The Unchosen: Procedural Fairness in Criminal Specialty Court Selection

Specialized criminal courts were created in an effort to offer nonpunitive responses to the commission of crime. The promise of these courts was that they would remove select populations from the traditional legal system and offer them something different, and perhaps better, than mere punishment and incapacitation.

However, the current selection processes for specialized courts—in which judges and prosecutors have almost completely unfettered discretion to decide both the criteria by which potential participants will be allowed to participate and whether any individual defendant meets such criteria—perverts the notion of providing specialized services to specific populations. Instead of selecting defendants based on neutral criteria, the discretionary bias inherent in the process invites judges and prosecutors to take extrajudicial considerations into account. These considerations result in the re-inscription of already existing privilege and, correspondingly, the reinforcement of biases that permeate much of the criminal legal system, such as those based on racial presentation.

Utilizing the framework of procedural fairness, this Article suggests that the presence of discretionary bias in the selection process for specialized courts threatens their continued legitimacy. Because specialized courts rely upon the freely-given cooperation of outside criminal legal stakeholders—such as social service providers and community organizations—in order to function, the courts must be perceived as a legitimate enterprise operating in an unbiased manner. Unless these courts begin to impose transparent, consistent, and procedurally just criteria for defendant selection, the entire project may cease to exist.

Introduction

Is it possible to conceive of a system in which the day that a factually guilty criminal offender is arrested is not one of the worst days of his life? In this imagining, perhaps it is the day that a drug addict gets admitted into free, intensive, in-patient drug treatment? Or the day that a lifelong sex worker who was trafficked as a teenager has her record expunged? It could be the day that a veteran suffering from PTSD has his symptoms recognized and is referred to a psychologist and a support group? This is the promise—although arguably unrealized—of specialized criminal courts.

Continue to imagine, for a second, that such a promise was real but was only available to certain defendants. And that those who were not selected were neither privy to the criteria underlying the decision nor given the opportunity to challenge the determination.

The scenario that you have just imagined happens every day in criminal courts around America: some defendants are promised something different—and perhaps better—than mere punishment, while others are not given any option other than to proceed through the undeniably dysfunctional, traditional criminal legal system.

For the last thirty years, specialized criminal courts have offered the promise of nonpunitive responses to the conditions underlying a defendant’s commission of crime. Thus, defendants who are diverted into specialized courts are, in effect, transferred out of a punishment-based system—with its array of collateral consequences—and into an assistance-based system.

Currently, judges and prosecutors serve as the gatekeepers to these courts. Astoundingly, the decision whether to offer a defendant diversion to a specialized criminal court is often made by a single judge or prosecutor. Sometimes the involvement of defense counsel is allowed, at other times not. Rarely is the defendant himself consulted about his eligibility. Selection criteria vary widely from court to court; in some, it is codified or at least discoverable. In others, the process is almost completely opaque and ad hoc, with the presiding judge and prosecutor’s office exercising virtually unfettered discretion. There are limited opportunities for advocacy preselection and no formal appeal rights upon denial.

Scholarship on judicial discretion has largely focused on sentencing. Similarly, much of the existing literature on prosecutorial discretion has focused on charging decisions and plea bargaining. Both apply primarily to traditional criminal court models. In a unique intervention, this Article examines the judicial and prosecutorial discretion underlying the selection of criminal defendants to be afforded the opportunity to have their cases adjudicated in specialized criminal courts.

This Article argues that the presumption that judicial and prosecutorial discretion in defendant selection will lead to just, or even desirable, outcomes is gravely flawed. Instead, the same harms that are attendant to exercises of discretion in charging, plea bargaining, and sentencing—racial disparities, social biases, and other forms of invidious discrimination—also inevitably appear in specialty court selection processes. Thus, relying on judicial and prosecutorial discretion in this arena can only replicate and intensify preexisting disadvantages among defendants: that is, the defendants who need diversion the least, because they are already better positioned for successful rehabilitation and reentry into society, are those most likely to be selected for it.

Complicating existing scholarship addressing the value and/or effectiveness of specialized criminal courts, this Article argues that, regardless of the quality of outcomes for defendants diverted into specialized courts, the model is to stay. Thus, the process through which defendants are selected should be unbiased, at least, and justifiable with reference to normative determinations, at best. Drawing from scholarship on procedural fairness, this Article argues that the perception—by defendants, the defense bar, treatment advocates, the social services community, and the public at large—that the selection processes are fair has an inherent practical value, and the current lack of constraints on discretionary bias leads to a perception of illegitimacy.

This Article’s normative suggestion is that judicial and prosecutorial selection discretion should be strictly limited, and that specialized courts should instead seek to adopt systemized selection models that presumptively favor the most disadvantaged defendants.

Part I recounts the rise of specialized criminal courts as alternative models to traditional criminal adjudication, exploring the promises of rehabilitation and reintegration into society offered by these models. Part II discusses existing thought on judicial and prosecutorial discretion and methods that have arisen for limiting discretion and (theoretically) reducing disparities arising from discretionary bias. Part III explores the ways in which discretionary bias impacts procedural fairness and raises challenges to the legitimacy of specialized courts. Lastly, Part IV proposes four models of specialized court selection that would constrain prosecutorial and judicial discretion and increase the perception of procedural fairness.

Procedural and administrative matters often play second-fiddle in scholarship regarding new or innovative approaches to the organization of criminal law. This is, simply put, a mistake. Proposed changes to substantive criminal statutes require, at the very least, legislative action, but also often involve forming coalitions of interested and powerful forces positioned to take up the cause of creating a movement to amend the law. In contrast, administrative or procedural “tweaks” are often a matter of simply implementing new court rules, a task that can sit in the hands of a single judge or judicial committee. Thus, a well-reasoned appeal for common sense rulemaking has the potential to generate better “bang for the buck” in terms of impact on the criminal legal system than more audacious suggestions for legislative reform. In the case of specialized criminal courts, improving the methods for selection could significantly impact the trajectory of tens of thousands of criminal defendants. And it all could be done with the sweep of a pen.

I. Specialization

A. Rise of Specialized Criminal Courts

The last thirty years have seen the rapid expansion of specialized courts—also referred to as “problem-solving” courts—within the criminal legal system.1See Allegra M. McLeod, Decarceration Courts: Possibilities and Perils of a Shifting Criminal Law, 100 Geo. L.J. 1587, 1605–07, 1610–11 (2012) (noting that, in 2012, there were more than 2,000 drug courts, “approximately 300 mental health courts, 200 domestic violence courts, thirty community courts, . . . and [over] 500 other specialized criminal courts”); Christine S. Scott-Hayward, Rethinking Federal Diversion: The Rise of Specialized Criminal Courts, 22 Berkeley J. Crim. L. 47, 51 (2017) (“Specialized criminal courts . . . have now become the focus of innovation at the front-end of the federal criminal justice system and appear to be the dominant form of diversion. These courts, which are variously called ‘alternative to incarceration’ programs, ‘court-involved pretrial diversion practices,’ and ‘diversion-based court programs’ now exist in at least 21 federal districts.” (footnotes omitted)). The earliest of these, and still perhaps the most recognizable model, was the drug-treatment court, where defendants charged with drug possession crimes could become eligible for nonpunitive resolutions to their cases, including referral to counseling services and admission to drug treatment programs.2See McLeod, supra note 1, at 1605–06.

States have expanded this specialized criminal court model to both offense-based domains and status-based defendants.3See Erin R. Collins, Status Courts, 105 Geo. L.J. 1481, 1484 (2017). New York State, as just one example, now has a Domestic Violence Court, a “Young Adult” Court, a Mental Health Court, a Human Trafficking Intervention Court, and multiple Community Courts, in addition to its Drug Court.4Problem-Solving Courts, N.Y. State Unified Ct. Sys., https://ww2.nycourts.gov/‌COURTS/‌problem‌_‌solving/‌index.shtml [https://perma.cc/‌4DBR-RAGC]. This Article does not specifically address the community court model, as its selection criteria often implicates a geographic catchment area as opposed to defendant-specific criteria. See Anthony C. Thompson, Courting Disorder: Some Thoughts on Community Courts, 10 Wash. U. J.L. & Pol’y 63, 85 (2002). And this list continually expands as additional populations are identified as worthy of being plucked from the traditional criminal legal system and its draconian sentencing structures and inability (or unwillingness) to provide rehabilitative services. The professed aim of specialized courts, in contrast to the traditional criminal adjudication model, is to offer a combination of treatment, monitoring, social services, and community service, in lieu of incarceration.5See discussion infra Section I.B; Collins, supra note 3, at 1484.

Eschewing the thinking that defendants warrant only punishment for having committed a crime6I use the phrase “only punishment” to acknowledge that scholars have compellingly argued that forcing defendants into therapy or drug treatment, for instance, is not only paternalistic, but can also be understood as punitive. See James L. Nolan, Jr., Redefining Criminal Courts: Problem-Solving and the Meaning of Justice, 40 Am. Crim. L. Rev. 1541, 1556 (2003) (“[S]imply altering the nomenclature . . . does not make therapeutic legal practices any less punitive.”).—and that courts are in the exclusive business of producing deterrence under the threat of punishment—the goal of specialized courts is problem-identification and resolution, almost always in collaboration with the community and relevant social service agencies.7See McLeod, supra note 1, at 1633 (discussing how courts employing a decarceration model “focus[] on deploying social structures separate from criminal law administrative components—such as local neighborhood networks, business organizations, and mental health, public health, job training, and other social services—to reduce criminal offending and to foster socially constructive citizenship behaviors”). As described by Allegra McLeod, the decarceration model of specialized courts “aims to identify those limited number of crimes for which criminal law intervention is most fitting,” which are left to be adjudicated in traditional courts, while “simultaneously . . . facilitating non-carceral responses to a range of other social ills.”8Allegra M. McLeod, Confronting Criminal Law’s Violence: The Possibilities of Unfinished Alternatives, 8 Unbound: Harv. J. Legal Left 109, 130 (2013). Specialized criminal courts “assign otherwise likely jail- or prison-bound defendants mental health and drug treatment, job and housing placement, along with other services in lieu of incarceration,”9McLeod, supra note 1, at 1595–96. and by doing so endeavor “to address the root causes of an individual’s involvement in the criminal justice system.”10Thompson, supra note 4, at 64.

While likely falling quite short of solving the underlying conditions leading to criminal behavior, specialized criminal courts can provide radical forms of relief for those who qualify. For example, New York State’s Human Trafficking Intervention Court11Human Trafficking Intervention Courts, N.Y. State Unified Ct. Sys., https://ww2.nycourts.gov/‌courts/‌problem‌_‌solving/‌htc/‌index.shtml [https://perma.cc/‌VEW9-XWFP]. New York also had a “prostitution diversion docket,” which predated the establishment of the human trafficking part. Id. created a comprehensive psychological assessment for defendants arrested for prostitution-related offenses, which informed judges as they made therapeutic and social service decisions.12See Katie Crank, Community Courts, Specialized Dockets, and Other Approaches to Address Sex Trafficking, in A Guide to Human Trafficking for State Courts 37, 43–44 (2014), http://www.htcourts.org/‌wp-content/‌uploads/Ch-‌2‌_‌140723‌_‌NACM‌_‌Guide‌_‌OnLineVersion‌_‌04.pdf [https://perma.cc/‌ZV3F-R3PX]. For defendants who can show that their arrests arose directly as a result of having been sex trafficked, the court can vacate all related charges—not just those currently in front of the court—but also earlier charges on the defendant’s record that can be linked to the trafficking.13See Edna Ishayik, Law Helps Those Who Escape Sex Trafficking Erase Their Criminal Record, N.Y. Times (Mar. 23, 2015), https://www.nytimes.com/​2015/​03/​24/​nyregion/​law-helps-those-who-escape-sex-trafficking-shed-its-stigma-too.html#:~:text=The%​20law%​2C%2​0​p​assed%​20in%202010,states%20have%20adopted%20similar%20statutes [https://perma.cc/​F2B8-PEYH] (discussing case of a Queens woman who had 129 convictions—mostly for prostitution and loitering—removed from her record). In another example, the Prostitution Court in Columbus, Ohio, offers a diversionary program that includes—in addition to counseling and supervision—a placement in safe housing.14Andrea Muraskin, For Victims of Sex Trafficking, A Therapeutic Court Provides a Way Out, WFYI Indianapolis (July 7, 2016), https://www.wfyi.org/‌news/‌articles/‌for-victims-of-sex-trafficking-a-therapeutic-court-provides-a-way-out [https://perma.cc/‌5EDJ-HA9B].

Today, most specialized courts offer iterations of the same forms of relief—counseling, connections to social service agencies, residential and outpatient drug treatment, and the like—but the possibilities are virtually endless. One can imagine a specialized criminal court that could order that a defendant receive an education in a trade, reunification with distant family, or medical treatment for a condition that may have limited employability. In this way, specialized criminal courts can constantly integrate new information about successful rehabilitative methods into their models in a way that traditional criminal courts cannot.15I make no claims that this continual revision is actually occurring, but the current model of the courts would allow it.

B. Specialized Court Models

Specialized criminal courts broadly fall under two models: offense-based and status-based.16The term “status courts” was coined by Erin Collins to describe “courts dedicated to offenders within a specific status group.” See Collins, supra note 3, at 1483. In the offense-based model, potentially eligible criminal defendants are screened for admission based on the nature of the defendant’s alleged criminal activity.17See David Jaros, Flawed Coalitions and the Politics of Crime, 99 Iowa L. Rev. 1473, 1505 n.164 (2014). The “oldest, most visible, widespread, and influential” example of this model is the drug court, where potential participants are identified based on having been accused of a drug-related offense.18Nolan, supra note 6, at 1542. For further discussion of drug courts and the discrimination that defendants often face, see Kalyn Heyen, Note, Drug Court Discrimination: Discretionary Eligibility Criteria Impedes the Legislative Goal to Provide Equal and Effective Access to Treatment Assistance, 43 Cardozo L. Rev. 2509 (2022) (also published in this Issue of Cardozo Law Review). Other examples include domestic violence courts,19Some experts have been hesitant to include domestic violence courts under the specialized court umbrella because the focus of these courts (vis-à-vis the defendant) can differ from other specialized courts. For an in-depth discussion, see Pamela M. Casey & David B. Rottman, Problem-Solving Courts: Models and Trends, Nat’l Ctr. for State Cts. 1, 4–5 (2003), https://ncsc.contentdm.oclc.org/‌digital/‌collection/‌spcts/‌id/‌169 [https://perma.cc/‌NCT5-MLP4]. DUI treatment courts,20DUI Courts, The Unified Jud. Sys. of Pa., https://www.pacourts.us/‌judicial-administration/‌court-programs/‌dui-court [https://perma.cc/‌83H2-U9LV]. and human (or sex) trafficking courts, which seek to resolve the cases of defendants who have criminal arrests related to having been sex trafficked, such as for solicitation or loitering.21While some scholars place sex trafficking courts under the rubric of “accountability courts,” I would argue that the newest iteration of sex trafficking courts, such as the model in New York State, which seeks to assist people who have been trafficked through holistic means similar to those in veterans court or young adult courts, should properly fall under the category of “status courts.” See generally Collins, supra note 3, at 1489–92.

