The International Competition Network, Its Leniency Best Practice and Legitimacy: An Argument for Introducing a Review System

Abstract

In his contribution The International Competition Network, its Leniency Best Practice and Legitimacy: An Argument for Introducing a Review System, Steven Van Uytsel argues that the International Competition Network, as an example of a transnational regulatory network, should set up a review system of its best practices. Best practices of transnational regulatory networks, are seen as a legitimate tool for influencing the regulatory behavior of their members. These best practices are, at the end, developed by experts in the field based upon the experiences of these experts with their respective legislation or practices. Nevertheless, this chapter shows that this may be problematic if the legislation or the practice with which these experts work exhibits flaws. This is an argument that can be made in the framework of the leniency program and its best practice under the Competition Network. Van Uytsel shows that the best practice finds its origin in the leniency program of two major jurisdictions, the United States and the European Union. The leniency programs of these two jurisdictions have recently been negatively scrutinized by several scholars. Therefore, the question arises on whether best practice is really reflecting a legitimate end-result for convergence. Suggesting that it is not, Van Uytsel argues that a review process could overcome the potential threat to legitimacy in this kind of transnational regulatory network and he also offers some ideas on how this review process could be institutionalized.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 1 Introduction

When discussing transnational regulatory networks (hereinafter, TRN), the International Competition Network (hereinafter, ICN) is often mentioned as an example.¹ This network, connecting 123 national and supra-national competition enforcement agencies directly with each other,² has arisen out of the failed attempts to establish a truly international competition law to combat the ever progressing globalization of anti-competitive behavior.³ It was believed that a more informal approach would enable the growth of a common understanding on competition law and its enforcement. The common understanding could then serve as the basis for the formulation of “best practices” on which the national legislations could converge.

Among the areas of competition law identified by the ICN to be part of this process to reach a common understanding was one of the most egregious forms of anti-competitive behavior, namely cartel formation. The competition enforcement agencies sought to share their experiences for identifying cartels. One of the tools most extensively discussed in this respect has been leniency programs. In the Anti-Cartel Enforcement Manual, the drafting and implementation of an effective leniency program received a separate chapter.⁴ Compared to other tools of detecting cartel behavior, which are only being discussed in smaller subsection of another chapter,⁵ this represents a substantial part of the total.

Devoting a whole chapter on the topic of leniency must indicate the importance that the competition enforcement agencies have attributed to the leniency program as an enforcement tool against cartels. Such a status can be read from many statements made by officials. James Griffin, summarizing the past 10 years of the operation of the United States (hereinafter, US) leniency program, praised this leniency program for leading to an increase in the number of cartel prosecutions and the amount of fines imposed.⁶ Philip Lowe, detailing the history of European cartel enforcement, indicated that the real change in enforcement came with the adoption of the leniency program.⁷

If the leniency program is that important for the enforcement of cartel laws, special attention is laudable. Nevertheless, the optimism of the competition enforcement agencies, worrying voices can be heard among scholars writing on the various leniency programs. The strategic use of leniency programs has been identified as problematic for a well-functioning leniency program.⁸ Leniency programs that are driven by foreign leniency applications are criticized for not offering a proper incentive scheme.⁹ Other studies have revealed the weakness of leniency programs to trigger a race to the enforcement agencies.¹⁰

The critique that has been formulated towards the leniency programs is indirectly also a critique of the ICN’s best practices. The best practices are being shaped by the experiences of the local competition enforcement agencies. If these experiences show flaws, they will automatically be reflected in the best practices. Through the best practices, the flaws will find their way to other competition laws or their enforcement regime. By using the example of the best practices on leniency, this chapter will concretize the idea developed by Yane Svetiev on the need for a review system of best practices.¹¹ Only by installing such a system, the ICN can keep its legitimacy as a norm setter for its members.

To develop this idea, this contribution will be structured as follows. The following section will give an idea why in the area of competition a transnational regulatory network has developed that aims at formulating best practices to converge on. Section 3 will provide details on the ICN’s best practices on the drafting and implementation of a successful leniency program. To identify that these best practices have their origin in several decades of experimenting, Sect. 4 will map the experience of the US and the European Union (hereinafter, EU). That these experiences do not necessarily reflect the best outcomes will be subject of Sect. 5. This section will summarize some of the critiques that have been formulated to the previously introduced leniency programs. Best practices cannot be conciliated with critique, unless one accepts that the best practices should be open for review. Only in this way, the best practices can keep their legitimacy. Section 6 will unravel this discussion by focusing both on legitimacy issues and review of best practices. Before concluding in Sect. 8, Sect. 7 details on how this review could be conceptualized. An argument is made that the evaluation should be done by impartial evaluators and extends beyond just reviewing what has been considered as best practice.

2 2 The Growth of a TRN in Competition Law

2.1 2.1 The Internationalization of Competition Law

In his article, Evolving toward What? The Development of International Antitrust, Harry First posits that 1982 marked a milestone for the enforcement of competition law.¹² It was the year in which William Baxter, then head of the Antitrust Division of the Department of Justice (hereinafter, DOJ), met with the head of the Commission’s Directorate-General for Competition to discuss a potential divergent outcome of the respective enforcement agency’s investigation into the behavior of IBM. The worry that drove Baxter’s action was that the Commission was about ‘to order IBM to disclose computer interface specifications, a remedy that Baxter thought was unwarranted’.¹³ This disclosure order would extend beyond the territory of the then European Community. In other words, the European Commission would not have been able “to localize the effects of what they do.”¹⁴

Even though First refers to this event as the fact that ‘antitrust was internationalizing’,¹⁵ the more important fact is that it Baxter’s visit showed the direction in which international competition law was heading. In the absence of an international competition law, and the increasingly willingness of enforcement authorities to assert jurisdiction over foreign conduct resorting effects in their territory, national enforcement authorities had to coordinate their activities in order to avoid conflicting outcomes.

The level of coordination was twofold. On the one hand, the knowledge of the existence of divergent views on competition law between different jurisdictions incentivized the enforcement authorities to start looking into the possible coordination of their general policies. On the other hand, the ever increasing liberalization of the world economy brought about more international orientated competition law cases. Many cartels had no longer just domestic effects, but effects worldwide. Merger increasingly occurred across borders. In order to streamline enforcement, either in terms of finding evidence or in terms of applications and substantive outcome, the enforcement authorities turned bilateral agreements to facilitate the investigations.

2.2 2.2 From Bilateral Cooperation Agreements to the International Competition Network

The bilateral agreements on cooperation and the informal channels for cooperation may have facilitated the consultations between the competition law enforcement authorities, the formal and informal contacts did not overcome all of the problems caused by the extraterritorial reach of competition laws. The incidents with the proposed acquisition of McDonnell Douglas by Boeing and of Honeywell by General Electric most eloquently illustrate this issue. The divergent views on the interpretation of the consequences of conglomerate mergers in both dossiers clearly indicated that focusing on the procedural aspects of competition law was not going to provide a satisfactory answer. With the one merger being substantially altered and the other one prohibited, these two ‘high profile and acrimonious cases were something of a wake-up call to the antitrust officials in both the United States and Europe’.¹⁶

Within the context of the dispute between the EU and the US on what is a better understanding of competition law, the International Competition Policy Advisory Committee (hereinafter, ICPAC) was established in 1997.¹⁷ This advisory committee was to assist the US DOJ in looking for the ‘new tools, tasks and concepts that will be needed to address the competition issues that are just arising on the horizon of the global economy’.¹⁸ The Committee was composed of 13 members. Besides the Committee chairpersons, James Rill and Paula Stern, and its executive director Merit Janow, seven business and foundation executives and three professors were attracted to take part in the hearings, deliberations and formulation of recommendations.¹⁹ Three years after its establishment, ICPAC submitted a report to Attorney General Reno and Assistant Attorney General Klein.²⁰ This report contained numerous recommendations, one of which was the establishment of a Global Competition Initiative (hereinafter, GCI).²¹

Eleanor Fox, one of the professors taking part of ICPAC, described the outset of the GCI as follows:

The GCI was envisioned as a virtual, voluntary forum with no ground address or secretariat, no power to make binding rules, and no power of adjudication. The idea for the enterprise stemmed from the realization that antitrust authorities, business people, and experts lacked a forum for the sharing of views and experiences, for close cooperation, and for exploration of common issues that could lead to convergence or harmonization.²²

It was Assistant Attorney General Klein who took the initiative to launch the proposal at the international level. At the Tenth Anniversary Conference for European Merger Control, Joel Klein indicated that the bilateral efforts were not a sufficient answer to the problems that were caused by globalization. There should also be a focus on the substantive part of competition law. However, he acknowledged that there was not forum suitable for this job.²³ Inspired by the proposal of setting up a GCI, Klein called for an initiative at the global level, something that may eventually pave the way for multilateralism within the field of competition law.²⁴

Even though the idea for a GCI was positively welcomed, the contours of this Initiative had still to be drawn. At a meeting in Ditchley Park, organized by the International Bar Association, with support from the American Bar Association Antitrust Law Section and the Fordham Corporate Law Institute,²⁵ it became clear that the government agencies wanted to take control over the GCI.²⁶ However, without support of the US, it was unclear whether the GCI would be a viable initiative. As early as the Ditchley Park meeting, there was no certainty on whether the newly inaugurated President George Bush and his administration would back up the initiative.²⁷

Attracted by the idea that the new initiative would only cover competition law and focus on issues for which solutions would be achievable, the newly appointed Assistant Attorney General, Charles James, and Timothy Muris, newly appointed to the position of Chairman of the Federal Trade Commission, were eager to support the GCI.²⁸ Once ascertained of this support, further consensus was being sought among other competition law enforcement authorities. By the time that the Fordham International Antitrust Conference was being held in 2001, this consensus was achieved among 14 jurisdictions.²⁹ The enforcement authorities of these 14 jurisdictions used the opportunity of the Fordham Conference to launch the initiative and named it the International Competition Network.

