Abstract
The New Penology paradigm stipulates that governments increasingly incarcerate ‘unruly classes’ in order to manage rather than punish these groups. Even more than in previous decades, post-industrial society is said to utilize the criminal sanction as a means of repressing the poor, urban, unemployed, and members of minority groups. Drawing on the New Penology framework, the current research uses the example of bail to assess whether risk management rationales have migrated into judges’ decision making despite a law that favors non-incarceration. Using a sample of 975 felony cases, this study investigates factors that impact bail decisions and outcomes in New Jersey, a state that does not permit preventive detention and in which the Constitution favors release on bail. Our findings suggest that the New Penology is generally not prevalent in New Jersey bail practice. Specifically, results suggest that judges do not seek to preventively detain the ‘urban underclass’ and that criminal history (traditionally used as a proxy for dangerousness) does not consistently emerge as a predictor of bail decisions. Race remains one of the strongest predictors of bail decisions, which is consistent with but certainly not limited to principles of the New Penology. These results call into question the inevitability of postmodern development and raise the issue of whether law can delay it.
Introduction
Law changes in response to evolving cultural shifts, but not uniformly. Law may be regarded as a dependent variable contingent on the movements of larger social forces and structures (Black, 1976), but the strengths and capacities of those social forces also vary. Law reflects any of a number of socio-political regimes which are based on different ideologies (or, different ‘knowledges’) (Foucault, 1972). It is well known that governments erect legal frameworks to achieve goals of their different political philosophies, but what is less understood is how these different laws interact, change, and have impact. When ideologies conflict, we would expect their corresponding laws to also conflict. When political ideologies change, the laws change with them; however, there may be a lag in updating the legal framework, particularly if the change is contested. Vestigial legal regimes may co-exist with new ideologies and new juridical requirements, though the old laws are likely to have little impact when socio-political developments have overtaken them.
When old law endures under new socio-political regimes, it may persist on the books but be dead in practice, or it may continue operating as it was initially designed to function. If the latter, it may be that old laws present opportunities for resisting movements toward new ideologies, and therefore constitute independent variables themselves – that is, that law affects socio-political movements, and does not simply passively reflect them. That law is an independent variable has always been the assumption of legal activists who seek to foment change.
This article examines a situation in which an old law persists on the books even though contemporary sensibilities have changed. The ideological shift examined here is the movement away from rehabilitative, Enlightenment-era, offender-centered correctional regimes to a New Penology in which criminal offenders are incapacitated and managed according to their predicted degree of dangerousness. The US federal and almost all state governments have changed their laws regarding pre-trial detention so as to conform to principles of the New Penology, but this legal reform was not universal. Unlike most other states, New Jersey did not update its bail laws to embrace incapacitative rationales. In the federal system of the United States, state law cannot violate the US Constitution. It does, however, grant states great leeway in designing and applying their own laws. These state laws can offer greater constitutional protections than the federal law requires, though not lesser protection (Michigan v. Long, 1983). As a result, it is possible for legal reform to bypass a state that holds fast to its own legal standards. The old law can survive – but will it work in different ways so as to serve the new ideological assumptions, or will the ‘law on the books’ resist and retard socio-political trends?
The current study investigates factors that affect bail decisions and outcomes in New Jersey, a state that favors release on bail. Because concerns associated with risk management and control of dangerous groups have become quite widespread in law, policy, and research, the current study aims to assess whether New Penological ‘waste management’ rationales have migrated into judges’ decision making despite the law’s mandate. We use a sample of 975 felony cases to investigate the influence of several legal and socio-demographic indicators on decisions to release defendants on bail, exploring whether the New Jersey bail system has indeed resisted adoption of the New Penology.
The rise of the New Penology
Penology has changed over the decades, and law has changed to serve it. The rehabilitative ethos announced at the turn of the 20th century in the manifesto of the American Correctional Association gave way in the 1970s to retributive (or ‘just desert’) rationales for punishment (Cullen and Gilbert, 1982; see also Allen, 1981). Retributivism rests on the fundamental requirement that a criminal should be punished according to the degree to which he is blameworthy, and therefore people who commit non-serious crimes will receive less severe punishments. However, in the 1980s, incapacitative ideology, which disregards blameworthiness, took hold. Under this paradigm, offenders convicted of non-violent crimes were incarcerated for periods of time inconceivable either in the rehabilitative era or under retributive rationales. Raw incapacitation became the often-unspoken goal in many areas of penological practice, propelling the increased use of risk management (Feeley and Simon, 1992). In bail policies, this risk management approach took the form of preventive detention.
According to Feeley and Simon (1992), surveillance and management of entire classes of the population – rather than either treatment or punishment of individuals – characterizes a New Penology. The authors state that the clearest example is selective incapacitation, which aims to ‘identify high-risk offenders and maintain long-term control over them while investing in shorter terms and less intrusive control for lower-risk offenders’ (Feeley and Simon, 1992: 458). Both low-risk and high-risk offenders are to be managed efficiently: ‘The new penology is neither about punishing nor about rehabilitation of individuals. It is about identifying and managing unruly groups… Its goal is not to eliminate crime but to make it tolerable through systemic coordination’ (Feeley and Simon, 1992: 455). This form of coordination includes techniques of control akin to ‘waste management’, in which the ‘urban underclass’ (regarded as a permanent danger to the broader society) is controlled by risk management strategies such as preventive detention and drug courier profiling (Feeley and Simon, 1992; Simon and Feeley, 1995: 166).
