In the past decade, major retailers nationwide have begun to employ a private, for-profit system to settle criminal disputes, extracting payment from shoplifting suspects in exchange for a promise not to call the police. This Article examines what retailers’ decisions reveal about our public system of criminal justice and the concerns of the agents who run it, the victims who rely on it, and the suspects whose lives it alters. The private policing of commercial spaces is well known, as is private incarceration of convicted offenders. This Article is the first, however, to document how industry has penetrated new parts of the criminal process, administering sanctions to resolve thousands of shoplifting allegations each year.

Proponents of private justice claim that everyone wins. Critics say it’s blackmail. The Article takes a tentative middle ground: While “retail justice” is not the American ideal, it may nonetheless be preferable to public criminal justice, at least if certain conditions are met. Rather than cancel the private justice experiment, therefore, as several states are poised to do, the government should aim to foster optimal conditions for its success.

Extending the central analysis, the Article then shows how the study of private justice leads to fresh perspectives on important criminal justice issues. It suggests, for example, that the costs to crime victims of assisting the prosecution may be a feature of the system, not a bug, if they encourage victims to invest in efficient crime-deterring precautions. It also complicates legal academic models of police and prosecutorial behavior built on maximizing arrests and convictions. The Article concludes by identifying conditions that conduce to private criminal justice and speculating about the next frontiers.


    1. Who Shoplifts and Why?
      1. A Century of Petty Theft
      2. Contemporary Evidence
    2. Private and Public Enforcement of Shoplifting Laws
      1. Ambivalence and Innovation
      2. The Persistence of Discretion
      3. The Patterns of Discretion
    3. The Rise of Retail Justice
    1. Are Suspects Worse Off?
      1. Economic Efficiency: Is Retail Justice a “Bad Deal” for Suspects?
      2. Equality: Are Certain Groups of Suspects Worse Off?
      3. Overenforcement: The Costs of Casting a Wider Net
    2. Are Victims Worse Off?
    3. Is Society Worse Off?
      1. The Social Costs of Crime
      2. Transparency
    4. Is This Blackmail?
    1. Lessons for Criminal Justice System Design
    2. Police and Prosecutorial Motives
    3. The Next Frontiers



Most shoplifters evade detection and many who fail are never formally punished. For decades, retailers battling theft have relied on a mixture of law enforcement and self-help, sending some suspected shoplifters to the station house and others to the street. Recently, a third option has emerged, raising basic questions about the interplay between public and private forces in American criminal justice. Some retailers now hand over shoplift­ing suspects not to the police but to a for-profit, specialist corporation like the Corrective Education Company (CEC). 1 For a detailed description of CEC and one of its main competitors, see infra section I.C. This “retail justice company” extracts payment from the alleged offender in exchange for a “restorative justice” course and a promise not to pursue criminal charges. The retailer pays nothing; in fact, in some cases, it reaps a portion of each suspect’s payment.

“Retail justice” is becoming a big business. CEC’s clients, for example, have included Walmart, Abercrombie & Fitch, Bloomingdale’s, DSW, and Burlington Coat Factory. 2 Restorative Justice for Retail Theft, CEC, (on file with the Columbia Law Review) [hereinafter CEC] (last visited Apr. 10, 2017). CEC reports an enrollment rate of roughly ninety percent, yielding thousands of “students” each year. 3 Id. It offers discounts and payment plans for suspects who cannot finance the entire $400 fee at once and “scholarships” for the poorest few. 4 Id. Terms like “students” and “scholarships” are placed in scare quotes to reflect ambivalence about their usage. Retail justice companies employ the terms to analogize their programs to other sorts of educational coursework. See infra notes 190–193 and accompanying text. Critics, however, resist the analogy in light of concerns about the voluntariness of the decision to enroll and pay the “tuition” that is discounted by the “scholarship.” See infra section II.A.1. CEC claims its program saves retailers time and money, relieves pressure on an overtaxed criminal justice system, and cuts recidivism by providing “life skills and motivation for reintegration.” 5 CEC, supra note 2.

The City of San Francisco, in contrast, alleges that CEC is little more than an extortion racket preying on the city’s residents. 6 See Complaint for Injunctive and Equitable Relief and Civil Penalties for Violations of Business and Professions Code Sections 17200 at 3, People ex rel. Herrera v. Corrective Educ. Co., No. CGC-15-549094 (Cal. Super. Ct. filed Nov. 23, 2015), 2015 WL 7709993 [hereinafter S.F. Complaint] (“[T]hreatening accused shoplifters with criminal prosecution unless they pay CEC hundreds of dollars violates California’s extortion laws.”). The Indiana Attorney General has reached the same conclusion, 7 See 2018 Op. Ind. Att’y Gen. No. 4, at 6–7, 2018 WL 1722444 (opining that CEC’s activities may fall within the Hobbs Act’s proscription of extortion). and a private-plaintiff class has sued participating retailers and CEC personnel for violating the Racketeer Influenced and Corrupt Organizations Act. 8 See Class Action Complaint for Violation of the Racketeer Influenced and Corrupt Organizations Act at 1, Doe v. Walmart Inc., No. 3:18-cv-2125 (N.D. Cal. filed Apr. 9, 2018), 2018 WL 1709735 [hereinafter Walmart Class Complaint].

There is no scholarly treatment, legal or otherwise, of this private “retail justice system.” Private dispute resolution is of course common on the civil side. 9 In addition to obvious examples like civil arbitration, recent work has highlighted the significant dispute-resolution function of corporate customer service departments and intermediaries like credit card companies. See Rory Van Loo, The Corporation as Courthouse, 33 Yale J. on Reg. 547, 551 (2016). And civil enforcement sometimes stands in for criminal jus­tice to rectify wrongs that violate parallel civil and criminal laws. Yet rhetoric imagining a state monopoly over enforcement of the criminal law persists. 10 See, e.g., Stephanos Bibas, The Machinery of Criminal Justice 84–88 (2012) (discussing “the state’s monopoly on criminal justice”); Adam B. Cox, Enforcement Redundancy and the Future of Immigration Law, 2012 Sup. Ct. Rev. 31, 35 (“[T]here are parts of the criminal law where a single state actor . . . has a meaningful monopoly on enforcement within a particular jurisdiction.”); Dan M. Kahan, Privatizing Criminal Law: Strategies for Private Norm Enforcement in the Inner City, 46 UCLA L. Rev. 1859, 1859 (1999) (“We are accustomed to seeing criminal law enforcement as an exclusive state prerogative.”); William M. Landes & Richard A. Posner, The Private Enforcement of Law, 4 J. Legal Stud. 1, 30–31 (1975) (“With few exceptions, there is a public monopoly—more precisely a series of public monopolies—of criminal-law enforcement.”); Ric Simmons, Private Criminal Justice, 42 Wake Forest L. Rev. 911, 918 (2007) [hereinafter Simmons, Private Criminal Justice] (“[T]he provisioning of criminal justice services, at least beyond the field of law enforcement, remains the exclusive province of the state.”). This supposed monopoly is itself of relatively recent vintage. 11 See Simmons, Private Criminal Justice, supra note 10, at 921–23, 971 (describing the historical evolution from private to public criminal law enforcement). And it has never been absolute. Consider, for example, the local diner that lets the neighborhood vandal repay a debt by washing dishes, 12 See generally Ric Simmons, Private Plea Bargains, 89 N.C. L. Rev. 1131 (2011) [hereinafter Simmons, Private Plea Bargains] (discussing various types of informal private bargaining between criminal offenders and victims). or the role of violent vigilantism in our national narrative. 13 See generally Edward L. Ayers, Vengeance and Justice: Crime and Punishment in the 19th-Century American South (1984) (exploring the history of violence and patterns of crime and punishment during slavery and Reconstruction in the American South); Richard Maxwell Brown, Strain of Violence: Historical Studies of American Violence and Vigilantism (1975) (tracing the history of violence and vigilantism from the American Revolution through 1970). Private adjuncts to criminal justice institutions are common, too, like diversion programs or private probation. What is novel here is the way “private justice”—wholly divorced from the criminal justice system—has become routinized and institutional­ized in a mass, for-profit industry, with buy-in from criminal justice actors.

Seminal work by Professors Elizabeth Joh, Ric Simmons, David Sklansky, and others has documented the extent to which “private police” prevent and investigate crime and apprehend suspected offenders. 14 See, e.g., Elizabeth E. Joh, Conceptualizing the Private Police, 2005 Utah L. Rev. 573, 609–15 [hereinafter Joh, Conceptualizing] (creating a typology of policing activities in which private police engage); Elizabeth E. Joh, The Paradox of Private Policing, 95 J. Crim. L. & Criminology 49, 51 (2004) [hereinafter Joh, Paradox] (arguing that “private police participate in much of the policing work that their counterparts do”); Simmons, Private Criminal Justice, supra note 10, at 919–21 (documenting the ubiquity of monitoring and investigating by private police); David A. Sklansky, The Private Police, 46 UCLA L. Rev. 1165, 1178–80 (1999) (discussing activities of private security personnel, including patrol work or acting as “cop[s] on the beat”). Sepa­rate research plumbs the private prison 15 See, e.g., Sharon Dolovich, State Punishment and Private Prisons, 55 Duke L.J. 437, 457–62 (2005) (arguing that private prisons increase the likelihood the state will impose punishments that are inhumane or gratuitously long); Malcolm M. Feeley, The Unconvincing Case Against Private Prisons, 89 Ind. L.J. 1401, 1419–21 (2014) (arguing private prison arrangements may “respect the dignity and minimize the deprivation of liberty of criminal offenders”). and probation 16 See, e.g., Laura I. Appleman, Nickel and Dimed into Incarceration: Cash-Register Justice in the Criminal System, 57 B.C. L. Rev. 1483, 1496–98, 1507–09 (2016) (arguing that private probation companies whose profits derive entirely from fees paid by offenders are a “classic conflict of interest”); Wayne A. Logan & Ronald F. Wright, Mercenary Criminal Justice, 2014 U. Ill. L. Rev. 1175, 1193 (“It is . . . common for profit-oriented private firms to contract with the government to provide probation services.”). See generally Christine S. Schloss & Leanne F. Alarid, Standards in the Privatization of Probation Services: A Statutory Analysis, 32 Crim. Just. Rev. 233 (2007) (examining the history of and statutory requirements for private probation). industries. But private criminal adjudication and sanctions are terra incognita—or maybe El Dorado, mythical altogether. 17 “In the 16th and 17th centuries, Europeans believed that somewhere in the New World was a place of immense wealth known as El Dorado.” Willie Drye, El Dorado, Nat’l Geographic,‌archaeology-and-history/‌archaeology/‌el-dorado [] (last visited Aug. 17, 2018). “Shopkeep­ers do not always report those they have caught,” one recent article begins, “but we have never heard of a shoplifter and storekeeper agree­ing to a payment beyond restitution to settle the matter confiden­tially.” 18 Saul Levmore & Frank Fagan, Semi-Confidential Settlements in Civil, Criminal, and Sexual Assault Cases, 103 Cornell L. Rev. 311, 345 (2018); see also, e.g., Simmons, Private Criminal Justice, supra note 10, at 911, 936 (asserting that “[p]rivate criminal law . . . is currently limited to the law enforcement stage of the process” and that “private entities perform . . . none of the adjudication and almost none of the dispositions in the criminal justice system”); cf. Sklansky, supra note 14, at 1277 (“If we know little about the private police, we know even less about private adjudication.”). One seemingly authoritative volume on “privatizing the United States justice system” is divided into three parts: police, adjudication, and corrections. While the sections on police and corrections are predictably rich, the brief section on adjudication contains only two pages touching on criminal law, which propose that crime victims be permitted to hire attorneys to assist the prosecution. See Tim Valentine, Private Prosecution, in Privatizing the United States Justice System 226, 226–28 (Gary W. Bowman et al. eds., 1992). CEC and its competitors, on the store­keeper’s behalf, do precisely that.

This Article begins, then, as a case study in the routinized, private settlement of a particular type of criminal dispute. The subject offense—shoplifting—is a minor crime with “major economic and social conse­quences.” 19 See, e.g., Paul Cromwell & Brian Withrow, The Dynamics of Petty Crime: An Analysis of Motivations to Shoplift, in Crime Types 242, 243 (Dean A. Dabney ed., 2d ed. 2013) (arguing that shoplifting “should be more widely and systematically studied”). The New York Times has called shoplifting “the nation’s most expensive crime.” 20 Susan Konig, Helping Shoplifters to Reform, N.Y. Times (Sept. 29, 1996), https://​ (on file with the Columbia Law Review); see also Rachel Shteir, The Steal: A Cultural History of Shoplifting 93 (2011) (noting that criminologist Ronald V. Clarke has called shoplifting “one of the ‘most common but least reported and detected crimes’ in the world”). Retailers’ losses from shoplifting approached $18 billion in 2016, or almost $50 million every day. 21 See Richard Hollinger, Nat’l Retail Fed’n, 2017 National Retail Security Survey 6, 8 (2017) [hereinafter Hollinger, 2017 Survey], []. The effects of shoplifting reach both far and deep. At least one in nine Americans shoplifts at some point in life; 22 Carlos Blanco et al., Prevalence and Correlates of Shoplifting in the United States: Results from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), 165 Am. J. Psychiatry 905, 909 (2008); see also id. at 911 (advising that this estimate “may represent a lower boundary of the true prevalence of shoplifting”). more than ten million people have been picked up in the last five years alone. 23 Shoplifting Statistics, Nat’l Ass’n for Shoplifting Prevention, http://www.​‌what-we-do/‌learning-resource-center/‌statistics [] (last visited Aug. 10, 2018). Shoplifting is a “crime of moral turpitude” that can catalyze exclu­sion or deportation for a noncitizen offender. 24 See, e.g., Bill Ong Hing, Deported for Shoplifting?, Wash. Post (Dec. 29, 2002), https://​​1a30b527-195a-433f-9d55-77e86fa4894e []; John E. Hogan & Amy B. Herbold, Collateral Consequences: The Potential for Deportation and Exclusion as a Result of a Municipal Court Shoplifting Conviction, N.J. Law. Mag., Dec. 2010, at 25, 25. And it famously triggered a life term for Gary Ewing, who left a pro shop with a trio of golf clubs lining his pants leg. 25 See Ewing v. California, 538 U.S. 11, 17–18, 20 (2003) (upholding a twenty-five year to life imprisonment sentence under California’s recidivist statute).

In addition to its parochial significance, this private justice industry raises and informs broader questions of legal theory and practice. The Article explores these questions, traveling from the local to the global across four parts. Part I begins with a social history of shoplifting and review of the pertinent criminological literature. The focus is on who offends, why they do it, and how industry and the legal system have traditionally responded. Each of these inquiries informs the normative analysis the Article later undertakes: The “who” identifies the population principally affected by retail justice, necessary, among other things, to weigh distributive effects. The “why” helps predict how shoplifters would likely react to various deterrent measures. And the “how” reveals the baseline against which to evaluate retail justice.

Part I then documents how retail justice has transformed the practice of shoplifting enforcement to date. Shifting legal and economic pressures make the retail justice universe highly dynamic. The facts described here are best understood as a snapshot of this nascent industry and its most prominent pioneer. By the time this Article goes to print, in fact, CEC may be defunct. 26 As of July 30, 2018, CEC’s website was no longer publicly accessible. For an archive of the previously accessible website, see CEC, supra note 2. There are signs of reincarnation, however, 27 CEC’s LinkedIn page describes a company called 3 Peak Solutions whose business model sounds identical to CEC’s. 3 Peak Solutions, LinkedIn, (on file with the Columbia Law Review) (last visited July 30, 2018). and CEC’s principal competitor continues to do business. 28 See Turning Point Justice, [] (last visited Aug. 10, 2018). Considering the forces that gave rise to retail justice in the first place, evolution, not extinction, is the most probable future course.

Part II conducts a preliminary evaluation of the retail justice system. If the media’s reception is any indication, the very concept will make some readers squirm. Likewise, the sole judicial decision on point characterizes the “irreducible core of CEC’s program” as “textbook extortion under California law.” 29 People ex rel. Herrera v. Corrective Educ. Co., No. CGC-15-549094, slip op. at 3 (Cal. Super. Ct. Aug. 14, 2017) (granting partial summary judgment in favor of plaintiff). These views are understandable but, upon deeper reflec­tion, more problematic than they first appear.

The “criminal compromise agreement” (or “restorative justice agree­ment”) the parties sign is a contract that, according to the standard Pareto assumption, should make them better off. 30 See Oren Bar-Gill & Omri Ben-Shahar, The Prisoners’ (Plea Bargain) Dilemma, 1 J. Legal Analysis 737, 738 (2009) [hereinafter Bar-Gill & Ben-Shahar, Plea Bargain] (discussing “the standard Pareto argument that a contract entered into freely by two parties necessarily improves the situation of both parties”). The Pareto assump­tion fails, however, when the conditions for efficient contracting are absent—if suspects are coerced or misinformed, for example, or if negative externali­ties, such as insufficient (or inefficient) general deter­rence, result. Part II exhaustively—yet tentatively, given our incomplete understanding of this evolving industry—analyzes these potential “market failures,” sidelining some and flagging others for lawmakers’ attention. Part II also considers the potential distributive effects of retail justice—whether we should expect its harms and benefits to be visited equally upon different social groups. Part II’s recurring theme is that, while retail justice may not be ideal, it may still be preferable to criminal justice. Private justice, in fact, is the predictable result of, and a potential palliative for, aggressive policing and harsh criminal penalties.

That is not to say the California court was wrong in concluding that retail justice meets the legal test for extortion, or blackmail. After all, the retailer allows the suspect to pay money in exchange for a promise not to report a crime. The justifications for criminalizing blackmail, however, do not support a ban on retail justice. Even if retail justice is blackmail, in other words, it should not be illegal. Regardless of what one thinks about black­mail generally—a question probed in an extensive legal and philosophical literature—the reasons for its prohibition are at their weakest in this setting. This argument plows no new ground. But it does show why the normative debate about blackmail matters more than previously thought. Rather than an occasional victim who seeks compensation by threatening to accuse the perpetrator of a crime, retail justice is a burgeoning, large-scale industry. Primarily “academic” debates about blackmail are academic no longer.

The upshot of Part II is that the normative valence of retail justice depends upon empirical facts about its implementation and the environments in which it operates. Rather than ignoring retail justice or trying to stamp it out, lawmakers can direct their efforts to ensuring that it works fairly and efficiently. Part III briefly makes recommendations toward this end, focused on retail justice companies’ communication with suspects, retailers’ crime-prevention initiatives, and the collection and reporting of aggregate crime data that retail justice masks.

Part IV extends. It demonstrates how the study of private justice gener­ates fresh perspectives on important criminal justice issues, such as the understudied role victims play in preventing crime. 31 See Omri Ben-Shahar & Alon Harel, The Economics of the Law of Criminal Attempts: A Victim-Centered Perspective, 145 U. Pa. L. Rev. 299, 303–08 (1996) [hereinafter Ben-Shahar & Harel, Economics] (criticizing legal scholarship’s “perpetrator-centered perspective”). Well-oiled criminal justice institutions, the Article contends, may actually discourage some victims from investing in socially desirable crime-deterring precau­tions. In other words, we may want the criminal justice system to be costly for victims when it would be cheaper for victims to prevent crime by taking precautions than for the public authorities to capture and punish offenders. We also want to concentrate public resources where they will not reduce private incentives to take precautions, such as when victims cannot afford precautions or will suffer irreparable harm from crime, and thus will purchase precautions regardless of public enforcement.

Part IV also highlights what retail justice teaches us about police and prosecutorial preferences. Critics claim that private justice usurps the prosecutor’s charging prerogative. Yet retail justice companies operate with the knowledge and (at least tacit) approval of criminal justice authori­ties. Where retail justice reigns, prosecutors still exercise discretion—they simply do so at a wholesale rather than retail level. The arrangement is more decriminalization than abdication. That prosecutors are willing to forgo so many easy cases, moreover, complicates academic models of prosecutors who seek to maximize convictions or conviction rates. 32 See, e.g., George Fisher, Plea Bargaining’s Triumph 48–49 (2003); Albert W. Alschuler, The Prosecutor’s Role in Plea Bargaining, 36 U. Chi. L. Rev. 50, 106–07 (1968); Stephanos Bibas, Plea Bargaining Outside the Shadow of Trial, 117 Harv. L. Rev. 2463, 2471 (2004); Josh Bowers, Punishing the Innocent, 156 U. Pa. L. Rev. 1117, 1128 (2008) [hereinafter Bowers, Punishing the Innocent]; George T. Felkenes, The Prosecutor: A Look at Reality, 7 Sw. U. L. Rev. 98, 114 (1975); Sanford C. Gordon & Gregory A. Huber, The Political Economy of Prosecution, 5 Ann. Rev. L. & Soc. Sci. 135, 143 (2009); Robert L. Rabin, Agency Criminal Referrals in the Federal System: An Empirical Study of Prosecutorial Discretion, 24 Stan. L. Rev. 1036, 1045, 1071 (1972); Eric Rasmusen et al., Convictions Versus Conviction Rates: The Prosecutor’s Choice, 11 Am. L. & Econ. Rev. 47, 52 (2009); Stephen J. Schulhofer, Criminal Justice Discretion as a Regulatory System, 17 J. Legal Stud. 43, 50–52 (1988); Alissa Pollitz Worden, Policymaking by Prosecutors: The Uses of Discretion in Regulating Plea Bargaining, 73 Judicature 335, 337 (1990). Other models assume that prosecutors seek to maximize sentence years. See, e.g., Oren Bar-Gill & Oren Gazal Ayal, Plea Bargains Only for the Guilty, 49 J.L. & Econ. 353, 357 (2006); Richard T. Boylan, What Do Prosecutors Maximize? Evidence from the Careers of U.S. Attorneys, 7 Am. L. & Econ. Rev. 379, 380 (2005); William M. Landes, An Economic Analysis of the Courts, 14 J.L. & Econ. 61, 63 (1971); Jennifer F. Reinganum, Sentencing Guidelines, Judicial Discretion, and Plea Bargaining, 31 RAND J. Econ. 62, 69 (2000). For a skeptical take on these models, see Daniel C. Richman, Old Chief v. United States: Stipulating Away Prosecutorial Accountability?, 83 Va. L. Rev. 939, 987–89 (1997). The same is true for police and arrests. 33 See, e.g., Inimai Chettiar et al., Brennan Ctr. for Justice, Reforming Funding to Reduce Mass Incarceration 24–25 (2013),‌sites/default/‌files/‌publications/‌REFORM_FUND_‌MASS_INCARC_web_0.pdf []; Jerome H. Skolnick & James J. Fyfe, Above the Law: Police and the Excessive Use of Force 189–90 (1993); Stephanos Bibas, Improve, Dynamite, or Dissolve the Criminal Regulatory State?, in The New Criminal Justice Thinking 61, 61 (Sharon Dolovich & Alexandra Natapoff eds., 2017) [hereinafter The New Criminal Justice Thinking]; Josh Bowers, Legal Guilt, Normative Innocence, and the Equitable Decision Not to Prosecute, 110 Colum. L. Rev. 1655, 1695 (2010) [hereinafter Bowers, Legal Guilt]; Stephen D. Mastrofski et al., Expectancy Theory and Police Productivity in DUI Enforcement, 28 Law & Soc’y Rev. 113, 117 (1994); Elina Treyger, Collateral Incentives to Arrest, 63 U. Kan. L. Rev. 557, 564 (2015); cf. Dhammika Dharmapala et al., Punitive Police? Agency Costs, Law Enforcement, and Criminal Procedure, 45 J. Legal Stud. 105, 107–08 (2016) (developing a model showing that police officers have unusually strong preferences for punishment and thus derive utility from making arrests).

Finally, Part IV begins to generalize and speculates about the future of private justice. CEC’s ambitious “vision is to reinvent the way petty crimes are handled, starting with retail theft.” 34 Appellant’s Opening Brief at 11, People ex rel. Herrera v. Corrective Educ. Co., No. A149195, 2017 WL 1366020 (Cal. Ct. App. Apr. 13, 2017), 2016 WL 6037455 (emphasis added). Part IV considers the conditions that conduce to a model of “offender-funded” private justice like CEC’s, to begin to identify where else the model might work. An “offender-funded” model is best supported when a small group of victims each suffers a large number of low-level, nonviolent harms by known offenders, and the existing options for deterring those harms are flawed. Part IV also suggests that other large institutions, like universities and employers, might support a distinct model of private justice that outsources adjudication rather than sanctions.

