Do EU and U.K. Antitrust “Bite”?: A Hard Look at “Soft” Enforcement and Negotiated Penalty Settlements

A. Hard Enforcement and the Deterrence-Based Approach

The origins of the deterrence-based approach can be traced to eighteenth-century contributions of classic utilitarian philosophers, such as Montesquieu, Beccaria, and Bentham. ²³ Modern variants of this approach are linked to the Law and Economics literature, originating from the seminal contribution of Nobel Prize economist Becker, later adapted to the field of antitrust by Landes. ²⁴ According to the deterrence-based approach, law obedience is not taken for granted; public and private enforcement are necessary both to prevent offenses and to apprehend offenders. Convictions alone, moreover, are not generally considered sufficient. Significant punishment is key to deterring law violations.

In EU and U.K. antitrust laws, like many other forms of economic regulation, punishment normally takes the form of an administrative fine. The enforcement system is expected to create a credible threat that the expected fines would be higher than the expected anticompetitive gains. Put differently, the fine must exceed the expected gain from the violation, multiplied by the inverse of the probability that a fine will be imposed. ²⁵ This enforcement approach, that is, formal findings of infringements and imposition of fines, is labeled in this article as hard enforcement.

Some studies attempted to capture the deterrent effect of hard enforcement in antitrust. A survey among British companies, for example, reported deterrence ratios of twenty-eight cartels deterred per detected by the British NCAs’ investigation, forty for other commercial agreements, and twelve for abuse of dominance. ²⁶ A survey among Dutch companies and their advisers reported a lower deterrence ratio for cartels, amounting to five cartels deterred per detected by the Dutch NCA. ²⁷

Hard enforcement, remarkably, is not only about deterrence. ²⁸ First, as a form of command-and-control regulation, it characterizes many social and economic regulations introduced in the late eighteenth century to facilitate large-scale industrialization (for example, sector and utility regulators, antitrust and environmental agencies). ²⁹ Under command-and-control, regulators fix the applicable standards for certain activities (command) and use the force of the law to prohibit conduct that does not meet such standards (control). In addition to deterring infringements, therefore, the force of the law is used to prohibit certain forms of conduct and demand positive action. Command-and-control regulation is believed to support the rule of law, as it relies upon the use of transparent and precisely formulated legal standards of general application. ³⁰

Second, hard enforcement is said to signal that law violations harm the social order. Fines are not just a “price” to be paid for non-compliance. They also attach a degree of condemnation. ³¹ This is reflected by the jurisprudence of the EU Courts, emphasizing that the fines imposed for antitrust infringements are not only aimed at deterrence or corrective justice, but also at punishment. ³²

Third, hard enforcement is a prerequisite for the development of a body of jurisprudence clarifying the scope of the law. A formal infringement decision resulting from a fully fledged adjudication process is not only important to correct a particular violation or increase general deterrence, but also to explicate and give force to legal rules, interpret the values upon which they are grounded, and bring reality into accord with them. ³³ Since only hard-enforcement actions are fully susceptible to judicial review, a steady flow of such decisions is essential for enhancing legal certainty and developing the law. ³⁴ To this end, hard enforcement can also prevent over-deterrence. ³⁵

Finally, hard enforcement is linked to the credibility of law enforcement authorities. In the words of Bernstein, “[o]ne of the critical tests of the effectiveness of a regulatory commission is its capacity to obtain the compliance of persons subject to the regulation and to enforce its regulations against violators. This capacity becomes, in the long run, a primary measuring rod of the ability of the agency to operate in the public interest.” ³⁶ He added that insufficient hard enforcement of laws may render them to “be more honoured in the breach than in the observance.” ³⁷

B. Soft Enforcement and the Compliance-Based Approach

Despite the benefits of hard enforcement, in practice, many regulatory authorities scarcely use their formal enforcement powers. Empirical studies reported that they more often recourse to education, advice, and persuasion. ³⁸ Ayres and Braithwaite famously advocated this approach as a normative strategy. Introducing the pyramid of regulatory enforcement, they sustained that regulators should seek to rely on education and compliance and that escalation to hard enforcement is only appropriate when such strategies have failed. ³⁹

Indeed, hard enforcement is not the only instrument in the antitrust authorities’ arsenal. In the United States, the use of compromise and negotiations in lieu of a full adjudication was introduced to civil public antitrust as early as 1906. ⁴⁰ Consent decrees ⁴¹ entrust the Federal Trade Commission (FTC) and the Department of Justice Antitrust Division (DoJ) to close a case by accepting concessions without determining whether an infringement has taken place or undertaking a market test. Becoming the hallmark of civil public enforcement, in recent years over 90 percent of the cases were settled via consent decrees. ⁴²

On the other side of the Atlantic too, the Commission and NCAs are equipped with a myriad of non-binding, non-adversarial enforcement instruments, such as informal opinions, guidance letters, press releases, ⁴³ and letters or notices warning against a specific possible infringement. They also include negotiated remedies in the form of commitment decisions (Article 9 of Regulation 1/2003, formal commitments) and decisions to close an investigation after the firms agreed to modify their potential anticompetitive behavior (voluntary commitments). ⁴⁴ The above instruments are collectively labeled as soft enforcement. ⁴⁵

Soft enforcement, as observed by the ECJ, “pursue[s] different objectives” from hard enforcement. ⁴⁶ Legal, economic, and social policy theories typically point to six aims of such strategy that are relevant in the field of antitrust:

First, soft instruments reflect a distinct theory of law enforcement. In place of focusing on deterrence and punishment, they use non-punitive measures to foster compliance. ⁴⁷ Advocates of the compliance-based approach maintain that soft enforcement may be more effective in preventing law violation because it generates less resistance from the regulatee. Its negotiated nature, free from the fear of sanctions, supports a culture of cooperation between firms and authorities, ⁴⁸ allowing authorities to act as consultants rather than policemen. ⁴⁹ Differing from the deterrence-based approach, firms are not seen as amoral calculators and law obedience is not exclusively conditional upon their self-interests. ⁵⁰ A shared commitment to compliance, moreover, may be constructive in the field of antitrust, bridging inevitable information asymmetries between authorities and markets. ⁵¹

Second, soft enforcement may strengthen legal certainty. Published and reasoned informal measures can clarify the law as applied in practice, thereby encouraging voluntary compliance. ⁵² While, as mentioned, hard enforcement is a prerequisite for the development of a body of jurisprudence clarifying the scope of the law and subject to judicial review, informal measures are of particular value to address new legal-economic questions or borderline infringements, where the full force of hard enforcement may be deemed inappropriate.

Third, soft enforcement offers unique legal remedies, unavailable under the hard-enforcement track. ⁵³ For instance, although the Commission may order any behavioral or structural remedy as part of an antitrust infringement decision, such remedy is limited to correcting the infringement and must be proportionate. ⁵⁴ Soft enforcement, by comparison, is not restricted to what is necessary to bring an infringement to an end. It facilitates a process of negotiation between firms and antitrust authorities, inviting tailored and case-specific solutions without determining if the law was infringed.

In addition to offering a wider range of remedies, this flexibility opens the door to tailor general standards or open-ended provisions of law to fit a particular set of circumstances or changing business and technological requirements, without the need to modify the law. ⁵⁵ In this sense, soft enforcement shifts antitrust from enforcement-based to a more regulatory nature ⁵⁶ ; antitrust rules are instrumentalized to meet the objectives of sector-specific regimes. ⁵⁷ This may be useful from the standpoint of authorities to “address issues that are politically or economically important but legally ambiguous.” ⁵⁸ As Abel observes, soft enforcement may neutralize economic and social conflict by responding to grievances in ways that prevent their transformation into serious challenges to the domination of state and capital. ⁵⁹

Fourth, soft enforcement is said to carry procedural economy benefits. ⁶⁰ Such instruments often neither entail a full-scale investigation nor reach definitive conclusions on matters of facts or law. They are also not subject to the rigid procedural safeguards imposed on formal infringement decisions. The saved resources can be directed to combat other infringements, ⁶¹ while providing the authority a quick tool to communicate its position and prevent further harm. ⁶² To this end, shortening the duration of proceedings may also positively affect the effectiveness of the decision. ⁶³

The limited controls imposed on soft enforcement are vindicated by their consent-based nature. The process of negotiation, the argument goes, transforms the enforcement from a public law paradigm (unilateral, top-down, hierarchical command by the authority) to a private contract-law kind. Both the general public interest and the interests of the suspected infringers are expected to be safeguarded through negotiations. ⁶⁴ This consent-based nature, in turn, brings additional procedural economy benefits because it justifies a more limited judicial review. Soft enforcement, therefore, overcomes a “war of attrition”; authorities conserve the resources required to manage long investigations over several years and to defend their positions in appeal, especially when well-resourced parties are involved. ⁶⁵