Feeley and Simon argue that three major shifts have occurred under the New Penology: the language of clinical diagnosis gave way to that of probability and risk; the goal of reducing recidivism perversely mutated such that criminal record became a positive indicator (enabling the justice system to intervene more often and more harshly); and techniques for targeting and punishing individual offenders morphed into methods of targeting deviants as an aggregate. Accompanying these new objectives is a new set of practitioner techniques. Auerhahn (1999) observes that incapacitation and risk management have become intertwined, as each of these objectives can be accomplished by utilizing the other. Moreover, using a ‘discourse of poverty and the underclass’, criminal justice agencies exercised heightened surveillance over specific segments of the population deemed to be high-risk: ‘largely black and Hispanic populations living in concentrated zones of poverty in central cities, separated physically and institutionally from the suburban locus of mainstream … America’ (Feeley and Simon, 1992: 467–468). These ‘large islands of third world misery in America’s major cities’ provide a permanent offender population to which low-cost management techniques can be constantly applied (Feeley and Simon, 1992: 467).
Others have observed that such developments are to be understood in the wider context of the sociology of ‘late modernity’. Garland describes political and economic developments of the past five decades as a march toward the ‘culture of control’, noting how neo-liberal disdain for welfarist approaches to various social problems (including crime) changed the sensibilities of policymakers and the public alike (Garland, 2001). Enlightenment rationality bred modern scientific and information-age revolutions as well as the ascendance of individualism and rights-based discourses, but in late modernity economic forces pushed the individual aside. Garland (2001: 100, emphases in original) argues that ‘the watchwords of post-war social democracy had been economic control and social liberation, but the new politics of the 1980s put in place a quite different framework of economic freedom and social control’.
The ‘management of dangerous classes’ would entail the incapacitation of a high proportion of defendants accused of non-violent crimes, even if only for short periods. The first opportunity for doing so, after arrest, would be to deny release on bail. A judge can release a defendant without requiring monetary bail (on own-recognizance, or ROR), or can refuse to release unless a monetary surety is provided, on the presumption that an accused who may otherwise not come back to court if released will return to get the money back. Arrestees who cannot pay the required amount are jailed while awaiting trial. The variety of options available in bail decision making renders this area of law useful for assessing whether trends toward greater control, surveillance, and risk management have in fact taken hold in the United States and if so, whether this is inevitable or can be resisted under old laws that do not embrace these ideological shifts.
The right to bail in the United States: Historical background and legal precedent
The Eighth Amendment of the United States Constitution provides that ‘excessive bail shall not be required’ and Federal Rule of Criminal Procedure 46 requires that ‘[a] person arrested for an offense not punishable by death shall be admitted to bail’. Some (e.g. Foote, 1954) have interpreted this as a constitutional guarantee of bail, a system that would be essentially identical to that of New Jersey. The Supreme Court in Stack v. Boyle (1951: 4) stated the reason for which release on bail has traditionally been regarded as essential to the criminal defendant: ‘This traditional right to freedom before conviction permits the unhampered preparation of a defense, and serves to prevent the infliction of punishment prior to conviction.’ The purpose of bail was initially to ensure a defendant’s appearance in court. Foote and other critics, however, suspected that orders to post monetary bail were being used as a method of detention. Their research exposed the inequitable practices of the system which had become dominated by money (Beeley, 1927; Foote, 1954). More specifically, Foote maintained that release decisions were based primarily on the nature of charges (with little consideration of factors related to the likelihood of appearance in court), that cash bail was being used to detain and punish the poor, and that defendants who were incarcerated before trial eventually received harsher sentences upon conviction than those who were granted pretrial release. Foote expected that these practices would increasingly come under legal challenge on Fourteenth Amendment equal protection grounds (Foote, 1965a, 1965b; Thomas, 1976), but instead reformers determined to address the inequities through administrative controls. Consistent with calls for more careful guidance of judicial discretion, guidelines for bail decision making were developed and implemented in many jurisdictions across the nation.
Bail guidelines were based on factors that predicted the defendant’s flight risk (Feeley, 1984; Thomas, 1976) and were thus consistent with reformers’ interpretation of the Eighth Amendment’s ‘right to bail’ language. Most notably, the Vera Foundation formed the Manhattan Bail Project in 1961, which led to a significant increase in the use of release on personal recognizance (ROR) (Ares et al., 1963). ROR refers to non-financial release (Eskridge, 1983) granted on the defendant’s promise to return for trial. The federal Bail Reform Act of 1966 required judges to grant ROR whenever possible; 1 an individual charged with a non-capital crime was presumed to be releasable on personal recognizance or with an unsecured personal bond, unless the release was believed to compromise the defendant’s return to court. If needed, additional conditions could be imposed to ensure the defendant’s appearance at trial, such as maintaining community and family and job ties, regular contact with the court, and so on (Bogomolny and Sonnenreich, 1969).
Even as the use of ROR increased, a change in the perceived purpose of bail was underway. A public campaign for harsher crime legislation began in the 1970s. One result was the enactment of the federal Bail Reform Act of 1984, which was radically different from the 1966 Act. This new law authorized preventive detention for potentially dangerous persons. According to 18 USC 3142(f), federal judges may preventively jail defendants charged with the following crimes: (1) a crime of violence; (2) an offense for which the maximum sentence is life imprisonment or death; (3) an offense for which a maximum term of imprisonment of ten years is prescribed in the Controlled Substances Act (21 U.S.C. § 801 et seq.) … or (4) any felony if the person has been convicted of two or more [serious] offenses …
How did the passage of this new bail legislation impact the use of own-recognizance release? The federal Bail Reform Act of 1984 permits ROR or unsecured personal bond for non-serious offenses if the release of a defendant is not believed to jeopardize the safety of the community. However, Section 3142(e) stipulates that if the judicial officer determines that no conditions of release will assure a defendant’s future court appearances or provide adequate protection of the community, the defendant (even if charged with a low-level crime) must be detained. Although, in principle, the 1984 Act strongly supported own-recognizance release, the emphasis on community safety introduced a new consideration into bail decisions and signaled a shift from the traditional purpose of bail. Moreover, the 1984 Act invited the judiciary to impose more stringent conditions even when release was granted, such as financial sureties in cases where own-recognizance release would previously have been expected.
