Emmanuel Combe, Constance Monnier-Schlumberger, Lutte contre les cartels et comportement des managers, novembre 2016, Concurrences Nº 4-2016, Art. N° 81860, pp. 51-68

This article was originally published in French. A translation in English is available below.

English Translation of the original article:

* [This article reflects the personal opinions of the author and does not commit the institutions to which he belongs.]

Introduction

1. Developed following the pioneering work of Gary Becker (1968), the economics of crime has mainly been mobilized in antitrust to analyze a firm’s decision to engage in a cartel: according to this approach, a firm is induced to form a cartel when the illicit gain resulting from this practice exceeds its expected cost, which depends both on the anticipated sanction and the probability of detection/conviction (see e.g. Connor and Lande [2012] for a synthesis of the literature).

This approach, which focuses on legal persons (the company being considered as a single, global entity), however, leaves aside the question of the specific incentives for "managers" [1] to engage in collusive practices, incentives which may not be aligned with those of shareholders or, conversely, may reinforce the latter’s interest in violating antitrust rules.

2. In this respect, the corporate governance approach, which highlights the existence of a possible divergence of interest between shareholders and managers, is proving particularly fruitful. According to this approach, the shareholders’ main objective is profit maximisation, which is reflected in the company’s stock market valuation in the long term; on the manager’s side, the objective may be relatively different: receive a bonus at the end of the year or quickly increase the share price, if he or she holds stock options or is granted free shares; benefit from internal or external promotion, with a focus on short-term performance; have increased power and prestige, by increasing the size of the company rather than its profitability; guard against the risk of bankruptcy, by diversifying the company’s activities through external growth, etc. Finally, other individual factors (risk appetite or aversion, psychological aspects or ethical behaviour) may influence an individual’s choice of whether or not to engage in a cartel practice. The existence of an agency relationship between shareholders and managers enables the latter to use the asymmetry of information they enjoy to pursue their personal objectives, especially if they manage to satisfy a minimum profit or steady earnings growth constraint, which shields them from too close scrutiny by shareholders.

3. The "corporate governance" approach has been applied to collusive practices since the 2000s (for a synthesis, see Buccirossi and Spagnolo [2008]), feeding a theoretical and empirical literature (see e.g. Alawi [2014]) with strong implications for public policies. Two fundamental questions are indeed raised: first, to what extent should managers be held liable for antitrust violations, beyond the sole responsibility of shareholders? Second, to what extent does the personal incentive for managers to engage in a cartel justify the intervention of public authorities, beyond the internal rules and incentives of the company?

4. As regards the incentives for managers and shareholders to engage in collusive practices, four different scenarios can be distinguished (Tables 1 and 2).

Table 1. Convergence/divergence of interests between shareholders and managers is available in the pdf version of this article.

Table 2. Penalties and incentives to comply with antitrust rules is available in the pdf version of this article.

5. A first case is one in which managers and shareholders have a common interest in complying with antitrust rules (case 1 in the tables). This scenario occurs when :

– sanctions against companies prove to be dissuasive - the expected cost exceeds the anticipated illegal gain - leading shareholders to demand strict compliance with competition rules by managers ;

– managers have no individual incentive to form a cartel, in particular because they fear internal or external sanctions (such as criminal prosecution), have a strong aversion to risk or a strong preference for ethical behaviour, or for any other reason (psychological aspects).

6. In a second scenario (case 2 in the tables), shareholders have an interest in managers engaging in anti-competitive practices even though they are reluctant to do so. This case can be envisaged if two conditions are cumulatively met:

– the penalties imposed on undertakings are insufficiently dissuasive, allowing the undertaking to make a net gain after the penalty ;

– managers risk strong individual sanctions, for example in the form of criminal prosecution, are highly risk-averse or place great importance on ethics in business life.

In this conflicting case, the manager may not form a cartel, or may engage in this practice against his or her will. The latter configuration is made possible by the fact that the manager maintains, through his employment contract, a hierarchical relationship with his employer: not implementing a practice, even if illicit, at the request of his superiors may constitute a breach of his professional obligations. In addition to the threat of dismissal or marginalisation within the company, shareholders can also mobilise positive incentives, for example by promising managers who engage in a cartel practice ex post compensation if they are prosecuted.

7. A third case is where managers have an interest in violating antitrust rules in their own personal interest, even though shareholders are opposed to it (case 3 in the tables). This scenario occurs when antitrust sanctions prove to be dissuasive (zero or even negative net gain) for companies, while managers incur low internal or external risks, compared to the gains resulting from a cartel practice, have little "risk aversion" or a low preference for ethics. It could be objected that this scenario is purely theoretical, insofar as shareholders can always align the behaviour of managers with their own interests, whether through positive incentives (stock options, for example) or negative ones (threat of dismissal). However, as we will see, shareholders have imperfect control over managers, given the asymmetries of information: managers can in some cases conceal the cartel practice from shareholders, arguing for example that outperformance comes from greater operational efficiency. Similarly, positive incentives are not always sufficient to discipline managers, since their motivation is not limited to remuneration alone, but also includes objectives such as, for example, rapid career progression.

8. Moreover, it is tempting for a company prosecuted by the antitrust authorities to invoke the isolated and concealed action of an employee in order to escape liability as a legal entity. For example, in France, in the hygiene and cleaning [2]products case, L’Oréal argued that the participation of two executives of the group in cartel meetings was a matter of personal initiative, of which their superiors were unaware, and that one of the two persons had deliberately concealed his actions. However, this argument was not upheld at the stage of the Authority’s decision.

9. Indeed, it should be remembered that such a defence strategy is unlikely to prosper. According to settled case-law, a company cannot exonerate itself from its own liability on the ground that the practices were carried out by managers, in isolation and without the knowledge of their employer, even if it was against the very interest of the company, as stated in the said decision (pt 1109): "(...) an executive of a company who takes part in a meeting concerning the activities of the undertaking which employs him is deemed to represent or commit that undertaking. In order to establish an undertaking’s participation in a cartel concluded at secret meetings, it is not necessary to determine whether the person who represented it at those meetings held a particular mandate or authority to that effect, but only to establish, on the basis of a sufficiently serious, precise and consistent body of evidence, that the undertaking was actually present at those meetings. ”

10. In Union law, the judgment of the Court of First Instance in the Parker [3] case follows the same approach. In that case, the applicants relied on the fact that the practice had been implemented by the director of the ’Oil & Gas’ unit of ITR Rubber, without the knowledge of his employer and in contradiction with the internal policy of the company; moreover, the director of the unit, who became an external consultant after 2006, acted in his personal interest and in that of his own companies. As such, according to the defence, the practice should not have been imputed to the company since the company itself was the victim. The Court of First Instance dismissed the plea, pointing out that the imputation to an undertaking of an infringement of Article 101 TFEU does not ’presuppose an action or even knowledge on the part of the members or principal managers of the undertaking concerned of that infringement, but the action of a person who is authorised to act on behalf of the undertaking’ [4]. It also notes that Parker never filed a complaint or took any action against its former employee.

11. The fact that a company cannot escape its own responsibility as a legal entity by invoking the alleged isolated and concealed behaviour of a manager does not mean, however, that it is easy to control the behaviour of employees internally, given the size of the company, its mode of governance and the asymmetry of information between shareholders and managers. In this respect, and as we will see in the second part, the implementation of public policies, by their very nature external to the company, can help to resolve any difficulties shareholders may have in controlling managers: in particular, the possibility of credible criminal proceedings against individuals who have taken part in a cartel is likely to "discipline" managers in the sense of respecting the law.

12. A fourth and final case is one in which the interests of shareholders and managers converge to infringe competition rules, although the motivations of each of these actors are different (case 4 in the tables). This situation occurs when two conditions are cumulatively met:

– the shareholders consider that the anticipated sanctions are not sufficiently dissuasive (the expected net gain from the infringement remains positive);

– managers do not incur any credible personal risk, external or internal to the company, when they break the rules of the game, while the cartel allows them to increase their own profits (remuneration, career advancement, etc.); and/or, these managers have specific individual characteristics that encourage them to engage in these practices (psychological profile, risk appetite, attitude towards ethics, etc.). Moreover, it is unlikely that the risk to managers is internal to the company, insofar as the shareholders, benefiting from the offence in the form of overprofit, have no incentive to "punish" them.

13. The latter case is particularly interesting: the alignment of interests between managers and shareholders in the sense of non-compliance with the law creates a double incentive to form a cartel, each party having its own determinants. The aim of this article is precisely to explain the main parameters that influence the incentive for a manager to engage in a cartel (I.), before analysing how shareholders (internal solutions) and public policy (external solutions) can help to dissuade him from doing so (II.).

I. The manager facing the cartel: what arbitration?

14. The literature on the "optimal sanction" assumes that the agent likely to violate antitrust laws balances - if not precisely calculates - the expected costs and benefits of a violation in order to make his choice. This reasoning can be applied to the behaviour of the manager, provided that behavioural factors are included and that the case in which the manager would act under the sole constraint of the shareholders is excluded.

