Melbourne University Law Review
|
|
Home
| Databases
| WorldLII
| Search
| Feedback
Melbourne University Law Review |
|
KAREN YEUNG[*]
[Despite the shortcomings of the monetary penalty as a form of punishment for corporate offenders, it is widely used in regulatory law and is likely to continue to be a feature of regulatory design. But how should the quantum of regulatory penalties be determined? In order to answer this question, this article critiques and contrasts two theories of punishment which have been expounded in relation to traditional crimes: deterrence theory (which claims that the purpose of punishment is to deter would-be offenders from committing crimes) and desert theory (which claims that the primary purpose of punishment is to communicate social blame or censure for the commission of crimes). Although commercial regulation is primarily concerned to deter undesirable behaviour, rather than to express moral condemnation of that behaviour, it is nevertheless suggested that a fair and effective regulatory penalty regime should combine the essential features of both models. Penalties should be set at a level which is sufficiently high to deter future contraventions of the law, provided that any given penalty award is not unfairly disproportionate to the seriousness of the offence. The final part of the paper reviews the approach taken by Australian courts in quantifying penalties for contraventions of Australian competition law in light of the suggested analytical framework.]
In ACCC v Pioneer Concrete (Qld) Pty Ltd,[1] three companies were ordered by the Federal Court to pay penalties of A$6.6 million each for fixing the price of pre-mixed concrete in Queensland. In addition, five of the companies’ senior employees were penalised A$50,000 each, and one employee was penalised A$100,000 for knowing involvement in the price fix.[2] The penalties ordered by the Court are the highest on record for contraventions of the prohibitions contained in Part IV of the Trade Practices Act 1974 (Cth) (‘TPA’), which forms the basis of Australian competition law. The size of the penalty awards, and a number of subsequent penalty awards of considerable magnitude, have caused public disquiet.[3] Some commentators, particularly those from the industry arena, have decried the levying of exorbitant penalties, while others claim that heavy financial penalties are essential in order to ensure that competition regulation is effective in protecting the welfare of consumers by maintaining the integrity of the competitive process.[4] Indeed, the current Chairman of the Australian Competition and Consumer Commission (‘ACCC’), Australia’s public competition and consumer law regulator, has suggested that even heavy financial penalties may be insufficient to deter unlawful anti-competitive practices, and has therefore suggested that further consideration should be given to the introduction of criminal sanctions for breach of Part IV of the Act.[5] The purpose of this paper is to consider how regulatory penalties should be quantified and, in particular, how penalties for contravening competition law should be quantified.
Although this analysis focuses on penalties for competition law violations, it is equally applicable to other areas of the law where financial penalties are payable for regulatory offences. As the twentieth century draws to a close, economic and social activity within industrialised societies is more heavily regulated than it has ever been since the onset of industrialisation.[6] The widespread adoption of policies of ‘deregulation’ by governments of many Western industrialised economies since the 1970s may have reduced the extent of direct state intervention in social and economic activity, but not its entire withdrawal.[7] State regulation continues to apply to a diverse spectrum of economic and social activities, ranging from health and safety, consumer protection, environmental pollution and the use of reproductive technology, through to financial services, public utilities and so-called ‘essential facilities’. Although the institutional design of regulatory regimes can vary considerably in the level of state intervention involved and in their reliance on either, or both, positive incentives to encourage the desired behaviour (‘carrots’) or through the imposition of sanctions for undesirable behaviour (‘sticks’), there is no doubt that the use of sanctions to deter undesirable behaviour (sometimes referred to as ‘command-and-control’ regulation) is likely to continue to be an enduring feature of regulatory design. The effectiveness of regulation in bringing about the desired form of behaviour by the regulated parties depends crucially, however, on both the level of penalty for non-compliance and the willingness of regulators to impose penalties for non-compliance with the law. This paper is concerned with the first of these issues: how should the quantum of regulatory penalties be determined?[8]
In order to answer this question, the first part of this examination draws upon theories of punishment expounded in the criminal sentencing context which have sought to explain not only why the law is justified in punishing those who commit crimes, but also why the quantum of punishment should be fixed at a particular level. Two alternative theoretical approaches are examined, each resting on entirely different philosophical foundations: the ‘deterrence’ approach as expounded by advocates of the economic analysis of law, which seeks to ‘price’ unlawful conduct in order to minimise social costs arising from unlawful conduct; and the ‘desert’ approach, which takes the view that those who break the law ‘justly deserve’ punishment, and therefore, the sanctions for violating the law should be commensurate with the nature of the wrongfulness of the sanctioned conduct. In other words, while the deterrence theory regards punishment as the ‘price’ of unlawful conduct, the desert theory regards punishment as communicating state censure and disapprobation of the offender.[9] It is argued here that neither model can be regarded as superior to the other. Both suffer from intractable problems, but are nevertheless capable of generating reasonable, practical outcomes — an effective penalty regime should be assessed both by its ability to deter undesirable conduct, while also ensuring that the penalties imposed in any individual case are fair and just. Thus, a regulatory penalty regime should be a hybrid, reflecting components of both models, rather than adhering to either model in its ‘pure’ form.
The second part of this analysis then attempts to illustrate how these theoretical approaches to penalty setting apply to the quantification of penalties for breach of competition law. The aim of competition regulation is to facilitate competitive markets based on the view that competition will lead to the efficient allocation of resources in society, thereby benefiting consumers. The Australian approach to quantifying competition law penalties is then examined in light of the Australian regulatory framework for competition regulation and the principles developed by the Australian Federal Court in order to identify whether these principles conform to either theoretical model. An examination of the case law reveals disagreement between individual judges concerning whether the proper aim of penalty setting for Australian competition law violations is ‘deterrence’ or ‘punishment’. This tension is largely attributable to whether the deterrence or desert model is the preferred theoretical basis for penalty setting. Although individual judges have attempted to articulate the theoretical basis for their decisions, it appears that they are not adhering exclusively to one or other of the two models, but are in fact adopting a combination of both models in a manner which largely conforms to the hybrid model advocated here.[10]
Because the analytical approach taken in this paper is not specific to competition regulation, but can be applied to any form of economic and social regulation which relies upon the use of legal sanctions for non-compliance, it is important to clarify what is meant by the term ‘regulatory law’. Although the term ‘regulation’ can be used to refer to any form of behavioural control, a narrower definition is used here. Regulation refers to state intervention in economic and social activity, aimed at directing or encouraging behaviour valued by the community, so as to facilitate the pursuit of collectivist goals which might not otherwise be realised, and which constitutes a form of ‘public law’ in the sense that it is generally for the state (or its agents) to enforce the obligations which cannot be overreached by private agreement between the parties concerned.[11] Thus, traditional areas of the criminal law do not fall within the domain of regulatory law because criminal activities are not ‘valued by the community’. What distinguishes regulatory offences from traditional crimes is that regulatory offences do not generally attract the same degree of moral condemnation typically associated with traditional crimes.[12] Regulated conduct is prohibited not primarily because of its moral improbity, but because it undermines the collectivist goals which the regulatory scheme seeks to facilitate.[13] Likewise, the area of law typically encompassed within the private law of obligations, such as laws relating to contract, property and tort, fall outside the realm of regulatory law because they are concerned to empower individuals and groups to pursue their own welfare goals in a market system of organisation, subject only to basic restraints, rather than a system of public law to implement collectivist goals. By contrast ‘regulatory law’, when used in the present context, denotes the law which implements the collectivist system which has a directive function — citizens are compelled by a superior authority, the state, to behave in particular ways with the threat of penal sanctions if they do not comply.[14]
In order to examine the theoretical basis for determining the quantum of penalties for violating the law (whether regulatory law or otherwise), it is first useful to examine, in general terms, the theoretical basis for punishing those who break the law (‘the offender’), because any attempt to answer the question ‘how much should we punish?’ is in part determined by the separate but related question, ‘why do we punish?’.[15] Although this paper is concerned with regulatory offences which, unlike traditional criminal offences, carry little or no connotations of moral improbity, both criminal and regulatory offences involve the application of the coercive power of the state to visit some form of ‘hard treatment’ on the offender.[16] Accordingly, the various theories which have been advocated in order to justify the imposition of punishment for criminal offences are directly relevant to the justification of punishment for regulatory offences.
In the criminal law context, five possible theoretical justifications for punishment have been identified: deterrence (which seeks to deter people from committing crimes); rehabilitation (which seeks to achieve rehabilitation of the offender); incapacitation (which seeks to incapacitate offenders so as to render them incapable of reoffending for substantial periods of time); desert (which seeks to ensure that offenders receive their ‘just deserts’ for crime committed); and restoration or reparation (which is concerned that victims of crime receive justice).[17] Three of these theories are prima facie unsuited to penalty setting in the context of regulatory law: rehabilitation seems to have little application to economic and social regulation where corporations are the primary offenders;[18] incapacitation is usually confined to particular groups, such as ‘dangerous’ offenders and career criminals and thus has little application to economic and social regulation; and restoration is the primary function of private actions to enforce regulatory law, rather than the basis for determining the quantum of punishment payable to the state in proceedings commenced by a public regulator.[19] This leaves the competing theories of deterrence and desert as possible theoretical bases for determining the quantum of penalties for regulatory offences.
Deterrence as a justification for the punishment of crime is based on the notion that punishment is warranted by reference to its crime prevention consequences.[20] The deterrence approach to punishment has historically been associated with classical utilitarian philosophy, famously advocated by the English jurist and philosopher Jeremy Bentham in the nineteenth century, based on utilitarian goals of deterring future unlawful behaviour.[21] According to Bentham, punishment was justified primarily in order to prevent others from committing the same offence, the disutility visited on the offender being outweighed by the utility to society arising from the deterrent effect of punishment, thereby preventing future crime.
A modern variant of the utilitarian theory of deterrence is advocated by proponents of the economic analysis of law.[22] An economic approach to sentencing is primarily concerned to minimise the social costs caused by crime. Because crime imposes costs on society both directly (in the damage wrought by criminal conduct) and indirectly (in the costs incurred by society in apprehending, convicting and punishing offenders), the aim of the law should be to set punishment at a level which deters crime to the level at which the total costs to society are minimised. Although not all advocates of the deterrence theory of punishment necessarily support the economic theory of crime deterrence, it is focused upon in this paper because it provides specific guidance for determining the quantum of punishment and therefore provides a theoretical basis for the quantification of regulatory penalties.
The economic variant of the deterrence approach to penalty setting (the ‘deterrence model’) regards conduct which violates the law as a commodity which may be effectively ‘purchased’ by requiring the offender to pay a penalty for breaking the law. The deterrence model depends crucially on the assumption that individuals and firms are rational actors who act in the pursuit of self-interest in order to maximise their individual welfare. In other words, it assumes that, when deciding whether to engage in a particular activity, each actor first calculates whether the estimated benefits to the actor of that activity will exceed the estimated costs to the actor of engaging in that activity. If so, then the actor will engage in the activity. Thus, if the estimated benefits of violating the law outweigh the estimated costs of engaging in unlawful behaviour, then the actor will rationally decide to break the law.[23]
Because conduct which violates the law may cause harm to others, thereby imposing costs on society, violating the law is to be discouraged. The imposition of penalties for violation of the law has the effect of increasing the estimated cost of violation. Penalties thus create financial incentives which deter firms and individuals from engaging in unlawful conduct.[24] A penalty is simply the ‘price’ of violating the law.
The fundamental aim of the deterrence model is the pursuit of economic efficiency. Economics is concerned with allocating society’s resources efficiently, based on the view that by so doing, the overall welfare of society is maximised. The essential decision-making criteria in economics is based on a simple cost-benefit analysis: if the benefits to society of a particular activity outweigh the costs to society, then the activity is efficient and should be encouraged because it enhances the overall welfare of society. If this analytical approach is applied to penalty setting, then the correct quantum of penalty is that which will deter unlawful conduct to an ‘efficient’ level.