In United States v. Salerno (1987), the Supreme Court upheld the 1984 Act and thus approved the notion that, under the Eighth Amendment, bail could be denied for purposes of preventive detention. 2 Subsequently, a second bail reform movement focusing primarily on the dangerousness of criminal defendants took hold (Baradaran and McIntyre, 2012; Fagan and Guggenheim, 1996), in which states passed laws that mimicked the federal Act (Wiseman, 2009). Today, 48 states have some law permitting judges to consider dangerousness or risk of reoffending in bail decisions, while only two (New Jersey and New York) have never amended their statutes (Baradaran and McIntyre, 2012: footnotes 52–55). Of the 48 states, most instituted pretrial services agencies. Originally mandated, under the VERA reforms, to investigate factors associated with the likelihood of returning to court, these agencies began to predict dangerousness instead; they assumed new duties of surveillance and drug testing of individuals on bail. New Jersey never instituted a pretrial services agency. Whether New Jersey’s ‘old law’ and court practices are effective in resisting what is often seen as an inevitable march toward incapacitation, heavy management of the urban underclass, and actuarialism for predicting dangerousness is investigated in this article.
Bail practices in the state of New Jersey
New Jersey has retained a scheme adhering to the original goal of bail: to return defendants to court to stand trial. Preventive detention in bail decisions is prohibited under the New Jersey Constitution; as such, every defendant is presumed to be a candidate for bail. However, judges have discretion to determine the type of bail release to be granted (i.e. own-recognizance release or money bond paid upfront either in full or at 10 percent). To guide judges in making this decision, the state’s judiciary adopted court rules consistent with the state Constitution and drawn from the 1966 (but not the 1984) federal Bail Act. The court rules are advisory and are to be applied consistently with the overarching Constitutional presumption that favors the granting of bail. Judges are advised to consider: 1) the seriousness of the crime charged against defendant, the apparent likelihood of conviction, and the extent of the punishment prescribed by the Legislature; (2) defendant’s criminal record, if any, and previous record on bail, if any; (3) defendant’s reputation, and mental condition; (4) the length of defendant’s residence in the community; (5) defendant’s family ties and relationships; (6) defendant’s employment status, record of employment, and financial condition; (7) the identity of responsible members of the community who would vouch for defendant’s reliability; (8) any other factors indicating defendant’s mode of life, or ties to the community or bearing on the risk of failure to appear, and, particularly, the general policy against unnecessary sureties and detention. (NJ Court Rule 3:26–1(a), emphasis added)
In considering the seriousness of the crime charged, New Jersey law allows the judge to apply bail restrictions (i.e. monetary conditions), but not preventive detention, to first and second degree felonies: crimes of violence, burglary, arson, child endangerment, and weapons use as well as first or second degree drug-related crimes or racketeering crimes. The court can also choose to release a defendant on personal recognizance or conditional release (NJ Court Rule 3:26–1(a)), and such release is assumed for defendants who are not charged with the serious offenses listed above. By contrast, federal law and state laws emulating it allow judges to detain individuals accused of crimes of low seriousness if deemed necessary for the protection of the community. This is the heart of the divide between New Jersey’s old law and bail laws predicated under the ‘New Penological’, post-modern ideology of risk management.
In addition to the New Jersey Court Rules, state case law (or lack thereof) reiterates that dangerousness is not to be the primary consideration in New Jersey’s bail scheme. While the federal courts have decided many cases approving the use of preventive detention, New Jersey’s constitutional prohibition of the practice has held firm. There is only one major case on the matter. In New Jersey v. Johnson (1972), the state Supreme Court held that lower courts cannot constitutionally deny bail in non-capital cases. This 1972 case was decided before the 1984 Bail Reform Act encouraged states to move to a more restrictive model, but there is no subsequent New Jersey case indicating that the state of New Jersey did so. As recently as 2001, in New Jersey v. Korecky, the state Supreme Court reiterated that the main goal of bail is to ensure a defendant’s presence in court; however, the court also noted that detention might be appropriate in cases where a defendant has engaged in misconduct during release on bail (in this case, threatening victims and witnesses). This is the extent to which New Jersey courts have addressed the issue of dangerousness, signaling that New Jersey strives to administer a bail system adhering to the initial purpose of bail (i.e. returning defendants for future court appearances) rather than a rationale primarily concerned with potential danger.