This was illustrated by the ECJ’s judgment in Alrosa. The Court accepted AG Kokott’s argument that the necessity of negotiated commitments “may be presumed as a matter of course in relation to the interests of the undertaking which has offered the commitments.” ⁶⁶ Hence, the “extent and content” of the principle of proportionality differ between commitments and hard enforcement. In the former, judicial review is limited to assessing whether the commitments were sufficient to address the competition concerns identified by the Commission. ⁶⁷ Later jurisprudence extended this obligation also to the protection of third parties’ rights. ⁶⁸

Whether soft enforcement leads to cost savings is a factual question, going beyond the scope of this study. It is sufficient to note that some have questioned this assumption. Black, for example, noted that effective soft enforcement requires significant resources. ⁶⁹ Others argued that in the field of EU antitrust, the procedural economy advantages of soft enforcement tend to be overstated. ⁷⁰ They showed, for instance, that the time elapsing between the initiation of a proceeding and the final disposition of a case is comparable under the hard enforcement and the formal commitments route. ⁷¹ Mariniello, in fact, found that adopting Article 102 TFEU commitments took longer than hard-enforcement decisions. ⁷² Monitoring the execution of remedies attached to commitments, moreover, imposes additional costs. ⁷³

Fifth, soft enforcement may serve authorities’ reputational interests. ⁷⁴ The limited availability of judicial review mediates the risk that legally or politically sensitive cases will be overturned by courts. In addition, in legal systems where hard enforcement is contingent upon a court’s approval, soft enforcement allows authorities to control the handling and the outcome of a case. ⁷⁵ This is of relevance for the antitrust authorities following a “judicial model” whereby only a court can adopt an infringement decision and impose fines on the basis of the NCAs’ recommendation (for example, in Austria, Finland, Denmark, Estonia, Sweden, and Malta). ⁷⁶ While those NCAs do not have the power to adopt infringement decisions independently, they can accept soft remedies without courts’ involvement. ⁷⁷

Finally, soft enforcement may increase deterrence by assisting harmed parties to detect infringements and bring private action. ⁷⁸ This feature was labeled in the United States as “sunshine” enforcement. According to this enforcement theory, a public authority may adequately intervene in markets by merely exposing information about anticompetitive behavior. Shining a bright light on such practice, in and of itself, is predicted to force infringing parties to comply with the law. ⁷⁹

C. Negotiated Penalty Settlements

Negotiated penalty settlements, also known as “transactional resolution mechanisms,” are an intermediate category between hard and soft enforcement. While they comprise of a formal finding of an infringement, the penalty is reduced or waived in favor of admitting to a violation (or to the facts involved) and cooperating with the authority.

Negotiated penalty settlements were introduced to U.S. antitrust in the late 1970s, upon the adoption of the DoJ’s first leniency program. ⁸⁰ European antitrust systems gradually adopted comparable instruments from the late 1990s. As elaborated below, there are three main forms of negotiated penalty settlements in the European Union and the United Kingdom—leniency, settlements for cartel cases, and settlements for non-cartel cases. ⁸¹

First, the Commission’s and NCAs’ leniency programs provide immunity from fines to the first firm disclosing its participation in a cartel and sharing relevant evidence and information. ⁸² Other firms may receive significant fine reductions upon disclosing their participation and providing evidence of added value, amounting to up to 50 percent according to the Commission’s program. ⁸³ The Commission’s and many of the national leniency programs were heavily inspired by the DoJ’s experience. Yet, they differ in several aspects, and criticized as providing firms less incentives to cooperate. ⁸⁴

Second, the Commission and some of the NCAs offer settlement programs, granting a fine reduction to all firms admitting to a cartel participation and accepting their liability. The Commission’s settlement program, which was first introduced in 2008, awards a 10-percent fine reduction in return for submission of a formal settlement, recognition of the infringement, the firm’s liability, acceptance of the range of likely fines, and waiver of further access to the file and an oral hearing. ⁸⁵ Notably, the European settlement programs are not equivalent to American consent decrees: they do not involve bargaining on the scope and details of the infringement or fine; are not subject to a court’s approval; and are separate from leniency, in terms of both their timing and the possibility for a cumulative fine reduction. ⁸⁶

Third, in 2019, the Commission introduced a new procedure for cooperation in non-cartel proceedings. Inspired by the settlement notice for cartels, it grants a similar fine reduction in vertical and abuse of dominance cases. ⁸⁷

The above forms of negotiated penalty settlements share some of the characteristics of hard enforcement, as they involve an official finding of an infringement and may be accompanied by a fine, at least for some of the involved firms. Yet, they involve many of the attributes of soft enforcement, as they depend on firms’ cooperation and are non-fully adversarial. Similarly to the promises of soft enforcement, negotiated penalty settlements are expected to provide procedural economy benefits. ⁸⁸ They can help antitrust authorities to quickly conclude their investigations and render their decisions. Cost savings are also associated with the avoidance of litigation costs. While both leniency and settlements are subject to judicial review in front of the EU and some of the national courts, firms have limited incentives to launch appeals. In addition, akin to soft enforcement, the cooperative nature of such negotiations is also said to reduce the resistance and hostility between the firms and antitrust authorities, thereby helping to promote long-term compliance. ⁸⁹

Negotiated penalty settlements also lay in between hard and soft enforcement when it comes to deterrence. On one hand, they may contribute to deterrence by increasing the number of detected and punished cartels (that is, the expected detection), thereby positively impacting general deterrence. On the other hand, the waiver or reduction of penalty may significantly reduce the specific deterrence of the apprehended cartel members (that is, the expected punishment). They run the risk that the penalty for an infringement will be lower than the anticompetitive gain. ⁹⁰

The threat to deterrence is reinforced in EU competition law due to limitations imposed on follow-on civil actions based on negotiated penalty settlements. While civil action is feasible, differing from the American system the information included in the firms’ leniency or settlement submissions cannot be used. ⁹¹ The leniency applicant in Europe, moreover, is generally exempted from joint and several liability for the entire harm caused by the cartel. ⁹²

Leniency and settlement programs, therefore, may negatively impact deterrence. ⁹³ For this reason, the ECJ held that they should be reserved to “strictly exceptional situations only,” ⁹⁴ that is, when the firms’ cooperation has been decisive in detecting and suppressing a cartel, and only so far as the program is implemented in a manner that does not undermine the effective and uniform application of the antitrust prohibition. ⁹⁵

The differences between hard and soft enforcement and negotiated penalty settlements are summarized in Table 1.

OPEN IN VIEWER

Table 1.

Differences between Hard Enforcement, Soft Enforcement, and Negotiated Penalty Settlements.

Law Enforcement Theory	Hard Enforcement	Negotiated Penalties Settlements	Soft Enforcement
Aims	Deterrence (general and specific)	Deterrence (focus on general)	Compliance
Process	Adversary	Consent-based	Consent-based
Type of enforcement action	Formal decision	Formal decision	Informal action
Finding of infringement	Yes	Yes	No
Possible use in follow-on private actions	Yes	Yes, subject to restrictions	No
Fines	Yes	Waived or reduced	No
Judicial review	Full	Limited	Highly limited

A. The Old Soft Enforcement Paradigm

The old enforcement framework of Regulation 17/62 strived to prevent antitrust violations by reducing the opportunity for committing infringements. This compliance-based approach prescribed a centralized notification regime, by which all potential anticompetitive agreements had to be notified to the Commission for authorization prior to their implementation. From the early days of EU antitrust enforcement, the Commission and ECJ interpreted the vague wording of the prohibition of Article 101(1) TFEU broadly. As a result, almost every agreement that could have restricted the parties’ commercial freedom, irrespective of the competitive situation of the market, had to be notified to the Commission.

