Abstract
In “Looking Beyond Risk in Paroling Denying Prisoners,” Dr. Dagan provided a refreshing critical analysis of the role of retributive considerations in parole proceedings. Its main argument is that if we are to make sense of the insistence on an admission of guilt, we cannot discuss parole proceedings merely in terms of prevention and risk. What is actually in play, at least in part, is retribution and moral evaluation. Taking retributive considerations into account enables a better understanding of the importance that admissions assume in parole proceedings and their role as a means of moral repentance and catharsis. In this brief commentary, we discuss Dagan’s thesis, offering a few remarks on the relationship between retribution and prevention. We provide some details on the nature of parole proceedings and on the practices that may increase the likelihood of moral and retributive considerations being taken into account contrary to the will of parliament and to the stated principles of parole.
In “The Catch-22 in Israel’s Parole Law” (Assy & Menashe, 2014), we offered a critical review of Israeli parole law, especially in relation to violent or sexual offenders. We showed there that early release from prison is not considered an entitlement but, rather, a privilege that seeks to encourage prisoners to behave well in prison and to engage in treatment and other activities that may help them reintegrate successfully back into society. The primary questions that should be examined in deciding whether parole should be granted is whether the prisoner is no longer dangerous to the public and whether an early release will help reduce recidivism. Parole boards are required by the Parole Act 2001 to assess the risk to the public posed by the prisoner and, to this end, they are aided by professional bodies whose task is to examine certain prisoners and issue detailed risk-assessment reports on them.
The professional body that provides risk assessment reports on sex offenders is the Forensic Psychiatry Division of the Mental Health Center (MHC), a medical division of the public hospital in Beer–Yacov that operates within the Prison Service and offers psychiatric services to prisoners. Incest offenders are risk-assessed by the Regional Committee for the Treatment of Incest, and domestic-violence criminals are assessed by the Inter-Ministerial Committee for the Prevention of Domestic Violence.
To improve their parole prospects, violent and sexual offenders are expected to undergo treatment and to embark on a new, more constructive lifestyle. However, to be allowed to enter such a corrective path, they must admit guilt, take responsibility for their actions, and show empathy for their victims’ suffering. The crucial fact pointed out in “The Catch-22…” is that the professional risk assessment bodies, led by the MHC, have adopted the position that an admission of guilt is the starting point for any successful treatment and rehabilitation plan. Accordingly, sex offenders who deny their crimes are usually excluded from treatment and, subsequently, assessed as being highly dangerous.
When these professional bodies do not recommend parole, the parole boards almost always dismiss the prisoner’s application, and courts almost always defer to such decisions. Early release is therefore rarely, if ever, ordered contrary to the recommendation of these professional bodies. One study found, for instance, that in the period 2008-2011, there was no single incidence in which an Israeli court refused an application to impose restrictions on a sex offender (under the Protecting the Public from Sex Offences Act 2006) against an official risk assessment report (Mashiach, 2012), and another reached similar findings in relation to parole decisions from 2001 to 2011 (Efodi, 2014).
From a legal point of view, the main problem is that the MHC—and, for this purpose, the other professional bodies—enjoys an almost absolute discretion that goes largely unscrutinized by courts. The Israeli Supreme Court has refrained from interfering in the MHC’s principled position on admissions of guilt and has fortified the MHC’s status by instructing lower courts and parole boards to defer to its reports as far as risk assessment is concerned. This means that lower courts and parole boards may not attempt to assess the risk of violent or sexual offenders independently, or to avail themselves to other sources of information or evaluation. Treating the MHC reports as decisive means rejecting all other indications that might shed light on the prisoner’s dangerousness, including other expert opinions, whether solicited by the prisoner or otherwise. This state of law places the MHC in an excessively powerful position, so much so that the MHC has felt comfortable in ignoring the Supreme Court’s several pleas over more than a decade for it to reconsider its position on the weight that should be attached to admissions of guilt (Assy & Menashe, 2014).