That New Jersey’s bail system seems ‘frozen in time’ is also indicated by the fact that the state has no pretrial services agency that investigates and reports to the court each defendant’s personal circumstances such as community ties, residence, employment, and so on, as listed in NJ Court Rule 3:26–1(a). At this very early pretrial stage of bail decision making, the only objective facts that a New Jersey judge possesses in assessing bail readiness is the seriousness of the charges and the arrestee’s criminal record. At the bail hearing, the defense attorney provides an overview of personal factors that could predict the likelihood of returning to court for trial, to the extent that they are known, and the prosecutor may respond using any available investigative facts from the police file. Judges can give these facts whatever credence that they wish. New Jersey courts do not systematically record defendants’ personal circumstances or community ties. Such information is not available to judges in New Jersey except as uncorroborated assertions offered by defense attorneys at the bail hearing. In addition, courts do not keep any records of the information known to or relied upon by judges. 3
By contrast, in states which have embraced the risk-management paradigm, courts have investigators and sometimes dedicated pretrial agencies to confirm facts about defendants’ personal circumstances; they often rank defendants on a scale measuring the likelihood of committing more crimes while on bail. The likelihood of returning for trial is not considered in these scales. The state of Virginia, for instance, might be seen as the opposite of New Jersey in this regard. With its ‘risk of reoffending’ scales prepared to guide judges in bail decisions and its pretrial agency conducting surveillance of defendants even if they are released on bail (Ostrom et al., 2002; VanNostrand, 2003), Virginia’s pretrial release system is surely an example of ‘New Penology’ in practice.
This is bail’s legal landscape. Since individuals who are incarcerated while awaiting trial have not yet been convicted, bail practices have great implications for issues of fairness and justice. Pretrial detention negatively impacts sentencing decisions. Discussions of the fairness of the law on bail inevitably return to Goldkamp’s (1979) classic observation that there are ‘two classes of accused’: those who have the financial means to be released, and those who do not. This low-level and rarely scrutinized stage of criminal processing has become married to a structure which demands money for freedom. Middle-class and wealthy defendants can afford to post bail, while working-class and poor defendants must rely on private bail bonds, a profit-making industry whose operations in almost every courthouse of every county of the nation is unique to the USA (Liptak, 2008). Of course, it has always been the case that defendants with limited financial resources are jailed while awaiting trial and those in better socioeconomic standing post bail money and are released (Wice, 1974). However, the postmodern difference is that the poor might now be held because judges predict that they are dangerous.
Criminological and sociolegal research on the New Penology in operation
Criminological research helped lay the groundwork for incapacitative rationales which find full expression in the New Penology. Selective incapacitation aims to reduce crime by identifying and incarcerating high-rate offenders, under the assumption that they are most likely to commit more crimes in the future (Von Hirsch, 1998). A landmark study on selective incapacitation conducted by Greenwood and Abrahamse (1982) used a survey of self-reported, high-rate incarcerated robbers and burglars to devise a seven-factor predictive index, including measures of criminal history and drug use. The scale has been widely criticized both on methodological and theoretical grounds (Auerhahn, 1999; Chaiken and Chaiken, 1984; Tonry, 1995; Von Hirsch, 1984, 1998; Zimring and Hawkins, 1997). Further research supported the ability of criminal justice agencies to manage and control dangerous populations through predictive methods (Moore et al., 1984). Actuarial models applied to criminal justice operations are often regarded as a hallmark of New Penological strategies (Field and Tata, 2010).
When applied to bail rather than sentencing, however, selective incapacitation is a difficult fit. Sentencers may be concerned with risks of reoffending, but bail decisions traditionally focus on flight risk. When imposing sentences, judges may not be well equipped to make accurate predictions about future behavior (Williams, 2003), but can at least rely on pre-sentence reports containing full information about the defendant’s personal circumstances. By contrast, bail hearings are held within days or even hours of the defendant’s arrest; as such, judges have very little information, which may lead to unreliable decisions and overprediction of danger (Gottfredson and Gottfredson, 1990 [1988]). Moreover, if a major critique of the New Penology is that it relies on actuarial methods to predict and control ‘dangerous populations’, it could be said that actuarialism in bail decisions could be used for an opposite purpose: to release defendants predicted to pose a limited flight risk. Thus, studying bail in the context of the New Penological paradigm holds promise for parsing out those elements that may be associated with postmodernist surveillance (such as control of dangerous classes) and those that may be neutral tools that could be applied to any task (i.e. managerialism and actuarialism) (see O’Malley, 2010; Silver and Miller, 2002).
Critics have also argued that selective incapacitation policies widen the gap in imprisonment rates between Black and White offenders (Tonry, 1996; Western, 2006). Attribution theorists argue that judges tend to see young Black men as dangerous, while regarding young White offenders as candidates for community sanctions. Applying these arguments to the bail decision, Schlesinger (2005) found that legal variables are the strongest predictors of judges’ pretrial decisions; however, she found that race and ethnicity best predict pretrial outcomes. The severity of charges and prior record of the defendant best predicted the decision to release. However, in cases where money bail was required, Black and Latino defendants were less likely than White defendants to be able to post bail, and were therefore more likely to be incarcerated during the pretrial period. Studies conducted in other jurisdictions have found that White defendants typically receive better pretrial decisions and outcomes when compared to Black and Hispanic defendants (Demuth and Steffensmeier, 2004; Katz and Spohn, 1995; Nagel, 1983; Schlesinger, 2005). Demuth and Steffensmeier (2004) found that in addition to race, prior criminal record and offense severity exert a strong effect on both pretrial decisions and outcomes.
These factors – race, prior criminal record, and offense severity – appear to be significant in explaining official decisions at all stages of criminal prosecution and, we argue, under differing penological approaches. They appear at the height of modernist rehabilitative punishment rationales and also in New Penological discourses. The stated reasons for relying on them – crime prevention versus criminal management – may differ, but the cold fact of their significance is omnipresent. Considering that these variables could evince both modern and postmodern approaches, it would not be surprising to find both at work simultaneously in observations of contemporary criminal justice operations.
Empirical observations of contemporary penal practices around the world find that ‘old and new discourses interact dynamically’. Local conditions, philosophies, and norms of official behavior can deflect the march toward managerialism and provide ‘counter-narratives to actuarialism… by evoking competing penal discourses’ (Field and Tata, 2010: 236). For instance, Wandall (2010: 329) looked for evidence of risk rationales in Danish courtroom decision making, finding that the process is formally and explicitly dominated by a modernist welfare account [but] … at the same time and at the level of informal and implicit communication … provides ample room for constructions echoing more recent penal change towards risk assessment.