In contrast, in the status-based model, candidates for diversion are screened on the basis of their identities or personal characteristics.22See id. at 1492. For example, veterans courts “address socially disruptive behavior on the part of veterans” based on the perception that many veterans experience post-traumatic stress disorder (PTSD), which leads to mental health and behavioral problems that are then criminalized.23McLeod, supra note 1, at 1608–09. Other examples of status-based courts include young adult and girls courts, which adjudicate juvenile offenses, and mental health courts, which deal with defendants who live with mental illness.24See Casey & Rottman, supra note 19, at 8; Collins, supra note 3, at 1492. Although some status-based courts may limit admission to only non-violent or petty offenders,25Casey & Rottman, supra note 19, at 7 (explaining some of the special issues involved in drug courts, including “[f]ederal guidelines that prohibit offenders from participating in drug courts if they have ever committed a violent offense”). others allow criminal defendants who have been accused of violent crimes,26McLeod, supra note 1, at 1608–09. including serious felonies.27See Kristine A. Huskey, Reconceptualizing “the Crime” in Veterans Treatment Courts, 27 Fed. Sent’g Rep. 178, 179 (2015).

In both models, the “moment” of diversion can occur either pre-adjudication or post-adjudication. In courts using a “deferred prosecution” system, selected defendants are transferred to the specialized court before pleading guilty to any charge or providing an allocution to any potentially inculpatory facts.28See Kevin S. Burke, Just What Made Drug Courts Successful?, 36 New Eng. J. on Crim. & Civ. Confinement 39, 41 (2010). In post-adjudication systems, the court requires the defendant to plead guilty to the pending charge (or stipulate to inculpatory facts)29See Thompson, supra note 4, at 73. before making treatment options available.30See Burke, supra note 28, at 41. Upon successful completion of the court-imposed conditions, the charges may then be dismissed. While the difference in these two systems matters a great deal to defendants and defense advocates (as it should), for the purposes of this Article, what is important is that, in either model, a successful defendant can proverbially walk out of the courthouse with no criminal record related to the charges for which they were arrested.31See id. (noting that the court “defers or suspends the defendant’s sentence while [they] participate[] in a drug-court program,” and if they “successfully complete[] the program, the sentence may be waived and the offense may even be expunged”); Thompson, supra note 4, at 73 (“In exchange for the defendant’s agreement to participate in the drug court process, the state holds the prosecution in abeyance with the agreement to dismiss the charges upon satisfactory completion of the process.”).

Defendants who are selected for, and opt into, specialized adjudication are afforded not only special forms of relief but also unique appellate rights.32See Burke, supra note 28, at 44–45 (“If a participant does not comply, for example by failing a drug test, they are not immediately kicked out of the program and sent back into the general criminal justice docket, but they may be reprimanded through admonishments in open court, increased testing, fines, mandatory community service, and escalating periods of jail confinement.”). Generally, the criminal legal system operates as a perverse game of Monopoly for defendants: get caught committing a crime, go directly to jail; be found to have violated parole, go directly to jail; fail to meet a condition of probation, go directly to jail. However, the more collaborative nature of the specialized court process, along with focus on the individual circumstances of the defendant and partnerships with expert agencies, all converge to challenge the zero-sum game that criminal defendants generally face. Thus, the structure of specialized courts often provides latitude to defendants who initially fail to follow the conditions imposed by the court.33See McLeod, supra note 1, at 1652; Thompson, supra note 4, at 75. See generally Crank, supra note 12.

Because a defendant’s relationship to the specialized court is more fluid than in the traditional model, the defendant (and thus defense counsel) is afforded multiple opportunities to plead the case for why they should be given a second, third, or fourth chance to remain in the court and, by extension, out of prison.34See McLeod, supra note 1, at 1652 (A judge at the New York City Midtown Community Court often uses graduated sanctions, such as mandating additional services or increasing the frequency of court visits to encourage compliance, rather than sentencing a defendant to jail. Throughout the process, defense counsel can continue to “play a role in safeguarding a defendant’s procedural rights and other interests . . . especially when motions for sanctions arise.”). Moreover, defendants who eventually fail to meet the conditions of the court, and are incarcerated as a result, are afforded the right to appeal on the terms of their initial specialized sentence.35That is the case unless they have signed an appeals waiver. For an example of the appellate framework often employed in specialized courts, see Tate v. State, 313 P.3d 274 (Okla. Crim. App. 2013); State v. Rogers, 170 P.3d 881 (Idaho 2007). As a practical matter, this means that appellate defense counsel can present less formally legalistic and more practical arguments on their client’s behalf, such as that the defendant should have been given a second chance in counseling, transferred to a different residential treatment center, or received a better‑tailored educational program.36These observations are based upon my past experience as an appellate criminal defense lawyer practicing in New York. In this way, the flexibility that these courts have in tailoring an initial program for a defendant extends to flexibility for the defendant to make his case to remain in the program.

Even if they are flagged as eligible, no criminal defendant is ever required to enter into a specialized court program. Participation is completely voluntary, and the ultimate decision, just as with any plea, lies with the defendant.37This is complicated in the context of mental health courts, where concerns have been raised about whether participants understand the voluntary nature of the diversion. See Casey & Rottman, supra note 19, at 46–49. All defendants retain their right to traditional methods of criminal adjudication, including a trial before a judge or jury.38This is to say that they retain that right unless they sign a waiver, which is required for admission to some specialized courts. See Eric Lane, Due Process and Problem-Solving Courts, 30 Fordham Urb. L.J. 955, 959 (2003) (“In each case, the defendant can refuse the alternative treatment.”); Trent Oram & Kara Gleckler, An Analysis of the Constitutional Issues Implicated in Drug Courts, 42 Idaho L. Rev. 471, 476 (2006) (“Participation in drug court programs is not mandatory.”). In this way, the option of being diverted into a specialized court is just that for a defendant—an additional option—not a final determination of the path down which his case will proceed.

C. Collaboration as a Central Component

Because specialized courts operate differently than traditional criminal courts, they require criminal legal stakeholders to engage in a substantially different process than is required in the traditional model. Instead of adversarialism, the lynchpin of traditional criminal courts, specialized courts operate under a collaborative model.39See Thompson, supra note 4, at 64 (explaining that drug courts, and other problem-solving courts that have evolved in their wake, “have all but abandoned conventional adversarial roles in the interest of providing a more therapeutic and less contentious environment for the resolution of issues”). In these courts,

Prosecutors and defense counsel engage in non-adversarial, team‑oriented roles designed to both support the judge and facilitate the progress of the defendant’s treatment. Instead of the adversarial contest characteristic of a traditional criminal proceeding, drug courts adopt an ethic of cooperation. . . . Rather than debating factual scenarios or legal implications, the principal players work together to determine the appropriate sanctions given the defendant’s circumstances.40Id. at 72 (footnote omitted).

The judge, who in the traditional legal system would assume a passive role overseeing the conduct of the prosecutor and defense counsel, instead assumes a more active—or, as some have characterized it, “invasive”—role, setting and regulating the terms of treatment throughout the entire period the defendant spends in that court.41See Eric J. Miller, Embracing Addiction: Drug Courts and the False Promise of Judicial Interventionism, 65 Ohio State L.J. 1479, 1492 (2004). Specialized court judges do not, at least initially, hand supervisory control over to agencies, corrections, or probation. Similarly, the prosecutor and defense counsel, relieved of their traditionally adversarial roles, “become partners collaborating in an effort to rehabilitate” the defendant for however long his case remains in the court.42Id.

External stakeholders, such as “treatment providers, law enforcement officers, probation officers, program coordinator[s], and case managers,” also actively participate as partners in the specialized court model, “operat[ing] as a team when addressing individual case issues.”43Casey & Rottman, supra note 19, at 44. Although not necessarily part of the active “team” in a defendant’s case, several other stakeholders also potentially play a role in the success of the specialized court enterprise: government agencies, which must provide additional funding and create an apparatus to monitor and support the relief granted by the court; social workers, counselors, and other medical professionals, who must agree to provide services under the observation and monitoring of the criminal legal system; community groups and civil rights organizations, who must support—or at least not undermine—the project of the courts; shelters, which provide housing options for defendants as part of court-ordered relief; police and police unions, who must agree to act in the interest of helping to ensure that a defendant can complete the programs ordered by the court; schools, daycares, churches and religious organizations, which are often the principle providers of social services in a region; and the media, to list a few.

Thus, for specialized courts to operate effectively, numerous stakeholders must buy into the model or, at the very least, not actively undermine it.44See Miller, supra note 41, at 1491 (“[D]rug court procedure embodies a non-adversarial partnership among the criminal justice, correctional, and treatment systems. This partnership ‘work[s] together to find care for defendants and to ensure that they remain in treatment.’” (alteration in original) (footnote omitted)). This not only includes the traditional or direct players in the criminal legal system; individuals and organizations that might naturally stand in tension—or even be at odds—with the traditional criminal adjudication system (and each other) must agree not only to support the goals of specialized courts but also to dedicate time and resources to assisting those courts. Without the cooperation of any of these parties, the model fails, regardless of the well-meaning intentions of any given judge or prosecutor.

D. Value Proposition

There is no great need to delve deeply into the shortcomings of the traditional criminal legal system.45See, e.g., Nolan, supra note 6, at 1541 (“Ubiquitous are complaints about overcrowded jails and prisons; the expense and burden of increasing court case loads; the ‘revolving door’ phenomenon of repeat offenders; the impersonal and assembly-line quality of ‘McJustice,’ or expedited case management; fatigue and job dissatisfaction among lawyers; the win-at-all-costs mentality of modern trial advocacy; and the adjudicative restrictions of hyper-proceduralism and mandatory minimum sentencing guidelines.”). For the purposes of this Article, however, the traditional criminal legal system’s primary negative constraint is that it is, by design, only able to offer simplistic punitive responses to what are almost always complex social, historical, and psychological problems.46See generally Michelle Alexander, The New Jim Crow: Mass Incarceration in the Age of Colorblindness (2010). In its most recent historical iteration, the traditional criminal legal system can offer only punishment, in its least punitive form, and incapacitation, in its most punitive form.47See Tom R. Tyler, Restorative Justice and Procedural Justice: Dealing with Rule Breaking, 62 J. Soc. Issues 307, 307 (2006) (“In the last several decades, America could perhaps best be characterized as a highly ‘punitive’ society. The focus of public attention has been on the need to punish rule-breakers and support has been high for harsh punishments for a wide variety of crimes, punishments including the death penalty and life in prison.” (citation omitted)). Any sentence short of permanent incapacitation also often carries with it a panoply of collateral consequences embodied in the form of a criminal record, which can continue to exert severe punishment on a defendant for decades after he has officially finished serving his time.48See Gabriel J. Chin, The New Civil Death: Rethinking Punishment in the Era of Mass Conviction, 160 U. Pa. L. Rev. 1789, 1790–92 (2012).

Moreover, because the traditional criminal legal system is primarily concerned with the crime—not the criminal49This is, at least in part, because, over time, there has been a “reconceptualization of the criminal offender as an individual who committed crimes because of independent, rational choice uninfluenced by extrinsic experiences or factors.” Collins, supra note 3, at 1484.— and because it only has one tool in its toolbox with which to address wrongdoing, it largely treats all people the same, employing a one-size-fits-all approach to adjudication.50See Thompson, supra note 4, at 64 (noting “the failure of traditional courts to address the individual circumstances of each offender’s life”).

In contrast, “[r]ather than seeking to punish and incapacitate criminals, problem-solving courts aim to address the deeper social issues that underlie many criminal cases by providing various services and incentives for defendants to improve their lives and avoid recidivating.”51Jaros, supra note 17, at 1504.

Scholars have raised more than a few substantive, ethical, and procedural critiques of the specialized court model. These critiques raise the possibility that diverted defendants may lose substantive constitutional rights that they would have had in the traditional system; that specialized court sentences can often result in a longer period of state supervision than would have resulted in the traditional system; that the model distracts from the structural factors that underlie crime, instead focusing on individual actions; and that these courts obscure the traditional roles of prosecutor, defense counsel, and judge in ways that could lead to the deprivation of a defendant’s rights.52See generally Oram & Gleckler, supra note 38; Thompson, supra note 4. Some scholars have also argued that these courts are simply not effective in meeting their stated rehabilitative goals.53See, e.g., Josh Bowers, Contraindicated Drug Courts, 55 UCLA L. Rev. 783, 786 (2008).

Setting these critiques to the side for a moment, admission to a specialized court has at least some value, at least for some defendants: defendants who successfully complete the conditions imposed by a specialized court are unquestionably better off than those who go through traditional courts in at least one concrete way—they can leave with no criminal record.54See McLeod, supra note 1, at 1632 (“Without requiring legislative repeal of particular criminal statutes, these courts provide a venue for suspending or dropping criminal charges in drug cases, a range of misdemeanor cases, and, in some instances, even in cases involving more serious felony charges as well as in a range of matters involving mentally ill offenders and veterans.”). Under the deferred prosecution model, charges are held in abeyance and dismissed upon successful completion, leaving no record of a conviction.55Thompson, supra note 4, at 73. Similarly, in the post-adjudication model, upon program completion, the court may agree to dismiss the charges56See Huskey, supra note 27, at 179; Nolan, supra note 6, at 1543; McLeod, supra note 1, at 1632. or expunge the defendant’s record.57See Burke, supra note 28, at 41 (“If the defendant successfully completes the program, the sentence may be waived and the offense may even be expunged.”).

While courts have been reluctant to acknowledge that collateral consequences of incarceration are, in fact, part and parcel of the punishment handed down at sentencing,58See Michael Pinard, An Integrated Perspective on the Collateral Consequences of Criminal Convictions and Reentry Issues Faced by Formerly Incarcerated Individuals, 86 B.U. L. Rev. 623, 630–31 (2006); Jeremy Travis, Invisible Punishment: An Instrument of Social Exclusion, in Invisible Punishment: The Collateral Consequences of Mass Imprisonment (Marc Mauer & Meda Chesney-Lind eds., 2002). it is undeniable that having a criminal record negatively affects a person in a whole host of ways, oftentimes for the rest of their life. A non-exhaustive list of collateral consequences includes deportation (for noncitizens), the deprivation of a citizen’s constitutional right to vote, up-to-lifelong public registry (for sex offenders), revocation of certain professional licenses, and ineligibility for public benefits.59Chin, supra note 48, at 1810.

Thus, admission to a specialized court constitutes what this Article will simply refer to as a “good,” meaning only that it is a thing that has some value.60I am not necessarily using the terms “good” or “allocation of goods” in the economic sense, but rather merely to convey the idea expressed. The question raised, then, is how this particular good is distributed.

II. Discretion

Much of the existing scholarship on specialized courts—particularly that which is critical—tends to treat these courts as if they are a transient phenomenon that will soon be exposed as ineffective (or worse) and dispensed with. However, there is no reason to believe that the criminal legal system responds to evidence of efficacy so quickly or surely.61See Michael C. Dorf & Jeffrey A. Fagan, Problem-Solving Courts: From Innovation to Institutionalization, 40 Am. Crim. L. Rev. 1501, 1501 (2003) (“The phenomenal growth of drug courts and other forms of ‘problem-solving’ courts has followed a pattern that is characteristic of many successful innovations: An individual or small group has or stumbles upon a new idea; the idea is put into practice and appears to work; a small number of other actors adopt the innovation and have similar experiences; if there is great demand for the innovation—for example, because it responds to a widely-perceived crisis or satisfies an institutional need and resolves tensions within organizations that adopt it—the innovation rapidly diffuses through the networks in which the early adopters interact. Eventually, what was originally an innovation becomes institutionalized.”). If these courts are going to exist and, as history suggests, continue to expand, then the procedural questions surrounding the selection of eligible defendants—or the allocation of the good of specialized courts—must be addressed.

