largely have a considerable margin of discretion in setting their enforcement agendas.See Annetje Ottow, 'Market and Competition Authorities: Good Agency Principles' (OUP, 2015), 159.
introduces our typology to structure the understanding of priority setting, distinguishing between seven aspects of priority setting in the pre-decision, decision, and post-decision stages of IRAs' decision-making.The degree of priority setting powers in each aspect is defined with reference to the external or internal controls imposed on the exercise of discretion.By presenting descriptive statistics and qualitative analysis of the operation of the CAs in Europe, it demonstrates the rich diversity of priority setting rules and practices across each of the seven aspects and the implications of IRAs' specific choices on the attainment of good governance principles.
The fifth section introduces four representative models of IRAs' prioritisation rules and practices emerging from the empirical and theoretical study.We argue that as priority setting rules and practices are deeply embedded in, and directly shaped by each IRA's respective legal system, identifying a single 'best' model for prioritisation is unfeasible.Outlining the four models of IRAs' prioritisation is, nevertheless, important as each model reflects a unique trade-off between the good governance principles.The models identify and visualise how a specific IRA could better align its priority setting practices to its powers as defined by law and hence, better comply with good governance principles.The sixth section concludes.

HISTORICAL ORIGINS OF IRAS AND THEIR PRIORITY SETTING POWERS
The nature and scope of IRAs' discretion to set priorities are inherently tied to the rationales justifying the emergence of economic regulation and its enforcement mode by IRAs.This section explores the historical development of each of these layers, starting from the nineteenth century.
As elaborated below, the emergence of IRAs was significantly influenced by two conflicting approaches to economic regulation: the Anglo-Saxon model of private ownership and the European model of public ownership.This section also analyses the degree to which IRAs' priority setting powers are governed by the external or internal controls imposed on the exercise of their discretion, adopted in parallel to their historical development. 29External controls are the limits on the exercise of discretion imposed on the IRA by the legislator, government, or judiciary.Internal controls, refer to self-adopted measures by the IRA overseeing and structuring the exercise of its discretion, for example by adopting binding or non-binding guidelines.

Anglo-Saxon model of private ownership, and the birth of IRAs in the US and UK
Economic governance in the US and the UK traditionally followed a private ownership model, leaving the ownership of industry to private market actors and limiting government intervention to cases of market failure. 30The regulation of utilities and other privately-owned sectors called for dedicated institutions to administer and enforce the legal rules under their jurisdiction.
In the UK, until the mid-nineteenth century, utilities were commonly regulated by 'Commissioners', who were the forerunners of modern IRAs.Commissioners had judicial, administrative, and regulatory responsibilities, which like their organisational structure varied considerably from one sector to another. 31The Commissioners' independence was seen as an important guarantee against arbitrary and unfair treatment by the British King and his ministers. 32 the second part of that century, some of this independence was lost as many of these bodies were incorporated into central or local governmental departments, which were subject to ministerial and political supervision. 33Yet, they still retained some independence, as besides being loyal to the minister, every civil servant was expected to be politically impartial. 34In parallel, from the early twentieth century, a specialised tribunal system, operating outside the ordinary courts system, was developed to handle disputes concerning matters of transport and competition, rent, social insurance and assistance. 35ese British institutions inspired the creation of modern IRAs.IRAs are an American regulatory innovation, that did not emerge as an intentional category of institutions, but as a group of agencies sharing common legal status despite having diverging structural and statutory characteristics. 36They appeared in the late-nineteenth and early-twentieth centuries with the rise of economic regulation.During that period, industrialisation and urbanisation stimulated economic growth given the increased mobility of workers and the expansion of regional and national markets.However, not all members of society felt they received a fair share.In particular, farmers, small businesses, and workers demanded government intervention to fight abusive practices by railroads.After legislation adopted by individual states failed, the 1887 Interstate Commerce Act (ICA) established the first federal independent regulator in the US: the Interstate Commerce Commission (ICC). 37As the design of the ICC was inspired by the British Railways Commission, a semi-judicial tribunal, 38 it acted as an administrative tribunal, operating reactively following case-by-case adjudication. 39e establishment of the ICC as an independent body was the result of an evolution rather than of doctrinal theory. 40During its first years of operation, it was not fully independent, but was placed under the US Department of the Interior. 41Granting independence to the ICC was a 'historical accident', originating from the disappointment of the drafters of the ICA with the appointed ICC President. 42Commentators pointed to various justifications for American IRAs' independence, including their quasi-judicial nature of the commissions, i.e. independence of regulatory agencies is akin to the independence of the judiciary; 43 developing independent expertise on technical and complex matters by separating regulatory functions and shielding IRAs from politics; advantages of geographical representation vis-à-vis executive departments, and taking up experimental tasks or tasks that did not fit with existing governmental departments. 44e prestige of the ICC stimulated the expansion of other IRAs. 45The ICC's structural features served as a template for other IRAs, in particular, antitrust agencies.The Sherman Act, which was adopted three years after the Interstate Commerce Act, did not set up an administrative commission and relied on enforcement by the Department of Justice (DoJ) and the courts.Yet, in 1914 the Federal Trade Commission (FTC) was established.Following the ICC structure, 46 it was adjudication-based. 47he first coherent legal-economic philosophy underlying IRAs was the 1930s New Deal, 48 which fuelled the spread of IRAs.IRAs were delegated broad powers, substantial discretion, and served as independent technical experts by insulating public officials from partisan pressures in the service of a long-term public interest.In 1935, the Supreme Court in Humphrey's Executor 49 recognised the constitutional status of independence, which distinguished IRAs from the executive. 50Fixed terms, for-cause removal, and multi-member board of experts were established as the cornerstones of IRAs. 51Upon the expansion of new agencies, the Administrative Procedure Act of 1946 (APA) standardised IRAs' administrative processes and controls, strengthening their powers and independence. 52ring the 1960s and 1970s, the ICC and other IRAs transformed from reactive adjudicators into proactive rule-makers and regulators. 53The focus on rulemaking, responded to scholars and judges advocacy, arguing that the general and prospective characteristics of rulemaking were more fair and efficient than the incremental, time-consuming adjudicatory approach. 54This proactive operational mode created priority setting powers, as it invited IRAs to set their regulatory agenda and granted them a functional advantage over the courts. 55itially inspired by the British Commissions, these American IRAs influenced economic governance in the UK.Following the Second World War, the UK abandoned its previous private ownership model and nationalised large parts of the industry. 56Government departments regulated and oversaw the operation of these nationalised sectors, aiming to ensure not only the functioning those which believed that large interstate commerce should be regulated.See   51 The Court ruled that the FTC was (1) non-political and nonpartisan, (2) uniquely expert, (3) 'quasi-legislative', and (4) 'quasi-judicial' and as such was an IRA, rather than an executive.Hence, it remained one of the core judicial pillar of the technocratic, independent administrative system by grounding the constitutionality of FTC Commissioner immunity from presidential removal for political reasons.Also see Daniel A Crane, 'Debunking Humphrey's Executor.'Geo.Wash.L. Rev and competitiveness of those services, but also other public polices, such as employment, economic growth, stable prices, and a balance of payments. 57These new tasks warranted new institutions.One early example was the British Monopolies and Restrictive Practices Commission (MRPC) in 1948, which responded to the 1944 Employment White Paper's call to introduce competition policy to achieve full employment. 58Unlike its American counterparts, the MRPC was advisory in nature.It held investigative and recommendation-making powers, but action could only be taken by the minister responsible for the relevant sector.As the management of public, private, and mixed entities became increasingly complex, a 1968 Report on Civil Service, called to reassess which activities should be performed by governmental departments and which should be moved to independent external bodies. 59Inspiration was drawn from how administration was organised in the US and France, where regulators involved scientists, engineers, and other specialists instead of the 'generalist' or 'amateur' British regulators. 60By highlighting the way the Swedish government was organised with public bodies enjoying independent status, 61 the Report encouraged the creation of the Office of Fair Trading (OFT) in 1973, a non-ministerial government department governing consumer protection and competition law.Inspired by the independence of the American FTC and the German Bundeskartellamt, the OFT was conferred priority setting powers, to start investigations and refer cases to the MMC. 62legating enforcement powers to IRAs was reinforced following the election of Thatcher's government.From the mid-1980s, the British economy was reorganised by privatising public utilities and liberalising others. 63This reform, in turn, called for a new mode of governance and the establishment of sector-specific regulators.Similarly to the OFT, many of these regulators were headed by a single Director General, operating free from political pressure. 64