These observations of court practices are confirmed in studies examining the influence of the New Penology in correctional practices. Hannah-Moffat (2005: 29) found that prison staff embraced approaches aimed at managing risk but that risk assessment combined with the assessed needs of the offender produced a ‘transformative risk subject… amenable to targeted therapeutic intervention’, which seemed to merge old-fashioned rehabilitative rationales with principles of the New Penology. Inderbitzin (2007) studied incarceration of US youth, concluding that correctional workers aimed to prepare youths for a life of lowered expectations consistent with future work opportunities available to the underclass. Management of the underclass is a hallmark of neo-liberal penal practices. Pretrial processes such as bail, however, presumably have different goals that do not include punishment. Kellough and Wortley (2002) tested the influence of the New Penology in bail practices in Canada – which, like New Jersey, has not passed legislation permitting judges to use the likelihood of recidivism as a criterion in bail decisions. The authors found that bail decision making had not been replaced by systematic risk management and risk assessments rationales. Decisions were based on individual assessments of the accused as set out in police reports. Police recommendations, not judges’ predictions of dangerousness, played the most important role in these decisions, and those recommendations were based on offenders’ individual situations as indicators of flight risk.
In sum, previous research indicates that New Penological practices, if evident at all, are woven into traditional correctional approaches and cannot easily be distinguished. However, these were studies of courts in Europe and Canada. The original formulation of the New Penology was offered by two US scholars observing the massive imprisonment boom in the United States beginning in the 1980s and thus may have over-reached in expecting that the trend would become normative worldwide. Even in the United States, subsequent research has shown that punitivity is significantly more varied among the various US states than what was commonly believed among those who decried ‘punitive American exceptionalism’ (Whitman, 2003). Kutateladze (2009) used multiple indicators to measure punitiveness in all 50 states. He found that the Northeast region is about as punitive as Western European nations, while the South provides the harsh practices that cause alarm over draconian US norms. In short, even in the United States, state and local politico-legal culture may mitigate New Penological tendencies. In the current study, we test this hypothesis in a liberal Northeast American jurisdiction with a state Constitution firmly rooted in Enlightenment principles.
The current study
This study investigates the factors that impact bail decisions and outcomes in a state where the law presumes release. Although we do not compare New Jersey’s bail system to that of a state which implemented legislation emulating the 1984 Bail Reform Act, we can hypothesize what a bail system would look like if it were operating under preventive detention assumptions and compare New Jersey’s bail decisions to this hypothetical model. Specifically, if the old ideology continues to guide judges’ decisions in New Jersey, we would expect the following findings to emerge from our analysis: first, consistent with New Jersey law, all defendants would be granted bail in some form (ROR or financial bail). Second, although it is expected that cases involving violent offenses would entail less ROR and higher bail amounts when compared to cases involving non-violent offenses, the amounts should not be exceedingly high as to constitute de facto detention. Third, contrary to New Penological predictions, it is hypothesized that young, urban, minority males from lower socioeconomic strata will not be subject to less favorable bail decisions and outcomes under New Jersey law. Fourth, if judges are following the mandate of New Jersey law, which does not regard dangerousness as a guiding principle in decisions to detain prior to trial, then it is expected that prior criminal record and other proxies for dangerousness will not emerge as strong predictors of bail decisions and outcomes.
If we confirm these expectations, this will be preliminary evidence that the law matters, that it is possible for a legal system to resist the adoption of principles of selective incapacitation, actuarial risk management, and heightened surveillance over the urban underclass, even in a political and social climate of ‘late modernity’. In short, it would confirm Kellough and Wortley’s (2002) Canadian observations, demonstrating that a counter-narrative to the New Penology can be maintained in the United States when the decentralized nature of US federalism provides the political space for it.
We assess bail operations in several ways. First, we examine the effects of legal and sociodemographic variables on bail decisions, examining the nature of the charged offense, urban/non-urban jurisdiction, criminal history, race, gender, and age. Next, we explore the outcomes of those decisions, especially the operation of money bail, to gauge the importance of economic resources in a state in which there is a legal presumption ‘against unnecessary sureties and detention’ (NJ Court Rule 3:26–1(a)). We explore the effects of these variables at each stage of bail decisions and outcomes, extending from own-recognizance release to money bail in its various incrementally increasing forms, to the imposition of high amounts of money bail (which defendants may not be able to afford, even with the help of private bail bondsmen, thus leading to pretrial detention).
Data
This study is based on individual-level data from the state of New Jersey’s Criminal Disposition Commission (CDC). The CDC was a statutorily mandated agency comprised of the chiefs of all principal state criminal justice agencies (NJ Stat. § 2C:48–1 (2008)). Its mission was to evaluate state criminal justice practices and policies to determine if systematic inequities were present. The CDC compiled a unique database in which it recorded every felony charge issued in New Jersey from 18 October 2004, to 24 October 2004. The week was randomly selected but assessed carefully in order to ensure that no unusual major events had affected court operations then. This sample of felony cases is representative of typical decisions made in New Jersey courts, providing generalizability of findings (at least for this state). The movement of these cases through the system, from arrest through sentencing, culminated in the creation of a database containing 2093 initial cases. Of these, 1118 were screened out as a result of the absence of bail decisions. These cases reflect instances in which: the judge dismissed the case at arraignment; defendants pled guilty, paid fines, and were released; prosecutors decided to either: (a) downgrade the charges to misdemeanors and remand the cases to municipal court, or (b) transfer the cases to a division outside of criminal court (juvenile, family, etc.). Because these arrestees were not subject to the bail process, their cases were not included in the analyses, resulting in a final sample of 975 felony cases.