A. Discretion in Specialty Selection Processes

The specialized court model confers an extraordinary amount of discretion on judges and prosecutors. This discretion includes the right to set general selection criteria for each court; to assign a weight to each factor; to determine whether any individual defendant meets the criteria; and to customize sentences based on the individual characteristics of each defendant.62Richard C. Boldt, Problem-Solving Courts and Pragmatism, 73 Md. L. Rev. 1120, 1147 (2014) (“[P]roblem-solving courts vest considerable discretion in the judges and other professionals who make crucial decisions with respect to the disposition of the criminal offenders subject to their jurisdiction.”). Yet, specialized court stakeholders have paid little attention to creating standardized, preannounced, and discoverable rules for defendant selection. Judge Kevin Burke, an early drug court judge, has commented that specialized courts “around the country operate in different ways and achieve a wide variety of outcomes. . . . [E]ach operates according to its own unique protocol” and creates its “own local legal culture.”63Burke, supra note 28, at 40. This observation is not only true of drug courts but of all specialized criminal courts.

Perhaps because of this inconsistency in selection protocols, much of the scholarship on specialized courts slips into the passive voice when discussing screening criteria and selection for admission: “[e]ligible participants are identified early and promptly placed in the Veterans Treatment Court Program;”64See, e.g., Huskey, supra note 27, at 179 (emphasis added). “the case is placed on a special docket with a dedicated judge and prosecutor;”65Tamar M. Meekins, Risky Business: Criminal Specialty Courts and the Ethical Obligations of the Zealous Criminal Defender, 12 Berkeley J. Crim. L. 75, 85 (2007) (emphasis added). “[w]hen a defendant is deemed eligible for a particular specialty court program or treatment alternative, the defendant’s case is placed on the specialty calendar.”66Tamar M. Meekins, “Specialized Justice”: The Over-Emergence of Specialty Courts and the Threat of a New Criminal Defense Paradigm, 40 Suffolk U. L. Rev. 1, 16 (2006) (emphasis added). This language bypasses the generative moment of a defendant’s experience with a specialized court—whether he is invited in or not.

The lack of attention paid to standardization can partially be attributed to the makeshift process through which these courts came into being. Although specialized courts are now ubiquitous throughout the criminal legal system, the creation of any individual court may have been driven by the efforts of a single judge within the jurisdiction who would then preside over the court.67See, e.g., Meekins, supra note 65, at 77 (“[N]ew specialty or problem-solving courts have been created by court administrators, judges, policymakers, and attorneys.” (footnote omitted)); Collins, supra note 3, at 1486 (“Some [specialty courts] come into existence through legislation, others by judicial fiat.”). For example, the Veterans Treatment Court in Buffalo, New York, was the brainchild and project of a single judge, as was the Mental Health Court in Hawaii.68See Huskey, supra note 27, at 178 (noting that the “first official VTC was established in Buffalo, New York, in 2008, by municipal Judge Robert Russell”); Casey & Rottman, supra note 19, at 8 (noting that Hawaiian Circuit Judge Marcia Waldorf started their mental health court). Even when these courts are created through legislative action, little attention is given to specifying the criteria that the courts will use to identify and select eligible defendants.69See, e.g., N.H. Rev. Stat. Ann. § 490-G:2 (2013) (stating that “participants (drug offenders) are identified early and promptly placed in the drug court program,” but outlining no method for identification, the stakeholder responsible for making the identification, or eligibility requirements).

In a report detailing the state of problem-solving courts, the National Center for State Courts (NTSC) highlighted variations across the most predominate models of specialized courts. The report noted jurisdictional variations in drug treatment courts, which included the “[p]arty or parties (e.g., drug court staff or coordinator, probation department, Treatment Alternatives to Street Crime agency, pretrial services agency, treatment provider, county health department) responsible for screening, assessment, case management, and treatment services.”70Casey & Rottman, supra note 19, at 7. Similarly, variations across mental health courts included the “[m]ental health criteria for program eligibility” and the “[o]ffense criteria for program eligibility.”71Id. at 9.

To add to the confusion, it is often unclear which parties are tasked with creating the selection criteria. In New York, for example, the Office of Court Administration advised the State’s “Sex Offense Courts [to] develop protocols to identify eligible cases” by “working with the stakeholders, such as the District Attorneys’ Offices and local law enforcement agencies, at the earliest possible stage in the court process.”72New York State Unified Court System Office of Policy and Planning, Sex Offense Courts: Mission and Goals, N.Y. Unified State Ct. Sys., https://www.Nycourts.Gov/​Courts/​Problem_Solving/​So/​Mission_Goals.Shtml#_Ftn1 [https://perma.cc/​9UFD-WE5S]. The Delaware prostitution court, known as the Victim Advocacy and Safety Enhancement project, took a similar approach—it brought together “court stakeholders” to implement “a screening tool tailored to trauma and prostitution issues.”73Crank, supra note 12, at 42. In both instances, the conception of stakeholders remains undefined but appears to exclude both defendants and the defense bar.

In contrast to New York and Delaware’s collaborative approach to creating sex trafficking screening criteria, in Portland, Oregon, “the Multnomah County prosecutor’s office gave one neighborhood-based prosecutor jurisdiction over all prostitution-related crimes,” who, therefore, had final say over admission criteria writ large and any decisions regarding individual defendants.74Id. at 44–45. Although perhaps surprising, a single judge or prosecutor having carte blanche over the gatekeeping functions for a specialized criminal court is not anomalous.75See, e.g., Colin Deppen, How Do Drug Courts Work? Are They Effective? A Look at the Revolution Sweeping the U.S. Legal System, PennLive (June 22, 2016, 8:46 PM), https://www.pennlive.com/​news/​2016/​06/​how_do_drug_courts_work_are_th.html [https://perma.cc/​43GD-RQ4J] (“In Potter County[, Pennsylvania], for example, district attorney Andy Watson said each possible entrant is discussed amongst local judges and law enforcement before a vote is taken on whether to admit them.”).

Related to the procedural issue of which parties are setting the admissions criteria is the substantive question of the content of those criteria; as one might anticipate, there is also little consistency in the requirements and restrictions set by different courts. In veterans treatment courts, for example, “there is substantial variation amongst the jurisdictions . . . . Some courts restrict eligibility to veterans who were honorably discharged . . . while others find eligible any defendant who has served in the military regardless of discharge status.”76Huskey, supra note 27, at 179. Other courts “are further restrictive and only allow veterans whose misconduct is specifically caused or related to combat trauma, PTSD, TBI, or other mental health issues,”77Id. (emphasis added). which is necessarily a highly subjective assessment.

Likewise, some mental illness courts also adopt the stance that the defendant’s mental illness must be determined to have “contributed to the commission of the offense.”78Nolan, supra note 6, at 1544 (quoting Arie Freiberg, Problem Solving Courts: Innovative Solutions to Intractable Problems?, 11 J. Jud. Admin. 7, 16 (2001)). Others have adopted more objective selection criteria, requiring only that the defendant have a documented history of mental illness.79McDaniel M. Kelly, Rehabilitation Through Empowerment: Adopting the Consumer-Participation Model for Treatment Planning in Mental Health Courts, 66 Case W. Res. L. Rev. 581, 591 (2015).

In many drug courts, the court will “accept defendants who have been charged with drug possession or another non-violent offense and who either tested positive for drugs or had a known substance abuse problem at the time of their arrest.”80Burke, supra note 28, at 41. However, in Washington, D.C., “defendants[] who have only tested positive for drugs on one occasion[] may not be eligible for entrance into drug court.”81Meekins, supra note 66, at 45. Instead, the court’s “guidelines specify that a defendant demonstrate a history of substance abuse, and the drug treatment professionals associated with this particular drug court do not believe that one positive test provides an adequate indication of addiction.”82Id. In this way, the court’s criteria require a selection screener to thinly slice the difference between an illegal drug user and an illegal drug abuser.

Some drug courts “exclude defendants with current or prior violent offenses.”83Burke, supra note 28, at 41. Still others will not divert defendants who are currently facing eligible charges if they have a “past, wholly unrelated offense,” regardless of whether violence was involved.84Id. (internal quotation marks omitted) (quoting Ryan S. King & Jill Pasquarella, Drug Courts: A Review of the Evidence 1 (2009)).

Informally—which is perhaps a distinction without a difference in this context—in limiting or opening selection to certain classes of defendants, judges often reference additional factors in the defendant’s favor, including the existence of family support structures, educational attainment, employment history, perceived “amenability to correction,” et cetera.85See, e.g., State v. Curry, 988 S.W.2d 153, 157–59 (Tenn. 1999). The Tennessee Supreme Court delineated the defendant’s “amenability to correction” and “potential for rehabilitation” as relevant considerations for a prosecutor’s decision whether to offer a defendant the option of pretrial diversion. See id.

In what are perhaps the most sweeping models of selection, some courts initially transfer all defendants charged with a particular offense to a specialized court. For example, the aforementioned New York stakeholders decided that all misdemeanor prostitution cases that continued past arraignment would be automatically transferred to the Human Trafficking Intervention Court; only after transfer would it be determined whether defendants continued in that part or were transferred back to the traditional court.86Human Trafficking Task Force e-Guide § 6.4, Human Trafficking Courts, Off. for Victims Crime Training & Tech. Assistance Ctr. [hereinafter Human Trafficking Courts], https://www.ovcttac.gov/​taskforceguide/​eguide/​6-the-role-of-courts/​64-innovative-court-responses/​human-trafficking-courts [https://perma.cc/​EBC5-N3H5].

Variations in screening procedures also extend to determinations about the moment of diversion and, like the variations in substantive admission criteria described above, are attributable to the fact that individual judges and prosecutors have outsized power to determine criteria. In this way, a specialized court’s decision to operate under a pre- or post-adjudication model can be based on the desires of the prosecutors who will practice within it:

During the initial design and development phase of a specialty court, when prosecution representatives require that pleas of guilty be instituted as a precondition of treatment, i.e., requiring that a jurisdiction embrace a post-adjudication model, they are acting as an adversary protective of its position in the controversy. The prosecutors argue for a post-adjudication model in order to protect their ability to go forward with prosecution should a defendant fail in treatment.87Meekins, supra note 66, at 47–48 (footnote omitted).

The effects of the centrality of the prosecutor’s role in selection cannot be understated. In addition to prosecutors often having unilateral power to make the “discretionary decision to divert a case from the traditional track in criminal court and to allow a defendant to participate” in a specialized court,88Thompson, supra note 4, at 80. many specialized courts “will only allow the defendant to enter treatment if the prosecution consents.”89Meekins, supra note 66, at 48. This grant of power to the prosecution, outside of the protections of the traditional adversarial system, challenges many of our precepts of fair justice. Prosecutors have a positionality and agenda that will not necessarily always align with the idea of nonpunitive rehabilitation of a criminal defendant. As scholars have pointed out, “[i]n determining whether to consent, prosecutors look at more than just the need for treatment or whether the defendant meets the program’s previously set guidelines.”90Id.

On an even more fundamental level, prosecutors are simply not tasked with representing the interests of criminal defendants—that is the job of defense counsel. And the defense bar is markedly peripheral to the creation of selection criteria or the implementation of screening procedures. The National Legal Aid and Defender Association (NLADA) has issued a guide entitled Ten Tenets of Fair and Effective Problem‑Solving Courts, which lays out aspirational rules for the operation of specialized courts.91See Ten Tenets of Fair and Effective Problem-Solving Courts, Nat’l Legal Aid & Def. Ass’n (2020), https://csgjusticecenter.org/​wp-content/​uploads/​2020/​08/​Ten​_​Tenets​_​Final​_​ACCD​_​version.pdf [https://perma.cc/​6BYY-HC2B]. First among these is the tenet that “[q]ualified representatives of the indigent defense bar shall have the opportunity to meaningfully participate in the design, implementation and operation of the court, including the determination of participant eligibility and selection of service providers.”92Id. (emphasis added). At the same time, NLADA acknowledged that “more often than not, defenders are excluded from the policymaking processes which accompany the design, implementation and on-going evaluation . . . of Problem Solving Courts.”93Id. There is no indication that this exclusion from the policy- and rule-making process has changed since NLADA’s guide was first issued.

This is not to imply that defense counsel is completely shut out of the hearing process determining whether an individual defendant will be admitted to one of these courts. But even when a specialized court allows defense participation, counsel has very limited—if any—opportunity to make meaningful arguments for why a defendant should be accepted into the program before the determination has been made.94See Burke, supra note 28, at 41 (“Deferred prosecution programs divert certain eligible defendants to the drug-court system before they plead to a charge. Post-adjudication programs, on the other hand, require a defendant to first plead guilty to the charge before making treatment options available. The drug court then defers or suspends the defendant’s sentence while he or she participates in a drug-court program. If the defendant successfully completes the program, the sentence may be waived and the offense may even be expunged. Defendants who fail to complete drug-court programs usually must return to the traditional criminal court for disposition of their criminal case.” (footnotes omitted)). By design, the decision whether to divert a defendant into a specialized court must be made at the very onset of the criminal case.95See Meekins, supra note 65, at 88 (“The decision to enter a guilty plea in order to receive treatment must be made early in a case . . . .”). This timing “puts enormous pressure on the defendant and defense counsel,” who must discuss and decide the issues related to diversion “before counsel has had time to investigate, research issues, file motions, or engage in significant discovery.”96Cait Clarke & James Neuhard, “From Day One”: Who’s in Control as Problem Solving and Client-Centered Sentencing Take Center Stage?, 29 N.Y.U. Rev. L. & Soc. Change 11, 30 (2004). Without having had time to obtain relevant facts or make informed arguments, defense counsel’s contribution cannot counterbalance the power that the court and prosecutor wield in the process.

Moreover, an invitation for defense counsel to participate in a discussion around selection does not equate to having any power to determine, or even sway, the ultimate outcome.97Allowing defense participation in discussions around admission should not be confused with the existence of a due process right. Almost by definition, if there are no set criteria for admission and no appeal structure for denials of admission, then any arguments made by defense counsel on behalf of a client are merely, in the words of a colleague, an exercise in “advance begging” in hopes of changing the discretionary decision of the judge or prosecution. The final decision in these courts rests completely with the presiding judge or prosecutor. As aptly stated by Eric Miller, “[t]he provision of a hearing is potentially useless unless the structure of that hearing is such that it ensures participation”—and I would add “meaningful” here—“in the process.”98Miller, supra note 41, at 1569. If, to the contrary, a hearing “has no potential to affect outcome, the perception that it is fair may be chimerical” as the existence of the hearing “does little more than exploit the cognitive biases that lead people to believe that merely by participating they can affect uncontrollable events.”99Josh Bowers & Paul H. Robinson, Perceptions of Fairness and Justice: The Shared Aims and Occasional Conflicts of Legitimacy and Moral Credibility, 47 Wake Forest L. Rev. 211, 250 (2012).

So, in sum, there is inconsistency in who sets selection criteria, inconsistency in the content of selection criteria, inconsistency in who determines whether any individual defendant meets those criteria, and inconsistency in the timing of when such decisions are made at all levels of specialized courts.

