From Arbitrariness to Coherency in Sentencing: Reducing the Rate of Imprisonment and Crime While Saving Billions of Taxpayer Dollars

Mirko Bagaric January, 2014

Dealing with criminals and preventing crime is a paramount public policy issue. Sentencing law and practice is the means through which we ultimately deal with criminal offenders. Despite its importance and wide-ranging reforms in recent decades, sentencing remains an intellectual and normative wasteland. This has resulted in serious human rights violations of both criminals and victims, incalculable public revenue wastage, and a failure to implement effective measures to reduce crime. This Article attempts to bridge the gulf that exists between knowledge and practice in sentencing and lays the groundwork for a fair and efficient sentencing system. The Article focuses on the sentencing systems in the United States and Australia. The suggested changes would result in a considerable reduction in incarceration numbers, lower crime, and a reduction in the expenditure on prisons. The key concrete recommendations of this Article are that the criminal justice system should move towards a bifurcated system of punishment, reserving imprisonment mainly for serious sexual and violent offenses and reducing the terms of imprisonment in general.

Introduction

Individuals who deliberately harm others constitute the greatest threat to the safety and security of other people and the community in general. Much is at stake in terms of how we deal with criminals. The sharp end of public policy relating to criminals is sentencing. Mistakes in sentencing potentially undermine our entire approach to dealing with those who present the greatest risk to our safety. Despite considerable reforms to the sentencing landscape in recent decades, there still remains a large gap between knowledge and practice in this realm.

Sentencing law and practice has changed , but it has not been strategically developed , and, on any rational measure, it has not improved . The most profound change has been the imposition of increasingly harsh penalties. The bottom line in terms of the effect of the changes has been a massive increase in prison numbers and a corresponding increase in the amount of public money required to fund prisons. The United States, for example, has the highest incarceration rate in the world—currently more than two million individuals are in confinement.¹

This has resulted in increasingly commonplace and loud calls for radical change.² Vivien Stern, former secretary general of Penal Reform International, for example, states, “Among mainstream politicians and commentators in Western Europe, it is a truism that the criminal justice system of the United States is an inexplicable deformity.”³

High imprisonment rates, however, are not necessarily a sure sign of sentencing policy failure. This is especially the case given that the increased incarceration rate has coincided with a significant reduction in the crime rate in the United States.⁴ It is incontestable that imprisonment forms a necessary component of any sentencing system, and there is no optimum incarceration rate. In theory, it may be that the United States should be imprisoning even more criminals.⁵

This Article lays the blueprint for a rational, evidence-based system of sentencing.⁶ It explains the appropriate objectives of sentencing, and, on the basis of existing empirical data, it details the manner in which these objectives can be best achieved.

I undertake this process against the backdrop of two existing sentencing systems: those currently operating in the United States and Australia. These systems are examined because they both have undergone considerable change over the past thirty years. The United States and Australia are flourishing, wealthy, highly-educated countries with the capacity to make informed, intelligent, and evidence-based public policy decisions. Yet their systems have greatly diverged when it comes to sentencing. The contrast is illuminating beyond revealing the apparent arbitrariness of sentencing policy. It transpires that there is much to learn from each system—and perhaps even more to avoid. A fusion of key aspects of each system could lead to a vastly improved overall sentencing system. The key is deciding which components of each system should be abolished and which should be retained with appropriate adaptations.

Comparing sentencing systems in Australia and the United States is also useful because the United States is a high incarceration nation, whereas, on an international scale, Australia is in the mid-range. The United States, in fact, has the highest incarceration rate in the world: nearly 750 per one hundred thousand in the adult population.⁷ In terms of expenditure, it costs taxpayers in the United States on average $45,000 to house a prisoner for one year. The total spending on prisons is now over $50 billion annually.⁸ By comparison, in Australia, the imprisonment rate is 168 people per one hundred thousand in the adult population;⁹ however, the cost of imprisonment in Australia is nearly double that of the United States: $79,000 per prisoner per year.¹⁰

In Part I of this Article, I set out the framework for a coherent and rational sentencing system and discuss some of the key obstacles that have frustrated the implementation of this model. In Part II, I provide an overview of the current sentencing systems in Australia and the United States, identifying their main flaws and strengths. Following this discussion, Part III examines the efficacy of these state systems at achieving the key stated goals of sentencing: incapacitation, general deterrence, specific deterrence and rehabilitation. This Part of the Article is critical to developing a coherent sentencing model.

The main formal change to sentencing determinations in the United States in the past few decades is the adoption of mandatory or presumptive sentences. The sentencing discretion in Australia, by contrast, remains largely unfettered. Thus I conclude in Part IV that a model sentencing system should embrace the fixed or presumptive system, but the content of the system should differ considerably from the current orthodoxy in the United States. Sentence lengths need to be principally guided by the proportionality principle: the harm caused by the offense should be matched by the harshness of the penalty. Proportionality finds far greater expression in the Australian sentencing system than in the United States,¹¹ and embedding proportionalism into sentencing is the main feature of the Australian system that should be transported to a model system of sentencing.

A major distinction between sentencing law and other legal areas— and one of the reasons for the ongoing policy dysfunction in sentencing— is the largely forward-looking nature of the sentencing inquiry. It has some backward-looking aspects in terms of focusing on offense severity and the profile of the offender (such as his or her criminal history), but objectives such as deterrence, incapacitation, and rehabilitation involve predictions and modelling which necessarily lead to approximations and speculation. Abolition of forward-looking aspects would make sentencing clearer and simpler. As discussed below, empirical data suggests we should be moving towards this end.

Part V thus concludes that most of the existing aggravating and mitigating considerations that currently exist should be abolished. There should not be any penalty enhancements or reductions in order to pursue any other objectives than the proportionality principle. The main qualification to this is that there should be a penalty premium in the order of 20 to 50 percent for recidivists convicted or serious sexual or violent offenses. The upshot of this model sentencing system would see a considerable reduction in the number of people imprisoned for nonsexual or nonviolent offenses and an increase in the efficiency and fairness of the system, thus releasing billions of public revenue dollars to be spent on activities other than prisons, while at the same time reducing crime (if some of the saved revenue is allocated to increasing police street presence). In Part VI, I outline the framework for a model sentencing system and in the concluding remarks I summarize the key aspects of my proposal.

I. Objectives of a Model Sentencing System

In this Part, I undertake an overarching examination of the considerations that should influence and shape sentencing policy. I first consider the manner in which theories of punishment impact how we deal with criminals. This is followed by a discussion of the role of public opinion in sentencing. After concluding that neither of these considerations greatly impact sentencing reform, I set out the four key objectives that a sentencing system should strive to achieve.

A. Justification of Punishment

At the most fundamental level, the starting point in setting a framework for sentencing is to establish that a system of state-imposed punishment is justifiable. Punishment in this context is the study of the connection between wrongdoing and state-imposed sanctions. The main issue raised by the concept of punishment is the basis upon which the harms administered by the state to offenders can be justified. Thus sentencing and punishment are inextricably linked, with punishment being the logically prior inquiry.

While, theoretically, the threshold issue regarding any sentencing system is the whether punishment is justifiable, for pragmatic reasons this discussion is not canvassed at length in this Article. Some commentators have advocated for the abolition of punishment altogether.¹² There are numerous reasons that abolitionists give for this view. One common theme is that a state that imposes criminal sanctions, rather than protecting society from harmful acts, usually provokes criminality and, in this way, punishment is destructive to society.¹³ It has also been charged that punishment is inherently unfair because it is employed mainly against the underprivileged and deprived sectors of the community: “rulers will never prosecute their own class associates. Or at least, it is very exceptional.”¹⁴

Irrespective of the cogency of such arguments, this debate has been lost. The institution of state-imposed sanctions for criminal behavior is such an entrenched part of our social fabric that no amount of philosophical persuasion is likely to lead to its eradication. It is more productive to make the system of punishment the best it can be, rather than seeking to rail against it.

This requires a greater understanding of the key rationales in support of punishment. To this I now turn. There are two main theories of punishment. Utilitarianism is the view that punishment is inherently bad due to the pain it causes the wrongdoer but is ultimately justified because this is outweighed by the good consequences stemming from it. These are traditionally thought to come in the form of incapacitation (i.e., imprisoning offenders and thereby preventing them from further offending), deterrence (discouraging further offending), and rehabilitation (inducing positive attitudinal reform).¹⁵ The utilitarian theory of punishment has fallen out of favor for two main reasons. The first is the perceived inability of the sentencing process to achieve the utilitarian penal objectives of incapacitation, deterrence, and rehabilitation.¹⁶ The second is the view that utilitarianism could lead to abhorrent practices, such as punishing the innocent.¹⁷

The main competing theory, and the one which enjoys the most contemporary support, is retributivism. Retributive theories of punishment are not clearly defined, and it is difficult to isolate a common thread running through theories carrying this label.¹⁸ All retributive theories assert that offenders deserve to suffer, and that the institution of punishment should inflict the suffering they deserve. However, they provide divergent accounts of why criminals deserve to suffer.¹⁹

While retributivism is the orthodox theory of punishment, I have previously argued that it is doctrinally flawed principally because it can only justify punishment by reference to consequential benefits stemming from punishment, mainly in the form of deterring crime.²⁰ I have also argued that the criticisms of a utilitarian theory of punishment have been overstated and that in fact utilitarianism is the most persuasive theory of punishment.²¹

It is important to be mindful of the differences between the utilitarian and retributive theories of punishment because of the possible implications they can have for sentencing offenders. However, as a result of the empirical data regarding the efficacy of punishment to achieve stated utilitarian objectives of punishment, it emerges that in fact there is no meaningful pragmatic difference between retributivism and utilitarianism so far as the design of a rational sentencing system is concerned. As noted in Part IV, the key focus of retributivism is to ensure there is proportionality between the punishment and the crime. This matching is potentially distorted in a utilitarian calculus by the need to achieve other objectives, namely general deterrence, specific deterrence, and rehabilitation. However, as is discussed at length in Part III of this Article, punishment is largely incapable of achieving these objectives (apart from absolute general deterrence) and hence the amount of punishment should not be influenced by these considerations. As is established in Part IV, both theories endorse the pursuit of proportionate sentences as a principal sentencing requirement. The potential theoretical divergence that the key theories of punishment have for sentencing has been negated by what the empirical data shows can be achieved through sentencing. Accordingly, it is unnecessary to explore this philosophical realm further for the purposes of this Article.

B. The Impact that Public Opinion Should Have on Sentencing Design

Sentencing is often a controversial issue that engages the community. Accordingly, any effective sentencing design must deal with democratic reality. Thus it is imperative to first ascertain how much weight should be given to public opinion, at least in principle, and then assess whether that calculation is achievable.

Some theorists have contended that public opinion should have a cardinal sentencing role. The satisfaction theory of retributivism provides that punishing wrongdoers satisfies “the feeling of hatred—call it revenge, resentment, or what you will—which the contemplation of such conduct excites in healthy constituted minds.”²² It thereby diminishes the prospect of harmful vendettas by victims and their associates, who may be tempted to exact their own revenge. On occasions it has received judicial support:

One of the objects of punishment, and by no means the least important object of punishment, is to prevent, so far as possible, the victims of crime from taking matters into their own hands. It is no great step from private vengeance to vendetta, and there is no knowing where the vendetta will stop.²³

However, at its highest, this theory merely justifies some punishment for offenders; it says nothing about how much punishment is appropriate or necessary. Accordingly, it does not provide a foundation for public opinion having a significant role in setting offense severity.

A closely associated view to the satisfaction theory is that public opinion must be factored into the sentencing regime—otherwise a lack of confidence in the courts will result and, ultimately, less respect for, and compliance with, the law. Certainly, numerous surveys across a number of different jurisdictions, including Scotland, the United States, Germany, and England, have shown that most members of the community believe sentences are too lenient.²⁴ Despite this, there is no empirical data showing that lenient sentences actually cause diminished legal compliance.²⁵

Moreover, when it comes to punishing criminals, pioneering studies into the human condition establish that the community is accepting of lower penalties if this is in its self-interest. An empirical study by Swissbased scientists suggests that humans are wired in such a way that we derive pleasure from punishing wrongdoers.²⁶ This is even so where there are no benefits stemming from the punishment to the punisher. The study is part of attempts in “neuroeconomics” and the “cognitive neuroscience of social behavior” to understand the social brain and the associated moral emotions.