Priority setting of Anglo-Saxon IRAs: limited controls
The transformation of the American and British IRAs into proactive regulators was a crucial step towards the emergence of their priority setting powers.Proactive operation invited IRAs to set their own agenda rather than reacting to cases brought in front of them.As elaborated below, this led to the emergence of the first priority setting governance models, defining IRAs' prioritisation powers and their controls.
The marginal attention devoted to the exercise of administrative discretion by Anglo-Saxon IRAs in general, and their priority setting powers in particular, can be explained by their history.
In the early days of British tribunals, the ultra vires doctrine was the main external control limiting the exercise of administrative discretion.This principle emphasises separation of powers and is based on a unitary concept of democracy. 65Originating from Dicey, who alerted to administrative discretion as a threat to the rule of law in England in the 1900s, 66 when the legislator delegates its power to an administrative agency, judicial intervention centres on ensuring that the agency does not transgress the legislator's will.The ultra vires doctrine served both as a justification for judicial intervention and prescribed its limits: the exercise of judicial control is limited to ensuring an agency respects its legislative boundaries; and when it operates within the scope of its delegated powers, courts avoid substituting their own views with that of the authority. 67The ultra vires doctrine was well-suited for English administrative law during the nineteenth century and the early American reactive-adjudicators agencies.These tribunals relied on adversarial systems, where private parties bring cases and evidence, leaving limited room for proactive action.Control over their operation, was, therefore, similar to judicial review by ordinary courts and focused on ensuring the protection of the private interests of the parties to the dispute. 68glo-Saxon IRAs had limited competences to adopt internal controls.British tribunals, who inspired the emergence of the US IRAs, were bound by the no-fettering rule, prescribing that public bodies with discretionary power are not entitled to base their decisions on a pre-determined rule without considering the merits of the individual case. 69This rule has two aims: safeguarding a fair 64  trial to ensure that individual cases are treated on their merits; and to promote administrative flexibility, adapting decisions to changing circumstances and priorities subject only to control of legality and reasonableness. 70In British Oxygen of 1971, the House of Lords held that IRAs are not prohibited from adopting internal controls if they retain discretion when applying them to a specific case. 71Arguably, jurisprudence in the UK did not only permit, but actually required, the adoption of such internal controls. 72 the US, too, the operation of IRAs was subject to limited judicial control.Until the 1930s, courts were highly reluctant to review administrative decisions unless authorised to do so by the law under which these agencies operated. 73Even after general reviewability of discretion was recognised, courts refrained from questioning both facts and policy choices made by IRAs limiting their scrutiny to questions of procedure and statutory interpretation. 74The New Deal confirmed both the limited judicial control of IRAs and left the review of enforcement priorities unchecked.
When IRAs adopted internal controls, they were often informal, not even binding the IRA itself. 75e limited external controls imposed on priority setting by British and American IRAs align with Anglo-Saxon traditions of criminal law granting wide, uncontrolled prosecutorial discretion. 76hering to the opportunity principle, the state is granted a choice not to start investigation, even when enforcement is technically and legally possible. 77This broad enforcement discretion is justified by rationales of procedural economy. 78As a 'first-come, first-serve' approach is undesirable, setting enforcement priorities is essential. 79Yet, the opportunity principle is not only a response to pragmatic considerations but reflects the belief that enforcement priorities are necessary in a democratic society, as enforcement should be avoided when it is unjustified. 80These justifications extend to administrative law enforcement, which is based on similar considerations of deterrence, seriousness of infringements, and norm concretisation. 81itish and American competition authorities illustrate two distinctive governance models of priority setting.They are both characterised by limited external controls (thus, leaving IRAs wide discretion), but differ in the degree of their internal controls.In the US, the power to set priorities is not only uncontrolled by the legislator, but neither DoJ 82 nor FTC 83 adopted internal controls.This reflects a model of high degree of prioritisation, with limited external and internal controls.As elaborated below, this model gives more weight to effectiveness, efficiency, and independence of IRAs over their transparency and accountability.In the UK, while few external controls are imposed on the exercise of the OFT/CMA's priority setting powers, they are required to publish an annual plan, explaining their priorities for the respective year. 84Moreover, they adopted internal prioritisation guidelines. 85This model forgoes effectiveness and efficiency, in favour of greater transparency and accountability.

European IRAs: origins and varying priority setting powers and controls
The control of key industries in Europe (beyond the UK) followed a public ownership model during the nineteenth century.Nationalisation of key industries endowed states to structure their economies by safeguarding the public interest against powerful private entities. 86Public ownership dominated the governance of utilities such as energy, transport, telecommunications and postal services.Having the characteristics of natural monopolies public ownership was advocated to eliminate economic inefficiencies, protect consumers, fight excessive political power, stimulate growth, favour specific regions or groups, and ensure national security. 8781 In practice, American scholars often made no distinction between enforcement discretion in criminal and administrative law, especially when sanctions are being imposed.This public ownership model meant that IRAs were created in Europe only by the late-1980s and 1990s with the 'rise of the regulatory state'. 88Following the footsteps of the UK, the rise of neo-liberal policies, globalisation, and increased international competition pushed states to deregulate and liberalise markets. 89The trend of establishing regulatory, arms-length agencies separated policy making from regulation and shifted from discretionary to rule-based instruments. 90Similar to the UK, IRAs emerged in telecommunications, energy, and financial sectors, and slowly spread to competition law. 91These developments are often traced back to the influence of American political-economy following the Second World War, but are also linked to the application of supranational EU rules in national legislation 92 and to the EU's pressure to insulate national decision-making from national politics and favouritism. 93As in Europe are characterised by varying institutional forms and powers.Some were explicitly made independent, while others operate as an institutional unit subject to the supervision of a ministry (eg the German Bundeskartellamt).Their nature and powers are deeply embedded in, and directly shaped by their respective administrative law systems and constitutional orders.
Contrasting administrative traditions and competing theoretical perspectives explain why governments established IRAs and delegated particular combinations of powers to them. 94untries and political systems with different types of capitalism rely on various degrees of economic coordination and define the IRAs' roles and discretion accordingly. 95tional criminal law traditions also influence the priority setting powers of IRAs.Unlike the opportunity principle characterising Anglo-Saxon jurisdictions, many civil-law jurisdictions adhered to the legality principle, requiring the state to act whenever sufficient evidence exists.Compulsory prosecution reflects the twin objectives of equality before the law and enhancing general deterrence. 97It prevents disregarding certain law infringements which make 'easy the arbitrary, the discriminatory, and the oppressive'. 98Decisions not to prosecute could be overturned by courts. 99From the 1960s, discretionary powers in criminal law expanded across European countries alongside embracing external and internal controls. 100Nevertheless, the varieties of legal traditions resulted in different legal controls over administrative discretion.For example, the degree of judicial review of discretion is marginal by English courts very restricted by German courts; and intensity of control depends on the subject matter in France or the procedure in Italy. 101 the case study of competition law demonstrates below, some IRAs followed the American/British models (high degree of prioritisation powers, either with or without controls), others considerably limited their competition authorities' prioritisation powers.These authorities are either required to investigate all potential infringements, or have limited prioritisation powers subject to various controls.While these models advance the legality principle, they come at the expense of IRAs' effectiveness, efficiency, and independence.