To insist on admission of guilt and to attach such weight to it are to overlook the findings of several studies that question the effectiveness of denial as a predictor of recidivism (Hanson & Morton-Bourgon, 2005; Yates, 2009). Other studies show that denial is an issue for treatment (Marshall, 1994; Maletzky, 1996), being complex and possibly with various motivations other than failure to realize the wrongness of certain actions (Assy & Menashe, 2014; Cooper, 2005).
Against this background, “Looking Beyond Risk in Paroling Denying Prisoners” provides a refreshing analysis that puts the topic in a new light (Dagan, 2015). Netanel Dagan suggests looking beyond the rhetoric of risk. His main argument is that if we are to make sense of the insistence on admission of guilt, we cannot discuss parole proceedings in terms merely of prevention and risk. Although the terminology is all about risk prevention, he says, what is actually in play, at least in part, is retribution and moral evaluation. “The voice is the voice of Jacob,” as it were, “but the hands are the hands of Esau” (Genesis 27:22).
This line of thinking may be described as a form of legal realism, which is a legal scholarship whose basic idea is that the law in action differs from the law in the book, and that judges make decisions based on considerations other than those that they state in their decisions (Freeman, 2014). From this standpoint, Dagan (2015) argues that we cannot entirely understand what parole boards and courts really do merely by looking at the risk-assessment reasoning in which they explain their decisions. We must also look beneath the written word. Taking retributive considerations into account enables a better understanding of the importance that admissions assume in parole proceedings and their role as a means of moral repentance and catharsis.
Drawing on Dagan’s analysis, retributive considerations may be said to transform the sentencing regime from a “determinate” one—where the just punishment is imposed by the court in advance, at the sentencing stage, as a function of the immorality or seriousness of the crime in question—to an “indeterminate” one, where the justness and proportionality of the punishment are measured against risk and therefore continue to be examined after the imposition of sentence by the court. Under this taxonomy, risk-based sentences may be shorter or longer than guilt-based ones.
However, irrespective of whether the sentence regime is risk-based or guilt-based, parole may provide another opportunity to re-evaluate the justice of the sentence in the light of developments during the post-conviction period. Confessions, treatment, and rehabilitation play an important role in either regime. In a risk-based one, confessions are believed to promote successful treatment, which would ultimately reduce risk and lead to shortening of the sentence. In a guilt-based regime, admission, responsibility, apology, and remorse are important for atonement and repentance. Thus, successful correction may improve the prisoner’s moral condition and that could in turn entail adjusting the sentence to ensure that it fits the prisoner’s present moral condition.
The attempt to reduce the risk of recidivism has a moral dimension because treatment may influence the prisoner’s personality, values, and chances to become a reformed and better member of society. Successful rehabilitation serves to reduce or even prevent recidivism and, accordingly, wrongfulness, suffering, and pain. The insistence on admission, remorse, and correction represents a desire to discipline the prisoner’s character. As Dagan explains, if confession is a sign of moral enhancement, denial is a sign of moral regression. Retribution therefore takes into account not only the seriousness of the offense in question but also of the culprit’s general “deservedness” or “moral quality.”
Dagan’s theory is descriptive, seeking to explain how parole boards actually make their decisions. The extent to which retributive considerations indeed influence the decision-making process despite the rhetoric of risk assessment and prevention is an empirical question that awaits empirical examination. Such examination is not easy to devise, though, given the double-identity or double-agency of the considerations in question: Most of the factors that may affect a risk assessment also affect a moral judgment (criminal record, the specifics of the crime, treatment, etc), and this makes it difficult to disentangle the two.
Toward the end of his article, Dagan goes on to make a brief normative argument, warning against the infusion of retributive considerations in parole proceedings. In his view, moral evaluations are unwelcome because they are always subjective and never accurate, not to mention the practical difficulty in distinguishing between genuine and fake remorse.
In this brief commentary, we offer a few remarks on the relationship between retribution and prevention. We provide some details on the nature of parole proceedings and on the practices that may increase the likelihood of moral and retributive considerations being taken into account contrary to the will of parliament and to the stated principles of parole.