Assuming that detecting and penalising violators is costless, then the efficient level of unlawful conduct occurs where the net costs to society arising from the unlawful conduct are minimised.[25] If the unlawful conduct in question has no beneficial effects on society while imposing costs on society, or if the costs of the unlawful conduct always outweigh its benefits, then the efficient level of unlawful conduct is zero.[26] Therefore, the aim of penalty setting should be ‘absolute deterrence’ of the unlawful conduct in question. On the other hand, if the unlawful conduct, although costly to society, also confers some benefits on society, then the aim is to set the quantum of penalties at a level which ensures that unlawful conduct which imposes a net cost on society is deterred but conduct which results in a net benefit to society is not deterred. Therefore, the aim of penalty setting in such cases should be ‘optimal’ rather than ‘absolute’ deterrence. These two possible interpretations of the effects of unlawful conduct reflect the two different policy prescriptions or variants of the deterrence model: the ‘unlawful gain’ model and the ‘injury to others’ model.
The ‘injury to others’ variant of the deterrence model of penalty setting aims to achieve ‘optimal’ deterrence of unlawful conduct. Optimal deterrence seeks to deter only inefficient violations of the law, while efficient violations of the law should not be deterred. This variant of the deterrence model assumes that conduct which violates the law may have some beneficial effects which may potentially outweigh its costs.[27] If the net gain to society arising from the unlawful conduct in any given instance exceeds the harm to society caused by that conduct, then the conduct is welfare-enhancing and should not be deterred. For example, a factory’s production operations may pollute a nearby river in breach of environmental laws, thereby causing harm to others. However, the factory’s production is also beneficial to society, because it generates profit for the factory owner and may provide consumers with goods which would not otherwise be available. If that benefit exceeds the harm so caused, then the factory should not be deterred from operating, albeit in a manner which pollutes the river. It is important to note that the benefits accruing to the offender as a result of the unlawful conduct can legitimately be characterised as a gain to society as a whole — the mere fact that the gains are generated as a result of unlawful conduct is immaterial.[28]
In order to achieve its objective of optimal deterrence, the penalty for violating the law should be set at a level which equals the total cost of the harm to others caused by the unlawful conduct. When an actor’s activity imposes costs on others which are not borne by the actor, these costs are known as ‘externalities’.[29] Legal regulation is required in order to ensure that these externalities are attributed to the actor. By setting the penalty level to reflect the injury caused to others, the imposition of a penalty causes the actor to ‘internalise’ the cost of the harms to others caused by the actor’s conduct. In other words, requiring the actor to pay a penalty for violating the law ensures that the actor bears the full cost of its unlawful action.
Because the unlawful activity may generate social benefits, the injury to others variant of the deterrence model is concerned to avoid the problem of ‘over-deterrence’ which arises if penalties are set at an excessively high level. Over-deterrence should be avoided because it results in the deterrence of unlawful behaviour which may be welfare-enhancing. Accordingly, it follows that the gain made by the violator and the ability of the violator to pay the penalty so levied are not relevant to the calculation of penalties. Taking account of the gain made by the violator in calculating the appropriate penalty would lead to over-deterrence, by deterring unlawful conduct which may generate a net gain to society. Taking account of the violator’s ability to pay would have the effect of over-deterring larger, more profitable firms, while under-deterring smaller, less profitable firms.[30]
The ‘unlawful gain’ variant of the deterrence model also seeks to deter firms and individuals from violating the law because of the harmful effects of such conduct. The unlawful gain model differs crucially from the ‘injury to others’ model, however, in that it aims to achieve ‘absolute’ deterrence of unlawful conduct, not merely ‘optimal’ deterrence. The aim of absolute deterrence follows from the premise that the unlawful conduct is either utterly without any redeeming social benefits, or the social benefits are always outweighed by the costs of unlawful conduct, and thus the total social costs are minimised by completely eliminating unlawful conduct.[31]
In order to achieve its objective of absolute deterrence, the penalty for violating the law should be set at a level which is at least equal to the net gain to the offender arising from the unlawful conduct.[32] By effectively disgorging the violator’s ill-gotten gain, this eliminates any incentive to violate the law. If the actor knows that there is no opportunity to gain from unlawful behaviour, then the rational actor will refrain from violating the law.
Because the unlawful gain variant of the economic model is based on the premise that the unlawful conduct in question always imposes a net cost on society, this model does not contemplate the possibility of over-deterrence arising from excessively high penalty levels.[33] The crucial determinant for setting the appropriate penalty level is the gain arising from the offender’s unlawful conduct. Accordingly, the harm to others caused by the unlawful conduct is irrelevant — even if the harm to others exceeds the gain to the offender, it is sufficient if the penalty is set at a level equal to the offender’s gain, for in so doing it eliminates any incentive to violate the law, thereby deterring all violations of the law. It also follows that the gain to the offender represents the minimum penalty level: a penalty which exceeds the gain to the offender is consistent with the unlawful gain model because it eliminates the incentive of potential offenders to act unlawfully.
The two variants of the deterrence model described above have assumed that the probability of detecting a violation of the law and levying a fine on the offender is 100 percent. It has also been assumed that there are no costs involved in enforcing the law. In other words, it is assumed that all violations of the law are detected without costs being incurred in the detection process. However, neither assumption is realistic. Accordingly, the deterrence model must be adjusted in order to take into account of the cost of enforcing the law (‘enforcement costs’) and the probability of detection.[34]
If the assumption of zero enforcement costs is relaxed, then the cost of enforcing the law must be factored into the determination of the optimal enforcement strategy required by the deterrence model. In relation to the ‘injury to others’ variant of the model, this is achieved by adding the enforcement costs to the penalty calculated under the assumption of zero enforcement costs, thus ensuring that those contemplating violating the law ‘internalise’ the enforcement costs involved in detecting and penalising the violation.[35] In relation to the ‘unlawful gain’ variant of the deterrence model, the existence of enforcement costs may affect the enforcement strategy adopted by the enforcer, rather than the quantum of penalty. Although the unlawful gain variant of the deterrence model should be applied to conduct which is always welfare-reducing so that it should be deterred absolutely, if the cost of harm caused by the conduct is less than the enforcement costs incurred in eliminating the conduct, then no enforcement action should be taken.
In addition, it is impossible in practice for all violations of the law to be detected and pursued, particularly given the limited resources available to enforcement agencies. In deciding whether or not to comply with the law, rational actors calculate whether the expected benefits of violating the law exceed the expected costs. Because the assessment is made ex ante, based on expected rather than the actual costs and benefits, the relevant costs to the actor of violating the law are based on the expected penalty from violation. Where the probability of detection for violating the law is less than 100 percent, this has the effect of lowering the actor’s expected costs of violating the law. It is therefore necessary to adjust the penalty according to the probability of detection. For example, if the optimum penalty assuming a 100 percent probability of detection is $1000, and if the probability of detection is 50 percent, the penalty must be increased to $2000 ($1000 ÷ 0.5). It follows that the same level of deterrence can be achieved by escalating the penalty level in response to a decline in the probability of detection.[36]
The preceding analysis has also assumed that offenders are risk neutral. However, if potential offenders are risk averse, they will be deterred by lower penalties, and thus the optimal penalty required to achieve the requisite level of deterrence should be reduced accordingly. Correlatively, if potential offenders are risk taking, they will need to face higher penalties before they will be deterred from violating the law, and thus the optimal penalty should be increased by the relevant risk aversion factor.[37]
Finally, the deterrence model of penalty setting assumes that legal error is avoided. In other words, the model assumes that legal rules are perfectly formulated and applied by courts so that all relevant conduct that inflicts harm on others is detected and penalised but that no benign conduct is penalised. However, some legal error is inevitable in practice, because legal decisions are inevitably based on incomplete information.[38] Thus, some conduct may be penalised although it may in fact be benign. Likewise, some conduct may escape penalty although it is in fact harmful. The possibility of legal error means that firms may substitute away from marginal conduct which is at risk of contravening the law if faced with the threat of penalties for contravention.[39] Legal error therefore results in the misallocation of resources and thus imposes costs on society. Moreover, the higher the penalty levels, the greater the costs arising from legal error as firms increasingly substitute away from marginal but benign activities.[40]
The desert theory of punishment differs radically from the deterrence theory due to its fundamentally different philosophical foundations. Desert theorists regard punishment for criminal conduct as a form of retributive justice. The essential concern of desert theory is that the offender should receive his or her ‘just deserts’ for violating the law.[41] There are several different approaches within desert theory which seek to explain why the law is justified in punishing those who commit crimes. For example, Moore argues that there is a fundamental intuitive connection between crime and punishment: we are justified in punishing because, and only because, offenders deserve it.[42] Andrew von Hirsch, one of the principal proponents of desert theory, relies on the communicative feature of punishment to explain why society is justified in punishing those who commit crimes and to argue that punishment primarily expresses blame or censure, thus conveying disapprobation of the offender and concurrently acknowledging to victims and potential victims that they have been wronged by criminal conduct.[43]
The principle of ‘just deserts’ not only provides the moral justification for punishment, but also provides the theoretical basis for determining the quantum of punishment. Desert theorists argue that the quantum of punishment must be determined by the principle of proportionality. The proportionality principle requires that the severity of the punishment must be commensurate with the seriousness of the offence.[44] According to von Hirsch, the requirement of proportionality provides the central role in determining the quantum of punishment because the severity of punishment communicates the degree of blame attributed to the crime. Hence, punishments should be ordered according to the degree of blameworthiness of the conduct — the more serious the conduct, the more severe the punishment.[45] Other desert theorists claim, however, that the principle of proportionality merely provides a limiting, rather than defining, principle to determine the quantum or severity of punishment. For example, Morris argues that proportionality is a limiting principle which allows the quantum of punishment to be based on other grounds provided that the quantum falls within the outer limits of proportionality: punishment must not be too severe so as to be oppressive, nor too lenient so as unduly to depreciate the seriousness of the offence committed. Although individual commentators have put forward several variants of the desert theory, for the purposes of this paper it will suffice to accept the core notion that punishment is a form of retributive justice which demands that the severity of the punishment should be commensurate with the seriousness of the offence.
Because the quantum of penalty must be proportionate to the seriousness of the offence, it is necessary to provide criteria which enable the ‘seriousness’ of offences to be determined. Desert theorists identify two key components to the concept of seriousness: harm and culpability.[46] Harm refers to the extent of the injury caused by the violation of the law. The greater the harm caused, the greater the appropriate penalty. Culpability refers to factors of intent, motive and the circumstances that bear on the degree of the actor’s blameworthiness.[47] It follows that violations of the law which are deliberate and calculated should be penalised more heavily than those which are committed carelessly or in ignorance of the law.
Although the two theoretical models described are primarily concerned with determining the severity of punishment for traditional criminal offences, they are general in nature and can be applied to any area of the law where financial penalties are payable for unlawful conduct. In this section, the strengths and shortcomings of each model are briefly compared and contrasted.
The most significant strength of the deterrence model is that it is theoretically capable of generating a specific quantum of penalty. Because the model is based on a series of theoretically quantifiable variables, it provides a non-arbitrary, principled and systematic approach to the practical problem of penalty setting. Accordingly, both variants of the deterrence model can encourage consistency in penalty setting and may reduce the appearance of arbitrariness.
Bilmes and Woodbury claim that the deterrence model is particularly valuable in acknowledging the need to factor the probability of detection into the calculation of penalty.[48] If violations of the law are difficult to detect, this justifies increasing the penalty. Factoring the probability of detection into the penalty setting equation does not, however, inexorably lead to the acceptance of the deterrence model of penalty setting. For example, it could plausibly be argued that the probability of detection is related to the ‘seriousness’ of the violation, which is a factor relevant to determining the penalty level under the desert model. Thus, the more covert the behaviour involved (and, therefore, the smaller the probability of detection), the greater the seriousness of the contravention, thereby justifying an escalation in the level of the penalty. Alternatively, the probability of detection could be used as a criterion for determining the allocation of enforcement resources, while utilising the desert model for penalty setting.[49]
Despite its appeal, the deterrence model suffers from several significant shortcomings, both in practice and in principle. Firstly, it is a behavioural model grounded on the crucial assumption of rationality. This assumption — that those who violate the law do so on the basis of an ex ante calculation that the expected benefits of violating the law outweigh the expected costs — may not accurately reflect reality.[50] In the criminal context, empirical studies have failed to provide reliable findings about the relative deterrent effects of various types and levels of penalty for various offences.[51] That said, the assumption of rationality may be more readily accommodated within the area of economic and social regulation which primarily concerns commercial behaviour by profit maximising firms rather than individuals.[52] But even in the regulatory context, empirical studies suggest that corporations are not solely driven by self-seeking individuals who are concerned exclusively with profit maximisation, and may be also motivated by non-financial concerns including a concern for social responsibility and respect for the rule of law.[53]
Secondly, while the deterrence model provides a theoretically coherent and complete basis for penalty setting, applying the model to real world circumstances generates intractable problems in practice. Quantifying variables such as loss, gain and probability of detection with any degree of precision may be impossible to achieve at reasonable cost. Indeed, precise objective calculation may be impossible — for example, calculating the extent of the harm arising from a violation may inevitably require some subjective assessment of harms that cannot be precisely quantified. As a result of such estimation problems, the penalties thereby calculated may in practice be no less arbitrary than penalties assessed under a less scientific model.[54]
Finally, the deterrence model is largely amoral. As a consequentialist model, its concern is solely with preventing future contraventions of the law. Notions of fairness and morality have no role to play. Accordingly, the deterrence model is prone to generating counter-intuitive results and outcomes which appear to be at odds with fairness and justice. For example:
The above examples provide powerful illustrations of various ways in which the deterrence model can generate results which are unfair and unjust. The deterrence model purports to predict and deter future behaviour, rationality is assumed, and considerations of fairness are ignored.
