Harvard Law Review Harvard Law Review Harvard Law Review

Criminal Justice System

Pessimistic Police Abolition

Introduction: Abolition’s Second Premise

The movement for police abolition seeks to eliminate, or massively downsize, American policing. Mariame Kaba’s Yes, We Mean Literally Abolish the Police1×1. Mariame Kaba, Opinion, Yes, We Mean Literally Abolish the Police, N.Y. Times (June 12, 2020), https://www.nytimes.com/2020/06/12/opinion/sunday/floyd-abolish-defund-police.html [https://perma.cc/UX6A-J6WG]. marks the movement’s new life in the mainstream.2×2. See Amna A. Akbar, An Abolitionist Horizon for (Police) Reform, 108 Calif. L. Rev. 1781, 1783 n.5 (2020). More detailed descriptions of abolition include Alex S. Vitale, The End of Policing (2017); Derecka Purnell, Becoming Abolitionists (2021); and Mariame Kaba, We Do This ’Til We Free Us (2021). Akbar’s article also situates abolition’s place within (and, as she notes, historically outside of) legal scholarship. See Akbar, supra, at 1785–86. Featured prominently in the New York Times, Kaba’s article presents the two premises that build to abolition.

First, the police “suppress[] marginalized populations.”3×3. Kaba, supra note 1. In the last decade alone, dozens of police murders have shaped public discourse. Hundreds more have not.4×4. Mapping Police Violence, https://mappingpoliceviolence.org [https://perma.cc/HW2W-VNHS]. See generally Franklin E. Zimring, When Police Kill (2017). And policing causes still more pain in other contexts — including through harassment,5×5. See, e.g., Christy Mallory, Amira Hasenbush & Brad Sears, Williams Inst., Discrimination and Harassment by Law Enforcement Officers in the LGBT Community 1 (2015), https://williamsinstitute.law.ucla.edu/wp-content/uploads/LGBT-Discrimination-by-Law-Enforcement-Mar-2015.pdf [https://perma.cc/S4QA-23CK]. stop and frisk,6×6. See, e.g., Stop-and-Frisk in the de Blasio Era, NYCLU (Mar. 14, 2019), https://www.nyclu.org/en/publications/stop-and-frisk-de-blasio-era-2019 [https://perma.cc/BJ43-58PC]. and rape.7×7. See, e.g., Crimes, Henry A. Wallace Police Crime Database, Bowling Green State Univ., https://policecrime.bgsu.edu/Home/Crimes# [https://perma.cc/KJ67-6URA]. This is true even if officers mean well and even if reforms sometimes work.8×8. A number of studies have suggested that reforms can have real, but limited, success. See, e.g., Emily Owens, David Weisburd, Karen L. Amendola & Geoffrey P. Alpert, Can You Build a Better Cop?, 17 Criminology & Pub. Pol’y 41, 43 (2018); Robin S. Engel, Hannah D. McManus & Tamara D. Herold, Does De-escalation Training Work?, 19 Criminology & Pub. Pol’y 721, 732–41 (2020); Anna Harvey & Taylor Mattia, Reducing Racial Disparities in Crime Victimization, J. Urb. Econ. (forthcoming) (manuscript at 1) (on file with the Harvard Law School Library). It is likely that individuals will do injustice when they are given weapons and told to enforce the law, especially in a country with a legacy of racism and strong notions of property. See, e.g., Vitale, supra note 2, at 7–30. The institution causes harm.9×9. See generally sources cited supra notes 1–2; Andrea J. Ritchie, Invisible No More: Police Violence Against Black Women and Women of Color (2017); Laurence Ralph, The Torture Letters: Reckoning with Police Violence (2020); Max Felker-Kantor, Policing Los Angeles (2018); Colin Gordon, Citizen Brown: Race, Democracy, and Inequality in the St. Louis Suburbs 120–44 (2019). But Kaba requires a second premise before reaching abolition. Namely, “police officers don’t do what you think they do,” like “catch the bad guys” or “find the serial killers.”10×10. Kaba, supra note 1. Together, then, abolition’s two premises are that (1) policing creates harm in a racist pattern, and (2) policing does not prevent harm. From here, abolition follows easily. Nobody thinks the state should needlessly harm people in a racist pattern.11×11. See, e.g., id. Arguing for abolition by doubting police effectiveness is common. See, e.g., Vitale, supra note 2, at 31 (“The Police Are Not Here to Protect You.”); Aya Gruber, Commentary, Policing and “Bluelining,” 58 Hous. L. Rev. 867, 887 (2021) (“Policing Does Not Fight Crime.”); Derecka Purnell, How I Became a Police Abolitionist, The Atlantic (July 20, 2020, 1:37 PM), https://www.theatlantic.com/ideas/archive/2020/07/how-i-became-police-abolitionist/613540 [https://perma.cc/EU42-DKGL] (“Police couldn’t do what we really needed.”).