Setting the Scene: The Risk Assessment Discourse
The present discourse in the parole legislation, case law, and literature in Israel mostly concerns risk prevention. The Israeli Parole Act 2001 states in express terms that conditional release is to be granted when it poses no risk to public order (§§2, 3, 5, 9). The Act provides a list of considerations that may be taken into account by parole boards in constructing their assessments of the risk posed by the prisoner to his family, victims, and the general public (§9). These include the prisoner’s conduct while incarcerated, his prospects for rehabilitation, the specifics of the applicant’s crime, his criminal record, any past paroles and the extent to which he abided by their conditions, the prisoner’s personal situation (including age, family and social conditions, drug use or treatment for addiction), misconduct or violation of rules, breach of order, harm to other inmates or officers, jail break or late arrival from furlough, suspected criminal activity, and attitude toward labor or other rehabilitative opportunities. The focus on risk is evident also in the nature of the restrictions imposed during the parole period, such as house arrest, electronic handcuffs, a commitment to therapy, and a work plan (§13).
As the Act does not prescribe the relative weight of each of the factors it enlists, the risk-assessment task is performed in a holistic fashion. The exercise is unstructured and conducted on an intuitive and subjective level. However, the expectation that the mentioned factors be taken into account for an assessment of risk but not of moral condition or quality may prove difficult because at least some of these considerations, if not most of them, are relevant also to a moral assessment of the prisoner. In other words, some of the considerations detailed in the Act may be applied even under an approach of retribution and vengeance.
The double-agency of risk-assessment factors makes it easier for decision makers to apply a retributive approach under the disguise of risk prevention. This may be made deliberately, in defiance of the will of Parliament, but it may also be made more subtly and unconsciously. Simply, decision makers may be unaware of the influence of moral repugnance on their assessment, or feel comfortable applying a strict risk assessment standard that judges the prisoner more harshly and with reference more to his level of non-cooperation with the authorities than to his real chances of recidivism.
Furthermore, some cognitive biases may occur even against the best efforts to avoid them. So even when decision makers are aware of the tacit influence of morality on their risk assessments and are determined to avoid it, they may fail to discard completely the moral dimension of these same factors and to focus only on their future-looking, risk-based contribution.
To sum this point up, the double-agency or double-identity of the factors that risk assessors tend to take into account makes it important to put their retributive or moral dimension on the table and requires a more candid and clearer discussion about what should or should not be taken into account. Furthermore, although the influence of moral considerations may be tacit and unintentional, putting them on the table and warning against them may help induce decision makers to endeavor to avoid the contamination.
To complete the picture, it is worth noting that the Israeli Parole Act does not completely exclude retributive considerations. It leaves a narrow window open for such considerations, but it limits their relevance to very specific circumstances—namely, when granting parole would render the time spent incarcerated so disproportionally short relative to the severity of the prisoner’s action that public confidence in the law enforcement authority might be substantially undermined (§10). The narrowness of this exception demonstrates the force of the general rule that in all ordinary circumstances parole boards ought to focus on risk reduction and recidivism prevention, not on retribution. It seems that the explicit mention of public confidence considerations seeks not so much to authorize parole boards to include them as to limit them to exceptional circumstances. In practice, parole boards do not apply the exception frequently (but a legal realist might still insist that this simply follows from the fact that such considerations are in play anyway, even when risk-assessment rhetoric is used).
The Problems Inherent in Unstructured Risk Assessments
As described already, retributive and moral considerations may influence parole decision makers unconsciously, as a result of cognitive bias or the difficulty of disentangling morality from prevention. Dagan rightly argues that the unstructured nature of parole discussions aggravates the difficulty of separating these considerations and therefore increases the influence of retributive and moral considerations on risk assessments. The task of risk assessment by parole boards is performed on a subjective and intuitive level, in relatively quick hearings that are not subject to the strict requirements of the rules of procedure and evidence followed in court. The leeway for retributive considerations, which this introduces, is broadened by the conservative judicial review approach of parole decisions.
