As opposed to other types of market conduct (such as, e.g., vertical distribution agreements or the unilateral use of market power), cartels are widely perceived to have few if any redeeming features. In recent years, one can clearly detect a firm commitment from antitrust enforcers around the globe to pursue rigorously the investigation, detection and prosecution of cartel activity.
Aligned with this development is a growing tendency in a wide variety of jurisdictions to hold individuals accountable for the creation and implementation of cartels, including through use of criminal law.
Unfortunately, the employment of criminal cartel sanctions is not without its problems. This paper, available here, seeks to evaluate some inherent problems associated with the use of criminal sanctions for cartel conduct to deter anticompetitive behaviour. It is structured as follows:
Section B outlines the deterrence-based justification for criminal cartel sanctions.
The primary rationale for the criminal cartel sanctions is economic deterrence. Unlike retribution, deterrence does not concern itself with punishment for punishment’s sake. The theory of deterrence holds that punishment can only be justified if it leads to the prevention or reduction of future crime.
Economic deterrence theory is a strand of deterrence theory that builds on economic efficiency and welfare concepts. It views criminal punishment as a method of maximising utility, to be employed only when the disutility of its imposition is less than the utility to society secured by its deterrent effect. Economists will usually look to the margins in order to determine the efficient amount of crime enforcement. Efficiency is obtained, and welfare maximized, where the marginal benefit of punishment is equal to its marginal cost.
The (economic) deterrence-based cartel criminalisation argument is essentially a two-step argument. With the first step, one demonstrates that an (administrative or criminal) fine on a company is unlikely to deter that company from engaging in cartel activity (or from encouraging its employees to engage in such activity). An antitrust authority must ensure that the level of the fine imposed for cartel activity is such that there will be a disincentive to engage in that activity. More specifically, given that cartel activity will rarely produce efficiencies, the authorities should focus on the expected gain from the cartel and, accordingly, set the fine at least equal to the expected financial benefit obtained from cartel activity divided by the probability of being caught and prosecuted.
The problem with this approach is that the size of the fine required to deter the company from cartel activity is far too large. Using reliable statistics for the average cartel mark-ups (of detected cartels), their average length, along with conservative estimates of the probability of detection and prosecution, one can demonstrate that the optimal fine should be in the region of 150% of annual turnover (according to Wils), if not more (Werden posits that the fine would have to amount to 200% of annual turnover). As a result, cartel fines cannot reach the deterrent-level, as they exceed the company’s ability to will in most cases pay. In fact, the available literature presents a figure of 18% as the percentage of companies convicted of cartel activity that would have sufficient resources to be able to pay the optimal fine.
The second step involves demonstrating that the deterrence gap can only be overcome by imposing individual criminal sanctions (i.e. custodial sentences) upon convicted cartelists. Focusing on individuals makes sense from an enforcement perspective. However, the point of criminalisation is that such individual sanctions should not be mere monetary (i.e. financial) sanctions, such as fines. The reason for this is that the corporation (which ultimately benefits from cartel activity) may wish to incentivise such activity among its staff, and may simply indemnify any sanctioned individual by paying the financial sanction for that individual. What needs to be found, so the argument runs, is a non-indemnifiable sanction that would clearly push the potential individual cartelist’s cost-benefit analysis in favour of compliance and away from cartel activity. This is where custodial sentences come in: they are widely seen by pro-criminalisation advocates as non-indemnifiable sanctions that are capable of pushing rational business executives away from cartel activity. Underpinning this assessment is the assumption that for business people, as opposed to hardened criminals, serving time in prison is to be avoided at all costs.
Section C critically analyses two important problems that arise when criminal sanctions (i.e., custodial sentences) are used to deter cartel activity.
A first problem is the difficulty of securing efficient competition law enforcement when criminal cartel sanctions are employed. Economic deterrence theory requires one to consider, in addition to the ability of a criminal sanction to deter, the actual costs of enforcement when deciding whether to criminalise a particular behaviour. Interestingly, the main focus of those who wish to advocate criminal cartel sanctions for (economic) deterrence purposes has been on the effectiveness per se of a custodial sentence in achieving deterrence. However, given that the use of criminal sanctions also involves costs, and that costs have a bearing on the efficiency of enforcement, such costs also need to be considered in order to determine whether economic deterrence theory can justify in principle the introduction and maintenance of criminal cartel sanctions.
What needs to be demonstrated is that criminal cartel sanctions can be imposed where the marginal benefit is equal to its margin cost. This is an extremely difficult, if not impossible, inquiry on which to come to a firm conclusion. There no systematic empirical evidence available to prove whether the marginal benefit of introducing sanctions against individuals (in the form of less harm from cartel activity) exceeds the additional costs that a system of criminal sanctions entails (including the costs of prosecution as well as of administering a prison system). There appears to be agreement that it would be virtually impossible to generate the relevant data. The best one can hope to do in this situation is to consider the arguments that underpin the assertion that custodial sentences for cartel activity are at least capable of generating more benefits than costs.