An inconvenient truth for police abolitionists is that their second premise is far less certain than their first. Empirical evidence has suggested, over and over again, that policing does decrease crime.12×12. See, e.g., Aaron Chalfin, Benjamin Hansen, Emily K. Weisburst & Morgan C. Williams, Jr., Police Force Size and Civilian Race 11 (Nat’l Bureau of Econ. Rsch., Working Paper No. 28202, 2020) (also finding a greater anticrime police effect on Black victims than white ones); Jonathan Klick & Alexander Tabarrok, Using Terror Alert Levels to Estimate the Effect of Police on Crime, 48 J.L. & Econ. 267, 277 (2005) (finding a general anticrime effect); John M. MacDonald, Jonathan Klick & Ben Grunwald, The Effect of Private Police on Crime: Evidence from a Geographic Regression Discontinuity Design, 179 J. Royal Stat. Soc’y 831, 838–42 (2015) (same); Steven Mello, More COPS, Less Crime, 172 J. Pub. Econ. 174, 189 (2019) (same); John MacDonald, Jeffrey Fagan & Amanda Geller, The Effects of Local Police Surges on Crime and Arrests in New York City, PLoS One, June 16, 2016, at 1, 10 (same); Patrick Sharkey, Why Do We Need the Police?, Wash. Post (June 12, 2020), https://www.washingtonpost.com/outlook/2020/06/12/defund-police-violent-crime [https://perma.cc/MX38-XJT2] (“One of the most robust, most uncomfortable findings in criminology is that putting more officers on the street leads to less violent crime.”). The empirical questions around prison abolition are distinct. Long prison sentences do not stop crime. See David Roodman, The Impacts of Incarceration on Crime 1 (Sept. 2017) (unpublished manuscript) (on file with the Harvard Law School Library). According to one leading study, hiring an additional officer prevents about 0.06 to 0.10 homicides.13×13. Chalfin et al., supra note 12, at 11. Because policing causes harm, everyone (upstanding officers included) should encourage the development of police alternatives.14×14. See, e.g., V. Noah Gimbel & Craig Muhammad, Are Police Obsolete? Breaking Cycles of Violence Through Abolition Democracy, 40 Cardozo L. Rev. 1453, 1509–26 (2019) (outlining some of these projects); Allegra M. McLeod, Envisioning Abolition Democracy, 132 Harv. L. Rev. 1613, 1616–17 (2019) (same, with a focus on Chicago). For some other examples, see Peter Hermann, He Used to Sell Drugs on D.C. Streets. Now He’s Paid to Make Them Safer., Wash. Post (Dec. 14, 2018, 6:00 AM), https://www.washingtonpost.com/local/public-safety/he-used-to-sell-drugs-on-dc-streets-now-hes-paid-to-make-them-safer/2018/12/13/eea91b66-e12e-11e8-ab2c-b31dcd53ca6b_story.html [https://perma.cc/BNG8-RCHA]; Healing Systemic and Interpersonal Violence, The Health All. for Violence Intervention, https://www.thehavi.org [https://perma.cc/K9VG-PFS4]. But those alternatives have not yet matched policing’s anticrime effect.15×15. Of abolitionist anticrime programs, Princeton sociologist and police critic Patrick Sharkey writes that “we have every reason to believe” they can work, but that they “obviously[] require[] rigorous testing.” Sharkey, supra note 12. Until that changes, the imagination-based16×16. E.g., Amna A. Akbar, Toward a Radical Imagination of Law, 93 N.Y.U. L. Rev. 405, 405–06 (2018); Melissa Stein, Toward the Abolitionist Imagination, Pol’y Rsch. Assocs. (Sept. 1, 2021), https://www.prainc.com/abolitionist-imagination [https://perma.cc/D62Z-4RU5]; Abolition and Imagination: Divesting from Police, Investing in Community Health + Wellbeing, Equi Inst. (June 8, 2020), https://www.equi-institute.org/blog/2020/6/5/abolition-and-imagination-divesting-from-police-investing-in-community-health-wellbeing [https://perma.cc/ZP9H-HVU6]. abolitionist has no good answer to those in the center (and even, quietly, on the left17×17. See Charles M. Blow, Opinion, “Defund the Police” Is Dead. Now What?, N.Y. Times (Aug. 31, 2022), https://www.nytimes.com/2022/08/31/opinion/defund-the-police.html [https://perma.cc/FSF4-4T5C]. ) who strike at abolition’s second premise with empirical data. These people do not doubt police injustice. They doubt police alternatives. Their caution is like a doctor’s during pharmaceutical trials: the disease hurts people, but the drug might be worse.