Moreover, not only are the parole hearings unstructured, but so is the professional examination method. The MHC enjoys by law an independent professional discretion over the professional methods to be used for risk assessment, and so far, it has opted for unstructured clinical examinations. It refuses to use any specific actuarial or otherwise structured risk-assessment tool, leaving the task to the assessor’s best efforts and judgment. The only use the MHC makes of the actuarial tools developed in the literature is to extract from them a set of common static and dynamic variables that have been found indicative of recidivism, requiring the risk assessors to take them into account in their examination of the subject. The evaluation of the relevant static and dynamic factors should be based both on the subject’s papers—welfare reports, prison’s reports, criminal record, the judgment in question, and so on (MHC’s Ethical and Professional Rules of Risk Assessment of Sex Offenders 2011, §3.2.4 [“Ethical and Professional Rules”])—and on an interview. When needed, the subject may be referred to a psychiatric examination (Ethical and Professional Rules, §3.2.5). Yet the final evaluation and relative weights that ought to be attached to each factor in the specific circumstances remain at the discretion of the specific assessor (Ethical and Professional Rules §3.2.1). The MHC justifies its reluctance to use actuarial tools on the grounds that such tools have not been validated and tested on Israeli, non-English speaking subjects (Assy & Menashe, 2015).
In a parliamentary session that was held before the Constitution and Law Committee on 11.8.2005 to discuss the work of the MHC, its representative detailed the kinds of considerations that tend to be taken into account by the risk assessors 1 : the phenomenology of the crime or criminal record; when the first offense was committed; the sort of victims involved (age, gender, relation to the subject); the subject’s personality, employment history, personal status, and family involvement; relevant medical diagnoses, including personality disorders and sexual divergences; the subject’s support circles; and any drug or alcohol dependencies the subject may have and general social functioning.
In the interview, the assessor will take note of the subject’s mode and rate of speech, intonation, and non-verbal communication such as gestures, movements, facial expressions, and, of course, speech content. The assessor will also attempt to understand the subject’s sexual world and, to this end, ask him about his sexual inclination, habits, fantasies, and other such aspects. The assessor may also ask the subject about his conviction, his attitude to the actions he committed, his motives, his thoughts at the time of committing the action, and so on. His attitude to the consequences of his actions for the victim is also relevant; whether or not he has empathy for the victim, and whether or not he regrets his actions.
The risk assessment is then concluded on two levels. One has to do with the degree of dangerousness, ranking low, low to medium, medium, medium to high, high, or very high. Another is concerned with the timescales to which the risk assessments refer. They may indicate short- and long-term threats, depending on the goals of the particular evaluation requested (whether, for example, it is sought for parole or, rather, furloughs). The assessor may also indicate certain restrictions that the subject should be to put under (Ethical and Professional Rules, §3).
This brief review of the risk assessment methodology makes it clear that the MHC regards the exercise as having nothing to do with retribution. The detailed discussion of the factors mentioned above cannot be easily dispensed with, nor can discussion of them be described categorically as a mask for retribution and moral condemnation. Indeed, Dagan (2015) is careful enough to stress that such considerations are not necessarily the only considerations in play and that risk remains the main issue.
Curiously, the MHC insists on using unstructured clinical interviews and risk assessment, although of methods examined in the literature, this method seems to be the least accurate and the least reliable. Actuarial tools or structured clinical examination seems to provide better risk assessment predictions (Monahan, 1981). An avowed legal realist may argue that this unstructured clinical assessment is used deliberately, despite its demonstrated inefficacy, to allow moral considerations to be taken into account. However, one could also say that at least on an unconscious level, such a method enhances the risk that retributive considerations influence the decision makers. Given the broad discretion that risk assessors have, the variety and multiplicity of factors that should be evaluated, and the impressionistic nature of their evaluation, they may be influenced by their moral judgment of the prisoner whether directly, in their final recommendation, or indirectly while ascribing the relative weight to the various risk factors. This tendency is aggravated by the cognitive difficulty of processing a large amount of information and evaluating multiple variables, and by the human tendency to avoid complexity and simplify tasks using heuristics (Saks & Kidd, 1980-1981; Simon, 1982).