However, inconsistency is not necessarily negative. To the contrary, it is generally accepted that the law can benefit from heterogeneity. Especially in the context of evolving factual circumstances or procedural interventions, variation can be a mechanism through which the courts test the efficacy of different approaches. But, in the context of specialized courts, the inconsistency is not experimentation in service of finding a compelling model for the courts to eventually harmonize around. Instead, it is arbitrariness grounded in an overreliance on discretion.

B. Discretion in the Traditional System

As described above, the only consistent thread throughout specialized court selection processes is that prosecutors and judges have almost unlimited discretion to determine which defendants will be granted admission to these courts. But why might this be of concern? At the most basic level, discretion merely describes the ability of a prosecutor or judge to bring personal opinions, judgment, and experience to bear on legal decisions.

But prosecutorial and judicial discretion can play a dangerous role in the criminal legal system, a system that otherwise rigidly prescribes the behavior of both defendants and defense counsel.100For defendants, behavior is prescribed through defining impermissible activity and limiting freedom. And for defense counsel, it is prescribed through professional rules and obligations, on constitutional grounds (including claims for ineffective assistance of counsel), and by court rules. The threat of discrimination resulting from such discretion (discretionary bias) has been recognized in other contexts in the criminal legal system—most notably during the charging, plea bargaining, and sentencing phases of proceedings. And when discretion is not subject to any accountability mechanisms—as is the case for specialized courts—the results have often been that the decisions of judges and prosecutors have been reached discriminatorily.

1. Prosecutorial Discretion in Charging and Plea Bargaining

Prosecutors have something close to absolute discretion in initiating criminal proceedings and selecting the charges that will be filed against a suspect. It is in the prosecutor’s sole discretion whether to bring charges against the suspect, how much and which evidence to present to a grand jury to obtain an indictment, whether to offer a plea bargain (along with the terms of the offer), and whether to voluntarily dismiss a prosecution.101Lissa Griffin & Ellen Yaroshefsky, Ministers of Justice and Mass Incarceration, 30 Geo. J. Legal Ethics 301, 315 (2017); H. Mitchell Caldwell, Coercive Plea Bargaining: The Unrecognized Scourge of the Justice System, 61 Cath. U. L. Rev. 63, 77 (2011). While a particular prosecutor may be required to answer to their superior on these questions, there is virtually no actor outside of that prosecutor’s office to provide a check on the prosecutor’s decision-making processes.102This level of discretion, while criticized by scholars, has been justified by an interpretation of the separation of powers doctrine. See Robert Heller, Comment, Selective Prosecution and the Federalization of Criminal Law: The Need for Meaningful Judicial Review of Prosecutorial Discretion, 145 U. Pa. L. Rev. 1309, 1326 (1997).

Although a layperson would likely be shocked to realize that a major component of our criminal legal system operates under such lax oversight, there is no reason to believe that the courts find the prosecutors’ “super powers” to be at all objectionable.103I borrow the term “super powers” from Paul Butler, who has used the term to describe the unfettered authority the courts ceded to police that encourage racially disparate treatment. See Paul Butler, The System is Working the Way It is Supposed to: The Limits of Criminal Justice Reform, 104 Geo. L.J. 1419, 1451 (2016). To the contrary, “[t]here is a broad and rather casual acceptance of the fact that prosecutors often exercise greater control over the administration of criminal justice than do other officials.”104James Vorenberg, Decent Restraint of Prosecutorial Power, 94 Harv. L. Rev. 1521, 1522 (1981). That acceptance has not been significantly challenged in the modern era—instead, with the increase of limitations on judicial discretion, prosecutorial discretion has blossomed even more robustly.105See Donald A. Dripps, Overcriminalization, Discretion, Waiver: A Survey of Possible Exit Strategies, 109 Penn State L. Rev. 1155, 1161–65 (2005).

Notionally, the courts can have some—though very limited—say in regulating prosecutorial decisions. But practically, there is very little probability that a prosecutor’s discretionary decisions will ever be disturbed post hoc. Courts have articulated their understanding that:

This broad discretion rests largely on the recognition that the decision to prosecute is particularly ill-suited to judicial review. Such factors as the strength of the case, the prosecution’s general deterrence value, the Government’s enforcement priorities, and the case’s relationship to the Government’s overall enforcement plan are not readily susceptible to the kind of analysis the courts are competent to undertake. Judicial supervision in this area, moreover, entails systemic costs of particular concern. Examining the basis of a prosecution [namely, discretionary decisions] . . . threatens to chill law enforcement by subjecting the prosecutor’s motives and decisionmaking to outside inquiry, and may undermine prosecutorial effectiveness by revealing the Government’s enforcement policy. All these are substantial concerns that make the courts properly hesitant to examine the decision whether to prosecute.106Wayte v. United States, 470 U.S. 598, 607–08 (1985).

Thus, absent some evidence of a malicious intent in charging, the prosecutor’s decisions regarding charging, presentation, and plea bargaining will never be upset. Such determinations, the courts have decided, are somehow outside of their expertise107This assumes a level of autonomy between the court and prosecutors that many practitioners might view skeptically. and thus are “particularly ill-suited to judicial review.”108Wayte, 470 U.S. at 607.

Solidifying the grant of this super power to prosecutors, courts have held that prosecutors are allowed to make charging decisions that have a discriminatory impact, as long as there is no intent to discriminate;109See McCleskey v. Kemp, 481 U.S. 279, 291–92, 312–13 (1987); cf. Washington v. Davis, 426 U.S. 229 (1976). that there is no requirement that prosecutors offer similar plea terms for comparable crimes or offenders; and that, even after negotiating a plea bargain, the prosecutor has sole discretion to decide whether a defendant has sufficiently complied with the terms such that the case will conclude.110Teah R. Lupton, Prosecutorial Discretion, 90 Geo. L.J. 1279, 1280 n.637 (2002).

Advocates for this level of discretion have argued that “[s]ignificantly curtailing prosecutorial discretion would accomplish consistency at the cost of individualized justice.”111Kenneth J. Melilli, Prosecutorial Discretion in an Adversary System, BYU L. Rev. 669, 674 (1992). This line of reasoning misses the perhaps obvious point that individualization is not always the antecedent to justice or equality. Sometimes it is, instead, the variable through which invidious discrimination can manifest.

While prosecutors themselves are not required to document the disparate impact of their charging decisions or even explain on the record the rationale for a charging or plea decision,112Griffin & Yaroshefsky, supra note 101, at 331; see Vorenberg, supra note 104, at 1554–55 (“Because prosecutors make decisions that can determine conviction and punishment, it is fair to test their process of decision against the standards imposed on other officials who make similarly critical judgments. Yet prosecutors are not held to anything remotely like what due process would require if they were engaged in an acknowledged rather than a hidden system of adjudication. No uniform, pre-announced rules inform the defendant and control the decisionmaker; a single official can invoke society’s harshest sanctions on the basis of ad hoc personal judgments. Prosecutors can and do accord different treatment—prison for some and probation or diversion for others—on grounds that are not written down anywhere and may not have been either rational, consistent, or discoverable in advance.” (footnote omitted)). quantitative and qualitative data from other sources show that unfettered prosecutorial discretion can, and does, lead to inequitable results.113See Developments in the Law: Race and the Criminal Process, 101 Harv. L. Rev. 1472, 1627, 1630–32 (1988) for a discussion of racial disparities in prosecutorial charging decisions. These findings will be discussed further in the following Sections.

2. Judicial Discretion in Sentencing

As Justice Breyer acknowledged in Blakely v. Washington, under indeterminate sentencing schemes, which provided the highest level of sentencing discretion to judges, “[t]he length of time a person spent in prison appeared to depend on ‘what the judge ate for breakfast’ on the day of sentencing, on which judge you got, or on other factors that should not have made a difference to the length of the sentence.”114Blakely v. Washington, 542 U.S. 296, 332 (2004) (Breyer, J., dissenting). As a result of this level of discretion, sentences for identical convictions within the same jurisdiction could vary wildly, and race all too often played a role in unfairly disparate sentencing.115Id. (“citing . . . studies indicating that, before the United States Sentencing Commission Guidelines were promulgated, punishments for identical crimes in the Second Circuit ranged from 3 to 20 years’ imprisonment and that sentences varied depending upon region, gender of the defendant, and race of the defendant”); see also Developments in the Law: Race and the Criminal Process, supra note 113, at 1627 (discussing racial disparities in noncapital sentencing).

Regardless of the particular sentencing regime under which a court operates, judicial discretion will always play a central role in sentencing decisions. As illustrated by Frank O. Bowman:

If conviction of Crime X generates a range of possible penalties from which a judge may choose, then a judge sentencing defendants convicted of Crime X can either declare that all persons convicted of Crime X in his courtroom will receive the same penalty or try to distinguish among those who have committed Crime X. If he takes the latter course and does so on any basis other than a lottery, he must identify—at least in his own mind—facts that distinguish the case before him from the universe of other cases involving convictions of Crime X. The facts deemed important by the judge might be facts about the offender (age, prior criminal record, prior good works, family ties, and the like) or facts about the offense that make this instance of Crime X more or less troublesome than other instances (violence, quantity of drugs, amount of loss, role in the offense, and so forth).116Frank O. Bowman, III, Debacle: How the Supreme Court Has Mangled American Sentencing Law and How It Might Yet Be Mended, 77 U. Chi. L. Rev. 367, 436–37 (2010).

Thus, unless discretion is completely excised from the model (e.g., all persons convicted of Crime X receive the same penalty regardless of mitigating or aggravating circumstances), it will necessarily continue to play an enormous role in sentencing decisions. And just as in the prosecutorial context, judicial discretion in sentencing has been shown to lead to disparities in how comparable defendants fare in the system.117Besiki Kutateladze, Vanessa Lynn & Edward Liang, Do Race and Ethnicity Matter in Prosecution? A Review of Empirical Studies, Vera Inst. of Just. 14–16 (2012), https://www.vera.org/​downloads/​Publications/​do-race-and-ethnicity-matter-in-prosecution-a-review-of-empirical-studies/​legacy_downloads/​race-and-ethnicity-in-prosecution-first-edition.pdf [https://perma.cc/​9E7S-LX57] (examining thirty-four empirical studies on prosecution and race). Senator Kennedy, in championing sentencing reform, emphasized that “judicial discretion worked to the disadvantage of those already disadvantaged by birth and social condition.”118Jaros, supra note 17, at 1492.

C. Extralegal Factors Involved in Discretionary Decision Making

Certainly, discretion is not necessarily inherently problematic, nor would it be possible to completely eliminate it from the criminal legal system, even if it were. But exercises in discretion can all too easily veer from the reasoned consideration of extrajudicial/extralegal factors to discretionary bias and discrimination. Research has shown that certain extralegal factors can influence a defendant’s outcome when discretion plays a role in the determination.

1. Racial Bias

Unsurprisingly to anyone familiar with the criminal legal system, racial disparities abound in almost every sphere where judicial and prosecutorial discretion play a role in determining outcomes.119This is not to imply that racial disparities are not also present in those areas where the jury, or other stakeholders, determine outcomes. These disparities appear most starkly at those stages of a criminal prosecution in which prosecutors and judges exercise virtually unchecked discretion.

At the charging stage, for example, empirical studies have correlated race with the severity of charges that a prosecutor initially brings against a criminal defendant.120Jerry Kang et al., Implicit Bias in the Courtroom, 59 UCLA L. Rev. 1124, 1140, 1142 (2012). Studies have also found that charges are more often dropped against nonwhite defendants than their white counterparts, which academics have posited is likely evidence that police arrest nonwhite defendants with less of an evidentiary basis than whites, and prosecutors pursue those unsupported charges to a greater extent than with white defendants.121John Hagan & Ruth D. Peterson, Criminal Inequality in America: Patterns and Consequences, in Crime and Inequality 14, 29 (John Hagan & Ruth D. Peterson eds., 1995). However, this statistic is perhaps also skewed by the likelihood that even in these insufficient evidence scenarios, prosecutors may still elicit a plea bargain from the defendant, in which case the relative evidentiary strength of the charges would be rendered invisible.

At the plea bargaining stage, controlled studies have shown that nonwhite defendants receive plea bargains less often than white defendants, and that the plea offers they receive are less favorable.122Id. A 2017 study of criminal cases in Wisconsin, for example, revealed that white defendants were twenty-five percent more likely than Black defendants to have their top charge dropped or reduced by prosecutors.123Carlos Berdejó, Criminalizing Race: Racial Disparities in Plea-Bargaining, 59 B.C. L. Rev. 1187, 1191 (2018) (analyzing more than 48,000 misdemeanor and felony cases in Wisconsin). Astonishingly, the data also showed that white defendants were nearly seventy-five percent more likely than Black defendants to see all misdemeanor charges carrying a potential sentence of incarceration dropped, dismissed, or amended to lesser charges.124Id.

At the level of judicial discretion, studies have found that Black defendants are more likely to have bail set at a higher amount and to be detained pending trial.125Cassia Spohn, Race, Sex, and Pretrial Detention in Federal Court: Indirect Effects and Cumulative Disadvantage, 57 U. Kan. L. Rev. 879, 888–89 (2009); see Jon Kleinberg, Himabindu Lakkaraju, Jure Leskovec, Jens Ludwig & Sendhil Mullainathan, Human Decisions and Machine Predictions, 133 Q.J. Econ. 237, 240–42 (2018). Judges’ decisions about who to release on pre-trial bail do not comport with algorithmic predictions regarding risk; in fact, many of the defendants that judges release fail to appear for trial and go on to commit new crimes. However, “a properly built algorithm can reduce crime and jail populations while simultaneously reducing racial disparities. In this case, the algorithm can be a force for racial equity.” Id. at 241. At sentencing, judges incarcerate Black defendants more frequently and sentence Black defendants to terms that are significantly longer than similarly situated white defendants.126U.S. Sent’g Comm’n, Demographic Differences in Sentencing: An Update to the 2012 Booker Report 2 (2017) [hereinafter Demographic Differences], https://www.ussc.gov/​sites/​default/​files/​pdf/​research-and-publications/​research-publications/​2017/​20171114_Demographics.pdf [https://perma.cc/​8MK6-JYS8]; see also Berdejó, supra note 123, at 1194. In the federal system, Black defendants are less likely to receive a downward departure from sentencing guidelines than white defendants.127Demographic Differences, supra note 126, at 2, 20.

Appallingly, research has shown that racial bias can play a role in sentencing even as between Black defendants. More than one study has shown that perceptions that a defendant has “Afrocentric features” can have a negative effect on sentencing decisions.128Friederike Funk & Alexander Todorov, Criminal Stereotypes in the Courtroom: Facial Tattoos Affect Guilt and Punishment Differently, 19 Psych. Pub. Pol’y & L. 466, 466 (2013).

None of this is to imply that prosecutors and judges are engaging in overt or conscious racial discrimination—just as none of this is to say that they are not.129See Kang et al., supra note 120, at 1172 (noting that, interestingly, judges overwhelmingly view themselves as exceptionally able to “avoid[] racial prejudice in decision making” even when data shows otherwise (quoting Jeffrey J. Rachlinski, Sheri Lynn Johnson, Andrew J. Wistrich & Chris Guthrie, Does Unconscious Racial Bias Affect Trial Judges?, 84 Notre Dame L. Rev. 1195, 1225 (2009))). In the arena of discretionary bias, the intent behind the discrimination is of no consequence to the defendants who are subject to it—it is the effect that matters. While the various manifestations of discrimination described in this Section may be a result of malice or apathy (or something else), they are, in the final analysis, structural and systemic in nature. Although criminal justice law and policy have morphed and evolved over the years, a maxim coined more than half a century ago remains as true today: “Wherever discretion exists in the legal system, it is likely to be exercised against [B]lacks.”130Norman Dorsen & Leon Friedman, Disorder in the Court: Report of the Association of the Bar of the City of New York, Special Committee on Courtroom Conduct 20 (1973).