In place of defining a limited and clear scope of prohibited conduct, this system of ex ante controls was based on the Commission’s discretion and a case-by-case assessment: when the Commission considered an agreement to be anticompetitive, it declared its incompatibility with Article 101 TFEU; when it found no ground for intervention, it would issue a negative clearance; and when an agreement fulfilled the conditions of Article 101(3) TFEU, it granted an exemption. Firms enjoyed immunity from fines if they duly notified their agreements, irrespective of the agreement’s content or potential harm to competition. ⁹⁶

When an agreement was not notified, hard enforcement was permissible. Regulation 17/62 stated that Articles 101 and 102 TFEU infringements must be enforceable by means of fines, and the ECJ emphasized that the Commission has the duty to ensure the deterrent effects of its fines. ⁹⁷ Yet, in practice, hard enforcement was a secondary strategy. The Commission confirmed that “[t]he aim of an effective competition policy is to have a direct effect on the conduct of undertakings rather than to create a multiplicity of lengthy procedures.” ⁹⁸ Instead of pursuing all serious antitrust infringements coming to its attention “with the utmost vigour,” the Commission declared it will choose the “appropriate action” in each case. ⁹⁹

The old enforcement regime was based on the premise that a system of controls—together with stating and clarifying the law via the Commission’s and the EU Courts’ decisional practice—would facilitate voluntary compliance. ¹⁰⁰ Even when an anticompetitive agreement was not duly notified, the Commission often refrained from imposing a fine, especially when the antitrust rules on the matter were not fully clear. ¹⁰¹

Although Regulation 17/62 did not explicitly refer to soft enforcement, such instruments were quickly created as a pragmatic solution to combat the administrative workload created by the notification system. Shortly after the Regulation came into force, the Commission was flooded with tens of thousands of notifications, ¹⁰² which continued to haunt its docket for many years later. ¹⁰³ From the early 1970s, the Commission introduced the tool of comfort letters, administrative letters signed by a Commission official stating that no action would be taken against a particular agreement. ¹⁰⁴ Such letters were used to handle cases “which at first sight raise no problems with respect to the rules of competition and do not require a formal decision.” ¹⁰⁵ Following their issuance, the proceedings were terminated without adopting a formal decision. ¹⁰⁶ Comfort letters were not binding on national courts, NCAs, or third parties, yet they constituted a matter of fact that national courts could have taken into account. ¹⁰⁷

In parallel to comfort letters, agreements raising antitrust concerns were mostly handled by accepting voluntary commitments. ¹⁰⁸ In such cases, the proceedings were terminated after a statement of objection was issued, without finding of an infringement or imposing fines. ¹⁰⁹ This enforcement strategy was justified as a means to bring a specific infringement to an end. The Commission stated that its “main aim is to ensure efficiently functioning markets. It is therefore prepared to accept undertakings from dominant companies to achieve this end.” ¹¹⁰

The procedure for accepting voluntary commitments was fully informal, meaning the Commission had no dedicated mechanism to ensure compliance or to protect the rights of third parties. ¹¹¹ The Commission had also not pronounced clear criteria for accepting such commitments or the appropriate type of cases. Strikingly, the acceptance of voluntary commitments was not made dependent on the seriousness of the infringement. They were used not only to deal with notified agreements involving minor or borderline infringements, but also to resolve serious infringements—such as price fixing, quota agreements, joint selling agencies, and setting the conditions of sale—as well as agreements that were not duly notified. ¹¹² While some investigations were terminated after an agreement was modified, terminated, or expired, ¹¹³ similar anticompetitive agreements were subject to hard enforcement and sanctions even if the parties agreed to modify or terminate them. ¹¹⁴

A similar compliance-based mechanism was embedded within the Commission’s non-opposition procedure. In the 1980s and 1990s, some of the European Union’s block exemption regulations specified that agreements containing terms that were neither explicitly exempted nor explicitly prohibited were exempted unless the Commission opposed them within a period of six months. ¹¹⁵ In practice, when an agreement raised concerns, the Commission mostly invited the firms to renegotiate its terms, and terminated the proceedings by soft means of issuing a comfort letter or accepting voluntary commitments. ¹¹⁶

The strategy of favoring soft over hard enforcement under the old enforcement regime is confirmed by Figure 1. ¹¹⁷ The top graph represents the number of soft enforcement actions, in the form of voluntary commitments (dotted area), comfort letters (white area), negotiations within the opposition procedure (black area), and Article 101(3) TFEU exemptions subject to commitments or conditions (gray area). The bottom graph represents the number of infringement decisions (hard enforcement and negotiated penalty settlements ¹¹⁸ ), either with and without fines (black and stripped bars, respectively). The graph covers the period between 1976 and 2004, namely from the first year when the full data on soft measures have been published and until the entry into force of Regulation 1/2003 modernizing the enforcement.

OPEN IN VIEWER

Figure 1.

Soft versus hard enforcement under the old enforcement regime, Commission (1976–2004). ¹¹⁹

The figure illustrates that the vast majority of enforcement under the old notification regime took the form of soft enforcement (upper graph). Even in the rare occasions where the Commission opted for a formal infringement decision (bottom graph), it was often not accompanied by fines, thereby considerably limiting its deterrent effect.

The figure also shows that the strategy of favoring soft over hard enforcement guided the Commission well into the 1990s. In 1997, for example, the Commission had not taken a single hard action against an Article 101 TFEU violation, noting that it was able to accept voluntary commitments instead. ¹²⁰

The Commission regularly justified its compliance-based approach with reference to procedural economy motives. Soft enforcement allowed it “to avoid the preparatory work and procedural delays required for the preparation of decisions once the Commission’s views were clearly established, or when a decision would not have added in any way to existing administrative case law.” ¹²¹ Along similar lines, the Commission argued that informal measures enabled a rapid response, as compared to the approximately twenty-four months required for issuing a formal decision (yet, it argued it was impossible to calculate the average period of time for such procedures). ¹²² The Commission also confirmed that procedural economy motives vindicated soft enforcement even in face of serious restrictions of competition. ¹²³

Notably, the Commission maintained that soft enforcement was particularly warranted in the areas where the law was relatively clear. In such instances, according to the Commission, “it increasingly suffices to draw a firm’s attention to actual or possible infringements of the rules of competition for the offending practice to be voluntarily terminated.” ¹²⁴ The Commission maintained that “[i]t is much to be regretted” that it must take hard-enforcement actions against practices that have already been clarified via case law. ¹²⁵ These statements stand in conflict with the theory of law enforcement presented in Part II, by which a compliance-based approach is justified in the face of new or borderline legal-economic questions to strengthen legal certainty, rather than for dealing with intentional infringements when the law is clear.

Commentators suggested that the reliance on soft enforcement could be explained by the relatively low legitimacy of EU antitrust in its early days. Gerber, proposed that emphasizing persuasion and compromise was necessary given the Commission’s weak enforcement tools and cautions political backing to EU antitrust. ¹²⁶ Temple Lang highlighted that the Commission’s enforcement priorities and instruments valued establishing the legal principles underlining EU antitrust law over correcting the operation of markets. ¹²⁷ This suggests that politics and lack of legitimacy, rather than enforcement theory, might have contributed to the prominence of soft enforcement.

B. Questioning the Commission’s Approach

Serving as the cornerstone of EU antitrust for many years, from the 1990s soft enforcement was increasingly criticized as ineffective.

First, it became apparent that the Commission’s selection of enforcement instruments was not based on a definite theory. As mentioned, there were no clear criteria for determining which cases were appropriate for soft instruments and which vindicated hard enforcement, resulting in similar types of infringements being subject to different treatments for no apparent reason. The European Parliament, for example, advocated for greater transparency, asking the Commission to provide more information “on the principles and criteria” guiding its strategy so as to “provide more guidance for affected undertakings.” ¹²⁸

Second, as the choice of soft instruments was neither based on a clear strategy nor bound to a clear procedural framework, such instruments have frequently failed to realize their potential benefits. Comfort letters and voluntary commitments mostly went unpublished, had no binding effect, and could not be challenged in front of the EU Courts. This stood in the way of enhancing compliance by increasing the awareness to antitrust rules, by shedding light on a specific infringement (the “sunshine” enforcement feature), or by clarifying the law as applied to new or borderline legal questions. ¹²⁹

Finally, the Commission’s enforcement strategy was condemned as costly and ineffective in preventing violations. ¹³⁰ In the over forty years in which Regulation 17/62 was in force, only nine notified agreements were prohibited without a complaint having been lodged against them. ¹³¹ In fact, by the late 1990s, only 0.5 percent of the notified agreements resulted in an infringement decision. ¹³² The notification system imposed immense costs on both firms and the Commission, ¹³³ and led to directing most of the enforcement resources to agreements raising no or very limited competition concerns, while serious infringements were left undetected.