2.3 2.3 The Role of ICN as a TRN in Convergence

The ICN has, from its establishment, aimed at ‘addressing antitrust enforcement and policy issues of common interest and formulate proposals for procedural and substantive convergence through a result-orientated agenda and structure’.³⁰ The idea of convergence has been complemented with the encouragement of ‘the dissemination of antitrust experiences and best practices’³¹ and promoting the ‘advocacy role of the antitrust agencies’.³² This initial aim has been restated in the Operational Framework that the Steering Committee formulated in 2012.³³ The ICN’s website incorporates a short restatement of this mission statement. In what it calls its mission statement on the top of the ICN website, the ICN advocates the ‘the adoption of superior standards and procedures in competition policy around the world, formulate proposals for procedural and substantive convergence, and seek to facilitate effective international cooperation to the benefit of member agencies, consumers and economics worldwide’.³⁴

Convergence is an idea that is pervasively present in all these documents. Convergence has therefore been given much more attention in the document elaborating the vision for the second decade of the ICN, convergence was given a context. In broad lines, the ICN Steering Committee defined convergence as the ‘voluntary adoption of widely-accepted norms of competition policy, substantive standards, procedures and levels of institutional capacity’.³⁵ Inherent in the concept of convergence is divergence. One can only move in the direction of widely accepted standards if the current practices are different from each other. This is also reflected in the vision on the second decade. Convergence is described as running through three different stages. Convergence, as also has been identified by Maurice Stucke, can only occur if there is an agreement on norms, standards and procedures that have been divergent in the past.³⁶ Therefore, the first stage towards convergence is the implementation of different norms, standards and procedures. This will allow for experimentation to see which of these norms, standards and procedures are more effective than the others. Sharing the information and experiences of the experimentation is the second stage. This part of the process will facilitate the identification of best practices, which could be then put forward as the benchmark.³⁷ The role of the ICN in this evolutionary process is to ‘promote the flow of information about different agencies’ ongoing experiments and feedback from these experiment’.³⁸ The third stage is that individual jurisdictions opt for the benchmark, the norm that has received consensus as being the best possible solution for specific problems.³⁹

The description of the ICN’s role in the convergence process colludes very well with what Anne-Marie Slaughter has termed an information network.⁴⁰ Across the various member organizations of the ICN, there is a lot of information on how they operate on a procedural and substantive level. The information is so overwhelmingly vast that it creates a ‘paradox of plenty’,⁴¹ in which not much attention is paid to the essence at stake. In order to come to some kind of common view, Robert Keohane and Joseph Nye emphasize the need to have ‘editors, filters, interpreters and clue-givers, as well as “evaluators” in distilling power for the plentitude of information’.⁴² The ICN fulfills this role. This TRN actually collect[s] and distill[s] information about how their members do business.⁴³ In order to have an operational end product, Slaughter has indicated that the TRN’s work usually end up in ‘a code of “best practice,” meaning a set of the best possible means for achieving a desired result’.⁴⁴

Stucke sees the ICN’s cartel leniency program as a good example of this process. Various members of the ICN had implemented a leniency program and have experimented with different formulations of the leniency programs. These experiments allow other members to learn in what format the leniency program will work well and see whether that format is well suited for their jurisdiction. The latter part of the convergence process has been guided by a best practice, which was, according to the ICN’s Steering Group, not too complex to formulate due to the relatively narrow differences in this area. Stucke notices, with reference to the ICN’s website, that this has led to a massive implementation of the leniency program in various jurisdictions.

3 3 The ICN’s Best Practices on Leniency

Before the Cartel Working Group started to compile the best practice on the implementation of a leniency program in Sydney in 2004,⁴⁵ several independent studies were already undertaken by economists to map out in what kind of situation a leniency program would operate efficiently. The economists made some general predictions, put forward guidelines related to the cartel member’s behavior and prescribed how enforcement authorities should handle leniency applications. Even though these economic studies rely on models, limiting their scope of applicability due to the dependency on the parameters within which these models were framed,⁴⁶ a common line is detectable in the conclusions of these studies. There is a general agreement that leniency programs, of which the actual content may differ according to the theoretical study, will induce cartel participants to come forward with information on illegal cartel activity.⁴⁷

The Cartel Working Group, that elaborated the best practices for the ICN, added more specifications to this analysis. The Cartel Working Group does not necessarily disagree with the economists’ conclusions, but this Working Group obviously only wants to focus on ‘best practices’, thus things that work properly. Therefore, the Working Group basically stated that a successful implementation of a leniency program requires several prerequisites. Among these prerequisites is the high risk of detection, significant sanctions, and transparency and certainty regarding the application of the leniency program.⁴⁸ Even though transparency and certainty regarding the application is not a prerequisite, but rather something that is part of the conceptualization of the leniency program itself, the Working Group recommends a leniency program for cases in which the enforcement authority is not aware of the cartel or where the authority is aware of the cartel but does not have sufficient evidence to proceed to adjudicate.⁴⁹

The economic studies also looked into the need to have a distinction between a pre-investigation and a post-investigation leniency application. Disagreement exists in relation to the moment leniency should be provided. Massimo Motta and Michele Polo assume that lenient treatment is not only efficient in the pre-investigation stage, but also in the post-investigation stage.⁵⁰ Giancarlo Spagnolo, on the contrary, argues that cartels are convicted to disappear once they are detected. Hence, there is only a need to focus on cartels that are not yet under investigation.⁵¹ However, starting from the presumption that a leniency program functions on the basis of a cost-benefit analysis, there would be no reason to exclude this kind of rational behavior from the post-investigation stage. A majority of the literature confirms this viewpoint.⁵²

The best practice developed by the ICN does not provide much detail on the distinction between a pre- and post-investigation leniency application. Indirectly, it is possible to deduct from the best practices that both a pre- and post-investigation leniency application could be a good practice. First, when discussing the full and frank disclosure requirement, the guiding text formulates that leniency may also be available after an investigation has commenced. However, this guiding text limits this possibility to the first eligible applicant.⁵³ Second, the best practice on subsequent applicants does not exclude the possibility that a leniency program extends to more than one applicant. Whether the subsequent applicants defect before or after a dawn raid is left open in the formulation of best practice.⁵⁴

In order to convince cartel members to defect the cartel and come forward with information leading to the breakdown of the cartel has been discussed in the economic literature in relation to the amount of leniency that should be offered. In an optimal situation, a generous reward is offered to the applicant.⁵⁵ Less aggressive leniency programs, which only offer a reduction, will be less effective.⁵⁶ Again, the more lenient treatment is offered in less courageous programs the more effective these moderate programs will be.⁵⁷ Whether the program is courageous or modest, the probability of reporting increases if the lenient treatment is restricted to a certain number of firms. The fewer the number, the more likely it will be that a cartel participant will come forward with information.⁵⁸ Some studies even point out that rewards for individuals will be more effective than the ones for corporations.⁵⁹

The ICN best practices are again not very informative about the level of leniency that would achieve the best result. Nevertheless, when discussing what the subsequent leniency applicants should receive, it is stated that these should receive less than full leniency.⁶⁰ Indirectly, we can conclude that the best practice for the first applicant is to give immunity and thus guarantee that no penalty will be imposed. How much the second applicant or any subsequent application should receive is not stipulated in the best practice. It has been stipulated that in some jurisdictions a 50 % reduction in fines is applied. Another option would be to make the level of leniency dependent on the quality of the information or evidence or the speed with which they report.⁶¹

Even though the observations of the economic studies focus mainly on a cost-benefit analysis of the cartel participants, indirectly they have also an impact on the behavior of the competition authorities. By concluding that certain incentives have a positive effect on reporting the illegal cartel activity, these incentives should not be jeopardized by actions of the competition authority. This has an impact on the substantive formulation of a leniency program. A leniency program should neither grant powers to competition authorities to “second-guess” the application, nor contain provisions obstructing or obscuring the application process.⁶²