2. Social Biases

While racial bias is a familiar lens through which to view the perils of discretion in the criminal legal system, other, perhaps more unconscious, biases also play a role in how certain defendants are treated in comparison to others. For example, innocent-looking people are likely to receive better outcomes in the criminal legal system than those who are perceived to look like criminals. Researchers have found that defendants with “untrustworthy-looking faces” are likely to fare less well in criminal courts than those with more trustworthy features.131Funk & Todorov, supra note 128, at 466. Similarly, “baby faced” defendants fare better than those who look more mature.132Id.

Attractiveness also plays an outsized role in criminal outcomes.133This is quite an understatement, as it should not play any role in criminal outcomes. See David L. Wiley, Beauty and the Beast: Physical Appearance Discrimination in American Criminal Trials, 27 St. Mary’s L.J. 193, 211–12 (1995). Attractive defendants fare better than unattractive ones.134Id. In simulations, research participants recommended lighter sentences for attractive defendants and heavier sentences for those perceived as unattractive, regardless of the severity or nature of the crime.135Id. at 212–14.

A defendant’s style choices (for lack of a better term) can also affect the determination of his guilt and the length of his sentence for a crime. A primary example of this is a defendant’s decision to display visible tattoos.136Funk & Todorov, supra note 128, at 466. Just like physical appearance and race, visible tattoos trigger stereotypes—quite often incorrect, but at any rate extralegal—of a defendant’s character and thus his general proclivity for criminality.137See John M. Hagedorn & Bradley A. MacLean, Breaking the Frame: Responding to Gang Stereotyping in Capital Cases, 42 U. Mem. L. Rev. 1027, 1034–35 (2012).

The initial perceptions of judges and prosecutors, just like those of jurors, are influenced by how they stereotypically understand the guilty to present.138See, e.g., Benjamin Weiser, A Judge’s Education, a Sentence at a Time, N.Y. Times (Oct. 7, 2011), https://www.nytimes.com/​2011/​10/​09/​nyregion/​judge-denny-chin-of-federal-court-discusses-sentencing.html [https://perma.cc/​622J-MGVZ] (discussing the factors Judge Denny Chin has considered in sentencing decisions, including that a defendant “seemed bright and articulate”). While it would be difficult to replicate these studies with real world judges and prosecutors, it is quite common to hear judges talk about a defendant’s physical appearance (e.g., clean cut), especially when handing down a lenient sentence. But taking even the smallest step back reveals that such considerations should play no part in a fair criminal legal system.

D. Efforts to Constrain Judicial and Prosecutorial Discretion

It has long been acknowledged that unconstrained discretion in the traditional legal system frequently generates racial disparities. Over the past several decades, as the prison population has exploded in the United States—and criminal legal stakeholders, politicians, and the general public have started to perceive mass incarceration as a problem to be solved, not a solution to a problem—stakeholders have attempted to impose methods to constrain prosecutorial and judicial discretion. These constraints arose both out of a concern for American exceptionalism in incarceration practices and as a direct response to discretionary bias in charging, plea negotiations, and sentencing. Methods of constraint included the issuance of charging guidance, the promulgation of sentencing guidelines, and the adoption of risk assessment tools. None of these approaches has managed to constrain discretion sufficiently so as to solve the bias problem.

1. Charging Guidance

The United States Attorney General, along with state district attorneys, have periodically issued charging guidance that is binding on the prosecutors operating under their respective banners. Other entities, such as the American Bar Association (ABA) or the Department of Justice, also issue guidance outlining certain standards and recommending—but only recommending—best practices in charging and negotiating plea bargains. Much of this guidance has directly addressed discretionary sentencing practices resulting in racial disparities.

As an example, in 2010, Attorney General Eric Holder issued a charging and sentencing memorandum to federal prosecutors.139Off. of Att’y Gen., U.S. Dep’t of Just., Memorandum to All Federal Prosecutors: Department Policy on Charging and Sentencing (2010). In that memo, he noted that “[u]nwarranted disparities” may have arisen in charging decisions, and he instructed prosecutors that “[p]ersons who commit similar crimes and have similar culpability should, to the extent possible, be treated similarly” and “without unwarranted consideration of such factors as race, gender, ethnicity, or sexual orientation.”140Id. at 1. Later, in 2013 and 2014, Holder issued additional charging guidance directed at reducing charging mandatory minimum–triggering offenses for nonviolent drug crimes and dissuading prosecutors from using the threat of sentencing enhancements to extract plea bargains.141Off. of Att’y Gen., U.S. Dep’t of Just., Memorandum to the United States Attorneys and Assistant Attorney General for the Criminal Division: Department Policy on Charging Mandatory Minimum Sentences and Recidivist Enhancements in Certain Drug Cases (2013); Off. of Att’y Gen., U.S. Dep’t of Just., Guidance Regarding § 851 Enhancements in Plea Negotiations (2014).

In 2017, newly appointed Attorney General Jeff Sessions issued a superseding charging and sentencing policy memorandum.142Off. of Att’y Gen., U.S. Dep’t of Just., Memorandum for All Federal Prosecutors: Department Charging and Sentencing Policy (2017). The Sessions memo directed prosecutors to charge defendants with the offense that “carr[ies] the most substantial guidelines sentence, including mandatory minimum sentences,” and categorized any lesser charge as an “exception” to the “core principle[s]” of the Justice Department that required supervisory approval.143Id. Thus, while giving lip service to prosecutorial discretion (noting “great confidence in our prosecutors and supervisors to apply [the directives] in a thoughtful and disciplined manner”), the Sessions charging memo stripped away a large part of what prosecutors have traditionally understood to fall under the purview of their discretion.144Id.

The ABA has also issued guidance to prosecutors in the form of its Criminal Justice Standards for the Prosecution Function. These standards include guidance about bias in charging, stating that a “prosecutor should strive to eliminate implicit biases, and act to mitigate any improper bias or prejudice” and that the “prosecutor’s office should be proactive in efforts to detect, investigate, and eliminate improper biases, with particular attention to historically persistent biases like race, in all of its work.”145Am. Bar Ass’n, Standard 3-1.6 Improper Bias Prohibited, in ABA Standards for Criminal Justice: Prosecution Function (4th ed. 2017). The ABA standards do not constitute binding guidance upon any prosecutor, create any right of action, or otherwise have the force of law; they are instead “aspirational” and describe “best practices.”146Am. Bar Ass’n, Standard 3-1.1 The Scope and Function of These Standards, in ABA Standards for Criminal Justice: Prosecution Function (4th ed. 2017).

Because of the transient nature of charging policy memos (i.e., the substance of the guidance is dependent on the issuer and frequently changes, as discussed above) and the nonbinding nature of other forms of guidance, the likelihood that these documents make a significant impact on discretionary bias is slim.147And the impact of these documents is very difficult to track, as prosecutors are not required to collect data on racial disparities in charging and plea bargaining.

2. Sentencing Guidelines

Mandatory sentencing guidelines received bipartisan support when they were initially introduced. Both political conservatives and liberals recognized that judicial discretion in sentencing posed a problem, although they certainly did not agree about the nature of that problem. “[L]iberals opposed indeterminate sentencing because they believed judicial discretion led to unjust sentencing disparities and racial bias, while conservatives argued that discretionary sentencing allowed lenient judges to give criminals inappropriately light sentences.”148Jaros, supra note 17, at 1491.

In quick succession, on both the state and federal level, a litany of mandatory sentencing bills was passed, “incorporating ideas such as the death penalty, ‘three strikes,’ mandatory minimums, victims’ bill of rights, and ‘truth in sentencing.’”149Clarke & Neuhard, supra note 96, at 26. For many proponents of guidelines, “the abolition of individualized sentencing was supposed to lead to uniform proportionate sentences of limited severity, an increased use of non-incarcerative penalties, and a reevaluation of the criminal code.”150Jaros, supra note 17, at 1492. This uniformity was also, not insignificantly, meant to diminish, if not extinguish, racial disparities in sentencing.151See id. at 1491.

The passage of time has shown that the implementation of mandatory guidelines achieved virtually none of the intended goals. Sentences, at least on paper, appeared both harsher and, at the same time, perhaps, more fairly imposed.152See Michael Tonry, Federal Sentencing “Reform” Since 1984: The Awful as Enemy of the Good, 44 Crime & Just. 99 (2015). And sentences certainly did become harsher, oftentimes to the point of absurdity.153See, e.g., Erwin Chemerinsky, Cruel and Unusual: The Story of Leandro Andrade, 52 Drake L. Rev. 1 (2003). However, in practice, both prosecutors and judges remained able to manipulate potential sentences through charging decisions and discretion in applying exceptions and sentencing departures.154See id. at 7 (“Giving tremendous discretion to prosecutors is characteristic of the three strikes laws and recidivist sentencing laws all over the country. It is also characteristic of federal sentencing guidelines giving prosecutors tremendous discretion in charging, which then influences the sentence received.”). As a result, racial disparities in sentencing did not see any significant decrease.155See Hum. Rts. Watch, Decades of Disparity: Drug Arrests and Race in the United States 16 (2009), https://www.hrw.org/​sites/​default/​files/​reports/​us0309web_1.pdf [https://perma.cc/​PS5K-9Y9V] (“Among [B]lack defendants convicted of drug offenses, 71 percent received sentences to incarceration in contrast to 63 percent of convicted white drug offenders. Human Rights Watch’s analysis of prison admission data for 2003 revealed that relative to population, [B]lacks are 10.1 times more likely than whites to be sent to prison for drug offenses.”); Am. Civ. Liberties Union, Written Submission of the American Civil Liberties Union on Racial Disparities in Sentencing (2014), https://www.aclu.org/​sites/​default/​files/​field_document/​141027_iachr_racial_disparities_aclu_submission_0.pdf [https://perma.cc/​5H7T-RKS6]. For these reasons, mandatory sentencing guidelines have roundly been considered a failed project,156See Griffin & Yaroshefsky, supra note 101, at 311 (“[D]ata shows that the proliferation of mandatory minimum sentencing and the exploding prison population coincide . . . .”). both by scholars and the general public, and have fallen out of use in recent years.157See United States v. Booker, 543 U.S. 220 (2005).

3. Risk Assessment Tools

In the last decade or so, jurists have begun to look to new methods of constraining judicial sentencing discretion.158See Sonja B. Starr, Evidence-Based Sentencing and the Scientific Rationalization of Discrimination, 66 Stan. L. Rev. 803, 811 (2014). This same period marked the rise of big data, which sought to systematically utilize the vast stores of publicly available information—in this case, on patterns of crime commission and characteristics of arrestees and defendants.159See J. C. Oleson, Risk in Sentencing: Constitutionally Suspect Variables and Evidence-Based Sentencing, 64 SMU L. Rev. 1329, 1343–44 (2011). It was out of this confluence that judges began employing risk-assessment tools, which are designed to “take information on recidivism rates for groups and use them to estimate the risk of recidivism for individuals possessing those same group characteristics.”160Dawinder S. Sidhu, Moneyball Sentencing, 56 B.C. L. Rev. 671, 673 (2015).

Risk assessment tools, which are generally propriety programs sold to court systems, are not intended to mechanically deliver a complete sentencing recommendation. Instead, the tools’ calculated risk of recidivism is used by the judge as one factor in sentencing, alongside presentence reports generated by departments of correction/probation, sentencing recommendation reports provided by the parties, and the judge’s own observations and insights.161Recent Case, Criminal Law—Sentencing Guidelines—Wisconsin Supreme Court Requires Warning Before Use of Algorithmic Risk Assessments in Sentencing.—State v. Loomis, 881 N.W.2d 749 (Wis. 2016), 130 Harv. L. Rev. 1530 (2017). As of this writing, most states use risk assessments, in some manner, in their sentencing regimes.162See Melissa Hamilton, Back to the Future: The Influence of Criminal History on Risk Assessments, 20 Berkeley J. Crim. L. 75, 88 (2015).

Judges were attracted to risk assessments because of their patina of objectivity and ostensible scientific underpinning. Presumably, they were tired of critiques like Justice Breyer’s suggesting that their sentencing decisions were not merely discretionary but were arbitrary and often discriminatory.163See Blakely v. Washington, 542 U.S. 296, 332 (2004) (Breyer, J., dissenting). Risk assessment tools, in this context, had the potential to inoculate judges from such criticism. However, just because tools are data-driven, it does not necessarily follow that “the information produced is objective, neutral, and valuable to society.”164Jessica M. Eaglin, Constructing Recidivism Risk, 67 Emory L.J. 59, 88 (2017). The use of objective data says nothing about whether the selection of which factors to privilege, and the decision of how to weigh them, are also objective.

Scholars have noted that the design choices made by developers of these tools mirror “socially accepted structural inequities in society”165Id. at 97. and thus systematically prejudice already disadvantaged minorities.166See Hamilton, supra note 162, at 130. An empirical study conducted by ProPublica showed that risk assessments exhibit “disparit[ies] between scores assigned to white defendants and those assigned to [B]lack defendants,” when controlled for other factors.167Katherine Freeman, Algorithmic Injustice: How the Wisconsin Supreme Court Failed to Protect Due Process Rights in State v. Loomis, 18 N.C. J.L. & Tech. 75, 84 (2016). Black defendants were deleteriously mis-scored at a higher rate than white defendants, and low risk scores skewed disproportionately in favor of white defendants.168Id.

Accordingly, this method of constraining judicial sentencing discretion, while perhaps solving the problem of evenhandedly estimating risk, does not obviate the problem of disparate sentences based on immutable characteristics like race.169See Eaglin, supra note 164, at 98 (“A tool may accurately estimate risk and consistently classify more minorities as higher risk even though they will not engage in future criminal behavior. Whether society accepts that [B]lack defendants will disproportionately bear the burden of additional supervision flowing from actuarial risk assessments is a normative decision.”).

* * *

Hence, risk assessments, like sentencing guidelines and charging guidance, while having the intent of eliminating discretionary bias, may instead continue to “exacerbate unwarranted and unjust disparities that are already far too common in our criminal justice system and in our society.”170Eric H. Holder, Jr., U.S. Att’y Gen., Remarks at the National Association of Criminal Defense Lawyers 57th Annual Meeting and 13th State Criminal Justice Network Conference, https://www.justice.gov/​opa/​speech/​attorney-general-eric-holder-speaks-national-association-criminal-defense-lawyers-57th [https://perma.cc/​8CVF-WWYC] (Aug. 18, 2015). Clearly, the balance between encouraging individualized determinations and reducing discretionary bias has, so far, been a difficult one to achieve.

III. Procedural Fairness

“As problem-solving courts join the mainstream, there is pressure to standardize practices across courts both to ensure fairness and equality and facilitate resource management and accountability.”171Casey & Rottman, supra note 19, at 11. However—more than thirty years into the development cycle of specialized courts—selection procedures in these courts are still completely subject to prosecutorial and judicial discretion. Not only have selection procedures not been standardized, but there is little indication that doing so is a priority for judges and prosecutors.