1. The decision to enter a cartel

1.1 The individual costs of cartel participation and their reduction

15. Three main types of costs will be taken into account by the individual when deciding whether or not to participate in a cartel.

16. The existence of criminal sanctions against natural persons is the first and main parameter. To assess the extent of these sanctions, the manager may refer to statistics on past cases. For example, in the American case, 235 individuals were sentenced to personal fines over the period 2006-2015 for a total amount of 34 million dollars, i.e. an average sanction per person of 146,000 dollars. In the case of firm prison sentences, the assessment of the penalty requires first converting the days of imprisonment into monetary equivalent: using the conversion method of Connor and Lande (2012), one year of imprisonment would be equivalent to an average of $1.5 million for a company director. Given that in the United States, the average length of imprisonment in the period 2006-2015 was 677 days, i.e. almost 2 years in prison, the average cost of prison for a manager sanctioned for participating in a cartel in the United States can be estimated at $2.8 million.

17. It should be noted that it is not so much the existence of criminal sanctions that is taken into account by the individual as their actual application: the fact that a cartel is detected and condemned does not necessarily imply that the manager is prosecuted in a personal capacity, even when such legal provisions exist. For example, in the United States, the threat of prison was long theoretical and it was not until the 1990s that a radical change of approach was made by the judicial authorities: today, every major cartel case gives rise not only to sanctions against legal persons, but also against managers. Conversely, in Europe, even when specific criminal provisions are provided for - as is the case in France [5] - their effectiveness remains weak, if not nil. In this case, it is unlikely that the parameter of criminal sanctions will be included in the individual’s cost-benefit calculation.

18. A second parameter taken into account by the individual is the degree of stigmatisation of cartel practices within the society in which he or she evolves. Indeed, beyond the legal framework, socio-cultural norms defining what is "licit" and "illicit" can play an important role in the individual’s behaviour: for the cartelist candidate, it is more difficult to assume a practice that is socially stigmatised as "immoral", especially if it enjoys high media visibility. While this cost is difficult to monetize, it is nonetheless an essential determinant of whether or not to break the law. More generally, breaking the law is likely to have a psychological cost for the individual, all the more so if he or she has a developed ethical or moral sense. We will return to these issues later.

19. A third parameter taken into account by the potential cartelist is the degree of adherence of his company to the competition rules. The presence within the company of a compliance programme is a signal in this respect; but to be truly perceived as credible by employees, the programme must go beyond a mere formal framework and be based on a firm commitment by senior management, accompanied by strong sanctions such as dismissal. For example, Bergman and Sokol (2015), in their study of the air freight cartel, show that the compliance programmes implemented at Lufthansa were operational to the extent that they were taken to the highest level by the group’s management. Once again, although this parameter is difficult to convert into a monetary equivalent, it is nonetheless significant in the decision of the individual, whose behaviour would, if necessary, run counter to the company’s compliance policy.

20. These different parameters determine the overall cost that the manager will have to bear if he breaks the law. In this respect, the individual’s perception of these various costs can be altered by the presence of "psychological biases", which can lead him to underestimate their magnitude.

21. Firstly, if the unlawful practice is carried out in consultation with the manager’s hierarchy, the latter will try to convince him of the merits of his action. Several experiments in experimental psychology show that the collective and hierarchical context of decision-making can favourably modify the individual’s incentive to engage in an illicit or immoral practice. In particular, Milgram’s (1963) famous experiment on obedience behaviour can be transposed to cartel situations (Stucke [2011]):

– the wrongfulness of a practice may be diminished if breaking the law is presented as a "necessary evil" to achieve a given result. In the case of a cartel, the individual may be convinced by his or her superiors that maintaining a high level of profitability requires recourse to a collusive practice. For example, Bryan Allison [6], a participant in the marine tubes cartel, states that the need to meet profit targets was a key factor in his decision to participate in the cartel (O’Kane [2011]). Similarly, in the chemical commodities case, the cartel was presented to the participants as a necessity in order to be able to invest, as reported by the former chairman of Solvadis: "(...) The substance of the debate was that the western region was stricken and therefore the activity required significant investments, and that a solution had to be found to allow margins to increase and to encourage these investments [7]";

– the seriousness of unlawful conduct can be blurred, provided it is presented in a positive light. In the case of cartels, there is even sometimes a reversal of values, as in the case of the lysine [8] cartel’s motto: "(...) our competitors are our friends; customers are our enemies". Similarly, the cartel may be justified in the eyes of the participants by the need to avoid too strong competition between companies, which would harm product quality or employment, or to re-establish a counterweight in commercial negotiations. In France, for example, one of the participants (a sales manager at SC Johnson) in the hygiene and cleaning products cartel stated that the cartel members were trying to "organise themselves to resist the demands of the large retailers ... because these were clearly excessive and unbearable for the suppliers" [9]. In the case of the steel products cartel condemned in France in 2008 [10], one of the cartel members (a sales manager at Descours et Cabaud) justified the practice by the crisis situation in the sector, explaining that "it was a question of survival".

22. Secondly, the fact that the damage caused is not directly visible to the perpetrators makes it easier to engage in unlawful conduct, unlike offences that target specific persons (e.g. snatching). In the case of a cartel, the damage is often diffuse or even indirect: cartels develop in intermediate product industries and affect a large number of customers, who will sometimes try to pass on the extra price to their own customers. Bryan Allison, a participant in the marine tubes cartel, said (O’Kane [2011]): "We are a tiny outfit, we are not involved with consumers, who are we hurting? (...) Who cares about us?"

23. Thirdly, participation in a cartel may be gradual and take place in several stages, starting with general discussions and ending with price-fixing meetings. Accustoming to the infringement then takes place gradually, which may reduce the perception of its illegality. Initially, the participant may even be invited to meetings where the anti-competitive object is not explicit or scheduled, before the content of such meetings is derived or disclosed to him. For example, in the case of the hygiene and cleaning products cartel, one of the participants (a sales manager at Unilever) stated: "The first time I was contacted for these meetings was at a forum where someone from Colgate offered me the opportunity to participate in meetings on topics of general interest that could be of interest to sales [11]people . "Similarly, the cartel may start out as an isolated, one-off event, affecting a specific customer, and then spread to all customers: in the chemical commodities case, for example, a member of the Western cartel said: ’I suggested in the tone of the joke that we call ourselves to consult on a customer ... We implemented this resolution and gradually extended it to other customers [12]... I was contacted for these meetings during a forum where someone from Colgate offered me the opportunity to participate in meetings on topics of general interest to sales people. ”

24. This alteration in the perception of the illicit nature of the cartel may be accentuated by the insufficient mastery by managers of the particularities of competition law, which differs from other areas of law, particularly contract law on the question of the formalisation of an agreement. In particular, the "passive" participation of a newcomer in meetings with an anti-competitive object may lead him to believe that his company’s will is not being carried away, on the grounds that he did not take an active part in the decision-making process. However, it should be recalled that, according to settled case-law, where an undertaking takes part, even without taking an active part, in meetings having an anti-competitive object and does not publicly distance itself from the content of those meetings [13], it is considered to be established that it is participating in the cartel resulting from those meetings.

25. Similarly, a manager may participate in a cartel organised in the context of meetings of a professional union, which may mitigate, in his view, the unlawfulness of his behaviour, insofar as the union’s primary mission is to defend the interests of its members. However, it should be recalled that, according to well-established case law, the participation of a trade union in a cartel does not in any way lessen the seriousness of the practice, quite the contrary: trade unions do not escape competition law when they assist in the conclusion or implementation of agreements having an anti-competitive [14]object and/or effect.

26. Once a list has been drawn up of the costs that he will have to bear by infringing competition law, the manager will assess the risk of detection of the cartel. This probability of detection has been the subject of a large body of empirical literature (see Connor and Lande [2012]), which reaches a fairly convergent result: despite different methodologies (surveys, statistical studies), the studies conclude that the probability of detection would range between 10% and 30%.

27. But again, this objective probability should not make us forget that the individual may be led to underestimate it, due to two cognitive biases.

28. The first bias is that of "availability": when an event makes an impression or is more visible than others, individuals tend to overestimate its occurrence. In the case of cartels, the small number of cases handled each year by antitrust authorities - between 5 and 10 decisions per year maximum in the case of the Commission - combined with their low media exposure, may lead individuals to underestimate this probability, or even to consider it almost nil.

29. The second bias is that of "overconfidence" (overconfidence): experimental psychology research shows that agents, faced with decisions whose outcome is uncertain, tend to underestimate the occurrence of unfavourable outcomes, thinking that they will do better than others, and to overweight favourable events. This bias is all the more marked when events are perceived as being controllable by agents (Korobkin and Ulen [2000]). Wils (2008) quotes a member of the American Bar Association who states that "most cartel participants believe that they will not be caught". In the case of a cartel, the probability of detection is not independent of the agents’ ability to conceal it: it therefore appears partly "controllable". Thus, to the question "Were you aware that participating in a cartel was a criminal offence? "Bryan Allison, convicted of participating in the marine tubes cartel, answers (O’Kane [2011]): "Yes, and that people could go to jail. But everyone had an aura of invincibility. "The experimental economics study by Damgaard et al. (2012) on encouraging managers to form a cartel [15] finds a result consistent with this interpretation: participants with a better knowledge of antitrust law are more likely to participate in a cartel than those who do not. This result can be explained by the sense of control that comes from mastering technical knowledge.