The study found that “most people seem to feel bad if they observe that norm violations are not punished, and they seem to feel relief and satisfaction if justice is established.”²⁷ The study focused on what is termed “altruistic punishment,” where the punishment involves costly acts that are of no benefit to the punisher but (potentially) confer benefits on other individuals by discouraging the wrongdoer from behaving in a similar manner in the future.²⁸ It found that the reason that individuals punish violators of widely approved norms, although they reap no offsetting material (or other) benefits themselves, is because they derive satisfaction from the punishment of norm violators. The part of the brain associated with enjoyment or satisfaction (the dorsal striatum) was activated when test subjects imposed a penalty on a wrongdoer.²⁹

The research further showed that altruistic punishment is not a reflexive response, like digesting food. Rather it requires a motivation and is based on an intention to do so. To this end, the researchers noted:

The typical proximate mechanism for inducing motivated action is that people derive satisfaction from the action. Most people seem to feel bad if they observe that norm violations are not punished, and they seem to feel relief and satisfaction if justice is established. Many languages even have proverbs indicating such feelings, for example, “Revenge is sweet.”³⁰

However, significantly, the experiment showed that when it comes to imposing punishment, we are not slaves to our feelings. A second part of the brain (the prefrontal cortex) is activated when players need to weigh the satisfaction derived from punishment against the monetary cost of punishing.³¹ Thus, although the desire to punish has a primal origin, in the end this can be controlled. As one of the co-authors of the study, Ernst Fehr, noted, “This all looks pretty rational. . . . People seem to trade off the expected satisfaction from punishing with the cost of punishing in quite a rational way.”³²

The results show that the higher the cost of the punishment, the lower the actual punishment imposed.³³ The main implication of the research findings relates to the relevance of public opinion to sentencing practice. The findings support the view that public sentiment, which seems to support increasing tougher sanctions, can be curtailed if the public is informed that punishment comes at a cost to the community. Accordingly, there is no demonstrable reason that public opinion must drive sentencing policy or procedure.

A more sophisticated justification for incorporating public opinion into sentencing is that the seriousness of a crime is, at least to some degree, determined by the extent to which it “offends community mores.”³⁴ Roberts provides examples of flag burning and hate crime offenses and concludes that the social nature of sentencing and the cultural relativity associated with crime severity provide a basis for public opinion to some degree informing offense severity.³⁵

It is clear that cultural considerations inform not only sentencing, but also criminality. However, it is not desirable that this practice should continue. Criminal sanctions involve the direct infliction of pain on individuals and, hence, require a moral justification. This cannot be achieved through norms based on regional and transient customs. We should be aspiring to an objective value system that is not dependent on historical and cultural customs and rituals.³⁶ Allowing community mores to inform crime and sentencing has the capacity to entrench and promulgate existing practices, which, by any measure, are inhumane and even savage. It is for this reason that in some parts of the world women accused of adultery are executed, including by stoning,³⁷ and people are imprisoned for homosexual acts in sixty-nine countries.³⁸

Sentencing is a purposive social endeavor that must be guided by rational inquiry, not raw impulse. It should have underlying principles that govern the way it ought to be administered. These are ascertained through a process of inductive and deductive logic and by analyzing the relevant empirical evidence to determine what objectives are, and are not, achievable through a system of state-imposed punishment. Guidance on sentencing matters should be sought from experts in the field, but not the uninformed.

Seeking public views on sentencing is analogous to doctors basing treatment decisions on what the community thinks is appropriate or to engineers building cars on the basis of what lay members of the community “reckon” seems about right rather than in accordance with the rules of physics. It is legal commentators, practitioners, and other experts (namely, criminologists, penologists, sociologists, moral philosophers, and econometricians) who should be educating the public about how to frame a sentencing system—not the other way around.

This is not to dismiss the important role of the community, especially victims of crime, in determining sentencing outcomes. While the unreflective views of victims and the community cannot drive sentencing outcomes, victims need to be listened to because, as discussed in Part IV of this Article, they provide us with the best information regarding the impact that criminal behavior has on well-being.

While it “feels” good to severely punish wrongdoers, satisfying our emotions does not provide a justification for deliberately inflicting hardships on (albeit flawed) members of our community. That is why we do not condone road rage, jealous lovers killing their partners, or frustrated parents beating their children. The level of civilization of a society roughly correlates with the extent to which it suppresses its feelings in relation to its law making process.

Accordingly, while reflexively it is assumed that public sentiment should be a powerful influence in sentencing standards, the reality is otherwise: the community is not incorrigibly punitive and there is no principle reason for allowing its (non-expert) views to guide sentencing policy.

C. The Rational Objectives of Punishment

Given that public opinion should not drive sentencing policy and that neither of the key theories of punishment mandate prescriptive objectives (apart from matching the punishment and the crime), there is a relatively blank canvas from which sentencing objectives can be established. To this end, it is essential that the objectives are clear and incontestable— framework errors threaten the viability of the entire structure.

Ultimately, the main objective relating to all aspects of the criminal justice system is to stop crime. If it cannot be stopped, then at least it should be reduced. The second goal is to punish criminals appropriately. As discussed in Part I.B of this Article, humans have an innate desire to punish wrongdoers—although this desire is not intractable. The concept of “appropriate” punishment is discussed at length in Part IV, which establishes that this is not only instinctively appealing but also doctrinally sound.

There are two other objectives of sentencing policy. These stem from the impact that sentencing necessarily has on matters of considerable community and individual importance. First, criminal sanctions involve public expenditure. This exceeds hundreds of billions of dollars annually.³⁹ Secondly, criminal sanctions by their very nature involve the deliberate infliction of pain.⁴⁰ Often this involves the deprivation of liberty—either total (imprisonment) or partial (probation). This injects moral considerations into the inquiry.

Thus, the sentencing system should aim to achieve (only) four goals:

(1) To stop or reduce crime; (2) To punish criminals appropriately; (3) To minimize the cost of the system; and (4) To ensure that the system does not violate important moral

prescriptions.

The rest of this Article sets out how best these goals can be achieved. A caveat to this is that I do not consider the desirability of capital punishment. The United States is the only developed nation apart from Japan that still imposes the death penalty. The literature and analysis regarding the desirability of the death penalty is voluminous. It can only be examined in the context of a stand-alone dissertation focusing on this issue. This is not a meaningful limitation to this paper given that not all states impose the death penalty⁴¹ and only a relatively very small number of criminals are executed in the United States.⁴²

II. Current sentencing systems—United States

and Australia

A. United States

Each state of the United States has its own sentencing system,⁴³ and there is considerable divergence across the respective regimes. The federal jurisdiction also has a discrete sentencing system, which is important because of the large number of offenders sentenced under this system and the important doctrinal influence it has at the state level.⁴⁴ Despite the sentencing variations across the United States, several key commonalties and themes exist. This Part summarizes these in order to provide a backdrop to the reform recommendations.

The key distinguishing aspect of the United States sentencing system compared to that of Australia (and most other sentencing systems in the world) is the wide-ranging use of mandatory minimum penalties, which in some form exist in all U.S. states.⁴⁵ As noted by Berman and Bobas, “[O]ver the last half-century, sentencing has lurched from a lawless morass of hidden, unreviewable discretion to a sometimes rigid and cumbersome collection of rules.”⁴⁶

These mandatory minimum penalties are often set out in sentencing grids, which typically use a criminal history score⁴⁷ and offense seriousness to calculate the appropriate penalty. None of these policies and practices emanated from a clear theoretical foundation, but rather stemmed from “back-of an-envelope calculations and collective intuitive judgements [sic].”⁴⁸ Despite this, there is a convergence of approach:

Modern sentencing reforms have repudiated rehabilitation as a dominant goal of sentencing. Many structured sentencing laws, including many guideline sentencing systems and severe mandatory minimum sentences, are designed principally to deter, incapacitate, and punish offenders.⁴⁹

The most extensively analyzed fixed penalty laws are the United States Sentencing Commission Guidelines Manual (Federal Sentencing Guidelines).⁵⁰ These Federal Sentencing Guidelines are no longer mandatory in nature, following the United States Supreme Court decision in United States v. Booker. ⁵¹ However, sentences within guideline ranges are still imposed in approximately 60 percent of cases.⁵² The set penalties apply to most types of offenses, including drug, fraud, and immigration crimes. A United States Sentencing Commission Report in 2011 noted that the number of offenses with set terms is increasing and the terms were increasing.⁵³

This comes against the backdrop of an increasing emphasis on incapacitation as the key sentencing objective. Perhaps the greatest indication of the harshness of United States sentencing is that life without parole is mandatory upon conviction of at least one specified offense in twentyseven states.⁵⁴ There are over forty thousand prisoners in the United States serving life without parole. This greatly exceeds the number of such prisoners in the rest of the world. In Australia, for example, there are only fifty-nine prisoners serving life without parole.⁵⁵

However, at least formally, incapacitation does not overwhelm the sentencing objectives in the United States. At the federal level deterrence and rehabilitation also guide sentencing policy. For example 18 U.S.C. § 3553(a) provides:

Factors To Be Considered in Imposing a Sentence.— the court shall impose a sentence sufficient, but not greater than necessary, to comply with the purposes set forth in paragraph (2) of this subsection. The court, in determining the particular sentence to be imposed, shall consider— . . . (2) The need for the sentence imposed— . . . (B) To afford adequate deterrence to criminal conduct; (C) To protect the public from further crimes of the defendant;

and

(D) To provide the defendant with needed educational or vocational training, medical care, or other correctional treatment in the most effective manner.

In terms of establishing the appropriate sentence, apart from the offense severity, the other key variable that determines the sanction is the prior history of the offender.⁵⁶ In relation to most offenses, a criminal history can approximately double the presumptive sentence. For example, an offense at level 14⁵⁷ in the Federal Sentencing Guidelines carries a presumptive penalty for a first offender of imprisonment for 15 to 21 months, which increases to 37 to 46 months for an offender with 13 or more criminal history points.⁵⁸ For an offense at level 36, a first offender has a presumptive penalty of 188 to 235 months, which increases to 324 to 405 months for an offender with the highest criminal history score. Thus, a bad criminal history can add between 136 to 170 months (over fourteen years) to a jail term.

A relatively well-known state-level presumptive sentencing system is the grid system in Minnesota,⁵⁹ which also utilizes the same two core variables in arriving at a sentence. The vertical axis of the grid lists the severity levels of offenses in descending order (there are ten different levels). The horizontal axis provides a (seven level) criminal history score, which reflects the offender’s criminal record.⁶⁰ The presumptive sentence appears in the cell of the grid at the intersection of the offense score and the offender score. Where the sentence is one of imprisonment, a precise period is indicated, as is a range within which a court can sentence an offender without it being regarded as a departure.⁶¹ The range allows for the operation of aggravating and mitigating circumstances, other than those relating to an offender’s prior criminal history. Sentences may only be imposed outside the range where substantial and compelling circumstances exist.⁶² In Minnesota,⁶³ prior convictions also can mean a considerable difference in ultimate disposition. For example, for a first offender convicted of theft (over $500) or non-residential burglary, a court can impose a non-custodial sentence,⁶⁴ whereas for an offender with “a criminal history score”⁶⁵ of six or more, the presumptive sentences are twenty-one months and thirty months, respectively.

Some of the harshest types of mandatory sentencing laws are the three-strikes laws, which have been adopted in over twenty states.⁶⁶ The California three-strikes laws,⁶⁷ which were reformed in 2012, are the most well-known.⁶⁸ Prior to these reforms, offenders convicted of any felony who had two or more relevant previous convictions were required to be sentenced to between twenty-five years to life imprisonment. The importance attributed to previous convictions was exemplified by the fact that the current offense did not have to be for a serious and violent felony— any felony would do. This meant that some offenders were sentenced to decades in imprisonment for relatively minor crimes. Defendants have been sentenced to twenty-five years to life where their last offense was for a minor theft (which, prior to the three-strikes regime, would normally have resulted in a non-custodial sentence). For example, Jerry Dewayne Williams, a twenty-seven-year-old Californian was ordered to be imprisoned for twenty-five years to life without parole for stealing a slice of pepperoni pizza from a group of four youths, based on his previous convictions.⁶⁹ Gary Ewing was sentenced to twenty-five years to life for stealing three golf clubs, each of which was worth $399. Prior to that he had been convicted of four serious or violent felonies.⁷⁰

The California three-strikes law was somewhat softened in 2012, such that a term of at least twenty-five years would only be required where the third offense was a serious or violent felony.⁷¹ In such cases, offenders continue to receive a significant premium—they must be sentenced to double the term they would have otherwise received for the instant offense.⁷² Thus, despite the softening of the laws, they still provide severe penalties for serious and violent offender third-strikers.

Hessick accurately observes that most legislation does not define aggravating or mitigating factors and instead simply states matters that increase or reduce offense severity.⁷³ She observes that the Federal Sentencing Guidelines stipulate thirteen considerations that are aggravating (i.e. justify an upward sentencing departure) and eight that are mitigating.⁷⁴ Aggravating considerations include whether the crime results in death or physical injury, the use of weapons, or gang activity.⁷⁵ Mitigating factors include assistance to authorities, the existence of coercion and duress as underpinning the crime, diminished capacity, and the voluntary disclosure of the crime.⁷⁶ In fact, most state non-capital sentencing regimes that identify aggravating and mitigating factors tend to stipulate more aggravating than mitigating considerations.⁷⁷

Proportionality is a requirement of the sentencing regimes of ten states in the United States.⁷⁸ The precise considerations that inform the proportionality principle vary in those jurisdictions, but generally there are six relevant criteria:

(1) Whether the penalty shocks a reasonable sense of decency; (2) The gravity of the crime; (3) The prior criminal history of the offender; (4) The legislative objective relating to the sanction; (5) A comparison of the sanction imposed on the accused with

the penalty that would be imposed in other jurisdictions; and

(6) A comparison of the sanction with other penalties for similar and related offenses in the same jurisdiction.⁷⁹

The prohibition against cruel and unusual punishment in the Eighth Amendment has been applied sparingly in the sentencing domain. To the extent that it has been applied in this area, it has been mainly in relation to proscribing the death penalty for non-homicide offenses⁸⁰ and juvenile offenders.⁸¹ In determining the scope of this limitation, the Supreme Court has taken into account international standards relating to appropriate levels of punishment.⁸²

In relation to non-capital sentences, the Supreme Court has endorsed the concept of proportionality as being a constraint on the level of punishment, but the concept has not been developed with any degree of precision and can only be invoked to prohibit sanctions which contain “gross disproportionality.”⁸³

Thus, the clear trend of United States sentencing changes over the past few decades has been a move to increasingly harsh penalties and a diminishing role for judicial discretion resulting in an inevitable significant increase in prison numbers.