The lack of standardized, discoverable, challengeable policies to ensure nondiscriminatory selection presents a threat to the entire project of specialized courts and the promise that they may hold. Specialized courts, to an extent previously unknown in the criminal legal system, rely on the compliance of not only defendants but also the defense bar, social service providers, and other community actors.172See supra Section I.C. If defendants do not view the courts as legitimate, two things can happen: those who are selected may not self-generate a desire to comply with the program, and those who are not selected may feel (additional) resentment and distrust toward the traditional courts to which they are relegated. Without cooperation from the defense bar, counsel may opt out of the “collaborative” process on which the specialized courts rely or advise clients not to accept diversion into such programs. Relatedly, if other community and social service stakeholders do not view the courts as legitimate, they may withhold their “buy-in.” Similarly, community agencies and social service providers, the last component of the specialized court partnership, may refuse to participate in what they perceive to be an unjust system.

A. Effects of Discretionary Bias in Specialized Court Selection

1. Documented Disparities

Consistent with data on discretionary bias in charging and sentencing, early evidence shows that Black and Hispanic defendants are less likely to be selected for specialized court admission than their white counterparts. For example, data from California State Drug Courts showed that, “[d]espite the disproportionate number of minorities who are charged with low-level drug offenses, minority offenders are less likely to be given the opportunity to enter drug court programs compared with white offenders because of the strict screening requirements of many drug courts.”173Joel Gross, The Effects of Net-Widening on Minority and Indigent Drug Offenders: A Critique of Drug Courts, 10 U. Md. L.J. Race, Religion, Gender & Class 161, 169 (2010). “The National Drug Court Institute has [reported] that state drug court participants are 62 percent [w]hite, 17 percent Black, and [10] percent Hispanic; by contrast, state prisoners are 32 percent [w]hite, 37 percent Black, and 22 percent Hispanic, and state probationers are 54 percent [w]hite, 30 percent Black, and 13 percent Hispanic.”174U.S. Sent’g Comm’n, Federal Alternative-to-Incarceration Court Programs 9 (2017). These statistics demonstrate that “[w]hite offenders are substantially overrepresented compared to their representation in the regular state probation and prison populations, while Black and Hispanic participants are underrepresented.” 175Id.

Of participants selected for admission to veterans court, 65.7% were white, while 26% were Black.176Jack Tsai, Andrea Finlay, Bessie Flatley, Wesley J. Kasprow & Sean Clark, A National Study of Veterans Treatment Court Participants: Who Benefits and Who Recidivates, 45 Admin. & Pol’y Mental Health 236, 239 (2018) (using data on a national sample of over 7,000 Veterans Treatment Court participants). Although this distribution roughly tracks the racial makeup of living veterans,177See Jason-Harold Lee & Julia B. Beckhusen, U.S. Census Bureau, Veterans’ Racial and Ethnic Composition and Place of Birth: 2011 (2012), https://www2.census.gov/​library/​publications/​2012/​acs/​acsbr11-22.pdf [https://perma.cc/​M4PE-RCFU]. it does not account for the disproportionate rate at which Black men are criminally charged and incarcerated.178See Elizabeth Hinton, LeShae Henderson & Cindy Reed, Vera Inst. of Just., An Unjust Burden: The Disparate Treatment of Black Americans in the Criminal Justice System (2018), https://www.vera.org/​downloads/​publications/​for-the-record-unjust-burden-racial-disparities.pdf [https://perma.cc/​WF2R-SSAE] (noting that “nearly 35 percent of all men who are under state or federal jurisdiction with a sentence of more than one year” are Black); Jennifer Bronson, Ann Carson, Margaret Noonan & Marcus Berzofsky, Veterans in Prison and Jail, 2011–12, at 3 (2015), https://bjs.ojp.gov/​content/​pub/​pdf/​vpj1112.pdf [https://perma.cc/​A49Z-Z95B] (noting that 50% of veterans in prison were white, 27% were Black, and 11% were Hispanic). An empirical study characterized the majority of participants in these courts as “[w]hite, unmarried, male, in their 40s, with at least a high school education, and a monthly income above $1000.”179See Tsai, Finlay, Flatley, Kasprow & Clark, supra note 176, at 239.

Similarly, the Changing Actions to Change Habits (CATCH) Court, a specialized court for sex trafficking victims in Columbus, Ohio, reported that its participants were “predominantly Caucasian and female.”180This characterization applies to the court as of 2012. See Karen Miner-Romanoff, An Evaluation Study of a Criminal Justice Reform Specialty Court—CATCH Court: Changing Actions to Change Habits, 3 J. Hum. Trafficking 1, 6 (2016). In contrast, during the same period, federal statistics revealed that sex trafficking victims were most likely to be Black (40%), followed by white (26%) and other races.181U.S. Dep’t of Just. Off. of Just. Programs, Most Suspected Incidents of Human Trafficking Involved Allegations of Prostitution of an Adult or Child (2011). And the Ohio Attorney General reported that the race of victims trafficked in Ohio before the age of eighteen broke down as follows: 26% were non-Hispanic white, 65% were Black, 3% were Hispanic, and 2% identified as Native American.182Celia Williamson, Tasha Perdue, Lisa Belton & Olivia Burns, Ohio Hum. Trafficking Comm’n, Domestic Sex Trafficking in Ohio 4 (2012). Of a smaller sample of “current adult victims of manipulation,” the Attorney General’s office found that 51% were white, 37% were Black, and 12% were Hispanic. See id. at 12–13. These racial breakdowns might also speak to who in the sex trade is considered a victim and who is considered a perpetrator, as many “traffickers” were also trafficked themselves.

Although empirical evidence is still scant, there is little reason to believe that such disparities will not continue to be revealed to exist across the spectrum of specialized courts, as these disparities arise whenever prosecutors and judges are bestowed this level of discretion. The same prosecutors and judges who are central to creating discretionary bias in the traditional criminal legal system are the actors who are vested with the power to write guidelines—or make ad hoc decisions—about which defendants are diverted into specialized courts. For this reason, even absent robust statistical evidence, stakeholders should have serious concerns “over the risk of racial disproportionality in who gets selected for the benefit and who succeeds in actually securing the benefit—and who does not and is subject to the hammer of harsher incarceration terms.”183Mary D. Fan, Beyond Budget-Cut Criminal Justice: The Future of Penal Law, 90 N.C. L. Rev. 581, 642 (2012).

In her work, Professor Barbara J. Flagg has discussed the idea of “policies that reinforce the existing racial distribution of key social goods.”184Barbara J. Flagg, “And Grace Will Lead Me Home”: The Case for Judicial Race Activism, 4 Ala. C.R. & C.L. L. Rev. 103, 124 (2013). In the case of specialized courts, the lack of neutral, characteristic-blind criteria is an example of an absence of policy that allows the exercise of discretion to the same end—reinforcing the existing racial distribution of a key social good: access to specialized courts.185See Gross, supra note 173, at 169 (“[M]inority offenders are less likely to be given the opportunity to enter drug court programs compared with white offenders because of the strict screening requirements of many drug courts.”); McLeod, supra note 1, at 1670 (“A related risk is that in removing more purportedly sympathetic defendants from conventional criminal courts, racial and class disproportion will increase, with more defendants of color and materially poor defendants remaining in the conventional courts.”).

2. The Possibility of Limited Status Recognition

In addition to the discretionary biases described above, there are at least two other ways that judges and prosecutors might fail to properly categorize defendants who potentially fall within the purview of the mission of a particular specialized court. First, judges may use a particular status as a stand-in for a characteristic, thus excluding defendants who share the characteristic but not the status. Second, judges may not equally recognize defendants who share a status, based on stereotypes surrounding that status.

The veterans courts can serve as an example of the first fallacy. These courts associate a specific personal characteristic of a defendant—in this instance, their status as a military veteran—with a special, less-culpable justification for the commission of an offense. Under this theory, the illegal conduct of veterans is more excusable—or better positioned for treatment instead of punishment—because that conduct may have been influenced by post-traumatic stress disorder.186See Kristine A. Huskey, Justice for Veterans: Does Theory Matter?, 59 Ariz. L. Rev. 697, 703 (2017) (noting that evidence unquestionably exists that suggests “a link between military service or exposure to combat; symptoms relating to PTSD, TBI, and other mental health problems; and criminal conduct”). . “These special courts recognize the negative impact of military service—particularly, exposure to combat and war zones . . . . [and] [t]hey acknowledge that veterans are a unique population, who could benefit from a treatment court tailored to their needs.”187Id. at 704–05.

Researchers have posited that:

[V]eterans with PTSD essentially reexperience trauma in “survivor mode,” which manifests in three distinct ways leading to crime: dissociative syndrome, sensation-seeking syndrome, and depression-suicide syndrome. Each of the three states exhibit particular behaviors such as aggression (dissociative), risk-taking (sensation-seeking), and reacting violently toward oneself or others due to perceived suffering (depression‑suicide), all of which can tend toward conduct that is criminalized. Put more simply, symptoms of PTSD may include experiencing flashbacks, increased perception of threats, anger, hypervigilance, exaggerated startle responses, emotional numbing or heightened emotional responses, all of which may lead to criminal behavior.188Huskey, supra note 27, at 181 (footnotes omitted); see also Huskey, supra note 186, at 703 (“[A]n individual with PTSD may re-experience trauma in survivor mode and consequently exhibit one or more behaviors, such as aggression, risk-taking, exaggerated startle response, or reacting violently to others or to oneself. People with TBI may experience depression, irritability, rage, aggression, mood swings, apathy, and acting on impulse.”).

This set of behaviors and motivations—understood to be manifestations of PTSD—would certainly sound familiar to any criminologist or urban sociologist, since they mirror those that have been documented in inner-city children of color raised in communities below the poverty line.189See Collins, supra note 3, at 1501 (“People who live in low-income urban neighborhoods are at a high risk for exposure to community violence. A 2011 study found that forty-three percent of the patients examined at Chicago’s Cook County Hospital displayed symptoms of PTSD, higher than even the highest estimates for veterans. Interviews of inner-city residents in Atlanta, Georgia revealed ‘rates of PTSD . . . as high or higher than Iraq, Afghanistan or Vietnam veterans.’” (footnotes omitted)). If the criminal legal system is to accept that defendants’ behavior that is induced by PTSD should be treated differently than other types of criminal behavior, then the question becomes why it should single out a particular cause of PTSD—serving in the military—instead of opening the court to all defendants who can show that they suffer from the disorder.190It is also notable that the “majority of veterans in prison (75%) and jail (69%) reported that they did not experience combat while serving in the U.S. military,” and only 23% reported “that a mental health professional ever told them they had post-traumatic stress disorder (PTSD).” See Bronson, Carson, Noonan & Berzofsky, supra note 178, at 7–8. So far, no judge has articulated a legitimate, principled justification for such a distinction.

Erin Collins suggests, quite convincingly, that there is no reasoned rationale for such distinctions. Instead, “[w]hat really distinguishes these offenders from others is not the impact of trauma per se, but rather a judgment that the criminal justice system should account for the impact of trauma upon certain populations because they are more deserving of such treatment.”191Collins, supra note 3, at 1500. Veterans, like people living with mental health challenges and people engaged in the solicitation side of sex work, are seen as victims of forces outside of their control. Regrettably, perhaps the largest population of people suffering from PTSD—young, poor, men of color—are simply not perceived in the same way.192Veterans are also thought to be necessary victims of government action to whom the government thus owes a duty. But the same can be said for young men of color who have been relegated to poor, segregated, under-resourced neighborhoods through the effect of years (if not centuries) of government policy. Not to be too on the nose, but these young men have been placed in war zones by the government just as surely as have military veterans. Accordingly, these young men are also owed a similar duty by the government, if such a duty exists.

The second error that judges can make is having a too-narrow conception of the population to which a specialized court is intended to attend. For example, “[t]he prevailing narrative about young people engaged in the sex trade is that they are young girls controlled by pimps.”193Rachel Swaner, Melissa Labriola, Michael Rempel, Allyson Walker & Joseph Spadafore, Youth Involvement in the Sex Trade: A National Study, at xvi (2016). However, a study from the Center for Court Innovation revealed that “young people who are engaged in the sex trade are a diverse population that does not conform to any particular stereotype.”194Id. at xv. And, although agencies that specialize in serving sex-trafficked teens “tended to perceive that the majority of these youth are female and work with a pimp,” the Center’s study found that, in fact, “[t]he population varies in gender, sexual orientation, and living situation, among other attributes.”195Id. at xiv–xv.

Thus, when judges presiding over specialized courts like the Prostitution Court in Columbus, Ohio, offer remedies such as placement in safe housing, they are unlikely to provide equivalent options for trafficked sex workers who do not fit the common conceptions of that population—defendants who are not women, or those who do not immediately appear to be in fragile or tenuous living conditions.196See Becca Kendis, Comment, Human Trafficking and Prostitution Courts: Problem Solving or Problematic?, 69 Case W. Rsrv. L. Rev. 805, 823–24, 836 (2019). By extension, because the remedies that the court offers are not appropriate for these types of defendants, they will not be extended the opportunity to have their cases adjudicated in that court.

In addition to simply failing to serve populations who are equally “deserving” of the intervention of a specialized court, judges also elevate and distinguish certain defendants over the teeming masses of offenders who will be pushed through the traditional criminal legal system. “[B]y invoking specious claims that the needs of these populations are unique,” those presiding over specialized criminal courts “obscure the connections between status court offenders and other offenders.”197Collins, supra note 3, at 1484.

B. Procedural Fairness

A potential solution to the discretionary bias problem in specialized court selection originates in the concept of procedural fairness. In the context of the criminal legal system, procedural fairness expresses a theory “that people will obey laws, without the threat of sanctions, when they experience the criminal justice system and its authorities as acting justly.”198Tyler, supra note 47, at 309. Here, what is just speaks to practices that people perceive as being morally appropriate and fair.199See id. at 308–09.

Fairness, under this conception, is not outcome-based but is instead process-based. The salient process considerations have been described as (i) voice—which speaks to whether a person has been given a meaningful opportunity to tell their story; (ii) neutrality—which specifically invokes the absence of bias; (iii) trustworthiness—which implicates the benevolence of the decisionmakers; and, finally, (iv) whether the person feels that they were treated with dignity and respect.200Michael M. O’Hear, Plea Bargaining and Procedural Justice, 42 Ga. L. Rev. 407, 420–22 (2008). If a participant is satisfied with these process considerations, that approval can “exert greater influence over acceptance of the result than do the outcomes themselves.”201Id. at 421. In fact, studies have shown that a defendant’s overall “satisfaction with case outcomes, the judge, and the court system are predicted first by procedural fairness and second by distributive justice.”202David B. Rottman, Adhere to Procedural Fairness in the Justice System, 6 Criminology & Pub. Pol’y 835, 838 (2007).

Conversely, when people experience the system as behaving randomly, irrationally, discriminatorily, or immorally, they are less likely to willingly submit to its authority.203See O’Hear, supra note 200, at 421–22 (“[T]he perception that legal authorities have legitimacy enhances the sense that the authorities are entitled to be obeyed. Fair procedures thus promote cooperation with the authorities and compliance with their directives, as well as the development of a more general sense of obligation to obey the law.”). Thus, it is only through acting justly that a court is considered legitimate. Legitimate authority is defined, for these purposes, as an authority, law, or institution “regarded by people as entitled to have their decisions and rules accepted and followed by others.”204Tyler, supra note 47, at 311.