C. The New Hard-Enforcement Aim and Rhetoric

Around the turn of the millennium, the Commission advocated a dramatic reform to antitrust enforcement. In its 1999 Modernization White Paper, it called for a shift to hard enforcement, as to allow it “to refocus its activities on the most serious infringements.” ¹³⁴ Instead of directing its efforts to provide comfort to specific agreements, the new enforcement regime of Regulation 1/2003 champions a deterrence-based approach, which is contingent on hard enforcement. The Regulation and the Commission’s policy papers accompanying it emphasize that while the new enforcement system is without prejudice to the power of using soft enforcement in limited and appropriate circumstances (that is, informal guidance ¹³⁵ and formal commitments ¹³⁶ ), hard enforcement directed at combating serious infringements should be the primary strategy. Deviating from the Commission’s prior practice, the Regulation declares that soft instruments such as commitments are not appropriate to address by-object restrictions, secret cartels, or cases where the Commission intends to impose a fine. ¹³⁷

The new enforcement regime was expected not only to refocus the Commission’s enforcement priorities on serious restrictions, but also to significantly boost the number of hard-enforcement actions. The Commission explained in its Modernization White Paper that as it “would be concentrating its attention on the most serious restrictions, the number of individual prohibition decisions can be expected to increase substantially.” ¹³⁸ This aim was repeated in the evaluation of the reform in 2009 and 2014. ¹³⁹

The centrality of the deterrence-based approach is also reflected by the rules for calculating fines. While the fining guidelines of 1998 already declared that the level of fines should ensure they have “a sufficiently deterrent effect,” ¹⁴⁰ the 2006 version further adds that the Commission will increase the fine when necessary to ensure that it exceeds the improperly made gains. Fines are expected to have

a sufficiently deterrent effect, not only in order to sanction the undertakings concerned (specific deterrence) but also in order to deter other undertakings from engaging in, or continuing, behaviour that is contrary to Articles [101] and [102] of the [EU] Treaty (general deterrence). ¹⁴¹

The new hard-enforcement rhetoric was embraced by all EU institutions. ¹⁴² The EU Courts emphasized that the antitrust prohibitions are ineffective when their breach is not accompanied by the imposition of fines. ¹⁴³ Pointing to the “intrinsic link” between prohibitions and fines, the ECJ observed that the imposition of fines by the Commission and NCAs is a precondition for the effectiveness of the rules. ¹⁴⁴ In Mamoli and Schenker, the ECJ held that fines are necessary to ensure that antitrust is applied effectively in the general interest. In striking contrast to past practice, the Court explained that while authorities are not precluded from not imposing a fine, such an approach must be reserved for “strictly exceptional situations only.” ¹⁴⁵

The shift to a deterrence-based approach was also reflected by EU secondary law, adopted by the European Parliament and the Council. The 2014 Damages Directive is premised on the notion that private enforcement is necessary to ensure the full deterrent effect of antitrust law. ¹⁴⁶ The 2019 ECN+ Directive expresses a similar approach. ¹⁴⁷ This shift was apparent during the negotiations preceding the adoption of the Directives, where the Commission stressed that hard enforcement “is a central enforcement tool,” ¹⁴⁸ a “key aspect” of the EU antitrust regime, ¹⁴⁹ and the “most efficient weapon in the Commission’s armoury to fight cartels.” ¹⁵⁰ At other occasions the European Parliament noted that although the Commission may issue informal opinions where “clarification is in the general interest,” this should be confined to exceptional cases. ¹⁵¹

The hard-enforcement rhetoric was not limited to enforcement by the Commission. The decentralization of the enforcement, coupled with the imposition of effective fines by a multiplicity of authorities, is deemed essential for strong, effective, and deterrent enforcement. ¹⁵² Therefore, the ECN+ Directive requires all Member States to at least match the level of fines imposed in infringement proceedings taking place in front of NCAs to those that can be imposed by the Commission. ¹⁵³ Moreover, in a bid to ensure the effectiveness of fines, it notes that firms “cannot compete on the merits if there are safe havens for anticompetitive practices, for example, because (. . .) undertakings are able to escape liability for fines.” ¹⁵⁴

The shift to a deterrence-based approach at the national level is further reflected by the trend of criminalization of serious antitrust violations by several Member States and/or the imposition of individual administrative sanctions. While criminal enforcement and individual administrative sanctions are not available as a matter of EU law, they may be imposed for the breach of Article 101 TFEU in national courts proceedings. ¹⁵⁵

Alongside this clear shift in aims and rhetoric, and despite the importance attributed to hard enforcement in the past twenty years, EU law does not direct or limit the selection of enforcement targets or instruments. The Commission and NCAs are expected to exercise self-restraint in the selection and application of soft enforcement instruments and negotiated penalty settlements, ¹⁵⁶ and their strategies and motivations remain mostly hidden. The next parts show that in practice, the Commission and NCAs have adopted diverse enforcement strategies, all in which hard enforcement remains the exception rather than the norm.

A. Infringement Decisions: Hard Enforcement and Negotiated Penalty Settlements

The hard-enforcement rhetoric—mirrored by the EU Regulations, Directives, jurisprudence, and the Commission’s policy papers—was not matched by a greater number of hard-enforcement actions in the Commission’s practice. The empirical findings indicate that not only was there no increase in the number of infringement decisions (as was foreseen by the Modernization White Paper), but that the Commission continues to follow a compliance-based approach and introduced new instruments of soft enforcement and negotiated penalty settlements.

Figure 2 outlines the total number of enforcement actions taken by the Commission since it rendered its first decision in 1964 and up to 2021. It specifies the number of hard-enforcement decisions (black bars), negotiated penalty settlements (gray bars), and published decisions reporting the use of soft enforcement (white bars). ¹⁵⁷

OPEN IN VIEWER

Figure 2.

Hard and soft enforcement per year, Commission (1964–2021).

The figure illustrates that contrary to the hard-enforcement rhetoric and unlike the initial expectation from Regulation 1/2003, ¹⁵⁸ freeing the Commission from the administrative burden of notifications did not result in a greater number of fully adversarial decisions. These findings are notable given the significant enlargement of the European Union since 2004, growing from fifteen Member States to twenty-eight (twenty-seven, following Brexit in 2020) and the increase in the staff and budget of the Directorate General for Competition. Although the geographic scope of EU antitrust law has expanded significantly and the Commission no longer has the burden of responding to notifications, there was no increase in the average number of infringement decisions per year (hard enforcement and negotiated penalty settlements).

The findings summarized by the figure demonstrate that since the entry into force of Regulation 1/2003 on May 1, 2004, 44 percent of the Commission’s proceedings involved negotiated penalty settlements: 15 percent a combination of leniency and settlements, 24 percent leniency, 1 percent settlements, and 4 percent cooperation procedure in non-cartel cases. Hard enforcement, by comparison, amounted to only 8 percent of the cases and even then, fines were imposed only in thirteen out of the eighteen cases. Put differently, 85 percent (!) of the Commission’s infringement decisions did not consist of a fully adversarial procedure. As forms of negotiated penalty settlements, they involved a significant degree of cooperation with the firms. This stands in striking contrast to the ECJ’s order to reserve such proceedings to “strictly exceptional situations only.” ¹⁵⁹

Moreover, the figure reveals that there was no increase in the average number of infringement decisions per year (gray and black bars) following modernization. By comparison, the number of infringement decisions grew since 2002, due to the increasing use of negotiated penalty settlements (gray bars). This appears to be associated with the introduction of the Commission’s new leniency program that year rather than the reform of Regulation 1/2003 in 2004 ¹⁶⁰ ; while between 2002 and 2004 the Commission issued an average of eight infringement decisions per year, from 2005 to 2021 this average dropped to 6.5.

The introduction of the Commission’s settlement program in 2008 and the cooperation procedure in non-cartel cases in 2019 did not appear to increase the number of infringement decisions. This challenges one of the main vindications of such programs, namely that the reduction of fines and specific deterrence is justified by the increased detection of infringements. ¹⁶¹

The limited number of infringement decisions and the prevalent use of negotiated penalty settlements do not correspond to the deterrence-based rhetoric following modernization and to the expected increase in the number of infringement decisions. Deterrence might be sub-optimal, especially when the above is taken together with the following observations ¹⁶² : First, most of the negotiated penalty settlement instruments are limited by definition to secret cartels. ¹⁶³ This means that the enforcement reform of Regulation 1/2003 did not have a significant positive effect on deterrence beyond cartels. The empirical findings imply that there is almost a negligible chance that firms will be sanctioned by the Commission for other types of antitrust infringements. In the limited number of cases when such practices were subject to a Commission’s investigation, they were more likely to be resolved by accepting formal commitments than by imposition of fines. ¹⁶⁴ Those findings are also in line with the limited number of the Commission’s own-initiative investigations, significantly declining since 2015. ¹⁶⁵

Second, members of secret cartels were relatively protected from sanctions where no leniency application has been filed. This corresponds to the finding of a previous study, showing that almost 70 percent of the firms investigated for cartel participation by the Commission have escaped a fine if they have not applied for immunity or leniency. ¹⁶⁶ In addition to hindering deterrence, this reduces the attractiveness of the Commission’s leniency program, and could explain the significant reduction of leniency applications submitted to the Commission over the past years. ¹⁶⁷ The limited risk of sanctions outside negotiated penalty settlements may also invite strategic use of leniency, by which firms threat to submit leniency applications to deter their competitors from deviating from a cartel. ¹⁶⁸ Along these lines, Stephan argues the Commission’s leniency program is not detecting active cartels, but rather uncovering failed inactive cartels. ¹⁶⁹

