• Specific Year
    Any

Gray, Anthony --- "Standard of Proof, Unpredictable Behaviour and the High Court of Australia's Verdict on Preventive Detention Laws" [2005] DeakinLawRw 9; (2005) 10(1) Deakin Law Review 177

STANDARD OF PROOF, UNPREDICTABLE BEHAVIOUR AND THE HIGH COURT OF AUSTRALIA’S VERDICT ON PREVENTIVE DETENTION LAWS

ANTHONY GRAY[*]

[Preventive detention laws authorize courts to order the continued detention in prison of a person who has served their allocated term of imprisonment, but who are thought to be at risk of re-offending if released. They raise fundamental issues about the separation of powers, the purpose of incarceration, and the standard of proof which is/should be required to authorize detention. They assume that it is possible to predict, with a satisfactory rate of success, whether or not a past offender would if released commit further offences. Recently, a majority of the High Court of Australia validated such legislation. The author in this article explains his reasons for disagreeing with the verdict of the Court in this matter.]

I INTRODUCTION

A recent attempt by a State Parliament to introduce preventive detention, allowing for the continued detention of an offender after their original term of sentence has expired, raises fundamental constitutional law and criminal procedure issues. Chief among them are whether the doctrine of separation of powers permits a State Court to be part of such a process, if so what the relevant standard of proof should be, and how it compares with the traditional criminal standard of beyond reasonable doubt, and current thinking on the predictability of future behaviour. This article addresses each of these issues, and submits that the recent majority High Court decision validating such preventive detention laws is, with respect, flawed on many counts.

In October 2004 the High Court of Australia, by a verdict of 6-1, upheld the validity of Queensland preventive detention legislation.[1] The 2003 law allows the Queensland Attorney-General to apply to the Supreme Court for an order that a person who has been convicted of a serious sexual offence be detained in prison for a further indefinite period after the end of the sentence which they were previously given.[2] In other words, further incarceration may be authorized not because of something the prisoner has done, but based on an assumption of what they may do if released. For many, this was an unexpected result.

In an earlier article in the Deakin Law Review,[3] published prior to the High Court’s ruling, the author expressed his strong belief that the legislation should, consistent with the High Court’s own ruling in the Kable decision in 1995,[4] be struck down on constitutional grounds. The legislation may also be said to run counter to fundamental legal principles, such as the rule of law and the common law right to liberty. Some see the legislation as running counter to the fundamental notion of presumption of innocence. It has been suggested that, as a result of the High Court’s decision in this case, other State Parliaments are planning preventive detention regimes. The issue of the validity of such schemes, and how far a State Parliament (or possibly the Federal Parliament) may constitutionally be able to proceed along this path, is likely to remain a contentious issue in the foreseeable future.

This article is split into several parts. Part II outlines the previous law in Australia on preventive detention regimes, based on the principle of separation of powers, and compares it with the High Court’s view in the Fardon case. An aspect of the law requires that the Queensland Government prove that the offender be likely to re-offend ‘at a high degree of probability’. The decision in Fardon implies that this is an acceptable standard of proof as a basis for further incarceration. Part III of the article outlines current models on the predictability of future behaviour. Part IV elaborates on the concept of reasonable doubt, and comparisons are made with the Fardon standard and the criminal standard of proof. Part V draws upon other arguments to suggest that such regimes are problematic and ill-conceived, including the suggestion that Fardon-type laws are in fact criminal in nature rather than civil. Other ways of dealing with the problem that those kinds of laws purport to address will be briefly considered.

II KABLE AND THE PRINCIPLE OF SEPARATION OF POWERS

A Kable

A similar law to the one challenged in Fardon was passed by the New South Wales in 1994. Specific details of that legislation challenged in the Kable case appear in the earlier Deakin Law Review article. Suffice to say for present purposes that the challenged law allowed the State Attorney-General to apply to that State’s Supreme Court for an order that (in the end, a particular named offender) be detained indefinitely, if the court was satisfied that the person was more likely than not to commit a violent offence if released. The Act was couched in terms of community protection.[5] Only the civil standard of proof was required in terms of making out a case under the legislation, and the court was not bound by traditional rules of evidence in assessing the application.

A majority of the High Court in that case struck down the legislation as being offensive to the Commonwealth Constitution. The judges accepted that there was no separation of powers in individual State Constitutions,[6] but a majority found that as State Courts exercised federal jurisdiction, they were part of the federal court structure, and as such the formal separation of powers that clearly did exist in the Commonwealth Constitution was drawn down to State courts in the federal hierarchy, including State Supreme Courts.

It is necessary to elaborate on the reasoning of the majority in that case in order to assess the judgments in the recent Fardon decision.[7]

1 Toohey J

His Honour noted that the preventive detention regime in the Act differed from existing recognized categories of non-punitive (or civil) detention, including mental illness, infectious disease, the traditional power to punish for contempt, or military discipline tribunals.[8] He noted that preventive detention was an end in itself. It was not an incident of the exclusively judicial function of adjudging and punishing criminal guilt. He found the Act infringed the Grollo principle,[9] in that it asked a judicial body to undertake non-judicial functions of such a nature that public confidence in the integrity of the judiciary as an institution would be diminished. He said that the law infringed the principle of separation of powers because

It requires the Supreme Court to participate in the making of a preventive detention order where no breach of the criminal law is alleged and where there has been no determination of guilt. On that ground I would hold the Act invalid.[10]

2 Gaudron J

Gaudron J agreed with the proposition that a State Parliament could not confer on its Supreme Court powers which were

repugnant to or incompatible with their exercise of the judicial power of the Commonwealth.[11]

She noted that the proceedings contemplated by the Act were not proceedings ordinarily known to the law. Acknowledging the Act attempted to dress them up as legal proceedings by referring to the rules of evidence, they did not partake of the nature of legal proceedings. This was because

they did not involve the resolution of a dispute between contesting parties as to their respective legal rights and obligations ... The applicant is not to be put on trial for any offence against the criminal law. Instead, the proceedings are directed to the making of a guess - perhaps an educated guess, but a guess nonetheless – whether, on the balance of probabilities, the appellant will commit an offence of ‘serious violence’ ... The power ... requires the making of an order .. depriving an individual of his liberty, not because he has breached any law, but because an opinion is formed, on the basis of material which does not necessarily constitute evidence admissible in legal proceedings, that he is more likely than not to ... commit a serious acts of violence. That is the antithesis of the judicial process, one of the central purposes of which ... is to protect the individual from arbitrary punishment and the arbitrary abrogation of rights by ensuring that punishment is not inflicted and rights are not interfered with other than in consequence of the fair and impartial application of the law to facts which have been properly ascertained. It is not a power that is properly characterized as a judicial function, notwithstanding that it is purportedly conferred on a court and its exercise is conditioned in terms usually associated with the judicial process ... the effect of the (Act) is ... to compromise the integrity of the Supreme Court ...[12]

She added that

Public confidence cannot be maintained in the courts and their criminal processes if .. the courts are required to deprive persons of their liberty, not on the basis that they have breached any law, but on the basis that an opinion is formed, by reference to material which may or may not be admissible in legal proceedings, that on the balance of probabilities, they may do so.[13]

Referring to the proceedings in the Act, she found that they were “dressed up as proceedings involving the judicial process. In so doing, the Act makes a mockery of that process, and inevitably weakens public confidence in it.”[14]

3 McHugh J

His Honour agreed that no government could act in a way that undermines public confidence in the impartial administration of the judicial functions of State Courts.[15] If it could, it would inevitably result in a lack of public confidence in the administration of justice by those courts. He found this Act did tend to undermine public confidence in the impartiality of the Court. McHugh J summarized the Act, concluding its object was to detain the defendant not for what he had done, but for what the Government feared he may do. He mentioned that the rules of evidence had been relaxed, and there was no need to meet the beyond a reasonable doubt test. This meant the law undermined the ordinary safeguards of the judicial process, made the Supreme Court the instrument of a legislative plan, initiated by the executive government, to imprison the appellant by a process that is far removed from ordinary judicial process.[16] The proceedings were dressed up as civil in nature, although the court was not asked to determine the existing rights and liabilities of any party. He concluded the Act compromised the integrity of the Supreme Court.

4 Gummow J

Gummow J noted the legislation provided “no determination of guilt solely by application of the law to past events being the facts as found. The consequence is that the legislature employs the Supreme Court to execute, to carry into effect, the legislature’s determination that the appellant be dealt with in a particular fashion, with deprivation of his liberty, if he answers specified criteria”. Referring to comments that “the legitimacy of the judicial branch ultimately depends on its reputation for impartiality and nonpartisanship”, Gummow J concluded that this reasoning “is particularly applicable where, as here, the Act draws in the Supreme Court of a State as an essential and indeterminate integer of a scheme whereby, by its order, an individual is incarcerated in a penal institution otherwise than by breach of the criminal law”.[17] His Honour’s conclusion was inevitable:

The Act requires the Supreme Court to inflict punishment without any anterior finding of criminal guilt by application of the law to past events, being the facts as found. Such an activity is said to be repugnant to judicial process. I agree”.[18]

He added that

the making ...of detention orders by the Supreme Court in the exercise of what the statute purports to classify as an augmentation of its ordinary jurisdiction, to the public mind, and in particular to those to be tried before the Supreme Court for offences against one or other or both or the State and Federal criminal law, is calculated to have a deleterious effect. This is that the political and policy decisions to which the Act seeks to give effect, involving the incarceration of a citizen by court order but not as punishment for a finding of criminal conduct, have been ratified by the reputation and authority of the Australian judiciary. The judiciary is apt to be seen as but an arm of the executive which implements the will of the legislature. Thereby a perception is created which trenches upon the appearance of institutional impartiality to which I have referred.[19]

He concluded that the most significant part of the legislation was that “whilst imprisonment pursuant to Supreme Court order is punitive in nature, it is not consequent upon any adjudgment by the court of criminal guilt ... not only is such an authority non-judicial in nature, it is repugnant to the judicial process in a fundamental degree.”[20]

The decision in Kable can thus be considered largely to be founded on a breach of the principles of separation of powers. Of the judges who heard the case, only Dawson J considered a further question of whether the exercise of statutory power was subject to some limits based on common law. He rejected such a proposition, raised tantalizingly as a possibility by the High Court in Union Steamship v King,[21] on the basis of the fundamental nature of Parliamentary supremacy. No argument was raised that the legislation was contrary to the rule of law,[22] or of international civil rights covenants.[23]

B High Court’s Decision in Fardon

One might have thought that the Kable case, and the recent Fardon decision, involved very similar issues and principles. Though, as will be explained, some of the details of the legislation differed, in essence the principle (and the objection) is considered to be the same. However, (problematically and unexpectedly) of the four members of the court in Kable who had struck down that legislation, the two who remained on the bench and heard the Fardon case both upheld the legislation as valid, as did more recent appointees.