While the primary weakness of the deterrence model is arguably its capacity for generating counter-intuitive results which are unfair, this constitutes the greatest strength of the desert approach to penalty setting. Because the aim of the desert model is to ensure that a ‘just’ penalty is imposed on the offender, the penalty level is designed to reflect social values of fairness.[60] Thus, repeat violations will lead to an escalation in penalty levels, so as to express moral outrage at the offender’s apparent wilful disregard of the law.
Furthermore, the more open-ended nature of the desert model’s inquiry into factors which bear upon the seriousness of the violation can potentially accommodate consideration of many of the variables which form the basis of the deterrence model. The notion of harm, which is one of the two components upon which the concept of seriousness is founded, is strongly linked to the injury-based variant of the deterrence model. Thus, where a violation of the law imposes large losses on others, this will increase the perceived seriousness of the violation and hence the level of penalty. Similarly, where the offender’s conduct is covert and deliberately concealed, thus lowering the probability of detection, the perceived culpability of the offender is increased and an escalation of the penalty is justified. While the desert model allows these variables to be factored into the penalty assessment process in the manner advocated by the deterrence model, it does not require such precise quantification of the relevant variables.
While the primary strength of the deterrence model is its ability to generate (at least in theory) a precise penalty figure, it is this quality which constitutes the principal weakness of the desert approach to penalty-setting. The desert approach is not theoretically capable of assigning a particular penalty to a particular offence.[61] The desert model is premised on the notion that offences can be ranked according to the degree of seriousness. This generates two difficulties, arising from the two dimensions of proportionality. Firstly, it assumes that a ranking of the relative seriousness of offences is possible (ordinal proportionality), when in practice this may be a controversial and complex task.[62] Secondly, even if it were possible to devise an accurate and acceptable ranking of offences on the basis of relative seriousness which thus enables the relative magnitude of the penalty to be determined, this cannot generate the absolute magnitude of the penalty (cardinal proportionality).[63]
The lack of precision inherent in the desert model also provides considerable scope for inconsistency and the appearance of arbitrariness in setting penalty levels. Concerns about inconsistency in penalty setting and sentencing generally have led to a variety of attempts by law and policy-makers to develop a more structured approach to sentencing.[64] While the desert model strives to set penalties at a fair and just level, inconsistency in the overall approach to penalty setting can potentially undermine this objective.
The different strengths and weakness of the two models are a product of their fundamentally different philosophical foundations.[65] The deterrence model assumes that people and firms are rational actors who make decisions according to a scientific cost-benefit calculus. Morality and popular conceptions of a ‘just society’ have no role to play. The deterrence model’s vision of society is one in which scarce resources are efficiently allocated — all consumers needs and wants, as indicated by their willingness to pay, are satisfied at the lowest possible cost. Within the deterrence model, penalties have no intrinsic value — they are simply signalling devices, indicating to potential offenders the likely cost of violating the law. The message conveyed by the imposition of a penalty as perceived by the deterrence model is, ‘your conduct is costly, and you must therefore pay’. It says nothing about the wrongfulness of the conduct. It is a consequentialist, forward-looking model — penalties serve no purpose other than to indicate to potential offenders the likely cost of committing an offence.
By contrast, the desert model is overtly moral. The proportionality principle, upon which the desert model is founded, reflects the moral premise that ‘the punishment must fit the crime’. Unlike the deterrence model, the desert model focuses not on the costs of the unlawful conduct, but on its wrongfulness. Punishing wrongful conduct is regarded as a necessary requirement of a just society.[66] The message thereby conveyed by the imposition of a penalty is ‘your conduct is wrongful, and you must therefore be punished’. It is primarily a backward looking model — penalties serve to censure or blame offenders for their wrongful conduct.[67]
The two models clash directly in one important respect. According to the deterrence model, if the probability of detection is less than 100 percent, then the quantum of penalty must be increased by dividing the penalty calculated (under the assumption that detection is 100 percent certain) by the probability of detection. Thus, the smaller the probability of detection, the greater the magnitude of the penalty. By inflating the penalty level in this manner, the deterrence model conflicts directly with the proportionality principle on which the desert model is founded. A penalty so magnified would be regarded by the desert model as ex hypothesi disproportionate to the extent of the offender’s wrongdoing. Because wrongfulness encapsulates the harm caused by the offender’s unlawful conduct, then a penalty which significantly departs from the size of the harm caused would fail to meet the proportionality criteria.[68]
As the preceding discussion illustrates, both models suffer from intractable problems but nevertheless possess characteristic strengths and can provide coherent guidance when attempting to resolve the practical problem of setting penalties. Neither model can clearly be regarded as superior to the other. Given the stark differences in philosophy and approach taken by the two models, which model is best suited to determining the quantum of penalties for violations of regulatory law?
The above theoretical models arose from attempts by legal theorists and philosophers to explain the justification and severity of punishment for traditional criminal offences.[69] There is, however, a perceived distinction which has been made by some judges between mala in se, conduct which is ‘wrongful in itself’ and which is thus treated as ‘criminal’ in societies with very different social and economic values, and mala prohibita, conduct which cannot be so characterised and which is the subject of specific proscription.[70] The paradigm of criminal responsibility involves the notion of individual wrongdoing, generally with a degree of awareness of the act or its consequences.[71] By contrast, the defining characteristic of laws which are aimed at regulating economic and social activity is that the breach of such laws does not commonly attract the same degree of moral condemnation which accompanies the commission of traditional crimes.[72] Regulatory law is primarily concerned to facilitate the achievement of collectivist goals by discouraging behavior which is considered to be inimical to those goals and thus detrimental to collective welfare. The punishment of regulatory offences is a practical means of controlling an activity, without necessarily implying the element of social condemnation which is characteristic of traditional crimes.[73]
Some commentators argue, however, that it is possible to attach moral opprobrium to regulatory offences, claiming that traditional crimes arouse moral disgust partly because they have been labeled criminal for so long.[74] Richardson, Ogus and Burrows attempt to find a middle-ground by arguing that an element of blame can be attached to many regulatory offences in so far as the commission of an offence amounts to a failure by the offender to take reasonable care in carrying out the regulated activity.[75] Others have distinguished regulatory laws which seek to protect economic interests, from regulatory laws which are concerned to promote other social interests (such as public health, industrial safety and environmental hygiene).[76] Kadish argues that the stigma of moral reprehensibility is not naturally associated with regulatory law which is primarily concerned to protect economic interests (such as competition law, securities regulation, export controls and the like), but does not make the same claim in relation to regulatory laws which seek to protect non-economic interests.[77]
Nonetheless, it cannot be denied that considerably less moral stigma attaches to the commission of regulatory offences when contrasted with traditional crimes, at least in relation to regulatory offences which are aimed at safeguarding collective economic interests or where the consequences of a regulatory offence are minor in their effects (such as car parking infringements and so forth). The lack of significant moral condemnation attaching to the commission of regulatory offences has implications for the theories of punishment outlined above. In particular, it presents desert theorists with a serious dilemma. If punishment is warranted because a just society is required to censure offenders for their blameworthy conduct, then it follows that punishment is not justified in the absence of any wrongdoing associated with the commission of a regulatory offence.[78] The dilemma is less acute where a regulatory offence is committed deliberately in wilful defiance of the law, for it is possible to argue that the deliberate flouting of the law provides an independently adequate ground for condemnation.[79] Deterrence theorists, on the other hand, face no such dilemma. Because the purpose of punishment is to prevent future harm and thereby safeguard collective interests, the punishment of regulatory offences is justified in order to deter others from such conduct.[80]
The dilemma faced by desert theorists is compounded by the fact that many regulatory offences are strict liability offences.[81] While criminal offences typically depend on proof of some form of mens rea, be it intent, knowledge or recklessness, in order to establish the commission of an offence, regulatory violations often discard mens rea so that a regulatory offence may be committed without the need to establish any mental element on the part of the offender. In other words, liability is visited on the offender for breach of regulatory law without proof of the offender’s culpability or subjective wickedness. The imposition of strict liability means that even the actor who takes all due care to avoid committing the offence, but nevertheless accidentally engages in the proscribed conduct, acts unlawfully and may therefore be punished.[82] For the desert theorist, the absence of any need to establish the offender’s culpability strongly suggests that punishment of the offender cannot be justified because it has not been shown that the offender is morally blameworthy. It has been argued, however, that the moral problems posed by strict liability offences are outweighed by the cost savings and practical expediency of dispensing with the requirement of proof of fault, particularly where the penalties are relatively small and the liberty of the individual is seldom at stake.[83]
The use of strict liability for regulatory offences may also have implications for deterrence theorists. Although it has generally been argued that strict liability strengthens the deterrent effect of the law by encouraging extreme care to be taken when carrying out the regulated activity, the deterrence impact of strict liability is unclear. It may be that strict liability reduces the incentive to take care, because liability may arise if the proscribed conduct occurs, however careful the actor may have been.[84] Kadish argues that the dilution of the requirement of culpability and the consequent proliferation of convictions without grounds for condemnation tends in the long run to impair the effectiveness of punitive sanctions both in the area of economic regulation and traditional crimes.[85]
Because regulatory law’s central concern is to modify behaviour by reference to its perceived undesirable effects on collective welfare, rather than to express moral condemnation of the offence, the theoretical justification for punishment of those who violate regulatory law appears to rest firmly on the deterrence theory of punishment. Does it therefore follow that, because the justification for punishment of regulatory offences rests primarily on deterrence rather than desert, the quantum of penalty for violation should be determined in accordance with the economic variant of deterrence theory?
If one subscribes to the view that a system of punishment should adhere to only one value, then this question must be answered in the affirmative.[86] However, as H L A Hart has powerfully argued, ‘any morally tolerable account’ of punishment ‘must exhibit it as a compromise between distinct and partly conflicting principles.’[87] Although the question of ‘why punish?’ is related to the question ‘how much to punish?’, they are nevertheless separate and independent questions.[88] Put differently, an important distinction must be made between questions of liability, in respect of which notions of fairness and morality arguably have little or no role to play in the regulatory context, and questions of sanction, which inevitably implicate the social and moral order.[89] It is therefore perfectly consistent to accept that the only justification for a system of punishment of regulatory offences is to deter future harms, but also to accept that the quantum of punishment for such offences should take issues other than deterrence into consideration.[90]
Accordingly, although the justification for punishing regulatory offences rests on deterrence, it does not necessarily follow that penalties for regulatory offences must be quantified in accordance with the deterrence theory of punishment, let alone the economic variant of deterrence theory. Not only can a system of state punishment seek to implement a plurality of values without falling into logical incoherence, but it is positively desirable that it should do so, thereby reflecting the multiplicity of societal goals which may legitimately be pursued by the state in punishing those who break the law. State punishment for regulatory offences should therefore encapsulate the essential concerns of both the deterrence and desert theories of punishment. Assuming that the state is justified in regulating the proscribed activity by the imposition of penal sanctions, it is important that the severity of those sanctions should be sufficient to deter potential offenders from engaging in the proscribed conduct, otherwise the regulatory scheme is unlikely to be effective. The deterrence justification for regulatory offences suggests that the quantum of regulatory penalties should, in the first instance, be set at a level that is sufficiently high to deter the offender and others from contravening the law.