Third, even where leniency applications have been submitted, the alleged cartel members had a high chance of escaping sanctions. As of 2020, only approximately 15 percent of the leniency applications led to a formal investigation by the Commission, and only 60 percent of these investigations ended with a prohibition decision and the imposition of fines. ¹⁷⁰

Fourth, commentators warned that even when penalties have been imposed, their level did not produce a sufficient deterrent effect. ¹⁷¹ A 2020 Court of Audits report observed that nearly two-thirds of the fines imposed by the Commission in cartel cases were lower than 0.99 percent of the parties’ global annual turnover. ¹⁷² Marvão and Spagnolo added that this trend has been aggregated over time, whereby over 70 percent of the fines imposed on convicted cartel members were lower than 1 percent of their turnover. ¹⁷³ Cautioning against a “leniency inflation,” they point to an increase in both the number of cartel members receiving full immunity and the level of fine reductions. ¹⁷⁴ This is particularly worrying when combined with the findings of Smuda, reporting that in 67 percent of cartel cases, the anticompetitive gains exceeded the maximum 10 percent worldwide turnover threshold prescribed by Regulation 1/2003 and the Commission’s Guidelines. ¹⁷⁵

Finally, the Commission does not engage in a comprehensive impact assessment of its enforcement strategies. It does not systematically reflect on the formation of its portfolio of cases, the selection of instruments, and the assessment of its performance. It is unclear how the Commission balances monitoring markets on its own initiative with responding to complaints, leniency, or settlements applications. ¹⁷⁶ When it comes to the imposition of penalties, while the Commission often emphasizes the high level of fine it imposed, it does not evaluate their deterrent effect. Undoubtedly, the Commission’s fines for antitrust infringements are among the highest in the world; yet, given the size and economic significance of the common market, this cannot alone inform a conclusion about their effectiveness. ¹⁷⁷

Taken together, the findings on the Commission’s practice question the realization of the basic aims of the European Union’s deterrence-based approach and the shift to hard-enforcement rhetoric, that is, increasing the number of infringement decisions and ensuring that the expected fine for an infringement is higher than the expected anticompetitive gains.

B. Soft Enforcement

In parallel to hard enforcement and negotiated penalty settlements, following the modernization, the Commission continued to resolve many of its investigations by means of soft enforcement. The new enforcement regime, in fact, has given rise to new soft enforcement instruments. This is demonstrated by Figure 3, specifying the makeup of the Commission’s Article 101 TFEU enforcement portfolio since the entry into force of Regulation 1/2003.

OPEN IN VIEWER

Figure 3.

Commission, ratio of hard and soft enforcement, May 2004–2021.

The figure shows that 11 percent of the Commission’s proceedings were closed by means of accepting formal commitments, a new enforcement instrument introduced by Regulation 1/2003. ¹⁷⁸ Formal commitments entrust the Commission (and NCAs) to terminate an investigation after making the behavioral or structural remedies proposed by a firm binding, and impose fines for failure to comply. ¹⁷⁹ In comparison to the voluntary commitments that were described above, the process for accepting formal commitments is more streamlined, transparent, and safeguards the rights of third parties. ¹⁸⁰ This instrument was inspired by the U.S. consent decrees, yet EU commitments do not involve the imposition of fines and are not being reviewed by courts prior to their adoption to determine whether the decree is in the public interest. ¹⁸¹

In addition, the Commission has continued to terminate investigations into potential infringements upon the acceptance of voluntary commitments. While the figure points to a limited use of this instrument (approximately 6% of the proceedings), their number is likely to be underreported because not all such cases are published.

Both formal and voluntary commitments were mostly accepted to settle potential infringements in highly regulated sectors such as utilities (telecom, media, energy, and transport), ¹⁸² financial services, ¹⁸³ and technology and digital markets. ¹⁸⁴ The Commission’s formal commitment decisions were relatively detailed, albeit avoiding a finding on the possible infringement. Voluntary commitments, however, were announced in short press releases (typically one to two pages), which do not offer a detailed factual or legal analysis.

The Commission made scarce or no use of the other soft instruments in its arsenal. It did not issue informal guidance or comfort letters, with the exception of a 2020 comfort letter concerning the supply of medicine during the Coronavirus pandemic. ¹⁸⁵ In late 2022, the Commission adopted a new notice on informal guidance, declaring that such an instrument may be used “where it considers it appropriate and subject to its enforcement priorities,” in particular, to clarify novel or unresolved questions and would provide added value to legal certainty and predictability. ¹⁸⁶ The new notice introduces a more formal process, including consultation with NCAs and an assurance that the Commission will not impose a fine if the firms relied in good faith on its guidance. This instrument, therefore, may be used more frequently in the future.

When viewing the Commission’s practice per year, Figure 2 may appear to suggest that there was a decrease in the use of soft enforcement since 2008. ¹⁸⁷ Yet, rather than concluding that the practice of accepting voluntary commitments or issuing informal guidance was abandoned, this figure may serve as an indication that the Commission stopped reporting such actions. There is evidence that the Commission still make use of such instruments. In its 2020 annual report, for example, it stated that it “engaged with companies (. . .) providing informal guidance for the types of cooperation that are likely to be unproblematic, and identified the necessary safeguards for such cooperation.” ¹⁸⁸ The little reporting of such discussions may stand in the way of realizing the benefits of soft enforcement, as will be elaborated below. ¹⁸⁹

The significant proportion of negotiated penalty settlements and soft enforcement implies that despite the hard-enforcement aims and rhetoric of modernization, the Commission continues to follow a compliance-based approach. As a result, a gap emerges between the stated aims of Regulation 1/2003 and the Commission’s enforcement practice. This gap is particularly alarming since the reliance on soft enforcement often goes underreported. The statistics published in the Commission’s reports on the functioning of Regulation 1/2003, for instance, include little information on soft enforcement. ¹⁹⁰ On the contrary, when surveying the allocation of the enforcement effort, the Commission concluded that the authorities have “prioritised the most serious and harmful anticompetitive practices, in particular, cartels, which account for a substantial proportion of their enforcement record.” ¹⁹¹ The Commission, likewise, does not generally distinguish between fully adversarial hard enforcement and negotiated penalty settlements in its annual reports and studies, and ignores the efforts allocated to soft enforcement.

Before turning to discuss the implications of this gap in Part VI, the next part demonstrates that the NCAs also made limited use of hard enforcement. In fact, contrary to the hard-enforcement narrative of Regulation 1/2003, some NCAs explicitly declared that soft enforcement should play a fundamental part in their enforcement strategy.

A. The Netherlands (ACM)

The ACM (and its predecessor until 2013, the NMa) has made frequent use of soft enforcement and compliance-based approach. Figure 4 indicated that it relied on soft enforcement to the greatest extent in comparison to the authorities presented in this study. Figure 5 further illustrates that this trend was reinforced since 2015. ¹⁹⁴ In particular, no findings of infringements concerning anticompetitive agreements were adopted in 2016, 2018, and 2019.

OPEN IN VIEWER

Figure 5.

Hard versus soft enforcement, Dutch NCA, per year.

Only 40 percent of the ACM’s proceedings ended with a finding of an infringement, from which about half involved negotiated penalty settlements and thus were not fully adversarial. In addition, four out of the thirty-six hard-enforcement actions did not include a fine. All those proceedings were adopted between 2004 and 2005, and the NCA justified the waiver of fines by the recent liberalization of the markets, lack of previous antitrust enforcement in those areas, ¹⁹⁵ or due to a governmental involvement in the infringement. ¹⁹⁶

Several soft enforcement instruments were used. Some instruments resulted in a detailed analysis that can enhance legal certainty and shed light to potential infringements. Others were limited to a laconic press release ¹⁹⁷ :

The ACM issued twelve formal commitment decisions and closed the proceedings in twenty-two cases after the parties agreed to modify their conduct (voluntary commitments). It explained that commitments are “highly recommended” as a device for quick and transparent termination of proceedings, while preventing future harm. ¹⁹⁸ Formal commitments were used to tackle controversial and politically sensitive matters related to professional associations, health care, ¹⁹⁹ and other highly regulated sectors. ²⁰⁰ Those decisions included a detailed analysis, ²⁰¹ clarifying the NCA’s position. Many of the voluntary commitments have also related to professional associations, health care, ²⁰² and other regulated sectors. ²⁰³ They were also adopted to address Coronavirus challenges. ²⁰⁴ Unlike the Dutch formal commitments, the reporting was short and concise, in a manner that is unlikely to realize the full benefits of soft enforcement.