So, when participants in the system perceive that procedural fairness exists, several things happen.205Tyler has identified several characteristics of institutions that exhibit procedural justice, including: (1) voice (the perception that your side of the story has been heard); (2) respect (perception that system players treat you with dignity and respect); (3) neutrality (perception that the decision-making process is unbiased and trustworthy); (4) understanding (comprehension of the process and how decisions are made); and (5) helpfulness (perception that system players are interested in your personal situation to the extent that the law allows). Emily Gold & Melissa Bradley, The Case for Procedural Justice: Fairness as a Crime Prevention Tool, 6 Cmty. Oriented Policing Servs. Off. (U.S. Dep’t of Just., Washington, D.C.), Sept. 2013, https://cops.usdoj.gov/​html/​dispatch/09-2013/​fairness_as_a_crime_prevention_tool.asp [https://perma.cc/​9U2R-2M72]. First, they express greater approval of the process, regardless of the outcome.206See Burke, supra note 28, at 125 (“No one likes to lose, but litigants recognize that they cannot always win. They accept losing more willingly if the procedure used is fair.”); Jonathan D. Casper, Tom Tyler & Bonnie Fisher, Procedural Justice in Felony Cases, 22 Law & Soc’y Rev. 483, 486 (1988) (“[L]itigants judge the fairness of their outcome on the basis of some abstract or principled criterion and . . . this evaluation affects their overall satisfaction with their court experience, beyond the impact of the favorability of the outcome itself.”). Second, general perceptions of racial bias in the judicial system, and the resultant distrust in the system, dissipate. Models have demonstrated that “race and other sociodemographic variables tend to become insignificant” when authorities exercise procedural fairness.207Rottman, supra note 202, at 838 (“If African Americans and, to a lesser extent, Latinos have less trust in the courts than other racial and ethnic groups, it is because they perceive less procedural fairness.”). This is true even when participants initially express race-based differences in their perceptions of the police and courts; and, specifically, this is also true of young, minority men who studies show are strongly motivated by procedural justice judgments.208Tom R. Tyler & Yuen J. Huo, Trust in the Law: Encouraging Public Cooperation with the Police and Courts 155 (Karen S. Cook, Russell Hardin & Margaret Levi eds., 2002).

Finally, when defendants perceive that they are being treated justly, they become more likely to obey the orders of the court,209See Rottman, supra note 202, at 838. which, in turn, sees more cooperation and “everyday compliance with the law.”210Tyler, supra note 47, at 312. This effect extends from the courtroom itself—procedural justice has been found to encourage deference and lessen the likelihood of spirals of conflict—to compliance with sentencing.211See id. at 308–09.

People who experience procedural fairness are more likely to “become self-regulating, taking on the personal responsibility for following social rules.”212Id. at 308. This last effect—that defendants more freely cooperate and self-regulate—is especially significant in the context of specialized courts, where effective treatment is premised upon the supposition that the defendant buys-in to the types of programs in which they are ordered to participate.

Tyler has identified two types of procedural justice—“justice in the quality of decision-making procedures and justice in the quality of treatment that people receive from others.”213Id. at 309. Specialized courts are already perceived as promoting the second type of procedural justice—treatment from others—or at least doing so to a greater extent than traditional courts.214See Casey & Rottman, supra note 19, at 11 (“Problem-solving court proceedings are rated more highly than traditional court proceedings on the dimensions of respect, neutrality, voice, and trustworthiness.”). However, these courts absolutely fall short of providing the first type of justice—quality of decision-making procedures.215An appeal to procedural justice necessarily raises questions surrounding the adequacy of due process and equal protection in this circumstance. In the criminal context, due process usually looks like a set of procedural rights (i.e., notice, hearings, testimony and confrontation, independent decisionmakers, and appellate rights). However, the moment of specialized court selection, much like many moments invoking prosecutorial discretion, is largely afforded none of these procedural protections. Similarly, equal protection rights are generally unenforceable in these areas of prosecutorial discretion, absent proof of malice, as their focus is on the reasons why police, prosecutors, and juries act in particular cases, not on group effects over time and space. Thus, equal protection arguments are ill suited to address subtle, systemic discrimination created by the seemingly neutral operations of discretion within the courts.

C. The Threat to Legitimacy

As long as specialized courts continue to fail to create or implement procedurally fair decision-making processes, they face the risk of delegitimization.216See Bowers & Robinson, supra note 99, at 211–12. This Article does not make a distinction between legitimacy stemming from procedural justice and moral credibility. For a discussion of those intersections, see generally id. “[E]ach encounter that people have with authorities is an instance of civic education, which teaches people about the law.”217Tyler, supra note 47, at 318. Following from this, each encounter that community and legal stakeholders have with the specialized court system is part of an education in the system’s fairness or lack thereof. If such stakeholders perceive the court’s decision-making process to be fundamentally unfair, either because of discretionary bias or because of racial discrimination, then the court’s legitimacy is diminished.218See id. at 312–14 (“When people view the authorities as engaging in practices that the public views as being morally appropriate, that heightens their sense that legal authorities are behaving morally. . . . [but when authorities] engage in racial profiling, which people view as an unfair procedure, they diminish their moral authority by showing that they do not share the public’s moral values.”). This Article does not make a distinction between legitimacy stemming from procedural justice and moral credibility. For a discussion of those intersections, see Bowers & Robinson, supra note 99.

The fact that judges created many of the specialized courts, as discussed above, would tend to demonstrate that they find them legitimate. Specialized courts were created as a direct response to judges’ “increasingly negative experience of the criminal justice system,” stemming from their observations of recidivism rates, the increase in mass incarceration, and the resultant burdens on judicial efficiency.219Miller, supra note 41, at 1564; see also Bowers, supra note 53, at 785 (“Supporters maintain that the courts effectively serve several goals: to provide second chances for nonviolent addicts, to preserve systemic resources, and to control crime by disrupting cycles of addiction and recidivism.”). Judges also cited, as a reason for supporting the creation of these courts, concerns about the “harshness” of sentencing laws and the lack of alternatives to such laws.220   Scott-Hayward, supra note 1, at 51 (“Specialized criminal courts . . . have now become the focus of innovation at the front-end of the federal criminal justice system and appear to be the dominant form of diversion. These courts, which are variously called ‘alternative to incarceration’ programs, ‘court-involved pretrial diversion practices,’ and ‘diversion-based court programs’ now exist in at least 21 federal districts. . . . Specialized criminal courts now appear to be the predominant response to continuing concerns among judges and other stakeholders about the harshness of federal sentencing laws and limited federal sentencing options.” (footnotes omitted)). Thus, if the existence of these courts successfully addresses these concerns, then there is no reason to believe that the system would lose buy-in from the judges that preside within it.221Or, at least there is a perception among judges that the courts address these concerns, regardless of whether such perception is borne out by evidence.

Prosecutors are easy enough to get, and keep, on board because—in addition to having the concerns shared by judges—participating prosecutors are assured an easy “win” whenever a defendant is diverted to a specialized court, particularly in post-adjudication jurisdictions. Especially in cases that might have otherwise posed a question of whether a defendant’s rights were violated, diverting a defendant into a specialized court eliminates a prosecutor’s duty to pursue (or alternatively dismiss) a case.222See Thompson, supra note 4, at 74 (“[P]articipants in [the deferred prosecution] model may be reluctant to challenge the legality of their detention or arrest after a lengthy participation in the court’s programs. In the post-disposition model . . . . [d]rug court participation is conditioned on a willingness to forego any legal challenges that the defendant might have raised regarding her arrest or the seizure of evidence.”); Nolan, supra note 6, at 1559 (noting that some jurisdictions require a defendant to “sign forms waiving a host of constitutional rights in order to participate in [a specialty] court, including the right to trial by jury, the right to a speedy trial, the right to a preliminary hearing, and the requirement of probable cause for a search and seizure”).

The incentives for the defense bar to collaborate in the specialized court process are more limited: defense counsel must believe that these courts actually confer a direct benefit on their individual clients (i.e., that they are preferable to the traditional legal system in at least one way).223Alternatively, defense counsel might acquiesce to participation in specialized courts out of a sense of ease. See Bowers, supra note 53, at 820–21 (“The indolent lawyer who wishes to dispose of the substantive case can convince readily the over-optimistic, acutely addicted, poor candidate to take the drug court option, because the offer seems to hold out the promise of everything the defendant could want: immediate freedom and the possibility of dismissal. The lawyer can then monitor compliance with little effort (and often from afar); and, if and when termination comes, the lawyer is required to appear only for a sentencing hearing at which no contested issues are litigated and for which no work need be done.”). A zealous defense lawyer’s acquiescence to and acceptance of the authority of a specialized court is no small task. Remember that, in doing so, defense counsel must advise their clients to make significant and potentially irreversible trade-offs. As discussed above, because of the collaborative structure of specialized courts, their existence fundamentally shifts the traditional role of defense counsel in a way that impacts counsel’s ability to advocate for the protection of the client’s constitutional rights.224See McLeod, supra note 1, at 1665. Moreover, defendants who enter specialized courts often have a much longer and more involved interaction with the court system225See U.S. Sent’g Comm’n, supra note 174, at 10. than those who have traveled the traditional route.226See Nolan, supra note 6, at 1555–56 (“Judges Hora and Schma acknowledge that drug court ‘requirements may prove more onerous than the equivalent traditional court sanctions for the same offenses’ and that this particular quality of the drug court ‘tends to disturb defense attorneys.’ Moreover, they recognize that drug courts ‘generally obligate a defendant to make more frequent court appearances and force the defendant to undertake forms of treatment which place more burdens on the defendant than normal probation.’ In response to the criticism that this may be unjust, they simply assert the preeminence of the therapeutic jurisprudence perspective.” (footnotes omitted)); Thompson, supra note 4, at 74. They also frequently receive longer periods of incarceration if they fail to meet the requirements of the programs to which they have been sentenced.227See Jaros, supra note 17, at 1511–12.

Similarly, participating social service providers can either be motivated by purely self-serving concerns—namely, being paid by the state—or by altruistic concerns. Scholars have suggested that such stakeholders articulate the latter reason to justify their participation—that they believe in these courts’ “social change goals and their own roles in furthering that social change.”228Rekha Mirchandani, What’s So Special About Specialized Courts? The State and Social Change in Salt Lake City’s Domestic Violence Court, 39 Law & Soc’y Rev. 379, 408 (2005). To put it bluntly, they participate because they want to be part of a good thing.

But what if these stakeholders came to believe that the courts were not a good thing, but instead, were a mechanism that only served to further invidious discrimination within the legal system? Remarkably, even at the far fringes—among people who were actually deemed to be white segregationists—“discrimination in the legal system was near the bottom” of their values.229Michael J. Klarman, The Racial Origins of Modern Criminal Procedure, 99 Mich. L. Rev. 48, 76 (2000) (describing Gunner Myrdal’s findings in An American Dilemma). Even those who were “intensely committed to racial segregation in education and to legal bans on interracial marriage did not endorse manifestly unfair trials for [B]lack criminal defendants.”230Id. In fact, “[t]he single clearest trend shown in studies of racial attitudes has involved a steady and sweeping movement toward general endorsement of the principles of racial equality . . . .”231Lawrence D. Bobo, Racial Attitudes and Relations at the Close of the Twentieth Century, in 1 America Becoming: Racial Trends and Their Consequences 264, 269 (Neil J. Smelser, William Julius Wilson & Faith Mitchell eds., 2001). People do not like to be racist. Or, perhaps more pointedly, people do not like to be considered racist or perceived as engaging in racist behaviors.

Because cooperation with a specialized court is not mandated for any nongovernmental stakeholder, such service providers must either have an incentive or a desire to participate in the endeavor. When specialized courts behave in ways that can be perceived as racially discriminatory, or against the general values of fairness and equality, then the motivation among these actors to participate necessarily decreases.232See Bowers, supra note 53, at 820 (noting that arbitrary actions by these courts, such as “[w]hen judges are free to construct rewards and sanctions out of whole cloth and to keep participants in treatment for indeterminate lengths,” can leave defense counsel feeling ineffective and uninvested in the process, “throw[ing] up her hands and tell[ing] her client: ‘Prepare to turn your life over to this judge and her whims for at least the next year or two’”). The good that they once perceived no longer seems good. As posed by Josh Bowers, if actors perceive the system “as unjust, what motivation would they have to assist it . . . or to defer to it . . . ? Common sense tells us that people are more likely to resist and subvert a criminal justice system that they see as unjust than they are to assist and defer to it.”233Bowers & Robinson, supra note 99, at 256.

There is an additional consideration that should impact the potential cooperation of defendants—the knowledge that, but for the existence of specialized courts, they may not have been arrested or prosecuted at all. Scholars have observed that the existence of specialized courts may cause a “net-widening effect,” where police make arrests and prosecutors charge crimes for which they would normally only issue a violation.234See Nolan, supra note 6, at 1561–62. Although it would be difficult (if not impossible) to prove empirically, there are certainly defendants who have been swept into the system only because a jurisdiction had a specialized court, but then were arbitrarily or discriminatorily rejected from participation in that court.

Certainly, scholars have raised questions about whether a causal link between perceptions of legitimacy and institutional deference has been definitively proven. However, while “[c]laims of deference may be nonfalsifiable . . . in settings where public-policy choices must be made, policy makers may be wise to assume that perceptions of legitimacy do have a sizable effect on rates of compliance and cooperation . . . .”235Bowers & Robinson, supra note 99, at 255.

So, the lack of procedural fairness in the selection stage has the possibility of undermining the authority and legitimacy of the specialized court model. Any sustained or large-scale move away from the traditional system toward rehabilitative solutions requires criminal defendants, the defense bar, and community partners to trust in the authority and legitimacy of the alternative courts.236See Tyler, supra note 47, at 313 (“[P]rocedural justice suggests that possibility of a legal system based more heavily upon voluntary cooperation—of process-based regulation.” (citation omitted)). So far, specialized courts have operated on the fringes of a behemoth traditional system, and discomfort about discretionary bias in selection processes has not been brought to the fore. However, as these courts continue to expand and touch an ever‑larger slice of the population, they will not be able to escape the problem of the inherent lack of procedural fairness in their structure.

Following this idea, the most straightforward way to increase both the existence and perception of procedural fairness in specialized court selection, and decrease discretionary bias and discrimination, involves instituting formal rules and standardized systems for selection.

IV. Models for Selection

If the first-order question is whether standardized selection criteria are necessary for the continued legitimacy of specialized courts, then a second-order question is what the content of those criteria should be. As Mari Matsuda has pointed out, “informality and oppression are frequent fellow-travelers,” and, as such, “measures to eliminate effects of oppression . . . . are best implemented through formal rules, formal procedures and formal concepts of rights.”237Mari J. Matsuda, Public Response to Racist Speech: Considering the Victim’s Story, 87 Mich. L. Rev. 2320, 2325 (1989).

As prior attempts to restrain discretionary bias within the traditional legal system have been largely unsuccessful, as discussed in Section II.D, it would be unwise for specialized courts to adopt any of the aforementioned methods. Instead, specialized courts should develop selection criteria based on a normative principle aligned with the purposes of the courts’ creation—to reduce recidivism through rehabilitation, decrease mass incarceration, and increase judicial efficiency.