B. Australia

I now provide an overview of the current Australian sentencing landscape. While sentencing law differs in each Australian jurisdiction, considerable convergence exists in relation to important areas. Sentencing in each of the nine Australian jurisdictions (the six states, the Northern Territory, the Australian Capital Territory and the Federal jurisdiction) is governed by a combination of legislation and the common law.

Each statute deals with three main dimensions of sentencing. First, it sets out the purposes and aims of sentencing.⁸⁴ Second, it describes aggravating and mitigating factors that affect sentencing. There is a considerable degree of variation in the extent to which these factors are set out in each individual statute. These considerations are set out most expansively in the Crimes (Sentencing Procedure) Act 1999 (New South Wales (NSW)),⁸⁵ which lists thirty relevant factors. Most sentencing statutes only sparsely deal with these considerations. This is not, however, indicative of a legal divergence between the respective jurisdictions. This is because, as is discussed below, aggravating and mitigating factors are mainly defined by the common law, which continues to apply in all jurisdictions.⁸⁶

The third main aspect covered by the sentencing statutes is the type of sanctions that can be imposed on offenders. These are similar across Australia.⁸⁷ Typically, there is a range of sanctions; however, in essence, there are four different types. The least serious is a finding of guilt without any further harshness being imposed on the offender apart from a promise to the sentencing court not to re-offend during the period of the order, which is typically in the range of twelve months. These sanctions are variously labeled as dismissals, discharges, or bonds.⁸⁸ The second, and most common, sanction imposed in Australia is a fine, which is a monetary exaction against the offender.⁸⁹ The third, and harshest, form of punishment in Australia is imprisonment.⁹⁰ The fourth general form of sanction consists of what are collectively known as intermediate punishments.⁹¹ These are generally imposed when the offense is too serious to be dealt with by a fine, but is not serious enough to warrant a term of imprisonment. Intermediate sanctions often involve a work component and an order to undertake some form of counseling or training, which is designed to have a rehabilitative effect. These come under various labels including community-based orders, home detention, suspended sentences, and intensive corrections orders.⁹²

The overarching methodology and conceptual approach that sentencing judges undertake in making sentencing decisions is the same in each jurisdiction. This approach is known as “instinctive synthesis.” The term originates from the Full Court of the Supreme Court of Victoria decision of R v Williscroft ,⁹³ nearly forty years ago where Justices Adam and Crockett stated, “Now, ultimately, every sentence imposed represents the sentencing judge’s instinctive synthesis of all the various aspects involved in the punitive process.”⁹⁴

The general methodology for reaching sentencing decisions has been considered by the High Court on several occasions, and the Court has consistently adopted the instinctive synthesis approach and rejected the alternative, which is normally referred to as the two-step approach. The alternative approach involves a court setting an appropriate sentence commensurate with the severity of the offense and then making allowances up and down, in light of relevant aggravating and mitigating circumstances.⁹⁵

In Wong v R ⁹⁶ most members of the High Court saw the process of sentencing as an exceptionally difficult task with a high degree of “complexity.”⁹⁷ Exactness is supposedly not possible because of the inherently multi-faceted nature of that activity.⁹⁸

As a result of this approach, there is no single sentence, which is accurate in a particular case. In Markarian v R , four justices of the High Court (Gleeson, CJ; Gummow, Hayne, and Callinan, JJ) noted that the sentencing judgment is a “discretionary judgment and . . . there is no single correct sentence.” On the contrary, there is a “sentencing range” within which views can reasonably differ as to the appropriate sentence.⁹⁹

Broadly, each jurisdiction endorses the same objectives of sentencing in the form of community protection (or incapacitation), general deterrence, specific deterrence, rehabilitation, and denunciation.¹⁰⁰ These overarching purposes, however, are generally too abstract to provide meaningful guidance. They often are in conflict¹⁰¹ and, as is further discussed below, it is not clear that in fact several of them are attainable through the sentencing process.

All jurisdictions have adopted the proportionality principle. In a clear statement of the principle of proportionality, the High Court said in Hoare v The Queen , “A basic principle of sentencing law is that a sentence of imprisonment imposed by a court should never exceed that which can be justified as appropriate or proportionate to the gravity of the crime considered in light of its objective circumstances.”¹⁰²

In Veen (No. 1) v The Queen ¹⁰³ and Veen (No. 2) v The Queen ,¹⁰⁴ the High Court also stated that proportionality is the primary aim of sentencing. In fact, it is considered so important that it cannot be trumped even by the goal of community protection, which at various times has also been declared as the most important aim of sentencing.¹⁰⁵ Thus, in the case of dangerous offenders, while community protection remains an important objective, at common law it cannot override the principle of proportionality. It is for this reason that preventive detention is not sanctioned by the common law.¹⁰⁶

Proportionality has also been given statutory recognition in all Australian jurisdictions.¹⁰⁷ However, there are a number of statutory incursions into the proportionality principle. These mainly stem from the trend towards tougher sentences. In Victoria, for example, serious sexual, drug, arson, or violent offenders¹⁰⁸ may receive sentences in excess of that which is proportionate to the offense. Indefinite jail terms may also be imposed for offenders convicted of “serious offenses,”¹⁰⁹ where the court is satisfied “to a high degree of probability that the offender is a serious danger to the community.”¹¹⁰ Similar provisions to those operating in Victoria regarding serious violent and sexual offenders¹¹¹ and indefinite sentences also exist in most jurisdictions for offenders who are assessed as constituting a serious danger to the community.¹¹²

It is important to emphasize that while the proportionality principle is a key aspect of the Australian sentencing system, this is more in abstract than reality. As noted in Part IV of this Article, the proportionality principle is lacking in content and is more akin to an aspirational ideal than a guiding prescription. Australian courts have steered clear of addressing the considerable jurisprudential challenges related to the proportionality doctrine, however this has not dampened their enthusiasm in the abstract for the doctrine.

Another commonality in all jurisdictions is that aggravating and mitigating factors operate relatively uniformly throughout the country, despite the different ways in which they are dealt with by statute. These considerations stem mainly from the common law and are continually evolving. There are between two hundred and three hundred such factors.¹¹³ Key mitigating considerations include: a plea of guilty; assistance to law enforcement authorities; remorse; voluntary cessation from offending; voluntary disclosure of crime; psychiatric and psychological illness; intellectual disability; youth; good prospects of rehabilitation; previous good character; onerous prison conditions; forgiveness of the victim; and an offense committed under duress.¹¹⁴ Important aggravating factors are: prior criminal record; significant level of injury or damage caused by the offense; vulnerability of victim; high level of planning; offenses committed while on bail or parole; offenses committed with others; and breach of trust.¹¹⁵

The large number of aggravating and mitigating factors is one of the key reasons that it is not possible to predict with confidence the exact sentence that will be imposed in any particular case. A degree of certainty is, however, injected by the fact that two factors attract a numerical discount. An earlier guilty plea can, in most jurisdictions, attract a 25 percent discount,¹¹⁶ while a guilty plea coupled with assistance to authorities can result in a 50 percent reduction in penalty.¹¹⁷

Fixed penalties for serious offenses in Australia are rare.¹¹⁸ However, two jurisdictions have “standard penalties” for a number of offenses. The most comprehensive standard penalty provisions are in New South Wales (NSW), which were enacted by the Crimes (Sentencing Procedure) Amendment (Standard Minimum Sentencing) Act 2002 (NSW).¹¹⁹ Division 1A of Part 4 (§ 54A(2)) sets standard non-parole periods for a range of offenses where an offender is found guilty at trial for an offense within the midrange of objective seriousness for an offense of that type.¹²⁰ By way of example, the penalty for murder is twenty years; the penalty increases to twenty-five years where the victim was a police officer, emergency services worker, correctional officer, judicial officer, or other designated official.¹²¹ The penalty for sexual assault is seven years, while the penalty for unauthorized possession or use of firearms is three years. The standard non-parole provisions commenced on February 1, 2003.¹²²

The standard non-parole periods apply where the offender is found guilty after a trial.¹²³ They are useful reference points where the offender pleads guilty, even when it is found that the objective seriousness of the offense is below the middle range. Pursuant to section 44(2) of the Crimes (Sentencing Procedure) Act 1999 (NSW), the parole period must not exceed one-third of the non-parole period for the sentence, unless special circumstances exist. Thus, in normal circumstances, the non-parole period cannot be less than 75 percent of the head sentence.¹²⁴

Similar provisions to those in NSW operate in South Australia. Section 32A of the Criminal Law (Sentencing Act) 1988 (SA) operates in a similar manner. This provision applies in relation to a range of serious offenses, mainly involving sex or violence.¹²⁵

The main problem with Australian sentencing is its unpredictability and inconsistency. The main reason for this is the instinctive synthesis approach to sentencing. The unfettered discretionary nature of Australian sentencing calculus is similar to the largely uncontrolled sentencing process in the parts of the United States approximately fifty years ago, which led Judge Marvin Frankel to describe the U.S. system as lawless.¹²⁶

The inconsistency in Australian sentencing is demonstrated in a number of ways. The Australian Law Reform Commission Report 103, Same Crime, Same Time: The Sentencing of Federal Offenders ,¹²⁷ looked at sentences across Australia involving the same offenses where the courts were all applying the federal sentencing regime¹²⁸ and noted considerable differences in penalties across jurisdictions. The offenses that the Commission focused on were drug offenses and fraud offenses.¹²⁹

For example, the report looked at sixty-three instances of trafficking a commercial quantity of MDMA (3,4-methylenedioxy-Nmethamphetamine, or Ecstasy) during the five-year period of 2000–2004.¹³⁰ The jurisdictions where most cases occurred were NSW, Western Australia (WA) and Victoria. Overall, the mean terms (maximum and minimum) combined for all three states were 136 and 66 months, while for each individual state they were as follows: NSW (154 and seventy-two months); WA (132 and sixty-nine months); Victoria (sixty-six and thirty-nine months).¹³¹

For a commercial quantity of heroin there were 155 cases, of which 86 percent involved this charge only.¹³² The mean term for these three states combined was eighty-seven and forty-eight months, but once again there were considerable regional differences for each of these three states, i.e. in NSW (eighty-one and forty-eight months); WA (169 and seventy months); Victoria (sixty-five and forty-three months).¹³³

The level of inconsistency is also demonstrated by numerous research reports which compare similarly placed offenders who are subjected to vastly different penalties. The most recent example of this is a 2013 report by the Victorian Advisory Council, Reoffending Following Sentencing in the Magistrates’ Court of Victoria .¹³⁴ One of the purposes of this report was to ascertain if offenders who were sentenced to imprisonment reoffended at different rates to those sentenced to other sanctions.¹³⁵ As noted later in this commentary, the empirical data clearly shows that harsh punishment does not discourage offenders from further offending and in fact may even result in a higher incidence of future offending. The theory of specific deterrence is false. The report supported this finding—offenders sentenced to imprisonment generally reoffended at a higher rate than those that were subjected to more lenient dispositions.¹³⁶ This information is not new.¹³⁷

The most illuminating aspect of the report is how that conclusion was derived. The methodology involved comparing the recidivism rates of offenders who had been sentenced to imprisonment with those who were subjected to more lenient dispositions, including wholly suspended sentences.¹³⁸ This involved controlling the respective samples for factors that could influence the result (e.g., prior criminal record, age, offense type, and sex). Thus, the methodology involved using matched subsamples.¹³⁹

It is striking that identically situated offenders could be subjected to such vastly different outcomes. This is one of the most unsatisfactory aspects of a largely unfettered judicial sentencing discretion.

Thus, sentencing in Australia has a number of objectives; however, there is no attempt to prioritize them. The sentencing calculus is largely uncontrolled, but it is at least superficially underpinned by the need to impose proportionate sentences. Sentences of more than ten years imprisonment are relatively rare,¹⁴⁰ and, on balance, they are much shorter than those in the United States, where the portion of prisoners serving a term of twenty years or more¹⁴¹ exceeds the portion of Australian prisoners serving sentences half that length. This is also reflected in the fact that the imprisonment rate in Australia is less than one-fifth of that in the United States.¹⁴²

III. Can Sentencing Deter, Rehabilitate,or Protect

the Community ?