The ACM issued thirty-two ex ante informal opinions, examining the compatibility of a specific agreement with the antitrust rules. Five of the informal opinions involved voluntary commitments offered by the parties. ²⁰⁵ This instrument was regularly used to examine agreements that enjoyed a degree of public or governmental support despite their (often serious) anticompetitive effect, ²⁰⁶ in particularly related to sustainability benefits. ²⁰⁷ Other informal opinions related to regulated markets ²⁰⁸ and Coronavirus challenges. ²⁰⁹ Informal opinions were mostly detailed, discussing factual and legal aspects of the case, yet some provided limited information. ²¹⁰

The ACM also reported warnings it has issued to parties, in a form of press releases. Those announcements were concise, limited to few, general details on the potential infringement and the involved parties. ²¹¹

The extensive reliance on soft enforcement was not just a matter of practice in the Netherlands. It is tied to the national approach to social and economic governance. Dubbed as the polder-model, Dutch decision-making process is heavily deliberative and consensus-based, inviting stakeholders to present their views and come to a joint solution. ²¹² The ACM justifies the reliance on soft enforcement with reference to its problem-solving supervisory style. It declares that it tackles market problems from a broad perspective. Instead of merely investigating alleged infringements, it seeks the root cause favoring norm-transmitting discussions and commitment decisions. ²¹³ Aiming at optimal compliance rather than maximum enforcement, ²¹⁴ trust forms the backbone of this supervision style. ²¹⁵

In its 2005 annual report, which presented a general account of the enforcement strategy, the NCA noted that it intends to use “as far as possible” alternatives to hard enforcement, involving consultation with market players’ representatives. ²¹⁶ It would refrain from imposing sanctions where: the infringement was terminated immediately upon notice or ended; is unlikely to be repeated; does not involve a hardcore infringement; and where alternative enforcement is likely to benefit consumers, generate sufficient prevention effects, and does not raise third-party objections. ²¹⁷ The soft enforcement philosophy was also advocated by the head of the authority during the Coronavirus crisis, noting that “we position ourselves these days more as a reasonable ‘market superintendent,’ rather than as cartel police or consumer union.” “The open and constructive dialogue with companies, consumers, and other authorities,” he added, “could be an efficient approach for other issues in the future as well.” ²¹⁸

Many aspects of the Dutch practice may contribute to realizing the benefits of soft enforcement. The commitment decisions and informal opinions often detailed the alleged infringement, antitrust concerns, and the accepted remedies to alleviate them. Yet, the extensive reliance on soft enforcement raises concerns as to the prevailing level of deterrence in the Netherlands, especially in the operation of professional associations, health care, and other highly regulated markets, in which soft instruments were almost exclusively used for many years.

This risk was highlighted by an external consultancy report examining the effectiveness of the NCA’s enforcement, commissioned in 2015 upon the request of the Dutch Minister of Economic Affairs. The report recommends granting further attention to the deterrence effects of the NCA’s efforts, analyzing the relationship between the enforcement instruments used to the long-term deterrent effects of its enforcement efforts. ²¹⁹ Nevertheless, as Figure 5 illustrated, soft enforcement only became more prominent in the following years.

B. Germany (Bundeskartellamt)

Contrary to the Netherlands and similarly to the Commission, the German NCA manifests a hard-enforcement rhetoric. It states that the aim of cartel prosecution is not only to uncover, terminate, and sanction cartels, but also to achieve as much deterrence as possible. A high level of prosecution and fines is essential, ²²⁰ the “key focus of the Bundeskartellamt’s work.” ²²¹ From the early 2020s, especially as a response to the Coronavirus and energy crises and the uncertainty surrounding sustainability-related cooperation agreements, soft rhetoric is becoming more common. The Bundeskartellamt affirmed that “we also consider ourselves a partner for businesses rather than exclusively a sanctioning or prohibiting authority.” ²²²

Figures 4 and 6 show that consistent with the hard-enforcement rhetoric, infringement decisions were prominent in Germany. Amounting to 58 percent of the proceedings, this ratio is the highest in comparison to the authorities examined in this study. Nevertheless, some promises of hard enforcement may not fully materialize in Germany, as out of the thirty-eight hard-enforcement actions, only twenty (53%) resulted in an imposition of a fine. In addition, 69 percent of the infringement decisions involved negotiated penalty settlements, and thus were not fully adversarial. This trend was more pronounced in recent years. Since 2017, only two fully adversarial hard-enforcement decisions were adopted, both were not accompanied by the imposition of fines. ²²³

OPEN IN VIEWER

Figure 6.

Hard versus soft enforcement, German NCA, per year.

Alongside the hard-enforcement rhetoric, soft enforcement was prominent:

The Bundeskartellamt adopted thirty-four formal commitments and two formal decisions declaring that it has no ground for action after the parties offered voluntary commitments. Similarly to the other authorities examined, such commitments related to practices in regulated and liberalized markets (especially energy ²²⁴ and sport broadcasting ²²⁵ ), which often raise complex questions about the relationship between the protection of competition and other public interests. ²²⁶ Such decisions were detailed and reasoned, allowing to clarify those challenges and increase legal certainty.

The Bundeskartellamt issued press releases reporting that its investigation was brought to an end after securing voluntary commitments in twenty cases. ²²⁷ This instrument was commonly used to resolve investigations into digital and technology markets, raising new legal and economic questions in fast-moving markets. ²²⁸ Here, too, there was no single form of publication. Most press releases consisted of a concise summary of the case, while others provided more details (two to four pages typically) focusing on the principles guiding the legal analysis. ²²⁹

The NCA provided guidance (informal opinions). ²³⁰ Most of this guidance went unpublished, making it difficult to assess how frequently this instrument was used. The sixteen published summaries (from which four declared that voluntary commitments were offered by the parties) ²³¹ commonly examine agreements related to broadcasting, ²³² digital and technology markets, ²³³ professional associations, ²³⁴ Coronavirus, ²³⁵ and sustainability arrangements. ²³⁶ The lack of or limited publication hinders the potential effect of this guidance. While the reported summaries pertained to cases in which the law appeared to be unclear or when the protection of competition should have been balanced against the protection of other public interests, many of the summaries did not sufficiently detail matters of facts and law to facilitate voluntary law compliance. This is illustrated by the 2017 press release on the Tierwohl animal welfare initiative. After providing general information about the initiative, the NCA merely declared that it will “tolerate the agreement only for a transitional period until 2020.” It did not specify the legal basis for “tolerating” the agreements, whether the initiative was likely to violate the antitrust rules, or fulfill the conditions for an exception. Consequently, the enforcement resources dedicated to this case could not realize the benefits of soft enforcement. ²³⁷

C. The United Kingdom (CMA and Sector Regulators)

U.K. antitrust has committed to a deterrence-based enforcement system. This was evident in the process of adopting the Enterprise Act of 2002, calling for a “strong deterrent effect” to root out anticompetitive behavior. ²³⁸ In addition to administrative hard enforcement, the Act strives to secure deterrence by criminalizing the participation in certain forms of cartels. ²³⁹ The centrality of a deterrence-based approach was repeated in more recent policy papers ²⁴⁰ and jurisprudence. ²⁴¹ In 2020, the then Chief Executive of the CMA, Andrea Coscelli, clarified that hard enforcement is the CMA’s “bread and butter,” and that the CMA “can only secure our legitimacy if we achieve robust enforcement outcomes on what the public believes to be the glaring injustices of the day.” ²⁴²

In parallel to this hard-enforcement emphasis, the CMA (and its predecessor, the Office of Fair Trading, OFT) explicitly invokes soft enforcement as an important strategic tool. ²⁴³ This was particularly apparent in regulated markets, where in the United Kingdom, unlike many other European legal systems, antitrust is being enforced also by sector regulators. The CMA noted that while a competitive environment “can be achieved in part through effective and efficient enforcement,” soft enforcement tools “can, in appropriate circumstances, also be effective.” ²⁴⁴

Despite the acknowledgment of both hard and soft enforcement, Figures 4 and 7 demonstrate that infringement decisions played a limited role in practice, especially beyond negotiated penalty settlements.

OPEN IN VIEWER

Figure 7.

Hard versus soft enforcement, U.K. antitrust authorities, per year (excluding warning and advisory letters).

Figure 7 records the strikingly low number of infringement decisions in the United Kingdom, especially when compared to other jurisdictions having a similar market size and antitrust authorities with a similar (or lower) budget and staff. ²⁴⁵ In contrast with the fifty-five infringement decisions adopted by the CMA and sector regulators between May 2004 and 2021 (hard enforcement and negotiated penalty settlements), the German and Dutch NCAs have adopted 124 and seventy-five infringement decisions, respectively.