Abolitionists ought to muster a response to empirical critics beyond faith in their second premise. If someone thinks policing does a lot of harm, but that abolition would increase crime, what should they do? Put differently, is there a police abolition for pessimists? This Note creates two analytical frameworks for approaching that question — one deontological and one consequentialist. These frameworks draw on prior work, though they aim to present that work in a newly analytic and persuasive model. Ultimately, both frameworks show why police violence is worse than private violence. They do not necessarily endorse abolition.

Most importantly, this Note sets an agenda. Abolitionists should consider what to do while police alternatives (hopefully) build empirical support. Doing so will help them convert empirical skeptics and maintain their own moral consistency. And to be sure, nonabolitionists share a duty to engage in the debate. Empirical pessimists ought to make sure that the police harm they willingly sanction is justified by police effectiveness.

I. Framing Police Abolition

Three initial points frame the discussion: (1) the United States has a violent crime problem, (2) this Note consciously explores a narrow policy space, and (3) police abolition has an underrecognized analog in death penalty abolition.

First, the United States has a lot of seriously harmful violent crime. Abolitionists often imply otherwise for the same reason they accept Kaba’s second premise: abolition would come easier if police work were useless. Michelle Alexander’s The New Jim Crow makes one version of this argument — doubting the presence of violence. She writes that “[t]he uncomfortable reality is that convictions for drug offenses — not violent crime — are the single most important cause of the prison boom in the United States.”18×18. Michelle Alexander, The New Jim Crow 99 (2010). The book deserves praise for spotlighting mass incarceration and the war on drugs. Conservatives may paint an unfair picture of American crime, but Alexander is wrong. As Professor James Forman Jr. writes, “more prisoners are locked up for violent offenses than for any other type, and just under 25% (550,000) of our nation’s 2.3 million prisoners are drug offenders.”19×19. James Forman, Jr., Racial Critiques of Mass Incarceration: Beyond the New Jim Crow, 87 N.Y.U. L. Rev. 21, 48 (2012). Professor John Pfaff similarly observes that most state prisoners have been convicted of a violent crime.20×20. John F. Pfaff, Locked In: The True Causes of Mass Incarceration — And How to Achieve Real Reform 32 tbl.1.1 (2017); see also Press Release, Wendy Sawyer & Peter Wagner, Prison Pol’y Initiative, Mass Incarceration: The Whole Pie 2022 (Mar. 14, 2022), https://www.prisonpolicy.org/reports/pie2022.html [https://perma.cc/W8TE-EPPX]. To be clear, this debate involves prisons, not police, but it illustrates the breadth of American violence. Consider murder. In 2010, “US homicide rates were 7.0 times higher than in other high-income countries.”21×21. Erin Grinshteyn & David Hemenway, Violent Death Rates: The US Compared with Other High-Income OECD Countries, 2010, 129 Am. J. Med. 266, 266 (2016). By 2015, they were 7.5 times higher.22×22. Erin Grinshteyn & David Hemenway, Violent Death Rates in the US Compared to Those of the Other High-Income Countries, 2015, 123 Preventative Med. 20, 22 (2019). And COVID-19 has only worsened the homicide epidemic.23×23. The murder rate has increased since the pandemic’s onset. See Richard Rosenfeld & Ernesto Lopez, Council on Crim. Just., Pandemic, Social Unrest, and Crime in U.S. Cities 3 (2022).