On the other hand, a system of punishment must not overlook the requirements of fairness in seeking to secure effective regulation.[91] The mere fact that there is little or no moral condemnation attached to the commission of a regulatory offence, nor a need to prove culpability by the offender to establish liability, does not necessarily imply that moral considerations and issues of culpability are irrelevant to the penalties levied thereunder.[92] Although conceptions of fairness and morality do not play a central justification for regulatory law, wholly ignoring considerations of fairness in penalising regulatory offenders would be objectionable in principle.[93] The fact remains that penalties for regulatory offences are, after all, a form of legal punishment. As such, penalties impose burdens on citizens which are backed by the coercive power of the state and thus possess an intrinsic moral dimension.[94] Penalties have significant consequences, both for the citizen on whom they are levied and for the society in which they operate. When a penalty is imposed for violating the law, it communicates the message ‘you broke the law, you must be penalised’.[95] Because penalties for violation of the law involve the imposition of ‘hard treatment’ on citizens by the state, it is essential in any liberal democratic society that the limits on the scope of punishment are properly circumscribed.[96] Democratic societies recognise the moral worth and autonomy of citizens and cannot therefore justify visiting disproportionately harsh punishment on those who break the law.[97] While the quantum of regulatory penalties should in the first instance be guided by the goal of deterrence, the proportionality principle is an important delimitation on the intrusive powers of the state.[98] Accordingly, exclusive reliance should not be placed on the deterrence model as the basis for determining penalty levels because this could theoretically permit the imposition of draconian, or unfairly lenient, sanctions on an individual offender.
Ignoring considerations of fairness would not only be unacceptable as a matter of principle, but would also be unwise as a matter of practice. As the preceding section demonstrates, the application of the deterrence model of penalty setting may generate unfair results. If a penalty scheme departs too greatly from moral consensus, this can (particularly over time) undermine the perceived legitimacy of the regulatory regime within which penalties operate.[99] Not only does this bring the law into disrepute, but may eventually undermine the legitimacy of the regulatory system.[100] Without the confidence and support of the regulated parties, and the society in which they operate, the regulatory system will inevitably fail to achieve its objectives.[101] The stigmatising power of punishment for unlawful conduct is one of society’s most powerful tools for encouraging compliance with the law.[102] A penalty scheme must not be too out of step with intuitive conceptions of fairness and justice lest it fail to inculcate in those who violate the law a sense of the wrongfulness of their conduct.
In summary, the deterrence justification for regulatory offences suggests that the quantum of regulatory penalties should, in the first instance, be set at a level that is sufficiently high to deter the offender and others from contravening the law. But, in any given case, if the goal of general deterrence requires the imposition of a penalty of a magnitude that would be disproportionate to the seriousness of the offence, then the goal of deterrence (be it ‘absolute’ or ‘optimal’, depending on whether or not the unlawful conduct is potentially welfare-enhancing) must give way to the moral principle that the ‘punishment must fit the crime’. In other words, the proportionality principle should circumscribe the outer limits of the range of acceptable penalties. Within that range, the deterrence model can be used to determine a particular quantum of penalty.[103] An effective penalty regime for regulatory offences should be based on a hybrid of the two models, so that the quantum of punishment reflects the goal of deterrence provided that it is not unfairly disproportionate to the seriousness of the offence.
The final section of this paper seeks to illustrate the application of the theoretical models examined above by focusing on penalties for violation of competition law, focusing in particular on the Australian experience of competition regulation. Laws aimed at regulating competition provide a prime example of regulatory law: the purpose of competition law is to eliminate certain restrictive commercial practices, not because such practices are worthy of moral condemnation, but because they lead to the inefficient allocation of resources by impairing the competitive process and thus diminish the overall welfare of society. It is particularly valuable to examine how Australian law has approached the task of penalty setting for competition law violations because a significant volume of case law has developed in relation to the penalty-setting process and because of recent controversy concerning the very substantial penalties which have been imposed on some offenders.[104]
Australian competition law is largely embodied in a series of statutory prohibitions contained in Part IV of the TPA. The Australian regulatory framework is based on a dual adjudication system: it is for courts to determine whether the Part IV prohibitions have been contravened and to impose pecuniary civil penalties on offenders, but it is for the ACCC as an independent administrative authority to determine whether conduct which would otherwise contravene the Act confers a net public benefit to the community and should therefore be granted legal immunity from the prohibitions.[105] The ACCC cannot grant such legal immunity unless the conduct has first been notified to the ACCC or the parties have applied to the ACCC for authorisation of the proposed conduct.[106] Thus, the legislative prohibitions contained in Part IV of the Act are not absolute. Immunity from the prohibitions is available in relation to net welfare-enhancing conduct (with the exception of monopolisation under s 46) via the authorisation and notification process.
Before attempting to identify which model best explains the approach taken by Australian courts when setting competition law penalties, it is useful first to consider how the deterrence and desert models could be expected to apply in the competition law context.
If the deterrence model is applied to competition law violations, should the aim of penalty setting be ‘optimal’ or ‘absolute’ deterrence? As previously noted, if the practices sought to be prohibited may in some circumstances be welfare-enhancing, then economic theory suggests that the ‘injury to others’ variant of the model should be applied so as not to deter welfare-enhancing violations. Apart from price fixing and bid-rigging agreements between competitors, which are universally condemned as economically inefficient, not all horizontal restraints (ie, where competition is restricted by co-ordination between firms which operate at the same level in the supply chain) are unequivocally anti-competitive and inefficient.[107] The same is true of vertical restraints (ie, where a restraint is imposed by one firm on another operating at a different level in the supply chain). A vertical restraint implemented by a firm with market power may be inefficient and thus welfare reducing, but the same type of restriction by a firm lacking market power may be benign, and may well have net pro-competitive effects.[108] Thus an evaluation of the market power of the firms engaged in the restrictive practice in the particular case may be required before the restraint can be condemned as inefficient.[109]
There is a considerable body of literature (largely written in the US) which is concerned with identifying the optimal basis for determining the penalty level for ‘antitrust’ (ie, competition law) violations so as to provide a measure of the appropriate level of damages which should be payable in a private antitrust suit against an offender. A theme that appears throughout the economic literature is a concern to avoid ‘over-deterrence’ which follows from the belief that welfare-enhancing violations of the law should not be deterred.[110] Posner cites by way of example a market-power enhancing merger which imposes costs of $1 million on society but confers benefits, as a result of the lower costs of the merged entity arising from economies of scale and scope, of $2 million. Posner argues that if the purpose of antitrust law is to promote efficiency, such a merger should be permitted but if the penalty is set substantially above the social cost of the violation then the merger will be deterred.[111] Thus, it is widely assumed by such theorists that some anti-competitive practices which violate the law may have net welfare-enhancing effects.[112]
If this strand of economic reasoning is adopted,[113] then it follows that competition law penalties should aim to achieve optimal deterrence, rather than absolute deterrence. The ‘injury to others’ variant of the deterrence model is therefore the preferred basis for penalty setting.[114] An exception to this approach should be made in cases of blatant horizontal price fixing and bid-rigging in respect of which penalties should be set so as to achieve absolute deterrence, thus utilising the ‘unlawful gain’ model, based on universal acceptance that such conduct is inefficient and confers no redeeming benefits on society.[115]
The goal of ‘optimal’ rather than ‘absolute’ deterrence appears to underpin the Australian institutional framework for regulating competition because restrictive practices which would otherwise breach the statutory prohibitions may be immune from penalty if the practice is notified to, or authorised by, the ACCC. It does not, however, follow that the courts should adopt the injury-to-others model (which aims to achieve optimal deterrence) as the basis for quantifying penalties for contraventions of the TPA. The institutional mechanism for ensuring that net welfare-enhancing anti-competitive practices are not deterred is via the authorisation and notification process. The task of evaluating whether the public benefits from proposed anti-competitive conduct does not rest with the courts, but with the ACCC as an independent competition regulator.[116] The task of courts when applying the legislative prohibitions is confined to considering only the effect or likely effect of the relevant conduct on competition; it does not extend to examining any other effects (whether welfare-enhancing or otherwise).[117]
Accordingly, it is necessary to ensure that penalties are structured so that firms proposing to engage in allegedly welfare-enhancing anti-competitive conduct face appropriate incentives to notify or to seek authorisation from the ACCC, thereby enabling the alleged social benefits to be evaluated and weighed against any potentially harmful anti-competitive effects. Therefore the penalties should be set so as to ensure that violations of the prohibitions are deterred absolutely, thus encouraging firms to seek authorisation of anti-competitive practices which provide net social benefits. Anti-competitive conduct which is welfare-enhancing should not be deterred, because authorisation can be sought and granted if net public benefit is established.
The only conduct which cannot be authorised is misuse of market power which is prohibited by s 46 of the TPA.[118] In order to establish a breach of s 46, the applicant must establish that the respondent: a) had a substantial degree of power in a market; b) has taken advantage of that power; c) for a proscribed purpose. If the conduct in question can be explained on grounds of productive efficiency then the respondent has not ‘taken advantage’ of its market power.[119] Thus the substantive prohibition contained in s 46 is structured with a built-in safeguard in order to ensure that conduct which is productively efficient and thus net welfare-enhancing is not caught by the prohibition. Assuming that legal error is avoided in applying s 46 so that only welfare-reducing conduct is prohibited,[120] then conduct which contravenes s 46 should be prohibited absolutely and penalties for breach should be assessed in accordance with the ‘unlawful gain’ model so as to secure absolute deterrence.[121] If this view is adopted, then penalties for contravening Australian competition law should be set according to the unlawful-gain variant of the economic model, rather than the injury to others variant.
If the desert model is applied to competition law penalties, it is necessary to establish both: 1) the seriousness of competition law violations relative to other violations of the law which are punishable by monetary penalty and the relative seriousness of specific competition law violations relative to each other (thereby establishing a scale of ordinal proportionality); and 2) the magnitude of the penalty scale itself, which involves specifying the monetary amount which fairly reflects society’s disapproval of the unlawful conduct (thereby establishing a scale of cardinal proportionality). Once ordinal and cardinal proportionality scales have been established, the quantum of penalty for any given violation of the prohibitions can be determined by reference to the seriousness of the violation.
It is immediately apparent, however, that the establishment of a penalty scale which ranks the severity of competition law violations relative to other violations of the law is a complex task which is practically impossible to achieve with any scientific precision. Nonetheless, in theory it may be possible to rank the seriousness of competition law violations relative to other types of commercial conduct which is subject to regulation and punishable by monetary sanction — such as laws relating to insider trading, disclosure of corporate information, financial services regulation and so forth, by reference to the potential harm caused by such practices valued in monetary terms. The relative culpability of offenders as indicated by the offender’s actual mental state, such as the deliberateness of the offence and knowledge of the illegality may also provide some assistance in ranking the ‘seriousness’ of specific violations of the law.
However, establishing the magnitude of the penalty scale in financial terms so as to reflect the extent of society’s disapproval cannot be achieved in any principled, non-arbitrary way. There is no quantum of penalty which suggests itself as uniquely appropriate for competition law violations.[122] Australian competition law sets the statutory maxima for penalties at A$10 million for corporations and A$500,000 for individuals, but these penalty ceilings cannot be explained other than as the legislature’s intuitive perception of the suitable magnitude of penalties.[123]
Assuming that reasonable solutions can be found to overcome the difficulties inherent in establishing a scale of proportionality, then a penalty scheme which is based on the just desert model could rest on the proportionality principle in order to determine the quantum of penalties, with the ‘seriousness’ of violations determined by the extent of the harm caused by the violation and the culpability of the offender. Hence the greater the loss caused by the violation, the greater the quantum of penalty. Similarly, an ‘innocent’ breach of the law would be penalised less severely than a breach arising from deliberate action taken by the offender in knowing violation of the law.