SETTING
Likely due to the incremental evolution of IRAs, so far, no normative framework has been developed to guide IRAs' priority setting rules and practices.As their prioritisation powers are deeply embedded in their respective national administrative and constitutional laws and institutional settings, no 'best-practice' principles guide the setting or control of enforcement priorities.Judicial oversight focuses on cases that IRAs decided to pursue, not on the process and the impact of selection.By comparison, in criminal law enforcement, the impact of choices of 101 English courts exemplified marginal review developing a doctrine of judicial self-restraint in deference to the sovereignty of Parliament and democratic institutions, German courts restricted the margins of unchallengeable discretion allowed to executive authorities, recognising discretion only when it is expressly granted by Parliament.Even when German decision-maker enjoy some discretion (Beurteilungssplielraum and Ermessen), it is limited by fundamental rights and general principles of administrative action.The stringency of review by German courts is compensated by the strict limits placed on standing, which are articulated around rights only, not lesser interests, give access to judicial protection.In France, the depth of judicial review varies according to the subject matter under review and hence, discretionary power is reviewed with different degrees of intensity, and Italian courts usually do not assess the merits of the case, but focus on matters of procedure and form which are consequently of great relevance.See Caranta n 17 above, 188-192.action and inaction were extensively analysed, 102 and there is wide consensus on the benefits of adopting guidelines to limit and control enforcement priorities. 103 the absence of a common benchmark to assess administrative prioritisation, this section fills this gap by examining two approaches for the oversight of administrative discretion in general: the narrow ultra vires doctrine and the broad public interest-based approach.We argue that the public interest-based approach is better suited for the examination of priority setting from the perspective of the rule of law and such an approach is justified given the institutional, practical, personal, and bureaucratic settings of priority setting.

Public interest-based approach to priority setting
The narrow ultra vires doctrine was well-suited for the early reactive-adjudicatory IRAs.Yet, as IRAs became proactive regulators by the late-1960s scholars warned against regulatory capture and private interest regulation emerging from the exercise of administrative discretion. 104In 1969, Davis called for confining, checking, and structuring discretion to prevent injustice and for setting internal controls that are externally reinforced by courts, a 'discovery' that was known to those who have studied criminal law. 105Davis, and other scholars, 106 advocated for a public interestbased approach making use of good governance principles to confine discretion. 107This broad approach is grounded on standards of legality to prevent abuse of power by public bodies, quasipublic, or private bodies with a degree of power. 108Besides focusing on the legislative will, judicial intervention controls the principles guiding administrative discretion and interprets legislation in conformity with fundamental rights and the public interest. 109is shift in the oversight of administrative discretion is significant.The narrow ultra vires principle is grounded on a negative conceptualisation of discretion.Accordingly, discretion is characterised by the absence of legal norms, a choice not legally determined: 110 'what is left of judicial control'. 111The public-interest approach extends beyond this negatively construed view.It acknowledges the autonomy of the administrative decision-maker 'to choose between different alternatives when concretising legal norms with a view to achieving the ends that those norms identify'. 112Instead of focusing on how far courts can go when they review discretion, it examines how legal norms operate in the spheres of discretion that those norms attribute to decision-makers and 'by virtue of absent or limited review, administrative discretion ought to be guided by legalnormative criteria '. 113 This approach can rely to an extent on existing mechanisms embedded in administrative processes to incorporate the public interest dimension of discretion in decision-making, 114 such as the duty of careful and impartial examination and the duty to give reasons.The duty of care orders institutions to examine carefully and impartially 'all the relevant aspects of the individual case', including, the relevant public interests implicated in decision-making. 115The duty to give reasons functions as a self-reflective tool for decision-makers, as it presupposes consideration of various aspects in a given situation and the implications of the chosen option.
By compensating for limited judicial review of administrative procedural guarantees, the principles of good administration 116 function as an 'aid' to the procedural and substantive requirements a modern administration has to comply with.Hence, the principles of good administration can structure the exercise of discretionary powers and are used as a 'standard of practice serving the attainment of administrative justice in compliance with Article 41 of the EU Charter of Fundamental Rights and CJEU case law. 117

Normative benchmarks guiding priority setting by IRAs
Building on the public-interest approach, we suggest five key good governance principles for evaluating priority setting by IRAs: effectiveness, efficiency, independence, transparency, and 111 Caranta, n 17 above.Such an approach, as Mendes n 1 above, 461 argued, fails to capture the complexity of the interaction between legal norms, discretion, and judicial review. 112Mendes, n 1 above, 462. 113ibid. 114What Mendes defends as a unitary concept of discretion which stresses not only the autonomy attributed by legal norms, but also the bounded nature of that autonomy.This unitary approach emphasizes the process of concretization of normative programs delineated in legal norms.In this process, the delimitation and verification of the conditions of action -whether dependent on value concepts or primarily on tools developed in specific scholarly fields -codetermines the definition of the legal solution.Ibid, 464. 115A careful and impartial examination of technically complex factual circumstances need to take into account and balance between various public interests.Careful examination would refer not only to factual assessments, but also to public interest appraisals, given the way in which both aspects are intertwined in the exercise of discretion.Case C-269/90 Technische Universität München EU:C:1991:438 at [14]; Mendes n 1 above, 466. 116See Nehl, n 17Error!Bookmark not defined.above for an overview and in-depth study on good administration as a concept in EU law. 117Ibid, 338; Mendes, n 106 above, 5; H Addink, 'Good Governance: Concept and Context (Oxford University Press, 2019), 251.
accountability.These principles are commonly used to assess the behaviour of public administration and administrative discretion. 118Given the extensive scholarship on good governance principles and their role in administrative procedures and the scope of this article, the following section merely introduces the implications of those principles in the context of priority setting.More specifically, as we explain below, we analyse which important trade-offs between these criteria are made in the various governance models of priority setting.As the case study of competition law enforcement in the next section demonstrates, making specific choices a complex exercise against the diverse national administrative and constitutional rules.

Effectiveness
Effective priority setting denotes IRAs' ability to meet the goals set by the legislation and focus its interventions on achieving these goals. 119Setting clear priorities that are built around a transparent strategy, enhances the credibility of IRAs' action. 120It is essential both for ensuring deterrence and for concretising the typically open-ended administrative provisions.Accordingly, the effectiveness of prioritisation should not only be assessed in quantitative terms (e.g., number of cases or level of the fines imposed), but calls for a balanced portfolio of cases, involving a mix of cases with various levels of complexity, size, short-and long-term effects 121 and risk balancing, enforcing 'classic' infringements and landmark cases that set a precedent and have a greater multiplier effect. 122

Efficiency
IRAs are bound by scarce financial, technical, and human resources unable to detect, investigate, and sanction every possible law infringement.Efficient priority setting rationalises the allocation of resources to deal optimally with cases within a reasonable time.This includes, in 118 At the global level, good governance principals were formulated by supranational organisations.Promoted by the United Nations since the late 1980s, good governance has become an important benchmark for the assessment of the process of decision making by governments and agencies.See OECD, 'OECD Best Practice Principles on the Governance of Regulators' (2012), http://www.oecd.org/gov/regulatorypolicy/governance-regulators.htm;Jennifer A. Elliott et al., 'The Making of Good Supervision: Learning to Say No' in Aditya Narain, Inci Ötker, and Ceyla Pazarbasioglu (eds), Building a More Resilient Financial Sector: Reforms in the Wake of the Global Crisis (International Monetary Fund, 2012). 119Ottow, n 28 above, 87, argues that the structure of the supervisory space has a major influence on the overall effectiveness of oversight.It requires a sufficiently clear mandate, optimal agency design, clear and efficient decisionmaking, appeal procedures and effective tools and instruments; A. addition to the selection of enforcement targets, the choice between (formal and informal) enforcement tools available to IRAs.While efficiency is an indispensable component of modern administration due to the need for speedy technocratic decision-making in a context of rapid market changes and increasingly complex socio-economic and technical issues, overreliance on this principle limits other good governance aspects of priority setting, for example, transparency.

Independence
IRAs' independence was traditionally justified by the technical complexity of regulated markets and the need for expertise. 123Its core is the regulator's legal and functional separation from both market parties and legislative and executive influence, 124 including 'the degree to which the day-to-day decisions of regulatory agencies are formed without the interference of politicians and/or consideration of politicians' preferences'. 125

Transparency
Transparent priority setting entails that IRA's decisions of action and inaction are based on clear and openly communicated legal-economic justification.Originating from the procedural duty to give reasons, transparency functions as a control of discretion. 126Providing sufficient evidence and grounds to justify IRAs' interventions strengthens procedural accountability. 127Clearly formulated, published, and reasoned priorities are important indicators of democratic and legitimate law enforcement. 128Transparency in setting enforcement priorities, moreover, strengthens predictability and allows relevant stakeholders to assess whether a certain behaviour is likely to result in regulatory intervention and to encourage self-compliance. 129Finally, transparency facilitates participation.Open consultations serve as an information-gathering tool, justifying regulatory action and inaction, and help IRAs balance competing public interests. 130Participation through submitting complaints can contribute to an accurate and legitimate decision-making, prevent capture, and invite IRAs to reflect on their position to prevent mismanagement or misuse of powers.