C Queensland Preventive Detention Law

Again, specific details of the challenged law appear in the earlier Deakin article. Essentially, the legislation applies to a prisoner in custody who is serving a period of imprisonment for a serious sexual offence. It applies to prisoners convicted of such an offence either before or after the legislation was introduced. A serious sexual offender is a person who has been convicted of an offence of a sexual nature involving violence or against children.[24]

On the application of the Attorney-General, if the Supreme Court is satisfied there is an unacceptable risk the person would commit a serious sexual offence if released, they may make an order in relation to the offender, including an order that the person be detained for a further period. The court must be satisfied of this risk ‘by acceptable, cogent evidence and to a high degree of probability.’ Medical and psychiatric evidence will be very important in assessing the application. In considering the application, the court is reminded that the object of the legislation is community protection, and this is the paramount consideration to be taken into account.[25]

D Similarities and Differences between the Kable Law and the Fardon Law

As discussed, the decision of 6 of the 7 High Court judges in the Fardon case was that the legislation was valid, and that the Kable principle had not been breached. Logically, since the correctness of the Kable decision was not challenged in Fardon or questioned by any members of the High Court, this different result can only be rationalized on the basis that:

(a)the legislative regimes are different in material respects; and/or
(b)the Kable principle has been reinterpreted, at least by a majority of the High Court. This result would be somewhat surprising, particularly by McHugh and Gummow JJ, who were part of the original Kable decision and spoke very strongly against the preventive detention law impugned in that case.

Let us consider whether the legislative regimes are different, as claimed.

E Are the Legislative Regimes Different in Material Respects?

Members of the majority of the High Court in Fardon drew some clear distinctions between the legislation in that case, and the New South Wales equivalent in the Kable case. Some differences can be conceded:

(a)The Kable law applied to one named offender, while the Queensland law applied to a category of offenders – namely those convicted of a serious sexual offence.
(b)The standard of proof differed, with the Kable law requiring only that the court be satisfied that the offender, if released, would be more likely than not to commit a further violent crime. The Queensland equivalent required the court to be satisfied to a high degree of probability.[26]
(c)The only possible orders that the Court in the New South Wales case could make were either that the person be detained for an initial six month period, or that the person not be detained. The Queensland equivalent allows the Supreme Court both of those options, but also a third option, namely that of supervised release.
(d)The New South Wales law applied whether or not the prisoner was incarcerated at the time of the application; the Queensland law applies only to those in incarceration at the time the application is made.
(e)The proceedings under the New South Wales law were stated to be civil in nature, while the Queensland proceedings are not classified as being either civil or criminal.
(f)The court was stated to be bound by the rules of evidence in the New South Wales law, but may order the production of reports etc which may have been held inadmissible in ordinary judicial proceedings.[27] The Queensland law allows the court to take into account ‘any relevant matter’ when assessing the application. McMurdo P of the Queensland Court of Appeal concluded this provision allowed the use of material which would not be admissible in an ordinary legal proceeding.[28]
(g)In both cases, a decision to order further detention could be appealed by the person affected.

There is a limited attempt made by the majority of the High Court in Fardon to rely on the above points of distinction to justify not applying the Kable precedent to the Queensland legislation challenged in this case. With respect, the author does not find the attempts to be convincing.

For example, Gleeson CJ noted that the Kable law applied only to one offender, whereas the Queensland law was not so confined.[29] He noted the Act conferred discretion as to whether an order should be made, and the kind of order. As much can be conceded. The question is whether these differences in material ones, when considering the ratio of the decision in Kable and the fundamental constitutional and civil liberties principles involved.[30] He noted that the rules of evidence applied, and the Attorney-General bore the onus of proof. Hearings were to be conducted in public, with a right to appeal. Each case was determined on its merits.

With respect, these are not considered to be satisfactory points on which to distinguish the present case from Kable. In the Kable case, the rules of evidence applied also. The Attorney-General bore the onus of proof in that case. Hearings were to be in public, with a right to appeal. There was nothing to suggest in that case that the matter would not be determined “on its merits”.[31]

However, perhaps the most interesting attempts to distinguish the Kable precedent come from those judges who participated in the majority verdict in that case, McHugh and Gummow JJ.

McHugh J claimed there were “substantial” differences between the law in Kable and the Fardon legislation.[32] He mentioned the Kable law’s focus on the individual person involved, rather than the category of offender, and the fact the law gave the court hearing the matter three options in terms of making orders, rather than the two options in the Kable case. These differences are admitted, but again, the author questions whether they are significant enough to justify an opposite conclusion. McHugh J referred to the higher standard of proof in the Fardon case. While admittedly there is a difference, the standard is still not at the criminal standard.[33] It is hard to justify the conclusion that a law requiring a more likely than not proof level compromises the integrity of the court, yet a law requiring a “high degree of probability” proof level does not.

His next claimed justification was that

in determining an application under the Act, the Supreme Court is exercising judicial power. It has to determine whether ... the court is satisfied that there is an unacceptable risk that the prisoner will commit a serious sexual offence if the prisoner is released from custody. That issue must be determined in accordance with the rules of evidence.

All of this may be true, but it is not considered to be justification for distinguishing the law in Kable. In Kable, the judges were asked to consider whether a person was more likely than not to re-offend. The rules of evidence were applicable.[34] Where is the difference?

The next claimed justification is a statement that “the Act is not designed to punish the prisoner. It is designed to protect the community against certain classes of convicted sexual offenders.” He then referred to a statement in the Act that its object was community protection.[35] Let us compare the community protection pedigree of both laws. The Kable law was entitled the Community Protection Act 1995 (NSW), and s3(2) stated that ‘in the construction of this Act, the need to protect the community is to be given paramount consideration’. The Queensland law states in s3 its object is to ensure adequate protection of the community, and in s13 that community protection is the paramount consideration. With respect, how can it be seriously argued that a real difference between the two laws is that only one of them is based on community protection?

The final supposed justification for the non-application of the Kable principle here was that there was nothing in the Act or the surrounding circumstances suggesting that the law was a disguised substitute for a legislative or executive plan. Nothing gave rise to the perception that the Supreme Court was acting in conjunction with, and not independently of, the Queensland legislature. Let it be noted that the Fardon law was passed in a rushed fashion, two days prior to the due release date of prisoner Fardon. Not surprisingly, he was the first prisoner about whom an application for further detention was made. It was successful, as have all other applications to date.[36] Is this a materially different situation from Kable, where the New South Wales government passed a law about a named individual, in response to community concerns that the offender may re-offend if he were released back into the community, as was imminent. The Supreme Court granted the application and he was detained further, until the law was struck out.

It is submitted that these kinds of extreme laws are usually passed in reaction to a specific perceived problem in the prison population. They are not usually pro-active in nature. Given this reality, does the State Parliament avoid constitutional problems by not naming the person or persons who have caused them to pass such extreme laws, but fall foul of difficulties if it happens to name the person who triggered the law being passed? Does it avoid problems if the Minister introducing the legislation avoids all reference to the factual context that in all likelihood led to the development of the law? Does it avoid constitutional problems by framing the law as applicable to a category of offender, albeit a very small category, rather than an individual? It seems, from comparing the judgments in the two High Court decisions, that the answer to these questions is yes.

Gummow J, also arriving at a different conclusion than the one he made in Kable, noted by way of justification the higher standard of proof in the current legislation.[37] An appeal was possible from a decision to grant an application.[38]

There is little reference to the principles Gummow J himself espoused in the Kable case. He referred with approval to the repugnancy test – whether the law was repugnant to or incompatible with the institutional integrity of state courts, such that it was likely to undermine public confidence in the courts.[39] Tantalisingly, he admitted that “detention by reason of apprehended conduct is of a different character (than other non-punitive detention such as mental illness and infectious disease) and is at odds with the central constitutional conception of detention as a consequence of judicial determination of engagement in past conduct.”[40] However, he was persuaded by the argument that since the prisoner had been guilty of a past crime, “there remained a connection between the operation of the Act and anterior conviction by the usual judicial processes”.[41] With respect, the same may also have been said about the Kable legislation. Mr Kable had been previously convicted of a criminal act. This did not prevent Gummow J in that case finding the legislation providing for the prisoner’s continued detention to be invalid. His Honour mentioned that the further detention was for a limited time.[42]

In lone dissent holding the legislation to be invalid was Kirby J. He held it to be fundamental in Australia that an individual not be imprisoned because of their beliefs, nor for future crimes which they may or may not commit. He found that the Kable principle extended to legislation which applied to a small number of offenders, rather than one offender, as was the case here.[43] Noting that liberty, while not absolute, had long been recognized in Australia as the most fundamental of rights, he found the Act did not confer judicial power, which involved the application of law to past events or conduct.[44] While there had been some attempt to dress this Act up as being civil in nature, and while he conceded that civil commitment was possible in strictly limited cases in Australia, the undeniable character of this law was punitive in nature.[45] Legislation of this kind did, he found, undermine public confidence in the judiciary in a fundamental way.[46]

The conclusion is that the argument that the legislation in Fardon and that in Kable are materially different is a difficult one to run. Differences between them are not seen by the author as justifying the different results reached in the cases. It is clear that the High Court has reinterpreted Kable in an extremely narrow way, and one would be skeptical that the principle, potentially a strong guarantee of judicial independence and civil liberties,[47] could be applied with much scope in future, at least with the current bench. As Kirby J noted, it may indeed be the “constitutional watchdog that would bark but once”.[48]

Apart from arguments about separation of powers and civil liberties, the main objections to legislation of this kind are considered to be

(a) the assumption that it is possible with a satisfactory degree of accuracy to determine whether or not a person may re-offend if released, and to use this determination as a basis for involuntary detention;
(b)the different standard of proof required under the preventive detention law for further incarceration; and
(c)arguable, the legislation is defacto criminal in nature, and runs counter to the orthodox legal view that it is better that guilty people go free than an innocent person be detained.

Parts III, IV and V of the article respectively deal with these issues in more depth.

III CURRENT THINKING ON THE PREDICTABILITY OF FUTURE CRIMINAL BEHAVIOUR

There is ample evidence of judicial notice[49] of the unpredictability of future criminal behaviour. In the Fardon case itself, Kirby J reiterated that ‘experts in law, psychology and criminology have long recognized the unreliability of predictions of criminal dangerousness’. He referred to expert evidence that psychiatrists notoriously overpredict. Predictions of dangerousness have been shown to only have a one-third to 50% success rate.[50] For the purposes of this article, the author researched the most popular current models for predicting future dangerous behaviour.[51]

One promising model for predicting future dangerous conduct is the HCR-20 Risk Assessment Scheme, developed in 1997 by a team of Canadian researchers.[52] The model considers three sets of factor when assessing likely future dangerousness, including historical factors, clinical and risk management factors. The factors are stated to be:

(a)Historical – previous violence, young age at first violent incident, relationship instability, employment problems, substance use problems, major mental illness, psychopathy, early maladjustment, personality disorder, and prior supervision failure (H scale);
(b)Clinical – lack of insight, negative attitudes, active symptoms of major mental illness, impulsivity, unresponsive to treatment (C scale);[53] and
(c)Risk Management – plans lack feasibility, exposure to destabilizers (eg a peadophile living close to a school, a drug addict close to a ready supply of drugs), lack of personal support, non-compliance with remediation attempts, and stress (R scale).