In developing a sentencing framework, it is necessary to identify the outcomes that can be achieved through a system of state-imposed punishment—it is obviously futile to pursue the unattainable, especially given the important interests at stake in this area. In assessing the objectives and capabilities of sentencing, as noted in Part I of this Article, the most important consideration relates to crime prevention. This consideration has in fact been instrumental in driving sentencing policy. However, something was forgotten in the process—namely, an examination of whether sentencing can in fact reduce crime and, if so, the manner in which this can be best achieved. This Part of the Article seeks to bridge this information gap.

General deterrence and incapacitation have been widely used to justify harsh criminal penalties. To a lesser degree, specific deterrence has been used in a similar manner. The main objective that has been invoked to support softer penalties is rehabilitation. I examine which of these objectives are achievable.

There is a vast body of literature in relation to each of these sentencing objectives. These topics have been the subject of extensive recent analysis.¹⁴³ The discussion below summarizes the main studies in relation to each relevant sentencing objective and the current state of knowledge. This is made easier by the fact that there is a relatively clear consensus in relation to each of the areas, with the exception of rehabilitation and, to a lesser degree, incapacitation. I start with the objectives in relation to which the evidence is most clear-cut.

A. Specific Deterrence Does Not Work

Specific deterrence aims to discourage crime by punishing individual offenders for their transgressions and, thereby, convincing them that crime does not pay. It attempts to dissuade offenders from reoffending by inflicting an unpleasant experience on them (normally imprisonment), which they will seek to avoid in the future.¹⁴⁴ The available empirical data suggests that specific deterrence does not work, so inflicting harsh sanctions on individuals does not make them less likely to reoffend in the future.¹⁴⁵ The level of certainty of this conclusion is very high; in fact, it is so high that specific deterrence should be abolished as a sentencing consideration.

There have been numerous studies across a wide range of jurisdictions and different time periods that come to this conclusion.¹⁴⁶ Daniel Nagin, Francis T. Cullen, and Cheryl L. Jonson provide the most recent extensive literature review regarding specific deterrence.¹⁴⁷ They reviewed the impact of custodial sanctions versus non-custodial sanctions and the effect of sentence length on reoffending. The review examined six experimental studies where custodial versus non-custodial sentences were randomly assigned;¹⁴⁸ eleven studies which involved matched pairs;¹⁴⁹ thirtyone studies which were regression based;¹⁵⁰ and seven other studies which did not neatly fit into any of those three categories and included naturally occurring social experiments which allowed inferences to be drawn regarding the capacity of imprisonment to deter offenders.

This last category of studies also included one based on clemency granted to over 20,000 prisoners in Italy in 2006. A condition of release was that, if those who were released reoffended within five years, they would be required to serve the remaining (residual) sentence plus the sentence for the new offense. It was noted that there was a 1.24 per cent reduction in reoffending for each month of the residual sentence. This observation can be explained on the basis that the threat of future imprisonment discouraged reoffending. However, it was also noted that offenders who had served longer sentences prior to being released had higher rates of reoffending, supporting the view that longer prison terms reduce the capacity for future imprisonment to shape behavior.¹⁵¹

Nagin et al. suggest that offenders who are sentenced to imprisonment do not have a lower rate of recidivism than those who receive a noncustodial penalty and, in fact, that some studies show that the rate of recidivism among offenders sentenced to imprisonment to be higher. They conclude that:

Taken as a whole, it is our judgment that the experimental studies point more toward a criminogenic [that is, the possible corrupting effects of punishment] rather than preventive effect of custodial sanctions. The evidence for this conclusion, however, is weak because it is based on only a small number of studies, and many of them point estimates are not statistically significant.¹⁵²

The review suggests that not only do longer jail terms not deter, but neither do tougher jail conditions. Studies also show that offenders who are sentenced to maximum security prisons, as opposed to minimum security conditions, do not reoffend less.¹⁵³

These general findings are supported by a more recent experimental study by Donald Green and Daniel Winik.¹⁵⁴ They observed the reoffending of 1,003 offenders who were initially sentenced for drug-related offenses between June 2002 and May 2003 by a number of different judges whose sentencing approaches varied significantly (some were described as “punitive” and others as “lenient”), resulting in differing terms of imprisonment and probation. The study concluded that neither the length of imprisonment nor probation had an effect on the rate of reoffending during the four-year follow-up period.¹⁵⁵

As noted above, a more recent study by the Victorian Sentencing Advisory Council, which compared similarly placed offenders who had been sentenced to imprisonment with those who had not been imprisoned, also found that those who were not sentenced to imprisonment had a lower recidivism rate.¹⁵⁶

Accordingly, the weight of evidence supports the view that subjecting offenders to harsh punishment is unlikely to increase the prospect that they will become law-abiding citizens in the future. It follows that the goal of specific deterrence cannot be achieved by the imposition of criminal sanctions and should not influence sentencing practice.

B. General Deterrence (Also) Does Not Work

The main form of deterrence that is used to justify harsh penalties is general deterrence. General deterrence seeks to dissuade potential offenders with the threat of anticipated punishment from committing similar offenses by illustrating the harsh consequences of offending.

There are two forms of general deterrence. Marginal general deterrence concerns the correlation between the severity of the sanction and the prevalence of an offense. Absolute general deterrence concerns the threshold question of whether there is any connection between criminal sanctions, of whatever nature, and the incidence of criminal conduct.¹⁵⁷ The evidence suggests that marginal deterrence does not work, while absolute general deterrence does work.¹⁵⁸ The findings regarding general deterrence are relatively settled.

The failure of even the death penalty to act as a marginal deterrence is exemplified by the experience in New Zealand. Between 1924 and 1962, there were periods of time when the death penalty (for murder) was in force, then abolished, then revived, and abolished again. The changes generally followed some level of public debate and were well publicized. Although the murder rates fluctuated during this period, they bore no correlation to the prevailing penalty, whether capital punishment or life imprisonment.¹⁵⁹

Similar findings have emerged in the United States.¹⁶⁰ Some commentators have attempted to establish a link between lower homicide rates and the death penalty in the United States.¹⁶¹ However, the evidence used in support of a connection between lower homicide rates and capital punishment has been debunked on the basis that the data upon which it is based is statistically insignificant and the evidence goes against the overwhelming trend of the data. As has been pointed out by Richard Berk, the main findings in support of the hypothesis that capital punishment is a deterrent are based on eleven instances out of a sample size of one thousand observations where the homicide rate dropped in a U.S. state following an execution in the previous year.¹⁶² The data is statistically meaningless and contrary to the trend of 99 percent of the observations. Beck states:

Whatever one makes of those 11 observations, it would be bad statistics and bad social policy to generalize from the 11 observations to the remaining 989. So, for the vast majority of states for the vast majority of years, there is no evidence for deterrence in these analyses. Even for the remaining 11 observations, credible evidence for deterrence is lacking.¹⁶³

Commenting on what clearly emerges from the literature, Beck concludes, “[I]t is apparent that for the vast majority of states in the vast majority of years, there is no evidence of a negative relationship between executions and homicides.”¹⁶⁴

The strongest evidence in support of the theory of marginal general deterrence stems from the considerable drop in serious crime levels in the United States over the past thirty years. As noted in the discussion below, the drop coincided with a significant increase in the imprisonment rate. The rate of violent crime in the United States dropped by more than 60 percent from 1993 to 2010.¹⁶⁵

These figures, at face value, suggest that imprisoning ever increasing numbers of offenders effectively reduces the crime rate. A number of detailed studies have been undertaken to examine and explain this causal connection. One analyst, William Spelman, has stated that up to 21 percent of crime reduction is attributable to the increased rate of imprisonment.¹⁶⁶ However, it is not clear whether this reduction is attributable to the incapacitation of offenders (who are thereby prevented from committing crimes whilst they are imprisoned) or to the effects of marginal deterrence.¹⁶⁷ Removing more than one million offenders from the community obviously makes it impossible for them to participate in crime and hence add to the crime statistics during their period of incarceration.¹⁶⁸

Further, it has been noted that similar crime reduction trends occurred in the United States’ nearest neighbor, Canada, over approximately the same period. During that period, the imprisonment rate in Canada actually fell.¹⁶⁹

Empirical evidence not only questions the causal link between higher penalties and lower crime, but also provides strong evidence of alternative explanations for falling crime rates. For example, economist Steven Levitt argues that up to 50 percent of the fall in the United States crime rate can be attributed to the legalization of abortion in the 1970s.¹⁷⁰ After that point, an increased number of women from disadvantaged groups (teenagers, the poor, and minority groups) were able to abort unwanted pregnancies, and Levitt argues that the children from those unwanted pregnancies would have been more likely to commit crimes as adults.¹⁷¹ This ostensibly incredible finding is supported by the fact that states with higher abortion rates in the 1970s had higher drops in crime in the 1990s: each 10 percent rise in abortions corresponds to a one percent drop in crimes two decades later.¹⁷²

Recent research from Germany is consistent with U.S. findings regarding the failure of marginal general deterrence.¹⁷³ At the Goethe University Frankfurt, Horst Entorf reviewed twenty-four years of criminal sentencing practices in West German states for correlates to the crime rate. Entorf sought to examine the effect of each stage of the prosecution process, from investigation to conviction, on the commission rates of two specific crimes (“major property” and “violent crimes”), in order to assess their relative contribution to the overall effect of the criminal prosecution process on crime rates.¹⁷⁴ The results were analyzed by the theoretical econometric analysis methodology,¹⁷⁵ which considered the deterrent effects of formal and informal, as well as custodial and non-custodial, sanctions.

It was discovered that a deterrent effect was found at “the first two stages of the criminal prosecution process” (charge and conviction) rather than at the “less robust” severity of punishment stage (sentencing).¹⁷⁶ Entorf also found that:

Results presented in [the] article suggest that crime is particularly deterred by certainty of conviction. Here, contrary to popular belief, neither police nor judges, but public prosecutors, play the leading role. Extending severity of sentences, however, does not seem to provide a suitable strategy for fighting crime. In particular, the length of the imprisonment terms proves insignificant. ¹⁷⁷

By contrast, the evidence relating to absolute general deterrence is more positive. The strongest empirical evidence in support of absolute deterrence comes from the United States, which (as noted above) over the past three decades has seen a marked increase in police numbers¹⁷⁸ and a sharp decrease in crime. The near universal trend of the data that outlines this link supports the view that more police, and hence the greater actual and perceived likelihood of detection, has contributed to the reduction in crime.¹⁷⁹

The connection is complex due to the multi-faceted nature of the changes that occurred during this period, which may also have had an effect on the crime rate. These changes include such things as better police methods, a generally improving economy, and other variables including abortion trends and the greater use of imprisonment. It has been noted that the greatest reduction in crime numbers occurs where police are highly visible.

This accords with the ostensible success of “zero tolerance”¹⁸⁰ policing in locations such as New York City, which saw the greatest number of extra police employed and the sharpest decline in crime.¹⁸¹ This trend was evident well over a decade ago. In a period of only several years following the introduction of zero tolerance policing, the rates of violent and property crime fell by approximately 35 percent.¹⁸²

After evaluating the large number of surveys analyzing the connection between more police and the crime rate, Raymond Paternoster concludes:

What we are left with, then, is that clearly police presence deters crime, but it is probably very difficult to say with any degree of precision how much it deters. Let us take Levitt’s estimate as a reasonable guess, that increasing the size of the police force by 10% will reduce crime by about 4% or 5%.¹⁸³

The link between lower crime rates and potential offenders’ higher perceptions of being caught supports the theory of absolute deterrence because such offenders know that if they are caught, some hardship awaits them. If rather than punishing offenders, police handed out lollies or movie tickets, the utilization of more police would result in more crime.

Thus, general deterrence does work, at least to the extent that if there were no real threat of punishment for engaging in unlawful conduct, the crime rate would soar. It follows that the threat of punishment discourages potential offenders from committing crime. This justifies the punishment of wrongdoers. The evidence does not support the view, however, that this relationship operates in a linear fashion; that is, the deterrent effect of sanctions does not increase in direct proportion to the severity of sanctions.

Accordingly, while the objective of deterrence justifies imposing punishment, it is at best a remote consideration when it comes down to the question of how much punishment should be imposed. Absolute general deterrence provides a justification for imposing punishment but it does not justify the imposition of penalties that exceed the objective gravity of the offense. It follows that the pursuit of general deterrence cannot justify the imposition of harsh penalties for offenders.

C. Incapacitation

1. Selective Incapacitation

Incapacitation is often used as a basis for justifying longer sentences. Incapacitation aims to protect the community by confining offenders to imprisonment during which time they can no longer commit offenses. The effectiveness of incapacitation cannot be judged by the height of the prison wall. Imprisonment as a means of community protection is only effective if, but for being imprisoned, the offender would have committed a further offense.¹⁸⁴

There are two forms of incapacitation. The first is selective incapacitation. This focuses on the individual offender, and its success is contingent upon distinguishing between offenders who will reoffend from those who will not. Most of the research in this area has been directed towards predicting which serious offenders will reoffend.¹⁸⁵ In this regard, the focus has been offenders who commit violent and sexual offenses.