In the end, understanding private justice sharpens our view of the criminal justice system. And understanding the criminal justice system—in all its manifold institutions, including the unconventional ones at its margins—is an essential step toward fixing it. 35 See Sharon Dolovich & Alexandra Natapoff, Mapping the New Criminal Justice Thinking, in The New Criminal Justice Thinking, supra note 33, at 1, 1 (“If we are to fix the current criminal system . . . we need a complete and nuanced understanding of what exactly this system is: What social and political institutions, what laws and policies, does it encompass?”). To be sure, private justice sits uncomfortably in the contemporary criminal justice landscape. Its ends, however, may turn out to justify its means. After all, the public system’s severity is the private system’s sustenance. The way out of private justice, for those who so desire, is not to squelch it but to starve it. Mollifying the criminal justice system would reduce both suspects’ demand for private alternatives and the ability of private intermediaries to extract rents in the form of hush money.

I. The Path to Private Justice

Leading economic and criminological theories struggle to explain the incidence of shoplifting because, unlike many crimes, shoplifting steamrolls race, class, age, and gender lines. 36 See Cromwell & Withrow, supra note 19, at 242 (“[Shoplifting] is widely distributed in the population and appears to cross racial, ethnic, gender and class lines.”). To sketch out a sense of who shoplifts and why, section I.A thus begins with a brief social history of shoplifting and overview of the criminological literature. This discussion identifies the individuals potentially affected by retail justice and helps to predict how they will react to the incentives retail justice, or its alternatives, provide. Section I.B describes how retailers and criminal justice authorities have traditionally responded to shoplifting, identifying the baseline against which to evaluate retail justice. Section I.C details how retail justice works and what it claims to accomplish.

A. Who Shoplifts and Why?

A caveat is required at the outset. It is difficult to determine, at any point in history, who is shoplifting and how much. Experts draw inferences from three imperfect sources: store apprehension statistics, criminal justice data, and self-report studies. Changes in the first two measures may reflect shifts in either commission or detection of shoplift­ing. Self-report studies may be more reliable, though respondents’ incentives to over- or underreport may vary with cultural norms (over time or among social groups) or even the manner in which a survey is administered. 37 See Lloyd W. Klemke, The Sociology of Shoplifting: Boosters and Snitches Today 7–9, 34–37 (1992) [hereinafter Klemke, Sociology of Shoplifting].

Specifically, every era since the 1870s has experienced a supposed shoplifting “epidemic.” 38 Id. at 19. Yet it is unclear, in each period, whether people were shoplifting more or retailers were catching them more often. Simi­lar difficulties plague the “who” question: There is a “range and variety of selec­tive factors that bring about the [apprehension] of a shoplifter and perhaps bring him to official attention.” 39 Mary Owen Cameron, The Booster and the Snitch: Department Store Shoplifting 25 (1964). Cameron’s study of shoplifting in 1940s Chicago department stores found that black shoppers were “kept under much closer observation than whites” and adolescents were “under almost constant observation.” Id. at 31. Such scrutiny biases upwards estimates of black and adolescent offending. As one commentator put the point, due to “self-fulfilling prophecies” about criminality, “the shoplifting statistics ‘cre­ated’ by security personnel may not accurately reflect shoplifting reality.” 40 Klemke, Sociology of Shoplifting, supra note 37, at 120.

These difficulties are handled in two ways. First, when possible, multiple data types and sources are used to triangulate the facts. Second, when sources conflict, descriptions are hedged accordingly.

1. A Century of Petty Theft. —Shoplifting first captured public attention shortly after the Civil War, as department stores proliferated. 41 See, e.g., Elaine S. Abelson, When Ladies Go A-Thieving: Middle-Class Shoplifters in the Victorian Department Store 4–5 (1989) [hereinafter Abelson, Middle-Class Shoplifters]; Kerry Segrave, Shoplifting: A Social History 3 (2001). In this new retail setting, shop owners could no longer monitor the entire premises, forcing them to rely on clerks who had relatively weaker incen­tives to prevent theft. At the same time, customers were newly allowed to browse unsupervised and goods were moved to open display, making them easier to secret away. 42 See Segrave, supra note 41, at 17–18. Women were thought to do most of the pilfering. 43 See Pilfering from Stores, N.Y. Times, Mar. 4, 1883, at 5, 5,​1883/03/04/archives/pilfering-from-stores-the-methods-and-results-of-shoplifting.html (on file with the Columbia Law Review) (quoting a store owner who estimated that, in 1883, “[f]ully nineteen-twentieths” of shoplifters were female). The best academic treatment is Abelson, Middle Class Shoplifters, supra note 41. Notorious shoplifters of the era included ladies with colorful pseudonyms like “light-fingered Sophie Lyons,” “Long Mary Moore,” “Frenchy Johnson,” “Black Lena” (who was white), and “Kid Glove Rosie.” See Segrave, supra note 41, at 3–6. “[F]emale fashion,” it was said, “afforded a lot of spacious hiding places for articles,” giving female thieves a technological advan­tage over men. 44 Segrave, supra note 41, at 3 (“Women carried purses of various sizes, wore outfits with long voluminous skirts, and were often decked out in shawls, gloves and muffs.”); see also id. at 15 (“One woman had a hollow heel fashioned in her shoes, another had puffs of hair lacquered to extra stiffness, to act as a receptacle for small items.”). Doubt­less more important, department stores success­fully cultivated an almost exclusively female customer base. 45 In 1904, for example, Macy’s estimated that as many as ninety percent of its patrons were women. Elaine S. Abelson, The Invention of Kleptomania, 15 Signs 123, 136 (1989).

Women charged with shoplifting often “accused the stores of permit­ting too much freedom: they became ‘over excited’ and over stimulated in the large stores[,]” which afforded them the “‘deplorable liberty’ to touch everything.” 46 Id. at 139 (quoting Paul Dubuisson, Les Voleuses des Grands Magasins, 16 Archives d’Anthropologie Criminelle 1, 342 (1901)). Some medical experts agreed—items on open display, they argued, were temptations “better than Satan himself could devise.” 47 Patricia O’Brien, The Kleptomania Diagnosis: Bourgeois Women and Theft in Late Nineteenth-Century France, 17 J. Soc. Hist. 65, 72 (1983) (internal quotation marks omitted) (quoting Paul Dubuisson, Les Voleuses des Grands Magasins 43 (1902)); see also Michael B. Miller, The Bon Marché: Bourgeois Culture and the Department Store, 1869–1920, at 200–05 (1981) (describing contemporary, nineteenth-century medical opinion about the irresistible nature and degrading effects of department stores on human—especially female—morality). These women, moreover, were increasingly “well-to-do, of good character.” 48 Segrave, supra note 41, at 7, 18–19; see also, e.g., Shoplifting in the Great Department Stores, N.Y. Times, Apr. 26, 1908, at 64, 64, https://timesmachine.​1908/04/26/104802462.pdf (on file with the Columbia Law Review) [hereinafter Shoplifting in the Great Department Stores] (“The shoplifters that we are afraid of are not the professional thieves, nor the poor people who steal from need . . . . The dangerous ones are the rich and influential women who either yield to a temporary impulse of temptation or are afflicted with a sort of degenerate tendency toward kleptomania.” (internal quotation marks omitted) (quoting a shop detective)).
Many retailers overlooked petty thefts by wealthy women or allowed the offenders to pay their way out of trouble. 49 See Segrave, supra note 41, at 7–8, 11 (describing how upper-class women in the United States and abroad were treated when arrested for shoplifting). Kleptomania—a “distinctive, irresistible tendency to steal,” 50 O’Brien, supra note 47, at 70 (internal quotation marks omitted) (quoting C.C.H. Marc, a French physician). thought principally to afflict women—came in and then out of fashion as a defense. 51 See Segrave, supra note 41, at 18–26. Individuals with kleptomania are now understood to constitute a small proportion of shoplifters and a very small fraction of the public. See, e.g., Brian L. Odlaug & Jon E. Grant, Impulse-Control Disorders in a College Sample: Results from the Self-Administered Minnesota Impulse Disorders Interview (MIDI), 12 Primary Care Companion J. Clinical Psychiatry, no. 2, 2010 (finding that, although 28.6% of college students surveyed reported having stolen an item in their lifetime, only 0.4% met the diagnostic criteria for kleptomania). Yet all the way into the 1950s, the middle- or upper-class woman remained the archetypal offender. 52 See, e.g., Segrave, supra note 41, at 18–26; Alex J. Arieff & Carol G. Bowie, Some Psychiatric Aspects of Shoplifting, 8 J. Clinical Psychopathology 565, 566 tbl.1, 567–68 (1947); Cracking Down on Shoplifters, Bus. Wk., Nov. 1, 1952, at 58, 61. Taking a slightly different position, Mary Owen Cameron argues, based on her 1940s downtown Chicago sample, that “well-to-do women shoplift in department stores considerably less frequently than middle- and lower-class women.” Cameron, supra note 39, at 119.

During the 1950s and ’60s, shoplifting gradually came to be seen as an adolescent problem, initially still concentrated among females in the middle and upper classes. 53 See, e.g., Cameron, supra note 39, at 102; Gerald D. Robin, Patterns of Department Store Shoplifting, 9 Crime & Delinq. 163, 167 (1963). For popular accounts, see Bill Davidson, They Steal Just for the Hell of It, Saturday Evening Post, May 18, 1968, at 23, 24; Fredelle Maynard, The Housewives’ Crime, Good Housekeeping, Oct. 1967, at 99, 154. Sociologists had documented adolescent involvement in shoplifting much earlier. See, e.g., Clifford R. Shaw, The Natural History of a Delinquent Career 57–58, 66–68 (1931). “[T]he under-21 group,” wrote one reporter, was then “on the greatest shoplifting spree in our history.” 54 Earl Selby, Youthful Shoplifting: A National Epidemic, Reader’s Dig., Apr. 1967, at 95, 95. A 1965 FBI report called shoplifting the nation’s fastest growing form of larceny, up ninety-three percent over five years. See John Edgar Hoover, Message from the Director, FBI L. Enforcement Bull., Dec. 1965, at 1, 1. Abbie Hoffman’s Steal This Book ushered the trend into the 1970s, 55 Abbie Hoffman, Steal This Book (1971); see also Mass Retailing Inst., Store Thieves and Their Impact 7 (1973) (reporting, based on a survey of 1,188 stores, that a majority of shoplifters were under eighteen years of age). Roughly one-quarter to one-half of juveniles polled admitted to having pilfered. Compare Changing Morality: The Two Americas, Time, June 6, 1969 (reporting that 23% of Americans admit to having shoplifted), and Michael D. Geurts et al., Researching Shoplifting and Other Deviant Customer Behavior Using the Randomized Response Research Design, 51 J. Retailing, Winter 1975–1976, at 43, 46 (reporting that 28.2% of survey participants with characteristics representative of young adults between the ages of fourteen and twenty-eight in Honolulu had shoplifted during the prior year), with Nat’l Coal. to Prevent Shoplifting, Shoplifting and the Law 8 (1980), https://​‌pdffiles1/‌Digitization/‌81811NCJRS.pdf [] (citing studies reporting that 49% of high school students admitted to having shoplifted at least once), Amin el-Dirghami, Shoplifting Among Students, 50 J. Retailing 33, 33 (1973) (citing a study reporting that roughly half of shoplifters are under twenty years of age), and Lloyd W. Klemke, Exploring Juvenile Shoplifting, 67 Soc. & Soc. Res. 59, 61 (1982) [hereinafter Klemke, Juvenile Shoplifting] (finding a majority of 1,189 high-school students had shoplifted at some point in their lives). when, for the first time, researchers also began to find that male offenders outpaced females. 56 See, e.g., Klemke, Juvenile Shoplifting, supra note 55, at 61 & tbl.1; Shoplifting Keeps Pace with Inflation, Security Mgmt., July 1978, at 27, 29 exh. 1. For a review of numerous additional studies from the 1970s and ’80s, see Klemke, Sociology of Shoplifting, supra note 37, at 47–50. In the 1980s, the number of thefts known to the police skyrocketed, though the causes are unclear. 57 Because self-reported shoplifting statistics held steady in the 1980s, Klemke concludes that “the increase shown in [known thefts] is more likely to be a product of changes in apprehension and reporting practices than a real increase in shoplifting behavior.” Klemke, Sociology of Shoplifting, supra note 37, at 8.

Early research on the racial and ethnic breakdown of shoplifters is scarce. Many stores did not collect (or release) these data and the premier shoplifting datasets excluded them. 58 See id. at 50–51. A handful of studies from the 1970s and ’80s found similar patterns of shoplifting activity across racial groups of youth. 59 See id. at 54 (reviewing studies); Saul Astor, Shoplifting: Far Greater than We Know?, Security World, Dec. 1969, at 12, 12 (“[R]ace seems to have nothing to do with theft.”). A single self-report study on adults found higher levels of general theft behavior among nonwhites than whites. 60 See Charles R. Tittle, Sanctions and Social Deviance: The Question of Deterrence 87 tbl.4.3 (1980).

2. Contemporary Evidence. —Shoplifting remains widespread today—recall that more than ten percent of the population has shoplifted at least once, generating $50 million in losses each day. 61 See supra text accompanying notes 21–23. On the “who” question, the modern view may consist of the unhelpful observation that “there is no ‘typical shoplifter.’ . . . [S]hoplifters come from varying social, age, and economic groups.” 62 Thomas Brad Bishop, The Law of Shoplifting: A Guide for Lawyers and Retail Merchants 6 (2007). Very likely there is geographic variation as well. Nevertheless, the best data support a few tentative generalizations. Two sources are especially useful: (1) sociologist Lloyd Klemke’s review of the social science literature through 1992; and (2) the National Epidemiological Survey on Alcohol and Related Conditions (NESARC), a large-scale, nation­ally representative survey from 2001 to 2002.

First, “there is a great deal of consensus that shoplifting is most fre­quent in the early part of the life cycle and that it declines as individuals move through the life cycle.” 63 Klemke, Sociology of Shoplifting, supra note 37, at 44. NESARC, for example, found that two-thirds of shoplifting cases occur before age fifteen. 64 Blanco et al., supra note 22, at 911.

Second, contemporary evidence “appears to overwhelmingly sup­port the conclusion that males are typically more active in shoplifting than fe­males.” 65 Klemke, Sociology of Shoplifting, supra note 37, at 50 (summarizing research). NESARC found that nearly sixty percent of shoplifters were male. 66 Blanco et al., supra note 22, at 906 tbl.1.

Third, “racial and ethnic patterns of who shoplifts” seem to “vary dramatically in different places and times.” 67 Klemke, Sociology of Shoplifting, supra note 37, at 55. “The limited research on race and ethnic variations in shoplifting,” Klemke finds, “suggests that only minor differences are evident in the population at large.” 68 Id. at 64. NESARC shows some­thing slightly different: “Native Americans had higher odds [of shoplifting] than whites, although blacks, Hispanics, and Asian Americans had lower odds of shoplifting than non-Hispanic whites.” 69 Blanco et al., supra note 22, at 909. Native-born Americans also reported shoplifting at higher rates than those who are foreign born. 70 See id. at 906 tbl.1.

Finally, the bulk of the evidence suggests that middle-class individu­als are most likely to shoplift. In NESARC, shoplifting “was significantly more common in individuals with at least some college education, among those with individual incomes over $35,000 and family incomes over $70,000, and . . . less common among those with public insur­ance.” 71 Id. at 909; see also Bishop, supra note 62, at i (“There are very few people stealing, as in Les Misérables, for a loaf of bread.”); Cameron, supra note 39, at 118 (reporting that apprehensions were not concentrated on individuals from the “slum and ‘ghetto’ areas of Chicago”). Still, Klemke does find “slight to moderate inverse relationships between social class and shoplifting behavior,” suggesting the truly wealthy are infrequently involved. 72 Klemke, Sociology of Shoplifting, supra note 37, at 64.

As for the cause, the evidence is mixed. Klemke, a sociologist, reads the evidence to support sociological explanations, while the psychiatrists inter­preting NESARC emphasize psychiatric ones. “[I]t is highly likely,” Klemke begins, “that some shoplifters fit the pathological conception, others are best seen as societal victims, and many others fit the frugal customer conception”—that is, their motivation “‘is the same as for normal shopping: the acquisition of goods at minimum cost.’” 73 Id. at 78 (quoting Robert E. Kraut, Deterrent and Definitional Influences on Shoplifting, 23 Soc. Probs. 358, 365 (1976)). But in general, Klemke writes, recent studies “conclude that most shoplifters are characterized by relatively normal psychological health and personalities that are indistin­guishable from non-shoplifters.” 74 Id. at 86 (emphasis omitted). They are not professionals “boosting” goods for resale. Klemke concludes that sociological theories stressing the individuals’ relationship to their environment can best explain who offends. 75 On the structural (macro) level, there is some evidence consistent with strain theory, which views deviance as animated by frustration with barriers to economic success. See id. at 87–88, 93. On the individual (micro) level, both economic and noneconomic motivations (such as peer pressure and sporting) seem to play a role. See id. at 88–93. And among juveniles, “shoplifting . . . is more frequently committed by youth who are less strongly bound to the social order (family and school).” Id. at 97. Further, shoplifting is more frequently committed by youth whose peers endorse the behavior. Id. at 97–105 (applying socialization reinforcement theories to shoplifting).

The NESARC data, however, challenge the notion that shoplifters resemble nonshoplifters along psychiatric dimensions. Researchers found that “[t]he prevalence of all antisocial behaviors was higher among individuals with a history of shoplifting than among those with no self-reported history of shoplifting.” 76 Blanco et al., supra note 22, at 909. In addition, almost ninety percent of individuals with a history of shoplifting had received at least one psychiatric diagnosis, compared to around fifty percent among nonshoplifters. Id. at 910. And because many of the behaviors asso­ciated with shoplifting can be “understood as a manifestation of impulsivity,” the authors concluded, “our findings are most consistent with the understanding of shoplifting as a behavioral manifestation of impaired impulse control.” 77 Id. at 911; see also Nat’l Coal. to Prevent Shoplifting, Program Guide 6 (1980),‌pdffiles1/‌Digitization/‌74730NCJRS.pdf [] [hereinafter Nat’l Coal. to Prevent Shoplifting, Program Guide] (finding that, of almost 25,000 students who had shoplifted, nearly seventy percent claimed to have decided to steal only after they were in the store).

B. Private and Public Enforcement of Shoplifting Laws

As societal understandings of who shoplifts and why have evolved, so have industry and state responses to the crime. This section traces the path that led to the creation of a market for retail justice companies. As is shown, the enforcement model has long been shot through with ambivalence and discretion. Recognizing this reality is crucial when evaluating the changes that retail justice has wrought.

1. Ambivalence and Innovation. — From the earliest public reports of shoplifting, retailers have been fickle and ambivalent about formal law enforcement. Shoplifting hurts the bottom line, but overly aggressive enforcement can too. Wrongful arrests can trigger lawsuits, 78 As early as 1878, the New York Times reported on a $150 damages award to a woman who had been wrongly accused of stealing a purse. See Editorial, N.Y. Times, May 28, 1878, at 8, 8, (on file with the Columbia Law Review). and even legitimate arrests might hurt business by scaring away customers who fear being falsely accused. 79 See The Woman Who Pilfers, N.Y Times, May 31, 1878, at 3, 3, https://timesmachine.​ (on file with the Columbia Law Review) (quoting a shopkeeper who heard from customers that they were almost too afraid of false accusations to come into his store). Many Progressive Era retailers, for these reasons, pressed charges only selectively. 80 One defense lawyer estimated in 1906 that, of the 4,000 individuals arrested for shop- lifting in New York each year, only about 700 made it into court and only 50 were convicted. Shoplifting in New York, N.Y. Times, Jan. 2, 1906, at 15, 15, http://timesmachine.​ (on file with the Columbia Law Review). They hired store detectives to help spot known shoplifters and pooled information with other stores. 81 See Segrave, supra note 41, at 12. Although retailers periodically resolved to toughen up, 82 See, e.g., Shoplifting in the Great Department Stores, supra note 48. they mostly released first-time suspects after making a record of the offense. 83 One major trade group in New York amassed a database of 55,000 known shoplifters. See Maude Miner Hadden, Shoplifters of Many Types Mingle with the Shoppers, N.Y. Times, Dec. 17, 1933, at 15, 15,​105831556.pdf (on file with the Columbia Law Review). Wealthy women, in particular, were often able to buy their way out of prosecution. 84 See George C. Henderson, Light-Fingered, Sunset Mag., Jan. 1927, at 14, 14.

By the 1950s, retailers could take advantage of new loss-prevention technologies like closed-circuit cameras. 85 See Segrave, supra note 41, at 46, 48. States, too, began to enact “mer­chant’s privilege” laws, shielding retailers from suit for false arrest as long as probable cause supported a suspect’s apprehension. 86 See id. at 47. Within ten years, almost every state had one. 87 Id. at 58. As crime rates then ballooned in the 1960s and ’70s, states raised criminal penalties 88 See, e.g., Nev. Anti-Shoplifting Comm., What to Do About Shoplifters 5 (1977), https://​‌pdffiles1/‌Digitization/‌40662NCJRS.pdf [] (describing a new statute authorizing up to ten years in prison and a fine of up to $5,000 if the property stolen is worth more than $100). and enacted “civil recovery” statutes authorizing retailers to obtain super-restitutionary dam­ages. 89 See, e.g., 1976 Cal. Stat. 5047 (codified at Cal. Penal Code § 490.5) (providing that an unemancipated minor who steals may be held civilly liable to the merchant for between fifty and five-hundred dollars, plus costs and the value of the item if it is not returned in its original condition). Arrests rose, too, 90 The number of shoplifting arrests essentially tripled between 1970 and 1976. Tis the Season to Be Wary, Time (Dec. 12, 1977),
article/​0,9171,915813,00.html (on file with the Columbia Law Review). And the number of retail thefts known to the police more than doubled between 1973 and 1980. Klemke, Sociology of Shoplifting, supra note 37, at 8 tbl.1; see also Rising Wave of Shoplifting—and No Solution in Sight, U.S. News & World Rep., Mar. 2, 1970, at 56, 56 (explaining how shoplifting “in all sections of the U.S. is going up at an alarming rate”).
despite lingering retailer ambivalence about justice-system involve­ment. 91 See, e.g., Inst. for Local Self-Gov’t, Private Security and the Public Interest 97 (1974) (“In our survey involving [private police], . . . 80% reported there were certain types of criminal incidents which were not reported to the police. These included . . . shoplifting . . . . [T]he most common practice in this private system of justice is to . . . release suspected shoplifters or maintain private listings of known criminals, especially shoplifters.”); Leonard E. Daykin, Your Profit May Be in the Customer’s Pocket, Progressive Grocer, Sept. 1968, at 54, 55, 60 (discussing a survey of 2,000 grocers finding that 27.8% always prosecuted, 34% occasionally prosecuted, 27% seldom prosecuted, and 17.2% never prosecuted shoplifters); see also Michael J. Hindelang, Decisions of Shoplifting Victims to Invoke the Criminal Justice Process, 21 Soc. Probs. 580, 583 tbl.1 (1974) (finding that the average rate of police referral during the 1960s hovered below 30%).

By 1988, thirty states had passed civil recovery laws, which typically granted retailers a substantial civil penalty in addition to actual damages. 92 See Delany J. Stinson, Attention Retailers: Civil Law Provides a Tonic, Security Mgmt., Sept. 1988, at 129, 129. Retailers—or specialist firms that serviced them—sent formal demand letters to suspected offenders, followed, when necessary, by suit in small claims court. 93 See id. at 131. For some retailers, civil recovery replaced criminal prosecu­tion, but others sought both remedies simultaneously. 94 See Richard C. Hollinger et al., 1998 National Retail Security Survey: Final Report 34 (1998) [hereinafter Hollinger et al., 1998 Survey]; see also Audrey Aronsohn, Teaching Criminals the Cost of Crime, Security Mgmt., May 1999, at 63, 63 (urging retailers to use both civil and criminal remedies whenever possible); Ryan P. Sullivan, Survey of State Civil Shoplifting Statutes (Apr. 26, 2016) (unpublished manuscript at 1),​abstract=2770583 (on file with the Columbia Law Review) [hereinafter Sullivan, Civil Shoplifting] (“[M]ost large retailers pursue criminal charges and a civil remedy.”). A 1998 survey found that retailers employed civil recovery around thirty percent of the time—roughly half the rate at which they sought criminal prosecution. 95 Hollinger et al., 1998 Survey, supra note 94, at 2–3. Today, every state has a civil recovery statute. The authorized recovery is typically $200 to $300 but can exceed $1,000, not counting attorney fees. 96 Sullivan, Civil Shoplifting, supra note 94, at 4; Bruce Mohl, Retailers’ Message to Shoplifters: Pay Up or Risk Prosecution, Bos. Globe (Dec. 11, 2005), (on file with the Columbia Law Review).