The current study investigates three dependent variables: bail type; bail amount; and bail outcome. Bail type is a binary variable indicating whether the judge granted release on own recognizance or imposed a financial bail condition. Bail amount refers to the amount of money required in cases where the judge imposed a cash bail condition; this variable was logged due to the skewness of the distribution. Bail outcome is a binary variable that designates whether, in those cases in which judges imposed cash bail conditions, the defendants were able to post bail and be released. If they could not, they were jailed.
The CDC database includes demographic and legal variables about each defendant, some of which were used as independent variables: age, gender, race (White, Black, Hispanic); jurisdiction type (urban/non-urban); and defense attorney type (public defender/private counsel/no attorney present at the bail hearing). 4 Jurisdictions were designated as urban/non-urban depending on population per square mile of each county (New Jersey is the densest state in the nation, with no ‘rural’ counties as in less populated states). The CDC data do not include specific information about arrestees’ socioeconomic status (because judges themselves lack access to reliable information relating to this indicator), and thus we use attorney type (private, public, and no attorney) as a proxy for the defendant’s financial situation. 5 The type of counsel has been included in prior research examining criminal justice decision making (see, for example, Petersilia and Turner, 1986). The New Penology holds that fewer young, poor, minority males from urban areas will gain release on bail. These are attributes that judges can observe at the bail hearings, even when they do not have any other personal information about the arrestees.
The data also include several other legal indicators relevant to bail decisions. Offense type is a four-category variable consisting of ‘person and weapons offenses’ (otherwise known as violent offenses), property offenses, drug offenses, and ‘other’ felonies. This roughly follows a crime severity (or ‘gravity’) scale, which is a common method of assessing the seriousness of criminal activity (Wolfgang et al., 1985) and its effects on sentencing (Johnson and Betsinger, 2009; Steffensmeier et al., 1993). 6 The ‘other’ category includes a range of non-violent public order offenses such as solicitation for prostitution, but also some white-collar ‘crimes of dishonesty’ such as forgery and perjury, all of which were indicted as felonies. The CDC data only include felony offenses, and not misdemeanors which in New Jersey are called ‘disorderly persons offenses’. The number of charges comprising a case, as a continuous measure of offense severity, is also included in the analyses. The number of charges may correspond to judges’ concerns about the potential threat that defendants may pose to the community (see Kellough and Wortley, 2002; Schlesinger, 2005). In addition to information about the seriousness of the offense, given that criminal history has been found to be the strongest predictor of judicial decisions at all stages of criminal procedure (Gottfredson and Gottfredson, 1990 [1988]), we include a measure of prior convictions (continuous variable). We opted for prior convictions as opposed to arrests; for reasons of due process, courts cannot use the number of prior arrests as a measure of criminal history. It is nonetheless noteworthy that the number of prior convictions was highly correlated with the number of prior arrests (r = 0.92, p < .001).
Sample description (n = 975)
The analyses that follow assess the effects of legal and extralegal factors on bail determinations and outcomes using binary logistic regression (separate analyses for bail type and bail posting status) and Tobit analysis (bail amount set by the court). Interactions between key offending measures (current offense type and number of prior convictions) and demographic variables (age, race, and gender) are also examined in all analyses, and are only presented in the models when significant interaction effects were found.
Findings
Bail type
Logistic regression analysis: Dependent variable = bail type (0 = ROR; 1 = financial bail) (n = 926)
Note: *p < .05; **p < .01; ***p < .001.
Reference categories are in parentheses.
Table 2 also shows that the number of charges (Wald = 15.26, odds = 1.14, p < .001) significantly predicts the odds of a money bail requirement; an increasing number of charges pending against a defendant reduces the odds that a defendant will be released on own recognizance. Of primary interest is the impact of offense type on conditions of bail imposed. When compared to violent offenses, the odds of having a financial bail requirement are 52 percent lower for property offenses, 59 percent lower for drug offenses, and 40 percent lower for other offenses. This is to be expected both under ‘old law’ (concerned with likelihood of returning for trial: defendants charged with more serious crimes face higher punishments and thus have more incentive to flee) and New Penological dangerousness assessments for community safety.
Two points are noteworthy. First, the number of prior convictions, often used as a proxy for dangerousness, is not a significant predictor of judges’ decisions to grant own-recognizance release versus imposing a financial bail requirement. Second, judges followed New Jersey law in granting at least financial bail to all defendants, thus rejecting preventive detention and its New Penological imperatives. Outright preventive detention of arrestees charged with the most serious crimes is not practiced. It may also be that if judges were setting impossibly high financial requirements, it would be evident de facto. Analysis of the next two variables indicates that this is not the case in these data.
Bail amount set
Tobit analysis: Dependent variable = bail amount (logged) (n = 914)
Note: *p < .05; **p < .01; ***p < .001.
Reference categories are in parentheses.
Outcome (ability to post bail)
Logistic regression analysis: Dependent variable = bail posting (0 = bail not posted; 1 = bail posted) (n = 661)
Note: *p < .05; **p < .01; ***p < .001.
Reference categories are in parentheses.