A wide-ranging analysis in the 1990s of the data regarding the capacity of any discipline to predict future criminal behavior noted that predictive techniques “tend to invite overestimation of the amount of incapacitation to be expected from marginal increments of imprisonment.”¹⁸⁶

More recent actuarial tools have been developed to score a person’s level of risk by mapping their profile to variables that are known risk factors. Structured professional judgment and criminogenic needs tools also use a range of variables¹⁸⁷ which are designed to be more nuanced than actuarial tools because they aim to not only predict the likelihood of violence, but also the imminence, severity, and possible targets of the risk.¹⁸⁸ These more recent attempts to accurately predict dangerousness in the context of violent and sexual offenses have also proven to be deficient.¹⁸⁹

All predictive tools use prior criminal history as a key variable. This has generally proven to be an unreliable indicator.¹⁹⁰ A New South Wales Study focused on offenders who committed serious violent offenses¹⁹¹ in 1994 and were released from custody no later than 2009. There were 435 such offenders and the tracking showed that by September 2011, seventythree offenders had committed another offense involving a serious degree of violence—meaning that 83 percent did not commit such an offense.¹⁹² The offenders that were least likely to reoffend were those convicted of drug offenses.¹⁹³

2. General Incapacitation

While selective incapacitation does not work, general incapacitation is more effective in reducing crime.¹⁹⁴ General incapacitation involves imprisoning offenders simply because they have committed a criminal offense on the basis that while in prison they cannot inflict harm in the general community. Little or no effort is normally made to predict future offending patterns, whether on the basis of previous criminal history or other considerations.¹⁹⁵ There is no clear line between selective and general incapacitation and the difference is often simply one of degree. Once large numbers of offenders are imprisoned on the basis of predictive criteria, which is demonstrably inaccurate, then a process that may have initially had the appearance of selective incapacitation turns into a system of general incapacitation. All jurisdictions punish recidivists more severely than first time offenders.¹⁹⁶ Often the extent of the premium is so significant that this has effectively evolved into a process of general incapacitation.¹⁹⁷

Theoretically, general incapacitation should work because the more people that are in prison, the less people there will necessarily be available who could commit crime in the general community. Accordingly, it should follow that this will reduce the crime rate in absolute terms. However, it should also reduce crime in a relative sense. This is because people who commit crime (even those who are non-recidivists) are disproportionately from one sector of the community: the lower socio-economic group. Poor people are grossly over-represented in prisons across the world.¹⁹⁸ Thus, imprisoning large numbers of poor people should reduce crime by reducing not only the number of crime offenses, but also the number of crimes per non-prison population.

Most of the research into the testing of the general incapacitation model has been undertaken in the United States, presumably because of the unprecedented increase in the prison population over the past thirty years. The weight of evidence supports the view that general incapacitation works. In the United States between 1993 and 2010:

(a) the rate of violent victimization rates decreased by 76%; (b) the decline in total household property crime victimization

was 64%.¹⁹⁹

During this period the imprisonment rate rose from 1.365 million to 2.27 million prisoners.²⁰⁰ At face value, these figures suggest a causal link between imprisoning greater numbers of offenders and an effective reduction in the crime rate.

As noted above, William Spelman has calculated that up to 21 percent of crime reduction is attributable to the increased rate of imprisonment.²⁰¹ Other studies support the success of incapacitation, but remain equally unclear about its precise impact. According to literature examined by Roger Warren, a 10 percent increase in imprisonment rates produces a 2 to 4 percent reduction in the crime rate; however, most of this relates only to non-violent offenses.²⁰²

While general incapacitation seems to have some validity, one constant finding is that it is usually most effective in relation to minor crime, although some success can also be achieved in relation to more serious forms of offending.²⁰³ This is because minor offenders reoffend more frequently than serious offenders.

The effectiveness of general incapacitation for relatively minor offenses is supported by an Australian study published in 2006 by Don Weatherburn et al., How Much Crime Does Prison Stop? The Incapacitation Effect of Prison on Burglary .²⁰⁴ The study measured the impact of imprisonment on burglary rates. It concluded:

[A]t least so far as burglary is concerned; prison does seem to be an effective crime control tool. Our best estimate of the incapacitation effect of prison on burglary (based on the assumption that burglars commit an average of 38 burglaries per year when free) is 26 per cent. This estimate does not appear to be overly sensitive to the value of offending frequency we assume . . . . These percentage effects might not seem large but, in absolute terms, an incapacitation effect of 26 percent is equivalent to preventing over 44,700 burglaries per annum.²⁰⁵

However, the report then noted that the cost associated with using imprisonment as a tool to reduce the burglary rate was too high:

The fact that prison is effective in preventing a large number of burglaries raises the question of whether increased use of imprisonment would be an effective way of further reducing the burglary rate. Our findings on this issue, like those of incapacitation studies in Britain and the United States (Cohen 1978; Tarling 1993), are not that encouraging. They suggest that a doubling of the sentence length for burglary would cost an additional $26 million per annum but would only reduce the annual number of burglaries by about eight percentage points. A doubling of the proportion of convicted burglars would produce a larger effect (about 12 percentage points) but only if those who are the subject of our new penal policy offend as frequently as those who are currently being imprisoned. Given what we know about the frequency of offending amongst burglars who do not currently receive a prison sentence, this seems highly unlikely.²⁰⁶

Similar findings are reported regarding sentence enhancements imposed on offenders in the Netherlands.²⁰⁷ A law passed in 2001 required increased sentence severity for offenders who had ten or more prior convictions.²⁰⁸ A key distinction between the enhancements imposed by these laws and those in some other jurisdictions was that they were relatively minor—no more than an additional three years’ imprisonment.²⁰⁹ By 2007, fourteen hundred offenders were sentenced under this regime, most of whom were non-violent offenders.²¹⁰ The result was a dramatic drop in the rate of burglary and car theft in the ten cities in which the law operated. The report concluded:

We find that in a situation of a relatively low rate of incarceration, sentence enhancements for a carefully selected group of prolific offenders can dramatically reduce the crime rate . . . . Although the group of offenders sentenced under the law accounted for only 5 percent of the prison population six years after its introduction, the sentencing policy lowered the rate of burglary and theft from cars by an estimated 40 percent through the incapacitation effect alone. When comparing the cost of enhanced prison sentences with the social benefits of lower crime rates, we find the benefits of the policy to exceed the costs. Even for this highly selective sentencing policy that only affected 1,400 offenders in the period 2001-2007, we find evidence for rapidly decreasing returns to scale. The marginal crime-reducing effect of incapacitating another prolific offender declines by more than half from the lowest to the highest rate of application of the law. The benefitcost ratio drops sharply when more offenders are serving time under the habitual offender law.²¹¹ More wide-ranging data also support the link between prior and future offending and the fact that the link is strongest in relation to minor offending. The most wide-ranging study of the trajectory of offenders in Australia was undertaken by the Australian Bureau of Statistics and released in August 2010, in a report entitled, “An Analysis of Repeat Imprisonment Trends in Australia.”²¹² The report is based on a fourteen-year longitudinal study for the period July 1, 1994 to June 30, 2007. The study grouped prisoners into two cohorts. The first consisted of those released between July 1, 1994 and June 30, 1997 and consisted of 28,584 people. The second comprised prisoners released between July 1, 2001 and June 30, 2004 and consisted of 26,700 people. The study compared recidivism rates from both cohorts within three years from release. It also examined the ten-year re-imprisonment rate for the earlier cohort.²¹³

The report noted that the number of prisoners with prior imprisonment grew at a rate of 3.2 percent each year,²¹⁴ although this rate was not steady, with the rate ranging from 56 percent to 62 percent.²¹⁵ The data on the portion of released prisoners who return to imprisonment within the respective three-year periods are even more illuminating. The report noted that for the 1994 to 1997 cohort, about 20 percent were re-imprisoned within two years; one-quarter were re-imprisoned within three years, and 40 percent were imprisoned by the end of the ten-year survey period.²¹⁶ A surprising finding (given that rehabilitative measures had presumably improved between 2004 and 2007)²¹⁷ was that prisoners released in the later cohort were more likely to be re-imprisoned than the earlier cohort over an equivalent three-year follow up period.²¹⁸ The re-imprisonment rate for the latter cohort was 17 percent higher than for the earlier one, i.e. the rate for 1994 to 1997 was 25.1 percent compared to 29.5 percent for the 2004 to 2007 cohort.²¹⁹

Thus most of the prison population examined was made up of people who had been in prison before. Moreover, the data showed that prisoners with prior imprisonment were twice as likely as first-timers to return to prison (50 percent compared to 25 percent imprisonment rates, respectively, from ten years after release).²²⁰ When a logistic regression was applied to this data, it emerged that the odds ratio that a prisoner with a number of previous prison terms would be imprisoned were 2.9 times that of a first-time prisoner.²²¹ Thus it does appear that first-time prisoners are less likely to be re-imprisoned than repeat prisoners.

In terms of re-imprisonment trends by offense type for the 1994 to 1997 release cohort, it was noted that by June 30, 2007, the offenders who were most commonly re-imprisoned were those sentenced for burglary (58 percent), theft (53 percent) and robbery (45 percent), while those who were least re-imprisoned were those convicted of drug offenses (24 percent) and sexual assault (21 percent).²²²

The recidivism levels ascertained by this report are high, but in reality they are considerably higher. The report did not focus on released offenders who committed crimes for which they were not imprisoned. Thus, there is a clear link between previous offending and an enhanced risk of future offending. This link justifies extra steps being taken to prevent offenders from committing further offenses. As we have seen, taking large numbers of people with prior convictions out of the community will reduce the crime rate.

Therefore, the complex question is: what response is appropriate and can be adapted to the above findings? The matter is complicated by the fact that prior offending is a stronger indicator of future offending in relation to minor offense than it is for serious offenses. In relation to relatively minor offenses, incapacitation works. However, while confining repeat minor criminals clearly disables them from committing further offenses in the community for a period of time, it almost certainly does not justify the unrestrained use of imprisonment to combat non-serious crime.²²³ These offenders should be subjected to an incapacitative penalty, but governments need to develop more intelligent alternatives to imprisonment that can monitor the activities of recidivist minor offenders at a fraction of the cost of imprisonment.²²⁴

Given the limits of predicting serious offending on the basis of prior convictions, selective incapacitation for serious offenses, on the other hand, seems to be flawed. However, there is stronger evidence that general incapacitation does work in relation to such offenses. While most serious offenders do not reoffend, individuals with previous convictions for serious offenses commit such crime at a much greater frequency than the rest of the criminal population. Further, as noted above, offenders with prior convictions for serious offenses reoffend more frequently than first-time offenders.

This leaves policy-makers with a difficult choice. Ultimately, the issue comes down to ascertaining the level and nature of risk and the appropriate burden placement—i.e. whether it should fall on the offender or prospective victims in the general community.

There is a degree of unfairness associated with imprisoning offenders for longer than is commensurate with the severity of their instant offense. However, it would be remiss of legislatures not to take all reasonable steps to prevent innocent people from being victimized. The fact that potential victims cannot be identified in advance does not negate the need to put in place mechanisms to limit serious encroachments on the human rights of citizens. Thus, the debate about incapacitation as an appropriate sentencing goal is not about balancing the utilitarian benefit of community safety against the right to liberty of offenders. It is about weighing competing rights: the liberty of offenders against the right to sexual and physical integrity of prospective victims.

In relation to risk allocation decisions, the weight of the burden should be disproportionately shouldered by the morally and legally culpable (offenders), as opposed to the innocent potential victims who have not played any role in creating the dilemma. The rights of the innocent trump those of the guilty, assuming the rights are of approximately equal importance.

Thus, the deprivation of liberty occasioned by longer sentences for recidivists is justified as a means of increasing the protection of the bodily and sexual integrity of other individuals. However, this does not justify greatly enhanced penalties. The balance that is appropriate must be proportionate to the objective that is sought and not gratuitously over-reach in order to satisfy the instinct to punish repeat offenders.

There is insufficient empirical data to enable accurate and forensic choices to be made about how much extra prison time should be imposed on recidivists. However, at some point there is a diminishing marginal return in terms of offenses prevented for each year of prison time. In addition, in any decision-making calculus, certain consequences (in the form of additional prison time) need to carry more weight than speculative outcomes (in the form of whether or not a particular offender would have actually re-offended). Therefore, the recidivist loading for serious offenses should be relatively minor, say 20 to 50 percent,²²⁵ and certainly nowhere near the oppressive levels that are manifest in some sentencing grids and three-strikes regimes.

It is important to note that the relevance of a prior record dissipates over time, such that if an offender remains crime-free for approximately seven years, his or her likelihood of reoffending is approximately the same as for a person without a criminal history.²²⁶ Thus, prior convictions should cease to be taken into account after this period.