The realization that Alexander’s picture is too simple inspires a second abolitionist move — doubting whether violent crime is so bad. After all, “ideas about violence embedded in the law are deeply entangled with race, with gender, with class.”24×24. David Alan Sklansky, A Pattern of Violence 5 (2021). Professor Sklansky’s book does not endorse abolition. These arguments are powerful, but properly defined, violence is still bad. Again consider murder. Biases are not why people want to stop murder. Because there is a lot of murder to be stopped in the United States, and policing stops crimes like murder, problematizing violence doesn’t entail abolition.

Second, for this Note’s purposes, abolishing the police means taking most of the resources currently given to the police and giving them elsewhere. Because abolitionists mean what they say,25×25. For instance, Kaba’s article is titled “Yes, We Mean Literally Abolish the Police.” Kaba, supra note 1. abolition is not a collection of reforms, such as demilitarization or improved training;26×26. See, e.g., #8cantwait, https://8cantwait.org [https://perma.cc/LP3Z-R4JZ] (campaign for eight police reforms); see also Lucas Manfield, Dallas Has Been Dispatching Social Workers to Some 911 Calls. It’s Working., Dall. Observer (Dec. 10, 2019, 4:00 AM), https://www.dallasobserver.com/news/dallas-has-been-dispatching-social-workers-to-some-911-calls-its-working-11810019 [https://perma.cc/JRY5-9LEH]. a rhetorical device aimed at “shift[ing] the Overton window”;27×27. Sean Illing, The “Abolish the Police” Movement, Explained by 7 Scholars and Activists, Vox (June 12, 2020, 11:00 AM), https://www.vox.com/policy-and-politics/2020/6/12/21283813/george-floyd-blm-abolish-the-police-8cantwait-minneapolis [https://perma.cc/KTY9-DKYU]. or even a pilot program financing limited police alternatives. Funding alternatives does not amount to abolition because cities can support them alongside the police. To wit, cities already fund their police even as they improve their schools, ideally decreasing the need for police.