The final section of this paper attempts to identify which model best explains the quantum of Australian competition law penalties. While the responsibility for initiating and conducting civil enforcement proceedings in order to recover penalties from suspected offenders lies with the ACCC,[124] it is the Federal Court of Australia which decides whether a breach of the Act has occurred and, if so, the quantum of penalty to be imposed.[125] Statutory guidance is provided to the Federal Court in determining the level of penalty for a breach of Part IV in s 76(1) of the Act which confers on the court a very wide discretion to determine the quantum of penalty but sets out a list of relevant matters which the court must take into account when exercising that discretion. For this purpose, relevant matters include, but are not limited to:
The statutory criteria do not, of themselves, appear strongly to reflect either of the theoretical models for penalty setting described above. However, it could be argued that these criteria fit more comfortably with the desert model of penalty setting, for all the stated criteria are matters which pertain to the ‘seriousness’ of the unlawful conduct. On the other hand, the listed criteria could be broadly interpreted as providing a basis for inclusion of the specific quantifiable variables which are utilised by the deterrence model. The statutory criteria are not exhaustive, however, and the court’s ultimate task is to determine a penalty ‘as it determines to be appropriate having regard to all relevant matters’. Accordingly, it is necessary to look more closely at the case law to identify how the Federal Court has approached the task of penalty setting under s 76 in order to identify the theoretical basis upon which the Court has proceeded.[127]
There have been a significant number of decisions in which the courts have considered the operation of s 76. The leading authority concerning penalty setting for contraventions of Part IV of the TPA is the judgment of French J in TPC v CSR Ltd.[128] In particular, French J elaborated on the criteria which are relevant to a determination of the appropriate level of penalty,[129] stating that:
The assessment of a penalty of appropriate deterrent value will have regard to a number of factors which have been canvassed in the cases. These include the following:
The first three factors are all expressly mentioned in s 76. They can be regarded as measures of the scope and impact of the conduct and it is conducive to deterrence that the greater the significance of these elements, the heavier the penalty should be.[135]
French J also made a number of strong statements of principle concerning the determination of penalties under Part IV of the Act which, at first blush, suggest that the deterrence model informs the Court’s approach to penalty setting. In his view, the sole purpose of such penalties is to deter future contraventions of the Act. Morality has no role to play.
Punishment for breaches of the criminal law traditionally involves three elements: deterrence, both general and individual, retribution and rehabilitation. Neither retribution nor rehabilitation, within the sense of the Old and New Testament moralities that imbue much of our criminal law, have any part to play in economic regulation of the kind contemplated by Pt IV. Nor, if it be necessary to say so, is there any compensatory element in the penalty fixing process ... The principal, and I think probably the only, object of the penalties imposed by s 76 is to attempt to put a price on contravention that is sufficiently high to deter repetition by the contravenor and by others who might be tempted to contravene the Act.[136]
The view that punishment has no role to play in the determination of penalties under s 76 was recently endorsed by the majority of the Full Federal Court in NW Frozen Foods Pty Ltd v ACCC.[137] The majority rejected the approach taken by Heerey J in comparing the culpability of the appellant with that of another corporation in an earlier price fixing case in which sizeable penalties were imposed. By so doing, Heerey J had, in the Full Court’s opinion, engaged in an exercise of punishment which was not the purpose of penalties imposed under s 76.[138]
By claiming that the only object of penalties imposed by s 76 is to ‘price’ contraventions so as to deter future contraventions by the offender (ie, specific deterrence) and other potential offenders (ie, general deterrence), French J appears to provide powerful endorsement of the deterrence model as the theoretical basis for penalty setting for breach of Part IV, although he does not make clear which variant of the deterrence model is to be applied. The express rejection of morality and punishment as having any relevance to the determination of penalties also appears to reject outright the desert model as the basis for penalty setting. Despite these statements, an examination of the general body of case law, the specific criteria enumerated by French J as relevant to penalty setting, and the reasoning adopted by French J in the CSR case itself, indicates that the court in fact pays more lip-service than adherence to the deterrence model of penalty setting and in practice adheres more closely to the principles underlying the desert model of penalty setting.
Although French J has stated that notions of morality and punishment have no role to play in the setting of penalties under s 76, the issue is by no means settled. In NW Frozen Foods, Carr J expressly stated that ‘the cases decided to date on the question of the assessment of pecuniary penalties have not ruled out or excluded punishment as one of the purposes of s 76 of the Trade Practices Act’, and deliberately left the issue open.[139] This comment was referred to in ACCC v Australian Safeway Stores[140] by Goldberg J who stated that, apart from the above-cited statement of French J in CSR, he was ‘unable to find any positive statement in any of the cases that the purpose of the imposition of penalties by s 76 “is not punishment”.’[141] Goldberg J considered the matter to be significant, because ‘in an appropriate case where there has been a flagrant and wilful contravention [a court] might take the view that a severe penalty was warranted, having regard to the deliberateness and wilfulness of the contravention.[142] In other words, the culpability of the offender may be relevant to determining the quantum of penalty. Similarly, Heerey J in ACCC v J McPhee & Son (Australia) Pty Ltd[143] stated that he did not necessarily agree with the statements of French J in CSR cited above. While Heerey J acknowledged that contravention of Part IV provisions are not criminal offences and that deterrence remains a primary objective of the imposition of penalties,
nevertheless s 76 imports into the penalty fixing process concepts of moral responsibility long known to the criminal law. In other words, the sources of the substantive provisions of Pt IV are doubtless economic policy and theory, but the penalties for contraventions are to be applied in a moral universe.[144]
A closer examination of the comments made by French J in CSR reveals that he conflated the substantive rules for determining whether a contravention of the TPA has occurred with the principles to be applied in determining the level of penalty for contravention. Put differently, he confused the question of breach with the question of penalty.[145] These are separate and independent questions. There are two strands of reasoning which led French J to his conclusion that morality has no role to play in penalty setting. First, because the purpose of the TPA is directed to regulating competition, the statutory provisions are regulatory rather than penal in nature. Secondly, the High Court in Queensland Wire Industries v Broken Hill Proprietary Ltd expressly rejected the view that morality was relevant when determining whether s 46 of the TPA (misuse of market power) has been contravened.[146] According to French J, it therefore followed that morality was irrelevant to setting penalties for contraventions of the TPA. But both strands of reasoning are concerned with the issue of breach: the first goes to the subject matter of the TPA, and the second to the substantive rules which define unlawful behaviour. While French J correctly stated that morality has no place in determining whether a breach of the TPA has occurred, it does not necessarily follow that morality has no place in determining the appropriate penalty which should apply in respect of a breach.[147] Indeed, it is argued that ignoring moral considerations in setting penalties for breach of Part IV would be undesirable and unwise, lest the regulatory regime lose the respect and support of the regulated parties.
Furthermore, on a closer examination of the nine factors cited by French J as relevant to a determination of penalty setting, six are concerned with matters of culpability and cannot be easily reconciled with the deterrence theory of penalty setting nor regarded as devoid of moral content. For example, the deliberateness of the contravention is irrelevant to the deterrence model. Indeed, the injury to others variant of the deterrence model is grounded on the notion that deliberate contraventions of the law are desirable and should occur if their net effect is welfare-enhancing. Because the deterrence model assumes that each actor is rational, all contraventions are assumed to be deliberate. By contrast, the deliberateness of a contravention lies at the core of the concept of culpability, which is fundamental to a determination of the penalties levied in accordance with the desert model. Similarly, the adoption of a compliance program and culture by the offender has no bearing on the calculation of penalties under the deterrence model. If the law is violated, then the offender must pay. All other things being equal, an offender who provides evidence of a culture of compliance should be penalised to the same extent as the offender who sets out deliberately to contravene the law.[148]
Furthermore, the comments of Smithers J in TPC v Stihl Chain Saws (Aust) Pty Ltd[149] were quoted by French J with approval, thus implicitly acknowledging that morality and fairness have a legitimate role in determining penalties under s 76:
The penalty should constitute a real punishment proportionate to the deliberation with which the defendant contravened the provisions of the Act. It should be sufficiently high to have a deterrent quality, and it should be kept in mind that the Act operates in a commercial environment where deterrence of those minded to contravene its provisions is not likely to be achieved by penalties which are not realistic. It should reflect the will of Parliament that the commercial standards laid down in the Act must be observed, but not be so high as to be oppressive.[150]
Accordingly, the Federal Court’s overall approach to determining the quantum of penalties under s 76 appears to be guided by a combination of the principles upon which the deterrence model and the desert model are founded. In the first instance, the Court claims to be guided by the principle of deterrence, which it defines as requiring penalties to be set sufficiently high so as to deter both the individual offender and those contemplating contraventions of the law in future. On the other hand, the Court has also referred to the importance of penalties being proportionate to the seriousness of the offence, thus indicating that the notion of ‘just desert’ which underlies the desert model should also guide the Court in determining the quantum of penalty. Some of the criteria which the Court has taken into account in penalty setting can be accommodated within the deterrence model, while other criteria appear to fit much more comfortably within the desert model.
Ultimately, the Court has correctly acknowledged that ‘the fixing of the quantum of penalty is not an exact science. It is not done by the application of a formula, and, within a certain range, courts have always recognised that one precise figure cannot be incontestably said to be preferable to another’.[151] In general, the framework of principles which the court has developed to determine the appropriate quantum of penalty strike a reasonable balance between the need to ensure that penalties serve to deter future contraventions, and the need to inculcate into offenders the sense that unlawful conduct is unacceptable and will not be tolerated, while avoiding penalties which would be regarded by society as disproportionate and oppressive.
This paper has examined two theoretical models which can be applied to determine the quantum of penalties for regulatory offences: the ‘deterrence’ model and the ‘desert’ model. The deterrence model is based on economic efficiency and seeks to ‘price’ unlawful conduct at a level which will deter firms from engaging in such conduct so as to minimise the social cost of unlawful conduct. Two variants of the deterrence model can be identified. The ‘injury to others’ model, which aims to deter unlawful conduct to its optimal level, by deterring only unlawful conduct which is net welfare-reducing to society. Thus, the relevant quantum of penalty is measured by the cost of harm to others caused by the offender’s wrongful conduct. On the other hand, the ‘unlawful gain’ model is concerned to achieve ‘absolute’ deterrence, in order completely to eliminate unlawful conduct, based on the assumption that such conduct is always net welfare-reducing. Thus, the relevant quantum of penalty is measured by the size of the offender’s unlawful gain, for by disgorging the offender’s ill-gotten gain, the incentive to violate the law is eliminated. By contrast, the ‘desert’ model regards punishment as a means of censuring offenders for wrongdoing and thus seeks to punish offenders in proportion to the wrongfulness of their unlawful conduct, which is reflected in the seriousness of the unlawful conduct as indicated by the offender’s culpability and the extent of the harm thereby caused.
Neither theoretical model is superior to the other. Both have certain strengths and weaknesses. The advantages of the deterrence model are its theoretical capacity to generate a specific quantum of penalty and its recognition of the need to take account of the probability of detection in calculating the penalty. On the other hand, the deterrence model assumes that firms and individuals are solely motivated by rational selfish welfare-maximisation, which may not accurately reflect reality, even in the commercial context. It also generates intractable problems in practice when attempting to quantify variables which are inherently incapable of precise quantification. Its main limitation, however, is its disregard for concepts of fairness and justice, so that penalties determined in accordance with the deterrence model may generate counter-intuitive outcomes which may undermine public confidence in the regulatory system. By contrast, the greatest strength of the desert model is its deliberate concern that penalty levels should be fair. However, its greatest weakness is its imprecision and inability to generate a precise quantum of penalty in any given situation, which can lead to inconsistency and arbitrariness in penalty setting.
Although it may be tempting to suggest that the deterrence model is preferable to the desert model as the basis for quantifying regulatory penalties because regulatory law is primarily concerned with achieving economic and social goals rather than expressing moral outrage at the proscribed conduct, this overlooks the important distinction between liability and penalty. While issues of fairness and morality may have little or no role to play in determining whether certain conduct is unlawful, such notions are inevitably implicated in the process of state penalisation of unlawful conduct. A penalty scheme must not be too out of step with societal notions of fairness and justice, lest it fail to inculcate those who violate the law with a sense of the wrongfulness of their conduct, and lead to an erosion of public confidence in the regulatory system. Accordingly, it is suggested that a fair and effective regulatory penalty regime should combine the essential features of both models: penalties should be set at a level which are sufficiently high to deter future contraventions of the law, provided that any given penalty award is not unfairly disproportionate to the seriousness of the offence.