The Cartel Working Group devoted much attention to this last aspect. Several conditions that could be attached to a leniency program are being discussed in the best practices. Full and frank disclosure of information is regarded as a best practice.⁶³ Similar, the ongoing cooperation of the leniency applicant is seen as necessary.⁶⁴ Confidentiality of the identity of the leniency applicant and his information will also contribute to a better functioning leniency program.⁶⁵ Another category of good practice in relation to the enforcement authorities link to the fact that there needs to be a guarantee that the enforcement authorities cannot jeopardize the application⁶⁶ and that there needs to be insurance on the certainty for applicants where investigations are closed without an enforcement action.⁶⁷ When engaging with the enforcement agency, a leniency applicant should be able to reserve its position in the queue and submit information after he has made the reservation. In other words, the creation of a marker system is a good practice.⁶⁸

4 4 Lessons Drawn from Three Decades of Experimentation

When the ICN started to elaborate the best practices on the leniency program, not that many jurisdictions had experimented with a leniency program. The US has the longest experience. The US could draw lessons from not only the conceptualization of its original leniency program, but also from legal changes to that program. The ICN could also revert to the relatively long experience of the EU with a leniency program. Similar as to the US, the EU also has amended its leniency program once before the Cartel Working Group started to concentrate on the issue of designing an efficient leniency program.

4.1 4.1 Pre- and Post-Investigation Leniency Applications

The only jurisdiction that has been experimenting with pre-investigation incentives has been the US. Their original Corporate Leniency Policy (1978 Leniency Policy),⁶⁹ which was established in 1978, offered a lenient treatment only in the pre-investigation stage. Corporations could not enjoy lenient treatment under this policy once the DOJ had started its investigation. When the 1978 Leniency Policy came up for revision in the 1990s,⁷⁰ one of the elements that changed in the policy was exactly this point of the scope of the leniency program.⁷¹ The program was expanded to leniency for the post-investigation stage, creating the presumption that this expansion would augment the likelihood of discovering and punishing illegal cartel activity.⁷²

If this policy change had been the only one, the subsequent increase of applications would definitely have been enough proof of the necessity to have a leniency program in the post-investigation stage.⁷³ However, as will be indicated below, many other reasons have prevented the 1978 Leniency Policy to be successful. Therefore, the policy change cannot be more than a presumption of the necessity to have a leniency program for the post-investigation stage.

Welfare considerations oblige to consider the installation of a leniency program in the post-investigation stage. Investigations are costly. Competition authorities, which have a suspicion on illegal cartel activity, will have to make the necessary human, financial and material resources available to start an investigation. Obtaining information from the cartel participants has its limitations, though. The inspections at business premises or private houses of company employees will only give positive results if physical evidence exists. Whether or not this evidence exists, the competition authorities will have to find out. Unless they have specific information about the existence of the information and the place to find it, they will have to spend a lot of time going through many documents with the risk of finding nothing at all. In the latter case, all the resources made available are wasted.⁷⁴

The costs of a certain method of investigation should be weighed against its benefits. If the costs outbalance the benefits, the use of that method is not justified.⁷⁵ A less costly method should be preferred. It seems clear that leniency programs can, if they are well designed, lower the search costs. Indeed, as Wouter Wils indicates, these costs will be shifted from the competition authority to the company and its staff.⁷⁶ Since they are more familiar with the illegal activity, collecting the relevant information will likewise be much cheaper. Whereas this analysis says something about the desirability of a leniency program in the post-investigation stage, nothing can be deducted from these considerations as to how the leniency program should look like in detail.

4.2 4.2 Immunity as an Incentive Inducing Cartel Defection

The explicit silence on the need to provide immunity for the first applicant in a leniency application as a best practice is somehow peculiar given the explicit lesson that could be drawn from the US and EU experience. It could be said that, as long as immunity is guaranteed in a pre-investigation stage, firms are willing to reveal information on illegal cartel activity. Providing guaranteed immunity in a post-investigation stage does not seem to be a necessity in order to induce firms to cooperate with competition authorities in the framework of a leniency program. A combination of the lack of guaranteed immunity with the existence of leniency in the post-investigation stage seems to support collusion. It is not possible to draw conclusions on whether this immunity should be restricted to the first applicant or whether reductions should be offered to any subsequent applicant.

These conclusions are based upon the following considerations. The US Leniency Policy, whether it was the 1978 or the 1993 version, limited the lenient treatment to the first company successfully applying. The difference between being the first and the second is immense in the US context. The first corporation will enjoy immunity, while the second, theoretically speaking,⁷⁷ will have to bear the consequences of a cartel prosecution.⁷⁸ This difference is supposed to set up a race between corporations to the door of the Antitrust Division of the DOJ.⁷⁹ This race, usually referred to as the “race to the courthouse door,”⁸⁰ will mortgage cartel activity. It will be very hard to establish the necessary trust among corporations to engage in cartel activity.⁸¹ In the 1993 format, the Leniency Policy got about three applications for immunity per month.⁸² However, in the two first years after the new policy, only 15 corporations applied.⁸³

Even though the 1996 Leniency Notice did not conceptualize a guaranteed immunity, its practice has some relevance for assessing the theoretical considerations above. Whereas the level of leniency may have been uncertain under this program, the differences between the amount of reductions was minimal. Of all cases during the period when the 1996 Leniency Notice was in force, only three cases were related to immunity.⁸⁴ All the rest were leniency cases after the Commission did a dawn-raid.⁸⁵ Even though the degree of leniency in the pre-investigation stage was still higher than in the post-investigation stage, undertakings did not attempt to reveal any information in the pre-investigation stage. This may indicate that the undertakings have been waiting until the moment they had to save their skin. In other words, absence of a clear incentive triggers a waiting game.

When immunity became established as a certainty in the 2002 Leniency Notice, a shift was noticeable from a waiting game to an assertive use of the leniency program. Rather than waiting and applying for reduction, the undertakings engaged in illegal cartel activity straightly applied for immunity.⁸⁶ The majority of leniency applications in the 3 years after the adoption of the 2002 Leniency Notice was for immunity and submitted before an investigation took place.⁸⁷ The data on the remaining part of the applications do not allow categorizing these applications for reduction in the pre- or post-investigation stage. It is not unthinkable, however, that at least a part of the applications situate in the post-investigation stage.

With an average of 25 applications for immunity per year, the 2002 Leniency Notice reaches 11 applications less than the 1993 Leniency Policy. Whether this difference is attributable to the fact that a second, third and even fourth undertaking can enjoy leniency thus causing a waiting game, is difficult to say. However, if we know that nearly half of the immunity applications in the US have occurred at the post-investigation stage, while the immunity application in the EU are all in the pre-investigation stage, the conclusion that leniency to more than one firm leads to a waiting game, most likely does not hold. The risk was not too high before.

The expected leniency discount needs to be sufficient in order to outweigh the possible gains of the cartel. In this respect, both the EU and the US have put a full immunity of 100 % forward. The higher the penalties that can be waived, the higher the success rate of a leniency program. This also implies that the infringer must be able to calculate the amount of the fine. In order to fortify the strength of the immunity, the US has also regulated that the treble damages, usually applicable to antitrust infringements, will be reduced to single damages.

4.3 4.3 Subsequent Leniency Applications Impairing the Race to Defect

The 1978 Leniency Policy granted complete immunity to the first successful applicant in a pre-investigation stage. There was no leniency provided under this program for any other cartel participant. The 1993 Corporate Leniency Policy (1993 Leniency Policy) did not change this. It only added immunity for the first successful applicant in the post-investigation stage.⁸⁸ A different approach was taken by the EU. The 1996 Leniency Notice provided for immunity or reduction for the first successful applicant in the pre-investigation stage. Subsequent applicants in this stage could enjoy leniency as well. In the post-investigation stage, only reduction was offered to the applicants.⁸⁹ The revision of this policy in 2002, only changed the format of immunity in the pre-investigation stage.⁹⁰ The 2006 revision did not change anything related to the incentives.⁹¹

To know whether the US’s approach towards leniency is more effective than the European one, both systems have to be contrasted with each other. In an empirical assessment of the 1996 Leniency Notice, Stephan Andreas investigated whether the Notice could induce undertakings to come forward and reveal an illegal cartel.⁹² In a period between 1996 and 2005, Stephan counted 33 cartel cases in which the Commission had taken a decision.⁹³ Out of the 33 cases, 20 were triggered by a leniency application.⁹⁴ The 20 cases could then be further divided in two categories: cases that have a US preceding or simultaneous investigation, or cases that were only investigated in the EU.⁹⁵ The former outnumbered the latter by eight, allowing Stephan to cautiously conclude that 14 EU leniency cases are likely to be on the back of a successful US Leniency Policy.⁹⁶ Indirectly, the author suggests that the US Leniency Policy was better conceptualized.