In terms of the predictive value of these factors, one group of researchers found that the H and C subscales were related to ward violence with moderate strength in a sample of 131 civilly-committed acute-care psychiatric patients.[54] A group of Swedish researchers using the HCR-20 model on 40 male forensic psychiatric patients found that recidivists scored significantly higher on the HCR-20 than did non-recidivists, particularly on C and R subscale items.[55] In a study involving 175 insanity acquitees, another team found that those who scored above the median on the HCR-20 were several times more likely that those scoring below the median to have histories of violent crimes.[56] Research by Douglas, Ogloff, Nicholls and Grant found the HCR-20 to be a strong predictor of violent behaviour - .76 for any violence and physical violence, .77 for threatening behaviour, and .80 for violent crime.[57] The HCR-20 model has been less useful in distinguishing between different levels of risk for different kinds of violence.[58]

Another model (known as the PCL:SV) [59] involves the use of factors classified as emotional/interpersonal traits[60]or behavioural factors.[61] Research by Douglas, Ogloff, Nicholls and Grant found the PCL:SV to be a reasonable predictor of violence, ranging from .68 for any violent and threatening behaviour to .73 for physical violence to .79 for violent crime.[62]

Other recent models propose using other factors. One common list of risk factors for violence is male gender, young age, previous violence, substance abuse, psychopathy, childhood abuse and maladjustment, positive psychotic symptoms, suicidality, impulsivity, anger, treatment non-compliance, lack of community support or supervision, poor family relations and stress.[63] Another team of researchers were able to correctly classify patients as violent or non-violent during a one year community follow-up at a 13% improvement in comparison with chance variables, by using such factors as arrest history, early family quality, current intimate relationship, admissions history and assault as part of the previous problem.[64] McNiel and Binder found that a five-item tool could predict inpatient violence at a level of 25% greater than chance. The factors included violence in the weeks prior to admission into a mental facility, absence of suicidality, manic or schizophrenic diagnosis, male gender, and married or cohabiting partner absent.[65]

Another model, known as the Psychology of Criminal Conduct model, was developed in the 1980s and refined over time,[66] and retains currency among some researchers. It embraces social learning theory, that through association with those who hold pro-criminal attitudes, people develop similar attitudes, drives, perceptions and motives. The principle of risk-needs-responsivity is recognized, whereby prediction of criminality relates to (a) risk factors relating to subsequent offending (b) whether the offender’s ‘criminogenic needs’ have been dealt with while in prison, or whether they remain at the time of release,[67] and (c) the offender’s responsivity to any treatment obtained.[68] The most recent research by Ogloff and Davis has found that treatment of offenders must be directed towards these criminogenic factors in order for the risk of re-offend to be substantially reduced.[69]

A Summary of Current Thinking on Predictability of Future Behaviour

The body of knowledge in this complex area is continuing to develop. There is some consensus on some of the factors increasing the likelihood of future violence, but some difference of opinion. Refined models that allow for prediction of future behaviour are producing reasonably accurate results, at least to date. There is debate about the emphasis in rehabilitation on the prospects of re-offence, and resultant uncertainty about a key element of future prediction – whether the offender has been adequately rehabilitated while incarcerated.

B Conclusion on Predictability of Future Behaviour

Predictive models for future behaviour are improving, but can only ever provide us with a probability, or informed guess, as to the likelihood of reoffence. Some models studied had claimed a 70-80% successful prediction rate for future offending. This is likely to meet the test for continued detention given by the Fardon law – a high degree of probability that the offender would, if released, re-offend. However, as the developers of the model themselves acknowledge, this still contains a high margin of error. Twenty to thirty percent of the time, even the advocates of the model acknowledge that they get it wrong. We should expressly recognize here also that a claimed 70-80% accuracy rate for some models is much less than the claimed rate for other models. A meta-study of 23 different predictive models in 2001 found an average success rate of 58%.[70] Recall that in his judgment in Fardon, Kirby J referred to a success rate of 33-50% in predicting future dangerousness.[71]

The difficulty with this uncertainty is that it is on an issue which is being assessed to consider whether or not to remove a fundamental right of a citizen, that of liberty.[72] How different is this high degree of probability test compared with the traditional criminal standard, and presuming they are different, can there be a justification (beyond mere pragmatism) for accepting a lower standard of proof for the offender’s continued incarceration? Why is (at best a) 70-80% likelihood of re-offence sufficient to keep someone in prison longer, when it would not likely have been good enough to put them there in the first place? We will turn to the issue of standard of proof presently in Part IV of the article.

Faced with this dilemma, how did the Court respond? Of the majority in Fardon, only the Chief Justice dealt with this key issue, and then only very briefly. Gleeson CJ opined that ‘no doubt predictions of future danger may be unreliable but ... they may also be right’.[73] The writer is not convinced that concerns about unpredictability can be dealt with in such a cavalier fashion. Suspicion that a person may have committed an offence in the past “may be unreliable but may also be right.” No-one would suggest that this would justify putting them behind bars in the first place. With great respect, why should suspicion about the person’s future behaviour be any different, in a coherent legal system?

IV STANDARD OF PROOF

Given that the Act allows for the continued detention of a person at a standard of proof different from the traditional criminal standard of beyond reasonable doubt, it is interesting to compare differences between the two standards. How different is the Fardon standard requiring a ‘high degree of probability’, compared with the orthodox criminal standard of “beyond reasonable doubt”?[74] If a 70-80% success rate is sufficient to meet the “high degree of probability” test, is it sufficient to meet the “beyond reasonable doubt” test?[75]

A Non-Mathematical Definitions of the Criminal Standard

The concept of proving criminal guilt beyond reasonable doubt has been described as a “golden thread (running) throughout the web of the English Criminal Law”,[76] but Australian courts have been somewhat reluctant to elaborate on its meaning.[77] It has been established that a reasonable doubt is not confined to a rational doubt or a doubt founded on reason.[78] It has been referred to as a “solicitude for certainty”.[79] There is some suggestion that the persuasion of guilt should amount to a moral certainty,[80] an abiding conviction[81], or a doubt that would cause a reasonable person to hesitate to act.[82] Sir Owen Dixon put it this way:

If an issue is to be proved by circumstantial evidence, facts subsidiary to or connected with the main fact must be established from which the conclusion follows as a rational inference. In the inculpation of an accused person the evidentiary circumstances must bear no other reasonable explanation. This means that, according to the common course of human affairs, the degree of probability that the occurrence of the facts proved would be accompanied by the fact to be proved is so high that the contrary cannot reasonably be supposed.[83]

Applying this definition to the case at hand, it would not seem that a 70-80% prediction of future behaviour (even placing the claimed accuracy rates at their highest), even if it could be attained, would meet the criminal standard of proof. It is submitted that a 20-30% failure rate does not equate to what cannot “reasonably be supposed”. It is concluded that the Fardon law requires a test for incarceration at a level substantially below that required by the criminal standard.[84]

B Mathematical Definitions of the Criminal Standard

Earlier in the article, reference was made to the most recent psychiatric literature, where model proponents estimated they could with a 70-80% degree of certainty determine whether or not an existing offender was likely to re-offend. As was indicated then, this could amount to the ‘high degree of probability’ required by the Fardon legislation as a basis for continuing to incarcerate a convicted sex offender. Intuitively, this level of accuracy, even if it could be obtained, seems less than the level required by the criminal standard of proof, that of beyond reasonable doubt. However, we should test this assertion.

One should acknowledge here that it remains a controversial idea that degrees of mathematical probability should be used in applying a legal standard of proof. While some judges expressly acknowledge the link between probability in the legal sense and probability mathematically,[85] others prefer to rely generally in evidence matters on an actual persuasion that the fact or occurrence is true.[86] While there is some recent judicial and academic support for a mathematical probability exercise in civil cases,[87] there has been little attempt to, and hostility towards, applying probability theory to criminal cases in the context of the ‘beyond a reasonable doubt’ test. In the early case of Brown v The King,[88] the High Court considered, and rejected, such a possibility. Isaacs and Powers JJ referred with apparent approval to a direction by an esteemed judge to a jury that ‘criminal cases must not be decided on the preponderance of probabilities, but on proof of guilt.’ Barton ACJ agreed that

where a juryman perceives such a preponderance (of probability) in a civil case sustaining the burden of proof, he is justified in deciding according to that greater weight of evidence. But the danger of applying a similar rule of action in criminal cases is manifest, because of the much more serious consequences ... which must result from a mistaken conclusion... (referring to another direction by a judge to a jury) in order to enable them to return a verdict against the prisoner, they must be satisfied, beyond any reasonable doubt, of his guilt; and this as a conviction created in their minds, not merely as a matter of probability; and if it is was only an impression of probability, their duty was to acquit.[89]

Debate continues in academic circles as to the extent, if any, to which mathematical probability may be useful in quantifying the concept of beyond reasonable doubt.[90] An interesting study in the 1970s conducted by sociologists Simon and Mahan[91] asked judges, jurors and sociology students about the level of probability generally they would require to be satisfied beyond reasonable doubt that a crime had been committed. Of judges, 4% would be satisfied at a probability level below .7, 33% satisfied at a probability level between .7 and .9, while 63% would require a probability level above .9. Some difference in understanding of the phrase ‘beyond reasonable doubt’ is evident when comparing these figures with jurors’ required levels of probability to meet the test.[92] About 26% would be satisfied with a probability level below .7, 20% satisfied with a probability level between .7 and .9, and 54% satisfied with a probability level above .9.[93] Some differences in the required level of probability according to the seriousness of the crime are also evident.[94]

The mismatch between judges’ understanding of beyond reasonable doubt and juries’ understanding of beyond reasonable doubt is a serious and interesting finding of this work that warrants attention by judges and further study, but the difference is not immediately relevant to this paper. What is considered relevant is the high degree of probability considered to be required to support a beyond reasonable doubt finding of guilt.

Some writers have been prepared to name a probability level, or indicative level, at which a jury might be convinced beyond a reasonable doubt that a person had committed an offence, albeit in different contexts. Nesson claimed that even if a jury knew there was a 24 chance out of 25 (96%) certainty that the offender was guilty, ‘there is no way, on this evidence, that a jury could form an abiding conviction that the defendant was guilty’.[95] Tribe gives us an example of a Swedish case involving a parking ticket, where the driver was accused of parking in a one-hour zone for too long. The driver claimed he had moved the vehicle in the time it was there, so had not committed an infringement. To prove it had not been moved, the government called an officer to testify that he recorded the positions of the tyre air-valves on one side of the car. Both before and after a period in excess of one hour, each of the wheel valves was pointing a particular way (eg one o’clock, eight o’clock). The driver insisted he had driven the car in the meantime, and upon his return it just so happened that the wheel valves were in the same place as they were the first time he had parked there. The probability that such an event took place was estimated at one chance in 144. This was enough to create reasonable doubt in that case.[96]

Probability levels were specifically referred to in the drink driving case of Samuels v Flavel.[97] There the defendant had a blood alcohol level of .18%. The offence was being in charge of a motor vehicle while incapable of controlling the vehicle. Police observed the defendant to be steady on his feet and neat in appearance, and answered questions rationally and politely. At the admitted level of .18% blood alcohol level, experts testified that 80% of the population would show clinical impairment of the kind referred to in the offence, while 20% would not. This was viewed by the Court in that case as reasonable doubt and the defendant was acquitted.[98]

The conclusion to this Part is that the meaning of reasonable doubt has been phrased in different ways, some with recourse to mathematical probabilities and some without. However, whether one embraces non-mathematical or mathematical formulae for determining reasonable doubt, the standard of proof required by the Queensland preventive detention law falls substantially short of the traditional criminal standard required. The standard prescribed in the Queensland law is not at the level of “moral certainty”. It does not equate to the 99% level of confidence in guilt, to which mention has been made.[99]

If a level of proof beyond reasonable doubt is required for original incarceration, for good reason, why is a demonstrably substantially less standard acceptable for continued incarceration in Australia? It is suggested to be illogical for substantially different standards of proof for original and continued incarceration. Analogously, the United States Supreme Court in Apprendi v New Jersey[100] concluded that “other than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt .. It is unconstitutional for a legislature to remove from the jury the assessment of facts that increase the prescribed range of penalties to which a criminal defendant is exposed. It is equally clear that such facts must be established by proof beyond a reasonable doubt”.[101] They have also insisted that, in the context of preventive detention laws, suggestions that the offender would be likely to be dangerous if released must be proven to the court beyond reasonable doubt. The civil standard has not been acceptable to that Court in the preventive detention context.[102]

It is suggested that if preventive detention laws are to be valid in Australia, at the very least the courts should require, as the United States Supreme Court has required, that the suggestion of dangerousness after release be proven at the higher criminal standard of proof. This can be rationalized as either an illustration of due process, or a step towards a coherent legal system, where both the original decision to incarcerate, and the subsequent decision to keep in prison, are based on the same standard of proof. The alternative Apprendi argument is that since a finding of likely future dangerousness increases jail time for the offender, that finding itself should be subject to proof beyond a reasonable doubt, before a jury if the offender so wishes.