2. The Persistence of Discretion. —Retailers today remain reluctant to call the police, often opting instead to exploit their property rights to sanction suspected thieves. 97 See, e.g., Joh, Conceptualizing, supra note 14, at 590 (“The private police department of Macy’s department store, in New York City, . . . reported to the public police only fifty-six percent of the 1900 people accused of shoplifting that it processed in a single year through its private detention center.”). Walmart has banned some shoplifters from all 4,540 of its locations. See Al Norman, Banned from 4,540 Walmarts, HuffPost (June 26, 2015),‌al-norman/‌banned-from-4540-walmarts_b_7147414.html []. In some instances, retailers have allegedly exceeded their property rights and levied nonrestitutionary fines, but this practice appears to have been limited. See, e.g., Brae Canlen, Insecurity Complex, Cal. Law., June 1998, at 30, 81 (“[S]everal parents and guardians of teenagers who were picked up for shoplifting [at Disneyland] claimed they were asked to pay a $275 to $500 fine to avoid criminal prosecution.”). But why? Theft, after all, is a classic, black-letter crime—a perfect fit, one might think, for criminal-justice-system attention. And retailers plainly regard shoplifting as a major trouble.

A few possibilities have been mentioned already, such as fear of suit for false arrest or of alienating customers. 98 See supra notes 78–80 and accompanying text; see also Alan D. Axelrod & Thomas Elkind, Note, Merchants’ Responses to Shoplifting: An Empirical Study, 28 Stan. L. Rev. 589, 589–90 (1976) (describing retailers’ stated reasons for caution, including “alienation of mistakenly accused customers, the possibility of injury to employees, . . . the costs of surveillance,” and “fears of civil suits for false arrest and false imprisonment” (footnotes omitted)). Surely, though, these concerns are diminished by strong merchants’ privilege protections and improved surveillance capabilities that lower the risk of erroneous accusations. Store security also have ways to minimize any visible disturbance when appre­hending suspects. 99 See Robin, supra note 53, at 164. Part II explores this question more deeply. A quick preview here, however, helps identify the problem that retail justice companies claim to solve.

The starting point is to appreciate the enormity of the challenge shoplifting presents. The sheer number of incidents in some major retailers is staggering. In four Florida counties, for example, Walmart stores—which, for some time, employed a “zero tolerance” policy—called the police 7,000 times in one year on suspected thefts. 100 Zachary T. Sampson et al., Walmart: Thousands of Police Calls. You Paid the Bill., Tampa Bay Times (May 11, 2016),‌projects/‌2016/‌public-safety/‌​walmart-police []. Walmart generated fourteen percent of the Camden, South Carolina, Police Department’s reports in one recent measure, most of which were for shoplifting. Josh Sanburn, Low Prices, High Crime: Inside Walmart’s Plan to Crack Down on Shoplifting, Time (Aug. 15, 2016),‌4439650/‌walmart-shoplifting-crime []. A single Walmart store in Tulsa averages over 1,000 calls per year. 101 Shannon Pettypiece & David Voreacos, Walmart’s Out-of-Control Crime Problem Is Driving Police Crazy, Bloomberg Businessweek (Aug. 17, 2016),​‌features/‌2016-walmart-crime (on file with the Columbia Law Review). A flow of cases this large has two principal effects.

First, it strains criminal justice system resources. “It’s hard to dedicate the manpower to process misdemeanor shoplifters,” explained one police administrator. 102 Emily Gold & Julius Lang, Ctr. for Court Innovation, Diverting Shoplifters: A Research Report and Planning Guide 5 (2012) (internal quotation marks omitted) (quoting Lieutenant Michael Brothers),​Diverting-Shoplifters-508.pdf []; see also Pettypiece & Voreacos, supra note 101 (quoting a Florida police captain who explained that “[t]he constant calls from Walmart are just draining” (internal quotation marks omitted)). A “typical theft costs the average police department over $2,100 to process,” according to one account. 103 Gold & Lang, supra note 102, at 4. Response times can be slow. 104 See Diane Ritchey, Walmart’s ‘Second Chance’ Program for Shoplifters, Security Mag. (Nov. 1, 2016), []. Nor is the bottleneck in the police alone: “There are courts in some of our markets,” reported one major discount retailer, “that tell us not to bring them our casual shoplifters.” 105 Gold & Lang, supra note 102, at 5 (internal quotation marks omitted) (quoting Randall Ferris, Senior Director of Loss Prevention at SUPERVALU). What­ever the reason, “[f]or the criminal justice system players, low-level retail theft often occupies a large percentage of misdemeanor caseloads, clogging the desks of everyone involved.” 106 Id. at 2. Not all shoplifting offenses are misdemeanors, moreover—in some states, the felony threshold zooms by quickly. 107 See Shteir, supra note 20, at 119 (describing the “enhancement” of state laws to allow felony charges in more circumstances).

Second, calling the police in every case taxes retailers as well. “By involving the public criminal justice system, the [retailer] loses control over the process, and the costs—both in time and money—to cooperate with the public police and courts can be significant.” 108 Simmons, Private Criminal Justice, supra note 10, at 925. Retailers are reluctant to have their employees miss work to meet with the police or testify in court, for example. 109 See Gold & Lang, supra note 102, at 4 (discussing the hidden costs of shoplifting for retailers). All the more so because the retailer receives no direct benefit from the offender’s punishment.

3. The Patterns of Discretion. — In the absence of retail justice, the fate of many shoplifting suspects is thus determined not by an exercise of police or prosecutorial discretion but rather by the retailer itself when deciding whether to alert the public authorities. On what basis do retailers make these consequential decisions?

Retailers have long employed “no prosecution limits,” contacting the police only when the stolen goods exceed some minimal value threshold. 110 See, e.g., Comment, Shoplifting Law: Constitutional Ramifications of Merchant Detention Statutes, 1 Hofstra L. Rev. 295, 297 (1973). Even Walmart, which, as noted, famously employed a “zero tolerance” policy for many years, eventually adopted a dollar-value cutoff—before contract­ing with retail justice companies. 111 See Michael Barbaro, Some Leeway for the Small Shoplifter, N.Y. Times (July 13, 2006), (on file with the Columbia Law Review) (noting Walmart’s decision to “no longer prosecute first-time thieves unless they are between 18 and 65 and steal merchandise worth at least $25”). In the wake of the California CEC litigation, Walmart recently removed the retail justice option from its stores pending review of the company’s theft-reduction programs. See Joe Palazzolo & Sarah Nassauer, Wal-Mart Suspends Shoplifting Punishment that Court Called “Extortion,” Wall St. J. (Dec. 21, 2017),‌articles/‌wal-mart-changes-how-it-punishes-shoplifters-1513864840 (on file with the Columbia Law Review). Strict cutoffs aside, (admittedly dated) research finds that the value of the suspect’s take powerfully predicts whether the case goes public. 112 Lawrence E. Cohen & Rodney Stark, Discriminatory Labeling and the Five-Finger Discount: An Empirical Analysis of Different Shoplifting Dispositions, 11 J. Res. Crime & Delinq. 25, 32 (1974); Melissa G. Davis et al., Private Corporate Justice: Store Police, Shoplifters, and Civil Recovery, 38 Soc. Probs. 395, 404 (1991); Hindelang, supra note 91, at 583; Richard J. Lundman, Shoplifting and Police Referral: A Reexamination, 69 J. Crim. L. & Criminology 395, 396 (1978). The quality of the evidence matters, too, presumably because retailers are reluctant to incur the criminal justice system’s costs when conviction is uncertain, and because they continue to fear liability for wrongful arrest. 113 Cameron, supra note 39, at 34–35; Segrave, supra note 41, at 37.

Researchers disagree on whether and how personal characteristics of the suspect play a role. 114 Disparate findings, of course, may reflect underlying variation among retailers. See, e.g., Dean G. Rojek, Private Justice Systems and Crime Reporting, 17 Criminology 100, 109 (1979) (“The frequency of arrests, along with the age and sex of the alleged offenders were seen to fluctuate with almost wild abandon from store to store.”). All agree that women and men are referred to the police at similar rates. 115 See, e.g., Cohen & Stark, supra note 112, at 35; Hindelang, supra note 91, at 583 & tbl.1; Lundman, supra note 112, at 397–98. Some have found that juveniles are treated leniently. 116 See, e.g., Hindelang, supra note 91, at 593; Lundman, supra note 112, at 398. Others find no age effects. See, e.g., Cohen & Stark, supra note 112, at 36; Rojek, supra note 114, at 109. The evidence on race, however, is sharply conflicted. 117 Compare Robin, supra note 53, at 169 (finding that black shoppers were both disproportionately apprehended and disproportionately referred to the police), and Cameron, supra note 39, at 136–37 (finding that arrested black shoppers were disproportionately prosecuted), with Cohen & Stark, supra note 112, at 34–35 (finding no significant race effects on referral rates when using “appropriate controls” including age and the value of the stolen goods), and Hindelang, supra note 91, at 591–92 (finding no significant race effects on referral rates when controlling for the value of the items stolen). Two additional papers that shared an author, Professor Richard Lundman, also conflicted—the first found a race effect while the second did not. Compare Lundman, supra note 112, at 397–99 (concluding that the race of the shoplifter has an effect on referral rates to the police, even when controlling for the value of the items stolen), with Davis et al., supra note 112, at 406 (finding no race effects in its sample after attempting numerous models).

Class may influence retailer decisions as well. One of the most recent academic studies found that poorer suspects are referred to the police more frequently. 118 See Davis et al., supra note 112, at 406. But cf. Cohen & Stark, supra note 112, at 25 (“Social class is also found to be insignificantly related to shoplifting dispositions, except for the high rate of prosecution of the unemployed.”). The motivation, however, may be neither animus toward the poor nor empathy for the affluent. The study’s authors conclude, instead, that “[s]tore police skim the affluent for civil recovery and ship the less affluent to the public criminal justice system.” 119 Davis et al., supra note 112, at 406. Retailers, in other words, may view civil recovery as the first-best deterrent sanction and resort to criminal justice, a second-best, only when civil recovery will be ineffectual because the suspect is insolvent. In this context, criminal law, just as economic analysis prescribes, essentially functions as tort law for the indigent. 120 See Alon Harel, Economic Analysis of Criminal Law: A Survey, in Research Handbook on the Economics of Criminal Law 10, 16 (Alon Harel & Keith N. Hylton eds., 2012) (noting that economic analysis leads to the conclusion that “monetary sanctions . . . ought to be used in cases where the wrongdoer is wealthy, while imprisonment ought to be used only where monetary sanctions cannot be used due to insolvency of the wrongdoer”).

In fact, in a controlled, experimental setting, with civil recovery out of the picture, retail security investigators were more likely to refer clean, well-dressed offenders for prosecution than dirty, poorly dressed ones. 121 See Andrey Feuerverger & Clifford D. Shearing, An Analysis of the Prosecution of Shoplifters, 20 Criminology 273, 285 (1982). “In accounting for this,” the study’s authors explain, “investigators commented that they were more likely to be sympathetic towards a shop­lifter who appeared to need what he stole than towards a shoplifter who appeared to be quite able to pay for the items involved.” 122 Id. Similarly, customers in a dif­ferent controlled study reported well-dressed shoplifters to store personnel twice as often as poorly dressed ones. 123 Fred Fedler & Bert Pryor, An Equity Theory Explanation of Bystanders’ Reactions to Shoplifting, 54 Psychol. Rep. 746, 746 (1984). As for what happens after the police become involved, see Kenneth Adams & Charles R. Cutshall, Refusing to Prosecute Minor Offenses: The Relative Influence of Legal and Extralegal Factors, 4 Just. Q. 595, 604–05 (1987) (finding that race and gender significantly predict prosecutorial decisions to dismiss charges, though prior arrests and number of charged offenses are stronger predictors).

C. The Rise of Retail Justice

Where many observers saw only a failing system of law enforce­ment—overtaxed, cumbersome, ineffectual, possibly discriminatory, and overly harsh toward those snagged in its net—entrepreneurs saw an opportunity for profit and Pareto improvement. The basic idea can be simply stated: Retail justice offers private settlement of criminal complaints. Instead of calling the police, the retailer looks to a retail justice company, which extracts payment from the alleged offender in exchange for “rehabilitative education” and a promise not to file a criminal complaint. The payment—and possibly the “restorative justice” course—reduce the likelihood that the suspect will offend in the future, providing the same (type of) benefit to the victim that public law enforcement would. The payment and education are nonetheless preferable, from the suspect’s perspective, to contact with the criminal justice system. Neither the victim nor the public authorities spend anything. One leading company touts that the “program enables first-time offenders to correct their mistakes and avoid prosecution,” allowing retail­ers to “reallocate loss prevention resources” and law enforcement to “focus their efforts in more effective ways for their individual communities.” 124 Our Story—Origin of CEC, CEC (Oct. 25, 2016),‌home/‌blog/‌our-story-origin-of-cec (on file with the Columbia Law Review) [hereinafter CEC, Our Story].

Here is how the process works at what appears to be the lead­ing outfit, CEC. CEC is a Utah-based corporation with a national presence, founded by a pair of Harvard Business School graduates in 2010. 125 Leon Neyfakh, Let’s Make a Deal, Slate (Feb. 26, 2015),
articles/news_and_politics/crime/2015/02/shoplifting_at_whole_foods_or_bloomingdale_s_pay_corrective_education_company.html [] [hereinafter Neyfakh, Make a Deal].
CEC has venture-capital backing and reportedly took in $7.6 mil­lion in 2017. 126 See Walmart Class Complaint, supra note 8, at 3; Corrective Education Company Receives Funding from Decathlon Partners, CEC (Feb. 24, 2015),‌home/‌blog/‌corrective-education-company-receives-funding-from-decathlon-partners (on file with the Columbia Law Review). CEC also boasts an impressive client list, which reportedly has included Walmart, Bloomingdale’s, DSW, Abercrombie & Fitch, Burlington Coat Factory, Whole Foods, American Apparel, Goodwill Industries, Sport Chalet, Kroger’s, Sportsman’s Warehouse, and H&M. 127 See Appellant’s Opening Brief, supra note 34, at 13; Neyfakh, Make a Deal, supra note 125.

Store management—not CEC personnel—retain responsibility for monitoring the retail premises and apprehending suspected shoplift­ers. 128 See Declaration of Darrell Huntsman in Support of Opposition to Motion for Summary Judgment at 2, People ex rel. Herrera v. Corrective Educ. Co., No. CGC-15-549094 (Cal. Super. Ct. Aug. 14, 2017) [hereinafter Huntsman Declaration] (“CEC does not apprehend suspected shoplifters, [and] does not determine who is apprehended . . . .”); Neyfakh, Make a Deal, supra note 125. Once apprehended, suspects are brought to a private room and screened for eligibility, typically including a criminal history check. 129 See Huntsman Declaration, supra note 128, at 2–3 (“In order to qualify for CEC’s education program, a determination is typically made as to whether or not the individual has an existing criminal record.”); see also Neyfakh, Make a Deal, supra note 125 (noting offenders are eligible when they have “been deemed significantly low-risk, based on their criminal history, by the store security guard who apprehended them”); CEC, supra note 2 (“CEC Connect™ qualifies offenders through a database of 50 billion records and meets the most rigorous security and privacy standards.”). Those who qualify are given the option to watch a CEC video explaining the company’s “restorative justice” program. 130 Huntsman Declaration, supra note 128, at 3–4 (detailing that “once a retailer determines that an individual should be offered CEC’s program, the retailer’s security personnel” offers the option to watch a video explaining the program). They are told that, if they choose not to complete CEC’s program, “CEC will refer this matter back to the retailer,” which “may pursue other legal rights to seek restitution and resolve this crime at their discretion.” 131 Id. at 4. Signing up costs $500, or $400 for those who pay in one lump sum, with “scholarships” available for the poorest few. 132 See id. at 5–6 (asserting that CEC “offers the program at a discount to more than 3–4% of course takers,” and that “the average financial assistance discount is . . . at least half the total course fee”); Neyfakh, Make a Deal, supra note 125 (“The company accepts credit cards, e-checks, and money orders, and while they offer a ‘scholarship’ program for people below the poverty line, . . . about 85 percent of offenders pay the full fee, and less than 2 percent qualify for a free ride.”). No money changes hands when the contract is signed. Suspects who agree to enroll are free to leave and have seventy-two hours to think it over, consult with a lawyer (if desired), and then pay a $50 deposit. 133 Huntsman Declaration, supra note 128, at 4–5, exh. C, at 1–2, 4. Even after paying the deposit and beginning the program, suspects can terminate the relationship and receive a partial refund. 134 Id. exh. C, at 2.

For years, suspects who enrolled also signed a “Criminal Compromise Agreement” and admitted guilt. 135 See Appellant’s Opening Brief, supra note 34, at 18–19. Today, the contract is styled as a “Restorative Justice Agreement” and no confession is required. 136 Huntsman Declaration, supra note 128, at 7, exh. C, at 15; see also Telephone Interview with Brian Ashton, CEO, CEC (Dec. 4, 2017). The agree­ment states that, if the suspect completes the pro­gram successfully, the retailer “will consider the matter closed for all purposes.” 137 Huntsman Declaration, supra note 128, exh. C, at 3. In particular, the retailer promises not to “pursue the matter with law enforcement” or seek civil recovery. 138 Id. exh. A, at 2. The contract cau­tions, though, that “law enforcement is not bound by this agreement” and covenants that, if the suspect is prosecuted notwithstanding successful completion of CEC’s program, CEC will refund all fees collected. 139 Id. exh. C, at 3.

More than ninety percent of the individuals presented with the choice during CEC’s first four years opted to enroll, generating approxi­mately 20,000 participants. An enrollee’s first actual contact with CEC is typically a call from a “life coach” who reaches out to “tell them about the course and make a payment plan.” 140 Neyfakh, Make a Deal, supra note 125. The core of CEC’s course, which most “students” take online over six to eight hours, was developed by a clinical psychologist and adapted by CEC “for the purpose of rehabilitating shoplifters.” 141 Id. It “focuses on helping accused shoplifters develop life skills, so that they are less likely to reoffend in the future.” 142 Id. On CEC’s own account:

There’s a chapter that helps them understand what could have happened if they’d gone through the traditional process. But after that, we [CEC] give them skills and the ability to actually go out and get a job . . . . These people that are getting apprehended typically haven’t been taught the life principles of how to build a resume, how to be presentable in an interview. They haven’t been given the skills to understand what a budget is, never mind how to manage their money. So as they’re going through the course, they build their own resume, they build their own budget, a work-out plan, an eating plan. 143 Id. (internal quotation marks omitted) (quoting E.J. Caffaro, CEC’s vice president for account management).

The retailer, for its part, saves time processing suspects, says CEC. “Studies have proven a 40% reduction in processing time when using CEC’s platform,” the company claims. 144 CEC, supra note 2. Initially, the retailer also collected a cut of CEC’s fee, around $40, each time it presented a suspect who enrolled. 145 Neyfakh, Make a Deal, supra note 125. CEC has reportedly ended this practice. See Huntsman Declaration, supra note 128, at 5–6. CEC charges the retailer nothing for its services, which it touts as being “completely offender funded.” 146 Neyfakh, Make a Deal, supra note 125 (internal quotation marks omitted) (quoting an earlier version of the CEC website). CEC’s business model may have changed slightly since Neyfakh reported on it. See CEC, supra note 2 (“CEC’s program is offender funded. The technology, database and professional services are all provided at little or no cost to the retailer.”). “Law enforcement agencies have also noticed our impact,” CEC maintains, “seeing as much as a 40 percent drop in the number of calls for service in their communities.” 147 CEC, Our Story, supra note 124; see also Arlington Police’s Walmart Restorative Justice Initiative Results in Three International Awards, Arlington Police Dep’t (Oct. 24, 2017), [] (describing a Walmart Restorative Justice Initiative with CEC resulting in “more than a 50 percent reduction in theft calls for service and overall arrests”); Natalie Crofts, 1st Time Shoplifters Can Keep Clean Record with Help of Utah Company, (Sept. 20, 2014), https://​‌index.php?nid=148&sid=31774717 [] (quoting supportive commentary from a Salt Lake City Police Department property crime detective). Journalists have found even larger effects. 148 See, e.g., Deborah Knapp, Walmart Cracks Down on Shoplifting by Taking Justice into Its Own Hands, KENS5 (San Antonio) (Nov. 12, 2016),‌
features/‌​walmart-cracks-down-on-shoplifting-by-taking-justice-into-its-own-hands/351317792 (on file with the Columbia Law Review) (reporting, based on a review of police records, drops of 33%, 49%, 56%, and 72% in shoplifting at four Texas Walmart stores).

CEC proudly advertises success in battling recidivism, claiming that it “reduces the likelihood that a shoplifter will come back to the store to steal again.” 149 Neyfakh, Make a Deal, supra note 125. “Less than 2% of shoplifters who complete the CEC educational program reoffend” at one of CEC’s retailers, “compared with estimates as high as 80% for those who do not participate in a restorative justice pro­gram.” 150 Telephone Interview with Brian Ashton, supra note 136; Statement of Corrective Education Company, CEC (June 8, 2016),‌home/​‌blog/‌statement-of-corrective-education-company (on file with the Columbia Law Review) [hereinafter CEC, Statement]. “CEC’s educational programs not only addresses [sic] behavioral issues, but provide life skills and motivation for reintegration,” the company’s website explains. 151 CEC, supra note 2. CEC is “continually reforming generations and changing lives, one day at a time,” it adds. 152 Id. Indeed, CEC even offers testimonials from “graduates” who claim the program helped them “create new values, attitudes, and goals” and “achieve self-responsibility and self-worth.” 153 Id. (testimonial of Eleanor, a CEC program participant).

Contracting with CEC commits retailers to sorting cases according to predetermined characteristics, without any on-the-scene discretion. CEC permits retailers to set eligibility criteria including criminal history, age, and item value, but excluding race, gender, language ability, or related character­istics. 154 Huntsman Declaration, supra note 128, at 2–3; Telephone Interview with Brian Ashton, supra note 136. Suspects who are too young or too old, 155 CEC uses a different “restorative justice” course for juveniles whom the retailer’s criteria do not exclude. See Corrective Education Company Announces Exclusive Partnership with D.A.R.E. at National Retail Federation Convention, CEC (June 29, 2015), https://​‌home/‌blog/‌corrective-education-company-announces-exclusive-partnership-with-dare-at-national-retail-federation-convention (on file with the Columbia Law Review). The juvenile contract is different as well. See Huntsman Declaration, supra note 128, exh. C, at 6. and those whose thefts are too small, may be released, for example, while suspects who steal big-ticket items may be referred to the police. Fewer than one in ten CEC stu­dents is a juvenile. 156 Email from Brian Ashton, CEO, CEC, to author (Jan. 15, 2018) (on file with the Columbia Law Review). Retailers set these criteria at the corporate level and embed them in a “black box”; the security guard simply enters basic infor­mation into a computer application and is instructed how to proceed. 157 Telephone Interview with Brian Ashton, supra note 136.

There is less public information about CEC’s competitors, like Turning Point Justice (TPJ), though enough to discern that the basic model seems similar. 158 See Retailers, Turning Point Justice, [] [hereinafter Turning Point Justice, Retailers] (last visited Aug. 10, 2018) (detailing TPJ’s business model, including the restitutionary payments that are collected by retailers). TPJ was founded in 2012 by a former district attorney from Salt Lake County, Utah, who had worked at CEC. 159 Palazzolo & Nassauer, supra note 111; Leadership Team, Turning Point Justice, [] (last visited Aug. 10, 2018). In an apparent effort to dis­tinguish itself from CEC, TPJ touts a “restorative justice” program developed by the National Association of Shoplifting Prevention and used by courts—as part of postarrest diversion programs—for over twenty years. 160 Turning Point Justice, Retailers, supra note 158. Suspects pay $400 to $425 to enroll with TPJ, roughly $50 to $75 of which is designated “restitution” and sent to the retailer. 161 Palazzolo & Nassauer, supra note 111. The precise amount of “restitution” is calculated by reference to the pertinent civil recovery statute. TPJ claims that retailers commonly donate this money to charity or to an “indigence fund” that covers program costs for suspected shoplifters who cannot afford the fees themselves. Telephone Interview with Paul Stewart, COO, Turning Point Justice (Feb. 9, 2018).