Accountability
Besides being independent, IRAs should be accountable for their enforcement choices and allocation of resources. 131IRAs are required to be 'vertically accountable' towards their political institutions, judiciary, regulates, and general public. 132Often, they are 'horizontally accountable' towards regional or international networks of regulators. 133Accountability can be understood formally through adopting procedural mechanisms to control IRAs' operation and informally, as a substantive requirement, focusing on meaningful interaction between the IRA and its audience. 134opting internal controls and clear communication on priority setting can facilitate the IRAs' formal and substantive accountability. 135

CASE STUDY: COMPETITION LAW IN EUROPE
This section presents the emergence of competition authorities ('CAs') in Europe, and the evolution of their prioritisation powers.As mentioned above, EU competition law offers important insights into priority setting as CAs are the second most common IRAs globally, and were among the first IRAs in the US and Europe, serving as a blueprint for other IRAs. 136The modernisation of EU competition law in 2004 was a defining moment in the development of priority setting.Prior modernisation, there was limited consideration of priority setting.While the Commission was entrusted with a high degree of prioritisation powers, similar to American IRAs, matters of prioritisations only became prominent after the implementation of Regulation 1/2003.The new enforcement system opened the door for setting priorities by the Commission and national competition authorities ('NCAs') while imposing few controls on the exercise of such discretion.

The emergence of competition laws and authorities in Europe
Competition law and policy were traditionally not seen as a core pillars of governments' economic governance in Europe. 137Modern competition law was gradually introduced in the aftermath of the Second World War.While drawing inspiration from the US, European enforcement systems were homegrown and differed considerably from one another.
Accordingly, until the 1990s, there were hardly any independent CAs in Europe.
Established in 1948, the British MRPC was the first CA, yet it had weak enforcement powers. 138t, British competition law remained cautious, incomplete, and under-enforced, until late 1990s. 139The Competition Act of 1998 and the Enterprise Act of 2002 were the first to introduce an effective competition regime, where investigation and decision-making powers are held by the OFT (from 2014 CMA), and subject to appeal to the Competition Appeal Tribunal (CAT).
In 1957, the German Federal Cartel Office (Bundeskartellamt) was established as the second competition authority in Europe under US pressure.Building on the Ordoliberal enforcement vision, the Bundeskartellamt was created as a highly independent CA, having the sole responsibility for enforcement and separated from the state bureaucracy. 140Its independence was regarded as a sine qua non of the modern Rechtsstaat, and justified its position outside the regular administrative hierarchy. 141The Ordoliberals' conviction that the office had to be autonomous was related to their experience during the Weimar period when the capacity of cartels and large corporations to attempts at control by putting political pressure on the executive branch.David Gerber 'Law and Competition in the Twentieth century' (Oxford University Press, 1998) 254-55. 141Gerber n 137 above, 282. 142The independence of the Bundeskartellamt originated from its internal organisation and procedures that protect it from political influence, and its juridical nature, which is reflected in its special status that is placed outside the regular administrative hierarchy.The role of the Bundeskartellamt in the German economy and legal and political system is well illustrated by Gerber arguing that '[T]he GWB was not just another law, and the Bundeskartellamt was not just another administrative office.Together, they symbolised rejection of a failed regime and belief in a democratic alternative.'Gerber n 137 above, 282.modern competition rules prohibiting cartels and abuse of a dominant position. 143When the enforcement of EU competition law was negotiated in 1961, Germany was the only Member State with experience and a clear enforcement vision.Therefore, German experience was prominent in drafting Regulation 17/62, and in particular, its institutional design of Directorate-General for Competition of the Commission (DG COMP, formally DG IV) as a centralised, independent, quasi-judicial body. 144From its early days, DG COMP stood out among other Directorates General because of its autonomy, judicial functions, and direct influence on the economy. 145Yet, when compared to its American counterpart, it does not function as a fully independent IRA: 146 final decisions are adopted through Commissioners' vote, who are political appointments of the Member States. 147Suggestions to transform DG COMP into a fully independent IRA were consistently rejected. 148Reflecting Ordoliberal views, Regulation 17/62 granted extensive enforcement powers to DG COMP.It established a centralised-notification system, with the Commission examining all potentially anti-competitive agreements before implementation, and having the sole power to grant exemptions.
By the 1990s, competition law became a 'common core' in all EU Member States, 149 due to successful market integration, the process of EU constitutionalisation, and strong supranational enforcement mechanisms under Regulation 17/62.As a form of 'Europeanisation', Member States adopted competition laws and created CAs following the EU model. 150These developments stimulated the 'modernisation' of EU competition law enforcement.Regulation 1/2003 created a multilevel enforcement system where the substantive EU provisions (Articles 101 and 102 TFEU) are enforced by the Commission and 27 CAs. 151The Regulation reformed the procedural rules governing the Commission's enforcement, however, it neither intervened with NCAs' procedures nor with their institutional design. 152It merely obliged each Member State to designate a CA responsible for the application of Articles 101 and 102 TFEU. 153As NCAs and national courts had no power to exempt an agreement under Article 101(3) TFEU, companies were incentivised to notify their agreements to the Commission to get legal certainty concerning compatibility of their responding to all notifications led to reactive enforcement and consumed much of its resources, leaving few opportunities for ex-officio investigations.During this period, the Commission's prioritisation was mostly limited to rejecting complaints 156 and selecting the order and tools of responding to notifications (formal-binding decision or informal comfort letters). 157While the centralised-notification system left little room for priority setting, it guaranteed uniformity and legal certainty for firms while developing expertise in a sensitive supranational setting. 158sembling the Anglo-Saxon model, the Commission's prioritisation was subject to limited external controls.The selection of enforcement targets was merely bound by the political control of the European Parliament and the Economic and Social Committee, which reviewed the Commission's activities as presented in its annual reports.Similar to American antitrust agencies discussed above, the Commission's prioritisation powers were not significantly overseen by internal controls.In 1963, the Commission adopted a resolution, setting informal-internal guidance for selecting its cases.Yet, given the Commission's reactive enforcement, this resolution had only limited impact. 159It was not fully published, hence did not increase the Commission's accountability nor transparency, and it merely summarised by the Commission's report, 160 and the Commission refrained from discussing its implementation in the subsequent years.
Nevertheless, the resolution is remarkable, as it demonstrated the complexity of the exercise.
The Commission declared that it would give priority where: (i) a decision is required to bring an infringement to an end; (ii) when actions are pending before national courts; (iii) to discover and examine agreement that were not notified; (iv) to respond to notifications.The Commission added substantive criteria, noting it would consider the 'type and gravity of the restriction of competition, its economic importance for the Common Market, an endeavor to spread the cases over the various economic sectors, and the effects of the subsequent decisions as a precedent for the interpretation and observance of the rules of competition, and thus for the clarity with which the law can be understood by enterprises'.The limited external and internal controls could also be explained by the Commission's institutional and procedural setting.While only limited information is available, at least up until the 1970s, choices not to bring a formal action was the responsibility of a single member of DG COMP's staff, and very few staff members were informed.Decisions to open an investigation, by comparison, were reviewed by at least 20 officials, and were more likely to generate debate on the selection of cases. 162This demonstrates how procedures influence case selection, as will be elaborated below.
The European Parliament pressured the Commission for greater transparency of its enforcement discretion.In 1986, for example, it called the Commission to clarify the criteria for case selection (3522 cases pending at the end of 1986) and the choice between a formal decision or informal settlement procedure. 163In response, the Commission disclosed its prioritisation practice in following year's annual report.Suggesting that the 1963 Resolution was not fully respected, the Commission declared that it would give priority to cases involving 'questions of broad political significance' and take cases ex officio or respond to complaints with reference to the 'seriousness' of the alleged infringement.When assessing the order of responding to complaints and notifications it would consider the urgency of the matter, for example when national legal proceedings are pending, but otherwise 'deal with them chronologically'. 164 a line of judgments in the 1990s, the EU Courts introduced important procedural controls on the Commission's discretion handling complaints 165 clarifying that the Commission had the power to reject complaints on priority grounds 166 despite the fact it only begun to assign degrees of priorities to complaints at the end of the 1980s. 167In its landmark Automec II judgment of 1992, the General Court (GC) discussed, for the first time, the Commission's priority setting powers and its limits. 168The GC rejected the applicant's submission that the Commission was bound by the legality principle when assessing complaints 169 acknowledging the Commission's wide discretion in this regard and explaining that the Commission could only effectively fulfil its task of implementing EU competition policy if it had the power to reject complaints. 170The Court limited its review to checking whether the Commission complied with the duty of care, namely that it examined carefully the factual and legal particulars brought to its notice, and no manifest error of law, appraisal, or misuse of powers took place. 171These principles were endorsed by the CJEU, in various cases later, 172 and are currently enshrined in the 2004 Notice on the handling of complaints. 173Following Automec II, the Commission declared that it would use this 'discretion with moderation' referring complainants to national authorities or courts more often than before, particularly where it was clear that the national enforcement enabled complainants to resolve the matter. 174 conclude, aside from some fundamental procedural guarantees for the rejection of complaints and the Parliament's pleas for greater transparency, matters of priority setting were overlooked in EU competition law enforcement prior to modernisation.