1.2 Gains from participation in a cartel

30. In most of the cartels detected, the individuals who implemented or participated in the practice held relatively high positions in their company’s hierarchy: very often they were sales managers, and sometimes even general managers. Connor and Lande (2012) thus examine the position of managers convicted of cartels in the United States between 1990 and 2008: they find that almost half of them were presidents or general managers, while the other half held a position as a manager in charge of sales or marketing (p. 440). In the case of France, several recent cartel cases highlight the involvement of senior executives. In the hygiene and cleaning products cartel, for example, it was mainly sales managers who made up the various negotiating [16]circles. The same is true in the case of the detergents [17] cartel and the steel [18]products trade. This high level decision-making does not then exclude the possibility that participation in meetings with competitors may be entrusted to managers at a lower hierarchical level, once the principle of the cartel has been established by the CEO (see on this point: Harrington [2006], chapter 4) [19].

31. This hierarchical positioning has a direct influence on the incentives of managers, in terms of career or remuneration, to engage in a cartel. In terms of expected gains, managers are assigned short-term objectives (such as revenue growth) by their hierarchy, on which their variable compensation depends in part. This is the case, for example, with the "bonus" system, the annual amount of which is often fixed in relation to a turnover target, particularly for managers in charge of sales functions. Similarly, top management may be granted stock options, the holding period of which before they can be exercised is limited, generally four to five years. The manager will then compare the profitability and speed of implementation of different strategies for increasing sales or profitability, some of which may prove to be illegal. Compared to other strategies such as launching innovations or prospecting for new customers, the advantage of cartel practice is that it is quick to implement: the increase in turnover and profit results from the "extra price" imposed by the cartel on customers. Moreover, provided that the cartel is stable, the amount of the gain can be approximated ex ante. If production costs between the cartel participants are similar, the cartel also ensures the same relative performance between companies in the same sector, which is often an internal evaluation criterion for a manager’s performance (Armstrong and Huck [2014]). Similarly, a cartel practice allows the manager to smooth performance, avoiding profit fluctuations, and to set it at a sufficiently high level to receive his bonus, without maximising his effort (Spagnolo [2005]). Finally, remuneration based on the share price has the effect of stabilising the cartel by limiting the manager’s incentives to "cheating": since cheating today may lead to a price war later on, it is no longer in the manager’s interest to deviate since the share price will be lowered in the future in the event of cheating, due to the fall in profits resulting from the possible price war (Spagnolo [2000]). In the case of a bonus, which is in principle capped, the incentive to deviate is also low since the gain for the company from the deviation will not be reflected proportionally in the manager’s bonus.

32. Beyond the perception of his variable compensation, the manager may want to take risks to outperform his peers internally and thus progress faster within the company.

33. If the company shows little regard for compliance with competition rules, the manager will anticipate that his illegal behaviour will not be sanctioned, or may even be "rewarded" with a promotion. For example, Stephan (2011) points out that one of British Airways’s senior managers was promoted after his company was convicted of cartel offences even though he was being prosecuted by the American courts. In the same vein, Connor and Lande (2012) analysed the fate of 35 people who had been sentenced to prison terms for cartel and show that some of them were still employed by their company. Managers can also rely on external recruitment, given the good operational performance they have achieved. Thus, Connor and Lande (2012), in the same sample, estimate that at least 9 managers had changed companies but remained within the same sector.

1.3 Cost-Benefit Comparison: Discounting, Risk Appetite and Ethical/Moral Sense

34. When the individual compares the expected costs and benefits of participating in a cartel, his choice will take into account three parameters: the effect of temporality, the degree of risk aversion, and the degree of adherence to business ethics (morality).

35. Firstly, there is a temporal asymmetry between the gains and costs of a cartel strategy: the gains in terms of promotion and remuneration occur in the short term, whereas the costs appear more distant, given the time taken to detect and investigate a case. The comparison of gains and costs must therefore include a discount factor: the more the future is depreciated in the eyes of the manager, the lower the perceived cost of the practice. This point particularly concerns managers’ remuneration, whether in the form of bonuses or even stock options: stock options can be exercised after 4 years, even though cartels detected in Europe by the competition authorities have an average life of 7 years (see Combe and Monnier [2012]). Moreover, as shown by the econometric study of Alawi (2014), CEOs who participate in a cartel have a higher age than the reference [20] group: it is therefore likely that their career will be over when the cartel is detected and their discount factor is therefore low. Finally, Gonzalez and Schmid (2012) find that CEOs who participate in a cartel exercise a higher proportion of their stock options during the cartel period than CEOs of similar companies who are not involved in a cartel: they therefore take advantage of the cartel practice to reap an immediate gain in the form of capital gains.

36. Second, the individual will make his or her decision, in a context of uncertainty, according to his or her degree of risk aversion, which will modify the average value of anticipated gains and costs. The hypothesis most frequently retained in microeconomics is that agents are risk-averse: between two choices that provide the same expected gains, the individual always opts for the most certain solution. But this hypothesis deserves to be discussed in the case of cartels, in the light of the experimental study by Damgaard et al. (2012): the authors show that some managers - always more than 10% and up to more than a quarter - always choose to form a cartel, regardless of the probability of detection, the amount of the fine and the overprofit associated with the cartel. This result can be interpreted as the fact that some managers, far from being risk-averse or risk-neutral, have a systematic taste for risk and/or illegal activities. Moreover, the influence of this parameter is difficult to assess insofar as its impact on the incentive to participate in a cartel is ambiguous: on the one hand, the risk-averse individual is likely to choose the cartel in order to reduce the risk of being dismissed or not achieving his or her performance objectives; but on the other hand, this same individual may be particularly reluctant to engage in an illegal practice associated with a risk of detection and sanctions (whether internal to the company or external such as criminal sanctions).

37. The question then arises as to what factors influence the degree of risk aversion/taste. Beyond individual characteristics, the empirical literature points out that demographic (gender, education, age), socio-economic (income, occupation) and personal (attitude, morality, etc.) attributes play a determining role. In particular, several empirical works in finance have shown that women are more risk averse than men. In particular, men and women differ in their investment behaviour, in lottery and/or betting situations, and in their attitudes, intentions and hypothetical choices regarding financial risk-taking (for a review of the literature on the subject, see Meier-Pesti and Penz [2008] and Croson and Gneezy [2009]).

38. Finally, an individual’s choice will be influenced by his or her degree of adherence to moral values and ethics in business life. The more developed the individual’s moral sense, the higher the subjective cost of cartel practice will be. Once again, beyond individual characteristics, the empirical literature has established a "portrait-robot" of the corporate fraudster: according to a recent study by PWC (2016), it is generally a man with a higher education diploma, of average age (in his 30s or 40s) and who already has 3 to 5 years of experience (see also PWC and London Business School [2015]). Other empirical studies show that women have a higher aversion to unethical behaviour than men (see not. Robinson et al. [2000]). In particular, the econometric study by Alawi (2014) shows that the presence of women in the management team significantly reduces the likelihood that a firm will engage or remain in a cartel, compared to a control group of similar firms that did not participate in a cartel. The experimental economics study by Hamaguchi et al. (2009) concludes that women are more likely than men to end a cartel by denouncing it through a leniency programme.

2. Keep the deal or cheat?

39. Once the individual enters the cartel, he will ask himself a second question: what is the incentive to respect the agreement? Economic science, through the theory of repeated games, has long since answered this question from the exclusive point of view of the incentive for a company to "cheat": the discount factor, the mechanisms for detecting cheating, the extent of the reprisals, the symmetry of the size and cost of companies will play a decisive role in the internal stability of a cartel (for a summary, see Combe [2016]). But socio-cultural and psychological factors specific to individuals can also intervene and reinforce the stability of cartels.

2.1 Profile Similarity

40. The members of a cartel can consider each other as "peers" as long as they share certain common characteristics: the fact that they work in the same sector of activity, with a strong interpersonal knowledge, the fact that the individuals have received the same academic training, are of the same nationality, evolve within the same circles, display the same social status, all favour the cooperation and stability of an organisation. In the interview devoted to him, Bryan Allison, an English national who took part in the underwater tubes cartel, insists on the role of nationality (O’Kane [2011]): "I had a reasonable degree of confidence in the Europeans. I was always a little wary of the Japanese. ”

41. Based on experimental studies, Glaeser et al. (2000) show that individuals of the same social status are more likely to cooperate with each other. Similarly, in the case of the UK sea freight cartel in the period 1879-1929, Podolny and Scott Morton (1999) study the probability of cartel members engaging in a price war if a new entrant enters the market; they show that predation and collective boycott are less likely if the new entrant is of the same social status as the cartel members or of English nationality. More recently in France, in the case of the hygiene and cleaning [21]products cartel, the entry of a new participant into the group was the result of a process of "adoubtement" which gave the practices a selective character. There was thus a co-opting mechanism based on the qualities of the employee and the characteristics of the company. The members of the steel [22] trading cartel, for their part, adopted measures to put pressure on foreign and/or non-regional companies.