II. Evaluating Retail Justice

This Part pivots from description to evaluation. Sections II.A through II.C examine whether retail justice seems likely to harm suspects, victims, or the broader public, respectively. Throughout, retail justice is compared to the real criminal justice system, warts and all, though potentially significant jurisdictional variation is necessarily (and unfortunately) obscured. 162 Even if retail justice is normatively preferable to criminal justice in many settings, there may be jurisdictions in which criminal justice is so lenient as to undermine this conclusion, or so draconian that it cannot form a morally appropriate baseline for comparative analysis. Cf. Robert Nozick, Coercion, in Philosophy, Science, and Method 440, 450–51 (Sidney Morgenbesser et al. eds., 1969) (“[W]hen the normal and morally expected courses of events diverge, the one of these which is to be used in deciding whether a conditional announcement of an action constitutes a threat or an offer is the course of events that the recipient of the action prefers.”). But see Oren Bar-Gill & Omri Ben-Shahar, Credible Coercion, 83 Tex. L. Rev. 717, 729–30 (2005) [hereinafter Bar-Gill & Ben-Shahar, Coercion] (arguing that, even when the baseline itself is morally intolerable, precluding the choice of an alternative disserves the threatened party’s interests). Section II.D then explains why leading theories for prohibiting blackmail do not justify a ban on retail justice.

A note on structure: There is no intellectual consensus on why blackmail is illegal. Myriad competing theories emphasize disparate values and interests. Accordingly, rather than address the blackmail question at the outset, the normative issues are taken up in their most natural order. This discussion nearly resolves the blackmail problem, leaving only a few loose ends to tie up in section II.D.

A. Are Suspects Worse Off?

Retail justice companies, critics argue, prey on vulnerable consumers, wielding the threat of criminal prosecution to extract confessions and hefty enrollment fees. The profit motive, moreover, creates incentives for overzealous enforcement, the brunt of which disfavored groups or, worse yet, the innocent, will bear. 163 See, e.g., Neyfakh, Make a Deal, supra note 125 (reporting that all four public defenders interviewed were “pretty appalled” by CEC’s business model). Susannah Karlsson, a Brooklyn Defender Services attorney, said with regard to CEC that “[t]here’s no judicial oversight, there are no constitutional protections, there’s no due process.” Id. (internal quotation marks omitted). Steven Wasserman of the Legal Aid Society’s Criminal Practice Special Litigation unit remarked that “it sounded like CEC was ‘flirting with the crime of coercion in the second degree.’” Id.; see also Lorelei Laird, Retail Justice: Are Private Education Programs for Shoplifters a Second Chance—or Extortion?, A.B.A. J., June 2016, at 18, 18 (reporting that “some observers are skeptical,” worrying that “the programs . . . could ensnare innocent people without due process”). Indeed, the City of San Francisco asserted the interests of its residents when it sued CEC in 2015, seeking to halt its operation. 164 See S.F. Complaint, supra note 6, at 8–16. And the trial court recently accepted the city’s argument. 165 People ex rel. Herrera v. Corrective Educ. Co., No. CGC-15-549094, slip op. at 3 (Cal. Super. Ct. Aug. 14, 2017) (holding that the “undisputed facts . . . establish that CEC’s diversion program runs afoul of California’s extortion laws”). But are shoplifting suspects really better off without the retail justice alternative?

This section begins to work through this question in three stages: First, is retail justice a “bad deal” for suspects in an economic sense, such that we should reject the standard assumption that, because suspects choose it, it makes them better off? Second, even if retail justice makes suspects better off in general, does it disadvantage particular groups of suspects, such as the poor or people of color? Third, does the retail justice model encourage overenforcement of shoplifting laws, potentially even ensnaring suspects who are legally or normatively innocent?

1. Economic Efficiency: Is Retail Justice a “Bad Deal” for Suspects? —The “standard Pareto argument,” applied here, is that retail justice “improves the situation” of the suspects who choose it. 166 Bar-Gill & Ben-Shahar, Plea Bargain, supra note 30, at 738. The agreement suspects sign to enroll with CEC, for example, is formally an offer to contract, which suspects are (at least ostensibly) free to reject. Sure enough, in marketing their services, retail justice companies emphasize how they help offenders by sheltering them from the criminal justice system and extending to them the proverbial “second chance.” 167 See, e.g., Fight Shoplifting, Grow Restitution, Integrity and Trust, Turning Point Justice, [​D56V-Y5XP] (last visited Aug. 17, 2018) (asserting that retailers that use TPJ’s service “save their communities millions, and provide a second chance for petit theft offenders”); CEC, supra note 2 (listing “Second Chances: Good people can make bad decisions” as a justification for the program).

The standard Pareto argument fails, however, when the conditions for efficient contracting are lacking. 168 Cf. Robert E. Scott & William J. Stuntz, Plea Bargaining as Contract, 101 Yale L.J. 1909, 1940–49 (1992) (identifying circumstances in which “the assumption of efficient contracting” fails in the plea bargaining context). Four possibilities are considered here. The first three correspond loosely to contract law’s concepts of undue influence, misrepresentation, and mistake of fact. The last entertains the notion that, even if retail justice benefits suspects individually, it harms them as a class by exploiting a collective action problem among them. Perhaps surprisingly given the demographic data reported above, fewer than ten percent of the “students” at one major retail justice company are juveniles. 169 See Email from Brian Ashton to author, supra note 156. The following analysis therefore assumes an adult suspect, remaining agnostic on the potentially quite different juvenile case.

a. Undue Influence. —“Free consent is . . . a predicate condition of presuming mutually valuable exchange.” 170 Alan Devlin, Fundamental Principles of Law and Economics 186 (2015). To many observers, this is the principal problem with retail justice: Suspects pay the retail justice companies’ fees only under serious pressure from the threat of arrest and criminal prosecution. There is truth to this critique; shoplifting suspects face an unenviable dilemma. It does not follow, however, that they are worse off for being offered the choice or that retail justice companies should be prohibited from offering it to them.

“Even highly coercive threats are present in many types of legitimate economic bargaining.” 171 James Lindgren, Unraveling the Paradox of Blackmail, 84 Colum. L. Rev. 670, 701 & n.162 (1984) [hereinafter Lindgren, Unraveling the Paradox]; see also Jennifer Gerarda Brown, Blackmail as Private Justice, 141 U. Pa. L. Rev. 1935, 1950 n.32 (1993) [hereinafter Brown, Blackmail as Private Justice] (“That the blackmailee may be faced with a hard choice . . . does not necessarily make the blackmail any more coercive than the choice facing many parties to wholly legitimate economic transactions.”). Duress is unlawful, but true duress occurs only when the “offeror” “manufactures a false choice for the offeree”: “your money or your life.” 172 Devlin, supra note 170, at 187. Retailers, though, are perfectly free to call the police on suspected shoplifters. The choice is not “false” in the relevant sense. The ultimate source of pressure on a suspect is neither the retailer nor the retail justice company but rather the erratic and draconian criminal justice system the suspect is desperate to avoid. 173 The question of false choice reappears infra section II.A.1.d in the discussion of suspects’ collective action problem. Some additional objections are dealt with there.

Yet unlawful coercion is not limited to duress alone. When circum­stances suggest that the pressured party acted under the domination of another and that his assent “does not reflect his preference” or is “contrary to self-interest,” the law may find that “undue influence” taints the deal. 174 Devlin, supra note 170, at 186; see also Restatement (Second) of Contracts § 177 cmt. b (Am. Law Inst. 1981) (directing attention to the “unfairness of the resulting bargain” in assessing whether an agreement resulted from “unfair persuasion”); Bar-Gill & Ben-Shahar, Coercion, supra note 162, at 744–46 (discussing a “substantive justice” approach to coercion that “focuses on the substantive fairness of the interaction” and “views a threat as coercive if it results in a one-sided transaction”). This does not seem to describe the basic retail justice landscape, however.

Consider the choice from the perspective of a typical guilty suspect. The suspect must weigh the cost of enrolling with a retail justice company and completing the required course, on the one hand, against the expected con­sequences of contact with the criminal justice system (and civil recovery), on the other. Arrest is not guaranteed—one recent figure pegs the arrest rate for shoplifting nationwide at fifty percent 175 Shoplifting Statistics, supra note 23. —and charges may never be filed. But consequences for the unlucky fifty percent include a full search of the suspect’s person and belongings, 176 E.g., United States v. Robinson, 414 U.S. 218, 221–24 (1973). along with the physical danger 177 See, e.g., Andrea M. Burch, Bureau of Justice Statistics, U.S. Dep’t of Justice, Arrest-Related Deaths 2003–2009, at 1, 4 (2011),‌content/‌pub/‌pdf/ard0309st.pdf/ [] (estimating 700 arrest-related deaths annually); Cal. Dep’t of Justice, Death in Custody: Arrest-Related, Open Justice,‌death-in-custody/‌arrest-related [] (last visited Aug. 10, 2018) (reporting that twenty percent of all deaths in custody are arrest-related). Sandra Bland’s death is, for many, most salient. See Leon Neyfakh, Why Was Sandra Bland Still in Jail?, Slate (July 23, 2015),‌articles/news_and_politics/‌crime/‌2015/‌07/‌sandra_​bland_is_the_bail_system_that_kept_her_in_prison_unconstitutional.html []. Disease is also a concern. See, e.g., Joseph A. Bick, Infection Control in Jails and Prisons, 45 Clinical Infectious Diseases 1047, 1047 (2007) (describing the “increased risk” of disease acquisition in carceral institutions). and collateral consequences of arrest 178 See, e.g., Jeff Grogger, Arrests, Persistent Youth Joblessness, and Black/White Employment Differentials, 74 Rev. Econ. & Stat. 100, 105–06 (1992) (finding that arrests help explain persistent nonemployment and the black/white employment gap); Rachel A. Harmon, Why Arrest?, 115 Mich. L. Rev. 307, 313–15 (2016) (“Arrestees lose income during the arrest, and sometimes their jobs . . . . An arrest can affect child custody rights, . . . trigger deportation, and . . . get a suspect kicked out of public housing . . . . [I]ndividuals with arrest records may have worse employment and financial prospects . . . even if the arrestee is never convicted of a crime.” (footnotes omitted)); Eisha Jain, Arrests as Regulation, 67 Stan. L. Rev. 809, 810–12 (2015) (describing how “immigration enforcement officials, public housing authorities, public benefits administrators, employers, licensing authorities, social services providers, and education officials, among others” use arrest information adversely against arrestees). and preventive detention 179 See, e.g., Will Dobbie et al., The Effects of Pretrial Detention on Conviction, Future Crime, and Employment: Evidence from Randomly Assigned Judges, 108 Am. Econ. Rev. 201, 201 (2018) (finding that “pretrial detention significantly increases the probability of conviction . . . [and] decreases formal sector employment and the receipt of employment- and tax-related government benefits”); Nick Pinto, The Bail Trap, N.Y. Times Mag. (Aug. 13, 2015), (on file with the Columbia Law Review) (describing a defendant who “lost three weeks of income, was subjected to brutal physical violence and missed Thanksgiving dinner with his family” before his charges were dismissed for lack of evidence); Cal. Dep’t of Justice, Death in Custody: Booking & Pre-Trial, Open Justice,‌death-in-custody/‌pretrial [] (last visited Aug. 10, 2018) (reporting that fifteen percent of all deaths in custody occurred among detainees awaiting arraignment or trial).
Pretrial detention can be lengthy even on minor charges. See, e.g., Steven B. Bright & Sia M. Sanneh, Fifty Years of Defiance and Resistance After Gideon v. Wainwright, 122 Yale L.J. 2150, 2162 (2013) (“A woman in Mississippi charged with shoplifting spent eleven months in jail before a lawyer was appointed to her case, and three additional months before entering a guilty plea.”).
(even if she is never prosecuted 180 See James B. Jacobs, The Eternal Criminal Record 201 (2015) (“It is standard practice for an arrest to remain on the rap sheet even if the case is dismissed, adjourned in contemplation of dismissal, or resolved by a not guilty verdict.”). ); a dizzying array of costs and fees (even if she is indigent); 181 For comprehensive taxonomies, see Appleman, supra note 16, at 1485 (describing “[p]rivate probation, bail fees, translation fees, indigent representation fees, dismissal fees, high interest rates, jail and prison costs, court fines, and community service charges”); Logan & Wright, supra note 16, at 1185–96 (detailing various legal financial obligations that criminal suspects and defendants incur at each stage of the criminal justice process); see also Konig, supra note 20 (reporting court and legal fees for shoplifting cases in the hundreds or thousands of dollars). See generally Alexes Harris, A Pound of Flesh (2016) (arguing that the use of monetary sanctions as a criminal sentencing tool in the United States allows courts to control individuals until they fully pay their debts, which reinforces existing inequalities). and the possibility of conviction and punishment (ranging from days to years in prison and hefty fines), 182 See, e.g., Ryan P. Sullivan et al., Stolen Profits: Civil Shoplifting Demands and the Misuse of Neb. Rev. Stat. § 25-21,194, 95 Neb. L. Rev. 28, 32 (2016) (reporting that shoplifting a $10 item in Florida could result in sixty days in jail and $500 in fines and fees, while shoplifting a candy bar in Nebraska could result in a six-month jail sentence and $1,000 in fines). For thefts of costlier items, substantially larger penalties may obtain. See, e.g., Va. Code § 18.2-95 (2018) (authorizing grand larceny sentences of up to twenty years); id. § 18.2-103 (defining theft of $500 or more as grand larceny). with additional collateral consequences. 183 See, e.g., John D. King, Beyond “Life and Liberty”: The Evolving Right to Counsel, 48 Harv. C.R.-C.L. L. Rev. 1, 23–34 (2013) (cataloging consequences of misdemeanor convictions, including “loss of the right to possess a firearm, to serve in the military, . . . to receive . . . public benefits, to drive a car legally, and to adopt a child”); Jenny Roberts, Why Misdemeanors Matter: Defining Effective Advocacy in Lower Courts, 45 U.C. Davis L. Rev. 277, 298–99 (2011) (explaining that some misdemeanor convictions can trigger deportation, loss of student-loan assistance, or eviction from public housing); see also Argersinger v. Hamlin, 407 U.S. 25, 48 & n.11 (1972) (Powell, J., concurring in the result) (listing the possible results of misdemeanor convictions, including social stigma, forfeiture of public office, disqualification from a licensed profession, and loss of pension rights).

It would be perfectly rational for self-interested suspects to prefer the retail justice option. All the more so if they are risk averse and thus benefit from the certainty retail justice provides. In fact, it would be similarly rational, if tragic, for innocent suspects as well. While the probability of prosecution and conviction should be lower for an innocent suspect, they are not negligible, 184 On the risks of wrongful conviction for misdemeanor defendants, see Alexandra Natapoff, Misdemeanors, 85 S. Cal. L. Rev. 1313, 1316, 1328–50 (2012) (describing the criminal justice system’s “high tolerance for wrongful petty convictions”). and the consequences of arrest are just the same as for the guilty in most jurisdictions. Indeed, we have understood since Professor Malcolm Feeley’s famous tome that, for many accused misdemeanants, “the process is the punishment.” 185 Malcolm M. Feeley, The Process Is the Punishment: Handling Cases in a Lower Criminal Court (1979); see also Bowers, Punishing the Innocent, supra note 32, at 1132–39 (describing how an innocent defendant’s experience with the criminal justice system may be punishment itself).

Note that, despite the foregoing analysis, this conclusion does not nec­essarily depend on fine empirical judgments about the cost of enrollment or the likelihood of arrest or prosecution in the absence of retail justice. On one leading account of coercion, a proposal is coercive only if it threatens to make its recipients worse off than they would have been in the “morally expected course of events.” 186 Nozick, supra note 162, at 450; see also Bar-Gill & Ben-Shahar, Coercion, supra note 162, at 740–44 (discussing “rights-based theories of coercion,” which posit that “[i]f B has a right to be free from situation X, then h[er] agreement to do Y in order to be freed from the threat of having X inflicted on her must result from, or, it is the definition of, coercion”). Yet the maximal threat retailers (acting through retail justice companies) could make is to call the police on everyone who declined to enroll. This would entail no rights violation, however, and no threat to make suspects worse off than they are morally entitled to be. Shoplifting suspects have no moral entitlement to a particular probability of arrest or prosecution and no moral complaint if retailers were to pursue charges in every case. 187 Cf. Nozick, supra note 162, at 452–53 (assuming that the police are “morally expected” to prosecute individuals they believe they can prove committed crimes). This also means that not until the price of “tuition” exceeded the cost of actual criminal justice sanctions could we infer that retail justice “students”—most of whom we would expect to reject the offer at this price—were getting a raw deal.

Notice that the criminal justice system’s severity is the fertile soil that nourishes the retail justice alternative. Public choice theory predicts as much. As Professor Keith Hylton has explained in a related context, “[a] system of harsh punishments encourages rent-seeking—for example, bribe-taking—on the part of law enforcement officials.” 188 Keith N. Hylton, Whom Should We Punish, and How? Rational Incentives and Criminal Justice Reform, 59 Wm. & Mary L. Rev. 2513, 2526 (2018) [hereinafter Hylton, Whom Should We Punish]. “As the harshness of penalties increases,” Hylton continues, “law enforcement agents have greater leverage with which to seek bribes, which can be demanded of the guilty and the innocent alike.” 189 Id. at 2527. Hylton is writing about public law enforce­ment agents, but private agents—with the power to stave off the public ones—can extract these rents as well.

b. Misrepresentation. — Retail justice companies boast about the transformational power of their “restorative justice” programs—which, they say, drive down recidivism—even posting testimonials from “students” describing their reformations. 190 See, e.g., CEC, Statement, supra note 150. These claims register as naïve, if not disin­genuous. The pertinent “student” population, recall, has been cleansed of most repeat offenders. Multiple studies have found that shoplifters seldom reoffend after their first apprehension, 191 See, e.g., Cameron, supra note 39, at 151 (“Among pilferers who are apprehended and interrogated by the store police but set free without formal charge, there is very little or no recidivism.” (emphasis omitted)); Cohen & Stark, supra note 112, at 30 (“[V]irtually no one continues shoplifting after being apprehended once . . . .” (emphasis omitted)); Admit Your Guilt, Serve No Time, Progressive Grocer, Nov. 1991, at 8, 8 (finding that only five percent of individuals who were apprehended for shoplifting and paid civil damages were later reapprehended). Klemke argues that “there are serious under-reporting biases in these studies.” Klemke, Sociology of Shoplifting, supra note 37, at 125; see also Lloyd W. Klemke, Does Apprehension for Shoplifting Amplify or Terminate Shoplifting Behavior?, 12 Law & Soc’y Rev. 391, 396 & tbl.1 (1978) [hereinafter Klemke, Apprehension for Shoplifting] (finding, using a self-report methodology, that forty percent of youths shoplifted again after having been apprehended). Of course, the same underreporting biases likely infect the retail justice companies’ recidivism figures as well. suggesting the baseline rate of recidivism may be low. Of those who do steal again, most will escape detection. 192 See Shoplifting Statistics, supra note 23 (“Shoplifters . . . are caught an average of only once in every 48 times they steal.”). Estimates of recidivism based on subsequent apprehension will therefore understate the true rate, potentially severely. 193 A fortiori if recidivism figures are based on subsequent arrest by the public police, as at least one CEC study appears to have been. See Neyfakh, Make a Deal, supra note 125 (“When the company presented a law enforcement agency in Florida with a list of several hundred people in their jurisdiction who had completed the CEC course during the previous two years, they were told that less than 1 percent of the sample had since been arrested for any crime . . . .”).

To be sure, some CEC “students” are probably more serious crimi­nals who have managed to keep their records clean. But the evidence suggests that rehabilitating this population is an enormous challenge. 194 See, e.g., Bruce Western, The Rehabilitation Paradox, New Yorker (May 9, 2016),‌news/‌news-desk/‌the-rehabilitation-paradox [] (“Rehabilitative programs are often too little, too late; we need to intercede early.”). An eight-hour online course is no brace against the deep-seated personal and struc­tural forces that precipitate serious criminality. 195 See, e.g., Lance Lochner & Enrico Moretti, The Effect of Education on Crime: Evidence from Prison Inmates, Arrests, and Self-Reports, 94 Am. Econ. Rev. 155, 183 (2004); Steven Raphael & Rudolf Winter-Ebmer, Identifying the Effect of Unemployment on Crime, 44 J.L. & Econ. 259, 262 (2001); John Paul Wright et al., Association of Prenatal and Childhood Blood Lead Concentrations with Criminal Arrests in Early Adulthood, 5 PLoS Med. 732, 732 (2008). Nor, for what it’s worth, does CEC’s course appear to incorporate even the most basic elements of the “restorative justice” movement in whose flag it is wrapped. 196 One expert commented, for example, that retail justice “does not sound like a model for restorative justice.” Laird, supra note 163, at 19 (citing Professor Mary Louise Frampton); see also, e.g., John Braithwaite, Restorative Justice & Responsive Regulation 11 (2002) (defining restorative justice as “a process whereby all the parties [the victim, offender, and affected community] with a stake in a particular offence come together to resolve collectively how to deal with the aftermath of the offence and its implications for the future” (internal quotation marks omitted) (quoting Tony Marshall)); Stephen P. Garvey, Restorative Justice, Punishment, and Atonement, 2003 Utah L. Rev. 303, 303 (listing “victim-offender mediation, family group conference, and sentencing circles” as “the processes associated with restorative justice”). Although “restorative justice theory uniformly endorses[ ] restitution as the primary remedial response to criminal acts[,] [t]he specific amount and form of this restitution is usually agreed upon by both the victim and the offender through some form of mediation process.” Garvey, supra, at 307. All this raises the possibility that retail justice companies misrepresent the benefits they deliver in exchange for the fees they collect. If any such misrepresenta­tion is inducing suspects to enroll, the Pareto assumption (that enrollment benefits suspects) may be misplaced. 197 See Devlin, supra note 170, at 188 (discussing how information asymmetries caused by misrepresentation can lead to inefficient outcomes).

It seems doubtful, however, that this is actually the case. More likely, the retail justice companies’ claims about recidivism are immaterial to suspects’ decisions to enroll. Reliance is likely lacking. Indeed, in attacking CEC’s business model, San Francisco alleges that people pay CEC’s fees for no other reason than to avoid contact with the criminal justice system. 198 See S.F. Complaint, supra note 6, at 10. That hardly suggests suspects are being hoodwinked—quite the contrary. If suspects do not expect anything of value from the course—and would pay the fees even if no course were offered, for instance—then the retail justice companies defraud them of nothing if the course turns out to be worthless.

It is worth acknowledging that an emerging literature finds that training in “noncognitive” skills may generate large reductions in crime under certain conditions. 199 See, e.g., Christopher Blattman et al., Reducing Crime and Violence: Experimental Evidence from Cognitive Behavioral Therapy in Liberia, 107 Am. Econ. Rev. 1165, 1182–204 (2017); Sara B. Heller et al., Thinking, Fast and Slow? Some Field Experiments to Reduce Crime and Dropout in Chicago, 132 Q.J. Econ. 1, 32–38 (2016). It’s not clear from available evidence that CEC’s courses incorporate this know-how. But it does raise the possibility that some forms of short-term “education” could have crime-reducing effects—a public good with positive externalities. We would expect the market to undersupply a good like this, making the normative question whether the state ought to provide a subsidy.

c. Mistake. —A third potential “market failure” that could undermine the efficient-contracting assumption stems from asymmetric information. Maybe suspects choose retail justice, the argument goes, because they har­bor misconceptions—which the retail justice companies exploit, or even foster—about their expected sanctions in the criminal justice system, par­ticularly as first-time offenders. 200 A court may be reluctant to allocate the risk of mistake to suspects given the circumstances in which their agreement is sought, cf. Restatement (Second) of Contracts § 154 (Am. Law Inst. 1981) (explaining the circumstances under which a contracting party may bear the risk of mistake), creating the possibility that the resulting contract is voidable by the suspect, cf. id. § 153 (stating that, under certain circumstances, a contract may be voidable by a mistaken party who does not bear the risk of mistake). Worse yet, some critics argue, retail justice companies afford none of the procedural rights criminal defendants enjoy—in particular, no judicial oversight and no assistance from counsel who might help them assess their odds before paying to enroll. 201 See, e.g., Neyfakh, Make a Deal, supra note 125; cf. Brown, Blackmail as Private Justice, supra note 171, at 1972 (arguing that public enforcement of the law might be a necessary condition to balance the individual’s interest in procedural protection with the public’s interest in the enforcement of substantive norms). If suspects understood the prosecutor was unlikely to pursue charges, for instance, they would have little reason to pay the fee. Indeed, San Francisco’s legal theory hinges partly on the purported leniency of its criminal justice system, which, it says, belies the threatening messages CEC delivers to shoplifting suspects. 202 See S.F. Complaint, supra note 6, at 1. Note that risk-averse suspects may prefer retail justice even when its expected sanctions are slightly higher, if contact with the criminal justice system brings more uncertainty. (CEC appears to have tempered its messaging since the lawsuit was filed, now saying less about what suspects can expect if they decline to enroll. 203 See Appellant’s Opening Brief, supra note 34, at 16–17. )

This argument suggests an important limit on the circumstances in which retail justice can be assumed beneficial based on suspect choice. The more retail justice companies exploit the misapprehensions of their “students” about the criminal justice system, the better the case for regulat­ing their activity. The risk may be highest where local justice is most lenient, suggesting, among other things, that retail justice companies might wish to tailor their messaging to local legal context.