Following modernisation: wide priority setting powers and national divergence
The entry into force of Regulation 1/2003 strengthened the Commission's competencies and ability to set priorities.Abolishing the notification system and granting NCAs enforcement powers allowed the Commission to select its cases and dedicate its resources to a proactive enforcement.
In particular, the Regulation aimed to reduce the number of complaints addressed to the Commission, when NCAs 175 could effectively deal with them, or when complainants could bring private actions before national courts. 176Enhancing the Commission's and NCAs' priority setting powers was, however, not accompanied by EU controls over the exercise of such powers: EU law does not impose external controls beyond codifying the jurisprudence on the Commission's powers to reject complaints, and alike, the Commission refrained from adopting internal controls.One exception is its Guidance on Article 102 TFEU enforcement priorities. 177spite its title, it lists substantive criteria for applying the Article, and does not set enforcement 171 ibid, para 80.To this end, the Commission 'should balance the significance of the alleged infringement as regards the functioning of the common market, the probability of establishing the existence of the infringement and the scope of the investigation required in order to fulfil, under the best possible conditions, its task of ensuring that Articles [101] and [102] are complied with'.Automec, n 156 above, para 81-86. 174European Commission, 'Twenty Third Report on Competition Policy' (1994), 120-121. 175Commission Notice on the Handling of Complaints, para 8. 176 ibid, para 21, 24-25, 36-39.The Commission may reject a complaint in accordance with Article 13 of Regulation 1/2003, on the grounds that a Member State CA is dealing or has dealt with the case. 177This informal and non-binding policy paper declares that although both exclusionary and exploitative conduct falls within the scope of this Article, the Commission will only focus on the former, which is typically more harmful to consumers.See Commission Guidance on Article 102, para 7-8.priorities. 178Hence, following modernisation the Commission still follows the model of high degree of priority setting powers, with little controls.Moreover, Regulation 1/2003 does not regulate NCAs' prioritisation powers.Pursuant to the EU principle of procedural autonomy, the powers, scope, and limits for setting priorities are determined by national procedural, administrative, and constitutional laws.As our empirical research shows below, certain Member States granted wide priority setting powers to NCAs, either without controls, akin to the Commission and Anglo-Saxon IRAs, while others limited those powers by imposing external controls.In some Member States, NCAs adopted internal controls to structure their discretion, while others left the exercise of discretion uncontrolled.
The Commission first voiced concerns over this diverging landscape of prioritisation rules and practices in its 2009 report on Regulation 1/2003.It noted that the NCAs' priority setting powers were an important aspect of divergence that 'may merit further examination and reflection'. 179ter, it called for harmonisation within the cooperation mechanism of the European Competition Network (ECN).The ECN's Working Group on Cooperation Issues and Due Process was to monitor convergence among the Member States and provided an overview of the different systems. 180In 2013, it adopted a Recommendation on the Power to Set Priorities, calling for harmonisation and converging towards the Commission's model.The Recommendation, however, neither indicated what analysis justified this choice, nor what its implications were.It simply argued that it would 'enhance effectiveness and efficiency in the enforcement (…) by allowing them to focus their action on the most serious infringements/sectors and areas most in need of their action, thereby increasing the impact of their action for the benefit of consumers'. 181bsequent to the Commission's public consultation on how to empower NCAs to be more effective enforcers, 182 the ECN+ Directive was the first to codify rules concerning NCAs' priority setting.However, these rules are limited to three aspects leaving core features of prioritisation unaffected.First, the Directive obliges Member States to enable their NCAs to have the power to set priorities for the enforcement of Articles 101 and 102 TFEU. 183This has limited effect, as prior to the Directive, all NCAs were legally competent to open ex-officio investigations. 184Second, NCAs should have the power to reject complaints on priority grounds.The legality principle, in other words, could no longer guide rejection of complaints.Third, NCAs should set their priorities independently, i.e., without taking instructions from public or private entities. 185

TYPOLOGY OF ENFORCEMENT DISCRETION
The lack of attention concerning IRAs' priority setting powers is evident through the lack of shared terminology and benchmarks.Our empirical study and interviews with CAs revealed diverging understandings and interpretations of priority setting and numerous (sometimes conflicting) meanings and objectives.For example, when asked about prioritisation, many authorities have not distinguished between 'agenda setting' and 'substantive criteria' as we define them below.This section introduces a new typology to guide the analysis of priority setting rules and practices.More specifically, it defines seven aspects of setting priorities and the possible external or internal controls guiding them, as summarised by Figure 1.These aspects were identified via a bottom-up approach systematically analysing rules and practices of CAs in Europe.The section presents how EU and national rules govern each of the seven aspects, their practical implementation and impact on good governance principles.

Figure 1: typology of priority setting
As elaborated below, these seven aspects are interdependent; choices concerning one aspect often affect others.For example, limited de jure competence frequently entails a formal procedure of priority setting and limits the IRA's ability to de facto select cases.At the same, time, pursuing an independent and efficient priority setting practice decreases transparency and accountability of prioritisation by limiting third parties' participation as well as the possibility for judicial review.

Agenda setting
Agenda setting is a list of ex-ante periodic enforcement agenda, publicly declaring that certain sectors or practices are a priority.It is often referred to as annual/action/work plan, or a strategy statement.
Setting an agenda requires IRAs to pronounce their enforcement strategies in advance, explaining how they plan to make use of their enforcement powers and budget.It can strengthen accountability, transparency, and predictability of their action. 186Agendas enhance IRAs' independence and legitimacy by allowing them, as expert-driven decision-making bodies to select their strategies free from external intervention.It fosters effectiveness and efficiency by encouraging proactive enforcement, instead of reacting to complaints or leniency applications The impact of an agenda on deterrence remains disputed.While an agenda may deter firms operating in the identified priority areas, it provides firms an opportunity to conceal evidence of infringements. 188The Greek NCA, for example, decided not to publish its internally-adopted agenda. 189Moreover, deterrence is threatened when an agenda is not regularly updated, focuses on a limited number of sectors and practices, or when enforcement focuses primarily on the identified areas and neglects others.A degree of uncertainty may generate higher compliance levels. 190e debate on the merits of agendas is reflected by our empirical findings.Only 48% of the CAs adopted agendas, all in the form of internal control. 19128% of the Member States obliged NCAs to adopt an agenda as a matter of national law, from which 3% had to report them to their parliament and 7% to their government. 192Reporting obligations can constrain CAs' independence, yet increase their accountability and legitimacy.17% of the CAs adopt an agenda following a public consultation.The Dutch NCA, for example, invites stakeholders to comment on its draft agenda via roundtable meetings, a dedicated online website, and social media.Public participation improves the quality of the agenda, increases effectiveness, and promotes accountability and transparency.