The number of charges was not significantly associated with the ability to post bail, but prior convictions did emerge as significant (Wald = 7.65, odds = 0.93, p < .01). With regard to offense type, the odds of posting bail are significantly lower for defendants charged with property offenses (Wald = 13.61, odds = 0.37, p < .001) and other felony offenses (Wald = 6.33, odds = 0.40, p < .05) when compared to defendants charged with violent offenses. No significant differences were observed between violent offenses and drug offenses. These findings may seem counterintuitive, as we expect judges to require very high financial bail in violent cases (and perhaps in drug cases as well) as a `back door' method of achieving preventive detention where the law prohibits it. Because bail amounts are higher for defendants charged with violent offenses (as evidenced in Table 3), it was reasonable to expect that the odds of posting bail would be lower for these individuals. Further analyses examined the outcomes of cases in which very high bail amounts are imposed (among the top quartile of the distribution of bail amount, i.e. above $22,500). We find that a higher proportion of defendants charged with violent (58.1 percent) and drug offenses (53.6 percent) are able to post these high bail amounts when compared to property (30 percent) and other felony (25 percent) offenses, χ2 (3, n = 268) = 15.20, p < .01. These results may partially explain the unexpected direction of association between offense type and bail posting status.
Most coefficients remain unchanged when introducing an interaction term between minority groups and prior convictions (see Model 2), with the exception of prior convictions, which is no longer significantly associated with the odds of posting bail. The significant interaction between minority status and prior convictions indicates that a more extensive criminal history affects the odds of posting bail more substantially for Black or Hispanic defendants when compared to White defendants.
Discussion
The main purpose of the current study was not to carry out a comprehensive analysis of predictors of bail outcomes; such an analysis would require detailed information about each case and in-depth qualitative interviews with judges. Rather, the study sought to determine whether contemporary criminal justice systems can follow the mandate of law without giving into the increasingly dominant principles of the New Penology paradigm, as discussed by Feeley and Simon (1992).
We explored New Penological indicators of risk management, surveillance, and prediction of risk to public safety – what Garland (1995: 193) has called ‘postmodern sightings in the penal field’. This links legal standards – the ‘law on the books’ – to a larger exploration of the role of legal coercion in the control of ‘dangerous classes’. Because risk management and preventive detention in court decision making have become common throughout the United States, it was plausible to hypothesize that principles of the New Penology would have seeped into the New Jersey bail system despite this state’s Constitutional prohibition of preventive detention and its court rules against unnecessary detention. However, our findings generally suggest that New Jersey is indeed implementing the law as intended and resisting post-modern pressures in important ways. Relying on principles of the New Penology, it was plausible to expect that judges would jail arrestees from lower socioeconomic classes and grant release to defendants with more money who, presumably, are not members of ‘unruly classes’. However, these were not the findings that emerged from New Jersey.
Our indicator of socioeconomic status, operationalized as the type of counsel, suggests that judges make decisions that are less favorable to defendants from the higher social stratum. The odds of receiving ROR are reduced and higher bail amounts are imposed on individuals with private counsel, controlling for other legally relevant and individual indicators. Despite this ostensible attempt on behalf of judges to consider the defendant’s financial capabilities and presumably control the wealthier defendants by requiring them to pay higher amounts, the outcome of financial bail requirements is unsurprising: defendants with private counsel have increased odds of posting bail, no doubt because they indeed have the money. This does not disprove Goldkamp’s classic observation that there are ‘two classes of accused’, since poorer defendants were still less likely to be able to pay financial bail even when set at lower amounts. This neither proves nor disproves use of New Penological ideologies, since 20th-century researchers (Foote, 1954; Goldkamp, 1979) found that the poor were more likely to get prison when compared to their wealthier counterparts, even before New Penological narratives developed.
Feeley and Simon also predicted that control will become more stringent toward defendants from urban areas. However, in this study, county type (urban/non-urban) did not emerge as a significant predictor of bail decisions or outcomes. Moreover, contrary to the assumption that young males are subject to more control than people from other groups, age and gender did not significantly predict bail decisions. Finally, while past offending is an important factor in sentencing decisions, it is noteworthy that prior convictions did not emerge as a consistent predictor of bail decisions in New Jersey. While this variable was associated with higher bail amounts once the judge had decided to require financial bail, it was not a significant factor in the judges’ decision to grant own-recognizance relief or financial bail. This result is particularly interesting, as prior offending is often regarded as a proxy for recidivism risk and dangerousness, and also as one of the most influential factors in criminal justice decisions (Gottfredson and Gottfredson, 1990 [1988]).
Two legal variables consistently emerged as significant predictors of bail decisions. Judges tend to grant own-recognizance release and lower amounts of money bail to those charged with less serious offenses. Moreover, bail decisions are less favorable to defendants with a high number of charges. These results may indicate the court’s interest in the nature of the case and are consistent with New Jersey law; the court rules permit judges to consider offense seriousness and number of charges in making bail decisions. Information about the number of charges is a thin basis for determining potential danger and is an ambiguous indicator of dangerousness or flight risk. It may signal the seriousness with which prosecutors regard the cases and/or how they rank the defendants’ dangerousness. It may also be that judges associate an increasing number of charges with higher flight risks. It is difficult to disentangle the meaning of these legal indicators for New Penological narratives, as opposed to old-fashioned concerns about failure to appear for trial. Garland has recognized these ambiguities, arguing that postmodernist interpretation of contemporary penal practices is ‘undermined by… the indeterminacy and ambiguity of some of the key concepts… [and] the fact that certain purportedly postmodern characteristics are actually typical of modernity rather than distinct from it’ (Garland, 1995: 197).