More important than the observation that EU leniency cases are preceded by investigations in other jurisdictions, the leniency applications in the EU were mainly after dawn-raids by the Commission were held.⁹⁷ In other words, the 1996 Leniency Notice was most successful in the post-investigation stage. This Notice was not conceptualized to induce undertakings to come forward with information in a pre-investigation stage. In fact, immunity was only granted in three cases over a period of 6 years. This is in stark contrast with the 2002 Leniency Notice, which was able to attract 20 applications for immunity in the first year of being in operation,⁹⁸ with a similar amount of applications in each of the next 2 years.⁹⁹ Unlike the 1996 Leniency Notice, the 2002 Leniency Notice establishes automatic immunity.

4.4 4.4 Discretional Granting of Leniency and Its Counterproductive Effects

The 1978 Leniency Policy attached several conditions for receiving immunity.¹⁰⁰ Most of the conditions were reasonable. In exchange for immunity, the applicant had to be the first to provide with candor and completeness previously unknown information, to promptly terminate its participation in the illegal activity, to continuously assist the Antitrust Division of the DOJ in their investigation, and to restitute the injured parties if possible. Further, the applicant should not have coerced others to participate or being the originator or leader of the illegal activity. However, one condition in specific was problematic since it was not in the control of the applicant. Even if the corporation would have met all the previously mentioned conditions, the DOJ could refuse immunity based on the criteria of “reasonable expectation.”

The condition of reasonable expectation implies that whenever the DOJ has a reasonable expectation that it would have discovered the reported illegal activity even if the corporation had not reported it, the DOJ would not grant any lenient treatment.¹⁰¹ The insecurity created by this provision was immense. The cost-benefit analysis to cooperate or to come forward with information could not be made anymore.¹⁰² For each calculation, the potential applicant had to predict the judgment of the DOJ. Without precedents, such a prediction is hard to make. Therefore, corporations chose to err on the side of caution and made the calculations on the worst presumptions, making the balance nearly always incline to the cost side. Nearly no corporation came forward with information.

Indeed, the 1978 Leniency Policy was barely used. According to the Antitrust Division of the DOJ, 17 corporations applied for leniency between 1978 and 1993.¹⁰³ Six requests were denied, one case was pending and ten corporations qualified for immunity at the time of revision.¹⁰⁴ Only four out of these ten corporations qualified for immunity before 1987, the year in which a leniency program for individuals started. All the other six requests followed, suggesting that the increased success of the Leniency Policy is partly due to the instigation of this policy.¹⁰⁵ The initial Leniency Policy had an average of approximately one leniency application per year.¹⁰⁶ Once the immunity was granted automatically, the application rate increased 20-fold.¹⁰⁷

The situation in the EU was slightly different. The 1996 Leniency Notice was indecisive as to the degree of leniency provided to a cooperating undertaking. Rather than stating that an undertaking that is the first to successfully cooperate would be granted immunity, the 1996 Leniency Notice left a discretional margin to the Commission. The first undertaking to report in a pre-investigation stage would benefit a reduction of 75 % or more. In the best case, this could amount to immunity. In a post-investigation stage, the first undertaking would enjoy a reduction between 50 and 75 %. The criterion to choose the degree of leniency was the decisiveness of the evidence to reveal the existence of an illegal cartel. Hence, it was up to the Commission to assess the value of the evidence provided.

Assessing the value of evidence provided is an internal process of the Commission. It entirely depends on how much evidence the Commission already has and what it will be able to acquire. Unless the applicant for leniency does not have a clear view on this aspect, as far as it is possible for evidence that may be acquired in the future, he will not be able to calculate his potential benefit of applying. From the viewpoint of a potential applicant, this leniency program will be perceived as ‘there might be some relief in relation to a potential fine from the Commission’.¹⁰⁸

The Commission saw the 1996 Leniency Notice as a success. Mario Monti, the at that time Competition Commissioner, stated in a press release in July 2001 that ‘[t]he Leniency Notice has played an instrumental role in uncovering and punishing secret cartels’.¹⁰⁹ Yet, this chapter has already put forward a study of Stephan to refute this viewpoint.¹¹⁰ Due to the uncertainty of obtaining immunity, there was no longer a need for undertakings to reveal any information in the pre-investigation stage. Taking the worst-case scenario in mind for the pre-investigation stage while making the calculus, the undertakings would have found out that it equaled with the worst-case scenario in the post-investigation stage. That is 10 % reduction of the penalty. Nearly all of the leniency applications thus also happened in the post-investigation stage, after the Commission started investigation.¹¹¹

Making immunity uncertain and putting the subsequent reduction penalties close to each other has consequently led to a major waiting game by the undertakings. Johan Carle, Pervan Lindeborg and Emma Segenmark somehow confirm this result of the 1996 Leniency Notice in the following terms:

[S]ince its entry into force in July 1996, the 1996 Notice has, as far as we are aware, merely been applied in approximately 16 cartel cases. In the majority of these cases the co-operating entity was only granted a reduction of 10–50 per cent. A very substantial reduction of 75 per cent has, as far as we have been aware of, been granted in a handful of cases under the 1996 Notice.¹¹²

Somewhat contradictory to his previous statement, Monti acknowledged that better results in this waiting game could be achieved by giving better incentives to the undertakings.¹¹³ However, it turned out that the way in which the incentives were conceptualized, was wrong.

4.5 4.5 The Importance of Clear Conditions

A leniency program is a complex web of conditions, related to information, order of application, time of application, obligations for the leniency applicants or the role the cartel participant has played. The broader the scope of the leniency program, the more complex this web will be.

Leniency programs are conceptualized in order to get information about illegal cartel activity. The US 1993 Leniency Policy stipulates in relation to information that the DOJ does not have received the information yet; the applicant reports it with candor and completeness.¹¹⁴ In a post-investigation stage, the information should be likely to result in a sustainable conviction.¹¹⁵ The EU 2006 Leniency Notice is much more detailed. The first information submitted should enable the Commission to carry out targeted inspections¹¹⁶ or to find an infringement of article 101 TFEU¹¹⁷ on the condition that the Commission does not have enough evidence yet to pursue either of them.¹¹⁸ The information should be complete.¹¹⁹ Further, the information should not be falsified nor disclosed to other persons.¹²⁰ For subsequent applications, the information should represent a significant added value.¹²¹

The time element in a leniency program points at the stage in which the applicant files for leniency. There are two stages; the application is filed either in the pre-investigation stage or in the post-investigation stage. The 1993 Leniency Policy does not stipulate the element making the difference between both stages, but these stages are clearly separated.¹²² In the 2006 Leniency Notice, the distinction between the two stages is less clearly described. The post-investigation stage is indirectly pointed at by stating that immunity can be obtained if the applicant provides information leading to the establishment of an infringement of article 101 TFEU EC, presuming that this can happen even after the Commission has done a targeted investigation.¹²³ Hence, a targeted investigation seems to be a lever between a pre- and a post-investigation stage.

Within the time element, it is important to know the order in which the applications are submitted to the competition authorities. The 1993 Leniency Policy determines that the first applicant can obtain immunity, whether it is in the pre- or post-investigation stage.¹²⁴ Similarly, the 2006 Leniency Notice mentions that the first applicant will be able to obtain immunity.¹²⁵ For the second, third and any other applicant, only a reduction of the penalty is available.¹²⁶ Both systems provide for a marker system to secure the first position in an immunity application.¹²⁷ For a reduction application in the EU, the order will be provisionally determined based on the order of submission on the condition that the information contains significant added value.¹²⁸ The order is final at the moment the Commission takes the final decision.¹²⁹

Within the obligation part, several conditions are grouped together. Some of the obligations are related to the illegal activity directly. The 1993 Leniency Policy requires the applicant to prompt and effective actions to terminate the illegal activity.¹³⁰ Similarly, the 2006 Leniency Notice requires the applicant to have ended its involvement in the alleged cartel immediately following its application.¹³¹ The obligations in relation to information have been discussed above. Other obligations relate to the cooperation with the competition authorities. Both the 1993 Leniency Policy and the 2006 Leniency Notice demand continuous cooperation with the competition authorities.¹³² Still another obligation relates to the injured parties, be it only in the US. The 1993 Leniency Policy asks for restitution of injured parties where possible.¹³³

In relation to the cartel participants, both the 1993 Leniency Policy and the 2006 Leniency Notice have provisions in relation to coercion. The programs do not allow immunity to be given to corporations that have coerced other parties to participate in the cartel.¹³⁴ The 1993 Leniency Policy also has one in relation to the ringleader. It stipulates that the leader or the originator of the illegal cartel activity cannot claim immunity.¹³⁵ In the post-investigation stage this is put under the general concept of unfairness.¹³⁶ The latter does not limit the scope of application for the ringleaders, but it does so for undertakings that have been coercing others to undertakings to participate. These undertakings will only be eligible to apply for reduction but not for immunity.