2.2 Creation of common values, "esprit de corps" and ratchet effect

42. Even when the members of a cartel do not initially present similar characteristics, the frequent interactions between them will help to build ties, foster mutual trust and even gradually create an "esprit de corps". Laboratory experiments confirm that collusion is strengthened when frequent social interactions between participants occur (Armstrong and Huck [2014]). A fortiori, cooperation and trust become possible when a third party actor, with strong legitimacy among the players, intervenes: in the case of a cartel, professional associations, whose primary mission is to defend the interests of their members, can contribute to the emergence of these common values. Stucke (2011) mentions the motto of one of them: "(...) competition is war, and war is hell. "These values can also have their origin in the history of the sector. Thus, in the case of the steel trading cartel in France, the members of the cartel refer to the existence in this sector of "a historical tradition of consultation" [23].

43. Cartel organizers can encourage loyalty among participants, much like the steel cartel in the United States at the beginning of the last century. In this case, cooperation was fostered by informal meetings described as "Gary’s Dinners" (named after the cartel organiser, Elbert Gary). In his speeches at the dinners Gary referred to the trust in the word given: "(...) we have better than a contract to guide and control ourselves, we as men, as ʻgentlemen’, as friends (...). We are committed to protecting each other, and we cannot act or omit to act without a clear sense that our honour is at stake. "More recently, in France, in the agreement on hygiene and cleaning products, trust in the word given - and therefore paradoxically honesty - was one of the cements of cooperation. When asked about the truthfulness of the information exchanged between participants during meetings, one of the members (a sales manager at Colgate-Palmolive) said: "Trust was the golden rule (...) if we spoke, we had to tell [24]the truth. "Similarly, in the laundry detergent [25]case, the cartel participants agreed to adopt a "code of good conduct".

44. Experimental economics - notably in the "ultimatum game" experiments (Armstrong and Huck [2014]) - has highlighted the fact that individuals who can cheat behave less aggressively or more loyally than theoretical models predict, fearing the "spirit of revenge" of other players. In the case of a cartel, the fear of all-out war in the event of cheating may promote internal stability. The mutual dependency between the cartel members, once the cartel is implemented, thus contributes to strengthening it. In the case of the steel trading cartel in France, one of the participants (sales manager at Descours et Cabaud) stated: "(...) our interests were linked and if someone deviated, everyone would fall and everyone would be in danger of losing their feathers. So there was a climate of trust that was maintained [26]. ”

45. This mutual dependence can be reinforced by the mobility of managers between companies within the same sector: the prospect of possibly being hired tomorrow by a competitor reduces the manager’s incentive to behave unfairly in the cartel. Thus, in the chemical commodities case, it was noted that there was a certain rotation of managers between companies, as one participant pointed out: "[]In addition, there was a lot of recruitment by competitors: at Solvadis, the directors of the Île-de-France, North and South-West regions came from Brenntag, so I couldn’t have been unaware of what was going on at Brenntag [27]. "In the same way, the practice of cross-directorships between companies in the same sector reinforces knowledge and interdependence between managers and thus limits the incentives to cheat in the cartel.

46. Once the group identity has been created within the cartel, the psychological "cost" of the deviation (cheating within the cartel, or even leaving the cartel) increases for the individual: through a kind of "ratchet effect", it becomes difficult for him or her to turn back. In particular, once inside the cartel, it becomes more difficult for a manager to "denounce" the cartel to the antitrust authorities, as leniency is a form of "betrayal" to his peers. For example, Bryan Allison, convicted for his involvement in the underwater tubes cartel, answers several questions about leniency programmes as follows (O’Kane [2011]): "I rather think that ʻgrassing people up’ isn’t really the done thing. Isn’t that a little unethical? (...) Go shop your fellow conspirators, it’s a bit below the belt. "Similarly, as with cheating, the fear of future stigmatisation following a whistleblower report can also discipline the individual, who may fear becoming "unemployable" in the sector following a leniency application from his or her company. As a result, the leniency procedure will be further mobilised by new managers/ shareholders of a cartel company (e.g. following a takeover). Individuals will indeed be more inclined to collaborate, since they come from other geographical horizons or relational circles than the hard core of cartel participants and/or have personally never taken part in the cartel: they will therefore not suffer the psychological cost of denouncing previous "allies".

II. How to discipline managers?

47. As we have seen above, as long as administrative sanctions against companies are sufficiently dissuasive, shareholders have no incentive for managers to violate antitrust laws: it is then in their interest to deter the commission of such practices within the company. Several levers can be mobilised to this end: executive compensation arrangements, a system of sanctions in employment contracts, reinforcement of compliance programmes, detection of "risk profiles". But even in this case, the company’s internal tools may prove insufficient, particularly in view of information asymmetries: the intervention of the "visible hand" of the public authorities then finds its full legitimacy, in addition to that of the shareholders.

48. A second case requires external intervention: when shareholders and managers display a convergence of interests to violate antitrust laws. In this case, it is the lack of internal incentives within the company that mainly justifies the implementation of public policies aimed at individuals. These public policies can be based on various levers, including the dissemination of a culture of competition, stigmatizing cartels, or the implementation of criminal sanctions against managers.

1. Internal company solutions

49. In order to discourage managers from engaging in cartel practices, shareholders can implement monitoring and sanction mechanisms as well as a management style that encourages ethical behaviour (Table 3).

Table 3. The different internal action levers aimed at disciplining managers is available in the pdf version of this article.

1.1 Compensation packages for executives

50. Performance-based compensation schemes are likely to encourage individuals to defraud, as has been highlighted by a large empirical literature, mainly focused on financial and accounting fraud. For example, Goldman and Slezak (2006) and Carson (2006) show that incentive compensation schemes not only stimulate executives’ efforts, but also increase their incentive to undertake fraudulent actions. Bergstresser and Philippon (2006) and Denis et al. (2006) show that the probability of fraud is higher in firms where executive compensation is more closely linked to share value and option ownership than in other firms. Bebchuk et al. (2010) provide a case study of compensation at Bear Stearns and Lehman (2000-2008) and conclude that the compensation schemes put in place may have encouraged managers to take excessive risks: in particular, the bonuses were not accompanied by a withdrawal clause in the event of the firm’s bankruptcy.

51. On the side of economic theory, several contributions have analysed the relationship between corporate fraud and managerial compensation, with contrasting results. Crocker and Slemrod (2007) show that providing compensation that encourages managers to reveal the real profits made by the company is particularly delicate. Andergassen (2008) shows that senior executives have a shorter-term vision than shareholders and are more inclined to manipulate information. Fleckinger et al (2013) show that shareholders can in some cases encourage managers to comply with the law through complex compensation schemes: in order to avoid the procollusive effects of performance-linked pay, compensation contracts should include suspensive clauses that consist in withdrawing the right to exercise stock options in the event of proven collusion. This alignment of interests between managers and executives in the sense of compliance with competition law nevertheless proves costly for shareholders, who have to cede an annuity to executives in the form of a higher salary. Under these conditions, if certain managers are particularly "amoral", their illegal behaviour cannot be avoided: the additional salary costs compatible with compliance with the law being prohibitive, shareholders will indeed have an interest in tolerating illegal practices emanating from these managers (the result of the aforementioned study by Damgaard et al. [2012] may be similar to such a case). Moreover, as Spagnolo (2000) points out, granting stock options to managers can reinforce the stability of a cartel: in fact, since the exercise of the option is deferred in time and since financial markets instantaneously integrate any new information, the gain of a "deviation" strategy within the cartel is offset by the fall in the share price, since the markets anticipate a price war tomorrow as a form of retaliation. Finally, given the delays in detecting and condemning cartels, it is possible that the manager/shareholder may have exercised his options before the practice is discovered by the antitrust authorities and thus realize a capital gain.

52. Ultimately, in order to encourage managers to better comply with competition rules, an internal company solution consists less in modifying the methods of variable remuneration than in modifying the conditions for granting or exercising such remuneration: rather than limiting performance-linked remuneration (which can have a positive effect on productivity and risk-taking), it is better to provide in the employment contract for the reimbursement of the bonus or the capital gain realised on stock options, if it turns out ex post that the manager has infringed competition rules. However, this "retroactive" solution comes up against the classic problem of "capturing" decision-making bodies: if managers control the procedures for setting remuneration within the company, it becomes difficult for shareholders to have this type of measure adopted. Bebchuk and Fried (2003) thus highlight that managerial power affects the design of compensation schemes in firms where ownership is separated from control. Conyon (2006), who studies changes in managerial compensation in the United States and the financial incentives put in place between 1993 and 2003, also shows that compensation and management committees have become increasingly independent of shareholders.

53. While the role of remuneration is crucial in encouraging managers to comply with competition law, it should not be forgotten that a manager’s personal interest is not limited to his or her remuneration alone: the desire to progress quickly in a career, outside the company, also comes into play, especially in a context of high managerial turnover between companies. In fact, other schemes may be used to complement compensation schemes that encourage compliance with competition rules.