The current criteria for admission to these courts largely centers around two interrelated concepts: desert and perceived potential for rehabilitation.238The penchant for relying upon the concept of desert as a normative principle is in line with a lay understanding of criminal punishment. See Bowers & Robinson, supra note 99, at 240 (“Studies confirm that laypeople think of criminal liability and punishment in terms of desert—the moral blameworthiness of the offender—and not in terms of other principles, such as general deterrence and incapacitation, which have been so popular with system designers during the past several decades.”). As discussed above, courts that select participants based on status are generally operating under a conception that some types of defendants deserve rehabilitative services more than others based on being a member of a certain group, be they veterans or people living with mental health diagnoses. When specialized courts limit participation to, for instance, nonviolent offenders, they are not only exhibiting a preference for deserving defendants but are also employing a model that gestures to whether a person is somehow well positioned for rehabilitation. In limiting or opening selection to certain classes of defendants, judges often reference additional factors in the defendants’ favor, including the existence of family support structures, educational attainment, employment history, etc.239See, e.g., State v. Curry, 988 S.W.2d 153, 157–59 (Tenn. 1999). This, too, speaks to those courts’ preference for defendants that the court believes have characteristics that increase the potential for successful rehabilitation.

In this Part, I argue that neither a sense of desert nor conceptions of amenability to easy rehabilitation lend themselves to a perception of procedural justice. Instead, if specialized courts want to be perceived as just in their selection processes, they must take a different normative stance about what constitutes fair selection.240The selection model implemented by New York’s Human Trafficking Court is arguably the most procedurally fair model in current use. Under that model, all defendants arrested on a particular set of charges—in that case, misdemeanor prostitution charges—are presumptively redirected into that court. They are then assessed (and, here, the criteria again become quite fuzzy) to determine whether they will be returned to the traditional criminal justice route. See Human Trafficking Courts, supra note 86. Moreover, specialized courts must include formerly incarcerated people, the defense bar, and social services providers in crafting selection criteria.241See Miller, supra note 41, at 1566 (discussing the value of “[p]articipation in the norm creation and norm enforcement process”).

In the following Sections, I outline several approaches to selection that are better aligned with perceptions of fairness: random selection, prioritizing admission by need, data-driven selection, and taking no approach to selection at all.

A. Proposed Models

1. Random Selection/Allocation

The purest form of procedurally fair selection—although perhaps not the best—would be the random allocation of specialty court “slots.” If defendants understand fairness to be equivalent to an equal opportunity to be selected for participation, then selection through a lottery would enhance the perception of the courts as acting in line with the moral understanding of its constituents.

Under this model, for example, a drug court could determine how many inpatient treatment beds would be available for the upcoming month (X slots) and then randomly select that number of new participants to enter drug court (X diversions) from among all drug-charged defendants arrested in the prior month. Through this method, every defendant with a qualifying charge would have an equal chance to receive the benefits of specialized court, regardless of race or other personal characteristics.242Ronen Perry & Tal Z. Zarsky, Queues in Law, 99 Iowa L. Rev. 1595, 1609 (2014) (“[R]andom selection . . . is totally indifferent to personal characteristics of any sort.”).

Randomization can be a reasonable approach to questions of fairness when “the objects to be assigned are indivisible and monetary transfers are . . . unavailable.”243Eric Budish, Yeon-Koo Che, Fuhito Kojima & Paul Milgrom, Designing Random Allocation Mechanisms: Theory and Applications, 103 Am. Econ. Rev. 585, 585 (2013). In the case of specialized courts, judicial resource allocation dictates that there will likely always be more defendants who inhabit the status or are charged with the relevant offense than availability to serve those defendants outside of the traditional criminal model. “Randomization can restore symmetry” between groups that have otherwise experienced asymmetrical treatment and, in doing so, also restore “a measure of fairness.”244Id.

However, criminologists have pointed out that random selection is not favored by inmates in the prison population as a method of ensuring fairness because it is perceived as being arbitrary or based on “luck.”245See Edna Erez, Random Assignment, The Least Fair of Them All: Prisoners’ Attitudes Toward Various Criteria of Selection, 23 Criminology 365, 365 (1985). Random selection also runs squarely against the articulated desire of judges that the criminal legal system dole out individualized sentences, which necessitates that the specific characteristics and actions of each defendant be considered when adjudicating their case. The law abhors arbitrariness just as much, and in fact more, than it rejects mandated uniformity. While “whether your number is called” is not exactly the same as “what the judge had for lunch,” for many it is likely too close for comfort.

While randomness has deficits when thought of as a long-term process for selection, it is much more appealing as a short-term or interim approach. In the absence of preferable solutions, replacing judicial and prosecutorial discretion with random selection could at least create a control group that could be used for empirical comparison. Random selection could tell us a great deal about a great many questions, including to what extent selection bias affects court outcomes; whether the non‑extrajudicial criteria used by the courts are borne out in the results of judicial intervention; and even, at the most basic level, whether these courts fare any better in outcomes (around recidivism and other factors) than the traditional legal system.246The necessity for data-driven solutions will be discussed further. See infra Section IV.A.3.

Thus, while random selection may not be an ideal solution, it may serve as a bridge to a long-term resolution.

2. Needs Assessments

When surveyed and presented with four models—randomness, need, merit, and queueing—federal prisoners selected “need” as the fairest way to allocate scarce resources among individuals.247Erez, supra note 245, at 365. This preference was expressed even by inmates who were given a scenario in which the resource on offer was not one that they professed to have a need for themselves; thus, their ranking of “need” as fairest did not seem to stem from self-interest, but instead from a sense of moral principle.

In the current model, “[m]any state programs focus on low-level, low-risk offenders who may not need intensive treatment to prevent further substance use, which has led critics to accuse those court programs of ‘cherry-picking’ in order to show low recidivism rates of their participants.”248U.S. Sent’g Comm’n, supra note 174, at 12–13. Some or many of these offenders may have, prior to the advent of specialized courts, actually been “screened out of a traditional diversion system,” but instead are now being swept up in a net-widening that exists only because of the systemic need for low-risk offenders.249See Miller, supra note 41, at 1553.

This practice of “cherry picking” is consistent with criminal justice’s focus on risk assessment and prioritizing those perceived to be at low risk for recommission of crime over those who are perceived to be at higher risk. But applying this risk assessment philosophy to specialty court selection means, in effect, that those defendants with the least need—because they are at the lowest risk of recidivism and thus have the highest chance of being restored to the noncriminal population regardless of intervention—will receive the most/best resources.

As Eric Miller has noted:

If [the liberal justification of diversion from prison] is to be borne out in practice, the court’s eligibility criteria must promote the diversion of individuals who are otherwise likely to spend a significant amount of time in prison rather than those likely to receive non-custodial sentences or sentences that require less institutional confinement than that meted out in [specialty courts].250Id. at 1553–54.

Following from this logic, instead of employing a risk assessment model for selection, a fairer process might be to employ a “need assessment”—that is, to identify those defendants who are most at risk for recidivism in the traditional criminal adjudication model and prioritize them for intervention in specialized courts. Experts in the field of drug treatment have made just this argument, that “drug court programs generally should be limited to high-risk, high-need defendants.”251U.S. Sent’g Comm’n, supra note 174, at 13.

What would this shift look like? Under a “risk” model, a drug court might choose a defendant who has a recent history of drug addiction, a light or nonexistent criminal record, moderate educational attainment, and strong family support. In contrast, under a “needs” regime, courts would be encouraged to select defendants who are most in need of rehabilitation and thus most likely to recidivate absent specialized intervention. This system would lead to choosing a defendant who perhaps had a long criminal history of drug-related crimes and less community and familial support—someone for whom rehabilitative intervention could make all the difference.

Pivoting to a “needs” paradigm has several benefits. First, as discussed above, it carries with it a perception of procedural fairness. Second, foregrounding need can move the specialized court system away from the concept of deserving defendants, at least how it is currently conceived. In a “needs” world, desert would be flipped on its head—defined by privations, not by status. Finally, a needs approach has the capacity to simultaneously address what we traditionally think of as risk (by centering rehabilitative efforts on those most likely to recidivate) while still being defendant-centered (also by centering rehabilitative efforts on those most likely to recidivate). If the goals of these courts are to reduce recidivism and mass incarceration, then the way to do it is to try to address the big, intractable problems, not merely to deal with those defendants who are most easily dealt with.

3. Combined Approach—Data Collection

An approach combining the two aforementioned methods might also help to move the system of selection closer to fairness. The principal barrier in determining how to balance creating a fair procedure for selection, limiting discretionary bias, and achieving the aims of these courts, is that there is a dearth of relevant data available in this area.

Because judges and prosecutors are not required to create a record of the rationale behind their decision making, either on the level of general criteria or individual determination, there is an impermeable barrier to understanding the complete nature of the problem, and thus the most appropriate solutions.252One of the biggest problems with the proliferation of front-end specialized criminal courts in the federal system is the fact that none have yet been evaluated, meaning that there are no systematic assessments of their design, curriculum, and short- and long-term outcomes. . . . In its review of these courts, the OIG noted that Executive Office of the U.S. Attorney has not conducted any evaluations or assessments of diversion programs, and, more worryingly, it “did not keep sufficient data to permit a comprehensive evaluation of the effectiveness of the USAOs’ uses of pretrial diversion programs and their participation in diversion-based court programs.” Scott-Hayward, supra note 1, at 91–92. Perhaps an initial step in the right direction would be to merely request that they do so and make these records accessible to those who would like to analyze them.

Judges and prosecutors are plainly operating under the presumption that the selection criteria they have set in a particular specialized court are aligned with the goals of that court and further the execution of those goals. But is there any data to show that this is accurate? For example, for those courts that limit admission to certain types of defendants, even within the subset of a particular status or offense (e.g., nonviolent), the belief appears to be that such defendants are more likely to achieve successful rehabilitation under the programs offered by the court than those defendants with a higher “need.” But, truth be told, nobody really knows because specialized criminal courts have yet to subject themselves, at least on any large scale, to evidence- and outcome‑based scrutiny.253“Without a proper evaluation, including a comparison group, it is not clear whether success either in the program or after the program can be attributed to the program itself.” Id. at 92.

This Article has proposed that defendants who might be considered less deserving of specialized treatment—the neediest cases—be prioritized to receive such treatment. This approach has the added benefit of providing much needed data on the effectiveness of these courts. Selecting defendants for whom rehabilitative intervention could be outcome transformative would actually test the system—it is easy to show rehabilitative successes when only considering the cases of people who need very little rehabilitation. But a system can only be proven effective when it is required to tackle the hard cases.

4. No Approach

Briefly, there is another approach to selection that would eliminate discretionary bias and increase procedural fairness. And that approach is no approach at all.

This Article would be remiss if it did not at least question whether there is, in fact, a need for selectivity at all. If specialized criminal courts generate better outcomes than traditional adjudication, then it would be reasonable to suggest that the specialized court model should be the predominate model of criminal adjudication, not an alternative model. Similarly, in a world where research found that specialized criminal courts created better outcomes for most defendants, but worse outcomes for some subset of defendants, then the selection presumption would flip—criteria would need to be set to determine who was not eligible for specialized court, instead of to determine who was. And if that were the case, then the selection problems raised in this Article would vanish.

Are specialized criminal courts more expensive than traditional courts in terms of hard-dollar costs per defendant? If the ratio of defendants adjudicated through traditional courts versus those adjudicated through specialized courts were flipped on its head, would that cost differential be offset by soft-dollar savings through the reduction of recidivism? Again, these are questions that have yet to be answered. They are also questions that are not ones of fairness or legitimacy, but of resource allocation.

If specialized courts really work, then the fairest outcome would be that we, as a society, figure out how to make them work for everyone.

B. Limitations and Open Issues

Certainly, more empirical studies on the demographics of defendants entering into specialized courts, and those being denied entry, are needed. States, which are often good at collecting data on specialized court outcomes,254See, e.g., Michigan’s Problem-Solving Courts: Solving Problems, Saving Lives, FY 2017 Annual Report on Performance Measures and Outcomes, Mich. Sup. Ct. (2017), https://www.matcp.org/​uploads/​8/​3/​0/​0/​83009324/​pscannualreport.pdf [https://perma.cc/​FB2E-FA6L]. should also begin to collect demographic data on defendants who were screened for specialized courts, or otherwise fell within their purview, but were not admitted. That data should include information on race, along with information on social indicators such as family background and economic status. If this wider data set continues to show that discretionary bias plays a role in selection, policymakers must be amenable to revisiting both the substance and process of selection.

Despite the benefits of implementing standardized specialty court selection criteria, doing so will never be sufficient to completely curtail prosecutorial discretionary bias. A zealous prosecutor, still imbued with unchallenged charging authority, could always charge a defendant in such a way that they fall outside of the catchment of specialized courts.255Susan P. Weinstein, Ethical Considerations for Prosecutors in Drug Courts, 15 Crim. Just. 26, 28 (2000) (“Determining what charge to attach to a defendant may unfairly preclude him or her from the drug court. These two situations—forgoing prosecution and exercising discretion in leveling charges—elicit some interesting issues that prosecutors have not had to address before the creation of drug courts.”). Therefore, as long as “[p]rosecutors control and almost predetermine the outcome of criminal cases through these two critical [charging and plea bargaining] decisions,” discretionary bias will never be eradicated.256Angela J. Davis, In Search of Racial Justice: The Role of the Prosecutor, 16 N.Y.U. J. Legis. & Pub. Pol’y 821, 832 (2013).

History has also taught us that once inconsistency in the application of a policy is made visible—perhaps especially inconsistency along racial lines—the inconsistency is often addressed by drastically leveling up or leveling down the policy. In this case, such leveling could look at least two ways: lawmakers could decide that specialized court selection processes cannot be fixed in such a way that the bias would be sufficiently moderated, so specialized courts should simply not exist. Alternatively, the proposed fix to the discretionary bias problem could be to codify selection criteria at such a level that most defendants would be excluded, regardless of race. If specialized courts are, in fact, a good, then either of these reactions would create a net negative outcome.

A final note on the case for “need”: this Article acknowledges that the political incentive to prioritize the worst offenders for the best treatment is likely not very high. “Tough on crime” lawmakers adopt that posture because it has political benefit. But we should remember that most people who are adjudicated in any part of the criminal legal system do not get a sentence of natural life. To the contrary, the vast majority will eventually be released. Directing specialized attention toward the neediest does, no doubt, require a higher risk tolerance than judges and prosecutors have demonstrated heretofore. That being said, if courts and lawmakers are as concerned about recidivism as they purport to be, then this is where the work will need to happen—not with those who need intervention the least, but with those who need it the most.

Conclusion

Minority defendants have “suspicion, distrust, and hostility . . . for the prosecutorial process,”257Hagan & Peterson, supra note 121, at 29. and rightfully so. The discretion inherently granted to prosecutors and judges in the traditional criminal legal system is often exercised to these defendants’ detriment. In order to ensure a perception of legitimacy, specialized criminal courts must begin to impose transparent, consistent, and procedurally just criteria for defendant selection. In many ways, the newness of the specialized court model, and the flexibility inherent in its design, make it easier for the stakeholders involved to create procedurally fair selection systems. As described by Allegra McLeod, specialized courts are “experimentalist institutions that are open to revision in light of ongoing empirical feedback—they are unfinished, self-correcting, reformist organizations. . . . decidedly unfinished, promising gradual reform rather than a bold new program fully specified in advance.”258McLeod, supra note 1, at 1637. Sufficiently constraining discretionary bias in selection should be among the first steps that specialized courts take down the path of realizing the goals of responding to crime in a way that avoids the pitfalls that continue to hobble the traditional system.