A review of Australian competition law and the principles emerging from the case law in determining the level of competition law penalties are replete with incantations of the aim of deterrence and the irrelevance of morality, suggesting that the Federal Court adheres to the deterrence model in quantifying penalties. On closer examination however, it appears that both the desert and deterrence models inform the Court’s approach to penalty setting, striking a reasonable balance between the need to deter future contraventions of the Act while avoiding oppressive or unfairly disproportionate penalties.
[*] Fellow in Law, Linnells’ Lecturer in Commercial Law, St Anne’s College, The University of Oxford. This article is based on a research report prepared by the author when acting as a consultant to the Australian Competition and Consumer Commission and is published with the kind permission of the Commission. The views expressed do not in any way represent the views of the Commission and are entirely those of the author. I am indebted to Frances Hanks, Timothy Endicott, Dan Prentice, Terry O’Shaughnessy and Liora Lazarus for their helpful comments on earlier drafts. Any errors remain my own.
[2] Ibid 41,583.
[3] See, eg, TPC v TNT Australia Pty Ltd [1995] ATPR 41-375; ACCC v Australian Safeway Stores Pty Ltd [1997] FCA 450; [1997] ATPR 41-562; ACCC v NW Frozen Foods Pty Ltd [1996] ATPR 41-515; ACCC v J McPhee & Son (Australia) Pty Ltd [1998] FCA 310; [1998] ATPR 41-628.
[4] R Featherston, ‘Negotiated Penalties — Trick or Treat?’ (Paper Presented at the Trade Practices Workshop, Business Law Section, Law Council of Australia, Fremantle, 1995). Cf Stephen Corones, ‘Collusive Tendering — Are the Penalties a Deterrent?’ (1996) 17 Proctor 16.
[5] See the comments of Professor Allan Fels referred to in Corones, above n 4, 16.
[6] Anthony Ogus, Regulation: Legal Form and Economic Theory (1994) ch 2.
[7] See generally Cosmo Graham and Tony Prosser, Privatizing Public Enterprises: Constitutions, the State and Regulation in Comparative Perspective (1991).
[8] The use of the monetary penalty as a sanction for corporate offenders has been criticised by a number of academic commentators as ineffective. For example, it has been argued that monetary penalties: do not pinch directly on the managerial nerves of corporate governance; do not necessarily result in corporate offenders taking internal disciplinary action against the individuals responsible; provide inadequate incentives for the corporation to revise its internal controls so as to guard against repeat offences; convey the impression that offences are purchasable commodities and tend to under-emphasise the socially undesirable nature of corporate offences; may have adverse spillover effects (externalities) on shareholders, workers, consumers and other third parties; are prone to evasion through the use of corporate group structures; and the penalty levels required to reflect the seriousness of the offence, or to effect deterrence against further offences, may exceed the capacity of the corporation to pay (the ‘deterrence and retribution trap’): see, eg, John Coffee Jr, ‘“No Soul to Damn, No Body to Kick”: An Unscandalized Inquiry into the Problem of Corporate Punishment’ (1981) 17 Michigan Law Review 386; Brent Fisse and John Braithwaite, ‘Sanctions Against Corporations: Dissolving the Monopoly of Fines’ in Roman Tomasic (ed), Business Regulation in Australia (1984) 129, 130–3; Brent Fisse and John Braithwaite, Corporations, Crime and Accountability (1993); Australian Law Reform Commission (‘ALRC’), Compliance with the Trade Practices Act 1974, Report No 68 (1994); ALRC, Sentencing: Penalties, Discussion Paper No 30 (1987). Despite these shortcomings, the monetary penalty occupies an important position within regulatory enforcement and is likely to continue to do so for the forseeable future. On the use of the monetary sanction in US regulatory enforcement, see Colin Diver, ‘The Assessment and Mitigation of Civil Money Penalties by Federal Administrative Agencies’ (1979) 79 Columbia Law Review 1435.
[9] See below Part II(C).
[10] However, this examination makes no empirical assessment of the extent to which the actual quantum of penalties imposed for contraventions properly fulfil the courts’ stated goals or whether the actual penalties which have hitherto been imposed are too high or too low. A recent empirical analysis of penalties for price fixing under the TPA indicated that the approach taken by courts largely conforms with economic reasoning, with the exception of the courts’ treatment of small business offenders: David Round, John Siegfried and Anna Baillie, ‘Collusive Markets in Australia: An Assessment of their Economic Characteristics and Judicial Penalties’ (1996) 24 Australian Business Law Review 292.
[11] Ogus, above n 6, 1–5.
[12] Andrew Ashworth, Principles of Criminal Law (1995) 50–1; Genevra Richardson, Anthony Ogus and Paul Burrows (eds), Policing Pollution (1983) 17; H L A Hart, Punishment and Responsibility (1968) 236. For further discussion, see below Part II(E).
[13] Hart, Punishment and Responsibility, above n 12, 235–6.
[14] The definition of regulation is largely drawn from that of Ogus: Ogus, above n 6, 2. Ogus does not, however, include competition regulation within the scope of his examination of regulatory law, because competition law is intended to reinforce rather than overreach the market system: see Ogus, above n 6, 30. However, within his analysis, Ogus nevertheless includes the regulation of natural monopolies, externalities and information deficits, which are all forms of market failure justifying the need for regulation. Anti-competitive commercial practices which are the subject of competition regulation are also instances of market failure which justify the need for regulation. Competition law is concerned to achieve the collectivist goal of efficient resource allocation, which would not otherwise be achieved in the absence of legal regulation, and thus displays the hallmarks of regulatory law.
[15] Hart, Punishment and Responsibility, above n 12, 14. See also H L A Hart, Law, Liberty and Morality (1963) 37.
[16] Cf Joel Feinberg, ‘The Expressive Function of Punishment’ (1965) 49 The Monist 397.
[17] Andrew Ashworth, Sentencing and Criminal Justice (1992) ch 3. Although these theories are concerned with the justification of punishment generally, the appropriate severity of punishment is, in part, determined by the general justification for punishment: Hart, Punishment and Responsibility, above n 12, 14.
[18] Jack Bilmes and John Woodbury, ‘Deterrence and Justice: Setting Civil Penalties in the Federal Trade Commission’ (1991) 14 Research in Law and Economics 191, 195. However, consider also Brent Fisse, ‘Rethinking Criminal Responsibility in a Corporate Society: An Accountability Model’ in Peter Grabosky and John Braithwaite (eds), Business Regulation and Australia’s Future (1993) 255 and Celia Wells, Corporations and Criminal Responsibility (1993) 21.
[19] Restorative theories of punishment are not solely or necessarily concerned that victims of crime should receive financial compensation. Restorative theories also encompass the perception that crime involves a conflict between the offender and victim which requires resolution — in which the victim is given rights to participate in the process of dispute resolution: Andrew von Hirsch and Andrew Ashworth, ‘Restorative Justice’ in Andrew von Hirsch and Andrew Ashworth (eds), Principled Sentencing (1998) 300–11.
[20] Ibid 395.
[21] Jeremy Bentham, ‘The Principles of Penal Law’ in John Bowring (ed), The Works of Jeremy Bentham (1838–43) 396.
[22] The pioneering work is by Gary Becker, ‘Crime and Punishment: An Economic Approach’ (1968) 76 Journal of Political Economy 169. See generally Robert Cooter and Thomas Ulen, Law and Economics (2nd ed, 1997); Richard Posner, Economic Analysis of Law (2nd ed, 1977).
[23] Cooter and Ulen, above n 22, 389–94.
[24] Ibid 391; Posner, Economic Analysis of Law, above n 22.
[25] Becker, above n 22; Posner, Economic Analysis of Law, above n 22; Cooter and Ulen, above n 22, 395–400.
[26] Becker, above n 22; Kenneth Dau-Schmidt, ‘An Economic Analysis of the Criminal Law as a Preference-Shaping Policy’ [1990] Duke Law Journal 1, 11–13.
[27] Becker, above n 22.
[28] The standard view among economists is that, even in the realm of traditional crimes, the criminal’s unlawful benefit should be factored into a determination of the optimal quantum of punishment: Cooter and Ulen, above n 22, 396.
[29] Ogus, above n 6, 33–5; Robert Officer and Philip Williams, ‘The Public Benefit Test in an Authorisation Decision’ in Philip Williams and Megan Richardson (eds), The Law and the Market (1995) 157, 159.
[30] Bilmes and Woodbury, above n 18, 202–3; Becker, above n 22.
[31] Dau-Schmidt, above n 26, 11–13.
[32] See also Bentham’s first rule: the value of the punishment must not be less in any case than what is sufficient to outweigh that of the profit of the offence: J H Burns and H L A Hart (eds), Jeremy Bentham: An Introduction to the Principles of Morals and Legislation (1970) 166.
[33] Robert Cooter, ‘Prices and Sanctions’ (1984) 84 Columbia Law Review 1523, 1532–3; Warren Schwartz, ‘An Overview of the Economics of Antitrust Enforcement’ (1980) 68 Georgetown Law Journal 1075, 1079–80.
[34] Cooter and Ulen, above n 22, 399–404.
[35] Ibid.
[36] Richard Posner, Antitrust Law: An Economic Perspective (1976) 223–4; Becker, above n 22. For a more detailed discussion of the multiplier principle, see Richard Crasswell, ‘Deterrence and Damages: The Multiplier Principle and its Alternatives’ (1999) 97 Michigan Law Review (forthcoming).
[37] Cooter and Ulen, above n 22; Becker, above n 22.
[38] Cooter and Ulen, above n 22.
[39] Richard Posner, ‘An Economic Theory of the Criminal Law’ (1985) 85 Columbia Law Review 1193.
[40] Cooter and Ulen, above n 22, 286–7; Michael Block and Joseph Sidak, ‘The Cost of Antitrust Deterrence: Why Not Hang a Price Fixer Now and Then?’ (1980) 68 Georgetown Law Journal 1131.
[41] Andrew von Hirsch and Andrew Ashworth, ‘Desert’ in Andrew von Hirsch and Andrew Ashworth (eds), Principled Sentencing (1998) 141–9.
[42] Michael Moore, ‘The Moral Worth of Retribution’ in Andrew von Hirsch and Andrew Ashworth (eds), Principled Sentencing (1998) 150. Classic retributive theory, chiefly associated with Kant, regards criminal responsibility not only as necessary, but also as a sufficient condition for punishment, which society is therefore obliged to impose:
Even if a civil society resolved to dissolve itself with the consent of all its members — as might be supposed in the case of a people inhabiting an island resolving to separate and scatter themselves throughout the whole world — the last murderer lying in prison ought to be executed before the resolution was carried out. This ought to be done in order that everyone might realize the desert of his deeds[.]
Immanuel Kant, The Philosophy of Law (Hastie trans, 1887) 194–8, cited by Paul Weiler, ‘The Reform of Punishment’ in Canadian Law Reform Commission, Studies on Sentencing (1976) 140.
[43] von Hirsch and Ashworth, Principled Sentencing, above n 19. von Hirsch also regards punishment as serving a secondary purpose of deterring others from crime. Penal ‘hard treatment’ provides an additional (modest) prudential supplement to the law’s moral voice by creating a disincentive to at least some of those who are insufficiently motivated by the moral tone of censure.
[44] Andrew von Hirsch, Censure and Sanctions (1993). In contrast, Feinberg argues that the proportionality principle requires that the seriousness of the crime should be proportionate to the public condemnation expressed by punishment, rather than the degree of ‘hard treatment’ imposed on the offender: Feinberg, above n 16.
[45] Andrew von Hirsch, ‘Doing Justice: The Principle of Commensurate Deserts’ in Hyman Gross and Andrew von Hirsch (eds), Sentencing (1981) 243.
[46] Ibid; Hart, Punishment and Responsibility, above n 12.
[47] Andrew von Hirsch and Kathleen Hanrahan, The Question of Parole: Retention, Reform or Abolition? (1979) 16.
[48] Bilmes and Woodbury, above n 18, 202.
[49] Ibid 208.
[50] Ibid 205.
[51] Ashworth, Sentencing and Criminal Justice, above n 17, 60–1; Deryck Beyleveld, ‘Deterrence Research as a Basis for Deterrence Policies’ (1979) 18 Howard Journal of Criminal Justice 135.