Clearly, the Fardon law does not provide that the likelihood that the offender would re-offend if released be shown at the beyond reasonable doubt level. At least some members of the High Court draw links between the preventive detention order and the previous offence, and certainly the offender is only eligible for a preventive detention order if he/she has committed a past offence. Consistent with the Supreme Court’s view in Apprendi, it is submitted that since a finding that a prisoner is likely to re-offend increases jail time, due process requires that fact should have to be proven beyond a reasonable doubt, to satisfactorily balance the rights of the community and the rights of the offender. Consistent with Hendricks, findings of dangerousness must be proven beyond reasonable doubt.

V OTHER ARGUMENTS AGAINST THE FARDON LAW

A The Proceedings are Criminal in Nature so the Criminal Standard Must Apply

Interestingly, the Queensland Fardon law does not identify itself as involving proceedings that are either civil or criminal in nature. This compares with the Kable legislation, which identified itself as involving a civil proceeding. Given the lower than reasonable doubt standard outlined in the Fardon legislation, however, one might assume that the proceedings are intended to be civil in nature.[103] Certainly, civil commitment is not unknown to the Australian legal system, though it has been allowed in only strictly limited circumstances.[104]

Let us proceed firstly on the assumption that the proceedings are civil in nature. What are the implications of such a finding? Interestingly, the United States Supreme Court considered this issue recently in Kansas v Crane.[105] It referred to its previous preventive detention decision in Hendricks as emphasizing ‘the importance of distinguishing a dangerous sexual offender subject to civil commitment from other dangerous persons who are perhaps more properly dealt with exclusively through criminal proceedings’. The Court reiterated in Crane that ‘that distinction is necessary lest ‘civil commitment’ become a mechanism for retribution or general deterrence – functions properly those of the criminal law, not civil commitment’. The Court in Crane required that in order for a preventive detention law to be valid,[106] the person against whom it was applied must be shown to have a special and serious lack of ability to control their behaviour.[107] The implication was that a person not within this category should be dealt with according to criminal process.

Consistently with the above requirement, it was important for the majority of the Court, in determining that the proceedings were civil in nature, that the aim of the statute was to provide treatment, not punishment. The offender would be committed in a facility set aside to provide psychiatric care, and segregated from the general prison population. Again, the implication was that had the aim of the Act been other than to provide treatment, or if the offender would be included within the general prison population, the proceedings might not be thought of as civil.[108]

The dissenting judges, in finding that the preventive legislation in Hendricks was in fact criminal in nature, noted that

one would expect a nonpunitive statutory scheme to confine, not simply in order to protect, but also in order to cure ... one would expect a nonpunitively motivated legislature that confines because of a dangerous mental abnormality to seek to help the individual himself to overcome that abnormality ... conversely a statutory scheme that provides confinement that does not reasonably fit a practically available, medically oriented treatment objective, more likely reflects a primarily punitive legislative purpose[109]

Taking the Fardon law on its face, it seems to be aimed more at community protection rather than treatment.[110] This is the stated aim of the law, and the paramount consideration for judges to bear in mind in cases of doubt. Substantively, the Fardon legislation prescribes no treatment for any person detained under it. There is nothing to suggest that a person about whom an application is successful will be segregated from the remainder of the prison population. As stated, these were necessary requirements in order for the United States Supreme Court to find the commitment to be civil in nature. They do not appear in the Fardon law.

If the proceedings are not in fact classified as civil in nature, they obviously must be criminal proceedings. The fact that the legislation prescribes something less than the criminal standard of proof as a basis for incarceration in criminal proceedings would then be likely to render it invalid.[111]

B Focus on Sexual Offences

While the author does not deny that sexual offences are very serious, one wonders at the reasoning behind confining the preventive detention regime only to those kinds of offenders. What of those convicted of manslaughter or armed robbery?[112] No evidence was presented when the law was introduced to justify why the law was targeted only at one category of offenders. The author has not seen any evidence suggesting that the recidivism rate for convicted sexual offenders is significantly higher than for other serious offences, which evidence might have provided some explanation or justification. Arguments based on community protection are not convincing – it is hard to say that preventive detention laws for sex offenders are justified on community protection grounds, while preventive detention laws for those convicted of manslaughter or armed robbery are not.

C Better that the Future Guilty Go Free than the Future Innocent Be (Further) Incarcerated

One of the most fundamental tenets of our criminal law is the notion that it is better that someone who is guilty goes unpunished, than for an innocent person to be wrongly incarcerated.[113] The idea has biblical overtones, and was embraced by many great philosophers including Blackstone and Aristotle, [114] as well as by the judiciary.[115] This reflects the very serious regard with which an individual’s liberty is held by our legal system, and is clearly consistent with other well-established rules of criminal procedure, including the presumption of innocence, and the right to due process.

By analogy in regards to preventive detention, surely the equivalent argument is that it is better that a person who will re-offend be released at the end of his time (a future “guilty” if you will), than to keep in prison a person who will not re-offend once released? (a future “innocent”). If this is not the equivalent argument, why not? Has the offender, by virtue of the fact that he/she in the past has committed a crime, given up fundamental criminal procedural rights, including the presumption of innocence and due process? If so, that would surely be a very dangerous assertion.[116]

VI ALTERNATIVES TO PREVENTIVE DETENTION IN AUSTRALIA AS CURRENTLY PRESCRIBED

A detailed discussion of alternatives to the current preventive detention scheme laid out in the Queensland law would be beyond the scope of this paper. However, it is hoped that in response to evidence that some of its inmates are not being rehabilitated by the corrective services system as it currently operates,[117] the Government is reviewing those programs.[118] It has already been noted that one of the unsatisfactory features of the Queensland preventive detention law is that it does not provide for any specific treatment of offenders detained under its provisions. While earlier it was argued this may mean the legislation is defacto criminal in nature rather than civil,[119] it is suggested that the Queensland Government should amend its legislation to provide for special, specific treatment for those detained under the law, as should any State considering preventive detention regimes in the light of the High Court’s view in Fardon. It is further suggested those prisoners detained under the law be segregated from inmates. Unless those State Governments implementing such regimes intend to make applications in relation to offenders every six months for the rest of the offenders’ lives, at some point the offenders will be released. At that point, we all hope that they will not re-offend, and Governments need to do what they can to make sure this does not happen. As a matter of public policy rather than law, keeping them in prison for a while longer is at best a band-aid solution to a serious problem

However, no-one pretends that it is always easy to rehabilitate offenders. The author would not have a difficulty with the increased use of indeterminate life sentences for the most serious sexual offenders, providing for later review of the offender’s progression and possible release once it is determined he/she is not a risk to society.[120] It should also be acknowledged that threats of further criminal behaviour may themselves amount to a further crime (usually assault), and may be dealt with under the criminal law. Perhaps there is an argument that the traditional definition of insanity is too narrow, and needs to be broadened to include further categories of offenders.[121]

VII CONCLUSION

The author believes that the High Court’s decision in Fardon fails to apply its own precedent set only 10 years ago in Kable, and offends the separation of powers principle, an important protection for human rights in the absence of an express bill of rights. The legislation assumes it is possible, with an acceptable degree of accuracy, to predict whether or not an offender would be dangerous if released. However, this is far from a settled issue.

Canvassing the vast literature on this issue, the author finds that while modern models have promise, models continue to vary wildly in their predictive accuracy. The legislation’s required standard of ‘high degree of probability’ certainly does not equate to the criminal standard, regardless of which approach is taken to explaining that standard. This is a problem, because it is inherently asymmetrical to require a different standard of proof for original and subsequent incarceration, it is inconsistent with the United States Supreme Court and other courts’ views on the boundaries of acceptable preventive detention, and preventive detention law like the Fardon version, which requires no particular treatment of the person detained and which includes them in the general prison population, may well be in substance a criminal rather than civil commitment.

Preventive detention law is an extreme response to a difficult problem, when there are alternatives less intrusive to human rights that should be considered.[122] Preventive detention laws are bad public policy, but more importantly for our purposes (and, contrary to Gleeson CJ’s view, as a non-sequitur),[123] should be struck down as unconstitutional. At the very least, the High Court should insist upon either a beyond reasonable doubt test on the allegation of future dangerousness, or specific evidence of treatment for those detained under the law and their segregation from those being imprisoned, to validate this kind of legislation.


[*] Acting Head of Department and Lecturer in Law, Faculty of Business, University of Southern Queensland.

[1] Fardon v Attorney-General (Queensland) [2004] HCA 46; [2004] 210 ALR 50 (Fardon).

[2] Dangerous Prisoners (Sexual Offenders) Act 2003, s13.

[3] Detaining Future Dangerous Offenders: Dangerous Legislation, 9 DEAKIN L. REV. 243 (2004).

[4] Kable v DPP (NSW) [1996] HCA 24; (1996) 189 CLR 51 (Kable).

[5] It was entitled the Community Protection Act 1994 (NSW), and s3 of the legislation stated its object was to protect the community, and that in interpreting the legislation, the need to protect the community was to be given paramount consideration.

[6] An orthodox view – see for example Clyne v East (1967) 68 SR (NSW) 385, Builders’ Labourers’ Federation of New South Wales v Minister for Industrial Relations (1986) 7 NSWLR 372.

[7] See also Peter Johnston and Rohan Hardcastle, State Courts: The Limits of Kable, [1998] SydLawRw 10; 20 SYDNEY L. REV. 216 (1998)

[8] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 98; considered in Chu Kheng Lim v. Minister for Immigration, (1992) 176 CLR 1.

[9] Grollo v. Palmer, [1995] HCA 26; (1995) 184 CLR 348,365.

[10] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 98.

[11] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 103.

[12] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 106-107.

[13] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 107.