The above-described conditions already reflect the experimentation with leniency programs for about three decades. Some of the conditions have not been problematic at all from the beginning. The conditions on coercion have been part of the earliest leniency programs of the US, the 1978 Leniency Policy, and the EU, the 1996 Leniency Policy, without much change. Similarly, the obligation to terminate the illegal activity and to continuously cooperate with the competition authorities has been part of the leniency programs since they were established in the US and the EU. The conditions in relation to information and order have caused more controversy and uncertainty for the application of the leniency programs. Besides, some concepts, such as ringleader and originator, have a history track of changes.

The problems in relation to the information provisions in the respective leniency programs are twofold. On the one hand, the applicant has to overcome the burden of finding out whether the illegal cartel activity has already been reported on or not. On the other hand, the applicant has to assess the meaning of general terms as “illegal activity,”¹³⁷ “sustainable conviction,”¹³⁸ “targeted inspection,”¹³⁹ “infringement of Article 81”¹⁴⁰ or “significant added value”.¹⁴¹ During the nearly three decades of experimenting with leniency, the US and the EU have learned a lot in this regard. Especially the EU has been paying attention to these aspects, as it revised its Leniency Notice in 2006 to reflect the necessity of creating transparency in relation to information.

The 1993 Leniency Notice requires previously unknown information from the first applicant in order to consider immunity from penalties. Something similar is inscribed in the 2006 Leniency Notice. The Commission may have already enough evidence for adopting the decision to carry out a dawn raid or for finding an infringement of EC, and so nullifying the right to obtain immunity. If the applicant is not aware of the deal that is on the table, the leniency application will be a poker game.¹⁴² However, it will be a distorted poker game. The competition authority would play with its cards close to its chest, while the applicant has to put all its cards on the table.¹⁴³ Much has been done to avoid this kind of situation, both in the US and the EU.

The US DOJ’s approach towards this problem has been to allow anonymous non-prejudicial immunity inquiries.¹⁴⁴ The inquiry only needs to reveal information about the particular industry or a specific area of economic activity.¹⁴⁵ The EU approach is different.¹⁴⁶ Anonymous inquiries are not accepted.¹⁴⁷ Instead, the Commission has devised a hypothetical application mechanism. This system exists since the 2002 Leniency Notice.¹⁴⁸ Unlike the US inquiry, the hypothetical application will need to supply quite a lot detailed information amounting to the level of evidence.¹⁴⁹ Whether one system should be preferred above the other, depends on the conception of the leniency program. Anonymous inquiries may result in an abuse when more firms can enjoy lenient treatment, while in a system creating a race to the courthouse door it may work perfectly.

Generally formulated conditions related to information constitute the other major problem. Again, each of the investigated jurisdictions has such concepts in its leniency programs. The 1993 Leniency Policy has such a general concept in the pre-investigation stage, “illegal activity,” and in the post-investigation stage, “sustainable conviction.” With “targeted investigation,” “infringement of Article 81,” or “significant added value,” the 2006 Leniency Notice has more generally formulated conditions. The two previous versions of the Leniency Notice had similar conditions included. However, unlike the 2006 Leniency Notice, the previous versions did not elaborate on the meaning of these generally formulated conditions. A case-by-case evaluation had to prosper the necessary precedents.¹⁵⁰ Judging from the Commission’s reaction in 2006, this work method did not provide the necessary clarity and certainty.

Left with a great deal of discretion, the Commission had to be ‘vigilant to ensure consistency’.¹⁵¹ Consistent treatment is not always easy to pursue. The EU practice has shown that the distinction between concepts started to blur, by asking more evidence than required under the one condition and less than required under the other condition.¹⁵² Therefore, the DOJ has adopted a twofold policy. First, the initial amount of information does not need to be more than a “good cartel story,”¹⁵³ which will expand, mainly driven by the DOJ, later on.¹⁵⁴ Second, the DOJ will “err in favour of the applicant where there is a genuinely close call.”¹⁵⁵ The Commission has never made statements in this regard. Instead, it has reformed its 2002 Leniency Notice in 2006 to create “upfront certainty on the part of a would-be leniency applicant as to the information and evidence required by the Commission.”¹⁵⁶ Conditions like “targeted investigation,”¹⁵⁷ “infringement of Article 81 EC,”¹⁵⁸ and “significant added value.”¹⁵⁹ By clarifying these concepts, the time element has become much more transparent then before.

Information that has to be submitted to the competition authorities in the US and the EU differs considerably. Whereas the US 1993 Leniency Policy reaches certainty by its simplicity, the EU 2006 Leniency Notice achieves it by a detailed description of what has to be submitted. This complexity has been extended to the issue of inquiring whether the Commission already has enough information about the cartel and so to check whether the applicant can still enjoy immunity. This can be explained to avoid any kind of abuse. However, the complexity surrounding the submission of information could explain why the Commission still attracts less leniency applications.

A firm, calculating whether it is profitable to defect the cartel, needs to be sure that it can win the race to the courthouse door. In other words, the leniency procedure needs to offer the firm the certainty that, when it makes the initial step, the position secured by this step does not get lost. The initial step may have to be taken in quite a rush. Yet, in order to obtain immunity, the firm has to come forward with enough information related to the illegal cartel. The hastiness, in which the initial step had to be taken, may have caused a lack of time to prepare the information as evidence sufficiently. The incompleteness of the application may not be a problem at first. However, when another firm realizes what happened, it may be inclined to submit an application containing more relevant information. Due to the high value of the second applicant’s information, he may supersede the first application. Such a situation can occur if the initial step does not secure the order of application.

The 1996 and 2002 Leniency Notice reflected this situation.¹⁶⁰ Undertakings applying for leniency derived benefit from submitting extensively documented leniency application to the Commission. Incomplete leniency applications were dangerous in two ways. First, the application could have been rejected on the ground that it did not fulfill all the substantive conditions.¹⁶¹ Second, another undertaking may be getting ahead and offer a “smoking gun”¹⁶² to the Commission.¹⁶³ Commission officials have pointed out that in the latter case it is the Commission’s practice that ‘the moment the second applicant submits evidence, the first applicant can no longer supplement its application with further evidence. Its application will be evaluated on the basis of the evidence it had submitted until the moment the second application was made’.¹⁶⁴ This uncertainty, combined with generally formulated conditions related to information, have without a doubt scared off risk averse undertakings to make use of the leniency program.

The US practice differs in that it allows the applicant to put down a marker, a sign keeping a place in the queue. This marker can be easily set. A call to the Antitrust Division of the DOJ requesting a marker with the explanation that the corporation needs more time to collect the evidence is usually sufficient.¹⁶⁵ The marker will be granted, almost always together with a time limit. Within this time limit, the applicant has, in principle,¹⁶⁶ to perform his promises; this is collecting and arranging information allowing him to make a proffer. The proffer is basically an outline of what the applicant is able to offer, and it does not need to be evidence.¹⁶⁷ At the end of the proffer, a conditional leniency letter can be asked. The actual grant of immunity will follow in short order.¹⁶⁸ In other words, evidence is looked for after the granting of immunity and largely driven by the DOJ.¹⁶⁹

A marker system contributes to the predictability of a leniency program. Several scholars have therefore argued that a similar system should be introduced in the Leniency Notice. This happened in 2006.¹⁷⁰ The marker system that the Commission introduced has a set of objective conditions to be fulfilled. Yet, they are not sufficient to guarantee that the marker will be granted. This is reflected in two elements. First, the Commission may grant a marker.¹⁷¹ Second, the applicant has to justify its request for the marker.¹⁷² The former will likely have an effect on risk-averse undertakings. Rather than applying for a marker, they probably prefer to make a full immunity application. In doing so, they may have lost the race or at least delayed the whole process. It is clear that the race to the regulator is undermined.¹⁷³ The latter puts the applicant in a defensive position. What else than disclosing a cartel could justify the request for a marker? How detailed does the applicant have to describe his inability to come forward with the necessary information at the moment?¹⁷⁴ For sure, without much more clarity on this aspect, undertakings may be dissuaded from approaching the Commission.

5 5 Critical Voices on the Existing Experience with Leniency Programs

Leniency programs have been generally praised for their success in detecting cartel behavior. Even though the general optimism on the function of leniency programs, some studies have pointed out the weaknesses of the operation of a leniency program. It is not the purpose of this contribution to review all critiques towards the leniency programs, but to summarize some of the most recent ones. The focus will be mainly on the US and EU leniency programs, as they have been the main source of inspiration for the ICN best practices.