[52] Ashworth, Sentencing and Criminal Justice, above n 17, 60; Abraham Goldstein, ‘White-Collar Crime and Civil Sanctions’ (1992) 101 Yale Law Journal 1895, 1898; Fisse and Braithwaite, ‘Sanctions Against Corporations’, above n 8.
[53] Ian Ayres and John Braithwaite, Responsive Regulation: Transcending the Deregulation Debate (1992) 19–23; John Martin, ‘Making the Giant Competitive Rather than Crushing — Industry Perspectives on Regulation Enforcement’ in Peter Grabosky and John Braithwaite (eds), Business Regulation and Australia’s Future (1993) 169. Note however, that the focus on the firm as economic actor compounds the rationality problem, because the firm is controlled by directors, whose objectives and motivations may not be wholly aligned with that of the firm. The relationship between company shareholders and managers generates a classic principal–agent problem. But arguably, the rationale is consistent — shareholders can incur greater monitoring costs as the penalty level rises: Bilmes and Woodbury, above n 18, 203; Fisse, above n 18, 255.
[54] Bilmes and Woodbury, above n 18, 208.
[55] This point is powerfully put by Sir Ian McKay, ‘Penalties, Remedies and Court Processes under the Commerce Act 1986: A Commentary on the Ministry of Commerce Discussion Document’ (Paper presented at the Ninth Annual Workshop of the Competition Law and Policy Institute of New Zealand, Wellington, 1998). See also Hart, Punishment and Responsibility, above n 12, 25. Even economists acknowledge that threatening competition law violators with draconian sanctions may not be optimal due to the inevitability of legal error: Block and Sidak, above n 40.
[56] This problem is avoided by the unlawful gain variant of the economic model, because it aims to achieve absolute deterrence. Thus, while under-deterrence is possible, over-deterrence is not. Note however, that the relevance of prior convictions has been the subject of debate between desert theorists: see Andrew von Hirsch, ‘Desert and Previous Convictions’ in Andrew von Hirsch and Andrew Ashworth (eds), Principled Sentencing (1998) 191.
[57] This can be avoided by increasing the penalty to reflect the greater of the loss to others or gain to the offender — because the absolute deterrence model prescribes a minimum penalty and knows no concept of over-deterrence.
[58] Ashworth, Sentencing and Criminal Justice, above n 17, 61; Andrew von Hirsch and Andrew Ashworth, ‘Deterrence’ in Andrew von Hirsch and Andrew Ashworth (eds), Principled Sentencing (1998) 47; Alan Goldman, ‘Deterrence Theory: Its Moral Problems’ in Andrew von Hirsch and Andrew Ashworth (eds), Principled Sentencing (1998) 80.
[59] von Hirsch and Ashworth, ‘Deterrence’, above n 58; cf Nigel Walker, ‘Deterrence’ in Nigel Walker (ed), Punishment, Danger and Stigma (1980).
[60] von Hirsch, Censure and Sanctions, above n 44, 104–5; Max Neiman, ‘The Virtues of Heavy-Handedness in Government’ (1980) 2 Law and Policy Quarterly 11. In contrast, Walker argues that the moral basis for penal censure is obscure unless the censure can be shown to have social utility, and fails to explain why the offender ‘ought to be punished’: Nigel Walker, ‘Modern Retributivism’ in Hyman Gross and Ros Harrison (eds), Jurisprudence: Cambridge Essays (1992).
[61] Bilmes and Woodbury, above n 18, 198; Ashworth, Sentencing and Criminal Justice, above n 17, 66–7.
[62] See, eg, ALRC, Sentencing of Federal Offenders, Discussion Paper No 15 (1980); ALRC, Sentencing: Penalties, above n 8. In the trade practices context, the ALRC stated that it did not accept that all contraventions of the TPA are equally serious and proposed that a maximum pecuniary penalty be determined in relation to each provision of the TPA where a breach is punishable, and that the determination be based on a considered analysis of the typical relative gravity of a contravention of each relevant provision: ALRC, Compliance with the Trade Practices Act 1974 (Cth), Discussion Paper No 56 (1993) Proposal 6.10.
[63] It could be argued that, according to the desert approach, the penalty should be at least as large as the offender’s gain, according to the moral principle that no-one should profit from a wrong: see Bilmes and Woodbury, above n 18, 199.
[64] ALRC, Sentencing of Federal Offenders, above n 62; ALRC, Sentencing: Penalties, above n 8; Federal Sentencing Guidelines 1998 (USA).
[65] The different values which underlie the two models are borne out in the decisions of the Australian Federal Court when attempting to set out principles which should guide the Court when setting penalties for violations of Australian competition law. Some judges have claimed that the purpose of penalty setting is deterrence and claimed that ‘morality’ has no role to play, whilst other judges have claimed that ‘punishment’ may be a legitimate aim in setting penalties for competition law violations: see below Part III.
[66] Hart, Punishment and Responsibility, above n 12, 222–3.
[67] Neiman, above n 60, 21.
[68] Richard Posner, ‘Retribution and Related Concepts of Punishment’ (1980) 9 Journal of Law and Economics 71.
[69] von Hirsch and Ashworth, Principled Sentencing, above n 19.
[70] Ogus, above n 6, 79; Hart, Punishment and Responsibility, above n 12, 236; Sherras v De Rutzen [1895] UKLawRpKQB 77; [1895] 1 QB 918, 922 (Wright J).
[71] Richardson, Ogus and Burrows, above n 12, 56.
[72] Ibid 17; Ashworth, Principles of Criminal Law, above n 12, 1; Hart, Punishment and Responsibility, above n 12, 236. Cf Wells, above n 18, 7–8.
[73] Ashworth, Principles of Criminal Law, above n 12, 1. Feinberg argues that penal sanctions for such regulatory offences do not constitute ‘punishment’, although the imposition of such sanctions visits ‘hard treatment’ on the offender because they do not convey reprobative symbolism, which he regards as the essence of punishment: Feinberg, above n 16.
[74] Vilheim Aubert, ‘White-Collar Crime and Social Structure’ (1952) 58 American Journal of Sociology 263; Wells, above n 18, 26.
[75] Richardson, Ogus and Burrows, above n 12, 18.
[76] Sanford Kadish, ‘Some Observations on the Use of Criminal Sanctions in Enforcing Economic Regulations’ (1963) 30 University of Chicago Law Review 423.
[77] Ibid 425–6.
[78] Ibid 445; Ashworth, Principles of Criminal Law, above n 12, 50; Richardson, Ogus and Burrows, above n 12, 17. Feinberg argues that the imposition of sanction without moral culpability does not generate any dilemma because sanctions for regulatory offences do not constitute ‘punishment’. This is because the essence of punishment is its reprobative symbolism, which is not attached to regulatory sanctions. However, he concedes that where the penalties are ‘severe’ then the argument loses cogency and cannot justify regarding the safeguards of culpability requirements and due process requirements as irrelevant: Feinberg, above n 16.
[79] Kadish, above n 76, 445. Cf Sir John Smith, Smith & Hogan Criminal Law (8th ed, 1996) 109.
[80] For this reason, Hart has observed that many modern retributivists concede that there ‘is a vast area of the criminal law where what is forbidden or enjoined by the law is so remote from the familiar requirements of morality’, that punishment is to be justified and measured mainly by utilitarian consequences: Hart, Punishment and Responsibility, above n 12, 236. See also Richardson, Ogus and Burrows, above n 12, 17.
[81] Richardson, Ogus and Burrows, above n 12, 16, 57. On the growth of strict liability offences, see Leonard Leigh, Strict and Vicarious Liability (1982) 14–16.
[82] Ashworth, Principles of Criminal Law, above n 12, 160.
[83] Richardson, Ogus and Burrows, above n 12, 17. Ashworth has pointed out the fallacy of this argument, however, on the basis that an assessment of fault is required (at least in relation to serious offences) in order for a court to pass sentence on a proper basis: Ashworth, Principles of Criminal Law, above n 12, 51, 163.
[84] Richardson, Ogus and Burrows, above n 12, 16–17.
[85] Kadish, above n 76, 444.
[86] The relevant variant of the deterrence model would then depend on whether the regulated conduct has the potential to be net welfare-enhancing (so that the ‘injury to others’ variant of the model should apply) or is always net welfare-reducing (so that the ‘unlawful gain’ variant of the model should apply).
[87] Hart, Punishment and Responsibility, above n 12, 1.
[88] Ibid 3.
[89] In TPC v CSR Ltd [1991] ATPR 41-076, the leading authority for the setting of Australian competition law penalties, French J fell into the trap of conflating the question of liability with the question of sanction. Thus he erroneously concluded that, because issues of morality are not relevant to the former question, they are not relevant to the latter question. See below Part III.
[90] Hart, Law, Liberty and Morality, above n 15, 90.
[91] Lucia Zedner, ‘Reparation and Retribution: Are They Reconcilable?’ (1994) 57 Modern Law Review 228, 243.
[92] The ALRC identified four different functions which a remedial court order for contravention of the TPA is designed to serve, which included ‘deterrence and, as a secondary or incidental outcome, retribution’: ALRC, Compliance with the Trade Practices Act 1974, above n 8, 29.
[93] Hadden refers to the British Law Commission’s study of strict liability for health and safety offences under the Factories Act (UK) which indicates that the presence of culpability was an important factor which enforcement officials took into account in exercising their discretion to prosecute: Tom Hadden, ‘Strict Liability and the Enforcement of Regulatory Legislation’ [1970] Criminal Law Review 496.
[94] John Brigham and Don Brown, ‘Distinguishing Penalties and Incentives’ (1980) 2 Law & Policy Quarterly 5.
[95] Feinberg, above n 16.
[96] ‘Two things are morally wrong: (1) to condemn a faultless man while inflicting pain or deprivation on him however slight (unjust punishment); and (2) to inflict unnecessary and severe suffering on a faultless man even in the absence of condemnation (unjust civil penalty)’: ibid.
[97] Hart, Punishment and Responsibility, above n 12, 25; Francis Allen, ‘The Decline of the Rehabilitative Ideal’ in Andrew von Hirsch and Andrew Ashworth (eds), Principled Sentencing (1998) 14, 17.
[98] This invocation of the proportionality principle differs slightly from the proportionality principle arising from the ‘just deserts’ model of punishment, because its primary concern is to avoid the use of oppressive or unfair state power against the citizen, rather than to ensure that the degree of censure or blame attributed to those who violate the law is commensurate with the seriousness of the violation: Hart, Punishment and Responsibility, above n 12, 181–2.
[99] Ibid 25.
[100] McKay, above n 55.
[101] Peter Grabosky, Clifford Shearing and John Braithwaite, ‘Introduction’ in Peter Grabosky and John Braithwaite (eds), Business Regulation and Australia’s Future (1993) 1, 6; Brigham and Brown, above n 94, 6.
[102] John Coffee Jr, ‘Does “Unlawful” Mean “Criminal”?: Reflections on the Disappearing Tort/Crime Distinction in American Law’ (1991) 71 Boston University Law Review 193.
[103] The use of the desert principle to locate the outer limits of penal sanctions has been advocated by several criminal law theorists: Frances Allen, The Decline of the Rehabilitative Ideal (1981); von Hirsch, ‘Doing Justice’, above n 45; cf von Hirsch, Censure and Sanctions, above n 44, ch 6. The proposed solution is a compromise one, and may be criticised by purists on that basis. However, I share the view of Judge Frank Easterbook who stated that,
there is no first best solution to the enforcement of anti-trust laws. If one existed in theory, it probably could not be achieved in the face of very expensive litigation. Fancy rules are costly to administer and they breed error. Only second-best solutions are available. This is my excuse for offering marginal adjustments in existing rules rather than some thoroughgoing revamp.
Frank Easterbrook, ‘Allocating Antitrust Decisionmaking Tasks’ (1987) 76 Georgetown Law Journal 305, 353.
[104] The use of negotiation between the regulator and the suspect firms and individuals has also been controversial: see Jeremy Thorpe, ‘Determining the Appropriate Role for Charge Bargaining in Part IV of the Trade Practices Act’ (1996) 4 Competition and Consumer Law Journal 69.
[105] Maureen Brunt, ‘The Australian Antitrust Law after 20 Years — A Stocktake’ in David Round (ed), Review of Industrial Organisation (1994) 483.