[14] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 108. Another ground of objection to this kind of law is the principle that electorally accountable branches of government should not be allowed to avoid political responsibility by seeking to delegate unpalatable functions to the judiciary. As the United States Supreme Court found in Mistretta v United States [1989] USSC 9; (1989) 488 US 361, ‘The judicial branch’s reputation for impartiality and nonpartisanship may not be borrowed by the political branches to cloak their work in the neutral colours of judicial action’. Elizabeth Handsley submits this is a preferable basis for the Kable principle than the undermining confidence line: Public Confidence in the Judiciary: A Red Herring for the Separation of Judicial Power[1998] SydLawRw 9; , 20 SYDNEY L. REV. 183, 201 (1998)

[15] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 98, 116.

[16] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 122.

[17] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 133.

[18] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 134.

[19] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 134.

[20] Kable v. DPP, (NSW) [1996] HCA 24; (1996) 189 CLR 51, 132.

[21] [1988] HCA 55; (1988) 166 CLR 1, 10: “Whether the exercise of that legislative power is subject to some restraints by reference to rights deeply rooted in our democratic system of government and the common law ... is another question which we need not explore”.

[22] As to which, see Martin Krygier, The Grammar of Colonial Legality: Subjects, Objects and the Rule of Law in AUSTRALIA RESHAPED: ESSAYS ON 200 YEARS OF INSTITUTIONAL TRANSFORMATION 220-260 (Geoffrey Brennan and Francis Castles ed , 2002), Martin Krygier, The Rule of Law, in INTERNATIONAL ENCYCLOPAEDIA OF THE SOCIAL AND BEHAVIOURAL SCIENCES vol 20 13403-408 (Neil Smelser and Paul Bates ed., 2001)

[23] This suggestion arose in the later decisions of Baker and Fardon, where Kirby J noted in the former case that (30) “the imposition of punishment, or added punishment, by the operation of a new law having retroactive effect is not only contrary to our legal tradition and offensive to its basic principles. It is also incompatible with the fundamental rules of universal human rights forbidding retroactive criminal punishment”; and Fardon (100) [FULL REFERENCE REQUIRED]. Refer also to George Zdenkowski Community Protection Through Imprisonment Without Conviction: Pragmatism Versus Justice AUST J H R [FULL REFERENCE REQUIRED]

[24] Schedule Dictionary, regardless whether the offence was committed in Queensland or not.

[25] s3 and s13(6) respectively

[26] However, both fell short of the orthodox criminal standard of beyond reasonable doubt. This point will be discussed in more detail in Part III of the article.

[27] Gaudron J noted this of the New South Wales law.

[28] Attorney-General (Queensland) v Fardon [2003] QCA 416, [91].

[29] Fardon supra n1, 56 [FULL CASE REFERENCE REQUIRED].

[30] Gleeson CJ quickly dismissed what might be thought to be legitimate civil libertarian concerns with legislation of the kind being considered: “There are important issues that could be raised about the legislative policy of continuing detention of offenders who have served their terms of imprisonment, and who are regarded as a danger to the community when released. Substantial questions of civil liberty arise. This case, however, is not concerned with these wider issues.” (52-53). One wonders why this case is not concerned with these wider issues – clearly the Act has great potential to derogate from perhaps the most fundamental civil liberty of all, the right to liberty. Will asking a court to participate in such a derogation of established civil liberties (namely ordering imprisonment without guilt) ever compromise the integrity and independence of the court, and undermine public confidence in it, according to the Chief Justice? The proposition that the separation of powers principle, to which Kable clearly relates, is integrally associated with civil liberties is unremarkable: Mistretta v United States [1989] USSC 9; (1989) 488 US 361, 380 “it was the central judgment of the framers of the Constitution that, within our political scheme, the separation of governmental powers into three co-ordinate branches is essential to the preservation of liberty’. One wonders how it is that the Australian Chief Justice can claim that ‘this case is not concerned with these wider issues”.

[31] Gleeson CJ raised two further justifications in his judgment, that (a) it cannot be a serious objection to the validity of the Act that the law which the Supreme Court is required to administer relates to a subject that is, or may be, politically divisive (57) [FULL CASE REFERENCE REQUIRED]; and (b) nothing would be more likely to damage public confidence in the integrity and impartiality of courts than judicial refusal to implement the provisions of a statute on the ground of an objection to legislative policy (58) [FULL CASE REFERENCE REQUIRED]. With respect, the author does not see the Kable principle as resting on either of these principles, so does not see the relevance of these points to resolution of this case. The challenge to the validity of the Act is based on arguments as to separation of powers and the independence of the judiciary; it is not about whether or not a government course of action is either politically desirable or sound from a public policy point of view. Higher and more important principles are at issue.

[32] 60 [FULL CASE REFERENCE REQUIRED]

[33] The United States Supreme Court has validated preventive detention laws, but only those requiring that a trial be held where the court would determine beyond a reasonable doubt whether the individual was likely to re-offend: Kansas v. Hendricks, [1997] USSC 63; (1997) 521 US 346.

[34] Section 17(1)(a) of the Community Protection Act 1994 (NSW) stated that the court was bound by the rules of evidence.

[35] Callinan and Heydon JJ agreed that the purpose of the detention here was for “community protection and not punishment” (109). Compare the view of the Supreme Court in United States v. Brown, [1965] USSC 134; (1965) 381 US 437, 458: “One of the reasons society imprisons those convicted of crimes is to keep them from inflicting future harm, but that does not make imprisonment any the less punishment”.

[36] Given that both Acts prescribe that community protection is the paramount consideration in assessing the application, it is submitted that judges will tend to err on the side of granting the further detention order sought by the Government. Further, no judge would wish for the situation where they refused the application, resulting in the offender being released back into the community, only to commit more offences. Some members of the community may blame the judge who refused the application.

[37] 77 [FULL CASE REFERENCE REQUIRED]

[38] although this was also provided for in the Community Protection Act 1994 (NSW).

[39] Id. at 78 [FULL CASE REFERENCE REQUIRED]

[40] Id. at 75. [FULL CASE REFERENCE REQUIRED] Hayne J had similar reservations – “I acknowledge the evident force in the proposition that to confine a person for what he or she might do, rather than what he or she has done, is at odds with identifying the central constitutional conception of detention as a consequence of judicial determination of engagement in past conduct.” (104)

[41] 80. [FULL CASE REFERENCE REQUIRED] Such a finding appears inconsistent with the recent finding of a majority of the United States Supreme Court in Apprendi v. New Jersey, [2000] USSC 57; (2000) 530 US 466 that “other than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt .. It is unconstitutional for a legislature to remove from the jury the assessment of facts that increase the prescribed range of penalties to which a criminal defendant is exposed. It is equally clear that such facts must be established by proof beyond a reasonable doubt” (Justice Stevens for the majority). Clearly, the likelihood that the offender would re-offend if released has not been proven at the beyond reasonable doubt level.

[42] 80 [FULL CASE REFERENCE REQUIRED]; “what is of present significance is provisions of Part 3 headed ‘Annual Reviews” ... less than (indefinite detention) is mandated by the Act (80). However, further detention under the impugned Community Protection Act 1995 (NSW) was also for a limited time, so it is difficult to see the basis of the distinction.

[43] Id. at 88.

[44] Id. at 95

[45] Kirby J also claimed the Act removed the fundamental right of presumption of innocence (94).

[46] 95. Indeed, Kirby J compared this legislation with laws passed by the Nazi Government in Germany in the 1930s. Punishment was addressed to the estimated character of the criminal, rather than the proven facts of a crime. The attention of the courts was diverted from allegedly wrongful past acts to a preoccupation with a ‘pictorial impression’ of the accused. Provision was made for punishment, or added punishment, based not on specified acts proven, but an inclination towards criminality such that the offender could not become a useful member of the community. A practice then developed whereby prisoners were not released from custody at the expiration of their sentences. Political prisoners and undesirables became increasingly subject to indeterminate detention (101-102). German preventive detention laws continue to this day. See for example Frieder Dunkel and Dirk Jan Zyl Smit Preventive Detention of Dangerous Offenders Re-Examined: A Comment on Two Decisions of the German Federal Constitutional Court and the Federal Draft Bill on Preventive Detention of 9 March 2004 5 GERMAN L J 619 (2004)

[47] For example, George Winterton The Separation of Judicial Power as an Implied Bill of Rights in Geoffrey Lindell ed FUTURE DIRECTIONS IN AUSTRALIAN CONSTITUTIONAL LAW (1994) “dividing government power is the oldest choice for restraining it, and thereby protecting liberty” (185), the United States Supreme Court in Mistretta v United States [1989] USSC 9; (1989) 488 US 361,380 “it was the central judgment of the Framers of the Constitution that, within our political scheme, the separation of governmental powers into three co-ordinate branches is essential to the preservation of libert” (emphasis added), Elizabeth Handsley Public Confidence in the Judiciary: A Red Herring for the Separation of Judicial Power [1998] SydLawRw 9; 20 Syd L R 183, 187 (1998) “an independent judiciary is widely regarded as essential for the protection of civil liberties, minority rights and/or the rule of law”; Sir Anthony Mason A New Perspective on Separation of Powers 82 CANBERRA B P A 1,2 (1996), and Mirko Bagaric and Tanya Lakic, Victorian Sentencing Turns Retrospective: The Constitutional Validity of Retrospective Criminal Legislation After Kable 23 CRIMINAL L J 145,149 (1999)

[48] Baker v R [2004] HCA 45; (2004) 210 ALR 1,17. Alternatively, it may be that a temporary muzzle has been placed on the watch-dog.

[49] Refer also Stephen J in Veen v The Queen [1979] HCA 7; (1979) 143 CLR 458, 464: “No doubt the whole question of prediction of behaviour in the future is a most difficult one. Its very difficulty is in itself a potent reason against undue weight being given to the protection of the community from what is predicted as the likely future violence of the convicted person. Predictions as to future violence, even when based upon extensive clinical investigations by teams of experienced psychiatrists, have recently been condemned as prone to very significant degress of error when matched against actuality”, McGarry v R [2001] HCA 62; (2001) 207 CLR 121, 141-142 (Kirby J), and McMurdo P of the Queensland Court of Appeal noted in the Fardon case that a prediction of dangerousness was “notoriously unreliable and ... must be based largely on opinions of psychiatrists” (supra n28, [91]). Of the majority in Fardon, Gleeson CJ opined that “no doubt predictions of future danger may be unreliable but ... they may also be right” (supra n1, 55). The writer is not convinced that concerns about the unpredictability can be dealt with so in such a glib fashion.

[50] 83; citing Kate Warner Sentencing Review 2002-2003 27 CRIMINAL L J 325, 338 (2003); to similar effect Barbara Underwood in Law and the Crystal Ball: Predicting Behaviour with Statistical Inference and Individualized Judgment 88 YALE L.J. 1408, 1410 (1979) “Existing techniques for parole prediction are highly inaccurate by any measure’, 1413 ‘no available predictive method is sufficiently accurate to satisfy the high degree of accuracy appropriate for the decision to incarcerate” (though acknowledging it may be possible to develop one). A German study concluded that between 60-90% of preventive detainees were ‘false positives’, in other words inappropriately diagnosed as being dangerous: Kinzig NJW 1457 (2001). A meta-study of 23 different prediction models found an average success rate of 58%: Alec Buchanan and Morven Leese Detention of People with Dangerous Severe Personality Disorders: A Systematic Review 358 Lancet 1955,1957 (2001). Stephen Morse concluded that ‘the ability of mental health professionals to predict future violence among mental patients may be better than chance, but it is still highly inaccurate’: Blame and Danger: An Essay on Preventive Detention 76 B.U.L REV 113,126 (1996); Radzinowicz and Hood reached a similar conclusion A Dangerous Direction in Sentencing Reform CRIM L R 713 (1981), Brown M Serious Violence and Dilemmas of Sentencing: A Comparison of Three Incapacitation Policies CRIMINAL L R 710 (1998).