D. Rehabilitation

While the goal of incapacitation can justify the infliction of more severe penalties for some offenders, should this be offset by the desire to rehabilitate offenders? The effectiveness of rehabilitation in lowering the rate of repeat offending has been the subject of a large number of studies. Though the evidence is not conclusive, as discussed below, on balance it seems that specific forms of intervention may be able to reduce recidivism. Unlike the other key sentencing goals that have been analyzed above, rehabilitation (if it is effective) serves normally to decrease rather than increase penalty severity.

Studies indicating that rehabilitation does not work provide the most damaging objection to rehabilitation as a suitable goal of sentencing. Following extensive research conducted between 1960 and 1974, Robert Martinson, in an influential paper, concluded that empirical studies had not established that any rehabilitative programs had worked in reducing recidivism.²²⁷ The Panel of the National Research Council in the United States, several years after this work, also noted that there were no significant differences between the subsequent recidivism rates of offenders regardless of the form of punishment. They concluded, “This suggests that neither rehabilitative nor criminogenic effects operate very strongly.”²²⁸

In recent years, the research has taken on a more optimistic note. Most Australian jurisdictions have devoted increasing resources to rehabilitation over the past decade. The most recent wide-ranging Australian study regarding the effectiveness of rehabilitation is a report by Karen Heseltine, Andrew Day, and Rick Sarre for the Australian Institute of Criminology, published in 2011.²²⁹ The report focused on changes and improvements to prison based correction rehabilitation programs in the custodial environment since 2004, when the previous report was issued.²³⁰

The report by Heseltine et al., while unable to evaluate the effectiveness of rehabilitation programs currently operating in Australian prisons, summarized recent cross-jurisdictional studies into the effectiveness of certain rehabilitation programs. It noted that while there were mixed results, there were some programs that reported positive outcomes.²³¹

This was especially the case in relation to sexual offender programs, where some studies showed that the recidivism rate of offenders completing the program was about half of that of other offenders.²³² The results of programs directed towards violent offenders were less positive, but a wideranging review of studies focusing on United Kingdom programs noted that reductions in offending for violent offenses by around 7 to 8 percent had occurred.²³³ Overseas studies reported some success with anger management programs, but an Australian study (of a shorter twenty hour program) showed limited positive outcomes related to program completion.²³⁴ There is no cogent evidence supporting the effectiveness of domestic violence or victim awareness programs.²³⁵ However, drug and alcohol programs have been shown to be effective at reducing substance abuse and reoffending.²³⁶

This assessment is consistent with the findings of Ojmarrh Mitchell, David B. Wilson and Doris L. MacKenzie, who undertook a major analysis of studies into the effectiveness of drug treatment programs in prison.²³⁷ The studies they focused on related to drug users and compared reoffending patterns of offenders who completed a drug rehabilitation program with those who did not complete a program, or completed only a minimum program between the years 1980 and 2004. They analyzed sixty-six studies in total. The report concluded, “Overall, this meta-analytic synthesis of evaluations of incarceration based drug treatment programs found that such programs are modestly effective in reducing recidivism.”²³⁸ Moreover, it noted that programs that dealt with the multiple problems of drug users (termed therapeutic communities) were the most successful, whereas there was no evidence to support good outcomes associated with “boot camp” programs.”²³⁹

A striking aspect of rehabilitation is how little is known about its effectiveness, especially in the Australian context where considerable (and increasing) resources are directed in the hope of rehabilitating offenders. As noted above, the weight of empirical data (though it is far from uniform or consistent) suggests that rehabilitative programs can reduce the likelihood of recidivism, especially for certain forms of offenses, such as sex-offences.

Moreover, it seems that successful rehabilitation programs can be administered within a custodial setting.²⁴⁰ If this is proven to be correct, it entails that, paradoxically, the goal of rehabilitation should lose its mitigatory impact, so far as being a basis for not imprisoning offenders or reducing the length of prison terms is concerned. However, the actual level of knowledge regarding the impact of rehabilitative programs on recidivism rates is so small that no policy or legal changes should be made at this point as far as rehabilitation is concerned.

E. Overview

The sentencing system should pursue only those objectives that empirical evidence (and in particular econometric research) shows are attainable through a system of punishing wrongdoers.

As we have seen, current empirical evidence provides no basis for confidence that punishment is capable of achieving the goal of specific deterrence. However, experience shows that absent the threat of punishment for criminal conduct, the social fabric of society would readily dissipate. Crime would escalate and overwhelmingly frustrate the capacity of people to lead happy and fulfilled lives. Thus, general deterrence works in the absolute sense—there is a connection between criminal sanctions and criminal conduct. However, there is insufficient evidence to support a direct correlation between higher penalties and a reduction in the crime rate. It follows that marginal deterrence (which is the theory that there is a direct correlation between the severity of the sanction and the prevalence of an offense) should be disregarded as a sentencing objective—at least, unless and until, there is proof that it works.

Incapacitation is generally flawed, since we are poor at predicting which offenders are likely to commit serious offenses in the future. However, there is a connection between prior offending and the commission of serious offenses. This justifies a sentencing enhancement being imposed on serious offenders. The emerging evidence in relation to rehabilitation is optimistic. While there are no far-reaching rehabilitative techniques which have proven to be successful at producing positive internal attitudinal change in all offenders, some techniques seem to work with certain classes of offenders. These programs seem to be effective even in an imprisonment setting. Thus, while sentencing should seek to rehabilitate offenders, there is no basis for assuming that this should greatly impact the type or length of sanction.

IV. The Golden Thread: Proportionality—and

Unpacking It

As noted in Part I, the key aim of both main theories of punishment is to ensure that the harshness of the penalty matches the severity of the crime. In this Part, I establish the reason for this proposition. I then provide meaningful content to the concept of proportionality.

A. Retributivism and Proportionality

Given the vast array of retributive theories, it is not possible to consider every retributive argument in favor of proportionality. I therefore only focus on two theories: the one that most expressly encapsulates proportionalism (i.e. lex talionis ) and the theory often regarded as the most influential contemporary retributive theory (i.e. censure).

1. Lex Talionis and proportionality

The retributive account that most clearly endorses the proportionality thesis is the lex talionis or the “eye for an eye, a tooth for a tooth” approach to punishment. This theory, however, provides little guidance regarding the proper workings of proportionalism. The lex talionis theory has no clear application in relation to most offenses: “what penalty would you inflict on a rapist, a blackmailer, a forger, a dope peddler, a multiple murderer, a smuggler, or a toothless fiend who has knocked somebody else’s teeth out?”²⁴¹

It has been suggested that a more plausible interpretation of the lex talionis theory is that the punishment and the crime should be equal or equivalent.²⁴² While this expands its potential scope of application, lex talionis is normally advanced as a stand-alone theory, devoid of a further rationale, and hence it is incapable of providing insight regarding the content of the proportionality limbs. What “equivalent” pain can be inflicted on an impecunious and homeless burglar?

Nevertheless, as discussed below, the lex talionis theory does provide some direction regarding the development of proportionalism. In order for the theory to gain traction and content, it is desirable to commence with an evaluation of offense severity in relation to offenses where there is an identifiable victim who suffers tangible harm and then move to other offense categories.

2. Censure and Proportionality

Perhaps the leading retributive theory is that advanced by Andrew von Hirsch. He contends that the principal justification of punishment is censure: that is, to convey blame or reprobation to those who have committed a wrongful act.²⁴³ Von Hirsch believes that censuring holds offenders responsible and accountable for their actions and that, by giving them an opportunity to respond to their misdeeds through acknowledging their wrongdoing in some form, it recognizes their moral agency.²⁴⁴

For von Hirsch, punishment has a dual objective. The other justification is to prevent crime. He believes that human nature is such that the normative reason for compliance must be complemented with a prudential one, otherwise “victimising [sic] conduct would become so prevalent as to make life nasty and brutish.”²⁴⁵ Also “it is the threatened penal deprivation that expresses the censure as well as serving as the prudential disincentive.”²⁴⁶ Although he believes that deterrence is not a sufficient reason for punishment, he claims it is a necessary one: “[I]f punishment has no usefulness in preventing crime, there should . . . not be a criminal sanction.”²⁴⁷ Instead there should be other means adopted to express censure.

Andrew von Hirsch believes that the following three steps justify the proportionality principle within his theory of punishment:

1. The State’s sanctions against proscribed conduct should take

a punitive form; that is, visit deprivations in a manner that expresses censure or blame.

2. The severity of a sanction expresses the stringency of the

blame.

3. Hence, punitive sanctions should be arrayed according to

the degree of blameworthiness (i.e., seriousness) of the conduct.²⁴⁸

However, on closer assessment, the argument is less convincing. As noted above, von Hirsch’s theory of punishment focuses heavily on the claim that the aim of punishment is to express censure—that is, to convey condemnation or blame directed at a responsible wrong-doer. However, more fully, for von Hirsch the purpose and function of punishment is two-fold: “(1) to discourage [criminal] conduct, and (2) to express disapproval of [criminal] conduct and its perpetrators.”²⁴⁹

Given that punishment, according to von Hirsch, has two purposes, it seems arbitrary to claim that the amount of punishment that is deserved is determined solely by its censuring goal. This, in turn, potentially undermines the credibility of the second premise, since logically both rationales for punishment must affect the inquiry of how much to punish. This being the case, it may be necessary to impose sanctions that are significantly more severe than is required to match the blameworthiness of criminal conduct.

Von Hirsch responds to this criticism by stating that despite his bifurcated account of punishment, prevention cannot be invoked in deciding how much to punish because proportionality would then be undermined.²⁵⁰ Arguably, this response is deficient because proportionality is not a justification for punishment; rather, it is merely a restraint on it, derived from the rationale for punishment.

A fuller assessment of the legitimacy of von Hirsch’s general theory of punishment and justification for proportionalism is beyond the scope of this paper.²⁵¹ What, however, is illuminating is the prominence with which he regards the proportionality principle. It is such a foundational aspect of his theory that it can trump one of its two justificatory pillars— general deterrence.

More interesting for current purposes is whether the justification that von Hirsch offers for his general theory can be used to inform the content of the proportionality principle. Ultimately, he contends that offense seriousness is gauged by the level of blameworthiness of the conduct.²⁵² Blameworthiness is a nebulous concept that has no fixed meaning and one which is incapable of accurate precision. Accordingly, it cannot shed meaningful light on offense severity. This is a point that seems to have been at least tacitly accepted by von Hirsch, for in developing a hierarchy of offense severity he ultimately relies solely on consequentialist considerations.

I now discuss whether a utilitarian theory of punishment is capable of providing firmer guidance regarding the content of the proportionality limbs.

B. Utilitarianism and Proportionality

Proportionality has traditionally been thought to have no role in a utilitarian theory of punishment. Rather than focusing on retrospective considerations to do with the nature of an offense to determine how much to punish, utilitarians place emphasis on prospective matters, such as the need for deterrence, rehabilitation, and so on. Given this, criticisms of utilitarianism have been made to the effect that it justifies substantial punishment for minor offenses, where this is necessary to reform or deter the offender.²⁵³

However, proportionalism and the utilitarian theory of punishment are not necessarily incompatible. Jeremy Bentham argued in favor of the proportionality principle on the basis that if crimes are to be committed, it is preferable that offenders commit less serious rather than more serious ones.²⁵⁴ In his view, sanctions should be graduated commensurate with the seriousness of the offense so that those disposed to crime will opt for less serious offenses. In the absence of proportionality, potential offenders would not be deterred from committing serious offenses any more than minor ones, and hence would just as readily commit them. This argument, however, has been persuasively criticized by von Hirsch. He points out that there is no evidence that offenders make comparisons regarding the level of punishment for various offenses.²⁵⁵ Further, as noted above, the weight of empirical evidence suggests that higher penalties do not result in less crime.