The point should not be overstated, however. Even factoring in the possibility of diversion, nonprosecution, and other channels of mercy, crim­inal justice system contact in most jurisdictions is something to be feared and avoided at virtually any cost. An arrest alone can be devastating, even in San Francisco. 204 On the consequences of arrest generally, see supra notes 178, 180, 185. Regarding San Francisco specifically, a pending class action lawsuit alleges that the city’s implementation of its bail schedule effects an unconstitutional wealth-based detention scheme. One of the lead plaintiffs, Riana Buffin, was allegedly arrested for theft from a department store and, unable to pay her $30,000 bail, held for forty-six hours before the prosecutor decided not to file charges, causing her to lose her job at the Oakland airport. See Third Amended Class Action Complaint at 6–7, Buffin v. City of San Francisco, No. 15-CV-4959 (N.D. Cal. filed May 27, 2016), 2016 WL 3587128.

As for procedural protections, we ought not to lionize this aspect of the criminal process either. Most misdemeanor prosecutions, argues Professor Alexandra Natapoff in her searing exposé, “baldly contradict the standard due process model of criminal adjudication,” lacking “the evidentiary and procedural protections that are supposed to ensure the guilt of the accused.” 205 Natapoff, supra note 184, at 1315–16. Perhaps most shockingly, there is “compelling evidence that . . . petty offenders in particular[ ] often do not get counsel even when they are legally entitled to it.” 206 Id. at 1341 (emphasis added); see also id. at 1340–42. The constitutional rule, evidently honored in the breach, mandates counsel whenever incarceration is imposed. See Scott v. Illinois, 440 U.S. 367, 373–74 (1979). Those who do get an attorney may receive only a few minutes of consultation before entering a plea. 207 See, e.g., Natapoff, supra note 184, at 1342–43 (describing misdemeanor represen­tation as often “a formality,” “better described as facilitating the guilty plea rather than checking the merits of the case”); Lisa C. Wood et al., Meet-and-Plead: The Inevitable Consequence of Crushing Defender Workloads, 42 Litig. 20, 23 (2016) (asserting that “attorneys engage in meet-and-plead dispositions in courtrooms across the country”). Judicial oversight is scarce. 208 See, e.g., Shima Baradaran Baughman, Subconstitutional Checks, 92 Notre Dame L. Rev. 1071, 1138 (2017) (lamenting “wrongful and pressured convictions by plea agreements without any judicial oversight”); Kate Levine, How We Prosecute the Police, 104 Geo. L.J. 745, 747–48 (2016) (“Reams of scholarship look at the lack of judicial oversight at every stage of the process, from plea bargains to sentencing decisions, and waivers that make pleas virtually unreviewable by appellate courts.”); see also Baldwin v. New York, 399 U.S. 66, 69 (1970) (reserving the jury trial right to crimes punishable by more than six months’ imprisonment). Yet the collateral consequences of conviction attach just the same. 209 Cf. Paul T. Crane, Charging on the Margin, 57 Wm. & Mary L. Rev. 775, 834–37 (2016) (arguing that courts should consider collateral consequences in determining which procedural protections to afford a defendant). It may well be that more people underestimate than overestimate the criminal justice system’s horrors. 210 Cf., e.g., Kirk R. Williams et al., Public Knowledge of Statutory Penalties: The Extent and Basis of Accurate Perception, 23 Pac. Soc. Rev. 105, 117 (1980) (concluding that the general public underestimates the severity of sanctions). Perceptions may vary by social class, depending upon the extent of peer contact with the system. Id. at 110.

d. Collective Action / Externalities. —Finally, it may be that, while any par­ticular suspect, viewed in isolation, benefits from an expanded choice set, the class of suspects as a whole is actually harmed. That is, retail jus­tice companies may appear to help suspected shoplifters only by exploiting a collective action problem that prevents them from banding together in resistance. Professors Oren Bar-Gill and Omri Ben-Shahar have modeled the point in the plea bargaining context. 211 See Bar-Gill & Ben-Shahar, Plea Bargain, supra note 30, at 740 (discussing “defendants’ collective action problem” in plea bargaining). The basic intui­tion is that, if all defendants could agree to insist on trial, they would overwhelm the criminal justice system and prosecutors would be forced to forgo prosecution in many cases. 212 See id. at 739–40. As Bar-Gill and Ben-Shahar note, “Plea bargains are contracts with externalities: each defendant who ac­cepts a plea frees prosecutorial resources to pursue other defendants.” 213 Id. at 743.

In fact, Bar-Gill and Ben-Shahar’s argument may in some ways be stronger in this context than in its original setting. Bar-Gill and Ben-Shahar rightly acknowledge, in the plea bargaining context, that prosecu­tors’ bud­gets are endogenous and may increase if plea bargaining were abolished, allowing prosecutors to pursue more cases. 214 See id. at 769. In the context of retail theft, we can actually observe the counterfactual—the world without retail justice. We know that, in the absence of retail justice, retail­ers and prosecutors in fact have declined to prosecute many suspects. 215 See supra notes 97–109.

At the same time, the arrest rate in this counterfactual world is still fairly high—fifty percent 216 Shoplifting Statistics, supra note 23. —and retailers can seek civil recovery as well. And the threat of even more arrests and civil recovery—if suspects were to boycott retail justice—is more credible than the threat of more prosecutions in the plea bargaining case, as arrests and civil recovery are cheaper than prosecutions. 217 See supra notes 102–103 and accompanying text. This matters because the expected costs of an arrest alone—with its concomitant physical risk and collateral consequences 218 See supra notes 177–183. —are likely higher for most suspects than the costs of paying a retail justice company and completing its course. All the more so once we factor in the chance of civil recovery and of prosecution conditional on arrest, and if most suspects presented with the dilemma, all of whom pass a criminal-history check, are risk averse.

2. Equality: Are Certain Groups of Suspects Worse Off? — Even if retail justice is a good deal for most suspects, concerns may persist if its effects are discriminatory on the basis of race, class, or some other morally irrelevant characteristic. Notice that one’s conclusion on the former issue frames the latter: If retail justice is a “good,” we should ensure that it is not being reserved to the privileged classes. If retail justice is a “bad,” however, our concern is that it is being forced upon disfavored groups. Based on the foregoing analysis, this section makes the former assump­tion: that retail justice is generally beneficial because it prevents harmful contact with the criminal justice system.

The appropriate baseline for an analysis of distributive effects is the manner in which the criminal justice system distributes criminality within the large pool of individuals who cause others harm. Sociologist John Hagan distinguishes between “suite” criminals, whose harmful and immoral acts are frequently treated as noncriminal matters, and “street” criminals, whose similarly motivated conduct is branded as deviant. 219 See generally John Hagan, Who Are the Criminals? (2012). A similar dynamic marks the early history of retail theft, where “kleptomania” or private payments shielded well-to-do ladies from the criminal justice system, while poorer offenders went to jail. 220 See supra notes 41–52 and accompanying text. Some con­temporary shoplifting research suggests that lower-class offenders continue to be treated more harshly than wealthier ones, and there may be age and race effects as well. 221 See supra notes 116–121 and accompanying text.

Against this baseline, retail justice companies appear, at first blush, to promote equality by “leveling up,” extending to low-status individuals the lenient treatment previously beyond their grasp. 222 See generally James Q. Whitman, Harsh Justice: Criminal Punishment and the Widening Divide Between America and Europe 10 (2003) (describing the gradual trend in continental Europe by which the forms of imprisonment previously extended only to aristocrats are now generalized and extended to the entire prison population). Retail justice, that is, shelters not “suite” criminals but “street” criminals, and not just wealthy shoplifters but poor ones as well. And on its face, at least, it applies even-handedly to suspects of every race and gender. 223 See Landes & Posner, supra note 10, at 41 (“[S]elective or discriminatory enforcement . . . would be eliminated under a regime of private enforcement. The law would be enforced against everyone who violated it and enforcement would not place a particular . . . individual at an unfair disadvantage.”).

There are, however, reasons to be cautious in this assessment. This section quickly sketches out five. Three of these can be safely put to bed, in the weak sense that they do not reveal retail justice to be worse than the system that operates in its absence. The same may be true of the remaining two, but the case is muddier. Accordingly, Part III recom­mends that lawmakers require retail justice companies to collect data that will facilitate close monitoring of these distributive effects.

The first concern stems from the fact that shoplifting is a “middle-class crime.” One might worry that retail justice coddles the middle class while neglecting the truly poor—who may commit other forms of theft—replicating the regressive class dynamic Hagan describes, just lower down the economic ladder. Yet, if the research is to be believed, retailers already favor the middle class over the poor. 224 See supra notes 118–120 and accompanying text. Perhaps more important, shoplifting is a “middle-class crime” only in the sense that the middle class is overrepresented in the population of offenders. 225 See supra note 71 and accompanying text. In absolute number, more shoplifters hail from the lowest income bracket than any other. 226 See Blanco et al., supra note 22, at 906 tbl.1 (reporting that 43.95% of shoplifters have personal incomes under $20,000 and 22.37% have family incomes under $20,000).

Second, shifting from public to “offender-funded” private justice trans­fers the costs of shoplifting, in a rough sense, from taxpayers to suspects. This cost structure may disproportionately burden the poorest suspects. More than that, retail justice may simply be inaccessible to the very poor, who cannot afford the enrollment fees. Retail justice companies purport to make their programs accessible to all through a combination of payment plans, discounts, and “tuition scholarships.” CEC says that ninety percent of suspects presented with the option choose to enroll. 227 See Neyfakh, Make a Deal, supra note 125. But we do not know why the other ten percent do not or, relevant here, how many decline due to financial constraints.

Yet the premise of this second point is false: Public justice, too, makes its “users” foot the bill. “As criminal justice costs have skyrocketed,” Professor Laura Appleman observes, criminal justice institutions have begun to impose “fees and fines at every turn,” and thus “the burden to fund the system has fallen largely on the system’s users, primarily the poor or indigent.” 228 Appleman, supra note 16, at 1485. Just as the fees paid to retail justice companies disproportion­ately harm the poorest suspects, so, too, do the criminal justice system’s fees. 229 See id. (“[F]unding of the criminal justice system has disproportionately fallen on those least able to pay.”). Even publicly funded counsel is no longer free in most jurisdic­tions. 230 See, e.g., Helen A. Anderson, Penalizing Poverty: Making Criminal Defendants Pay for Their Court-Appointed Counsel Through Recoupment and Contribution, 42 U. Mich. J.L. Reform 323, 324 (2009); Beth A. Colgan, Paying for Gideon, 99 Iowa L. Rev. 1929, 1929 (2014). For a state-by-state list of so-called recoupment laws, see Harris, supra note 181, at 28–41 & tbl.2.4. Those who truly cannot pay spend years fighting their debts or, worse yet, are jailed for nonpayment, leading observers to lament the return of debtors’ prisons. 231 See, e.g., Harris, supra note 181, at 50–51 (describing the social consequences of imposing monetary sanctions and incarcerating debtors for nonpayment); Appleman, supra note 16, at 1489–92 (describing the “long and unsavory history” of “incarcerating the impoverished” in the United States).

Third, it is not known where, geographically, retail justice companies operate. We do know that they have serviced retailers frequented by both the wealthy (Bloomingdale’s, for example) and less well off (such as Walmart). But we do not know which Walmart stores, for example, used their services. If the stores are situated primarily in (relatively) wealthier regions or areas with racial demographics that skew white, then retail justice may be sheltering those populations disproportionately, with economically or racially regressive effects. And because private industry is choosing whom to protect, there is no obvious mechanism through which an angry public can hold the responsible parties to account.

We are not so demanding of criminal justice institutions in this respect, however. While selective enforcement and prosecution are prohibited, courts have made discrimination virtually impossible for defendants to prove. 232 See United States v. Armstrong, 517 U.S. 456, 463 (1996) (requiring litigants who allege discriminatory prosecution to meet an intentionally demanding standard); Bonita R. Gardner, Separate and Unequal: Federal Tough-on-Guns Program Targets Minority Communities for Selective Enforcement, 12 Mich. J. Race & L. 305, 318–19 (2007) (analogizing the standard of review for selective prosecution claims to the standard for equal protection claims, which requires a showing that any discriminatory impact have a discriminatory purpose). And we certainly do not demand equality across jurisdictions as opposed to within them. If the criminal justice authorities in heavily black City A decide to enforce shoplifting laws to the hilt, for instance, while mostly white City B’s authorities are far more lenient, a black shoplifting defendant in City A cannot point to City B in support of a selective-prosecution claim. (That is not to say the law has this right, however. Part III advocates data collection that would permit review of precisely this type of disparity within retail justice.)

Now for the two more nagging concerns: One is that the retail justice apparatus may enable or encourage store security to be more aggressive in ways that bear disproportionately on disfavored groups. This is not inevita­ble—one can imagine a world in which store security do not change how they patrol when retailers shift from public to private justice. If retail justice companies pay retailers or security firms for each enrollee, however, these payments may, as San Francisco alleges, “encourage[ ] security companies . . . to target not just individuals who may have shoplifted, but those who are most likely to fear getting turned over to the police” (who, presumably, are most likely to enroll). 233 S.F. Complaint, supra note 6, at 2; see also id. at 5 (“At particular risk of exploitation are those individuals who a security guard perceives would be most likely to buckle under CEC’s high-pressure tactics, perhaps out of fear of being turned over to law enforcement.”).

It is not entirely clear what San Francisco has in mind as the target pop­ulation here. The dynamics are complex. Different social groups may dread police contact for different reasons—the fear of physical mistreatment (per­haps highest among young black males), for example, versus the fear of deep social and professional embarrassment (perhaps highest among wealthy, middle-aged, white professionals), versus the fear of deportation (for undocumented immigrants). One might just as well assume security guards would target individuals they think are most likely to be able to pay retail justice enrollment costs—those who appear to have money to spare. 234 Cf. Elliott Ash et al., Local Public Finance and Discriminatory Policing: Evidence from Traffic Stops in Missouri 3–4 (Columbia Pub. Law Research Paper No. 14-591, 2018), (on file with the Columbia Law Review) (finding that, when local governments are under budget stress, traffic citation rates increase, with increases concentrated on white drivers and largest where white-to-black income ratio is highest). Data collection is necessary to resolve this concern with any confidence.

Finally, there is the fact that, after querying both internal and public records, retail justice companies refer repeat offenders to the police. This is a sensible approach to deterrence: Prior sanctions failed, suggest­ing the need for harsher medicine. Yet the practice bakes in whatever biases infected earlier interactions with enforcement authorities. If we believe the police (public or private) disproportionately target black men, for exam­ple 235 For evidence suggesting that rates of shoplifting are similar across racial groups, see supra notes 67–68 and accompanying text. For evidence that members of racial minority groups are nonetheless apprehended for shoplifting at disproportionate rates, see, e.g., Cohen & Stark, supra note 112, at 33 fig.1 (finding that black and Mexican American shoppers were apprehended out of proportion with their presence in the shopping public); see also George Won & George Yamamoto, Social Structure and Deviant Behavior: A Study in Shoplifting, 53 Soc. & Soc. Res. 44, 52 (1968) (same for native Hawaiians); Nat’l Coal. to Prevent Shoplifting, Program Guide, supra note 77, at 4 (discussing a survey of 3,550 retailers in which forty-six percent opined that racial minorities were more prone to shoplift than whites). —or that customers are more likely to report apparent thefts by blacks than whites 236 Cf. Max C. Dertke et al., Observer’s Reporting of Shoplifting as a Function of Thief’s Race and Sex, 94 J. Soc. Psychol. 213, 217–18 (1974) (finding no statistically significant difference in white customers reporting thefts by black versus white thieves, but that, when prompted to confirm having witnessed a theft, thefts by blacks were confirmed more often). —then a suspect’s prior record depends in part on his race. By relying on this measure, the retail justice companies discriminate as well.

How this compares to the alternative, however, is an empirical question. Relative to a retailer employing a zero-tolerance policy, which calls the police on every suspected shoplifter, retail justice companies may fare poorly. But zero-tolerance policies appear to be rare. More often, retailers make discretionary decisions, or apply some simple rules of thumb, to determine when to make the call. 237 See supra section I.B. Given the discretion involved—and that one rule of thumb in fact has been to call the police on repeat offenders—it would not be surprising if bias played a more, not less, important role when retail justice is absent. Retail justice companies, in other words, introduce (or perhaps exacerbate) one potential bias (from prior police interactions) but eliminate another (from the discretionary decision of when to call the police). The latter bias may well outweigh the former, but it is hard be sure. Again, more data are necessary to answer this question.

3. Overenforcement: The Costs of Casting a Wider Net. —The profit motive, critics contend, creates incentives for overzealous enforcement of the law, possibly sweeping in innocent as well as guilty defendants. San Francisco’s lawsuit alleges, in this vein, that CEC’s “payment structure creates a powerful incentive to pressure people to enroll in CEC, regardless of the evidence, if any, of their guilt.” 238 S.F. Complaint, supra note 6, at 2.

Academics have debated the basic point for decades. In the 1970s, economists Gary Becker and George Stigler sketched out a system in which private citizens investigate crimes, apprehend and try suspected offenders, and retain the proceeds, such as fines convicted defendants pay. 239 Gary S. Becker & George J. Stigler, Law Enforcement, Malfeasance, and Compensation of Enforcers, 3 J. Legal Stud. 1, 13–17 (1974). Professors William Landes and Richard Posner countered that a public monopoly on criminal law enforcement may be preferable because it ena­bles “discretionary nonenforcement” of the law by prosecu­tors. 240 Landes & Posner, supra note 10, at 38–41. Discretionary nonenforcement is an efficient way to tem­per the (inevitable) overinclusivity of criminal statutes without creating loopholes for defendants. 241 See id.

Decades later, the Landes–Posner notion of “discretionary nonenforce­ment” became the linchpin of Professor Ric Simmons’s argument to prohibit (most) private criminal settlements, like the ones retail justice com­panies facilitate. Simmons argues that private criminal settlements should be prohibited “because they remove the prosecutor from the settlement process.” 242 Simmons, Private Plea Bargains, supra note 12, at 1131. “[T]he prosecutor,” Simmons reasons, “plays a critical role in selecting which cases should be prosecuted, how they should be charged, and what sentence is appropriate.” 243 Id. “[T]he parties who negotiate a private criminal settlement,” in contrast, “do not practice discretionary nonenforcement.” 244 Id. at 1187.

As a descriptive matter, it is misleading to say that retail justice compa­nies usurp the prosecutorial role or edge prosecutors out of the picture. Prosecutors simply exercise their discretion at the wholesale level. 245 Not always—apparently, in some jurisdictions, prosecutors preapprove each individual suspect for participation in a retail justice program. Telephone Interview with Paul Stewart, supra note 161. They are not unaware of what retail justice companies are doing; 246 See Huntsman Declaration, supra note 128, at 12 (explaining that CEC’s policy is to “reach[ ] out to local criminal justice system stakeholders about the program” before conducting business in a particular jurisdiction); Telephone Interview with Brian Ashton, supra note 136 (explaining that CEC “always” informs police and prosecutors when it enters a new jurisdiction). indeed, at least some prosecutors actively encourage it. 247 See Huntsman Declaration, supra note 128, exh. D (letter of support for a trial period from the San Diego County District Attorney); CEC Inform, CEC,​‌home/‌cec-inform (on file with the Columbia Law Review) (last visited Sept. 15, 2017) (“The States Attorneys [sic] office is supportive of any Retailer’s decisions to implement a Restorative Justice Program.” (internal quotation marks omitted) (quoting Eighth Judicial Circuit of Florida)); Jessica Pishko, ‘Restorative Justice’ for Shoplifting? A Court Calls It Extortion, The Marshall Project (Oct. 30, 2017),‌2017/‌​10/‌30/‌restorative-justice-for-shoplifting-a-court-calls-it-extortion [] (“Law enforcement . . . is supportive in most states.”); cf. Arlington Police, supra note 147 (“Walmart and CEC have been incredible partners in looking at innovative ways to reduce criminalization and look for alternative methods in dealing with the ongoing issues related to misdemeanor thefts.” (internal quotation marks omitted) (quoting Police Chief Will Johnson)); Knapp, supra note 148 (“Our department and departments across the country are in favor of programs that are like this.” (internal quotation marks omitted) (quoting a police official)). Prosecutors bent instead on preserving their monopoly could simply subpoena retail justice companies’ business records and prosecute some of their “students,” which would quickly put an end to the business endeavor. 248 Cf. Simmons, Private Criminal Justice, supra note 10, at 960 (“[A] private criminal justice system will only exist . . . at the suffrage of the criminal justice system. If the public authorities . . . believe that a private criminal justice system is inappropriate, all they need to do is . . . bring formal criminal charges against all the defendants regardless of the outcome . . . .”). In a sense, San Francisco’s lawsuit, which seeks to rout retail justice companies from the jurisdiction, is the exception that proves the rule. 249 Even in San Francisco, it seems, CEC operated with the knowledge of the District Attorney, as opposed to the City Attorney who filed suit. Telephone Interview with Brian Ashton, supra note 136.

This descriptive point, however, even if correct, does not resolve any normative concerns about overenforcement—most significantly, whether retail justice encourages the enforcement of shoplifting laws against inno­cent suspects, as San Francisco alleges. It helps here to distinguish between “actually innocent” defendants, who have not in fact violated the law, and “normatively innocent” defendants, who “did it . . . [but] did not thereby offend the public’s moral code.” 250 Bowers, Legal Guilt, supra note 33, at 1658 (alterations in original) (internal quotation marks omitted) (quoting Akhil Reed Amar, The Constitution and Criminal Procedure: First Principles 90 (1997)). Normative innocence is a state of “relative blameworthiness,” resulting from a normative judgment of whether the defendant “ought to be charged.” 251 Id.

There are powerful disincentives for store security to target actually innocent individuals. The prospect of tort liability for false imprisonment and related harms undoubtedly looms large, especially when, as is common, surveillance video has captured the pertinent events. Retailers also wish to avoid the negative publicity wrongful accusations bring about. Indeed, this pair of concerns has long been thought to motivate retailers’ relatively lax approach toward shoplifting detection. There is no persuasive evidence that the incentives retail justice creates, at least to date, are strong enough to turn the tide.

The harder question concerns the effects of retail justice on normatively innocent suspects, for whom the prospect of increased enforcement seems more plausible. The evidence suggests, however, that such suspects are rare. In their discussion of overenforcement, Landes and Posner were concerned with prohibited conduct “that the legislature . . . did not in fact want to forbid.” 252 Landes & Posner, supra note 10, at 38. Their examples involve “minor infractions of the traffic code” and “violations of building-code provisions that, if enforced, would prevent the construction of new buildings in urban areas.” 253 Id. Likewise, in his work on normative innocence, Professor Josh Bowers focuses on “petty crimes that typically lack concrete victims,” many of which are “mala prohibita offenses.” 254 Bowers, Legal Guilt, supra note 33, at 1659, 1666. Bowers gives examples like “an indigent man . . . arrested for hopping a turnstile to get to his first day of work” and “an elderly man . . . arrested for selling ice pops without a license on a hot summer day.” 255 Id. at 1658.