Michael Petrunik Models of Dangerousness: A Cross Jurisdictional Review of Dangerousness Legislation and Practice 1994-2002 (paper prepared for the Ministry of the Solicitor-General of Canada)(www.psepc-sppcc.gc.ca/publications/corrections/199402_e.asp, accessed 12/1/05) “research based largely on samples of involuntary mental patients indicates that false positive predictions of violence exceed true positive predictions at a rate of approximately 2 to 1”, John Monahan found that clinical predictions were accurate in only one in three cases (The Prediction of Violent Behaviour: Toward a Second Generation of Theory and Policy 10,41 AMER J of PSYCH 10 (1984)). Note the comments of the Supreme Court in Schall v Martin [1984] USSC 120; (1984) 467 US 253, 278 “from a legal point of view there is nothing inherently unattainable about a prediction of future criminal conduct”, but cf Burger CJ of the United States Supreme Court in Addington v Texas [1979] USSC 77; (1979) 441 US 418,429 “there is a serious question as to whether the state could ever prove beyond a reasonable doubt that an individual is both mentally ill and likely to be dangerous”.

[51] There remains a lively debate as to whether a clinical or statistical method should be used in making the assessment (usually separately, though some models use elements of both). A clinical prediction occurs when a decisionmaker attempts to assess offenders as individuals, taking into account particular individual characteristics. Statistical techniques are used to generate presumptions and probabilities based on prior cases and experience: Barbara Underwood, supra n49, 1420-1424

[52] C D Webster, Kevin Douglas, D Eaves and S D Hart, from the Mental Health, Law and Policy Institute, Simon Fraser University, British Columbia.

[53] Clinical methods of predicting future behaviour are often credited to Michael Foucault About the Concept of the Dangerous Individual in 19th Century Legal Psychiatry 1 INTERNATIONAL J LAW PSYCH 1 (1978) but have now largely been discounted as an effective means (at least of itself) in predicting future behaviour: Richard Wettstein Psychiatric Perspective on Washington’s Sexually Violent Predators Statute 15 UNI PUG SD L R 597 (1992). Some researchers have found statistical data to be more reliable than clinical: Quinsey, L, Rice, M E, Harris, G T and Cormier, C A VIOLENT OFFENDERS: APPRAISING AND MANAGING THE RISK (1998), Milner, J S and Campbell, J C Prediction Issues for Practitioners in Campbell, J C ed ASSESSING DANGEROUSNESS: VIOLENCE BY SEXUAL OFFENDERS, BATTERERS AND CHILD ABUSERS (1995)

[54] Ross, D J, Stephen Hart and Webster, C D Facts and Fates: Testing the HCR-20 against aggressive behaviour in hospital and community (unpublished manuscript), referred to in Kevin Douglas et al Assessing Risk for Violence Among Psychiatric Patients: The HCR-20 Violence Risk Assessment Scheme and the Psychopathy Checklist: Screening Version 67 JOURNAL CONSULTING CLINICAL PSYCHOLOGY 917-930 (1999)

[55] Strand, S, Belfrage, H, Fransson G and Levander, S Clinical and Risk Management Factors in Risk Prediction of Mentally Disordered Offenders: More Important than Actuarial Data? 4 LEGAL CRIMINOLOGICAL PSYCH 67-76

[56] Kevin Douglas et al Psychometric Properties of HCR-20 Violence Risk Assessment Scheme in Insanity Acquitees (paper presented at the 106th Annual Convention of the American Psychological Association, San Francisco).

[57] Kevin Douglas et al Assessing Risk for Violence Among Psychiatric Patients: The HCR-20 Violence Risk Assessment Scheme and the Psychopathy Checklist: Screening Version 67 JOURNAL CONSULTING CLINICAL PSYCHOLOGY 917, 923 (1999), although this problem may be related to the confidence of the risk judgment made: Kevin Douglas and James Ogloff The Impact of Confidence on the Accuracy of Structured Professional and Actuarial Violence Risk Judgments in a Sample of Forensic Psychiatric Patients 27 LAW and HUMAN BEHAV 573 (2003). The HCR-20 was found to be a reasonably accurate predictor of violence in Kevin Douglas, James Ogloff and Stephen Hart Evaluation of a Model of Violence Risk Assessment Among Forensic Psychiatric Patients 54 PSYCH SERVICES 1372-1379 (2003).

[58] Kevin Douglas and James Ogloff Multiple Facets of Risk for Violence: The Impact of Judgmental Specificity on Structured Decisions About Violence Risk 2 INTERNATIONAL J FORENSIC MENTAL HEALTH p19-34 (2003).

[59] Stephen D Hart, D Cox and R D Hare The Hare Psychopathy Checklist: Screening Version, Toronto

[60] including (a) superficiality (b) grandiosity (c) deceitful (d) lacking in remorse (e) lacking in empathy (f) refusal to take responsibility.

[61] including (a) impulsivity (b) poor behaviour controls (c) lacking in goals (d) irresponsibility (e) adolescent anti-social behaviour (f) adult anti-social behaviour.

[62] n44

[63] Kevin Douglas and Webster, C D Predicting Violence in Mentally and Personality Disordered Individuals in James Ogloff ed PSYCHOLOGY AND LAW: THE STATE OF THE DISCIPLINE p175-239. Others view risk factors as falling into two categories, static risk and dynamic risk factors. Static risk factors are generally historical markers that cannot be changed, such as criminal history and age at first offence. Dynamic risk factors are potentially changeable aspects of the individual, such as accommodation, employment, education and anti-social attitudes. Both sets of factors are said to have predictive validity: Gendreau, P, Little, T and Goggin, C A meta analysis of the predictors of adult offender recidivism: what works (1996) 34 CRIMINOLOGY 575-607

[64] Klassen and O’Connor Assessing the Risk of Violence in Abused Mental Patients: A Cross- Validation Study: PSYCHOLOGICAL ASSESSMENT: JOURNAL CONSULTING CLINICAL PSYCHOLOGY 1, 75-81

[65] McNiel, D E and Binder, R L Screening for Risk of Inpatient Violence: Validation of an Actuarial Tool LAW AND HUMAN BEHAVIOUR 18, 586.

[66] Andrews, D A and Bonta, J THE PSYCHOLOGY OF CRIMINAL CONDUCT (2003) 3rd ed

[67] Criminogenic factors include pro-criminal attitudes, criminal associates, substance abuse, anti-social personality, problem-solving skills, and hostility/anger; non-criminogenic factors include self-esteem, anxiety, feelings of alienation, psychological discomfort, group cohesion and neighbourhood improvement.

[68] Affected by factors such as intellectual functioning, self-esteem, motivation levels, staff characteristics, therapeutic relationships, environmental support, program content and delivery.

[69] James Ogloff and Michael Davis Advances in Offender Assessment and Rehabilitation: Contributions of the Risk-Needs-Responsivity Approach PSYCHOLOGY, CRIME AND LAW 229-242 (2004). Many of these studies involved patients detained in mental institutions rather than prisons; it is not thought that the degree of predictability of future behaviour would differ markedly according to whether the person has a history of mental illness or not, given most of these studies consider the mental illness in assessing likely recidivist behaviour. However, further research would be required to validate this assertion. Another model, the Sex Offender Risk Appraisal Guide, claims a prediction score of 62%: Rice, M E and Harris, G T Cross Validation and Extension of the Violence Risk Appraisal Guide for Child Molesters and Rapists (1997) 21 LAW AND HUMAN BEHAVIOR 231.

[70] Alec Buchanan and Morven Lesse Detention of People With Dangerous Severe Personality Disorders: A Systematic Review 358 THE LANCET 1955,1957 (2001); another summary of the various studies and their predictive value is found in Hazel Kemshall RISK ASSESSMENT AND MANAGEMENT OF SERIOUS VIOLENT AND SEXUAL OFFENDERS: A REVIEW OF CURRENT ISSUES (2002), a report commissioned by the Scottish Executive.

[71] Supra n1, 83

[72] Perhaps an even worse example was the legislation challenged unsuccessfully in Baker v R, supra n48. The law there allowed those sentenced to life imprisonment to apply to the Supreme Court for an adjustment to be made to their sentence, so that the offender would be ordered to serve a definite jail term, after which time they would be released. This law was expressed not to apply to a category of offenders, to which Baker belonged. This category was those about whom a recommendation that the prisoner never be released had been made at the time of sentence. At the time those recommendations were made, they had no legal impact. This legislation was upheld as a valid law by the High Court although the trial judge had found that the statistical risk of recidivism of Baker was extremely low and his chances of rehabilitation very high. He concluded the offender’s conduct while in prison had been exemplary, with no suggestion of any conduct which would pose a risk to the community. Despite this, legislation which denied Baker the opportunity to commute his sentence to a definite one, while allowing other life prisoners to bring such an application (including one triple murderer, because he was not subject to a non-release recommendation), was declared to be valid.

[73] Supra n1, 55

[74] In considering the validity of preventive detention regimes in that country, the United States Supreme Court has only validated legislation which allows the question of the likelihood that the offender will re-offend to be determined by a judge or jury beyond reasonable doubt. No lesser standard, such as the one in the Queensland legislation, has been allowed: Kansas v Hendricks [1997] USSC 63; (1997) 521 US 346. Most recently, the Court has required that there also be proof of serious difficulty in controlling behaviour: Kansas v Crane [2002] USSC 10; (2002) 534 US 407.

Recent examples of the High Court applying the reasonable doubt test to a criminal matter include RPS v The Queen [2000] HCA 3; (2000) 199 CLR 620, 630 (Gaudron ACJ Gummow Kirby and Hayne JJ) and Azzopardi v The Queen, Davis v The Queen [2001] HCA 25; (2001) 205 CLR 50, 64 (Gaudron Gummow Kirby and Hayne JJ)

[75] It is worth reiterating here that Kirby J in Fardon referred to literature in the area maintaining that predictions of future dangerousness had only a one-third to 50% success rate, so a claimed success rate of 70-80% in predicting future dangerousness is the highest that can currently be put, but this figure is by no means consistently obtained, or universally claimed. The author has taken the highest practical success rate as a benchmark for argument.