1.2 The threat of dismissal

54. Another option is to put in place a policy of internal sanctions, including dismissal or compensation. Thus, shareholders could ex post sanction executives who have formed a cartel, by dismissing them or blocking their career progression within the company. In practice, however, there is a problem of temporality in this policy. Indeed, assuming that shareholders want to "punish" deviant executives by dismissing them, it is unlikely that the individuals who participated in the practice are still employed in the company, given the high turnover of executives. For example, Kaplan and Minton (2012) (estimate the turnover rate at 15.8% in large US companies over the period 1992-2007, implying an average length of time in management positions of less than 7 years. Over the more recent period, from 2000 onwards, this turnover even increases to 16.8%, implying that executives remain in their positions for an average of less than 6 years. If we focus on CEOs alone, the term of office is even shorter: according to Schwab and Thomas (2006), it is in 81% of cases less than 5 years, with an average of 3.6 years. Similarly, out of a sample of 184 CEOs of companies listed on the New York Stock Exchange, Gillian et al. (2009) obtain a median of 3.4 years and an average of 3 years, which is very short [28]. However, the time between the start of anti-competitive practices and their detection and subsequent condemnation by competition authorities is generally quite long. According to Stephan (2011), during the period 2008-2010, the total time between the start of a cartel and its condemnation by the Commission ranged from 5 to 40 years. For example, in the case of the car windscreen cartel, the sanction was imposed 9 years after the infringement had ceased. Similarly, if we consider the 111 cartels condemned by the Commission between 1969 and 2009, their average lifespan was 7 years before they were detected (Combe and Monnier [2012]): if we take into account the time taken to investigate a cartel case, the decision to impose a penalty will be taken - at best - 9 to 10 years after the start of the practice. The manager can therefore anticipate, at the time he commits to a cartel practice, that he is unlikely to suffer internally from the consequences of his behaviour. The econometric studies by Han (2010) and Alawi (2014) confirm the existence of a negative relationship between the length of CEOs’ terms of office and the likelihood of the company engaging in a cartel: managers with a short life span in the company are more likely to engage in this type of illicit practice.

55. The company’s shareholders could also turn against the managers, suing them civilly or criminally for breach of duty and claiming compensation. However, the ability of individuals to pay often remains limited compared to the damage that the cartel will have caused to the company.

1.3 Strengthening Compliance Programs

56. Another solution is to strengthen compliance programs, through training sessions, "good conduct" manuals and contractual commitments, to encourage managers to denounce any anti-competitive practices of which they become aware within the company, and to conduct periodic and surprise audits to detect possible breaches of competition rules. Compliance programs, which are part of a culture of corporate self-regulation, can help to reduce the cognitive and psychological biases mentioned in section I, in particular by preventing the proliferation within the company of values contrary to antitrust law and a diminished perception of the illegal nature of cartels. In this respect, the frequency of training is a key factor in maintaining a "competition culture" within the company on a regular basis, as is the possibility for employees to inform their senior management, anonymously, of a possible infringement of competition law within the company (the so-called "whistleblower" procedure).

57. However, while compliance programs reduce the risk of infringement, they may not be sufficient on their own to deter employees from engaging in anti-competitive practices for several reasons (see Stephan [2010]):

– In many cases, the employees who took an active part in the cartel were well aware of the illegality of their behaviour. The lysine or graphite electrode cartels are textbook cases in this respect: in the first case, the cartel members ostensibly mocked the antitrust authorities and the FBI, while in the second case they continued their practice even though the Commission had opened an investigation ;

– Internal audits are not always suitable for detecting secret practices, for which cartel members can deploy treasures of ingenuity. For example, in the case of the graphite electrodes, a complex system of code names was used to mask the identity of the companies and individuals involved in the cartel ;

– a compliance program is only truly effective if it is carried and embodied at the highest level of the organization. However, Stephan (2011) shows, on the basis of a sample of 40 cartels, that sales or marketing directors and sometimes even senior management (vice-president, CEO) were directly involved in setting up a cartel. In this case, the compliance programme appears to be purely "cosmetic", in that it is not respected by the very people who should embody it;

– A liability regime that is centred on the company and conditional on the implementation of a compliance program implies monitoring costs for the company (see Aubert et al. [2006]). If these costs are too high, shareholders will refrain from setting up such a system;

– The effectiveness of a compliance program also depends on the governance structure of the company. Abrantes-Metz and Sokol (2015) show that the complexity of the organisation - associated for example with a larger company size - increases the costs of monitoring agents and thus contributes to encouraging illegal behaviour. Gonzalez et al (2014) show that companies involved in cartels have more "busy" directors (sitting on many management bodies) and tend not to change auditors. Compliance programmes should therefore provide for a more regular change of auditors, which also generates a cost.

1.4 The managerial functioning of the company

58. The type of management implemented within the company can influence the incentive to defraud: for example, recent research by PWC and the London Business School (2015) on the promotion of ethical behaviour in the financial [29]services sector shows that management based on a "tough" and punitive approach in the event of poor performance generates a climate of fear, which in turn leads to unethical behaviour. Moreover, if shareholders set unsustainable profitability targets, they implicitly invite managers to resort to fraudulent means to "meet the targets".

59. The PWC study also shows that, to avoid fraudulent behaviour, management rules must be clearly defined and transparent, given managers’ aversion to ambiguity. Ambiguity describes a situation in which an individual has to make choices when the probabilities of events that may occur are difficult to estimate (doubts, inaccuracies and/or missing data). The manifestation of this ambiguity can result from contradictory or incomplete information, which leads to a lack of confidence in the individual’s own expectations. Work in psychology and experimental economics has shown that individuals’ aversion to ambiguity is stronger than risk aversion (for a synthesis, see Camerer and Weber [1992]), as has some work in the neurosciences (for a synthesis of this work, see Schmidt [1996]; for an overview of the literature, see Cabantous and Hilton [2006]). For the manager, anxious to receive his bonus or to benefit from a promotion, a cartel strategy will enable him to reduce ambiguity, by limiting uncertainty about possible events: the cartel ensures a fairly predictable level of profitability, stable over time and comparable to that observed among its competitors who are members of the cartel. Cartels would thus be more likely to develop in companies for which the information relating to the conditions and rules for granting bonuses and promotions is perceived by managers as incomplete or contradictory. While antitrust sanctions against companies are not a deterrent, shareholders even have an interest in deliberately maintaining a high degree of ambiguity among managers, for example concerning the risk of dismissal, the likelihood of promotion, possible internal transfers, in order to induce them to engage in a cartel. This policy of "management by ambiguity" has the advantage for shareholders of inviting managers, without the need to tell them, let alone write them down, to break the rules while making them bear full responsibility for their illegal actions. Conversely, when antitrust sanctions against companies are dissuasive, shareholders must make the remuneration of managers and their career development predictable (through systems of grids, indexes, etc.) if they want to dissuade managers from infringing antitrust rules.

1.5 The "profiling" of managers during recruitment and promotions

60. As a subsidiary option, the company may also include criteria relating to ethical behaviour and risk-taking in tests for the recruitment or promotion of managers. Indeed, as the experimental study by Damgaard et al (2012) has shown, some individuals, who could be described as "hotheads", prove to be insensitive to the values of detection and sanction parameters and always choose to form a cartel. This raises the question of the ex ante identification of these individuals, who have a "risk-anything" and perfectly "amoral" profile. During interviews and job tests, where initiative is highly valued, it would be useful to ensure that compliance with the rules of the game is part of the criteria for evaluating a candidate. Furthermore, promoting a better representation of women in management bodies could reduce the risk of cartels, given their greater adherence to ethics.

2. Solutions external to the company

61. It appears that internal company solutions, while necessary, are both difficult for shareholders to implement and probably insufficient to dissuade managers from engaging in a cartel.

62. Under these conditions, the intervention of the "visible hand" of the public authorities takes on its full meaning: by spreading a "competition culture" stigmatizing cartels, by resorting to penal/administrative sanctions against individuals (Table 4), public action "complements" that of private companies. This public action directed at individuals may also find internal legitimacy, given the limits of the policy of sanctions against companies: it is possible that the amount of the sanctions that make it possible to ensure the deterrence of cartel practices exceeds the companies’ capacity to pay (see Wils [2001]). In these circumstances, the introduction of criminal sanctions against natural persons, although of a different legal nature, "complements" the action against legal persons to ensure the deterrent effect of antitrust policy.

Table 4. The different external levers for disciplining managers is available in the pdf version of this article.

2.1 The spread of an "anti-cartel" culture

63. According to a recent PCW study (2016), 44% of the companies surveyed stated that the issue of worker morality is the main issue in corporate fraud (and lack of ethics, the main cause of fraud). If we want to dissuade managers from engaging in cartel practices, an external solution therefore involves building business ethics, which leads to stigmatising this type of practice, thus increasing the "psychological cost" of violating competition rules. In particular, in the case of cartels, it is possible to equate this practice with theft, insofar as a cartel implies, among other effects, an unjustified transfer of wealth from customers to cartel members. However, this diffusion of a "culture of competition" is a long-term process, as shown by the information campaigns against drink-driving, which have very slowly changed people’s perceptions and behaviour. The dissemination of an "anti-cartel" culture can be supported by vectors such as the media, political decision-makers, academic training or competition authorities.