However, there is yet another basis upon which proportionality may have a role in utilitarian punishment. Some suggest that disproportionate sentences risk placing the criminal justice system into disrepute because such sentences would offend the principle that privileges and hardships ought to be distributed roughly in accordance with the degree of merit or blame attributable to each individual.²⁵⁶ Violations of this principle lead to antipathy towards institutions or practices that condone such outcomes. Harding and Ireland, for example, believe:

Proportion in punishment is a widely found and deeply-rooted principle in many penal contexts. It is . . . integral to many conceptions of justice and as such the principle of proportion in punishment seen generally acts to annul, rather than to exacerbate, social dysfunction.²⁵⁷

Similarly, others feel that one of the main reasons for the success of the Finnish criminal justice system²⁵⁸ is the emphasis placed on the principle of proportionality: “principles of proportionality and perceived procedural fairness are key factors that influence the willingness of the people to conform to the law.”²⁵⁹

Empirical studies appear to support this sentiment. After a 1984 study of approximately 1,500 people who lived in Chicago regarding their contact with legal authorities, scholars noted that normative issues are closely linked with compliance with the law.²⁶⁰ People do not merely obey the law because it is in their self-interest to do so; they also obey the law because they believe it is proper to do so. The judgment that it is appropriate to obey the law, not only is affected by the internal content of the law, but by the attitude of the community towards those who enforce the law. Thus, the perception of a legitimate police force makes it more likely that laws will be observed. The concept of tying criminal sanctions up with the community’s moral disapproval of the act is strongly embodied in the Scandinavian sentencing system, where it is assumed that:

The disapproval expressed in punishment is assumed to influence the values and moral views of individuals. As a result of this process, the norms of criminal law and the values they reflect are internalized; people refrain from illegal behaviour, not because it is followed by unpleasant punishment, but because the behaviour itself is regarded as morally blameworthy. . . . The effective functioning of criminal law is not based on fear, but on legitimacy and acceptance.²⁶¹

Accordingly, there is a utilitarian foundation for proportionalism if the proportionalist ideal is so inherently ingrained in the human psyche that non-observance of the doctrine will disincline individuals from complying with legal norms.²⁶² In light of the above points, there is considerable theoretical merit in this argument. However, there is no empirical proof of this hypothesis. Anecdotal evidence exists in the fact that the Scandinavian nations, which ostensibly place most weight on the proportionality principle,²⁶³ also have among the lowest crime rates in the world. However, it is notoriously difficult to conclusively identify the cause and effect systems in operation regarding wide-ranging societal practices. The main reason for this is the number of other variables that contribute to individual and collective behavior. The reason for the lower crime rate in Scandinavia could be unrelated to proportionalism and could, for example, relate to the wide social welfare net in those countries, which reduces the incidence of poverty, which is known to increase criminal behavior.

Nevertheless, at this point of human learning, enough has been said to suggest that proportionalism has at least more than a tenable foundation in the context of a utilitarian theory of punishment. This is important because, as discussed below, a consequentialist approach to proportionalism is the most promising way forward to providing clear and stable criteria that inform both proportionality limbs.

Thus while proportionalism is on face value an ingrained part of most retributive theories, such theories provide little guidance regarding the content of the proportionality limbs. Greater guidance emerges against a utilitarian backdrop to punishment, although the empirical validity of a proportionalist principle in this context remains to be proven.²⁶⁴

C. Unpacking Proportionality

Despite the near universal endorsement of proportionality, there is no convergence in sentences either within or across jurisdictions—even those that ostensibly place cardinal emphasis on proportionality in sentencing determinations.²⁶⁵ The vagaries are so pronounced that it is verging on doctrinal and intellectual fiction to suggest that an objective answer can be given to common sentencing dilemmas, such as how many years of imprisonment is equivalent to the pain felt by an assault victim, or whether a burglar should be dealt with by way of imprisonment or fine, or the appropriate sanction for a drug trafficker. Certainly, there is no demonstrable violation of proportionality if a mugger, robber, or drug trafficker is sentenced to either twelve months or twelve years imprisonment.

Some commentators have argued that proportionality is so vague as to be meaningless, in light of the fact that there is no stable and clear manner in which the punishment can be matched to the crime.²⁶⁶ The more complex inquiry is how legislatures can go about matching the two limbs of the proportionality principle, but this is a difficult task. How many years of imprisonment correlate to the pain endured by a rape victim? As noted by Andrew von Hirsch and Nils Jareborg, in developing their “living standard approach” to offense seriousness,²⁶⁷ “virtually no legal doctrines have been developed on how the gravity of harms can be compared.”²⁶⁸ Jesper Ryberg notes that one of the key and damaging criticisms of proportionality is that it “presupposes something which is not there, namely, some objective measure of appropriateness between crime and punishment.”²⁶⁹ He further notes that to give content to the theory, it is necessary to rank crimes, rank punishments, and anchor the scales.²⁷⁰

Given that there has been no systematic, doctrinally sound approach to defining the factors that are relevant to proportionality, a logical place to start in elaborating on the meaning of proportionality is to examine the manner in which it has been developed by the courts. However, this inquiry is of limited utility. The courts have not attempted to comprehensively explain either proportionality limb. In relation to offense severity, rather than positively defining the factors that are relevant to offense severity, the courts have focused on dismissing some considerations that are irrelevant. Factors such as “good character, . . . repentance, restitution, possible rehabilitation and intransigence”²⁷¹ have been excluded.²⁷² However, some factors have been positively identified as relevant to offense seriousness. These include: the consequences of the offense, as well as the level of harm; the victim’s vulnerability and the method of the offense;²⁷³ the offender’s culpability, which turns on such factors as the offender’s mental state²⁷⁴ and his or her level of intelligence; the level of sophistication involved;²⁷⁵ the protection of society;²⁷⁶ and even the offender’s previous criminal history.²⁷⁷

The problem with such a list is that, despite its non-exhaustive character, it is too particular, and it is no more than a non-exhaustive list of common aggravating factors. Once considerations such as the method of the offense and the victim’s vulnerability are included, there appears to be no logical basis for not including other considerations that are typically thought to increase the severity of an offense such as breach of trust, the prevalence of the offense, profits derived from the offense, and an offender’s degree of participation. Such an approach is devoid of an overarching justification and is, ultimately, baseless.

Accordingly, in order to inject content into proportionality, it is necessary to revert to a fundamental principle. The key aspect of the principle is that it has two limbs. The first is the seriousness of the crime and the second is the harshness of the sanction. Further, the principle has a quantitative component—the two limbs must be matched. In order for the principle to be satisfied, the seriousness of the crime must be equal to the harshness of the penalty.

The main difficulty here is that the two currencies are different. The interests typically violated by criminal offenses are physical integrity and property rights. At the upper end of criminal sanctions, the currency is (deprivation of) freedom. The only conceivable way to give content to the proportionality principle is to ascertain the extent to which offenders and victims are set back by various offense and penalty types.²⁷⁸ However, the problem is not intractable. It has been suggested that content can be injected into the proportionality principle if both aspects of the equation focus on the extent to which the interests of victims and offenders are set back as a result of the crime and punishment, respectively.

1. First Limb: Evaluating the Seriousness of the Offense

To this end, a number of studies have measured the impact of certain crime offense categories on victims. As noted below, the best information available suggests that, typically, victims of crime suffer considerably and, in fact, suffer more than is manifest from the obvious and direct effects of crime. The problem with some studies is that they do not distinguish adequately between different types of crime to determine the relative impact of criminal offense types. However, the data available suggests that victims of violent crime and sexual crime have their well-being more significantly set back than for other types of crime.

Rochelle Hanson, Genelle Sawyer, Angela Begle, and Grace Hubel reviewed the existing literature regarding the effects of violent and sexual crimes on key quality of life indices.²⁷⁹ The crimes examined included rape, sexual assault, aggravated assault, survivors of homicide (i.e., relatives of those killed) and intimate partner violence. The key quality of life indicia examined were: role function (i.e., capacity to perform in the roles of parenting and intimate relationships and to function in the social and occupational domains); reported levels of life satisfaction; and well-being and social-material conditions (i.e., physical and mental health conditions). The report demonstrated that many victims suffered considerably across a range of well-being indicia, well after the physical signs had passed. The report concluded:

In sum, findings from the well-established literature on general trauma and the emerging research on crime victimization indicate significant functional impact on the quality of life for victims. However, more research is necessary to understand the mechanisms of these relationships and differences among types of crime victimization, gender, and racial/ethnic groups.²⁸⁰

Findings showed that victims of violent crime and sexual crime in particular have:

(1) Difficulty in being involved in intimate relationship and far

higher divorce rates;²⁸¹

(2) Diminished parenting skills (although this finding was not

universal);²⁸²

(3) Lower levels of success in the employment setting (especially in relation to victims who had been abused by their

partners) and much higher levels of unemployment;²⁸³

(4) Considerable impairment and dysfunction in social and leisure activities, with many victims retreating from conventional social supports;²⁸⁴ and (5) High levels of direct medical costs associated with violent

crime (over $24,353 for an assault requiring hospitalization).²⁸⁵

A study published in 2006, focusing on victims in the United States, found both that victims of violent crime were 2.6 times as likely as nonvictims to suffer from depression and 1.8 times as likely to exhibit hostile behavior five years after the original offense,²⁸⁶ and that for 52 percent of women who had been seriously sexually assaulted in their lives, their experience led to either depression or other emotional problems. For one in twenty women, it led to attempted suicide.²⁸⁷ And in another study examining the effects of either violent or property crime on the health of 2,430 respondents,²⁸⁸ Chester L. Britt noted, “Victims of violent crime reported lower levels of perceived health and physical well-being, controlling for measures of injury and for socio-demographic characteristics.”²⁸⁹ Further, these findings were not confined to violent crime: victims of property crime also reported reduced levels of perceived well-being, but it was less profound than in the case of violent crime.²⁹⁰

The ranking of crime is “complicated by the fact that seriousness is held to be a function of both harm and culpability and, at least according to some theorists, also of prior criminal record.”²⁹¹ There is, however, no principled reason for infusing either of these considerations into an assessment of offense severity.

As we saw earlier, this variable-rich approach to offense severity which involves importing aggravating and mitigating considerations into proportionality calculations is consistent with the manner in which courts have often interpreted that proportionality principle. However, it is flawed. There are several problems with allowing factors not directly related to the offense to have a role in evaluating offense seriousness.

First, it is contradictory to claim that the principle of proportionality means the punishment should be commensurate with the objective seriousness of the offense and then allow considerations external to the offense to have a role in determining how much punishment is appropriate. Once the inquiry extends to matters not even remotely connected with the crime, such as the offender’s upbringing or previous convictions, the parameters of the offense have been clearly exhausted.

Secondly, by importing other considerations (especially aggravating and mitigating factors) into proportionalism, much of the splendor of the principle of proportionality is dissipated. The principle then cannot be claimed as being indicative of anything: to ascertain how much to punish, the appealing idea of looking only at the objective seriousness of the offense is abandoned and the inquiry must move elsewhere—and, indeed, everywhere. Giving content to the principle of proportionality would become unworkable—as is currently the case. In each particular sentencing inquiry the principle would need to be flexible enough to accommodate not only the objective circumstances of the offense, but also the mitigating circumstances. Given the uniqueness of each offender’s personal circumstances and the vast number of variables which are supposedly relevant to such an inquiry and the fact that mitigating factors often pull in a diametrically opposite direction to the objective factors relevant to the offense, any attempt to provide a workable principle of proportionality must fail.

Some commentators have suggested that culpability in particular is part of the proportionality thesis.²⁹² The better position, however, is that it stands outside the doctrine. Culpability is a broad concept and involves varying degrees of blameworthiness. The broadest demarcation normally focuses on whether the offense was committed intentionally, recklessly, or, in some cases, negligently. However, within these categories are numerous sub-categories. Hence, intentional offenses which are planned are normally regarded as more blameworthy than those committed on the spur of the moment. Incorporating this range of considerations into the principle would considerably negate the capacity for measurability and is not doctrinally coherent because it relates to a subjective consideration of the offenders’ mental state and does not relate to the harm experienced by the victim. A person who is crippled by a crime suffers no less depending on whether the act is negligent or intentional.

This is not to deny that culpability is relevant to sentencing. It is relevant to overall offense seriousness but not to the objective seriousness of the offense. Once the harm stemming from the offense is quantified, a premium can be added to incorporate different levels of offense blameworthiness. The culpability adjustment should not be too great given that, from the victim’s perspective, pain is pain, but it is not clear that the motivation behind the act causing the pain changes its intensity or duration.²⁹³ Similar considerations apply regarding the role that aggravating and mitigating factors have in the sentencing calculus. There may still be a role for circumstances, which are not relevant to the objective seriousness of the offense in the determination of the appropriate penalty. However, the basis for their relevance must stem from other considerations thought to be integral to the sentencing calculus,²⁹⁴ such as rehabilitation and community protection.

The above analysis provides some guidance regarding measuring offense seriousness. Further clarity will emerge if the focus on graduating offenses commences with offenses that have identifiable victims and, once a degree of consensus is obtained in that context, assessments should then be made in relation to offenses that have less identifiable and more remote forms of harm, such as drug and motor traffic offenses and offenses that potentially undermine important institutional structures and processes, such as perjury.²⁹⁵

2. Second Limb: Evaluating the Hardship of Sanctions

While there has been some consideration given to measuring crime severity, there has been less attention given to the other side of the proportionality equation, i.e. measuring punishment severity. Ryberg contends this is due to the underlying belief that the “answer is pretty straightforward”²⁹⁶—with imprisonment being clearly the harshest disposition. As Ryberg notes, the answer would seem to rest on the “negative impact on the well-being of the offender.”²⁹⁷ Von Hirsch and Ashworth also believe that it is less complex to rank punishments because the appropriate reference point seems to be the degree of suffering or inconvenience caused to the offender.²⁹⁸

The starting point is to evaluate the adverse impact of imprisonment, given that it is the harshest sanction and the one that probably has the least amount of diversity in its application—in all societies it minimally involves physical confinement. It is surprising how little research has been conducted into the extent to which this sanction actually sets back well-being.