Consider, for example, Zuckerman’s argument in relation to fact finding in court adjudication. He argues that factual inferences are not always drawn on a firm epistemological basis but may be influenced by social, cultural, subjective, and personal perceptions and biases (Zuckerman, 2014). This concern does not necessarily entail a post-modernist disbelief in a single universal truth. It only highlights the complexity of the fact-finding process and its vulnerability to subjective biases. These vulnerabilities are likely to be even more intense in clinical evaluations that are, by their nature, subjective.
Dagan recounts a number of factors that help keep this mechanism immune from public criticism. Most parole hearings occur away from the public gaze and, furthermore, while a wrong decision to grant parole may attract acid public criticism when the prisoner reoffends, a wrong decision to refuse parole remains behind the scenes and its consequences remain speculative or unknown.
Possible Interactions Between Retribution and Prevention
Although retribution may be the main justification for sentencing (Ghanayim, 2006; Kant, 1996), it may also serve other social ends, including incapacitation, deterrence, and rehabilitation (Duff, 1986; Hart, 2008). The relative weight of each of these may be debated as a matter of principle and also in individual cases when judges come to prescribe the deserved punishment of a given culprit. Judges need to strike a balance in the light of the specific circumstances of both the action and the actor standing before them.
Parole systems are often introduced as a separate stage, distinct from sentencing. They are forward-looking, concerned with what happens after imprisonment. The hope of obtaining early release operates as an incentive to behave well in the prison and to endorse different norms that may reduce recidivism on release.
The relationship between retribution and prevention—and, as a consequence, between sentencing and parole—may take different forms. For example, it is possible to say that retributive considerations set both minimum and maximum limits of sentence, which delineate the parole period. On this account, the period of sentence that has to pass before a prisoner becomes eligible for parole constitutes the minimum sentence retributive considerations can tolerate, whereas the full sentence presents the maximum that is entailed by them. The proper length within this range could be determined either solely by risk considerations or by weighing these against retributive considerations.
In this model, which may be referred to as the “Range Model,” when parole is exclusively based on prevention, only risk considerations may be allowed, and they operate within the range delineated by retributive considerations that have to do with the seriousness of the crime of which the prisoner was convicted. Here, admissions of guilt, remorse, and rehabilitation can be considered only for prevention purposes, that is, if they are good predictors of recidivism. In the other scenario, when the parole criteria include retribution as well as prevention, the retributive considerations may go beyond the seriousness of the crime and encompass also post-crime and post-conviction developments. Confession, treatment, and rehabilitation efforts may indicate that the prisoner has become a better person than before, which entails a sentence reduction. In this model, the moral condition of the prisoner continues to be considered even after the sentence is imposed, in a way that seeks to adjust the sentence to the prisoner’s current condition.
Another way to describe the interplay between retribution and prevention is to say that when courts prescribe a sentence, they actually determine two components: One is dictated by retribution and is fixed and irreducible (for example, two thirds, in the Israeli system), and another represents dangerousness and prevention, which is the flexible and dynamic part (one third, in Israel). In this model, only risk considerations are permitted at the parole stage. Naturally, risk is a dynamic and flexible factor, and therefore requires continuous evaluation. 2
In a parole system that resembles this model, which may be referred to as the Fixed Formula Model, admission, remorse, and rehabilitation efforts may be considered only for prevention purposes, and if they are reliable indicators for recidivism. Specifying the fixed proportions of retribution and prevention in the sentence enhances certainty and prevents confusion, inconsistencies, and arbitrariness. However, the relative proportions of retribution and prevention, which make up the sentence as a whole, may be debated, as can be the very idea of having a fixed ration (2:1 in Israel). This may be challenged on the ground that it does not allow for judicial discretion and undermines the ability to tailor the punishment to individual circumstances. This, it might be said, is disrespectful of human individuality and diversity (thus echoing the debates on structured punishment, as in the United States; Stith & Cabranes, 1998).