Legal competence to prioritise (de jure)
The de jure competence to priorities refers to the IRAs' ability based on law, to choose which cases to pursue and which to disregard.Law enforcement theories distinguish between IRAs who follow the opportunity principle and enjoy full de jure competence with high degree of discretion; and those bound by the legality principle, obliging them to initiate an investigation into any potential infringement coming to their attention. 193The degree of legal competence to prioritise reflects a jurisdiction-specific trade-off between efficiency and independence on the one hand, and equality before the law, accountability, and transparency on the other.
Our As mentioned above, until recently the NCAs' legal competence to prioritise was not addressed by EU law.The ECN+ Directive now links de jure competence to efficiency, effectiveness, and independence of NCAs. 196However, the obligations in the Directive are drafted in general terms, leaving the degree of prioritisation discretion and concrete prioritisation competences to national preferences.
The empirical findings as summarised in Figure 2 point to great divergence among jurisdictions and demonstrate that some CAs who generally enjoy wide prioritisation powers (opportunity principle or the public interest test) are limited when assessing complaints.This represents a trade-off between the effectiveness of broad de jure competence and the accountability and transparency embedded in third parties' participation.In Finland, for example, the authority enjoys a wide discretion to prioritise, but will reject a complaint only when the 'matter is manifestly unjustified'. 197In Cyprus, the authority may 'push-back in the line' a complaint on priority grounds,

Ability to set priorities (de facto)
The power to set priorities is often implicitly constrained by practical settings in which decisions are made. 200The de facto ability to set priorities refers to IRAs' practical (human, financial, and technical), institutional, and organisational capabilities that affect their course of action.
Adequate resources and capabilities are not only essential for effective and efficient enforcement, but safeguard IRA's independence (i.e., budget allocation without prejudice to national budgetary rules and procedures), transparency and accountability.competence.An obligation to respond to notifications, complaints, or referrals, for example, leave IRAs limited capacity to autonomously select its priorities.Their institutional design also influences de facto capacity.Prioritisation can be more complex in multi-function IRAs, which must allocate resources across a broader range of activities, some of which may involve obligatory (e.g.regulatory) tasks, while others may be more discretionary. 203spite the importance of de facto ability, and significant challenges reported during our interviews, 204 it is subject to limited control.The EU Courts confirmed that case allocation within the ECN does not take into account the actual ability of an NCA to deal with the case. 205In addition, the ECN+ Directive acknowledges the link between the level of enforcement and NCAs' budget, skills, 206 and independence, 207 yet only prescribes a vague obligation, requiring Member States to ensure that NCAs have the 'necessary resources to perform their tasks'. 208In practice, NCAs' budgets vary considerably, 209 even between countries having similar GDP. 210The Directive, nevertheless, introduced national external controls requiring NCAs to submit periodic reports on their activities and resources to national governments or parliaments.Those publicly available reports should include information about the resources that were allocated in the relevant year, and any changes compared to previous years. 211

Procedure to prioritise
Administrative procedures and institutional processes significantly impact prioritisation choices, in particular, those related to (i) the type of prioritisation decisions (reasoning, publication and availability of judicial review; (ii) time limits; (iii) participation of third parties; and (iv) the institutional setting of the decision-making.Similarly, to the competence to prioritise, the figure shows that many CAs are subject to different procedural rules when they reject complaints.Some CAs that are not required to investigate each complaint (see Figure 2), including the Commission, 212 are under the duty to examine all matters of fact and law brought to their attention and to provide reasons (CAs marked with asterisk in Figure 3).Second, prioritisation is affected by time limits for adopting such decisions.The effects of time limits vary; fixed limits may encourage efficient prioritisation process, but also incentivise IRAs to refrain from investigating complex cases.Placing no -or very long -limits, may provide a legally anchored way to delay proceedings.Around one-third of the CAs are subject to external or self-imposed time constraints, whose duration and scope vary considerably. 214For example, the Italian CA is obliged to specify a deadline for completing investigation, 215 and failure to comply may lead to the annulment of the decision altogether. 216Italian jurisprudence linked this obligation to the right to a fair trial (Article 6 ECHR) and good administration (Article 41 CFR).
Third, third party participation also affects the selection of cases.In many jurisdictions, parties having a legally relevant interest can report a possible violation and participate in the decisionmaking procedure.They can appeal CAs' decisions not to pursue a case.Participation of third parties is grounded on the instrumental function of their intervention.It helps CAs to identify an accurate representation of the factual situation and reach a materially correct decision, corresponding to the facts and the public interest. 217Participation rights serve as external control enhancing the transparency and accountability of prioritisation.Complementing the function of judicial review, third parties may challenge the CAs' proposed action and warn against errors. 218ile, granting third parties' access and participation rights comes at the expense of efficiency, it enhances deterrence and law compliance.
The conditions of access to the procedure and third parties' participation rights during the CAs' procedures vary considerably from one legal system to another. 219While many CAs have explicit rules in relation to formal complaints, 220 national approaches diverge as to whom would be recognised as an interested party and whether their participation rights also extend to other interested third parties. 221inally, IRAs' institutional organisation also shapes priority setting, for instance, by how decision-making tasks are separated or integrated from other enforcement tasks.Institutional choices involve trade-offs between the swiftness of decision-making and expertise vis-à-vis quality control, transparency and legitimacy. 222Prioritising choices of IRAs whose investigation and decision-making (adjudication) tasks are divided between two different institutions (eg administrative authorities and courts, external tribunals), are subject to review already within this first stage of the procedure. 223milar considerations arise from the leadership model of the authority. 224Some IRAs, take prioritisation choices by a single person, e.g. the agency head or a specific unit.This unitary executive model has the advantage of faster and consistent decision-making.Other IRAs, most EU NCAs take decisions by a group of staff members -either the management or a designated group -by a vote, consensus, or a combination of these two.Multi-member decision-making allows for greater expertise, transparency, and legitimacy in the decision-making and may better shield against political influence. 225

Substantive criteria
Substantive criteria refer to external or internal criteria guiding IRAs' decisions whether to pursue or disregard a case.Unlike agenda setting, this aspect does not refer to a specific sector or practice, but to case-specific circumstances.
Setting substantive criteria streamlines the exercise of the IRAs' discretion, promotes efficient use of resources, focus enforcement efforts on deterrence and clarify the rules.It prevents both underenforcement (by encouraging enforcement in cases of legal or doctrinal importance) and overenforcement (discouraging enforcing practices having only limited impact on consumers and markets).Publishing substantive criteria enhances IRAs' accountability and predictability requiring them to articulate their strategy in advance by explaining how it plans to make use of its enforcement powers and budget.Imposing external substantive criteria limits IRAs' independence, but enhances its accountability towards political institutions and allows stakeholders' participation via public consultation.
In competition law, while EU law does not offer substantive criteria, some Member States or NCAs adopted such criteria.According to   While different models may characterise other types of IRAs, we believe that the case study of competition law demonstrates the impact of prioritisation choices on the trade-offs across the five good governance principles.

Model I -high degree of prioritisation, with external or internal constraints
Under the first model, IRAs enjoy a relatively high degree of prioritisation powers.They have both de jure competence and de facto ability to select cases.Their prioritisation discretion is, nevertheless, structured and controlled by a set of formal or informal rules, either imposed externally by law (Greek CA) or jurisprudence (Dutch CA), or adopted by IRAs internally (CAs in Finland, Netherlands, UK).These control mechanisms may focus on various procedural aspects of prioritisation (e.g., unit deciding on priorities, selection process, publication, and reasoning requirements) or substance (agenda and the substantive criteria for selecting cases).
A high degree of prioritisation powers allows IRAs to effectively select cases and reject lowpriority complaints while focusing on matters of legal or doctrinal importance.Balanced case selection -based on carefully curated criteria that are clearly communicated to stakeholders and the larger public; is subject to consultation and periodically reviewed -leads to more effective The strength of this model is increased accountability and transparency, which are expected to ensure that prioritisation decisions are taken in the public interest.Strong external controls and third parties' participation are important benefits, in particular, for newly established IRAs lacking strong reputation, legitimacy, or experience.The limitations placed on choices not to pursue a potential infringement may contribute to a balanced portfolio of cases, as long as the IRA's resources reasonably match its workload.
The weakness of this model is the costs of extensive external controls that reduce IRAs' independence (effectiveness and efficiency) and lead to a reactive enforcement system leaving little room for strategic planning.Such an effect could undermine IRA's power to use its expertise to select cases based on the public interest.When combined with limited resources, this model may result in wasting efforts on insignificant cases.