As sympathetic as some shoplifters surely are, few, if any, of them are normatively innocent under these frameworks. There can be little doubt that the legislature did intend to criminalize even small-ticket retail thefts by individuals in great need. Nor is shoplifting merely a victimless, regula­tory offense. That so many shoplifters historically have escaped prosecution likely reflects judgments about prosecutorial resource constraints and priorities rather than judgments about normative innocence the retail justice companies now upset. 256 Indeed, Bowers argues that prosecutors typically ignore issues of normative innocence in petty crime cases “because of keen institutional pressures to charge reflexively.” Id. at 1661.

A separate overenforcement concern involves concededly guilty sus­pects. If the profit motive of retail justice motivates more aggressive enforcement of shoplifting laws—or greater efficiency enables it—then retail justice may feed the beast of overcriminalization. For example, a retailer that, before contracting with a retail justice company, only rarely called the police, may now sanction a much larger percentage of suspected shoplifters. This is troubling for those who believe that society’s principal criminal justice problem is not underenforcement but its opposite.

There are several reasons, however, that the overcriminalization argu­ment is more complicated, and probably weaker, than it first ap­pears. As an initial matter, the fact that society has criminalized shoplift­ing might be thought to signal that, setting aside enforcement costs, the efficient level of shoplifting is zero. 257 See Robert Cooter, Prices and Sanctions, 84 Colum. L. Rev. 1523, 1548–50 (1984); see also Devlin, supra note 170, at 106–08 (discussing “coercive transfers that take place in low transaction cost environments” as a type of crime that, “in the absence of enforcement costs, society ought to eliminate on the basis of efficiency”). That more guilty individu­als are sanctioned, the argument goes, moves us closer to that ideal and cannot count as a demerit for retail justice. 258 See Michael Gorr, Liberalism and the Paradox of Blackmail, 21 Phil. & Pub. Aff. 43, 62 (1992) (“That the ‘victim’ [of incriminatory blackmail] might be made worse off by the prohibition of blackmail is irrelevant for the obvious reason that crimes do not ordinarily deserve to be concealed.”); cf. id. (“[D]isclosure is in fact the morally preferred state of affairs.”).

Skeptics should consider two additional points. First, as mentioned earlier, retailers that work with retail justice companies—at least with CEC—specify eligibility criteria including age and item value. 259 See notes 151–154 and accompanying text. Reportedly, many retailers do not turn suspects over to CEC whom, based on these criteria, they would not have referred to the police. 260 Telephone Interview with Brian Ashton, supra note 136. Retail justice, in other words, does not necessarily lead retailers to cast a wider net. Second, even if some retailers do now cast a wider net, the consequences of being ensnared, which involve no contact with the criminal justice system, are less severe. 261 Conditional on being apprehended, suspects’ expected sanctions must be less than in the criminal justice system or else they will reject any private settlement offer. Private sanctions, in theory, should reflect the probability that the prosecutor would decline to prosecute. Cf. Brown, Blackmail as Private Justice, supra note 171, at 1956–57 (demonstrating that “allowing blackmail will not necessarily overdeter when prosecutors exercise discretion to decline prosecution of the blackmailee’s crime,” because “the probability of prosecution can affect the blackmailer’s demand”). Retail justice thus distributes punishment more equitably across the population of offenders. To put the point differently, we might just as well say that the net has narrowed in the sense that fewer, rather than more, shoplifting suspects will enter the criminal justice system.

B. Are Victims Worse Off?

It might seem odd that retailers, which long have lamented the scourge of shoplifting, would abandon the criminal justice system in favor of a more lenient, private alternative. Why would retailers not want the strongest deterrent sanctions available? 262 The assumption here is that criminal punishment generally tends to deter shoplifting. The possibility that it does not is addressed in section II.C.2.a. See also, e.g., Klemke, Apprehension for Shoplifting, supra note 191, at 401 (finding serious reasons to doubt that criminal punishment deters shoplifting). Is it possible that, in opting for retail justice, retailers actually act against self-interest?

It is possible, but unlikely. These retailers are sophisticated entities that contract for retail justice under calm conditions with ample information. The interesting question is not whether retail justice makes retailers better off but why it seems to do so. Some possibilities have been mentioned above, but a deeper exploration is now in order.

As an initial matter, it is not actually clear that the premise of the ques­tion posed here—that is, that retailers are embracing a more lenient approach—is true. The reasons recall the preceding discussion of overen­forcement. It does seem fair to assume that, in an individual case, criminal sanctions inflict more disutility on the suspect than the retail justice compa­nies’ fees and coursework. 263 Although the retail justice companies would likely resist the characterization of their fees as “sanctions,” from the offender’s perspective, certainly they are that. Cf. Klemke, Sociology of Shoplifting, supra note 37, at 128 (“These more positive approaches may still be viewed as punishment by involuntary subjects . . . .”); Levmore & Fagan, supra note 18, at 312–13 (“[T]he higher price a party pays for secrecy might deter misbehavior as successfully as any legal remedy, and the latter normally comes at a greater social cost.”). Otherwise, we would expect there to be a market, outside the setting discussed here, for the “restorative justice” courses the retail justice companies administer. From the prospective offender’s viewpoint, however, the relevant question concerns not the actual but rather the expected sanctions in each system. Expected sanc­tions, of course, fold in not only the anticipated magnitude of sanctions but also the likelihood of apprehension and the likelihood of sanctions conditional on apprehension.

Because the mechanism for apprehension—the retailer’s private police—is held constant across the two systems, the likelihood of apprehen­sion should not differ drastically. But it should be slightly higher in the retail justice setting if the private police use time they save processing offenders to apprehend additional suspects, 264 See Turning Point Justice, Retailers, supra note 158 (claiming this benefit). or if they are given financial incentives to procure retail justice “students.” 265 See, e.g., S.F. Complaint, supra note 6, at 2 (alleging that CEC paid private police firms ten dollars per enrollee). The likelihood of sanctions conditional on apprehension may also rise with the shift to retail justice, but heterogeneously across retailers, depending upon the criteria used to establish eligibility for enrollment.

All things considered, then, retail justice likely substitutes weaker but more certain sanctions for stronger, less certain ones. Retail justice compa­nies also dispense sanctions more swiftly. 266 See Turning Point Justice, supra note 28 (emphasizing the program’s “immediate consequences that are proportionate with the offense”). Empirical research con­sistently attributes deterrence more to the certainty and celerity of sanctions than to their severity. 267 See, e.g., Michael Tonry, An Honest Politician’s Guide to Deterrence: Certainty, Severity, Celerity, and Parsimony, in 23 Deterrence, Choice, and Crime 365, 374 (Daniel S. Nagin et al. eds., 2018) (summarizing research). This is not inconsistent with the supposition made elsewhere that first-time shoplifting suspects may tend to be risk averse. See Murat C. Mungan & Jonathan Klick, Identifying Criminals’ Risk Preferences, 91 Ind. L.J. 791, 793–94 (2016). Accordingly, the retail justice compa­nies’ claims to effective deterrence may be more plausible than they initially appear. 268 Cf. Levmore & Fagan, supra note 18, at 328 (“There will, therefore, be cases where private agreements provide more deterrence than the criminal law and it is for this reason that the law ought to tolerate some experimentation with private agreements in lieu of criminal charges.”). (The comparative claims may be misleading nonetheless, given the low baseline rates of recidivism for first-time shoplifters who are apprehended. 269 See supra note 191 and accompanying text (surveying sources highlighting the low recidivism rates after first apprehension). )

In any event, regardless which system better deters potential shoplift­ers, the choice to opt for retail justice does make one thing clear: Retailers believe their return on investment in deterrence is higher in the private than the public system. But this, too, might seem odd, for economists have long regarded crime deterrence as a classic public good. 270 See, e.g., Alon Harel, Efficiency and Fairness in Criminal Law: The Case for a Criminal Law Principle of Comparative Fault, 82 Calif. L. Rev. 1181, 1201 (1994) [hereinafter Harel, Efficiency and Fairness] (describing the “widely held belief that protection [from crime] is a pure public good”). Government must provide a criminal justice system, the argu­ment goes, precisely because private parties will view the returns as insufficient to motivate adequate investment in deterring crime. This is because private actors cannot capture all the benefits of their expendi­tures on deterrence—if I hire a security guard to patrol in front of my house at night, my next-door neighbors benefit as well, with no obligation to contribute. 271 See id. at 1202 n.55.

In fact, retailers have already paid taxes to finance the police, prosecutors, courts, and prisons. Why bypass these public institutions? Retailers that opt out cannot, of course, demand a tax refund. The obvious answer might be the kickback from retail justice companies, when available, but this is ultimately a sideshow; many retailers continue to work with CEC even absent this minor remuneration. The better explanation is that, notwithstanding that retailers have paid taxes, prosecution of shoplifters in the criminal justice system remains costly and inefficient from the retailers’ perspective. The taxes paid are sunk costs and what drives retailers is the price of going forward.

Assisting in the public prosecution of shoplifting suspects is costly to retailers in several ways. 272 See, e.g., Joh, Conceptualizing, supra note 14, at 590 (“Assisting in a formal police investigation [costs] . . . time and effort . . . , as well as the potential for negative publicity . . . which may be more costly than the initial disruption.”). Employees may need to testify or sit for interviews with police detectives during business hours. The merchandise in question may languish in evidence lockers, unavailable for sale. Frequent visits from the police, to take suspects into custody, could drive away other customers who see the police as a threat or as a signal that criminal activity is afoot. Retailers may even be reluctant to lose the business of the suspected shoplifter himself, who may double as a paying customer or whose family and friends may be regulars. 273 My colleague and Dean Tom Miles suggested that, in this way, we might view retail justice as a kind of customer loyalty program.

What the retailers receive from the criminal justice system in return, moreover, often proves too paltry to justify the costs. In many jurisdic­tions, shoplifting is a low-priority crime. Police response to calls can be slow—requiring the retailer to maintain prolonged custody over the suspect—and prosecution infrequent. 274 See Laird, supra note 163, at 18–19. The criminal justice system simply does not provide a service valuable enough—or deterrence strong enough—to justify the costs of participating in it. Of course, the quantity of criminal justice resources available to fight shoplifting is not fixed. In theory, the state could respond to these constraints by sending help. Indeed, some police departments have hired additional officers specifically to respond to calls from Walmart. 275 See Barbaro, supra note 111 (explaining the burden that Walmart’s zero-tolerance policy placed on local police departments); Pettypiece & Voreacos, supra note 101 (describing an officer known to his colleagues as “Officer Walmart”). Yet this seems not to be the norm.

In addition, just as in the initial public-good analysis, the retailer cannot capture all the benefits of its expenditures on deterrence here, inside the criminal justice system, either. A retailer’s participation benefits other outlets the suspect may have targeted. In other words, because victims are required to expend resources (beyond background taxation) to obtain deterrence through the criminal justice system, and because those expenditures produce positive externalities, the criminal justice system has not solved the public-good problem after all. 276 See Philip J. Cook & John MacDonald, The Role of Private Action in Controlling Crime, in Controlling Crime: Strategies and Tradeoffs 331, 331, 346 (Philip J. Cook et al. eds., 2011).

*   *   *

To nail down the point and to tee up the next section’s analysis, consider who bears the costs of shoplifting under three different systems of law enforcement. 277 I set aside the costs of the private police, which are constant. In the first, retailers do nothing other than expel red-handed shoplifters with a warning. They then attempt to shift the costs of their inventory “shrinkage”—which are relatively high, given their lackluster efforts at deterrence—onto their customers in the form of increased prices. To the extent that market competition prevents them from transferring all the expenses, this do-nothing approach effectively allocates the costs of shoplifting to the retailers and their customers together.

The second approach—criminal prosecution of some suspects—spreads the costs more widely. Here, the taxpayers bear the expense of the justice system’s deterrence-producing institutions, partially offset by “user fees” collected from suspects and defendants. Retailers assume the costs of assisting the prosecution as well as the residual costs of shrinkage, some of which they pass through to their customers.

Finally, consider the retail justice model. Retail justice, recall, is entirely “offender funded.” Taxpayers bear no enforcement costs because the criminal justice system’s institutions are not involved in any way. Retailers bear no enforcement costs either, as they simply hand off suspects to the retail justice companies. In fact, the retailers may make money in the form of kickbacks. The net effect is to shift costs from retailers, customers, and taxpayers onto the suspects themselves. It is not hard to understand why retailers would prefer this option. And it looks appealing from society’s perspective as well, if it produces deterrence more efficiently. 278 See Levmore & Fagan, supra note 18, at 313 (arguing that private settlement can deter wrongdoers at a lower cost than the legal system would).

C. Is Society Worse Off?

This last point turns out to be more complex than it first appears. This section contemplates the effects of retail justice on social welfare beyond the interests of suspects and victims. It begins with a discussion of the social costs of crime, focusing on prevention and enforcement costs, before pivoting to consider the more diffuse effects of decreased transparency.

1. The Social Costs of Crime. —Society’s goal, within a utilitarian framework, is to minimize the total social costs of crime, which include the direct costs incurred by victims, the costs of prevention and enforce­ment, and any unjustified costs imposed on defendants and their families. 279 See, e.g., Gary S. Becker, Crime and Punishment: An Economic Approach, in Essays in the Economics of Crime and Punishment 1, 40 (Gary S. Becker & William M. Landes eds., 1974). In some settings, private settlement of criminal disputes would raise concerns about insufficient general deterrence, if victims settle for too little when prosecution was otherwise likely. 280 Cf. Landes & Posner, supra note 10, at 42 (arguing that legalizing private forms of enforcement, like blackmail, might lead “the blackmailer [to] sell his incriminating information to the offender for a price lower than the statutory cost of punishment to the criminal, which would reduce the effective cost of punishment . . . below the level set by the legislature”). Underdeterrence seems improbable here, where the “fine” is a multiple of the average shoplifting take, 281 See Clint Rainey & Allegra Hobbs, Been Caught Stealing, N.Y. Mag. (Dec. 8, 2013), [] (reporting an average retail theft by customers of $129). and where retailers have continued contact with the offenders. Especially outside the biggest cities, individuals who shoplift from major retailers likely continue, out of practical necessity, to patronize those same stores in the future. This means retailers at least partly internalize the expected costs of monitoring and recidivism, motivating them to seek socially efficient levels of deterrence. 282 See Levmore & Fagan, supra note 18, at 323–24.

The more interesting issue concerns prevention and enforcement costs. Retail justice companies may minimize these costs to retailers, but they do so partly by externalizing them—that is, by shifting them to suspects. This cost externalization may, in some cases, create incentives for retailers that are perverse from society’s perspective. The crucial point is that the availability of a costless (or even profitable) mechanism for adjudicating and sanctioning shoplifting encourages retailers to favor enforcement when prevention (that is, victim precautions) might deter crime more efficiently.

The choice between prevention and enforcement is ubiquitous in society, yet the pertinent legal literature is surprisingly thin. 283 See, e.g., Ben-Shahar & Harel, Economics, supra note 31, at 299–301 (exploring how sanctions imposed on pre-crime activities alter victims’ incentives to take preventive measures); Harel, Efficiency and Fairness, supra note 270, at 1196 (suggesting that criminal law should embrace the principle of comparative fault to encourage victims to take precautions against crime); Robert A. Mikos, “Eggshell” Victims, Private Precautions, and the Societal Benefits of Shifting Crime, 105 Mich. L. Rev. 307, 308–09 (2006) (suggesting that, while crime prevention measures might displace crime rather than reduce the amount of total crime, such displacement may be beneficial). There is also an economics literature on point. See, e.g., Ann P. Bartel, An Analysis of Firm Demand for Protection Against Crime, 4 J. Legal. Stud. 443, 443–44 (1975) (exploring what factors impact a firm’s decision to make private expenditures to prevent crime); Omri Ben-Shahar & Alon Harel, Blaming the Victim: Optimal Incentives for Private Precautions Against Crime, 11 J.L. Econ. & Org. 434, 434–35 (1995) (arguing that “victims of crime may also be in a position to take enforcement measures that may deter crime and substitute or complement the government’s effort” and discussing how to make victims’ incentives more efficient); Charles T. Clotfelter, Private Security and the Public Safety, 5 J. Urb. Econ. 388, 390–91 (1978) (examining the demand for private protection and the impact of such protection on crime rates); Steven Shavell, Individual Precautions to Prevent Theft: Private Versus Socially Optimal Behavior, 11 Int’l Rev. L. & Econ. 123, 123 (1991) (showing why crime victims may not take the socially optimal level of precaution). Potential victims often employ the two strategies simultaneously. I rely on the criminal justice system to deter home invasions by catching and punishing burglars, for example, but I also lock my door. The hard question concerns the socially optimal level of precautions—should I not only lock my door but also build a moat? Victims, it turns out, commonly take too few or too many precautions, depending on the circumstances.

Victims overinvest in precautions that generate negative externali­ties, such as diverting crime toward other victims. 284 See Ben-Shahar & Harel, Economics, supra note 31, at 309–11 (“One factor that may lead to excessive investment in precautions, and which has gained much attention in the literature, is the diversion or displacement of crime.”). But see Mikos, supra note 283, at 310 (arguing that crime diversion or displacement is socially beneficial because it shifts crime away from “eggshell victims,” who suffer relatively high levels of harm and thus invest more in precautions). Victims underinvest, in contrast, in precautions that generate positive externalities, such as deterring crime against other victims. 285 See Ben-Shahar & Harel, Economics, supra note 31, at 311–12. The classic example involves competing precautions against auto theft: The Club, a pole-like device that locks the steering wheel, and Lojack, a radio transmitter that allows police to locate a stolen car. The Club merely displaces crime to other owners. Not so for Lojack—in fact, because potential thieves cannot tell which cars have Lojack, Lojack reduces theft for every car that might have Lojack. The predictable result? “People buy too many Clubs and not enough Lojacks.” 286 Barry Nalebuff & Ian Ayres, Why Not?: How to Use Everyday Ingenuity to Solve Problems Big and Small 24 (2006); see also Ian Ayres & Steven D. Levitt, Measuring Positive Externalities from Unobservable Victim Precaution: An Empirical Analysis of Lojack, 113 Q.J. Econ. 43, 44–45 (1998) (“An individual car owner’s decision to install a Lojack only trivially affects the likelihood of his or her own vehicle being stolen since thieves base their theft decisions on mean Lojack installation rates.”).

Familiar precautions for retailers include strategic lighting and store layout, security cameras, greeters, well-spaced personnel, item placement (with small, valuable items out of customer reach), and security tags. 287 See, e.g., Klemke, Sociology of Shoplifting, supra note 37, at 130–33; Shteir, supra note 20, at 171–95. Retailers employ these precautions to varying degrees, and they seem to make a difference. Indeed, data from observational studies of custom­ers show great variance among stores in rates of shoplifting. One inter­pretation of these data is that “certain characteristics of stores should . . . be consid­ered as an important variable influencing the amount of shoplifting that occurs.” 288 Klemke, Sociology of Shoplifting, supra note 37, at 10–11 (discussing Abigail Buckle & David P. Farrington, An Observational Study of Shoplifting, 24 Brit. J. Criminology 63 (1984)). “Certain stores,” that is, “may be viewed as prime targets for shoplifting because of the nature or quality of merchandise, or because they are seen as having poor security.” 289 Id. For direct evidence that precautions reduce shoplifting, see David P. Farrington et al., An Experiment on the Prevention of Shoplifting, 1 Crime Prevention Stud. 93, 107–10 (1993) (finding that electronic tagging caused a lasting decrease in shoplifting, store redesign caused an immediate decrease that wore off, and uniformed guards had no effect); M. Patrick McNees et al., Shoplifting Prevention: Providing Information Through Signs, 9 J. Applied Behav. Analysis 399, 402–03 (1976) (finding that both general and item-specific signage reduced theft).

While it is hard to be sure, there is reason to think at least some retailers employ these measures too sparingly. In an echo of Victorian era rhetoric about the devilish temptations of department stores, that is certainly the sense one gets when reading media coverage of Walmart’s enormous demands on the police. Walmart, some experts say, “lays out its stores in a way that invites trouble and often doesn’t have enough uniformed employees to make sure everything runs smoothly.” 290 Sampson et al., supra note 100; see also Pettypiece & Voreacos, supra note 101 (quoting police criticism of Walmart’s theft-prevention efforts); Scott E. Wolfe & David C. Pyrooz, Rolling Back Prices and Raising Crime Rates? The Walmart Effect on Crime in the United States, 54 Brit. J. Criminology 199, 210–14 (2014) (finding that crime fell more slowly in the 1990s in counties where Walmart built stores). More generally, see Ian J. Abramson, Shoplifting: Fastest-Growing, Hardest-to-Control Crime, Volume Retail Merchandising, Feb. 1983, at 2, 2 (blaming retailers for “creating an atmosphere for impulse buyers without adequate security” and thereby “encouraging possible customers to indulge in ‘impulse shoplifting’”). Its stores “can feel messy and disheveled,” allowing “troublemakers to rationalize that the company doesn’t care.” 291 Sampson et al., supra note 100; see also Charles Fishman, The Wal-Mart Effect 202–04 (2006) (describing Walmart’s shopping environment). Journalists found that Walmart stores call the police far more often than Target stores in the same jurisdiction, even when controlling for store size and hours of operation. 292 See Sampson et al., supra note 100; see also Pettypiece & Voreacos, supra note 101. This discrepancy suggests that there is ample room within a successful business model for greater investment in precautions. And given that most shoplifting is situational and impulsive, rather than premeditated, 293 See Blanco et al., supra note 22, at 911–12. such investment ought to reduce aggregate theft rather than merely displace it. 294 See generally Rob T. Guerette & Kate J. Bowers, Assessing the Extent of Crime Displacement and Diffusion of Benefits: A Review of Situational Crime Prevention Evaluations, 47 Criminology 1331, 1356–57 (2009) (reviewing 102 studies and concluding that “crime displacement seems to be the exception rather than the rule, and it is sometimes more likely that diffusion of crime-control benefit will occur”).

Most often, as in the home-invasion example, the prevention-enforcement question concerns the allocation of crime-deterrence respon­sibility between private victims and public authorities. 295 See generally Keith N. Hylton, Optimal Law Enforcement and Victim Precaution, 27 RAND J. Econ. 197, 197–98 (1996) [hereinafter Hylton, Optimal Law Enforcement] (arguing that victims have insufficient incentives to take precautions against crime because they externalize the cost of public law enforcement). In the retail justice context, however, the analysis differs. As noted, retail justice companies permit retailers to shift deterrence-related costs from themselves (and their customers) to suspects. Whether this makes society better off depends on whether retailers or suspects are the “least cost avoiders.” It is tempting to think the answer is obvious—it must be suspects, who can avoid the costs simply by refraining from offending (or arousing suspicion). But this misapprehends the real issue.

As Professor Alon Harel argues, “The identification of the criminal as the ‘cheaper cost avoider’ does indeed mean that it is socially desirable that criminals avoid carrying out crimes. But given the persistence of criminal activity, the salient question is who should bear the costs of preventing such activity.” 296 Harel, Efficiency and Fairness, supra note 270, at 1198–99. For a different way of framing the issue, see Hylton, Optimal Law Enforcement, supra note 295, at 198–99 & n.9 (defining “offender precaution” to include “the profits forgone by an offender who chooses not to commit a crime,” and arguing that “[i]f victim precaution is cheaper than offender precaution (or forbearance), then an optimal punishment policy might require more precautionary effort from victims and less from offenders”). In the traditional case, again, we compare the cost of victim pre­cautions with the cost of state enforcement. 297 See Richard A. Posner, Economic Analysis of Law 239 (7th ed. 2007) (“[T]hrowing responsibility onto the victim might minimize aggregate social costs. It costs something, though very little, to lock one’s car—less than it would cost the criminal justice system to bring the thief to justice.”). Here, instead, we com­pare the cost of victim precautions with the cost of retail justice enforcement, which, of course, is ultimately financed by suspects’ fees. Costs in the retail justice model include the expense of developing and administering the “restorative justice” course, the creation and maintenance of supportive technology, 298 See, e.g., CEC, supra note 2 (describing program technologies available on multiple computer platforms). and the time spent by suspects taking the course. It seems plausible that these costs exceed the expense to retailers of some of the common precautions mentioned above. In sum, while retail justice may reduce enforcement costs relative to the criminal justice system, victim precautions might—in some situations, even if not in many—reduce them even further.