This means that the answer to the question of whether a jurisdiction should introduce criminal cartel sanctions is not a precise one, and that the debate on the optimal anti-cartel strategy is unlikely to cease. This is reflected in the growing number of academic conferences and research dedicated to cartel criminalisation. What is clear is that any jurisdiction contemplating cartel criminalisation should be aware that such a project is ‘a long-term, front-end loaded investment’ which may require a wide range of criminal prosecutions to ‘give birth to a new culture’. In any event, the main arguments for and against criminalisation are as follows:
- An important argument is that if cartels (or at least the most serious and harmful cartels) are deterred through the use of custodial sentences, then the positive impact on efficiency could be immense because cartels cause great harm. The relevant detriment caused by cartels is not only the overcharge – which can theoretically be recovered under private enforcement – but also the deadweight loss (allocative inefficiency). Fortunately for advocates of antitrust criminalisation, allocative inefficiency in the context of cartels is not trivial – it is estimated to be between 3% and 50% of the overcharge.
- As regards costs of cartel criminalisation, they include those involved in investigating, prosecuting and incarcerating individual cartelists, including the cost of taking otherwise productive individuals out of the economy. These costs too may be immense. Fortunately, there are a number of techniques that can be used to try to keep such costs at a minimum. First, one can reduce the costs involved in incarceration by ensuring that only short terms of imprisonment are imposed. Second, prosecutors could focus solely on the most serious of cartels, in order to send out a deterrent message to the most destructive elements in the economy without incurring unnecessary and frivolous costs. Third, the successful use of plea-bargaining, leniency/immunity programmes and bounties has the potential to reduce significantly the costs involved in investigating cartels to a criminal standard. Fourth, a criminalised jurisdiction could impose cost orders upon convicted cartelists, as occurs in Ireland. The trouble, however, is that the above-identified mechanisms for reducing costs in the context of cartel criminalisation suffer from a number of drawbacks. First, the use of short sentences to keep the costs of criminal antitrust enforcement low runs counter to efforts to create and maintain a moral norm against cartel activity. Secondly, the use of plea-bargaining in the criminal law context is a controversial cost-saving mechanism for many jurisdictions. Third, some of the methods identified are already employed in administrative regimes (e.g., leniency programmes and bounties). To operate as a mechanism for further reductions in costs, they must therefore demonstrate additional advantages in the context of the employment of criminal punishment – and these seem to arise mainly, if not solely, as regards personal criminal immunity.
The second main problem with criminalising cartels is the need for connecting the criminalised cartel activity to morally wrongful behaviour. It is true that a finding of ‘moral wrongfulness’ is not required in order to support a deterrence-based criminalisation argument, and that deterrence theory supports the adoption of a morally-neutral criminal cartel offence. However, even if criminal cartel sanctions are pursued for the purposes of deterrence (as opposed to, say, retribution), it does not follow that the link between criminalised cartel activity and morally wrongful activity is irrelevant. Establishing a link between criminalised cartel activity and morally wrongful behaviour is in fact very important because, if criminal law is applied to morally-neutral (cartel) activity, it may then be perceived as being unjust.
In effect, it is widely held that criminal law should not be understood merely as ‘a device for promoting particular economic and social ends’, but rather as a law ‘directed to moral standards of society’. In the absence of such a restraint, so the argument runs, the moral authority of the law can be undermined, the meaning of criminality may change, and criminal law may begin to lose its legitimacy. Against this, it can be argued that criminal law is not solely used to reflect the morality of a given society, but that it can be used to create a moral reaction to behaviour deemed by the lawmaker to be objectionable. However, this view has a number of problems, which means it would be advisable to ensure that the criminalised conduct lines up generally (if crudely) with general perceptions of morally wrongful behaviour.
There is an additional advantage to this approach: if the criminal cartel offence is perceived to be legitimate (due to the fact that it reflects society’s view of the moral wrongfulness of cartel activity) then it is likely that compliance with that law will be more pronounced. From this standpoint, the difficulty lies in creating a criminal cartel offence that inevitability captures the ‘criminality’ of cartel activity. Authors have identified three norms against which the moral wrongfulness of cartel activity can be judged: the norms against stealing, deception and cheating. Central to this argument is the claim that consumers assume that business people do not unlawfully engage in anticompetitive practices. Fortunately for advocates of cartel criminalisation, there is some (limited) empirical support for this claim.
This is a short yet comprehensive overview of debates about the justifications and conditions for criminalising cartel conduct. It contains a very good overview of the topic and of the literature for anyone interested in the topic, and provides a good introduction to those for whom the topic is new.