In the US, Daniel Sokol,¹⁷⁵ based upon earlier studies of Nathan Miller,¹⁷⁶ has argued that the leniency program may not be used properly. Before critiquing the use of the leniency program in the US, Sokol evaluates the benefits of having a leniency program. Leniency programs have been praised for their ability to incentivize cartel members to defect. In the case of the US, this incentive constitutes the escape of criminal conviction and treble damages. The condition for receiving this lenient treatment is, of course, the provision of full cooperation with the DOJ in its cartel enforcement.¹⁷⁷ The US further tries to improve its enforcement by offering an Amnesty Plus program.¹⁷⁸ This program allows for firms, already convicted for their role in a different cartel to benefit from a sentencing discount for revealing its role in a still undetected cartel. Of course, the lenient treatment will apply to this newly revealed cartel.

Sokol suggests that the success of the leniency program may be deducted from the increase in cartel fines, the number of people who spent time in jail and the number of days people have spent time in jail.¹⁷⁹ Nevertheless, there is still skepticism whether the existing US cartel regime offers optimal deterrence. Even with the leniency program in place, the formation of new cartels is not necessarily prevented or the stability in existing cartels is not necessarily negatively affected.

The sub-optimality of the US leniency program is based upon the idea that too generous leniency programs may create strategic behavior. This strategic behavior is to provide the enforcement authorities with information on behavior that is not a clear cut violation of the competition law. An empirical survey conducted by Sokol suggests that competitors often are reluctant to defend themselves against these cartel charges.¹⁸⁰ Rather than taking the risk of a fully litigated trial, Sokol indicates, that these competitors often opt for a settlement. In other words, firms are using the leniency program to punish other competitors.¹⁸¹ Problematic in this scheme is the DOJ’s willingness to settle on behavior that is not a clear cut violation.¹⁸²

A further critique derived from a survey that Sokol has held among practitioners, is the transparency and legal certainty surrounding the procedure of the leniency application.¹⁸³ A lack of transparency will increase the risk of dealing with the DOJ and consequently lead to the continuation of the cartel. Besides a reference to the placing of the marker, it is not immediately clear which of the provisions need more transparency. Other elements that contributed to a risk of coming forward, and that applies to individuals, is the distorted trust towards in-house counsels after Stolt-Nielsen.¹⁸⁴

The leniency applications in the US will also be affected by follow-up prosecutions in other jurisdictions. The complex web of leniency applications that a firm may have to go through in other jurisdictions may create additional uncertainty. It is not for sure that leniency can be obtained in all the other jurisdiction or that similar conditions or rules apply in all these different jurisdictions. One of the elements that has been identified as problematic in another jurisdiction is the absence of an attorney-client privilege in the Europe for in-house lawyers.¹⁸⁵

Cross-border cases enable a link with the critique towards the European leniency program. In a recent case, the Pfleiderer case, it was made clear that the content of a leniency application is not protected against a discovery procedure that could be started in the US.¹⁸⁶ The discovery procedure, which would be requested in the framework of a private damages claim, will make firms cautious of applying for leniency in Europe if they haven’t applied in the US.

The waiting game that may result from the potential discovery procedure is not the only one operating under the European leniency program. Studies done by Marie Goppelsroeder, Maarten Pieter Schinkel, and Jan Tuinstra on the one hand,¹⁸⁷ and Dennis Gärtner and Jun Zhou on the other hand,¹⁸⁸ hold that there may be no race at all under the European leniency program. More in specific, with econometric tests, these scholars reveal that a majority of the leniency applications are done when the cartel has already collapsed or is on the verge of dying. It has been identified by these studies that it is a phenomenon running over nearly two decades, meaning that it is not specific to one of the different formats that the European leniency program has known. This kind of critique has followed the much earlier detected form of delay, which existed in the fact that the European leniency program mainly triggered leniency application following an application in the US.¹⁸⁹ Margaret Bloom, however, does not agree with the conclusion that these follow-on applications show the weakness of the European leniency program, but rather reveal the more effectiveness of criminal sanctions.¹⁹⁰

Nicolo Zingales, theorizing about the possibility that a leniency program could render sanctions ineffective, formulates another critique. The idea behind his reasoning rests on the generosity of leniency programs. If a leniency program offers lenient treatment well beyond the first applicant, there is an inherent danger that the total amount of sanctions decreases, with a possible stabilizing effect on cartels.¹⁹¹ Due to lower sanctions, firms will not be as deterred anymore as they used to be. Even though Zingales does not offer any kind of data on the effect of the leniency program on the amount of fines, he opines that it would be better for the EU to follow the American example by limiting the leniency to much less applicants.¹⁹² Some years earlier to Zingales study, Cento Veljanovski has conducted a survey on the effect of the leniency program on the average fine for a cartel.¹⁹³ His survey revealed that there was a substantial reduction in the average overall fine and the average fine for a firm.¹⁹⁴ This negative effect on deterrence is even further aggravated by appeal judgments.¹⁹⁵

6 6 The Leniency Good Practices as the Benchmark for Convergence

6.1 6.1 Critique to the Source of the Best Practices Affecting Its Legitimacy

The ICN aims at norm setting or public policy making. Even though the ICN would like to steer the behavior or determine the freedom of the ICN’s members, it is not operating within the framework of what is the usual practice for international law making. The ICN is a forum, which is not created by an international treaty, allowing the competition agencies to engage with its foreign counterparts. There is no reliance anymore on the head of state or the foreign ministry to lead the negotiations.¹⁹⁶ The contacts with the foreign counterparts happen now directly, even though these agencies are not allowed to bind the State as understood under Article 7 of the Vienna Convention.¹⁹⁷ The ICN does further not aim at creating a formal treaty or any other kind of traditional source of international law. The ICN operates by formulating guidelines that, in the best case, take the format of best practices. For this purpose, the ICN, and thus the enforcement agencies, is assisted by private actors and other international organizations.

Operating outside the framework of what is the usual practice for international law making has made the transnational regulatory networks open for criticism. The aspiration of the ICN to steer the behavior of its members raises the question whether this network should be eligible to exert this kind of power over its member agencies. Indirectly, this will affect the public in general. However, unlike in a parliamentary system, the general public cannot give its approval or disapproval over the policies pursued by the ICN. Any control, through voting for example, is being denied to the general public. Furthermore, as these networks operate transnationally, there is also a problem of a global general public, which most likely does not exist as of today. This has been the basis for a legitimacy critique.¹⁹⁸

To formulate an answer to the growing legitimacy problematique of these transnational regulatory networks, the concept of legitimacy has been widely debated and rethought. It is not the purpose to “get bogged down in the particularities of the debates”.¹⁹⁹ As Chris Brummer states, it is sufficient to concentrate on the dominant approach that splits legitimacy into two categories: input and output legitimacy.²⁰⁰ Input legitimacy is concerned with the involvement of the governed.²⁰¹ Output legitimacy deals with the quality of the rules and whether they are any effective.²⁰² Organizational qualities are thus not the only elements that can contribute to legitimacy, but also the organization’s accomplishments. Legitimacy can thus also be derived from the ability to solve problems.²⁰³

The need for the ICN to rely on output legitimacy has been pointed out by several scholars.²⁰⁴ Even though the ICN is a non-exclusionary organization, allowing each enforcement agency to become member,²⁰⁵ and gives the opportunity to each member to fully participate,²⁰⁶ it cannot be denied that mature enforcement agencies are the main driver behind of the ICN’s activities and policy setting.²⁰⁷ This inequality cannot be balanced by the presence of non-governmental advisors, because their role is somehow controlled. NGAs can participate in the Annual Conference. However, the Steering Group and the host agency of the Conference can control the participation of the NGAs in terms of geographical origin or background.²⁰⁸ This possibility has been included to prevent that the Conference would be captured by a certain interest group.²⁰⁹ The Steering Group has also power to rely on NGAs for other purposes than the Conference. NGAs can be consulted for a particular or potential project, for issues to be considered in a Working Group or by the Steering Group, or for assisting in the drafting of work products of the Working Groups.²¹⁰ ICN member agencies are free to consult with NGAs at their own discretion and this to seek information or expertise. Calvin Goldman, Robert Kwinter and Navin Joneja stipulate that NGA input is encouraged.²¹¹ However, the member agencies are the main driving force behind the ICN objectives and work products. NGAs do not only come from North-America, Europe or Japan, but also from jurisdictions with newly established agencies.²¹² Structures have been built into the ICN operational framework not to prioritize any of the NGAs or to let one particular interest be overrepresented.²¹³ Nevertheless, as Fox warns, the defense bar and the industry associations are predominantly represented.

Due to the deficit of input legitimacy, the ICN should not yield to cease its existence. The ICN could have moral authority because it is a forum where expert knowledge is gathered. The knowledge on how to solve competition law disputes is a common good for global affairs.²¹⁴ However, in order not to lose its authority, the ICN should show that it has the ability to solve problems and that it is not captured or manipulated by interest groups, private or public, when solving these problems. In other words, the ICN may be vulnerable by threats to its reputation.²¹⁵ In order to prevent the loss of moral authority, the quality of the norms, standards and procedures suggested should be a fact.