2. Transparency. —By operating wholly outside official institutions, critics argue, retail justice undermines “the community’s collective interest in the administration of justice as a public event that binds and defines us.” 299 Brown, Blackmail as Private Justice, supra note 171, at 1967; see also S.F. Complaint, supra note 6, at 15 (“The fact that CEC operates entirely outside the criminal justice system strips district attorneys’ offices of any and all oversight functions . . . .”); Neyfakh, Make a Deal, supra note 125 (“There’s no judicial oversight, there are no constitutional protections, there’s no due process.” (internal quotation marks omitted) (quoting Susannah Karlsson)). More concretely, retail justice frustrates popular oversight of the criminal process, deprives citizens of valuable information about offenders in their midst, and silences public condemnation which, on some theories, differentiates the criminal process from all others. These three critiques are considered in turn, followed by a discussion of the effects of retail justice on official crime data.

Initially, notice that, without secrecy, the retailer and suspect never reach a bargain. Secrecy is what the suspect pays for. For these transparency considerations to win out, then, the benefits of publicity must outweigh any welfare gains already discussed. 300 See Levmore & Fagan, supra note 18, at 313 (“Information is valuable, to be sure, but the higher price a party pays for secrecy might deter misbehavior as successfully as any legal remedy, and the latter normally comes at greater social cost.”).

a. The Publicity Norm in Criminal Cases. —The public jury trial remains the gold standard for American criminal justice. In reality, however, the trial’s rarity, not its quality, makes it precious. Only a frac­tion of shoplifting charges lead to public trials; prosecutors drop some cases, divert others, and plead out most of the rest. 301 Jocelyn Simonson, The Criminal Court Audience in a Post-Trial World, 127 Harv. L. Rev. 2173, 2174–75 (2014). Plea bargaining nominally takes place under the auspices of a public system, but the deals themselves are struck behind closed doors. 302 Id. at 2175. Increasingly, the public is excluded even from the parts of the plea process that are supposed to be transparent. 303 See id. (“[T]he public . . . receives little information about the behind-the-scenes decisions and negotiations that lead to these plea bargains.”).

It turns out, then, that the public’s ability to audit the criminal process by observing its institutions at work is illusory, at least in the mine-run misdemeanor case. 304 Indeed, I have criticized plea bargaining for precisely this reason. See John Rappaport, Unbundling Criminal Trial Rights, 82 U. Chi. L. Rev. 181, 198 (2015) (noting that trials, unlike plea bargaining, “shine a light on investigatory behavior and the exercise of governmental power more generally”). It is far from clear that the shift from plea bargaining to retail justice meaningfully exacerbates the problem. At most, the difference is one of degree.

On the second point—that retail justice deprives the public of useful information about offenders, preventing, for example, the identification of potentially troublesome recidivists 305 This is a point Professors Levmore and Fagan stress. See Levmore & Fagan, supra note 18, at 312–13 (noting that settlements and private arbitration proceedings “deprive the world of information that judicial decisions might convey . . . [and] third parties may not even know of hazards or facts that are known to the settling parties but costly for others to rediscover”); see also Murat C. Mungan, The Scope of Criminal Law, in Research Handbook on the Economics of Criminal Law 51, 57 (Alon Harel & Keith N. Hylton eds., 2012) (“Criminal law . . . has the function of producing information concerning a convict’s attitude towards the rest of society and his preferences. . . . Hence, criminal law may allow other members of society to alter their behavior towards the ex-convict and take low-cost targeted precautions against him.”). —the realities of criminal justice again blunt the critique. Recall, first, that, in a world without retail justice, many retailers call the police infrequently, often banishing suspected offenders instead. 306 See supra note 97 and accompanying text. Of the suspects who are arrested, many will escape formal charges or obtain diversion, leaving behind records of questionable utility for tracking repeat offenders. 307 See, e.g., Klemke, Sociology of Shoplifting, supra note 37, at 128; Konig, supra note 20.

Moreover, retail justice companies keep records for precisely this purpose—to identify and screen out repeat offenders. To be sure, unless retailers pool information, this technique will catch only shoplifters who reoffend at the same retailer or another retailer that contracts with the same retail justice company, 308 See Grover C. Trask, Best Best & Krieger, LLP, California’s Business Community Restorative Justice Precomplaint Education Program for Petty Theft Offenders 3 (July 31, 2013),‌home/‌uploads/‌BBK-White-Paper-MJM-Edits-13.07.31.pdf (on file with the Columbia Law Review) (discussing CEC’s data-gathering systems). so coverage is admittedly imperfect. But the system likely captures a great number of cases, and the ones that slip through the cracks probably present little serious danger. 309 See Blanco et al., supra note 22, at 907 tbl.2 (showing that individuals who self-report shoplifting report low rates of violent behavior); Gold & Lang, supra note 102, at 2 (“[M]ost non-professional shoplifters do not commit other types of crimes.”). Confidentiality of offense records, however, does have another implication that is not so easily minimized. Confidentiality may inadvertently increase statistical discrimination against black males seeking employment. See Amanda Agan & Sonja Starr, Ban the Box, Criminal Records, and Racial Discrimination: A Field Experiment, 133 Q.J. Econ. 191, 195 (2018) (“[W]hen employers lack individualized information, they tend to generalize that black applicants, but not white applicants, are likely to have records.”); Harry J. Holzer et al., Perceived Criminality, Criminal Background Checks, and the Racial Hiring Practices of Employers, 49 J.L. & Econ. 451, 452 (2006) (“If accessibility to criminal history information is limited . . . employers may infer the likelihood of past criminal activity from such traits as gender, race, or age [which] disproportionately impact[s] African Americans.”). If prospective employers—particularly retailers, one might think—cannot be confident whether an applicant has shoplifted, they may rely on less accurate and more odious proxies for criminality, such as race. Landlords, colleges, and other institutions that screen for criminal records may do the same. In short, even if confidentiality helps black males who have shoplifted once, the evidence suggests it harms black males whose records are clean. Because confidentiality “helps black men with records while hurting black men without records, the net effect on black male employment would depend on the real-world sizes of these groups.” Agan & Starr, supra, at 229.

The third transparency concern: Professor Henry Hart famously theorized that conduct is “criminal” precisely when (and because) it incurs the “moral condemnation of the community.” 310 Henry M. Hart, Jr., The Aims of the Criminal Law, 23 Law & Contemp. Probs. 401, 405 (1958). Retail justice precludes the collective act of public disdain. 311 See Brown, Blackmail as Private Justice, supra note 171, at 1970–71 (arguing that private settlement of criminal matters is inappropriate because it removes the official sanction and condemnation of the state); Simmons, Private Plea Bargains, supra note 12, at 1165–66 (same). Is this a serious demerit?

We might ask first whether, in the most practical sense, retail justice dampens the message of condemnation the legislature sought to convey by criminalizing shoplifting. This is an empirical question and, for reasons related to the discussion of deterrence above, the answer may be no. If, absent retail justice, the criminal prohibition against shoplifting is grossly underenforced, then retail justice, by potentially reaching more offenders, may actually amplify rather than muffle the criminal law’s message—at least to the principal audience whose behavior it is designed to control. One thing retail justice companies do, after all, is underscore the social harms that shoplifting inflicts. 312 See, e.g., C.A. Partnership Program, Turning Point Justice, [] (last visited Aug. 11, 2018) (“With many retailers only making a few cents on the dollar in profits, stealing even a chocolate bar can quickly reduce store profits that lead to store closings and job losses in local communities.”). And this seems, significantly, to satisfy retailers’ retributive thirst.

Still, the administration of retail justice involves no public disapproval of the offender. One might assume this is a problem. For one thing, holding all else equal, it weakens deterrence. 313 See Hylton, Whom Should We Punish, supra note 188, at 2533 (“[E]xposure of a crime is a separate punishment by itself.”). It also interferes with “one of the state’s most important tasks in articulating and enforcing the criminal law: declaring societal norms in public and labeling as ‘criminal’ the behavior that runs afoul of them.” 314 Brown, Blackmail as Private Justice, supra note 171, at 1970–71. Yet maybe that is the wrong inference to draw. Perhaps the stronger inference is that the public authorities, which (with some exceptions) seem at least complicit in the operation of retail justice, do not believe that most first-time shoplifters deserve public condemnation.

As Part I showed, the historical record reflects longstanding ambiva­lence toward the offense. Legislatures could, after all, enact mandatory reporting statutes to ensure the private police send every suspected shoplifter to the public system. But they do not. Nor has any state legislature made a move to ban retail justice from the marketplace. 315 The Minnesota legislature, however, is considering the question. See H.R. 1520, 2017 Leg., 90th Sess. (Minn. 2017). Perhaps, then, it is best to conceptualize what is happening here as a novel species of decriminalization. There are entirely sensible reasons one might favor such a policy, including the belief that it will reduce, rather than elevate, the rate of subsequent criminal offending. 316 See, e.g., Cameron, supra note 39, at 165 (arguing that there may be “a strong argument in favor of keeping pilferers out of jail lest they receive there the kinds of knowledge and emotional support they need to become ‘successful’ commercial thieves”); Klemke, Apprehension for Shoplifting, supra note 191, at 401 (finding that apprehension and police contact each increased subsequent shoplifting among juveniles); see also Samuel Walker, Reform the Law: Decriminalization, in Deviant Behavior 678, 679 (Edward J. Clarke ed., 7th ed. 2008) (describing the theory that “any contact with the system . . . imposes a ‘criminal label’ on the individual,” who “internalizes the label and proceeds to act out the role, committing additional and more serious crimes”).

b. Crime Data. —A somewhat different concern is that retail justice dis­torts official crime data. Every suspect retail justice companies poach is a statistic that will not show up in public data recording offenses known to the police. Jurisdictions in which retail justice companies thrive will there­fore publish crime rates that are artificially depressed. Shoplifting is reported in official FBI crime data as “larceny-theft,” which, in turn, consti­tutes the largest component of the umbrella category “property crime.” 317 See FBI, U.S. Dep’t of Justice, Uniform Crime Report: Crime in the United States, 2010, Larceny-theft 1 (Sept. 2011), []; FBI, U.S. Dep’t of Justice, Uniform Crime Report: Crime in the United States, 2010, Property Crime 2 (Sept. 2011),​propertycrimemain.pdf [] (“Larceny-theft accounted for 68.1 percent of all property crimes in 2010.”).

To be sure, even absent the effects of retail justice, selective reporting and prosecution plague official shoplifting data. 318 This is a point Mary Owen Cameron rightly stressed in her influential 1964 work. See Cameron, supra note 39, at 23–24. Various shoplifting “epidemics,” for instance, may not have been epidemics at all but rather the manifestation of changes in retailer behavior, such as increases in the rate of apprehension and police referral. 319 See, e.g., Klemke, Sociology of Shoplifting, supra note 37, at 8 (“[T]he increase shown in FBI data is more likely to be a product of changes in apprehension and reporting practices than a real increase in shoplifting behavior.”); Hindelang, supra note 91, at 584 (finding that, due to changes in the rate at which retailers called the police, the police would have observed a 32% increase in shoplifting from 1963 to 1968 even if the number of offenders had not changed). To some extent, though, this dynamic exists for all but the most serious crimes, which are reported and prosecuted more consistently. 320 See Jayne E. Robinson & Michael R. Rand, U.S. Dep’t of Justice, Criminal Victimization in the United States, 2008 Statistical Tables 98 tbl.91 (2011),​pub/pdf/cvus08.pdf [] (estimating, for example, that victims reported 47.1% of crimes of violence but only 33.6% of thefts). There is no reason to think these background measurement errors are distributed unevenly across jurisdictions in any significant way. Retail justice, in contrast, concentrates and magnifies the effects. If a small city’s Walmart switches from an aggressive police referral policy to CEC, for example, the drop in official property crime statistics could be instantaneous and significant without any change in the underlying crime rate.

Crime data are central to a variety of personal and policy decisions in contemporary society. 321 See generally Modernizing Crime Statistics 85–102 (Janet L. Lauritsen & Daniel L. Cork eds., 2016) (providing an overview of the users and uses of crime statistics). Families may consult crime rates when choosing where to settle. 322 Online tools provide detailed information about crime rates for those seeking to move. See, e.g., Crime Statistics, City-Data,‌crime [] (last visited Aug. 11, 2018) (providing a map of crime statistics viewable based on city and state). Academics use them to research the determi­nants of crime and potential solutions. 323 See Cameron, supra note 39, at 61, 174 (asserting that “the operation of private police represents a challenging problem in the field of criminology” because of its effect on crime statistics, which are the “basic data on which theories of crime causation are built”). Governments allocate funds with crime rates in mind. 324 See Modernizing Crime Statistics, supra note 321, at 91–94. Police officials are evaluated partly on their ability to drive crime down. 325 See generally John A. Eterno & Eli B. Silverman, The Crime Numbers Game: Management by Manipulation (2012) (describing incentives for police to manipulate crime statistics). When retail justice distorts crime data, it warps these personal and policy decisions too. That is not to say the effect cuts in any clear direction—it depends on local needs and incentives. Law enforcement, for example, may support retail justice when it perceives the benefits from depressing crime rates to outweigh the cost of resources forgone—or vice versa.

D. Is This Blackmail?

According to the California court ruling on San Francisco’s lawsuit, CEC’s business model is “textbook extortion.” 326 People ex rel. Herrera v. Corrective Educ. Co., No. CGC-15-549094, slip op. at 3 (Cal. Super. Ct. Aug. 14, 2017) (granting summary judgment in part). Whether that decision is correct as a matter of California law is a question better left to other fora. Answering it, in any event, would not resolve the legality of retail justice under other states’ blackmail (or extortion) statutes. 327 Cf. Simmons, Private Plea Bargains, supra note 12, at 1191 (“[M]ost states prohibit incriminating blackmail altogether, but a few allow the practice in limited circumstances.”). The more fruitful inquiry is whether the justifications for prohibiting blackmail support banning retail justice. Even if retail justice is blackmail, in other words, should it be unlawful? There is substantial dissensus on whether and why blackmail should be illegal in the first place. 328 See Mitchell N. Berman, Blackmail, in The Oxford Handbook of Philosophy of Criminal Law 37, 37 (John Deigh & David Dolinko eds., 2011) [hereinafter Berman, Blackmail] (“[E]xplaining why blackmail is properly criminalized remains ‘one of the most elusive intellectual puzzles in all of law.’” (quoting James Lindgren, Blackmail: An Afterword, 141 U. Pa. L. Rev. 1975, 1975 (1993))). This section explains why several leading blackmail theories fail to justify a prohibition on retail justice. Professor James Lindgren’s influential theory comes closest but ultimately it, too, falls short.

Assuming for present purposes that the agreements between CEC and shoplifting suspects are indeed blackmail, they are what commenta­tors call “opportunistic” and “incriminating” (or “crime-exposure”) blackmail. They are “opportunistic” because the putative blackmailer, CEC, does not expend resources to acquire the incriminating infor­mation; the retailers gather it in the course of ongoing security efforts. 329 See Mike Hepworth, Blackmail: Publicity and Secrecy in Everyday Life 74–77 (1975) (describing four types of blackmail: entrepreneurial, opportunistic, commercial research, and participant). They are “incriminating” (or “crime-exposing”) because the information suspects wish to suppress relates to criminal activity, as opposed to, say, noncriminal but embarrassing personal facts. 330 See Berman, Blackmail, supra note 328, at 74 (discussing “crime-exposure blackmail”); Brown, Blackmail as Private Justice, supra note 171, at 1935–36 (defining and discussing “incriminating blackmail”). Many leading theories of blackmail simply do not apply to opportunistic and incriminating blackmail. 331 One prominent theory centers on the resources wasted acquiring, and then suppressing, information about the blackmail victim. See, e.g., Ronald H. Coase, The 1987 McCorkle Lecture: Blackmail, 74 Va. L. Rev. 655, 674 (1988); Douglas H. Ginsburg & Paul Shechtman, Blackmail: An Economic Analysis of the Law, 141 U. Pa. L. Rev. 1849, 1865 (1993); Steven Shavell, An Economic Analysis of Threats and Their Illegality: Blackmail, Extortion, and Robbery, 141 U. Pa. L. Rev. 1877, 1879–80 (1993) [hereinafter Shavall, Economic Analysis of Threats]. But opportunistic blackmail involves no such purported waste. See James Lindgren, Blackmail: On Waste, Morals, and Ronald Coase, 36 UCLA L. Rev. 597, 601 (1989). The same response dispatches the notion that permitting blackmail encourages invasions of privacy in search of blackmail material. See Jeffrie G. Murphy, Blackmail: A Preliminary Inquiry, 63 Monist 156, 160 (1980). Another theory worries that “to avoid being blackmailed by [opportunistic blackmailers] . . . potential victims will . . . reduce their level of innocent, yet embarrassing, activities.” Shavell, Economic Analysis of Threats, supra, at 1903 (emphasis omitted). But incriminating blackmail does not affect incentives to engage in innocent conduct. See Berman, Blackmail, supra note 328, at 44 (“It is unclear that the prospect of being blackmailed over innocent activities would be sufficiently great in a regime of legalized blackmail to have any significant effect on the incidence or manner of their performance.”). The discussion below considers only ones that do.

Professor Mitchell Berman’s “evidentiary theory” would criminalize opportunistic, incriminating blackmail because “it tends both (a) to cause or threaten identifiable harm, and (b) to be undertaken by a morally blameworthy actor.” 332 Mitchell N. Berman, The Evidentiary Theory of Blackmail: Taking Motives Seriously, 65 U. Chi. L. Rev. 795, 836 (1998) [hereinafter Berman, Evidentiary Theory]. Berman’s theory is unpersuasive largely for the reasons Ric Simmons lays out: First, in general, motive is (rightly) irrelevant to criminality; second, and more important, incrimi­nating blackmail does not necessarily reveal a “morally blameworthy” motive. 333 See Simmons, Private Plea Bargains, supra note 12, at 1157–61. As Simmons observes, both the Model Penal Code and some state statutes actually condone certain forms of incriminat­ing blackmail. 334 Id. at 1158. Ohio, for example, pro­vides a defense to blackmail when the black­mailer’s “purpose was limited to . . . [c]ompelling another to refrain from misconduct or to desist from fur­ther misconduct [or] [p]reventing or redressing a wrong or injustice.” 335 Ohio Rev. Code Ann. § 2905.12(C)(1)–(2) (2018) (internal statutory subsection references omitted).

As Simmons shows, Berman’s insistence that incriminating blackmail is morally blameworthy “is contingent upon the premise that we all have a civic duty to report crime, and it is morally wrong to fail in that duty in order to pursue one’s goals.” 336 Simmons, Private Plea Bargains, supra note 12, at 1159; see also, e.g., Joel Feinberg, The Paradox of Blackmail, 1 Ratio Juris 83, 85–88 (1988) [hereinafter Feinberg, Paradox of Blackmail] (contending that members of society owe a civic duty to report crime, even if not a legal one); Ginsburg & Shechtman, supra note 331, at 1858 (arguing that incriminating blackmail “default[s] on a duty” to the state to report crime). According to Berman, such failure “tends to bespeak a disregard for the common good and the concrete interests of actual and potential victims.” 337 Berman, Evidentiary Theory, supra note 332, at 861.

There is certainly not, in the ordinary case, any legal duty upon citizens to report crime. And whether there is a civic or moral duty seems question­able, at best. Less than half of all criminal victimizations are reported (by victims) to the police; 338 Jennifer L. Truman & Rachel E. Morgan, Bureau of Justice Statistics, U.S. Dep’t of Justice, Criminal Victimization, 2015, at 6 tbl.4 (2018),‌content/‌pub/​‌pdf/‌cv15.pdf [] (finding that 46.5% of violent victimizations and 34.6% of property victimizations were reported to police in 2015). witnesses likely report crime far less often than that. Any civic reporting duty is honored only in the breach. Berman disregards, moreover, the possibility that incriminating blackmail in fact serves the interests of “actual and potential victims” by deterring crime, thus fulfilling the very same obligations the putative reporting duty imposes. 339 Berman, Evidentiary Theory, supra note 332, at 861. Likewise, it is far from clear that blackmail meets Berman’s first criterion for criminalization—causing or threatening identifiable harm—when the blackmailer is “achieving some level of punishment and deterrence at no cost to the state.” See id. at 1161–62 & n.114.

In the shoplifting context, specifically, it would be strange to maintain that retailers violate a civic duty to report crime at the same time public authorities complain they report too often—even threatening to fine retailers for reporting. 340 The mayor of Beech Grove, Indiana, labeled the city’s Walmart a “public nuisance,” threatening to fine the company for each subsequent call to the police. See Laird, supra note 163, at 19. A different account reports that Walmart would be fined only for subsequent petty-shoplifting calls. See Sampson et al., supra note 100; see also Michael Gonzalez, Portage Considers Fining Businesses for Calling the Police Too Often, Chi. Trib.: Post-Trib. (Apr. 10, 2018), (on file with the Columbia Law Review). It would also be odd to say that CEC disregards victims’ interests when it works at the behest of those very same victims. And how, moreover, are we to discern CEC’s true motives? Profit is surely a principal motive, but CEC professes other plausible motives as well. Berman’s theory cannot resolve these questions.

Landes and Posner argue that “the decision to discourage blackmail follows directly from the decision to rely on a public monopoly of law enforcement in . . . criminal law.” 341 Landes & Posner, supra note 10, at 42. That public monopoly, in turn, traces to concerns about under- or overdeterrence: underdeterrence if blackmailers sell their incriminating information too cheaply; overdeterrence if they blackmail individuals whom prosecutors, in their exercise of discretionary nonenforcement, would have left alone. 342 See id. at 38, 42. Part II has responded to these arguments already. 343 See supra sections II.C–.D. So has Professor Jennifer Gerarda Brown in her cogent critique of the Landes–Posner position. 344 See Brown, Blackmail as Private Justice, supra note 171, at 1953–58 (challenging key assumptions Landes and Posner make in their argument). Brown, in turn, would rest the illegality of incriminatory blackmail largely on the “transparency” arguments I also dispatched earlier. Compare id. at 1966–73 (identifying public enforcement as “crucial to the declarative function of criminal law”), with supra section II.C.2 (describing how criminal justice processes largely exclude the public, thereby undermining transparency-related arguments). Brown touches upon “[r]esource disparities,” Brown, Blackmail as Private Justice, supra note 171, at 1970, and “procedural protections,” id. at 1972, which I address at supra sections I.A and II.A.1, respectively. Simmons joins me in rejecting Brown’s arguments. See Simmons, Private Plea Bargains, supra note 12, at 1165–71. In short, when there is no significant problem of judgment-proof defendants, and little if any efficiency of scale from centralized public enforcement, there is scant reason to assume the public monopoly on criminal justice is justified. 345 See, e.g., Steven Shavell, The Optimal Structure of Law Enforcement, 36 J.L. & Econ. 255, 276–79 (1993).

The blackmail theory that comes closest to supporting a ban on retail justice is Lindgren’s. Lindgren understands blackmail as “the seeking of an advantage by threatening to press an actual or potential dispute that is primarily between the blackmail victim and someone else.” 346 Lindgren, Unraveling the Paradox, supra note 171, at 703. Incriminat­ing blackmail, to Lindgren, is “bargaining with the state’s chip,” 347 Id. at 702; cf. Landes & Posner, supra note 10, at 42 (discussing how blackmail potentially interferes with state objectives). which is “unfair in that the threatener uses leverage that is less his than someone else’s.” 348 Lindgren, Unraveling the Paradox, supra note 171, at 703. It also “involves suppressing the state’s interests.” 349 Id. at 672. According to Brown, Lindgren “would outlaw blackmail because it harms third parties”—here, the state—“by compromising their rights.” 350 Brown, Blackmail as Private Justice, supra note 171, at 1963.

Note that, in acting through the retail justice companies, retailers may not actually gain any net leverage if they also waive their civil-recovery rights. Even putting that aside, Lindgren’s critics, including both Brown and Simmons, have the stronger position. They point out that “the blackmailer ‘appropriates’ the state’s leverage but also creates some deterrence value that inures to the benefit of the general public.” 351 Id. at 1965. “Nothing is actually being ‘stolen’ from the state,” in other words; rather, the blackmailer “is advancing the state’s goals (at least part of the way) and saving the state money.” 352 Simmons, Private Plea Bargains, supra note 12, at 1156. Indeed, it is not completely clear that Lindgren would oppose the retail justice model. Lindgren allows that “permit[ting] one to threaten exposure of a . . . crime when honestly seeking restitution in a matter related to the exposure . . . may be an appropriate rule” under his theory, “since the threatener’s personal interest is likely to be substantial when the claim pressed is related to the information threatened to be exposed.” Lindgren, Unraveling the Paradox, supra note 171, at 715. The issue, to Lindgren, would largely turn on whether the retail justice companies’ fees exceed “any reasonable restitution.” Id.