The role of race
One key finding concurs with Feeley and Simon’s argument: the substantial impact of race on bail decisions, both in bail type granted (ROR versus financial) and bail amount required. The odds of being granted release on own recognizance were lower for minorities when compared to White defendants, even after controlling for other legal and sociodemographic indicators. Moreover, when financial bail was required, minority defendants were likely to have higher bail amounts imposed. These results support assertions made under the theory of the New Penology, namely that more stringent levels of coercive legal control are applied to minority groups. However, this observation is scarcely limited to the New Penological discourse, and raises even broader questions about the enduring role of race and its impact on public officials’ decision making.
While the race effect in itself is not a particularly novel finding, it is an important one particularly because decisions at the earliest stages of criminal processing significantly impact subsequent decisions and outcomes in the case. The role of race in bail decisions can be regarded as evidence of a New Penology or could be interpreted under the racial formation perspective, the focal concerns theory of judicial decision making, or the racial threat perspective, among other interpretive frameworks. Schlesinger (2005) argues that judges have insufficient time and information to make decisions and will therefore rely on legal factors as well as distinct attributions related to a defendant’s social characteristics during pretrial processing.
The current study does not allow us to ascertain whether New Jersey judges are imputing dangerousness to minority defendants or whether other factors are being considered. We do not know why judges do what they do, but we can empirically observe what they do: minorities consistently receive less advantageous decisions throughout the various stages of the bail process when compared to White defendants. This can potentially affect decisions rendered at subsequent phases of the case for minority defendants, such as the likelihood of receiving beneficial offers in return for guilty pleas (especially when the defendant is jailed prior to trial and is eager for release), or sentencing outcomes. The race effect is not, however, an observation that is unique to the New Penology.
Bail outcome
We found that the bail amount set by the court is strongly associated with the defendant’s ability to post bail: the higher the amount, the less likely the defendant to post bail. This result, however predictable, was important to an analysis of preventive detention because it showed that New Jersey judges did not set bail so high that defendants charged with violent offenses (those who would be preventively detained in other states) could not meet the financial requirements. Furthermore, although defendants with public defenders received more favorable bail requirements when compared to those with private attorneys, the former group was less likely to post financial bail if imposed; these amounts were not sufficiently low to enable defendants with public counsel to be released on bail. This dynamic is also reflected in the relationship between offense type and the ability to post bail: despite the higher bail amounts imposed, defendants charged with violent offenses were more likely to post bail when compared to defendants charged with other types of offenses. Further analyses showed that in cases where very high bail amounts were imposed, defendants charged with violent and drug offenses were more likely to post these high bail amounts when compared to those charged with property and other felony offenses. If the New Penology was evident in New Jersey bail practices, we would find that defendants charged with violent offenses would be more likely to be detained prior to trial. However, this is not the case. Very clearly, preventive detention is not used in New Jersey, either de jure or de facto, which is not consistent with New Penological predictions.
The outcomes of bail judgments (i.e. the ability to post bail) paint a different portrait from bail decisions. Despite the intentions of New Jersey judges, the bail system ultimately exerts a higher level of control over members of minority groups, poor but not necessarily urban, and charged with less serious offenses. While judges attempted to impose more favorable conditions for defendants with more limited financial resources as well as for individuals charged with less serious offenses, these groups were less likely to post bail and therefore more likely to be detained before trial. It may be that even the lowest bail amounts were too high for some defendants.
Conclusion
Revisiting these findings in light of the initial hypotheses, which rested on the presumption that the old penology continues to guide bail decision making in New Jersey, our first hypothesis is supported. All defendants, even those involved in the most serious forms of offending, were granted the opportunity for bail in one form or another. Second, if the law is implemented as intended, we hypothesized that bail amounts imposed in cases involving violent offenses would not be so high as to result in de facto detention. This is indeed what we observed in our data; in fact, defendants charged with violent offenses were more likely to post bail when compared to other defendants. Third, we hypothesized that sociodemographic characteristics (young, urban, minority males from lower socioeconomic strata) would not be subject to less favorable bail decisions under New Jersey law. With the exception of race, findings are generally in line with this hypothesis: county type (urban or non-urban), gender, and age were not significantly associated with bail decisions. Minority defendants were less often granted ROR and were required to pay higher bail amounts, though race did not emerge as a significant predictor of the ability to post bail. Moreover, it remains unclear whether race effects reflect an inclination toward the New Penology; other frameworks have also provided explanations for the over-representation of minorities in the criminal justice system. With regard to socioeconomic status, while defendants with a public attorney were less likely to post bail, judges appeared to grant more ROR and lower bail amounts to these individuals. In sum, our findings do not illustrate any blatant attempts to exert greater control over the young, poor, predominantly minority males who comprise the ‘urban underclass’. Finally, criminal history did not emerge as a consistent predictor of bail decisions and outcomes, suggesting that New Jersey judges are not making bail decisions solely on the basis of dangerousness.
In conclusion, we did not find compelling evidence to support the claim that the New Penology has seeped into the New Jersey legal system. New Jersey’s law does not appear to condone the rhetoric of dangerousness, which alone may be the most powerful indication that a postmodern New Penology is not inevitable. This raises the question about the extent to which the implementation of the New Penology depends on changes in law that support the paradigm. New Jersey has resisted the adoption of such legislation. However, Black and Hispanic defendants in New Jersey nevertheless fared worse than their White counterparts. It may be that the enduring effect of race on bail outcomes indicates that the blunt realities of poverty and race may trump ideology, both past and present.
Does law inevitably reflect and serve social and political trends? Can it resist these trends, or might it even be able to affect them independently? Our study suggests that, far from being inevitable, postmodern development can be shaped and even delayed by local political culture and laws. Old laws may have contemporary relevance; they can resist ideological shifts in important ways. It seems that in New Jersey, at least in regards to bail, law does matter after all.