Admission of Guilt and Repentance
From a moral point of view, it is possible to say that a culprit who takes responsibility for his actions, recognizes their reprehensible nature, and regrets their consequences, is morally better than one who denies the crime or its culpability, or otherwise refuses to take responsibility for his actions. Usually, these considerations are taken into account at the sentencing stage.
However, the relationship between admission of guilt and repentance is more complex. To begin with, not all admissions reflect genuine regret or repentance. A culprit who pays lips service or admits guilt merely to obtain a privilege remains morally blameworthy (even more so given his insincerity). Moreover, a culprit may realize the offensive and immoral nature of his actions and come to regret their consequences, yet refuse to express this attitude for various reasons—such as to avoid shame, a marriage breakdown, harassment from other inmates, appearing inconsistent with his trial account, and so on. In other words, genuinely repentant culprits may be found among both those who admit guilt and those who do not, just as unrepentant culprits may be found among both those who deny the crime and those who do not.
Depending on the actual ratio here (which at present is empirically unknown), reliance on admission as a rule of thumb for repentance may arguably be justified. For example, if the number of repentant culprits turns out to be higher among those who admit guilt than among those who deny their crimes—in other words, if admission more often results from repentance than from a desire to obtain privileges, and if repentance is more likely to result in admission than in denial—then a reliance on admission may in the long run produce more good than evil. The problem is, however, that relying on admissions alone to achieve more good than evil is a lazy shortcut that fails to pay sufficient respect to each culprit as an individual who deserves to be examined profoundly and judged by more than an external expression of guilt and repentance.
It might be thought that admission is socially valuable and should be encouraged even when it is insincere, because it might have a positive effect on victims and enhance public confidence in the rehabilitation system. Clearly, such a positive effect would be deceptive and therefore morally questionable, as it would mislead victims and the public into a false assurance that a culprit has been found, punished, and successfully rehabilitated.
Prevention and Retribution Reconsidered
Crime prevention and risk considerations do not necessarily compete with retribution. Early release encourages prisoners to behave well in prison and, later on, to reintegrate in society. Although, under one variation of the Range Model, this comes at a price in terms of retribution for the specific crime, prevention of crime generally is a desirable social policy from a broad retributive standpoint. Because crime is wrong and expresses moral diversion, a social policy that attempts to reduce crime rates is desirable because it reduces the incidences of wrongdoing and with it the need for retribution. There is some risk, though, that such a policy may undermine deterrence, because criminals will know that there is always time for repentance and that they can obtain shortened sentences if they agreed to change and cooperate later on. After all, in certain circumstances, a criminal action, however morally repugnant, may only be a one-off incidence that does not demonstrate an anti-social inclination.
Conclusion
In “The Catch-22…,” we described a situation in which innocent prisoners are faced with a tough choice: Either insist on innocence and be denied parole, or lie by admitting a crime, fake regret, and undergo unnecessary treatment. Dagan’s important response reveals the moral dimension of this position. On the one hand, an innocent prisoner is morally justified in insisting on his innocence (and may even be required to do so). On the other hand, wrongful conviction and sentence are morally bad, so attempting to shorten the sentence reduces the wrong. The moral difficulty lies in the means: succumbing to the pressure of imprisonment and the temptation of parole by making false statements about one’s guilt and remorse. Admitting guilt contrary to the facts deceives victims (and the public) by fostering their belief that the criminal who offended them was found, punished, and reformed. In such circumstances, granting parole when no admission is made is better even from a retributive perspective. The problem, of course, is still that whether the prisoner is innocent or not is ultimately unknown. Even so, and although a prisoner may refuse to admit guilt for various reasons—irrationality, shame, family considerations—it may be that the very fact a prisoner defies the system and refuses to admit guilt even when he has no case in court to win and has the potential of parole to lose may indicate innocence.
Footnotes
Acknowledgements
We are greatly indebted to Nevine Emmanuel and Ayal Groner for their excellent contributions.