Negotiated penalty settlements amounted to 82 percent of the infringement decisions (and 49% of the total enforcement efforts recorded in the figure), meaning that only ten proceedings were fully adversarial (from which only eight resulted in an imposition of a fine). Up to 2019, all of the enforcement efforts of the sector regulators were resolved by soft enforcement instruments. ²⁴⁶

The limited level of infringement decisions in the United Kingdom, especially in important regulated sectors, was criticized by the National Audit Office in its 2010 report. ²⁴⁷ Subsequently, the government announced far-reaching changes to the enforcement regime, including the Enterprise and Regulatory Reform Act 2013. This reform, however, did not significantly increase hard enforcement. This was observed by the 2016 National Audit Office report, calling to “build up a steady flow of successful high-profile cases, decisions and fines that withstand appeal.” Findings of infringement and the imposition of financial penalties were seen as critical for deterrence and for clarifying the law. ²⁴⁸ In light of the above, the CMA declared that it would move to “step-up” its enforcement. ²⁴⁹ Figure 7 confirms the increase in infringement decisions since, albeit mostly in the form of negotiated penalty settlements.

The relatively limited number of infringement decisions has a few possible explanations. It was suggested that sector regulators find it easier and more effective (at least in the short term) to use their regulatory rather than antitrust powers. ²⁵⁰ The CMA added that, unlike other European countries, the British approach to economic regulation supports early structural reforms and regulatory frameworks. ²⁵¹ The limited number of infringement decisions, especially in fully adversarial settings, was also explained by reference to the national rigors appeal system. The National Audit Office stated that stakeholders noted that “the UK was the best jurisdiction in the world to defend a competition case.” ²⁵² The length of the enforcement process and uncertainty of its outcome in appeal may have reduced the appetite of the antitrust authorities to use their hard-enforcement powers. ²⁵³ More recently, the CMA linked the limited hard enforcement to Brexit, noting that working toward the EU exit from the European Union in 2020 was a complex and resource-intensive undertaking, resulting in diverting focus and resources away from enforcement. ²⁵⁴

In parallel to hard enforcement and negotiated penalty settlements, various types of soft enforcement instruments were used in the United Kingdom. Like the other authorities examined, not all of those instruments were equally likely to realize the benefits of soft enforcement:

The CMA and sector regulators adopted seven formal commitments and terminated additional nine proceedings after the parties agreed to modify their conduct (voluntary commitments). Like the German and Dutch NCAs, the commitments focused on practices in digital markets, ²⁵⁵ regulated sectors, and trade associations. ²⁵⁶ Differing from the other NCAs, however, all formal and some of the voluntary commitments included a full and detailed analysis, clarifying the underlying factual and legal circumstances. This is particularly noteworthy as the CMA maintained that despite its thorough analysis, securing commitments have produced important procedural economy benefits. ²⁵⁷

The NCA issued short-form opinions. This instrument is similar to the Dutch informal opinions and to the Commission’s informal guidance in that it offers an ex ante and detailed analysis of a specific agreement raising novel or unresolved questions. ²⁵⁸ As per its declared policy, the NCA had made scarce use of this instrument, issuing only two short-form opinions (both by the OFT) and such instrument has seem to be abandoned in recent years. ²⁵⁹

Perhaps most importantly, the CMA has sent warning and advisory letters notifying businesses that they might be breaking antitrust law. ²⁶⁰ Those letters do not involve findings of law or fact. They only spell out the CMA’s concerns and recommend firms to self-assess their practice. ²⁶¹ Because there is no complete and accurate data on the use of warning and advisory letters, they are not recorded in the figures above. However, since 2014, the CMA published some aggregated information, suggesting that between 2014 and 2021, 390 warning letters and 167 advisory letters were issued as to possible infringements of EU and national antitrust laws. ²⁶²

The effectiveness of advisory and warning letters as a form of soft enforcement is questionable. In theory, such letters can encourage compliance by bringing about the benefits associated with sunshine enforcement. Yet, even after the CMA began to publish the use of advisory and warning letters, the register includes only very general information. Aiming to protect the identity of the parties involved and avoid “naming and shaming,” this information is limited to the type of the agreement and the market concerned. ²⁶³ Such reporting, therefore, is insufficient to effectively facilitate private action.

The choice of cases to be handled by warning and advisory letters is not fully clear. The CMA has deferred to a vague standard, noting it will make use of such letters where it believes there is or may be illegal activity, but does not find “it appropriate to pursue a full investigation.” ²⁶⁴ A review of the register suggests that most cases involved resale price maintenance, and that at times also serious infringements were handled by means of letters rather than hard enforcement. ²⁶⁵ In 2021, for example, the CMA sent advisory letters to a number of firms as a response to an allegation of price fixing by suppliers of goods and services to disabled university students. ²⁶⁶

A. Deterrence and Law Compliance

The findings presented above suggest that antitrust enforcement in the European Union and the United Kingdom failed to meet the ends of the modernization reform, namely shifting from a soft to hard-enforcement regime and increasing the number of infringement decisions. The findings may also suggest that the level of deterrence of EU and U.K. antitrust enforcement is sub-optimal, that is, that there is no sufficient threat that the expected fine for infringements would be higher than the expected anticompetitive gains. This results from a combination of the low excepted probability for detection and prosecution (Points 1–5) and the expected fine (Points 5 and 6):

The limited number and proportion of hard-enforcement decisions are not decisive, in themselves, to conclude that the level of deterrence in the European Union and the United Kingdom is insufficient. Yet, such a conclusion may be supported when considering that the modernization (and the following secondary and soft laws) aimed to increase the number of hard-enforcement decisions. It may also be reinforced as the geographic markets subject to EU competition law have significantly increased in parallel to the entry into force of Regulation 1/2003 (implying an increase in the number of anticompetitive activity, see Part IV.A), and since there are no theoretical or empirical indications that the level of anticompetitive activity in the European Union and the United Kingdom has decreased. It is also supported by the findings that the introduction of settlement programs for cartels and non-cartels has not led to a greater number of infringement decisions. The presence of a sub-optimal level of deterrence compared to what was envisioned by Regulation 1/2003 may also be inferred from the observation that in contrast to the assumption that as a rule hard enforcement should be used to ensure that antitrust is applied effectively and in the general interest, hard enforcement has become the exception. As long as private enforcement in the European Union and the United Kingdom remains underdeveloped and confined to follow-on actions, firms might have little incentive to act in compliance. There is a serious risk that EU and U.K. antitrust laws do not “bite.”

A sufficient and steady threat of hard enforcement is not only a precondition for a deterrence-based approach. It is also essential for compliance-based approach and negotiated penalty settlements. Law enforcement literature—in particular Ayres and Braithwaite’s pyramid of regulatory enforcement—emphasizes that the threat of hard-enforcement underpins the functioning of voluntary compliance. ²⁶⁸ Firms are more likely to effectively cooperate with antitrust authorities when a failure to do so is likely to result in severe sanctions. ²⁶⁹ Lowering the expected detection and/or sanctions diminishes the firms’ incentives to cooperate. Over-reliance on negotiated penalty settlements, notably, was said to weaken the effectiveness of leniency programs. Hammond, the Deputy Assistant Attorney General for Criminal Enforcement in the DoJ, famously predicted that a leniency program will only be successful if there is a threat of severe sanctions; a high risk of detection; and a high degree of transparency and predictability in relation to the award of leniency. ²⁷⁰ It is doubtful whether the first two prerequisites have materialized in the European Union and the United Kingdom.

B. Failure to Achieve the Benefits of Soft Enforcement

While soft enforcement played an important part in the Commission’s and NCAs’ practice, there are no or little guiding rules on what cases are appropriate for soft enforcement and how such instruments should be designed. For example, although EU law entrusts NCAs to adopt formal commitments, it leaves their particularities to be determined by national laws. Other, less formal forms of soft enforcement are subject to even fewer controls. In contrast with the U.S. consent decrees, ²⁷¹ and the European Union’s and the United Kingdom’s formal commitments, many soft instruments are not subject to a clear procedure or substantive test, and are only confined by general principles of law.

The qualitative systematic review of the soft instruments used by the Commission and NCAs, as presented in Parts IV.B and V, demonstrated that there was no alignment between the theory of law enforcement and the particularities of the soft instruments used by the authorities. This suggests that soft enforcement has often failed to achieve the full promises associated with such instruments.