The Note also does not refer to abolition as an act of revolutionizing taxes to support unprecedented social services,28×28. See, e.g., Purnell, supra note 11 (“If we are committed to eliminating this harm long-term, then society must offer quality housing, food, day care, transit, employment, debt cancellation, and free college so that people will not be stuck in unhealthy relationships because they need food, money, health insurance, or a place to live.”). or as a philosophy of “cooperation instead of individualism, . . . mutual aid instead of self-preservation.”29×29. Kaba, supra note 1. In abolitionist terms, it refers to W.E.B. Du Bois’s “negative” vision, not the “positive” one in which fundamental “institutions and ideas . . . would be remade.”30×30. Allegra M. McLeod, Prison Abolition and Grounded Justice, 62 UCLA L. Rev. 1156, 1162 (2015); see W.E. Burghardt Du Bois, Black Reconstruction in America 189 (1935) (explaining that, to become truly free, former slaves “must have land; they must have education”). Abolitionists often fight this narrow version of their proposals.31×31. See, e.g., Fabiola Cineas, What the Public Is Getting Right — And Wrong — About Police Abolition, Vox (Oct. 30, 2020, 11:30 AM), https://www.vox.com/21529335/abolish-the-police-movement [https://perma.cc/2762-Z7N3]. And improved social services might well decrease crime. The problem is that actual decisionmakers face narrow options. In today’s police budgets, policymakers find pots of money far too small to revolutionize society. New York City’s notoriously expensive32×32. See, e.g., Abby Narishkin, Kuwilileni Hauwanga, Hailey Gavin, Jessica Chou & Andrea Schmitz, The Real Cost of the Police, And Why the NYPD’s Actual Price Tag Is $10 Billion a Year, Bus. Insider (Aug. 12, 2020, 3:28 PM), https://www.businessinsider.com/the-real-cost-of-police-nypd-actually-10-billion-year-2020-8 [https://perma.cc/9RB9-CW4S]. police department receives less than ten percent of the city’s budget, and almost no city commits the majority of its funds to policing.33×33. What Policing Costs, Vera Inst. of Just., https://www.vera.org/publications/what-policing-costs-in-americas-biggest-cities [https://perma.cc/LLX2-YDEX]. New York City already maintains a Department of Education budget over three times larger than its Police Department budget.34×34. Compare Vera Inst. of Just., A Look Inside the New York City Police Department Budget 1 (2020), https://www.vera.org/downloads/publications/a-look-inside-the-new-york-city-police-department-budget.pdf [https://perma.cc/4YCM-ZSF3] (stating that the “annual cost of the NYPD is $11 billion”), with Funding Our Schools, N.Y.C. Dep’t of Educ., https://www.schools.nyc.gov/about-us/funding/funding-our-schools [https://perma.cc/3YQJ-5KB6] (stating that the 2022–2023 school year budget is $38 billion). More broadly, the nation spends about $115 billion on police each year.35×35. What Policing Costs, supra note 33. But the federal government spent $803 billion in direct household stimulus during COVID-19.36×36. See Three Rounds of Stimulus Checks. See How Many Went Out and for How Much, Pandemic Oversight (Feb. 17, 2022), https://www.pandemicoversight.gov/data-interactive-tools/data-stories/three-rounds-stimulus-checks-see-how-many-went-out-and-how-much [https://perma.cc/BUQ8-D7KU]. That relief failed to overhaul American social services. Finally, the United States employs fewer police officers per capita than the average nation.37×37. Council of Econ. Advisers, Exec. Off. of the President, Economic Perspectives on Incarceration and the Criminal Justice System 45 (2016). This all means reinvesting police resources will not lead to strong social services. It will lead to slightly stronger, but still quite weak social services.

Because abolitionists do make demands on narrowly bound actors — Kaba demands an “immediate” fifty percent cut in the police force38×38. Kaba, supra note 1.  — they must engage with the narrow question (should cities abolish policing today?), in addition to the ambitious one (should the United States abolish policing as part of a future turn toward socialism and cooperation?). Today, if not tomorrow, reinvesting police resources will not reorder society. It will increase crime. Should cities do it anyway?

An abolitionist might fold their ambitious program into the narrow question by arguing that short-term abolition will make long-term socialism more likely. The future payoff could justify an immediate increase in crime. This Note would encourage such an argument, as long as it is honest about abolition’s short-term empirical limitations, the terrors of crime, and the need to answer the narrow question. But the argument still faces serious difficulties: it is not obvious why eliminating the police would make socialism more feasible,39×39. People might feel less altruistic as crime rises. why a social revolution should start with policing, whether next week’s beneficiaries outweigh tomorrow’s crime victims, or even whether a socialist revolution is desirable. This Note won’t take on those lofty questions. Instead, it will explore two surprisingly intuitive reasons why empirical pessimists might embrace abolition as it could look today.