64. With regard to the role of the media, it must be noted that cartel practices are the subject of relatively limited media coverage. Thus, Sokol (2012) shows that during the period 1994-2008, cartel decisions gave rise in the United States to few press articles compared to other "white-collar crimes": in a sample of 815 American newspapers, the number of press articles per cartel case varies between 1 and 3.7, whereas it can exceed 200 in accounting fraud cases and for the New York Times alone.

65. Stephan (2012) makes a similar observation in the United Kingdom and explains this low media coverage of cartel cases by several factors:

– Cartel practices are collective practices, even though the media are rather attracted by the personalisation of "white-collar crimes", in the image of insider trading. The so-called "Kerviel" case is a recent illustration of this;

– Most cartels involve intermediate goods, which do not directly affect consumers and are therefore not of obvious media interest.

66. Policy-makers can also play a role in the fight against cartels by putting the subject on their political agenda. Thus, in the Australian case, Beaton-Wells and Haines (2010) analyse how the political class, in the 2000s, successfully transformed a narrowly defined subject - the fight against cartels - into a real issue for society, culminating in 2009 in a new law criminalizing these practices. The fight against cartels was presented as a measure to protect consumers and promote purchasing power, in a context of liberalisation and international openness of the Australian economy. Some policies have also made it a measure to defend small producers against large international cartels.

67. As far as academic training is concerned, it has to be said that compliance with competition rules is still rarely addressed in the curricula of management students, even though courses on business ethics are nowadays given, mainly on corruption or accounting and financial fraud.

68. Competition authorities can also play an advocacy role, by widely disseminating the results of their litigation activity, promoting dialogue and meetings with businesses, administrations, political staff, professional organisations and even the general public. In particular, in the case of France, the Competition Authority may, in the course of sector inquiries or opinions, reiterate the principles governing competition, particularly in the area of cartels. In addition, it is essential to assess the general public’s perception of cartels and the policy to combat them in order to target the points on which the authorities need to place greater emphasis. But this educational role will be all the more convincing if it is fuelled by contentious cases: it is therefore imperative to maintain a sufficiently high level of detection of cartels. Indeed, the stigmatization of collusion behaviour, the dissemination of pro-competitive norms and anti-trust rules will be all the easier to achieve if cartels are regularly condemned and if these condemnations are relayed in the media. Sustained and regular detection thus helps to reduce the cognitive biases studied previously, and in particular the underestimation of the risk of detection.

2.2 Criminal sanctions in the form of fines/prison sentences

69. Public authorities may also choose to introduce criminal sanctions against natural persons in addition to administrative sanctions against companies. In addition to overcoming the company’s inability to pay the "optimal" fine, where appropriate, criminal sanctions can enhance the attractiveness of leniency programmes when immunity from punishment is extended to criminal proceedings against employees of the first whistleblower company; they also send a strong signal about the social stigma of cartel practices (Wills [2008]).

70. Criminal sanctions mainly take the form of fines or prison sentences against individuals.

71. A fine, as long as its legal ceiling is not set too low, has the advantage of establishing a monetary equivalence between the personal illegal gain made by the manager because of the cartel (receipt of a bonus, faster promotion) and the sanction: the latter must, at the very least, confiscate it. But a personal fine has several disadvantages:

– it is always transferable to the shareholders, explicitly in the employment contract or de facto ;

– its stigmatizing effect is weak: the manager simply has to pay a cheque, which removes part of the intrinsic motivation to comply with the law ;

– in practice, the fine ceilings are set at a relatively low level compared to the illegal gain the manager may have made. Thus, in the case of France, Article L. 420-6 limits the fine to EUR 75,000.

72. Criminal sanctions in the form of prison sentences make it possible to remedy the first two disadvantages mentioned above: the prison sentence is directly borne by the individual in question and has a strong stigmatizing effect on the latter, confronted with the eyes of his or her relatives and peers. In the United States, the choice has clearly been made since the 1990s to strengthen criminal sanctions in the form of prison sentences (see Hammond [2010], Kovacic [2016]). This policy is supported by a renewed legal arsenal: revision in 1993 of the leniency programme with immunity from criminal prosecution for employees of the first whistleblower company; adoption in 2004 of the Antitrust Criminal Penalty Enhancement and Reform Act, which raises the ceiling of sanctions against individuals for violations of the Sherman Act to $1 million and 10 years in prison. In practice, the American justice system has also tended to broaden the scope of criminal liability by prosecuting several employees of the same company in cartel cases: in the case of the vitamins cartel, six Hoffmann-La Roche executives were thus charged. Moreover, since the vitamins cartel (1999), no fewer than 40 non-US executives have served prison sentences on US soil in international cartel cases.

73. Beyond the American case, several countries have developed a criminal arsenal to combat cartel practices, such as Ireland (1996), the United Kingdom (in 2002) and Australia (2009). However, with the exception of the American case, the effectiveness of the criminal solution remains very limited to date. In France, for example, an average of two criminal convictions per year were handed down under Article L. 420-6 of the Commercial Code during the period 1986-2006, most of them in the form of fines or suspended prison sentences, and in the context of broader criminal proceedings, particularly in relation to the offence of favouritism (Viros [2016]).

74. The English case is particularly instructive in this respect: the criminal scheme provided for in the Enterprise Act 2002 resulted in only one prison sentence in 12 years, in the case of the marine pipe cartel, against three senior executives. The low effectiveness of this system can be explained by the standard of proof required for conviction. Thus, until 2014, the judge had to prove that the participants in the cartel had acted "dishonestly"; the concealment and secrecy of the practice were not in themselves sufficient to prove the willingness to act dishonestly. For example, in the case of the cartel of galvanized steel tanks, both accused were acquitted for lack of sufficient evidence of "dishonest" conduct (Wagner-von Papp [2016]) [30].

75. In France, although the standard of proof is different, it is nonetheless very restrictive: it is not enough that the natural person was involved in the cartel; the judge must also show cumulatively that he or she took a "personal and decisive" part in the conception, organisation or implementation of the cartel, and that his or her intention was fraudulent (Viros [2016]).

76. More fundamentally, with the exception of the United States, prison sentences for cartel practices suffer from a lack of legitimacy among the population. The Cartel Project empirical study on the Australian case is instructive in this respect (Beaton-Wells and Haines [2010]): on the basis of a large questionnaire survey, the researchers show that while 62 per cent of respondents consider cartels to be a "dishonest" process, only 42 per cent consider that they should be treated as criminal practices. Moreover, prison appears to them to be clearly disproportionate to the seriousness of the offence: sanctions such as fines on individuals, publicity about their misdeeds (naming and shaming) and disqualification are more popular (Figure 1). 77. Similarly, Stephan (2015, 2016), who analyzes studies from the United States, Germany, Italy, and the United Kingdom, notes that with the exception of the Italian case, the majority of respondents consider the cartel to be as serious a practice as fraud. However, only a minority of respondents were in favour of prison sentences: 26% in Italy; 27% in the United Kingdom; 28% in Germany; 36% in the United States [31].

Figure 1: Appropriate sanctions for cartel participants (Australian investigation) is available in the pdf version of this article (Source: Beaton Wells and Haines (2010)).

2.3 Disqualification of managers

78. Faced with the lack of legitimacy of prison sentences and given the low effectiveness of criminal fines, measures for declaring "incapacity" constitute a middle way that deserves to be explored: a manager or corporate officer who has directly contributed to violating competition rules or who has knowingly "allowed" managers to do so may be banned from exercising any position of responsibility within a company for a certain period of time. This is the approach taken by the United Kingdom through the Competition Disqualification Orders, which were reformed in 2010 and provide for a maximum period of incapacity of 15 years.

79. There are several arguments in favour of disqualification measures:

– disqualification negatively affects the leader’s reputation and career prospects, while minimizing costs to the community (no incarceration). A study by Deloitte for the OFT (2007) shows that companies consider disqualification of executives as a particularly dissuasive measure, just after prison and before antitrust fines;

– This measure seems to be more accepted by the public than imprisonment: 48% of respondents in the United Kingdom are in favour of it, compared with 11% for prison sentences (Stephan [2012]). Disqualification thus combines strong social acceptability with a deterrent effect (Table 5);

– the scope of disqualification may be broader than that of criminal sanctions: in the United Kingdom, the new 2010 disqualification provisions provide that a director may be prosecuted for failing to report a cartel of which he could not reasonably have been unaware;

– it is more difficult for the disqualified person to circumvent the effectiveness of this measure, compared to a personal fine, which can be compensated by the employer.