[76] Woolmington v The Director of Public Prosecutions [1935] AC 462, 481 (Lord Chancellor Viscount Sankey)

[77] For example, Isaacs and Powers JJ in Brown v The King [1913] HCA 70; (1913) 17 CLR 570,594: “The words ‘reasonable doubt’ are in themselves so far self-explanatory that no further explanation is considered strictly necessary. Usually attempts to elucidate them do not add to their clearness” (Barton ACJ to like effect (584); Thomas v The Queen (1960) 102 CLR ,595 per Kitto J: “Whether a doubt is reasonable is for the jury to say; and the danger that invests an attempt to explain what ‘reasonable’ means is that the attempt not only may prove unhelpful but may obscure the vital point that the accused must be given the benefit of any doubt which the jury considers reasonable”, and Windeyer J (604-605) “attempts to paraphrase and embellishment to explain to juries what is meant by satisfaction beyond reasonable doubt are not always helpful”; and Dawson v The Queen [1961] HCA 74; (1961) 106 CLR 1, 18 per Dixon CJ: “it is a mistake to depart from the time honoured formula. It is .. used by ordinary people and is understood well enough by the average man in the community. The attempts to substitute other expressions, of which there have been many examples not only here but in England, have never prospered”.

[78] Green v The Queen (1971) 126 CLR ,33 (Green)(Barwick CJ McTiernan and Owen JJ), La Fontaine v R [1976] HCA 52; (1976) 11 ALR 507, 514 (Barwick CJ); cf Latham CJ in Burrows v The King [1937] HCA 56; (1937) 58 CLR 249,256 “doubt such as would be entertained by reasonable men”

[79] Green, 33 and similarly the United States Supreme Court in Victor v Nebraska [1994] USSC 15; (1994) 511 US 1,6 “near certitude of the guilt of the accused” (Justice O’Connor for the majority)

[80] Brown v The King supra n77, 585 (Barton ACJ) and 595 (Isaacs and Powers JJ), Commonwealth v Costley (1875) 118 Mass 1,24 approved by the United States Supreme Court in Fidelity Mutual Life Association v Mettler [1902] USSC 114; (1902) 185 US 308 and Miles v United States [1880] USSC 118; (1881) 103 US 304. The definition of reasonable doubt on law.com is similar, defining “reasonable doubt” as “not being sure of a criminal defendant’s guilt to a moral certainty”(<www.law.com>, accessed 4 January 2005); similarly <www.findlaw.com>, accessed 4 January 2005, defines reasonable doubt as being when a factfinder cannot say with moral certainty that a person is guilty or a particular fact exists. It is trite to point out that a high standard of proof in criminal matters is justified by the serious consequences for the accused of a wrongful conviction.

[81] Hopt v Utah 120 US at 439, Victor v Nebraska [1994] USSC 15; (1994) 511 US 1,6 (Justice O’Connor for the majority)

[82] Holland v United States 348 US at 140.

[83] Martin v Osborne [1936] HCA 23; (1936) 55 CLR 367,375. For a comprehensive discussion of the Australian authorities, refer to Frank Bates Describing the Indescribable – Evaluating the Standard of Proof in Criminal Cases 13 CRIM L J 330 (1989)

[84] This observation was surprisingly made by only a few of the judges hearing the preventive detention appeals: McHugh J in Kable, supra n1, (120): “proof of the case against the appellant was made much easier by removing any need to prove beyond reasonable doubt that he was likely to commit a serious act of violence”. McMurdo P in Fardon, supra n28, in the Queensland Court of Appeal also noted the standard of proof was not at the level of beyond reasonable doubt.

[85] In Malec v J C Hutton Pty Ltd (1990) 169 CLR 638, 642-643 Deane Gaudron and McHugh JJ noted “A common law court determines on the balance of probabilities whether an event has occurred. If the probability of the event having occurred is greater than it not having occurred, the occurrence of the event is treated as certain”. To like effect was Lord Simon in Davies v Taylor [1974] AC 207,219 “Beneath the legal concept of probability lies the mathematical theory of probability. Only occasionally does this break surface – apart from the concept of proof on a balance of probabilities, which can be restated as the burden of showing odds of at least 51 to 49 that such-and-such has taken place or will do so”. These comments were obviously raised in the context of a civil case, and it is an even more controversial issue whether probabilities can be applied in the criminal context, as we will discuss further.

[86] For example, Dixon J in Briginshaw v Briginshaw [1938] HCA 34; (1938) 60 CLR 336,361-362 (Briginshaw) stated that “when the law requires the proof of any fact, the tribunal must feel an actual persuasion of its occurrence or existence before it can be found. It cannot be found as a result of a mere mechanical comparison of probabilities independently of any belief in its reality ... At common law it is enough that the affirmation of an allegation is made out to the reasonable satisfaction of the tribunal.” (He also suggested that the gravity of the consequences flowing from a particular finding, the seriousness of the allegation made were also relevant in deciding whether the issue has been sufficiently proven). The requirement of reasonable satisfaction was also applied by Stephen, Mason, Aickin and Wilson JJ in West v Government Insurance Office of New South Wales [1981] HCA 38; (1981) 148 CLR 62,66

[87] Respectively, Justice David Hodgson The Scales of Justice: Probability and Proof in Legal Fact-Finding 69 ALJ 731, 732 (1995): “If, on the basis of adequate material concerning circumstances of a particular case, the tribunal believes that an event occurred, with the strength of that belief being at least such as would be indicated by a probability in excess of 50 per cent, then the civil onus is discharged”, Justice David Hodgson A Lawyer Looks at Bayes’ Theorem 76 AUSTRALIAN LJ 109 (2002), and David Hamer The Civil Standard of Proof Uncertainty: Probability, Belief and Justice 16 SYDNEY L R 506, 507 [1994] “the most weighty considerations of justice in the civil trial are accuracy and equality, which are best recognized through a mathematical standard of 50 per cent”. Some cases where probability was specifically applied by the judges in a civil case are Rose v Abbey Orchard Property Investments Pty Ltd (1987) Aust Torts Reports 80-121 (probability that an accident scene had not been properly inspected by the defendant before the accident), and Tenax Steamship Co Ltd v The Brimnes (Owner) [1974] EWCA Civ 15; [1975] QB 929 (probability that payment was made before notice had been given); and note Murphy J’s controversial approach in TNT Management Pty Ltd v Brooks (1979) 53 ALJR 267 – of the three possible explanations for an accident, two stood in favour (at least partly) of the plaintiff, so the plaintiff should be successful

[88] supra n77, 595

[89] 585-586. This issue divided the United States Supreme Court in Victor v Nebraska [1994] USSC 15; (1994) 511 US 1. The majority were in favour of casting the test in probability terms: p6, 10 respectively “The beyond a reasonable doubt standard is itself probabilistic ... (noting) the very high level of probability required by the Constitution in criminal cases”, and “those probabilities must be so strong as to exclude any reasonable doubt” (Justice O’Connor for the majority); while the minority were not in favour: p5 “the word probability brings to mind terms such as ‘chance’, ‘possibility’, ‘likelihood’ and ‘plausibility’ – none of which appear to suggest the high level of certainty which is required to be convinced of a defendant’s guilt ‘beyond a reasonable doubt”.

[90] Negatively, Tribe says the concept of guilt beyond a reasonable doubt “signifies not any mathematical measure of the precise degree of certitude we require of juries in criminal cases, but a subtle compromise between the knowledge, on one hand, that we cannot realistically insist on acquittal wherever guilt is less than absolutely certain, and the realization, on the other hand, that the cost of spelling that out explicitly and with calculated precision in the trial itself would be too high” (Trial by Mathematics (1971) 84 HARV L R 1329, 1801, 1810); Charles Nesson claims that “precise attempts to define the concept of reasonable doubt undercut its function”, because if it becomes too precise, what citizens thought was a common understanding among them as to what it means will be shown to be erroneous: Reasonable Doubt and Permissive Inferences: The Value of Complexity (1979) 92 HARV L R 1187, 1196-1197, and Alan Tyree Probability Theory and the Law of Evidence 8 CRIMINAL L J 224,233 (1984) “the jury is implicitly being instructed to accept a fixed level of improper convictions”.

Positively, see Lindley Probabilities and the Law in UTILITY, PROBABILITY AND HUMAN DECISION MAKING 223, 229-230 (1975), H J Walls What is Reasonable Doubt? CRIMINAL L R 458,469 [1971], who suggests that in order to determine what sort of probability figure represents beyond reasonable doubt, we should give details of cases involving probability to different teams of statisticians, judges and lawyers. A comparison of the results would, he argues, allow us to determine what exactly “beyond a reasonable doubt” could be quantified as, in terms of probability; and Michael Finkelstein and William Fairley A Bayesian Approach to Identification Evidence 83 HARV L R 489, 503 (1970) “The standard beyond a reasonable doubt ... import probabilistic notions”

[91] Rita James Simon and Linda Mahan Quantifying Burdens of Proof 5 LAW AND SOCIETY REVIEW 319 (1971)

[92] This perhaps casts doubt on the bold statement of Isaacs and Powers JJ in Brown that “the words reasonable doubt are in themselves so far self-explanatory that no further explanation is considered strictly necessary” (594).

[93] Of the sociology students surveyed, 7% would have been satisfied with a probability level below .7, 28% satisfied with a probability level between .7 and .9, with 65% only satisfied if the probability level was greater than .9

[94] For murder, judges surveyed required a .92 degree of probability (jurors .86), rape .91 (jurors .75), assault .88 (jurors .75) and stealing .87 (jurors .74). Some judges have similarly commented that the test of beyond reasonable doubt varies in its strictness according to the nature of the offence: Denning LJ in Bater v Bater [1951] P 35, 36-37, Blyth v Blyth [1966] AC 643, 673 and Dixon J in Briginshaw supra n86 (361-362)

[95] Charles Nesson Reasonable Doubt and Permissive Inferences: The Value of Complexity 92 HARV L R 1187,1193 (1979)(using an example where a prison guard is supervising 25 identically garbed prisoners in an exercise yard, where 24 participate in a murder and one sits out. The prison guard cannot see clearly enough which of the 24 participate in the murder and which one sits out) A Burton Bass, H Davidson Gesser and K Stephan Mount Scientific Statistical Methodology 5 DALHOUSIE L J 350, 361 (1979) concluded that “we should not, in all circumstances, remain convinced that a 99% probability equates to an abiding conviction to a moral certainty (or beyond reasonable doubt). Yet our courts do appear to remain so convinced”.

[96] Laurence Tribe supra n90, 1340 (1971)(the Swedish court also concluded that, if another assumption were made and the probability that the driver’s story was true was 1 in 20, 736, that would have satisfied they beyond reasonable doubt test). Jonathan Cohen found that even at a 99% confidence of guilt level, reasonable doubt might still exist: The Logic of Proof (1980) CRIMINAL L R 102; as did Glanville Williams The Mathematics of Proof CRIM L R 297,306 (1979) “If a juror is able to calculate that he is 99% sure, then he already has a doubt, in legal theory, and ought not to convict”, and Lindley estimated that beyond reasonable doubt implied odds of at least 100 to 1: Probabilities and the Law in UTILITY, PROBABILITY AND HUMAN DECISION MAKING 223,229, 230 (1975). Anecdotal evidence is consistent with this. Richard Eggleston refers to a member of a Victorian jury in a murder trial saying that some members of the jury thought beyond reasonable doubt meant 98% certain; the others thought it meant 100% certain: EVIDENCE, PROOF AND PROBABILITY (1983) p114

[97] (1970) 37 SASR 256

[98] Refer to Sir Richard Eggleston Focusing on the Defendant (1987) 61 ALJ 58 for a fuller discussion.