The ICN can only guarantee that the suggested norms, standards and procedures are effective when it is monitoring the outcome of the convergence process. Monitoring the convergence process entails the possibility of also questioning the best practices. Thus, just like Stucke understands, the ICN’s best practices should not be seen as the creation of a fixed end product.²¹⁶ The adoption of best practices by the members does happen voluntarily and divergence may occur in this process. The divergence may be inspired by the different legal structure of the adopting jurisdiction,²¹⁷ a different economic structure of the adopting jurisdiction,²¹⁸ or simply the idea that it will function better in a different way.²¹⁹ Evaluation has to make this clear. However, the evaluation needs to go further. It may well be that there is no divergence from the best practice, but that the leniency program has not the desired outcome.²²⁰ Also in this kind of cases, it is necessary to evaluate the best practices and correct shortcomings. These could exist in gaps, unexpected shifts, or wrong predictions.

The process of convergence is a constant evolutionary process, forcing the ICN to periodically revise its best practices to reflect the continuous experimentation. Continuous experimentation should then also enable the ICN to reflect what the best practices are taking the specificities of the members, such as, among others, the size of the economy or the stage of development, into consideration.

6.2 6.2 Evaluating the Benchmark

Svetiev has also acknowledged the need for constant monitoring of the best practices.²²¹ When developing his framework against which he tries to legitimacy the operation of the ICN, Svetiev heavily relies on benchmarking in a corporate environment. In a corporate environment, Svetiev argues, the benchmarking will seldom be about just copying.²²² Several restraints, such as IP rights or business sensitive information, will not allow for creating a model benchmark identical to what is being applied in a certain firm. The absence of these restraints in a regulatory context and the often willingness of jurisdictions to voluntarily offer their rules and institutions as model creates a risk of ending up with sub-optimal equilibriums due to the just following what has been identified as the best practice. The end result may be sameness, with as a consequence that the adopted best practice is not necessarily optimal for the jurisdiction implementing the best practice. Even though these concerns exist, Svetiev indicates that even a mere transplantation of rules seldom leads to identical outcomes in the receiving jurisdiction.²²³

In spite of this conclusion, Svetiev indicates that the best practice selection in the ICN may in fact lead to sub-optimal practices or to practices that are not suitable for the entire membership.²²⁴ One of the element identified as the cause is the minoritarian bias in the agenda setting of the ICN and its working groups.²²⁵ Financially well-resourced enforcement agencies, which are often the more mature enforcement agencies, have more possibilities to influence the agenda setting. Once issues are on the agenda, other problems arise. Less mature enforcement agencies may not have enough experience to discuss and evaluate the proposals.²²⁶ Very often there even exists already a consensus over the issues that have been put on the agenda, so that no profound discussion takes place.²²⁷ In some cases, the issues on the agenda are salient to the majority of the jurisdictions that they even do not participate in the discussion.²²⁸

Svetiev does not seem to find the minoritarian bias entirely problematic. The ICN is not conceptualized to look for a well suited solution towards a specific problem, but for experiences, substantive or procedural, that work in other jurisdictions.²²⁹ If such experiences are identified, enforcement agencies looking for solutions should transform these experiences to suit the task and the local environment in which they are supposed to operate. The ICN best practice benchmarking is thus not going to lead to ‘disruptive innovation in antitrust implementation strategies’²³⁰ or ‘broad dissemination, and improvement, of practices actually useful to the participating authorities in their day-today work’.²³¹

Even though the ICN is not seeking for innovations to the enforcement strategies or substantive law, the outcome of the search for best practices may still be affected by the minoritarian bias. The best practice benchmarking may be constrained by framing the problem in a certain direction or by offering an already existing consensus. This existing consensus may, for example be built upon wrong premises or even blunt errors. This kind of problems can only be overcome if a structure of revision for these best practices is implemented. Revision possibilities would render best practices provisional, ‘in the sense that they can be revised in the light of data from reports about implementation and the outcomes achieved’.²³² When writing his paper, revision procedures were not in place in the ICN structure. This has not changed yet. However, the ICN has instigated that it would revise its best practices during the second decade of its existence.²³³

Svetiev argues that the ICN should formalize some form of processes of ‘gathering and distributing knowledge about how those best practices are implemented’.²³⁴ In other words, best practices should be only a first step in formalizing the interactions in a network. This control on implementation should not be about counting the number of jurisdictions that have followed the best practices, but rather of what the outcome is ‘from either following or diverging from the endorsed practice’.²³⁵ To effectuate the evaluation model, the ICN has to introduce a duty to report. This does not mean that the non-binding nature of the best practices should be affected. It can still exist next to duty to report.²³⁶

7 7 Evaluating Leniency Best Practices

7.1 7.1 An Argument for Non-Governmental Advisors as Evaluators

Enforcement agencies have a tendency to overstate the success of their leniency programs. The general attitude of the enforcement agencies is to emphasize one aspect of the leniency program to reiterate its success. The US DOJ refers to the ever-increasing level of the total amount of fines.²³⁷ In the US, where there are custodial sentences, there has also been a reference to the significant increase in the total number of days spent in jail.²³⁸ The European Commission admits that most of the cartels that have been detected by the European Commission are detected after one cartel member has confessed and asked for leniency.²³⁹ In making this statement, there is an automatic reduction of the measure of the success of a leniency program to the number of decisions.²⁴⁰

However, the data these officials rely on could be misleading. The data they look at to make these statements are usually the data numbering the total amount of applications the total amount of decisions taken or the increase of the total amount of fines. This data does not reveal what is driving the leniency program. It is not possible to detect whether the applications for leniency are sincere applications or part of a strategy. The data also does not reveal whether leniency is mainly driven by foreign follow-on actions or not. Displaying only the scale of the fines does not reveal a possible change in the method of calculating the fine.

The tendency to overstate the success of the leniency program and downplay any kind of criticism has been explained by public choice theory.²⁴¹ The DOJ is in need of budget for its antitrust division. Being able to present improvements in detection due to the leniency program and, subsequently, a high rate of settlements of these detected cases, the DOJ is able to paint a positive picture of its activities. The positive picture will enable the DOJ to justify its budget, even if in other parts of the enforcement there may be declines noticeable. Sokol evokes the idea that ‘[t]o suggest that the “golden child,” as one practitioner described the leniency program, makes the DOJ less worthy of political and financial support’.²⁴² The DOJ has, still according to empirical research of Sokol, ‘shown an unwillingness to reexamine the leniency program and responds overwhelmingly negatively to any criticism of the program’.²⁴³

Having conducted interviews with both officials from the JFTC and the EU Commission’s DG Comp, the current author cannot but agree with the finding that enforcement authorities tend to positively evaluate their leniency program and see no point in formulating any critique towards their leniency program.²⁴⁴ Whether public choice reasoning is really driving their enthusiasm is not immediately obvious. It is for sure though that the increased number of decisions or the smoothness with which these decision can be reached are part of the explanation to view the leniency program as a necessity.

Against the backdrop that enforcement agencies tend to mask flaws in their leniency programs and have reasons to do so, an argument could be made that these enforcement agencies would not be the best placed to evaluate best practice. Indeed, questioning the best practices would inherently implicate that the leniency programs that are the basis for these best practices are also flawed. This does not downplay the role of the enforcement agencies. They hold the key to lots of information, which is necessary for the evaluation of any kind of best practice.

The information that is held by the enforcement agencies may not be enough to come to profound conclusions. Part of the implementation of the legislation that is based upon the ICN’s best practices is done by lawyers and other legal advisors. These actors are, in the framework of the ICN, categorized as non-governmental actors. The extensive survey that Sokol has held among lawyers in the US is an example of how lawyers hold information regarding competition law and its enforcement.²⁴⁵ The kind of information these lawyers have provided to Sokol is something that would most likely not become apparent by just analyzing data from enforcement agencies. Dave Anderson, a practicing competition lawyer and non-governmental advisor to the ICN, seems to agree with this analysis. In his prospect for the future, Anderson sees the role of lawyers as, among others, to keep the best practices relevant and alive.²⁴⁶ At the end, lawyers have a “behind-the-scene view”²⁴⁷ on the facts that lead to the operation of best practice inspired legislation.

Enforcement agencies could function as an agent for information disclosure on the application of the best practice. Reports will unavoidably lead to another paradox of plenty, which has to be edited down to useful conclusions. The lack of a formal secretariat inside the ICN, would make the enforcement agencies responsible for handling this information. The problem is that, unlike with the generation of best practices, these reports have to be screened not for commonalities but for eventual disruptions in the application of the legislations reflecting the best practices. The occupational priorities of the enforcement agencies would probably mean that they are not best placed to engage in this activity. Furthermore, the data may display the need for new trends in the best practices, eventually requiring engaging in innovative legislative work. Among the non-governmental advisors, academics or research institutes would be best placed perform this task. This kind of suggestion is not completely new. It has been recognized during the ICN Conference in Seoul that NGAs could be attributed a greater role.²⁴⁸ Certain substantive areas allow for more input from the NGA. In specific, the criteria for leniency and amnesty were mentioned as an area in which NGA could contribute.