Model IV -low degree of prioritisation
IRAs of the fourth model have limited priority setting powers.Bound by the legality principle, they are obliged to investigate every possible law infringement coming to their attention.These decisions must be taken formally, be reasoned and published, and are subject to judicial review.
The strength of this model is linked to the justification of the legality principle: safeguarding equality before the law and deterrence, including sensitive and complex cases.IRAs enjoy the highest degree of transparency and accountability and provide considerable room for third parties in the decision making.
The weakness of this model is the high procedural costs.The obligation to investigate all potential violations can lead to inefficient use of resources and often results in limited de facto ability to start ex officio investigations.Even when IRAs decide to set an enforcement agenda or substantive criteria to guide priority setting these will have limited effects in practice.Responding to every possible violation can reduce their independence and credibility. 241CONCLUSIONS Priority setting by IRAs is an invisible, yet essential component of regulatory enforcement.Scholars, policy-makers, and enforcers 'implicitly assume that laws are somehow self-enforcing and that there is full compliance' 242 and, as the blindspot of administrative discretion, they overlook how IRAs decide which cases they pursue and which they disregard.Still, those choices are vital given IRAs' finite resources, and as a form of concretising open-ended administrative norms.The power to set priorities allows IRAs to effectively enforce regulation that require complex socio-economic and technical assessments, based on technocratic expertise, and free from external intervention.
Intertwined with the rise of the modern regulatory state, IRAs were born as a historical accident.They emerged from Anglo-Saxon adjudication-based 'Commissioners' and their priority setting powers were incrementally developed as IRAs transformed into proactive regulators.These early IRAs were delegated wide discretionary powers to provide expert decision-making by independently setting their own priorities and subject to only limited external controls prescribed by the ultra vires principle.As IRAs expanded to Europe and beyond, new models emerged that were profoundly shaped by national administrative and criminal laws as well as by their non-legal context, for example, institutional setting.While many jurisdictions followed the Anglo-Saxo blueprint, others adhered to the legality principle and considerably limited their IRAs' prioritisation powers.However, no specific legal or regulatory theory exist that shaped this incremental and disperse development of priority setting powers, and that could define the various aspects, aims, and controls of prioritisation.This article opens the 'black-box' of priority setting.First, it makes priority setting discernible by shedding light on the historical development of IRAs' priority setting powers.Second, the article deconstructs its composite nature by offering a novel typology of seven aspects of priority setting in IRAs' pre-decisional, decisional, and post-decisional stages.Finally, it defines normative benchmarks to analyse and evaluate IRAs' priority setting rules and practices against the principles of good governance.While historically, oversight of priority setting rules and practices was negatively construed, focusing on control via judicial review, we advocate for a broad public interest approach.Complying with the rule of law, it incorporates legal-normative criteria of good administration.It offers an overarching approach to assess IRAs' priority setting, while being sensitive to its composite nature and to their national institutional, practical, and bureaucratic embeddedness.
Our analysis points to four concluding observations.First, the Anglo-Saxon and EU IRAs that emerged since the end of the nineteenth century and significantly influenced prioritisation models elsewhere, have been characterized by wide priority setting powers and a focus on independence, efficiency, and effectiveness, while vastly overlooking the democratic rationales of transparency and accountability.Besides being subject to limited external and internal controls, courts also refrained from reviewing their prioritisation decisions as long as IRAs stayed within the boundaries of their legislative mandate.
In 1956, the Restrictive Trade Practices Act created new institutions and enforcement powers.It established a Registrar of Restrictive Trading Agreements and imposed the duty to notify certain anti-competitive agreements.Adjudication powers were granted to the Restrictive Practices Court (RPC), a newly established judicial body, independent of political and economic pressure.In 1965, the MRPC was transformed into the Monopolies and Mergers Commission (MMC), upon expanding its powers of investigation also to merger controls.
Héritier and D. Lehmkuhl, 'Governing in the Shadow of Hierarchy: New Modes of Governance in Regulation', in Héritier and M. Rhodes (eds.)New Modes of Governance in Europe: Governing in the Shadow of Hierarch (Palgrave Macmillan, 2011), 66. 120 Ottow, n 28 above, 76. 121 Ibid, 160. 122Commission Staff Working Document 'Impact Assessment' accompanying the document proposal for a Directive of the European Parliament and of the Council to empower the competition authorities of the Member States to be more effective enforcers and to ensure (SWD/2017/0114 final, 22.3.2017),part I, 46.
15 years after modernisation, Directive 2019/1 (the 'ECN+ Directive') of 2019 harmonised NCAs' powers and institutional settings to a limited extent.Aiming to create more effective national enforcement, the Directive obliges Member States to provide NCAs certain investigative and enforcement powers, and introduces general provisions to safeguard NCAs' independence, and accountability.For this purpose, it includes few provisions on priority setting, which will be discussed below.Nevertheless, beyond a minimum level of harmonisation, the Directive does not substantially converge national institutional and procedural settings. 154rior modernisation: limited attention to priority setting Prior to modernisation, the centralised-notification system limited the relevance of prioritisation.Given the lack of competition law culture and NCAs' limited powers, Member States had scarce enforcement, 155 which in turn, left prioritisation unaddressed.For the Commission, the burden of 151 Council Regulation (EC) No 1/2003 of 16 December 2002 on the implementation of the rules on competition laud down in Articles 81 and 82 of the Treaty [2003] OJ L1/1.Also see KJ Cseres 'Competition Law Enforcement Beyond the Nation-State: A Model for Transnational Enforcement Mechanisms?' in H-W Micklitz and A Wechsel (eds), The Transformation of Enforcement (Hart, 2015), 319-339.On the transformation of EU competition law, see F Cengiz, 'Regulation 1/2003 Revisited' (2009) 42 TILEC Discussion Paper 26, 30; K.J. Cseres and A. Outhuisje, 'Parallel Enforcement and Accountability: The Case of EU Competition Law' in M. Scholten and M. Luchtman (eds.),Law Enforcement by EU Authorities: Political and Judicial Accountability in a Shared Legal Order (Edward Elgar, 2017), 82-114.152 The Regulation, nevertheless, includes certain provisions affecting the powers of NCAs.Article 5 lists the powers of NCAs when they apply Articles 101 and 102 TFEU and what type of decisions the NCAs can take in such cases, without harmonising the procedural rules to be followed by the NCAs.As national procedures for the application of Articles 101 and 102 TFEU were not harmonised by the Regulation, they remained subject to general principles of EU law, in particular, the principles of effectiveness and equivalence and the observance of fundamental rights enshrined in the Charter of Fundamental Rights of the European Union and the European Convention on Human Rights.The procedural differences had been addressed to some extent in Articles 11 and 12 of Regulation 1/2003 through the cooperation of the NCAs within the ECN.European Commission, 'Commission Staff Working Paper accompanying the Report on Regulation 1/2003' SEC (2009) 574 final, para 200; European Commission, 'Commission Staff Working Paper, Enhancing Competition Enforcement by the Member States' Competition Authorities: Institutional and Procedural Issues' COM (2014)453, para 43.153 These authorities could be administrative or judicial in nature, as long as, they could guarantee that the provisions of Regulation 1/2003 were effectively complied with. Reglation 1/2003, Article 35.Point 2 of the Notice on cooperation.Case C-176/03, Commission of the European Communities v Council of the European Union, ECR I-7879, paras 46-55.154 K.J. Cseres, 'The Implementation of the ECN+ Directive in Hungary and Lessons Beyond' (2019) 12(20) Yearbook of Antitrust and Regulatory Studies (YARS) 55.