Third, death penalty discourse already contains many of the arguments this Note brings to policing discourse. It is unsurprising that policing and the death penalty are cousins. Both are prominent and controversial arenas in which the state uses force against its citizens. And one can imagine an argument for death penalty abolition that hinges on empirics, like the common argument for police abolition. Perhaps the death penalty is unnecessary because it fails to deter murders.40×40. To be clear, it probably does fail to deter murders. See, e.g., The Death Penalty: Questions and Answers, ACLU, https://www.aclu.org/other/death-penalty-questions-and-answers [https://perma.cc/L4UY-NPV7]. But death penalty abolitionists have learned to supplement that framing. On the dominant view, the death penalty is unacceptable because it violates people’s rights, or is “shocking to [the] conscience.”41×41. Furman v. Georgia, 408 U.S. 238, 369 (1972) (Marshall, J., concurring); see also Hugo Adam Bedau, ACLU, The Case Against the Death Penalty (rev. ed. 2012), https://www.aclu.org/other/case-against-death-penalty [https://perma.cc/45Q2-7JXE]; Cass R. Sunstein & Adrian Vermeule, Is Capital Punishment Morally Required? Acts, Omissions, and Life-Life Tradeoffs, 58 Stan. L. Rev. 703, 704 (2005); Death Penalty, Amnesty Int’l, https://www.amnesty.org/en/what-we-do/death-penalty [https://perma.cc/V794-K7YW]. This position is durable. It survives even if the death penalty is an excellent deterrent, and it has contributed to nearly half of America believing murderers should be spared.42×42. Death Penalty, Gallup, https://news.gallup.com/poll/1606/death-penalty.aspx [https://perma.cc/GDM6-X8EE]. This Note explores how police abolitionists can build something similar for themselves. Even where the Note does not explicitly reference the death penalty debate, readers are encouraged to find parallels to it and to pull intuitions from it.

II. A Deontological Argument for Police Abolition: The Police Do Harm

One reason to abolish the police — if the police stop crime — sounds in the intuition that doing harm is worse than allowing harm. If police officers are the state, then when an officer kills, the state kills. Meanwhile, when the state defunds the police, it merely allows additional murders. If the state should never kill but may sometimes allow murders, then it should abolish the police — or so the argument would go. Note, as above, that death penalty abolitionists have developed precisely the same argument against state-sponsored killing.43×43. See, e.g., Sunstein & Vermeule, supra note 41, at 707. Before digging in, this argument is “deontological” because it reflects the intuition that people are bound by certain duties, independent of consequences.44×44. For an introductory survey of positions here, see Larry Alexander & Michael Moore, Deontological Ethics, in Stanford Encyclopedia of Philosophy (Edward N. Zalta ed., Winter 2021 ed.), https://plato.stanford.edu/entries/ethics-deontological [https://perma.cc/4SEH-6M45]. This Note will explore neither precisely what deontology entails, nor exactly how the doing/allowing distinction works.45×45. For an introductory survey of positions here, too, see Fiona Woollard & Frances Howard-Snyder, Doing vs. Allowing Harm, in Stanford Encyclopedia of Philosophy (Edward N. Zalta ed., Fall 2021 ed.), https://plato.stanford.edu/entries/doing-allowing [https://perma.cc/9KLU-L2B7]. It will instead draw out the distinction’s intuitive appeal, situate it within law and morality, and apply it to policing.

Consider a canonical thought experiment developed by Professors Philippa Foot and Judith Jarvis Thomson — the “Transplant Problem.”46×46. See Judith Jarvis Thomson, The Trolley Problem, 94 Yale L.J. 1395, 1396 (1985) (naming the Transplant Problem alongside similar thought experiments); Philippa Foot, The Problem of Abortion and the Doctrine of the Double Effect, 5 Oxford Rev. 5, 9 (1967) (presenting the problem). Suppose Rachel strikes five people with her car. All five barely survive. They appear in the emergency room each needing a different organ. Can Sarah, the doctor on call, kill a single sleeping patient in order to harvest their organs and save the five?47×47. See Thomson, supra note 46, at 1396. No. As Thomson writes, “[e]verybody to whom I have put this . . . hypothetical case says, No, it would not be morally permissible.”48×48. Id.; see also Edmond Awad, Sohan Dsouza, Azim Shariff, Iyad Rahwan & Jean-François Bonnefon, Universals and Variations in Moral Decisions Made in 42 Countries by 70,000 Participants, 117 PNAS 2332, 2334–36 (2020) (studying common responses to the Trolley problem, which suggest that deontology has a common pull). The reason is that doing harm to the one is worse than allowing harm to the five.49×49. See Foot, supra note 46, at 10. Importantly, too, this doing/allowing distinction has pedigree in the law. Very briefly, in both civil and criminal law, allowing harm creates liability in many fewer situations than doing harm does.50×50. Liability for allowing harm is generally limited to special relationships. See, e.g., Tarasoff v. Regents of the Univ. of Cal., 551 P.2d 334, 343 (Cal. 1976); Model Penal Code § 2.01(3) (Am. L. Inst. 1985).