First, many of the soft enforcement efforts were not fully published and/or reasoned. Even when informal commitments, opinions, guidance, warning letters, and press releases were published, the analysis typically neither made a finding about the existence of an infringement nor was sufficiently detailed on matters of facts and law to assist such a conclusion. The reporting of soft enforcement, therefore, could not substantively clarify the authorities’ interpretation or be generalized to inform other cases. Furthermore, the highly informal nature of such actions implies that judicial review is impossible or unlikely. Taken together with the observation that much of the soft enforcement pertained to politically sensitive markets and those raising complicated legal and economic challenges, this entails that important areas of EU and U.K. antitrust are being developed by a line of soft enforcement actions that are never tested judicially.

Second, the particularities of the soft instruments used could have only a modest effect in assisting harmed parties to detect infringement and bring follow-up private actions (the “sunshine” feature of soft enforcement). ²⁷² The reporting of the enforcement actions is particularly crucial in the European Union, where third parties do not have a right to access the Commission’s investigation files. This is also true for negotiated penalty settlements, whereby both the file and the leniency and settlements applications remain confidential. ²⁷³ The Commission refuses access requests to protect the cooperative nature of its programs, and this policy was affirmed by the courts. ²⁷⁴ Because the vast majority of private actions are brought as follow-on actions, the large number of soft enforcement actions—and to some degree also negotiated penalty settlements—stands in the way of effective private litigation. ²⁷⁵

Third, the reported lack of a linear connection between the gravity of the infringement and the enforcement instrument used may deprive antitrust of its meaning. This could be the case when soft enforcement is used to settle hard-core infringements (see Parts IV and V), or when the negotiated remedies go beyond what is necessary to end the alleged infringement. In such situations, soft enforcement may serve as sector regulation, ²⁷⁶ overly focusing on the remedy instead of prevention of the wrong. ²⁷⁷

C. Underreported and Underexplored

Perhaps above all, the existence, frequency, and effects of soft enforcement and negotiated penalty settlements are underreported and underexplored in Europe. EU and national antitrust are mostly silent about the desired enforcement strategy, appropriate proportion between hard and soft enforcement, suitable type and the number of cases for negotiated penalty settlements, criteria guiding the selection of instruments, and their impact. Such a debate is strikingly missing from the Commission’s and NCAs’ annual reports, policy papers evaluating the success of Regulation 1/2003, and the Damages and ECN+ Directives. They often did not disclose the degree of reliance and efforts allocated to soft enforcement, and generally did not distinguish between fully adversarial hard enforcement and negotiated penalty settlements. This distorts the proper understanding of the enforcement system and the assessment of its effectiveness.

This point was highlighted by the U.K. National Audit Office. Its 2016 report noted that while the national antitrust authority has well-established prioritization criteria for the selection of cases, it is difficult to take robust prioritization decisions about the selection of instruments given the limited evidence about their relative effectiveness. ²⁷⁸ A similar observation was made by the Dutch NCA, noting that it does not have a method to quantify the impact of soft enforcement. ²⁷⁹

The little attention provided to the non-fully adversarial instruments hinders the proper assessment of the effectiveness of EU and U.K. competition law and the success and legitimacy of the antitrust authorities’ enforcement efforts. It also leads to overlooking the gap between what the authorities do to what they are saying they are doing—that is, between practice and the hard-enforcement aims and rhetoric. NCAs that strategically rely on soft enforcement, in particular, were not yet required to explain if and how such an approach is compatible with the shift to hard enforcement embraced by EU law and jurisprudence. All of this may result in an enforcement practice out of line with the enforcement strategy enshrined in EU and U.K. antitrust.

A. Allocation of Enforcement Efforts

Antitrust authorities should direct their enforcement efforts to pursuing the long-term public interest. They should avoid the temptation to use soft enforcement and negotiated penalty settlements to produce short-term or reputational benefits. In particular, when the general interest favors a different allocation of enforcement efforts, such instruments should not be used to inflate the number of successfully handled cases, which, as previous studies have observed, ²⁸³ is the common reputational measure of authority’s performance.

Accordingly, by aligning law enforcement theory to practice, non-fully adversarial tools should not be used to pursue: (1) low-hanging fruits, namely infringements causing a modest anticompetitive harm or that do not significantly contribute to a deterrent effect; (2) “weak” cases unlikely to result in findings of infringements if conducted in adversarial settings. ²⁸⁴ Soft instruments and negotiated penalty settlements should be reserved to cases based on an accepted theory of harm and supported by substantial evidence ²⁸⁵ ; (3) serious or repeated infringements, ²⁸⁶ in particular when the cooperative and friendly relations between authorities and firms may lead to capture ²⁸⁷ ; and (4) to circumvent judicial review. When the probability of a reversal of an authority’s decisions is high—when the law is unclear, when the case requires the balancing of conflicting interests, or when a new remedy or theory of harm are being tested—non-fully adversarial tools may be attractive from the viewpoint of an authority. Yet, they should not be used if adjudication better serves the public interest. ²⁸⁸

B. Choice of Instrument and Form of Publication

The selection of instruments and form of publication should correspond to enforcement theory. First, soft enforcement and negotiated penalty settlements should only be used when there is a genuine commitment for compliance and trust. This shared commitment needs to be tested in each case, rather than assumed. ²⁸⁹ For example, soft enforcement is only appropriate when the potential infringement was terminated immediately after the firms became aware of its potential concerns, is unlikely to be repeated, and does not involve a serious restriction of competition. ²⁹⁰

Moreover, antitrust authorities must ensure that firms are not overly pressured to cooperate given the threat of high fines, commercial and litigation costs, negative publicity, or if they lack sufficient information about the allegations they face. ²⁹¹ Especially as the Commission and most of the European NCAs function both as prosecutors and adjudicators, firms may be assured that refusal to compromise will not result in harsher charges and sanctions, ²⁹² and that they could not impact the outcome of the case during the process of negotiation. ²⁹³ The timing for selecting the instrument is key. Antitrust authorities should avoid entering negotiations before they formulate their preliminary concerns and should provide firms with a degree of depth and quality that can inform their choice to cooperate or litigate. ²⁹⁴

Similarly, authorities should be mindful that the consequences of negotiated penalty settlements and soft enforcement may extend beyond the negotiating parties. Such enforcement efforts may carry a precedential value, guiding future conduct of the authority and other firms. ²⁹⁵ In the European Union’s multi-leveled governance system, moreover, soft instruments can be used to block enforcement by other authorities. They may prevent (soft and hard) enforcement ²⁹⁶ by another authority, even when the choice of soft enforcement was informed by enforcement priorities rather than a belief that no infringement had taken place. ²⁹⁷ In cases where such effects are likely to materialized, hard enforcement may be more appropriate.

Second, procedural economy benefits should not overly influence the selection of instruments. Other effects of negotiated penalty settlements and soft enforcement should also be weighted, such as increasing compliance, legal certainty, predictability, and assisting in bringing private actions. The selection of instrument should reflect the nature of the infringement, distributional inequalities, and social implications of the infringements, intention and behavior of the violating firms, frequency of violations, and the need for an authoritative law interpretation or supervision of compliance following the termination of the procedure. ²⁹⁸

Enforcement theory should guide the form of publication and the participation rights of third parties. Authorities may be incentivized to limit the detail of reporting and public consultations in non-fully adversarial proceedings to realize procedural economy benefits and avoid resistance from the cooperating firms. Yet, such considerations should be assessed on a case-by-case basis against the other benefits of negotiated penalty settlements and soft enforcement. Realizing the full potential of a compliance-based approach, in particular, may tilt the balance in favor of opting for formal commitments over informal voluntary commitments, opinions, guidance, or warning letters.

C. A Balanced and Transparent Portfolio of Cases

Antitrust authorities should mindfully allocate their efforts across hard enforcement, negotiated penalty settlements, and soft enforcement. The selection of instrument should not only match the characteristics of the suspected infringement (the “micro” level), but also the authority’s overall strategy (the “macro” level). The benefits of using non-fully adversarial instruments to handle a specific case should be balanced against the deterrent effect of the enforcement system as a whole and maintaining the firms’ incentives to effectively cooperate with the authority in appropriate cases.

In the absence of detailed guidance to inform the selection of enforcement efforts and instruments, antitrust authorities are expected to exercise self-restraint. Hence, the attempt to align theory to practice may be supported by increasing the transparency and accountability of such choices. Authorities should explain why a particular instrument was chosen to handle each case, and annually report the total allocation of their efforts across the instruments (number of actions pursued, sanction imposed, and allocation of budget and staff). In particular, they should safeguard against an “inflation” of leniency applications and settlements.

Openly sharing the allocation of enforcement efforts invites additional checks and balances. It is expected to facilitate a discussion internally within the authority, as well as external scrutiny and public debate. ²⁹⁹ In the longer term, such debate may also help bridge the gap between the hard-enforcement aims and rhetoric and the soft practice.