One can also read Lindgren, however, as worried about the black­mailer’s unjust gain rather than the blackmail victim’s loss. 353 See Joel Feinberg, Harmless Wrongdoing 363 n.45 (1988); Brown, Blackmail as Private Justice, supra note 171, at 1963–64. Indeed, this reading, more than any other theory, likely captures our intuition about the “wrongness” of retail justice: Even if we posit that retail justice benefits suspects, victims, and society more broadly (by deterring shoplifting efficiently), why should retail justice companies be permitted to profit from appropriating the state’s leverage?

As Berman has argued, however, and as Lindgren later conceded, 354 James Lindgren, Blackmail: An Afterword, 141 U. Pa. L. Rev. 1975, 1988 (1993). Lindgren’s theory is better at describing blackmail than justifying its prohibition. “Lindgren provides no reason,” Berman writes, “why use of someone else’s leverage for individual gain should be made unlawful, let alone criminal.” 355 Berman, Blackmail, supra note 328, at 54 (emphasis added); see also Feinberg, Paradox of Blackmail, supra note 336, at 83–85 (arguing that an undeserved gain is insufficient, in a liberal society, to justify criminalization when there is no corresponding harm); cf. Hanoch Dagan, The Law and Ethics of Restitution 26 (2004) (arguing that unjust enrichment is not a legal argument but a “loose framework as well as an invitation for normative inquiry”); Emily Sherwin, Restitution and Equity: An Analysis of the Principle of Unjust Enrichment, 79 Tex. L. Rev. 2083, 2110 (2001) (advocating an understanding of unjust enrichment “as an organizing principle rather than a decisional standard”). “Furthermore,” Berman adds, “if the use of such leverage is wrongful, it’s not clear why the squandering of another’s chips—by deciding neither to threaten nor to make a given disclosure—is not likewise wrongful and thus properly criminalizable.” 356 Berman, Blackmail, supra note 328, at 54–55.

In the end, that retail justice may qualify as blackmail under some (perhaps many) state statutes does not tell us that the reasons for its prohibition are well founded. Blackmail laws have always rested on shaky theoretical footing. And the theories that may justify prohibiting blackmail in certain settings do not extend persuasively to the case of retail justice.

III. Summary and Recommendations

Close examination of retail justice paints a finer picture than popular press accounts can afford. Even a casual observer, to be sure, can compile a laundry list of qualms about the industry. But no shorter is the list of (well-founded) grievances about public criminal justice. After careful reflection, it is not clear that retail justice is worse than its public counterpart, and in several important respects it may be better.

This conclusion holds, however, only if certain conditions—suggested in the preceding discussion and crystallized in this Part—are met. The normative valence of retail justice, that is, depends upon empirical facts about how it is implemented and the environments in which it operates. Rather than ignoring retail justice or trying to stamp it out, lawmakers can help ensure that it operates fairly by regulating it in the following respects.

First, in most circumstances, the availability of retail justice makes shoplifting suspects better off by allowing them to opt out of the criminal justice system, with all its dangers and lingering legal consequences. The exceptional case is one in which the suspect harbors misconceptions about the criminal justice system—believing the expected sanctions to be harsher than they really are—and thus artificially inflates the benefits of avoidance. So, while retail justice should not be prohibited under the banner of protecting suspects’ interests, the state should ensure that retail justice companies do not mislead suspects about the severity of the criminal justice option. And if resources for enforcing this antifraud rule are scarce, they should be focused on jurisdictions in which criminal justice is particularly lenient, where the risk of misapprehension is highest. 357 A more aggressive approach might be to require Miranda-style warnings that inform the suspect about the rate at which prosecutors in the jurisdiction pursue charges against first-time shoplifters, the range and distribution of sentences upon conviction, and the procedural rights available to arrestees and defendants.

Second, in many settings, society might prefer retail justice to criminal justice because it generates deterrence more efficiently. Essentially, the “tuition” fees suspects pay to retail justice companies serve as fines that are administered at lower cost than criminal justice sanctions. If, however, retailers could prevent crime even more cheaply by investing in precautions, then retail justice allows retailers to pursue private gains at the expense of social welfare and should be curtailed. The difficulty is how to identify these settings. One possibility is to require retailers to demonstrate compliance with industry best practices for loss prevention as a precondition to opting into retail justice. 358 See, e.g., Loss Prevention Found., How Loss Prevention Audits Are Used, LPM Insider (May 17, 2018),‌insider/‌inventory-shrinkage/lp101-how-loss-prevention-audits-are-used-2 [] (“The purpose [of a loss-prevention audit] is to reinforce guiding principles and practices that should already be familiar to those being audited and evaluate compliance with known standards of performance.”). In the long run, one could even imagine a municipal “loss prevention code,” akin to a fire code, applicable to retailers above a certain size.

Third, because it buries criminal violations, retail justice distorts official crime data. Given the interest in, and manifold uses of, these data in contemporary society, states should require retail justice companies to publish aggregate data about the cases they process. An analogy is the Clery Act, 359 Jeanne Clery Disclosure of Campus Security Policy and Campus Crime Statistics Act, Pub. L. No. 101-542, 104 Stat. 2384 (1990) (codified at 20 U.S.C. § 1092(f) (2012)). which regulates the reporting of campus crime. While colleges and universities are not required to call the local police whenever they learn about crimes committed on campus, the Clery Act does obligate them to keep and disclose information about certain offenses. Institutions subject to the Clery Act must track and publish crime statistics and maintain public logs with details about each such incident brought to their attention. 360 See id. Clery Act offenses are more serious than shoplifting; nevertheless, given the volume and economic impact of retail theft, reasonable people (say, prospective business owners) may value aggregate shoplifting data in planning their affairs. 361 See, e.g., Modernizing Crime Statistics, supra note 321, at 98 (“Businesses may use [Uniform Crime Report] crime data to learn about the nature and extent of problems in the cities or communities in which they operate or are considering for expansion or relocation opportunities.”).

Fourth, the distributive effects of retail justice are indeter­minate. Many of the potential racial and economic biases that concern critics of retail justice manifest to the same degree in the criminal justice system. It is possible that retail justice could exacerbate bias—for instance, by incorporating bias from suspects’ prior interactions with the police. Yet this is true only if this bias is stronger than those of retailers that exercise discretion in determining when to call the police, many of which may look at the suspect’s record in addition to other personal characteristics. This seems unlikely, though vigilance is appropriate given the stakes. Data reporting by retail justice companies should therefore include information on the race and gender of suspects, as well as the locations of operation, to allow for state and public monitoring of disparate impacts.

Finally, because, under ordinary circumstances, retail justice has the potential to make everyone better off, prosecutors should refrain from enforcing blackmail laws against the retail justice industry, absent particularized concerns including those flagged throughout Part II. In the long run, if the retail justice experiment is successful, lawmakers might consider amending blackmail laws, where necessary, to ensure they do not prohibit the practice.

IV. Extensions

Beyond its parochial significance, the study of retail justice also generates fresh perspectives on several important criminal justice issues. This Part draws out these points. Section IV.A extracts lessons for institutional design from the discussion in section II.C of victims’ incentives to take precautions. Section IV.B demonstrates how retail justice challenges conventional views about police and prosecutorial motives. Finally, section IV.C speculates about the possible next frontiers of private criminal dispute resolution.

A. Lessons for Criminal Justice System Design

Recall the conclusion that, while retail justice may reduce the sum of prevention and enforcement costs relative to the criminal justice system, victim precautions might reduce costs even further. This insight extends in interesting directions. Most economically minded legal scholarship on criminal justice, as noted earlier, analyzes offenders’ incentives alone. Only a handful of writers discuss the role of victim precautions in reducing crime. They have emphasized the ways in which the substantive criminal law is, or could be, used to encourage efficient precautions. For example, the tendency of the criminal law to punish attempts more leniently than completed crimes can be understood as a way of discouraging excessive private investment in precautions, which, by thwarting some completed crimes, turn them into mere attempts. 362 See Ben-Shahar & Harel, Economics, supra note 31, at 341–42; see also Harel, Efficiency and Fairness, supra note 270, at 1211–26 (discussing provocation, the “no retreat” rule, and the classification of property crimes).

What is true of the substantive criminal law is also true of criminal procedure and the design of criminal justice institutions. Just as the rules of criminal law can alter victims’ incentives to invest in precautions, so can other aspects of the criminal process. When returns on participation in the criminal justice system increase sufficiently from the victim’s perspective, victims will begin to rely on the criminal justice system when they would have previously invested in precautions—even where precautions remain more cost-effective from society’s vantage point. Precautions and public law enforcement are substitutes, as others have noted. 363 See, e.g., Charles T. Clotfelter, Public Services, Private Substitutes, and the Demand for Protection Against Crime, 5 J. Urb. Econ. 388, 393 (1978); Tomas J. Philipson & Richard A. Posner, The Economic Epidemiology of Crime, 39 J.L. & Econ. 405, 407–11 (1996).

This substitution effect, in turn, generates two observations. First, in contexts in which cost-effective precautions against crime are available, a criminal process that is costly and cumbersome from the victim’s perspective—while doubtless frustrating to the victim—may be socially beneficial if it encourages victims to invest in efficient precautions. 364 A similar point may hold for criminal procedure protections like the reasonable doubt rule, which reduce expected returns on victim participation in the criminal process. For a similar argument in the property law context, see Abraham Bell & Gideon Parchamovsky, The Case for Imperfect Enforcement of Property Rights, 160 U. Pa. L. Rev. 1927, 1932–40 (2012). Second, wholly apart from the many other downsides of an overweening criminal justice system, 365 See, e.g., Exec. Office of the President of the U.S., Economic Perspectives on Incarceration and the Criminal Justice System 43–45, 50–52 (2016), https://obamawhitehouse.‌sites/‌default/‌files/‌page/‌files/‌20160423_cea_incarceration_criminal_justice.pdf [] (documenting direct government spending on the criminal justice system along with costs imposed on offenders’ families and communities); Dorothy E. Roberts, The Social and Moral Cost of Mass Incarceration in African American Communities, 56 Stan. L. Rev. 1271, 1281–300, 1304 (2004) (discussing how mass incarceration is “damaging social networks, distorting social norms, and destroying social citizenship” in African American communities).
an outsized apparatus may have the additional, unrecognized disadvantage of discouraging certain victims from taking socially efficient precautions. There are important (and illuminating) limits on the reach of this argument, but first consider the basic point.

Suppose the police in my neighborhood excel at catching bicycle thieves—they have ample manpower and spare no expense in tracking down a stolen bike. Suppose they have also made it simple for me to file a bicycle-theft report via a smartphone application. In this world, it would be sorely tempting not to spend money on a professional-grade lock for my bike and not to go out of my way to find a rack every time I need to park. If my bike disappears, I will file a report and the police will get my bike back—using taxpayer resources far greater than what the lock and minor route deviations would have cost me.

Now suppose there is no convenient smartphone app—to file a theft report, I have to get myself to the police station, wait in line, provide owner­ship documentation, and fill out a stack of paperwork. That lock and those bike racks start to look much more attractive, a salutary shift from society’s perspective. Similarly, suppose that filing a report remains easy but the likelihood of recovery is greatly reduced because the police department is resource constrained. Again, investment in private precautions begins to sound more appealing, which is socially beneficial.

That the criminal justice system and private precautions are substitutes tells us something not only about the size or type of criminal justice system we might want but also how to allocate enforcement resources within that system. More specifically, the desire to avoid dis­torting victim incentives supplies an argument for directing enforcement resources toward certain victims and crimes rather than others. Regarding victims, there is no concern about reducing investment incentives among individuals who lack the resources to make the investments in the first place. In other words, while a cheap, effective criminal process may reduce investment in socially efficient precautions among the wealthy (like retailers), it is unlikely to affect the spending patterns of the poor. The debate about whether the criminal law is over- or underenforced in poor communities—especially poor communities of color—is fraught and complex. 366 See, e.g., Randall Kennedy, Race, Crime, and the Law 10–12, 364–86 (1997) (discussing different ideological positions on policing and crime in African American communities); Tracey L. Meares & Dan M. Kahan, The Wages of Antiquated Procedural Thinking: A Critique of Chicago v. Morales, 1998 U. Chi. Legal F. 197, 206–11 (arguing that some concerns around police overenforcement in inner-city and minority communities erroneously assume that those communities oppose higher levels of policing); David Thacher, The Distribution of Police Protection, 27 J. Quantitative Criminology 275, 284–86 (2011). The suggestion here is only that the victim-precaution angle supplies one argument in favor of those who would prefer greater enforcement in these settings.

Just as the substitution argument may apply to some victims but not others, the same is true for crimes. Here, too, it makes sense to allocate criminal justice enforcement resources where they will not lead victims to forgo spending on socially efficient precautions. In some contexts, for example, there may be no accessible, efficient precautions because available precautions are ineffective or unduly expensive (in either financial or personal terms). In other contexts, the private harms of victimization—which ordinarily are not compensated through the criminal process—may be so substantial that victims will continue to take precautions even when law enforcement is effective and inexpensive for the victim.

*   *   *

The preceding discussion may help make sense of public hostility toward Walmart’s heavy demands on the criminal justice system. As a taxpayer, Walmart is entitled to some basic level of public protection, regardless whether it takes socially efficient precautions to prevent crime on its premises. But when it begins to make excessive demands on the criminal justice system—demands, perhaps, beyond those it would make if it did take efficient precautions—then continuing to meet those demands begins to look more like (regressive) redistribution than merely spreading the costs of crime across the tax base. 367 See Harel, Efficiency and Fairness, supra note 270, at 1207–08 (discussing an “equal costs” model for distribution of protection). At this point, it only seems right that Walmart itself, or the consumers who choose to patronize it, pay the excess. By reducing the value of criminal justice assistance to Walmart, we may encourage Walmart to shift its deterrence strategy away from enforcement and toward prevention.

B. Police and Prosecutorial Motives

Acquiescence in retail justice, section II.C.2 argued, may best be understood as an exercise, rather than abdication, of police and prosecutorial discretion. It is discretion at the wholesale level. But why are the public authorities exercising their discretion in this way? Put differently, what can acquiescence in retail justice tell us about police and prosecutorial preferences and priorities?

Legal scholars tend to assume that prosecutors seek to maximize either their convictions or conviction rates, often because these are thought to be the measures by which they are evaluated for promotion or election. 368 See supra note 32 and accompanying text. The retail justice story challenges this position. Most shoplifting cases are easy wins for prosecutors—the stolen goods are recovered and surveillance footage captures the offender in the act. Plea deals come quickly. Prosecutors fixated on collecting convictions would be foolish to cut off this flow of easy wins. Likewise, if police aimed to maximize arrests, as some maintain, 369 See supra note 33 and accompanying text. why would they not push hard against, rather than encourage, an arrangement that steals easy ones out from under their gaze?

What retail justice suggests instead is something like a crime-control model of prosecutorial and police behavior. As long as they are assured that retailers are taking care of the problem and not allowing thieves to run rampant through their stores, prosecutors and police seem generally content to focus their attention and resources on other problems. Again, this is not because there are not arrests to be made and convictions to be counted—there are (and cheap ones). But the resources are better spent elsewhere. This behavior seems less consistent with a conviction-maximizing hypothesis than with models in which the goal is to maximize social welfare or deterrence subject to a budget constraint. 370 See, e.g., Frank Easterbrook, Criminal Procedure as a Market System, 12 J. Legal Stud. 289, 295–96 (1983). It is also consistent with a model in which law enforcement authorities maximize both social welfare and their own career prospects. See, e.g., Edward L. Glaeser et al., What Do Prosecutors Maximize? An Analysis of the Federalization of Drug Crimes, 2 Am. L. & Econ. Rev. 259, 288 (2000). Sticking it to shoplifters probably never made any cop or prosecutor’s career.

C. The Next Frontiers

Recall CEC’s ultimate vision: “to reinvent the way petty crimes are handled, starting with retail theft.” 371 Appellant’s Opening Brief, supra note 34, at 11 (emphasis added). If retail theft is just the beginning for the industry, what comes next? Put another way, what characteristics of the retail setting have fostered these private justice institutions, and are there other settings that share the same traits? Several conditions, it seems, con­duce to the particular model of “offender-funded” private justice the retail justice companies embody. The claim is not that each of these conditions is strictly necessary on its own, but rather that, the more that are satisfied, the more likely it is that “offender-funded” private justice will work.

First, the stakes are low in the typical shoplifting case—the average take is $129. 372 See Rainey & Hobbs, supra note 281. This allows retail justice companies to extract fees sufficiently high to deter future thefts and thus protect victim and third-party interests. Given offenders’ solvency limits, the same is unlikely to be true for more serious crimes. 373 See Landes & Posner, supra note 10, at 42; Levmore & Fagan, supra note 18, at 326, 328. Moreover, because retailers at least partly internalize the risks of recidivism, they are unlikely to opt for private justice when it would underdeter; 374 See supra notes 280–281 and accompanying text. police and prosecutors would be less likely to step aside in such a case as well.

Second, shoplifters are typically nonviolent. Were the contrary true, private justice companies might be more reluctant to take responsibility for suspects because of physical safety risks to their employees and poten­tial future victims. 375 See Brown, Blackmail as Private Justice, supra note 171, at 1940. Any concerns about underdeterrence would be heightened as well.

Third, shoplifting has identifiable victims, in contrast with so-called “victimless” crimes like drug use or prostitution. Victims are the most obvious candidates to initiate the private justice process. 376 See Landes & Posner, supra note 10, at 32 (discussing the need for a system of bounties to create incentives for private enforcement of laws prohibiting “victimless” crimes). Moreover, shoplifting victims often know who the offenders are. Perhaps not “often” in an absolute sense, if it is true that shoplifters are apprehended only one time in forty-eight. 377 See supra notes 91–93 and accompanying text. But when an offender’s identity will ever be known, the chances are high the victim knows it through in-store surveillance—only infrequently will police investigation supply the information. 378 Cf. Landes & Posner, supra note 10, at 31–32 (comparing the probability of a tortfeasor being apprehended with the probability of a party in breach of contract being apprehended).

Fourth, a relatively small number of victims each suffers a huge number of thefts. 379 See supra notes 35–36 and accompanying text. This reduces transaction costs by allowing private justice companies to profit from contracting with a few major clients. Were the number of incidents much lower and dispersed among victims, not only would transaction costs rise, but also private justice companies might need to raise enrollment fees. This would reduce the rate at which suspects enrolled, as private justice became unaffordable or, at the least, less appealing relative to the criminal justice system. It might also heighten the sense that private justice companies are exploiting suspects, attracting state attention.

Finally, shoplifting victims have few appealing options for sanction­ing offenders other than calling the police. “Offender-funded” private justice is, in a sense, a method of outsourcing sanctions. In some settings, the nature of the relationship between offender and victim will allow for flexible and tailored (nonlegal) sanctions internal to the relationship. This is not the case for retailers and their arms-length customers.

Where else, then, might these conditions be present? The retail setting itself supplies one obvious example, as shoplifting is not the only prevalent form of retail theft. Theft by employees only narrowly trails shoplifting as the leading source of inventory shrinkage. 380 See Hollinger, 2017 Survey, supra note 21, at 8. The average take is higher for employees than shoplifters, 381 See Rainey & Hobbs, supra note 281 (reporting that employees stole an average of $715 compared to $129 stolen by customers). which might suggest the need for a higher fee to achieve adequate deterrence. At the same time, the employment relationship allows the employer to monitor its employees closely, achieving some deterrence through means other than the fee.

It is possible that many retailers prefer to use job-related sanctions—such as demotions, job transfers, or pay reductions—to punish employees who steal. This could explain why CEC’s employee-targeted program appears to be slow to launch. 382 CEC seems to be trying to penetrate the employment market. In 2014, the company announced a program called “CEC Return” to help retailers address employee theft. See Press Release, CEC, Corrective Education Company Announces CEC Return, a Restorative Justice Education Program that Addresses Employee Theft, MarketWired (Oct. 13, 2014),‌press-release/‌Corrective-Education-Company-Announces-CEC-Return-Restorative-Justice-Education-1956856.htm/ []. There is little information available about CEC Return. In 2016, the company then premiered a program called “CEC Retain,” “an employee engagement offering used to prevent employee theft and more fully engage new employees.” See Corrective Education Company (CEC) Expands Its Product Line and Sales Team, CEC (June 16, 2016), https://www.​‌home/‌blog/‌corrective-education-company-cec-expands-its-product-line-and-sales-team (on file with the Columbia Law Review). But it is also not implausible to imagine employers that would prefer to keep discipline separate from the terms and conditions of employment, to maximize the chances of restoring and preserving the employment relationship.

Another example might be a landlord who oversees a large number of rental units in an urban setting. The landlord, just by the odds, likely encounters all sorts of nonviolent, low-level criminal offenses on her properties, including drug use (or distribution) and noise violations. Doing nothing may be unappealing, as these crimes likely depress rental values, making the landlord a “victim” of even so-called “victimless” crimes. Calling the police, as in the retail environment, may be          time-consuming, ineffectual, and frightening to current or prospective tenants. Eviction is surprisingly expensive. 383 See, e.g., Lucas Hall, The True Cost of Eviction Is More than $5,000, Landlordology (July 17, 2015),‌cost-to-evict-a-tenant []; see also Matthew Desmond, Evicted 114 (2016) (pegging the cost of eviction at $600 in low-income areas of Milwaukee from 2008 to 2009). Compared to these more traditional options, an “offender-funded” model of private justice has a certain appeal.

There is also the possibility of a different, plausible model of for-profit private justice. Instead of outsourcing sanctions, private justice could outsource adjudication. Consider two examples. University disci­plin­ary committees frequently adjudicate student crimes committed on campus—from the mundane, like vandalism, to the explosive, like sexual assault. 384 See, e.g., Collin Binkley et al., College Disciplinary Boards Impose Slight Penalties for Serious Crimes, Columbus Dispatch (Nov. 23, 2014),‌content/​‌stories/‌local/‌2014/‌11/‌23/‌campus-injustice.html [] (“Colleges . . . use campus disciplinary boards to pass judgment on students accused of violent crimes, including rape and assault. Sometimes, schools handle crime and punishment without ever reporting violations to police. Most cases never go to court.”); see also id. (discussing cases involving trespassing, theft, and property destruction). The school’s continuing relationship with its students makes available sanctioning mechanisms—like academic probation or suspension—that may be preferable to what an external provider (including a court) can offer. If, however, a private justice company can adjudicate cases more efficiently, schools might pay them to handle this part of the process.

Likewise, many employers—not just retailers battling employee theft—handle on-the-job crimes in-house. 385 See, e.g., C.D. Shearing & P.C. Stenning, Private Security and Private Justice: The Challenge of the 80s, at 31 (1983) (“Systems which have been established within industry and commerce for disciplining workers are the epitome of . . . ‘systems of private justice’ and are increasingly supplanting the criminal justice process . . . as the means of dealing with crime in industry.”). It is possible that some would prefer to outsource adjudication and then impose job-related sanctions on those found guilty. Of course, nothing would stop an employer (or university) from outsourcing both adjudication and sanctions. The decision presumably would turn on a comparison of the sanctioning alternatives, likely focusing on cost-effectiveness, including consequences for the future of the relationship.


Legal scholars long have focused on the role of police and prosecutors as gatekeepers for the criminal justice system. 386 See, e.g., Stephanos Bibas, The Need for Prosecutorial Discretion, 19 Temp. Pol. & C.R. L. Rev. 369, 369 (2010) (describing prosecutors as the “key gatekeepers who ration criminal justice”); Charlie Gerstein & J.J. Prescott, Process Costs and Police Discretion, 128 Harv. L. Rev. Forum 268, 272, 276–78 (2015),​04/vol128_GersteinPrescott.pdf [] (describing how police exercise of the discretionary power to arrest shapes criminal justice outcomes). How these actors exercise their discretion determines who escapes the criminal justice system’s net and who is entangled within it. This Article highlights how late to the game police and prosecutors can be—the station house phone rings only when the private gatekeepers that precede them place the call. In these settings, the question is not whether individuals suspected of crime will enter the justice system but rather which justice system—public or private—will assess their guilt and administer any necessary sanctions.