[99] Though, again acknowledging it is controversial to put a probability percentage on reasonable doubt, and even if it is acceptable to do so, the precise figure said to equate to the criminal standard is a matter of conjecture

[100] [2000] USSC 57; (2000) 530 US 466. The Canadian approach is similar in its insistence on proof beyond reasonable doubt for preventive detention regimes: Kirkland (1957) SCR 3; refer to Michael Petrunik Models of Dangerousness: A Cross Jurisdictional Review of Dangerousness Legislation and Practice 1994-2002 (paper prepared for Ministry of the Solicitor-General of Canada)( <http://www.psepc-sppcc.gc.ca/publications/corrections/199402_e.asp> , accessed 12 January 2005)

[101] Justice Stevens, for the majority. Consistently, they have insisted upon a finding that the offender would be likely to re-offend at the beyond reasonable doubt standard, in the context of preventive detention laws: Kansas v Hendricks [1997] USSC 63; (1997) 521 US 346.

[102] Ibid. This standard was required although the majority judges found the proceedings to be civil in nature. The argument that the beyond reasonable doubt standard should be required is suggested to be even stronger if the proceedings are classified as substantively criminal, as the minority of the Supreme Court in that case so classified the preventive detention law.

[103] Only Kirby J addressed this issue in Fardon; he found that while the legislation was dressed up as being civil, in fact it was criminal (punitive)(supra n1, 94). Presumably the other judges did not find it necessary to decide the point, or assumed it was civil in both style and substance.

[104] Chu Kheng Lim v Minister for Immigration, Local Government and Ethnic Affairs (1992) 110 ALR 97

[105] [2002] USSC 10; (2002) 534 US 407; Kansas v Hendricks supra n101

[106] The court in determining whether or not proceedings were civil or criminal in nature referred to Allen v Illinois (1986) 478 US at 369, and Kennedy v Mendoza-Martinez [1963] USSC 21; (1963) 372 US 144, the “Kennedy factors”: whether a sanction involves an affirmative restraint, how history has regarded it, whether it applies to behaviour already a crime, the need for a finding of scienter, its relationship to a traditional aim of punishment, the presence of a non-punitive alternative purpose, and whether it is excessive in relation to that purpose.

[107] This suggestion has not escaped criticism: Stephen Morse Uncontrollable Urges and Irrational People 88 Va LR 1025, 1026-1027 (2002), Eric Janus Hendricks and the Moral Terrain of Police Power Civil Commitment 4 PSYCHOLOGY, PUBLIC POLICY AND LAW 297, 298 (1998) and Christopher Slobogin A Jurisprudence of Dangerousness 98 NORTHWESTERN UNI SCHOOL OF LAW 1,35 (2003)

[108] Similarly, the German Constitutional Court recently, in assessing the validity of a preventive detention regime in that country, insisted that the detention not merely amount to the warehousing of detainees. Adequate treatment had to be provided to those detained under the legislation: Frieder Dunkel and Dirk Van Zyl Smit Preventive Detention of Dangerous Offenders Re-Examined: A Comment on Two Decisions of the German Federal Constitutional Court (BVerfG – 2 BvR 2029/1 oF 5 February 2004 and BVerfG – 2 BvR 834/2 – 2 BvR 1588/02 of 10 February 2004) and the Federal Draft Bill on Preventive Detention of 9 March 2004 5 GERMAN L J 619 (2004). Paul Robinson makes the same point: “Acknowledgement of the preventive nature of the detention also logically suggests a right to treatment, non-punitive conditions, and the principle of minimum restraint, meaning greater freedom among those who are detained”: Punishing Dangerousness: Cloaking Preventative Detention as Criminal Justice ( <http://lsr.nellco.org/upenn/wps/papers/41> ). He also suggests that prisoners detained under civil commitment legislation should be segregated from those in prison as punishment (29). Many of the preventive detention statutes operative in the United States do require treatment of an offender soon after he has been apprehended and charged with a sexual offence, or at least while they are detained (eg the Illinois Sexually Dangerous Persons Act 1973)

[109] Justices Breyer, Stevens, Souter and Ginsburg, noting also that the legislation only purported to provide treatment for the relevant offenders at the time they neared their prescribed jail term, giving rise to a conclusion that the legislation was not in fact civil but punitive.

[110] As the dissenting judges said in Hendricks, “one of the reasons society imprisons those convicted of crimes is to keep them from inflicting future harm, but that does not make imprisonment any the less punishment”. To like effect in Australia, refer to Paul Fairall Violent Offenders and Community Protection in Victoria – The Gary David Experience 17 CRIM L J 40, 50 (1993) “The argument that preventive detention is not intended as, and therefore does not amount to, punishment, may be dismissed out of hand. People are not sent to prison for punishment, but as punishment”, and Robert Williams Psychopathy, Mental Illness and Preventive Detention: Issues Arising From the David Case [1990] MonashULawRw 10; 16 MONASH L R 161,179 (1990) “the essence of incarceration form a punitive point of view is the deprivation of liberty, and this is in no way lessened by claiming the incarceration is civil ... such (preventive) incarceration is ... properly classified as a form of preventive detention akin to imprisonment”.

[111] The United States Supreme Court in Addington v Texas [1979] USSC 77; (1979) 441 US 418 thought the beyond reasonable doubt test was unnecessarily stringent where non-punitive confinement was concerned, and see Christopher Slobogin A Jurisprudence of Dangerousness 98 NORTHWESTERN UNI L R 1, 6-7 (2003) “the reasonable doubt standard is overly stringent when the state’s goal is to prevent rather than to punish”. However, by corollary, if in fact punitive detention was being administered, as arguably it is, the criminal standard of proof would be required. This would surely lead to a more coherent response where both the original finding of guilt of the accused, and subsequent arguments about their dangerousness, would be judged by the same standard of proof, that of beyond reasonable doubt. The author finds it difficult to defend a system whereby the original incarceration is judged by a higher standard of proof than the decision about continued incarceration.

[112] The problem does not arise with those convicted of the most serious crime of murder. That offence is punished by mandatory life imprisonment, affording a Parole Board the opportunity to be satisfied of the offender’s rehabilitation before allowing their release.

[113] However, the number of guilty persons of whom it would be better to acquit than to convict one innocent person tends to vary from 5 to 1000 depending on the individual views of the proponent.

[114] 4 William Blackstone COMMENTARIES 358 “better than 10 guilty persons escape than that one innocent suffer”, Benjamin Franklin “it is better (one hundred) guilty persons should escape than that one innocent person should suffer” (letter to Benjamin Vaughan 14/3/`785) in 9 Benjamin Franklin WORKS (1970) 293, and arguably Aristotle, PROBLEMS, bk 29.13 at 951a37-b8 “it is a serious matter to decide in the case of a slave that he is free, but it is much more serious to condemn a free man as a slave”. Refer to Alexander Volokh Guilty Men 146 U PA L R 173 (1997).

[115] “I would rather wish twenty evil doers to escape death through pity, than one man to be unjustly condemned” (English Chief Justice John Fortescue A LEARNED COMMENDATION OF THE POLITIQUE LAWS OF ENGLAND ch 27 at 63, 1567 (1471), Robert Mulcaster translated, “better five guilty persons should escape unpunished than one innocent person should die” (Lord Hale (1678) 2 HALE P.C.290, by Justice Marshall in Furman v Georgia [1972] USSC 170; (1972) 408 US 238 and Justice Blackmun in Ballew v Georgia [1978] USSC 40; (1978) 435 US 223, 234, and by Murphy J of the Australian High Court in Repatriation Commission v Law [1980] FCA 92; (1981) 147 CLR 635, 639.

[116] This debate continues in the United States. While one academic thought that even if the risk of recidivism was only 50%, preventive detention was permissible because “a mistaken decision to confine, however painful to the offender involved, is ... simply not morally equivalent to a mistaken decision to release ... One is much less harmful than the other”: Alexander Brooks The Constitutionality and Morality of Civilly Committing Violent Sexual Predators 15 UNI PUG SD L R 709, 752 (1992), others vehemently disagree. For example, Professor LaFond responds that “Suddenly, the fundamental assumption of American criminal justice that it is far worse to convict an innocent man than to let a guilty man go free has been transformed into a first principle worthy of George Orwell’s 1984. Now it is far better, according to Brooks, that at least half, and maybe more, of the people confined to a psychiatric prison indefinitely be harmless in order to ‘incapacitate’ those who may commit a future crime. Even better, why not convert our criminal sentencing system into a game of chance? Release from prison would be decided by the toss of a coin. At least this lottery will be more accurate than the one Brooks embraces”: John LaFond Washington’s Sexually Violent Predator Law: A Deliberate Misuse of the Therapeutic State for Social Control 15 UNI PUG SD L R 655, 698-699 (1992).

[117] For example, Fardon had been in jail for 28 of his 44 years before allegedly threatening to commit further offences if released.

[118] Some recent research has found cognitive behavioural techniques to be useful in rehabilitating sexual offenders eg Grubin, D and Thornton, D A national program for the assessment and treatment of sex offenders in the English prison system 21 CRIMINAL JUSTICE AND BEHAV 55 (1994), Hall, G C N Sexual offender recidivism revisited: a meta-analysis of recent treatment studies 63 JOURNAL OF CONS and CLINICAL PSYCH 802 (1995), Beech, A, Fisher, D and Beckett, R STEP 3: AN EVALUATION OF THE PRISON SEX OFFENDER TREATMENT PROGRAMM: A REPORT FOR THE HOME OFFICE BY THE STEP TEAM (1999), Hedderman, C and Sugg, D Does Treating Sex Offenders Reduce Re-Offending? 45 RESEARCH FINDINGS (1996), Home Office Research, Bush, J Teaching Self-Risk Management to Violent Offenders in J McGuire ed WHAT WORKS: REDUCING RE-OFFENDING: GUIDELINES FROM RESEARCH AND PRACTICE 139 (1995), reporting a 45% reoffend rate for those completing the program, against a 77% reoffend rate for those who had not. Refer also to Hazel Kemshall RISK ASSESSMENT AND MANAGEMENT OF SERIOUS VIOLENT AND SEXUAL OFFENDERS: A REVIEW OF CURRENT ISSUES (2002), ch 5 Risk Management of Sexual and Violent Offenders, paper prepared for the Scottish Executive

[119] Relying on the United States Supreme Court discussions on the distinction between civil and criminal proceedings in Hendricks and Crane.

[120] Such regime avoids the legal problems evident in the system of preventive detention by not involving a court in the decision whether or not to continue detention, and not ordering an initial determinate sentence, followed by a last-minute indeterminate sentence where the prisoner is left in limbo for six months at a time. Such a sentence would only follow upon a conviction for an offence proved beyond reasonable doubt.

[121] As Slobogin puts it, speaking of Garry David and other multiple sexual offenders, and proposing a test of undeterrability to justify preventive detention: “all of these people can be considered undeterrable, even though they are not insane under traditional definitions, because their desire for the ‘benefit’ they receive from crime is demonstrably greater than their fear of significant punishment”: supra n 101, 44 (2001)

[122] It is certainly not an answer to the problem since, unless the State Government makes application every six months until the prisoner dies, eventually the offenders detained under the controversial law must be released.

[123] N 31.

Download

No downloadable files available