80. However, the disqualification procedure also has several limitations:

– it may have an adverse effect on the effectiveness of leniency programmes: if the leader can escape disqualification by collaborating with antitrust authorities in the context of leniency, his personal ex ante incentive to form a cartel is strengthened. In order to avoid this adverse effect, it is necessary to limit immunity from disqualification to executives of the first company to collaborate with the antitrust authorities;

– the disqualification procedure may affect company executives of a high age, close to retirement, thus limiting its effectiveness;

– it is always possible for shareholders to compensate the disqualified executive, for example by granting him or her generous compensation to compensate for the loss of income resulting from the forced discontinuation of his or her business;

– it cannot be ruled out that disqualification may lead to experienced and successful managers being excluded from the conduct of companies.

Table 5. Social acceptability and deterrent effect of different individual sanctions is available in the pdf version of this article.

2.4 Strengthening leniency programmes and the remuneration of whistleblowers

81. The leniency programmes set up in a large number of countries aim to destabilise cartels by granting partial or total immunity from sanctions to companies that cooperate with the antitrust authorities in the detection of cartels in which they have taken part. These programmes have proven to be effective in detecting cartels, particularly in view of the "first come first served" rule, which grants full immunity from sanctions to the first cooperating company. In the American case, this total immunity even extends to criminal proceedings against individuals who took part in the cartel. But in order to dissuade managers from engaging in a cartel, another solution is to grant a reward to individuals - and in particular managers - who collaborate with the antitrust authorities.

82. Indeed, Buccirossi and Spagnolo (2008) show that rewards for whistle-blowing employees have similar deterrent effects to those of leniency programs, which involve increased transaction costs for firms involved in cartels. Indeed, a firm participating in a cartel must increase the remuneration of employees who are informed of its firm’s unlawful conduct, so that they do not report it to the authorities. This arrangement thus improves the deterrence of cartels by directly increasing the cost of collusion. According to these authors, these systems, if well designed, have deterrent effects complementary to those induced by leniency programmes (see also Aubert et al. [2006]).

83. While the United States has rejected the use of such reward mechanisms in antitrust offences, they exist and work effectively for other types of fraud. For example, the False Claim Act in the United States is a reward mechanism for individuals who report fraud to the government. Under this scheme, whistleblowers can receive up to 30 per cent of the funds recovered, which can be a substantial amount: the United States government was able to recover $3.5 billion between 1986 and 2000 (Stephan 2014). Similarly, the Dodd-Frank Wall Street Reform and Consumer Protection Act provides for a reward mechanism when whistleblowers provide information that enables successful investigations to be conducted before the Securities and Exchange Commission (SEC). More specifically, the SEC has a reward program in place since 2011 for individuals - whether or not they are company managers - who report fraud over $1 million. The reward can be between 10 and 30% of the amount collected, with the amount awarded varying according to the importance of the information provided, the degree of involvement of the whistleblower in the commission of the fraudulent practice, the time frame, and the interference with the company’s internal reporting. Over the period 2011-2015, the US agency in charge of this programme (Office of the Whistleblower) has awarded no less than $54 million to 22 whistleblowers, with a reward exceeding $30 million in 2014.

84. Such a reward system was introduced in the fight against cartels in South Korea in 2002. This system was reformed in 2004 by the adoption of an amendment that defines a formal reward mechanism for cartel informants (for the first informant only) [32]. The amount of the reward is decided by a committee whose role is to ensure transparency in the awarding of the reward. Confidentiality was also guaranteed and the amount of the reward increased. This reform was implemented in response to a lack of whistle-blowing, attributed to cultural barriers, fear of reprisals and the low levels of rewards available. Since 2006, the program has been applied eight times, in some cases that would not have been detected without this procedure (the case of the welding rod cartel, where the whistleblower received a reward of US$62,657). In December 2012, the maximum amount of the reward was awarded to a whistleblower, i.e. US$2.8 million (see Stephan [2014] for more details).

85. Based on an analysis of the experience in South Korea, Stephan (2014) highlights the conditions for the effectiveness of a reward policy. Although the authorities would prefer to reward only employees not directly involved in the cartel, reward mechanisms should not exclude individuals directly involved in the illicit practice, since the employees of the company in question may be the only possible informants (they hold the evidence useful for convicting the cartel) or may have been pressured to participate in the offence. The measures adopted to protect whistleblowers from possible reprisals by their employers are not sufficient to ensure that individuals report fraud. This is particularly important in view of the significant costs and risks involved. These risks include the risk of dismissal and a reduction in the individual’s future employability, repercussions for the family and on personal and social life. In order to counter these costs and risks, the reward amounts must be very large ($4 to $5 million in Stephan’s hypothetical example). These sums are justified if the denunciation leads to the conviction of a cartel (and the imposition of a fine) that could not have been detected without the denunciation.

86. It is true that whistleblower remuneration policy can have an undesirable long-term effect on business ethics: indeed, the denunciation of a fraudulent practice is no longer guided by an intrinsic motivation ("not breaking the law") but extrinsic, based on the lure of gain. However, it seems, in the light of the American experience, that extrinsic motivation does not replace but complements intrinsic motivation: whistleblowers initially reported the fraud to their employer without the latter drawing the consequences. Moreover, the potentially perverse effects of this type of system (exaggerated or fraudulent whistle-blowing) can be minimised by corroborating the whistle-blowing with information obtained through leniency programmes once the investigation is launched (Stephan [2014]).

Conclusion

87. Based on a cost-benefit analysis conducted at the individual level, we have shown that a manager’s incentives to form a cartel are based on specific factors, including a strong behavioural and psychological dimension. They may converge or diverge with those of shareholders. If shareholders want to deter the commission of cartel practices within their company, they can mobilize different levers, such as the terms of executive compensation. But internal tools may prove insufficient, given the asymmetry of information, thus justifying government intervention in addition to that of shareholders. When shareholders and managers display a convergence of interests to violate antitrust laws, it is the lack of internal incentives within the company that mainly justifies the implementation of public policies targeting individuals. These public policies can be based on various levers, including the dissemination of a culture of competition, stigmatizing cartels, or the implementation of criminal sanctions against managers. 

Bibliography

ABRANTES-METZ R., SOKOL D. (2015), Antitrust Corporate Governance and Compliance, in The Oxford Handbook of International Antitrust Economics, vol. 2, R. D. Blair and D. (2000), Antitrust Corporate Governance and Compliance, in The Oxford Handbook of International Antitrust Economics, Vol. 2, R. D. Blair and D. Sokol (eds.), Oxford University Press, pp. 586-618.

ALAWI S. (2014), Corporate Governance and Cartel Formation", Working Paper, 2nd Economics & Finance conference, Vienna, 36 p.

ANDERGASSEN R. (2008), High-powered incentives and fraudulent behavior: Stock-based versus stock option-based compensation, Economics Letters, 101(2), 122-125.

ARMSTRONG M., HUCK S. (2014), Behavioral economics and antitrust, in The Oxford handbook of international antitrust economics, vol. 1, R. Blair and D. Sokol (eds.), Oxford University Press, pp. 205-228.

AUBERT C., REY P., KOVACIC W. (2006), The Impact of Leniency Programs On Cartels, International Journal of Industrial Organization, 24(6), 1241-1266.

BEATON-WELLS V., HATES F. (2010), The Australian conversion: how the case for cartel criminalization was made, New Journal of European Criminal Law, 1(4), 499-521.

BEBCHUK L., FRIED J. (2003) Executive compensation as an agency problem, Journal of Economic Perspectives, 17(3), pp. 71-92.

BEBCHUK L., COHEN A., SPAMANN H. (2010) The Wages of Failure: Executive Compensation at Bear Stearns and Lehman 2000-2008, Yale Journal on Regulation, vol. 27, pp. 257-282

BECKER G. (1968), Crime and Punishment: An Economic Approach, Journal of Political Economy, 76(2), 169-217.

BECKMANN D., MENKHOFF L. (2008), Will Women be Women? Analyzing the Gender Difference Among Financial Experts, Kyklos, 61(3), 364-384.

BERGMAN H., SOKOL D. (2015), The Air Cargo Cartel: Lessons for Compliance, in Anti-Cartel Enforcement in a Contemporary age - Leniency religion, C. Beaton-Wells and C. Tran (eds.), Hart Publishing, 330 pp.

BERGSTRESSER D., PHILIPPON T. (2006), CEO incentives and earnings management, Journal of Financial Economics, 80(3), 511-529.

BUCCIROSSI P., SPAGNOLO G. (2008), Corporate Governance and Collusive Behavior, in Issues in Competition Law and Policy, vol. 2 (ABA Section of Antitrust Law), pp. 1219-1240.

CABANTOUS L., HILTON D. (2006), De l’aversion à l’ambiguïté aux attitudes face à l’ambiguïté. Les apports d’une perspective psychologique en économie, Revue économique, 57(2), 259-280.

CAMERER C., WEBER M. (1992), Recent developments in modeling preferences: Uncertainty and ambiguity, (Journal of Risk and Uncertainty, 5(4), 325-370.

CARSON T. (2006), Self-Interest and Business Ethics: Some Lessons of the Recent Corporate Scandals, Journal of Business Ethics, 43(4), 389-394.