The direct adverse impact of prison conditions has been well documented. And it has been known for several decades that the “pains” of imprisonment extend far beyond the deprivation of liberty. Other negative consequences of imprisonment include:

(1) The deprivation of goods and services;²⁹⁹ (2) The deprivation of heterosexual relationships;³⁰⁰ (3) The deprivation of autonomy;³⁰¹ and (4) The deprivation of security.³⁰²

What is less well understood is how these deprivations affect the life trajectories of prisoners. The evidence available indicates that it has a considerable negative impact that transcends the actual term of imprisonment. Imprisonment seems to have an adverse effect on well-being measures after the conclusion of the sentence, even to the point of significantly reducing life expectancy.

A study that examined the 15.5-year survival rate of 23,510 ex-prisoners in the U.S. state of Georgia, found much higher mortality rates for ex-prisoners than for the rest of the population.³⁰³ There were 2,650 deaths in total, which was a 43 percent higher mortality rate than normally expected (799 more ex-prisoners died than expected). The main causes for the increased mortality rates were: homicide, transportation accidents, accidental poisoning (which included drug overdoses), and suicide.³⁰⁴

Many offenders released from prison continue to have their wellbeing set back in more ways than increased mortality rates. A recent New Zealand study, for example, showed that post-release offenders displayed vulnerabilities associated with financial matters; drug temptations; decision-making; and social interactions.³⁰⁵

Former prisoners without strong social networks were especially vulnerable and often had difficulty meeting their own basic needs, including experiencing hunger, homelessness, and being unable to access health care.³⁰⁶

This data, although only cursory, suggests that imprisonment is a more painful disposition than it appears on face value. It is even more complex to make an assessment of the severity of other sanctions, such as probation, community work orders, and fines because of their variability. But at least, in theory, the problem is not insurmountable. The severity of sanctions would be evaluated by reference to their level of “onerousness.” Ryberg uses similar terminology in suggesting that the answer would seem to rest on the “negative impact on the well-being of the offender.”³⁰⁷ This requires the same types of considerations as those involved in the assessment of the other limb of the proportionality thesis.

It has been suggested that one cannot grade the severity of penalties because painfulness is a subjective concept.³⁰⁸ A taxi-driver who is deprived of his or her driver’s license feels the pain more severely than a person who works from home. This is no doubt true, but the same applies regarding the harm caused by criminal offenses. Pickpocketing five dollars from Bill Gates is hardly likely to cause him even the slightest angst, whereas stealing the last five dollars from a hungry homeless person may have a devastating effect upon him or her. Despite the enormous difference in the impact of these offenses, the law has no difficulty in making theft an offense and, secondly, it has not refrained from evaluating the general seriousness of such conduct.³⁰⁹ This is because in relation to any branch of law, generalizations must be made about the things that people value and the typical effect of certain behavior on those interests.³¹⁰

3. Matching the Punishment to the Offense: Worst Crimes to the Worst Forms of Punishments

The final problem regarding proportionality is how to match the severity of the punishment with the seriousness of the offense. The type and degree of punishment imposed on offenders should cause them to have their well-being set back to an amount equal to that which the crime set back the well-being of the victim. This enables a theoretical matching at least to be made. However, there is currently insufficient data to allow a precise ranking.

Nevertheless, some tentative conclusions can be drawn. The above data indicates that the effects of being either a victim of a serious sexual or violent crime, or an inmate of a prison may both have been previously underestimated. These experiences all seem to have profoundly negative effects on life trajectories that continue well after the immediate event has ceased. On this crude measure they are matched in terms of the relative negative impacts, hence, imprisonment is an appropriate disposition for serious sexual and violent offenses. Of course, this says nothing about the length of imprisonment that is appropriate for certain categories of sexual and violent offenses. Yet, this crude empirically-based technique is preferable to the randomness that currently exists in relation to offense and sanction matching.

This approach suggests that, as a general principle, imprisonment should be reserved for serious and violent offenders. This would constitute a profound change to prison demographics and result in a large reduction in prison numbers.³¹¹

V. Mitigating and aggravating considerations

An important step in developing a coherent sentencing model involves addressing the issue of aggravating and mitigating sentencing variables. As we have seen, in Australia there are several hundred such considerations. They operate less significantly in the United States but can nevertheless still impact considerably on the sentence.

There is no clear rationale for most of these considerations. However, logically, if they are to be retained they must have a demonstrable connection to the objectives of sentencing. All aggravating and mitigating considerations should be abolished unless a cogent justification is given for them. That is, they must either further a sentencing aim (in which case they would normally be mitigating) or frustrate a sentencing objective (in which case they would normally be aggravating).

To justify the retention or development of an aggravating or mitigating consideration, it is necessary to: (1) state the sentencing aim that is being invoked; and (2) show how the consideration will assist in promoting the aim. Thus, for example, remorse should only be a sentencing consideration if it is established the remorseful offenders are less likely to reoffend, in which case the need for incapacitation is lessened. But, until this type of connection is empirically validated, remorse should be excluded from the sentencing calculus.

It could be contended that abolishing sentencing variables could lead to tougher sentences.³¹² Tougher sentencing is not, however, the inevitable by-product of abolishing sentencing variables. This is because, to the extent that it emerges that certain variables are in fact irrelevant, there is no reason to think that they will come mostly from the mitigating, as opposed to the aggravating, side of the calculus. While variables such as remorse might become redundant, so too might variables such as prior criminal record which often serve to greatly increase the sentence. It is important to emphasize that the main aim of presumptive sentences is not to achieve consistency. This would be a mistake. As has been noted previously: “[I]t is a major mistake to construct a system of guidelines whose primary structural objective is to minimize inter-judge sentencing disparity.”³¹³

VI. The framework of a model sentencing system

The analysis in this Article makes for a vastly different sentencing system to that which presently exists in the United States or Australia. The main reform that should occur in the United States is substantive: there should be a reduction in the severity of most sanctions. The main reform that should occur in Australian is formal: that is, sentences should be made more consistent and predictable by increasing the number of presumptive penalties.

The most profound systematic change that is necessary stems from the fact that reliance on imprisonment should be significantly diminished given that “old favorites,” such as specific deterrence, general deterrence, and incapacitation (apart from cases of prior criminality) can no longer be invoked to justify incarceration because they do not work and need to be discarded as sentencing goals. It follows that what we should be doing is watering down the severity of punishment for many offenses.

The nature and length of sentences needs to be primarily guided by the principle of proportionality. We should move towards a two-pronged system of punishment. We should deal more harshly with sex and violent offenders and show more leniency to other types of offenders, given the differential impact of those offenses. This is because the current empirical data shows that sex and criminal offenses devastate the lives of victims, while other offenses are normally less damaging. As noted earlier, given that presently only approximately 50 percent of prisoners are in jail for sex and violent offenses, this approach would result in a reduction in the prison rate. At the same time, we should be striving for a lower crime rate. The key to this is to divert some of the funds saved from prison expenditure to policing, so as to achieve a higher visible police presence.

The goals of less punitive sanctions and less crime may seem overly ambitious, but it is not unattainable. Any intuitive unease that lowering imprisonment rates would inevitably lead to increased crime rates is to some extent allayed by a comparison of sentencing practice in jurisdictions such as the Scandinavian countries. There, sentencing premiums are not attached to pursue the aims of incapacitation or deterrence, and the main determinant in setting criminal sanctions is the principle of proportionality.³¹⁴

These systems are founded on a (relatively) forensic approach to sentencing that have many similarities to the model we have proposed. The Swedish and Finnish criminal justice systems have both managed to achieve the twin goals of low crime rates and low incarceration levels.³¹⁵

The main rationales underpinning Swedish sentencing law are proportionality and equivalence.³¹⁶ The seriousness of an offense (its “penal value”) is determined by reference to “the harm, offense or risk which the conduct involved, what the accused realized or should have realized about it, and his intentions or motives.”³¹⁷ This penal value is mapped onto a scale, ranging from a small fine to life imprisonment.

There is a presumption against imprisonment that can be rebutted in a number of circumstances, one of which is that the offender has previous convictions, especially if they are recent (within four years); similar to the current offense (both in terms of offense type and gravity); and there is more than one previous offense.³¹⁸ Thus, prior convictions can influence the choice of sanction and, further, can also affect the length of a prison sentence. However, when this does occur, the increase in length is “usually not excessive.”³¹⁹

In relation to the Finnish model, it has been noted that the “approach is driven by an analysis of what sentencing should aim to achieve, tied in closely with a cost-benefit analysis of the effectiveness of various penalties in achieving the aims of general social policy, such as a reduction in crime through prevention, deterrence, and rehabilitation.” The prison rate in Finland is about half of that in Australia. Moreover, the crime rate in Finland is about 35 percent lower than that in Australia.³²⁰

It is important to emphasize that the Finnish did not achieve success in sentencing by accident. Thirty years ago, Finland had a strict criminal justice regime inherited from neighboring Russia, and one of the highest rates of imprisonment in Europe. However, academics provoked a fundamental rethinking of penal policy, urging that it should reflect the region’s liberal theories of social organization. The Director of the Finnish National Research Institute of Legal Policy, Tapio Lappi-Seppala, stated that the “Finnish criminal policy is exceptionally expert-oriented.”³²¹ He went on to say, “We believe in the moral-creating and value-shaping effect of punishment instead of punishment as retribution.” Over the past two decades, more than forty thousand Finns have been spared prison, $20 million in costs have been saved, and the crime rate has gone down to relatively low Scandinavian levels. The Finns openly state that “[w]e don’t believe in an eye for an eye, we are a bit more civilized than that, I hope.”³²²

The Scandinavian experience is a victory of principle over expedience and is a model that should be followed by United States and Australian law-makers. I do not suggest that transporting the Finnish or Swedish models to the United States and Australia will necessarily have the same outcomes. There are many variables that impact the workings of criminal justice initiatives, including social, political, and environmental considerations. To this end, it is important to note that Finland and Sweden are relatively classless cultures with a Scandinavian belief in the benevolence of the state and a trust in its civic institutions.³²³ However, the Finnish and Swedish experiences do provide a strong foundation for confidence that if the development of sentencing policy and practice is guided by expert analysis, it will become a far more sophisticated and workable system.

One controversial aspect of the above proposal is the endorsement of mandatory minimum penalties. These have been heavily criticized in the United States. For example, an international comparative report on sentencing by the University of San Francisco Center for Law and Global Justice states:

Mandatory minimum sentences provide “one size fits all” punishments that do not permit consideration of the individual circumstances of the crime, the offender, their past history, or other considerations traditionally reserved for the legal decision maker.³²⁴

As noted in the report, the American Bar Association has recommended the repeal of mandatory minimum sentences because they do not allow for sentences to be “both uniform among similarly situated offenders and proportional to the crime that is the basis of the conviction.”³²⁵

These criticisms are overstated. The United States’ fixed-penalty sentencing regime does not provide for “one size fits all” given that it is broken down to offense type and previous history is a major determinant regarding penalty.

The real problem with the current fixed penalty regime in many current United States jurisdictions is the level at which the penalties are set. A United States Sentencing Commission survey found that 62 percent of federal judges believed that the penalties were too severe.³²⁶

If proportionality is used to underpin a fixed penalty regime, this criticism is likely to dissipate. There is no necessary link between fixed penalties and harsh penalties. For example, several Scandinavian countries (Denmark and Norway) have mandatory minimum penalties for murder, but they are no more than six years.³²⁷

Conclusion

Sentencing is the final stage of the criminal justice process, and it has a profound impact on the integrity and effectiveness of how society deals with criminals. Currently, the large gulf between knowledge and practice has led to poor implementation of sentencing procedures. This is because there is a large gulf between knowledge and practice. Such a chasm emerged and became entrenched to some extent because the treatment of criminals does not command extensive or thorough analysis. To the contrary, there is an instinct in democracies to simply penalize criminals. Mistakes and intellectual sloppiness are thus tolerated in this environment.

Given the crude manner in which sentencing has evolved, there is room for considerable skepticism that, at least in the foreseeable future, sentencing will be transformed from a confused and fragmented practice into something akin to a social institution underpinned by a body of empirical and normative knowledge. In order for such a transformation to occur, it is imperative to take an overarching approach to sentencing and not make any assumptions about the validity of existing sentencing policies and practices.

Lawmakers may ultimately choose to ignore the recommendations stemming from such an approach. The populist nature of the democracies in the United States and Australia is a considerable obstacle for principle to trump unbridled community passion in this area. In order for a new evidence-based sentencing regime to be implemented, it is necessary to convey to politicians the self-defeating nature of current sentencing law and to develop a coherent and clear framework for sentencing reform. In simple terms, the framework is:

• Sentences should be adjusted so that the harshness of the

sanction matches the harm caused by the crime;

• Sentencing premiums should not be added to pursue the

aims of incapacitation or deterrence;

• The most severe punishments should be reserved for serious

sexual and violent offenders;

• A recidivist premium of approximately 20 to 50 percent

should be added to serious offenders;

• All sentences should be determined by way of pre-determined presumptive grids; • Deviations should only occur from this for demonstrable

reasons.

The best way to reduce crime is to increase the visible police presence. Some of the taxpayer dollars saved on prisons should be diverted to policing. This will result in a massive reduction in the expenditure of prisons, allowing more money to be spent to productive community services, such as health and education.