[107] Presley Warner and Michael Trebilcock, ‘Rethinking Price-Fixing Law’ (1993) 38 McGill Law Journal 679; Frank Easterbrook, ‘Detrebling Antitrust Damages’ (1985) 28 Journal of Law & Economics 445.
[108] For an analysis of vertical restraints, see F M Scherer and David Ross, Industrial Market Structure and Economic Performance (1990) 541–69.
[109] This distinction may be reflected in the way in which substantive rules are constructed. For example, horizontal price fixing is illegal per se (ie, on proof that the price fixing conduct occurred) while exclusive dealing is only unlawful if it has the purpose or likely effect of substantially lessening competition in a market: TPA ss 45, 45A, 47. See Richard Posner and Isaac Ehrlich, ‘An Economic Analysis of Legal Rulemaking’ (1974) 3 Journal of Legal Studies 257.
[110] Easterbrook, ‘Detrebling Antitrust Damages’, above n 107, 449–50.
[111] Posner, Antitrust Law, above n 36, 222.
[112] William Landes, ‘Optimal Sanctions for Antitrust Violations’ (1983) 50 University of Chicago Law Review 652, 653; Posner, Antitrust Law, above n 36; Easterbrook, ‘Detrebling Antitrust Damages’, above n 107; Kenneth Elzinga and William Breit, ‘Private Antitrust Enforcement: The New Learning’ (1985) 28 Journal of Law and Economics 405. The underlying reason why some economists claim that anti-competitive conduct may, in certain circumstances, be efficient and thus welfare-enhancing occurs because efficiency is not a unitary concept. Efficiency has three possible dimensions: allocative efficiency, productive efficiency and dynamic efficiency: see Re 7 Eleven Stores Pty Ltd [1994] ATPR 41-357, 42,677 (‘Re 7 Eleven’). Allocative efficiency arises when economic resources are allocated between different goods and services in precisely the quantity which consumers wish (their desires being expressed by the price they are prepared to pay on the market). Productive efficiency occurs when goods and services are produced at the lowest cost possible using current technology. Dynamic efficiency refers to the capacity of the market to adapt to change and incorporate new processes and products. There may be circumstances in which a reduction in allocative efficiency resulting from anti-competitive action may be outweighed by the gain in productive or dynamic efficiency accompanying such action.
[113] Some economists take the view that monopoly pricing can never be welfare enhancing. For example, Schwartz is unconvinced that the exercise of monopoly power could result in cost savings to the monopolist and thus generate social benefits, basing his view on the standard deadweight loss argument against monopoly: Schwartz, above n 33, 1081–2. On this view, it follows that competition law penalties should aim to ensure absolute deterrence, and thus the unlawful gain model for penalty setting is to be preferred. Yet Schwartz then goes on to argue that a loss-based penalty measure is to be preferred because a gain-based remedy will not fully reflect the harm to society caused by the monopolist’s conduct, and thus fails to achieve the objective of efficiently allocating enforcement resources to minimise the harm associated with monopoly. However, he concedes that if a loss-based remedy is adopted, there is a possibility that the gain to the violator may exceed the loss to others and therefore fail to deter monopoly conduct, but he claims that this will rarely occur in practice: Schwartz, above n 33, 1083. In other words, although Schwartz favours absolute deterrence of monopoly conduct, he nonetheless supports the injury to others model for calculating penalties rather than the unlawful gain model. Schwartz seems to conflate two distinct ideas: i) the quantum of penalty; and ii) the allocation of enforcement resources. He implicitly assumes that the allocation of enforcement resources will logically be determined by penalty size. But this is not an inevitable or necessary link.
[114] Posner favours the injury to others variant of the deterrence model, recognising that monopoly pricing may yield cost savings for the monopolist, so that if such cost savings exceed the deadweight loss to society which is caused by monopoly pricing, it is inappropriate to deter the conduct because it is net welfare-enhancing. The New Zealand Ministry of Commerce, however, rejects Posner’s reasoning without providing any adequate justification for so doing, simply stating that ‘[g]iven that this is a contentious argument and that the size of the deadweight loss is usually small compared with the monopoly profit, for the purposes of this document we have adopted the illegal gain rather than the “total harm” approach’: New Zealand Ministry of Commerce, Penalties, Remedies and Court Processes under The Commerce Act 1986: A Discussion Document (1998) 9.
[115] The adoption of two different penalty standards could generate some counter-intuitive results. If the gain to the violator in a price fixing case is very small, but the loss inflicted is very large, according to the absolute deterrence standard, the penalty should be that of the small gain. By contrast, if in a rule of reason case, the same dimensions of gain and loss arise, then according to the optimal deterrence model the penalty should be the large loss. From an outsider’s perspective, the result seems inequitable. But, this could be adjusted to allow for the greater loss-based penalty to be awarded in the price fixing case because the absolute deterrence model knows no concept of over-deterrence. It would therefore be possible to set penalties for price fixing at whichever is greater of the loss to others or the gain to the violators.
[116] Officer and Williams, above n 29, 157.
[117] For a discussion of the Australian institutional approach, see Brunt, above n 105, 510–12.
[118] However, conduct which is authorised or notified in relation to conduct prohibited by TPA ss 45, 45B, 47 and 50 is immune from contravening s 46: TPA s 46(6).
[119] Queensland Wire Industries Pty Ltd v The Broken Hill Proprietary Company Ltd [1989] HCA 6; (1989) 167 CLR 177; Frances Hanks and Philip Williams, ‘Implications of the Decision of the High Court in Queensland Wire’ [1990] MelbULawRw 4; (1990) 17 Melbourne University Law Review 437, 444–6.
[120] It is conceded that the assumption of zero legal error is unrealistic in practice: see above n 38 and accompanying text. Conduct which does not satisfy the ‘take advantage’ test and thus does not breach s 46 of the Act is not necessarily welfare-enhancing.
[121] Economists might criticise the Australian institutional scheme on the basis that requiring the public competition regulator to scrutinise ex ante all conduct alleged to be welfare-enhancing (assuming that an authorisation application is made) is more costly than a regime in which courts evaluate whether particular anti-competitive conduct is welfare-enhancing if challenged in litigation. The latter regime underlies the approach taken to misuse of monopoly power in s 46 of the Act — it places the onus on individual firms to assess for themselves whether their conduct is likely to be efficiency-enhancing: Warner and Trebilcock, above n 107. But such a criticism is empirical in nature, and it is by no means certain that an alternative institutional regime would be less costly in total than the existing regime.
[122] von Hirsch, Censure and Sanctions, above n 44, 19.
[123] Under European Community competition law, the negligent or intentional infringement of the prohibitions contained in arts 85(1) and 86 of the Treaty Establishing the European Economic Community as Amended by Subsequent Treaties, opened for signature 25 March 1957, 298 UNTS 3 (entered into force 1 January 1958) range from ECU 1000 to ECU 1 million or up to 10 percent of the turnover in the preceding business year of each of the undertakings participating in the infringement: Council Regulation No 17/62, OJ L13, 21 February 1962. Under US antitrust law, breach of the Sherman Antitrust Act of 1890, 15 USC ss1–7 (1993) is punishable by fines of up to US$10 million (for corporations) and US$350,000 (for individuals). In the US, the setting of fines for antitrust offences is subject to the guidelines laid down in the Federal Sentencing Guidelines 1998 (USA) pt R.
[124] The TPA expressly stipulates that penalties imposed for breach of Part IV of the TPA are not criminal in nature. Hence the rules of civil rather than criminal procedure apply in the conduct of penalty proceedings: TPA s 78.
[125] In recent years, however, the quantum of penalty has largely been determined by negotiation between the ACCC and the respondent, subject to the court’s approval: Thorpe, above n 104.
[126] The ALRC sought public submissions on whether the TPA should provide guidelines for courts imposing a penalty for contravention of the Act and what the content of those guidelines should be. In its final report, the ALRC noted that many of the submissions opposed the proposal on the ground that it would intrude too much on judicial discretion and was unnecessary. The ALRC concluded that detailed guidelines such as the Federal Sentencing Guidelines (USA) were unnecessarily complicated and inflexible. However, it favoured some guidance and proposed that three additional factors should be required to be taken into account by courts in assessing penalties: (if the contravener is a corporation) whether a revision of the contravener’s compliance controls is necessary and the extent to which defective internal controls have been revised in light of the lessons learned from the circumstances in which the contravention occurred; how much, if any, profit was made as a result of the contravention; and whether the contravener continued the relevant conduct notwithstanding a written request from the ACCC to discontinue it: ALRC, Compliance with the Trade Practices Act 1974 (1993), above n 62, Proposal 6.9; ALRC, Compliance with the Trade Practices Act 1974 (1987), above n 8, 124.
[127] See generally Arie Frieberg, ‘Monetary Penalties Under the TPA’ (1983) 11 Australian Business Law Review 4.
[128] [1991] ATPR 41-076 (‘CSR’).
[129] The factors considered by courts to be relevant to a determination of penalties for price fixing agreements have been analysed in detail by economists Round, Siegfried and Baillie, above n 10, 300–1.
[130] In particular, courts may consider the economic impact of the contravening conduct on the market — a greater effect warranting a heavier penalty. Cf TPC v Allied Mills Pty Ltd [1981] FCA 156; (1981) 37 ALR 256 with TPC v Malleys Ltd [1979] ATPR 40-118.
[131] Courts have taken the view that the larger the corporation involved in the contravention, the larger the appropriate penalty, on the basis that what would deter a small company might have little effect on a very large one: TPC v Carlton and United Breweries Ltd [1990] FCA 248; [1990] ATPR 41-037; TPC v TNT Australia Pty Ltd [1995] ATPR 41-375, 40,168 (‘TPC v TNT’).
[132] Systematic, deliberate defiance of the law is considered to warrant heavier penalties: TPC v TNT [1995] ATPR 41-375.
[133] The higher the level of management responsible for participation in the contravention, the more serious the infringement and the higher the level of penalty: Ibid 40,167.
[134] Cooperation with enforcement authorities has been regarded as a mitigating factor which justifies a reduction in penalty: Ibid 40,169–70.
[135] CSR [1991] ATPR 41-076, 52,152–3.
[136] Ibid 52,152.
[137] [1996] FCA 1134; (1996) 71 FCR 285 (‘NW Frozen Foods’).
[138] Ibid 296–7.
[139] Ibid 299.
[140] [1997] FCA 450; [1997] ATPR 41-562, 43,810.
[141] Ibid 43,811. The disagreement between individual Federal Court judges concerning whether the goal of penalty setting pursuant to s 76 of the TPA is limited to ‘deterrence’ or also includes ‘punishment’ seems to suggest that the courts are assuming deterrence to mean ‘absolute’ rather than ‘optimal’ deterrence. On this view, absolute deterrence arises from setting the quantum of penalty at the level at which the offender would be indifferent to committing the offence or not thereby eliminating any incentive to offend. A penalty which sought to inflict ‘punishment’, on the other hand, would be a sum of money which would make the offender prefer not to have committed the offence.
[142] Ibid 43,811.
[143] [1998] FCA 310; [1998] ATPR 41-628.
[144] Ibid 40,891.
[145] This error is referred to by Goldberg J in ACCC v Australian Safeway Stores Pty Ltd [1997] FCA 450; [1997] ATPR 41-562, 43,811.
[146] [1989] HCA 6; (1989) 167 CLR 177, 194 (Deane J). See also TPC v Commodore Business Machines Pty Ltd [1989] ATPR 40-976, 50,672; TPC v TNT [1995] ATPR 41-375.
[147] See above Part II(E).
[148] The other criteria which appear to have no apparent relevance to the economic model are: the nature and extent of the contravening conduct, the circumstances in which the conduct took place, whether the contravention arose out of the conduct of senior management or at a lower level, and whether the company has shown a disposition to cooperate with the authorities responsible for the enforcement of the Act in relation to a contravention. However, note that in Round’s study, economists could simply assign economic value to judicial criteria as proxies for economically relevant variables: Round, Siegfried and Baillie, above n 10.
[150] CSR [1991] ATPR 41-076, 52,152.
[151] TPC v TNT [1995] ATPR 41-375, 40,165 (Burchett J).
AustLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.austlii.edu.au/au/journals/MelbULawRw/1999/18.html