7.2 7.2 Evaluating the Implementation and Not the Form

There have been suggestions as how to evaluate the best practices of the ICN. Some propose to compare the legislative instruments with the best practices and see to what extent the best practices are reflected in the legislative instrument.²⁴⁹ This kind of exercise provides an overview of the degree of convergence in form. However, it is not possible to deduct from this exercise whether either the convergence or divergence has led to results that are better or worse than what legal instruments after which the best practices were modeled, are generating.²⁵⁰

In a different area than the leniency program, Maria Coppola and Cynthia Lagdameo looked into the extent to which the ICN members implemented the best practices for merger notification and review.²⁵¹ Their study starts from a table representing the main best practices and mapping out which country’s legislation is in conformity with these best practices. The table clearly shows that there is a huge discrepancy between the best practice and the current legislative landscape. This leads to the conclusion that either the best practices are not universal or that considerable work needs to be done. Based upon the overwhelming support for the best practices, which is reflected in the tiny number of changes that do not conform the best practices and the absence of any regime compliant with the best practices that changed to a non-compliant regime, the authors opt for the latter conclusion.²⁵²

It needs to be stipulated that before reaching this conclusion, Coppola and Lagdameo have highlighted some barriers to implementation.²⁵³ Several of those barriers, such the lack of resources, unclear formulation of the best practices and the non-suitability of the best practices, may imply that some reflection on the best practices is needed as well. It is not sufficient to only look at the degree of convergence; the reasons for divergence have to be revealed as well. Only then proper conclusions can be drawn.

Other scholars have argued along the lines of Coppola and Lagdameo for evaluating the implementation of the best practices. For the leniency program, Kirtikumar Mehta and Ewoud Sakkers stipulate, for example, that the global promulgation of leniency programs based on the standards developed and promoted by the ICN is one of the better examples of how the ICN has contributed to convergence of law.²⁵⁴ The convergence in form may well be a fact; it is, nevertheless, not possible to deny the critique on the operation of several of the leniency programs.

In order to evaluate the current best practices on leniency, the ICN could design questionnaires to be sent to the enforcement agencies. Detailed information could be collected on various issues related to the leniency program, besides the already available data on the number of leniency applications and decisions following these applications.

In order to evaluate the operation of a leniency program, data could be collected on the following issues. In order to see whether there is a race among the cartel participants, more detailed information is necessary on the timing of application. In other words, data needs to be collected on whether the applicant applied before or after the down-raid. In systems that allow for more than one firm to apply before the down-raid, it may be worthwhile to see how many firms did in fact apply in the pre-investigation stage.

Deterrence could be deducted from various other variables. Information on the duration of the cartel could be an element contributing to assess the deterrent effect of the leniency program. Deterrence could also be evaluated by collecting data on whether the cartel is still operational at the time of the leniency application. The report could also make a distinction between leniency applications that have an endogenous cause. Having many of this kind of applications suggests could also suggest the lack of deterrent effect.

More difficult to assess are the conditions attached to a leniency program. Data could be collected on how many firms succeed in maintaining their position in the queue. Other data that could is the speed with which the leniency application can proceed to a formal decision. Since continuous cooperation is usually required, how many times firms have to be contacted before an investigation can lead to a decision.

Leniency programs may trigger a huge number of applications. Not all of these applications necessarily lead to a decision. In order to make an evaluation of the leniency program, it may be worthwhile to have data on why there is a discrepancy between the two. It could indicate a case overload due to an overly easy conceptualization of the leniency program. Nevertheless, it would also allow for making an argument that the leniency program is an invitation for submitting applications on behavior that is not a clear-cut violation of competition law. Connected to this, data could be collected on whether the leniency applications lead to a decision of the enforcement agency or to a settlement.

The enforcement agencies could further submit data the kind of cartels that are being revealed by the leniency program. This kind of data could be supplemented by data on whether there is a decline or increase in the detection of certain types of cartels. Having this data would enable to draw conclusions on whether a more diversified set of enforcement tools is necessary or not.

The reasons for applying for leniency may be more difficult to deduce from the data available to the enforcement agencies. Nevertheless, the enforcement agencies will be able to indicate whether firms have been under investigation in more than one cartel and whether these investigations link with each other through leniency applications. This is probably the extent to which the enforcement agencies can assist in detecting the reasons on why firms come defect a cartel. Due to the limited capacity of the agencies in this respect, the information provided to the ICN could be supplemented by information that is available to among lawyers. This information could be collected through academics, as Sokol did in his empirical research on the functioning of competition law, or by lawyers themselves. Both can contribute this information in their capacity of non-governmental advisors.

7.3 7.3 Evaluation to Fill Gaps

The leniency program may have to be evaluated for what it is. However, the critiques formulated towards the operation of a leniency program are not always a direct critique to the formulated best practices. The negative assessment of the US leniency program is not directed at the way how the leniency program is conceptualized. The critique is rather oriented at the operational link between the leniency program and settlements. It only requires a brief look at the best practices to know that the Cartel Working Group has not provided any information on this issue.

The interaction of the leniency program with other sanctioning systems has been almost left untouched. The best practices formulate guidance on how a leniency program should be conceptualized in a bifurcated enforcement model or on how leniency will work in a parallel civil and criminal model.²⁵⁵ In the former model, the best practice advocates a consistent leniency policy between the different authorities investigating and prosecuting cartels.²⁵⁶ The latter model requires a clear articulation of the application policy in order to avoid unpredictability and uncertainty.²⁵⁷

The interaction of the leniency program with other sanctions or enforcement tools is just one example of issues that are not covered in the best practices. Another element that is not elaborated in detail is the need to have post-investigation leniency program, even though many of the leniency programs have been experimenting with this kind of leniency as well. Much more guidance may be required in this respect. At the end, it has been pointed out that too generous leniency programs can either make the leniency program ineffective in terms of deterrence or create a waiting game in order to come forward with information.

The best practice also only focuses on a single format of a leniency program. There has been no attention paid to a distinction between the level of development of the competition law or its enforcement agency. A leniency program may have a serious impact on the enforcement agency. Alan Riley commented that the European leniency program has serious implications for the European enforcement policies. Riley summarizes this issue by stating that:

DG Competition is now in many ways the victim of its own success; leniency applicants are flowing through the door of its Rue Joseph II offices, and as a result the small Cartel Directorate is overwhelmed with work. … It is open to question whether a Cartel Directorate consisting of only approximately 60 staff is really sufficient for the Commission to tackle the 50 cartels now on its books.²⁵⁸

Therefore, the best practices could offer guidelines on what the impact may be of the conceptualization of a leniency program in a certain way. The ICN best practices could give guidelines in one way or another. Another option would be to provide guidelines on how to prioritize cases.

Other gaps than the ones described above may be uncovered by analyzing the operation of a diverse set of leniency programs. In order to increase its appeal, the ICN should try to close these gaps or even provide a more diverse set of best practices. In doing so, the ICN could increase its acceptance among its members, and so eventually to greater convergence of competition law.

8 8 Conclusion

National competition law enforcement agencies have organized themselves in an informal framework to elaborate common principles on competition law and its enforcement, which they would like to use to influence each other’s behavior. In doing so, they created a transnational regulatory network, which they named the ICN. Within the framework of the ICN, information on the operation of various competition law practices is being gathered and analyzed in order to formulate best practices. These best practices should inspire the ICN members to voluntarily adjust their substantive law or their enforcement practices. Convergence is thus the ultimate goal.

Convergence only can occur when an agreement can be reached that a certain practice is better than the others. Hence, it is argued that convergence always presupposes a period of experimentation. However, even if there has been a period of experimentation followed by a period of convergence to an agreed standard, it may turn out that the standard to which was converged is not a proper standard. To avoid such a situation and to increase the legitimacy of the transnational network as a regulator, it has been argued that a revision of these standards should be possible.

By relying on the ICN’s best practice of the leniency program, it has been argued that the standards are arrived from decades of experience in the US and the EU. This experience is, however, subject to critique. Indirectly, this critique will also affect the best practices. In order to prevent that such critique would jeopardize the best practices, a revision system should be set up. It has been further argued that this revision system should be impartial, due to which a certain category of NGAs would be better placed to perform this evaluating exercise. Another element of the argument has been that the evaluation should not be restricted to what currently exists as a best practice, but that also gaps and shortcomings should be identified. Only if this kind of system is established could the ICN keep its legitimacy as an effective regulator.