CSERES K., SCHINKEL M., VOGELAAR F. (2006), Law and Economics of Criminal Antitrust Enforcement: An Introduction, in Criminalization of Competition Law Enforcement - Economic and Legal Implications for the EU Member States, Edward Elgar Publishing, 368 pp.

COATES J., GURNELL M., SARNYAI Z. (2010), From molecule to market: steroid hormones and financial risk-taking, Philosophical Transactions of the Royal Society B, Biological Sciences, 365(1538), 331-343.

COMBE E. (2016), La politique de la concurrence, La Découverte, coll. "Repères", 128 p.

COMBE E., MONNIER C. (2011), Fines Against Hard Core Cartels in Europe: The Myth of Over Enforcement, Antitrust Bulletin, 56(2), 235-275.

COMBE E., MONNIER C. (2012), Les cartels en Europe : une analyse empirique, Revue française d’économie, 27(2), 187-226.

CONNOR J., LANDE R. (2012), Cartels as Rational Business Strategy: Crime Pays, Cardozo Law Review, 34, 455-462.

CONYON M. (2006), Executive Compensation and Incentives, Academy of Management Perspective, 20(1), 25-44.

CROCKER K., SLEMROD J. The Economics of Earnings Manipulation and Managerial Compensation. The RAND Journal of Economics, 38(3), 698-713.

CROSON R., GNEEZY U. (2009), Gender Differences in Preferences, Journal of Economic Literature, 47(2), 1-27.

DAMGAARD M., DUKE C., HUCK S., WALLACE B. (2012), The impact of competition interventions on compliance and deterrence: A controlled economic experiment, London Economics for the OFT, 81 pp.

DENIS D. HANOUNA P., SARIN A. (2006), Is There a Dark Side to Incentive Compensation? Journal of Corporate Finance, 12(3), 467-488.

FLECKINGER P., LAFAY T., MONNIER C. (2013), Executive compensation and corporate fraud risk, Economic Review, 64(3), 457-467.

GILLIAN S., HARTZELL J., PARRINO R. (2009) Explicit versus implicit contracts: evidence from CEO employment agreements, Journal of Finance, 64, pp. 1629-1655

GLAESER E., LAIBSON D., SCHEINKMAN J., SOUTTER C. (2000), Measuring Trust, Quarterly Journal of Economics, 115(3), 811-846.

GOLDMAN E., SLEZAK S. (2006), An equilibrium model of incentive contracts in the presence of information manipulation. Journal of Financial Economics, 80(3), 603-626.

GONZALEZ T. A., SCHMID M. (2012), Corporate Governance and Antitrust Behavior, Working Paper, University of St. Gallen, 25 p.

GONZALEZ, T. A., SCHMID, M., YERMACK D. (2014), "Smokescreen: How Managers Behave When They Have Something to Hide," NYU Working Paper No. FIN-13-002.

HAMAGUCHI Y., KAWAGOE T., SHIBATA A. (2009), Group size effects on cartel formation and the enforcement power of leniency programs, International Journal of Industrial Organization, 27(2), 145-165.

HAMMOND S. (2010), The Evolution of Criminal Antitrust Enforcement Over the Last Two Decades, speech to the National Institute on White Collar Crime, http://www.justice.gov/atr/public/speeches/255515.htm.

HAN M. A. (2010), Short-Term Managerial Contracts Facilitate Cartels, Amsterdam Center for Law and Economics, Working Paper No. 2010-16. HARRINGTON J. (2006), How Do Cartels Operate, Foundations and Trends in Microeconomics, 2(1), 1-105.

JENSEN, M., MECKLING W. Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics 3(4), 305-360.

JONES A., WILLIAMS R. (2014), The UK Response to the Global Effort Against Cartels: Is Criminalization Really the Solution? Journal of Antitrust Enforcement, 2(1), 100-125.

KAPLAN S., B. MINTON B. (2012), How Has CEO Turnover Changed? International Review of Finance, 12(1), 57-87.

KOROBKIN R., ULEN T. (2000), Law and behavioral science: removing the rationality assumption from law and economics, California Law review, 88(4), 1051-1144.

KOVACIC W. (2016), Criminal enforcement of competition law: Implications of US experience, Concurrences No. 2, pp. 33-40.

KRAY L., HASELHUHN M. (2012), Male pragmatism in negotiators’ ethical reasoning, Journal of Experimental Social Psychology, 48(5), 1124-1131.

LEE G., FARGHER N. (2013), Companies’ Use of Whistle-Blowing to Detect Fraud: An Examination of Corporate Whistle-Blowing Policies, Journal of Business Ethics, 114(2), 283-295.

MEIER-PESTI K., PENZ E. (2008), Sex or Gender? Expanding the sex-based view by introducing masculinity and femininity as predictors of financial risk taking. Journal of Economic Psychology, 29(2), 180-196.

MILGRAM S. (1963), Behavioral study of obedience, The Journal of Abnormal and Social Psychology, 67(4), 371-378.

O’KANE M. (2011), Does prison work for cartelists? The view from behind bars, Antitrust Bulletin, 56(2), 483-500.

PODOLNY J., SCOTT MORTON F. (1999), Social status, entry and predation: the case of British shipping cartels 1879-1929, Journal of Industrial Economics, 47(1), 41-67.

PWC (2016), Global Economic Crime Survey, Adjusting the Lens on Economic Crime Preparations brings opportunity back into focus, http://www.pwc.com/gx/en/economic-crime survey/pdf/GlobalEconomicCrimeSurvey2016.pdf

PWC and London Business School (2015), Stand out for the right reasons, June, http://pwc.blogs.com/files/why-you-cant-scare-bankers-into-doing-the-right-thing.pdf.

ROBINSON R., LEWICKI J., DONAHUE E. Extending and testing a five factor model of ethical and unethical bargaining tactics: introducing the SINS scale. Journal of Organizational Behavior, 21(6), 649-664.

SCHMIDT C. (1996), Risk and uncertainty: A knightian distinction revisited, in Uncertainty in Economic Thought, London, Edward Elgar, 256 pp.

SCHWAB S., THOMAS R. (2006), An Empirical Analysis of CEO Employment Contracts: What Do Top Executives Bargain For? Washington and Lee Law Review, 63, 231.

SOKOL D. (2012), Cartels, Corporate Compliance and What Practitioners Really Think About Enforcement, Antitrust Law Journal, 78, 201-240.

SOLTANI B. (2014), The Anatomy of Corporate Fraud: A Comparative Analysis of High Profile American and European Corporate Scandals, Journal of Business Ethics, 120(2), 251-274.

SPAGNOLO G. (2000) Stock-Related Compensation and Product-Market Competition, The RAND Journal of Economics, No. 25.

SPAGNOLO G. (2005), Managerial incentives and collusive behavior, European Economic Review, 49(6), 1501-1523

STEPHAN A. (2010), See no evil: Cartels and the limits of antitrust compliance programs, The Company Lawyer, 31(8), 231-239.

STEPHAN A. (2011), The Battle for Hearts and Minds: The Role of the Media in Treating Cartels as Criminal, in Criminalizing Cartels: A Critical Interdisciplinary Study of an International Regulatory Movement, C. Beaton-Wells and A. Ezrachi (eds.), Hart Publishing, Oxford, 472 p.

STEPHAN A. (2012) Survey of public attitudes To Price-Fixing and Cartel Enforcement in Britain, CCP Working Paper 07-12.

STEPHAN A. (2014), Is the Korean Innovation of Individual Informant Rewards a Viable Cartel Detection Tool? CCP Working Paper 14-3.

STEPHAN A. (2015), Survey of Public Attitudes to Price Fixing in the UK, Germany, Italy and the USA, Centre for Competition Policy UEA Law School University of East Anglia, CCP Working Paper 15-8.

STEPHAN A. (2016), Is there public support for cartel criminalization? Concurrences No. 2, pp. 40-44.

STEFFENSMEIER D., SCHWARTZ J., ROCHE M. [2013], Gender and Twenty-First-Century Corporate Crime: Female Involvement and the Gender Gap in Enron-Era Corporate Frauds, American Sociological Review, 78(3), 448-476.

STUCKE M. (2011), Am I a Price-Fixer? A Behavioral Economics Analysis of Cartels, in Criminalizing Cartels: A Critical Interdisciplinary Study of an International Regulatory Movement, C. Beaton-Wells and A. Ezrachi (eds.), Hart Publishing, Oxford, 472 p.

VAN STAVEREN I. (2014), The Lehman Sisters hypothesis, Cambridge Journal of Economics, 38(5), 995-1014.

VIROS D. (2016), Individual criminal sanctions in France, Concurrences No. 2, pp. 24-27.

WAGNER-VON PAPP F. (2016), Individual sanctions for competition law infringements: Pros, cons and challenges, Concurrences No. 2, pp.14-23.

WILS W. (2001), Does the effective enforcement of Articles 81 and 82 EC require not only fines on undertaking but also individual penalties, in particular imprisonment, in European Competition Law Annual 2001: Effective Private Enforcement of EC Antitrust Law.

WILS W. (2008), Efficiency and Justice in European Antitrust Enforcement, Hart Publishing, Oxford and Portland Oregon, 206 p.