Figure 2 :
Figure 2: de jure competence (prior to the implementation of the ECN+ Directive) Opportunity principle (high discretion) Public interest (medium discretion) Legality principle (low discretion) (a) All procedures, except complaints (b) Rejecting complaints D Beyond IRA's resources and its staff skills, 202 de facto ability is often tied to the de jure

First, the type
of prioritisation decisions NCAs adopts and the external and internal controls imposed on the reasoning and publication of such decisions are assessed.As summarised by Figure3, some CAs are required to adopt a formal, reasoned, and published decision explaining why they disregarded a specific case.Such a decision is subject to judicial review.Prioritisation choices of other CAs are informal, unreasoned, and not published.As a matter of internal procedure, such choices are generally unreviewable by courts.The nature of prioritisation decisions represents a trade-off between efficiency, transparency, and accountability: taking a formal, reasoned, and published decision reduces efficiency, but helps communicate the IRA's position to the public and can be reviewed by courts.

Figure 3 :
Figure 3: type of decision, reasoning, and publication Notes: * Those CAs adhere to different rules when rejecting complaints + There are two types of decisions concerning case initiation in the Netherlands.One is originating from so-called enforcement request (handhavingsverzoek, defined in Article 1:3 (3) of Dutch Administrative Act) and as such they will always take the form of formal decisions that are reasoned and partly or fully published.The other type of decision concerns informal signals that are only internally reasoned and not published.++With the exception of the rejection of super-complaints.

Figure 4 ,
Figure 4, 38% of the CAs are guided by external substantive criteria, set by the national legislature, government, or judiciary as external control, 68% by internal substantive criteria, adopted by CAs as internal control (mostly published and publicly available), and 20% are not guided by any external or internal criteria at all.

Figure 4 :
Figure 4: external and internal substantive criteria

( a )
External criteria (adopted by the legislature or government) (b) Internal criteria (adopted by the CA)
33Ibid.34ibid.35PaulP.Craig, 'Administrative Law' (Sweet & Maxwell, 2016), 60-61.36Thisexplains why key theories and insights of the regulatory processes find their foundations in the history of US independent regulators.See C Carrigan M Febrizio, 'Tracing the Development of U.S. Independent Regulators in Martino Maggetti, Fabrizio Di Mascio, and Alessandro Natalini (eds.),Handbook of Regulatory Authorities (Edward Elgar Publishing, 2022).

was a Supreme Court of the United States case decided regarding whether the United States President has the power to remove executive officials of a quasi-legislative or quasi-judicial administrative body for reasons other than what is allowed by Congress. The Court held that the President did not have this power
. Carrigan and Febrizio, n 36 above, 13.
Albert Venn Dicey, 'Introduction to the Law of the Constitution' (Macmillan, 1959).
96 88 Majone, n 30 above. 89Fabrizio Gilardi, Delegation in the Regulatory state: Independent Regulatory Agencies in Western Europe.(Cheltenham: Edward Elgar Publishing, 2008). 90G Majone 'The rise of the regulatory state in Europe' (1994) 17 West European Politics 77; M Loughlin and C Scott 'The regulatory state' in P Dunleavy et al (eds) Developments in British Politics 5 (London: Macmillan, 1997). 91First to France the Netherlands, Spain and other jurisdictions.See Mark Thatcher, 'Delegation to Independent Regulatory Agencies: Pressures, Functions and Contextual Mediation' (2002) 25(1) West European Politics 125, 128. 92Ibid, 133.Delegating those tasks to IRAs benefited national governments by shifting the reasonability of adopting and applying complicated or unpopular EU laws, especially when they counter national standards. 93Majone, n 30 above, 6; David Levi-Faur, 'The Global Diffusion of Regulatory Capitalism' (2005) 598 The Annals of the American Academy of Political and Social Science 12; Fabrizio Gilardi, Jacint Jordana, and David Levi-Faur.'Regulation in the Age of Globalization: the Diffusion of Regulatory Agencies Across Europe and Latin America' in Graeme A. Hodge (ed.), Privatization and Market Development: Global Movements in Public Policy Ideas (Edward Elgar Publishing, 2006); Thatcher, n 91 above, 133; S Wilks and I Bartle, 'The Unanticipated Consequences of Creating Independent Competition Agencies' (2022) 25(1) West European Politics, 148, 151. 94Ibid.These theories are as old as the study of modern administration. 95P.A. Hall and D. Soskice 'An Introduction to Varieties of Capitalism' in P.A. Hall and D. Soskice (eds), Varieties of Capitalism: The Institutional Foundations of Comparative Advantage (Oxford University Press, 2001). 96Davis, note 5 above; J. Herrmann, 'The German Prosecutor' in K.C Davis (ed), Discretionary Justice in Europe and America (University of Illinois Press, 1967).In Germany, for example, the legality principle was a key component of the establishment of the Reich in 1871.Tak, n 77 above, 2. Shawn Marie Boyne, 'The Cultural Limits on Uniformity and Formalism in the German Penal Code' (2012) 58 Crime, Law and Social Change 251, 252.

142
Besides those two early national examples, competition law enforcement was limited to supranational application by the European Commission.The first supranational competition law provisions were included in the Treaty on the European Coal and Steel Community (ECSC) of 1951.When the Treaty of Rome was signed in 1957, none of the six signatory countries had 137 Majone,, n 30 above, 50. 138Heald, n 56 above, 38-39.Also see Giandomenico Majone, 'The Rise of Statutory Regulation in Europe' in G Majone (ed) Regulating Europe (London: Routledge, 1996), 47-8. 139With the reforms of 1998 and 2000 the competition provisions in the Fair Trading Act 1973, the Restrictive Trade Practices Act 1976, the Resale Prices Act 1976 and the Competition Act 1980 were swept away.See Scott, n 58 above.
161agreement with EU competition law.European Commission, White Paper on modernisation of the rules implementing Articles 85 and 86 of the EC Treaty (Brussels, 28 April 1999), para 6. 156 Case T-24/90 Automec Srl v Commission EU:T:1992:97 157 The power to choose the order of responding to notifications was confirmed by the GC in Case T-5/93 Tremblay and Others v Commission, ECLI:EU:T:1995:12, para 60; Case T-62/99 Sodima v Commission, ECLI:EU:T:2001:53, para 36.For the latter, the Commission had to choose between responding to a notification in a formal-binding decision or by means of informal comfort letters.It is estimated that approximately 96% of the cases were resolved by informal means, see Ivo Van Bael, 'The Antitrust Settlement Practice of the EC Commission' (1986) 23 Common Market L. Rev. 61. 158C Harlow and R Rawlings, 'Process and Procedure in EU Administration' (Hart Publishing, 2014), 199. 159As reported in Commission of the European Communities, 'Seventeenth Report on Competition Policy' (1988), 23-24. 160EEC.Commission, Seventh General Report on the Activities of the Community (1 April 1963-31 March 1964), 68-69.Also see Meessen n 126 above, 92. 161Ibid.

186
187stein, C. R., Kahneman, D., Schkade, and D Ritov, I., 'Predictably Incoherent Judgments' (2002) Stanford Law Review, 54, 1153-1215.regardless of their impact on markets and society.187Anagenda guides staff members in deciding whether to open an ex-officio investigation or to reject a complaint, and what enforcement tools to use in respective cases.
empirical findings identified a third, intermediate category of IRAs whose discretion to prioritise is subject to a public interest test.Many Central and Eastern European CAs, including Hungary, Croatia, and the Czech Republic fall into this category.According to this approach, CAs can choose not to pursue a case only when such decision complies with the public interest.This imposes both external and internal controls.What amounts to the public interest varies from one jurisdiction to another.The Dutch CA, for example, has a 'duty to enforce', a general obligation requiring administrative authorities to take action against all potential law violations except for specific circumstances.The Dutch Council of State acknowledged IRAs' powers to set priorities, as long as this does not lead to 'never enforcing' low-priority cases, save exceptional circumstances.194Dutchcourtsinterpreted this duty as imposing an increased duty to motivate on the Dutch CA when it rejects complaints and does not pursue a case.195 cannot reject it altogether. 198 more subtle example is present in the UK, where the authority is bound to assess 'super-complaints' launched by a designated consumer body.199 but Enterprise Act 2022 c. 40, Section 11. 200 Schmidt and Scott, n 9 above. 201Cf.Eric Biber, 'The Importance of Resource Allocation in Administrative Law' (2008), 60 ADMIN.L. REV. 1.