Another of Foot’s hypotheticals drives the point home and turns the analysis toward policing and the death penalty. A judge faces “rioters demanding that a culprit be found for a certain crime and threatening otherwise to take their own bloody revenge on a particular section of the community. The real culprit being unknown, the judge sees himself as able to prevent the bloodshed only by framing some innocent person . . . .”51×51. Foot, supra note 46, at 8. It strikes most people as obvious that the judge should not frame an innocent person (doing harm to them) even to save the riot victims. The parallel to abolition is straightforward. The state cannot sacrifice innocent people simply to prevent crime — whether through judges (Foot), the death penalty (elsewhere), or police (here).

Philosophers can also justify the doing/allowing distinction from first principles. To Thomson, “[t]he reason why the surgeon may not proceed in Transplant is that if he proceeds, he maximizes utility . . . but in so doing he would infringe a right of the young man’s.”52×52. Thomson, supra note 46, at 1404 (emphasis added) (emphasis omitted). In classical Kantian language, perhaps one cannot use the young man simply as a means.53×53. For a helpful explanation of Kant’s theory here, see Samuel Kerstein, Treating Persons as Means, in Stanford Encyclopedia of Philosophy (Edward N. Zalta ed., Summer 2019 ed.), https://plato.stanford.edu/entries/persons-means [https://perma.cc/QJY5-8TVG]. That is, maybe agents ought to respect individuals as individuals, not treat them as faceless and replaceable members of the public. Alternatively, the doing/allowing distinction might hold because the duty not to harm someone is simply greater than the duty to help them, perhaps for reasons of respect or noninterference. In Foot’s words, “even where the strictest duty of positive aid exists, this still does not weigh as if a negative duty were involved.”54×54. Foot, supra note 46, at 12.

The argument has now taken a rough shape. The state should not do harm through policing, even if policing prevents harm, because doing harm is worse than allowing it. This commitment is supported by various intuitions; it has parallels in existing law and in the death penalty debate; and it can be justified.

A. Shaping Deontological Abolition

This position must be presented carefully. A crude version would, for instance, forbid the state from taxing.55×55. Christopher Lewis & Adaner Usmani, The Injustice of Under-Policing in America, 2 Am. J.L. & Equal. 85, 104–05 (2022). Taxation works only because the state backs it with the threat of property seizure. And taking property from someone, or removing them from their property, does harm to them. Police abolitionists don’t want to be tax abolitionists.

Abolitionists can dodge libertarianism in a few ways. First, the abolitionist can draw their doing/allowing distinction at bodily violence. The state may collect taxes by taking property, limited by a per se rule against touching a body. But issues would persist. Do people want to forbid the state from touching active shooters? A more effective position forbids only unjust state violence. On this view, violence is sometimes allowed — for example, to stop an active shooter. But when an act of violence is unjust, it is much worse to do it than to allow it. Whether violence is just might depend on factors including the legal breach that the violence addresses, the violence’s utility, the violence’s severity, and the violence’s community effect.56×56. Consider some examples of how these factors apply: jaywalking is a minor violation, while murder is a major one; violence against an incapacitated offender has little utility, while violence against an active offender has more; taxation is minor violence, while shooting is severe; and violence against a Black community may have a strong group effect, while violence against a white community may have a weaker one. On these criteria, tax-based property takings may be perfectly just. Police officers, meanwhile, often hurt innocent people,57×57. See, e.g., Richard A. Oppel Jr., Derrick Bryson Taylor & Nicholas Bogel-Burroughs, What to Know About Breonna Taylor’s Death, N.Y. Times (Aug. 23, 2022), https://www.nytimes.com/article/breonna-taylor-police.html [https://perma.cc/A338-4WNS]. hurt people for no good reason,58×58. See, e.g., Justin Fenton, These Two Baltimore Cops Robbed, Sold Drugs and Disgraced Their Badges. But Is a Combined 454 Years in Prison Fair?, Balt. Sun (Jan. 11, 2021, 5:00 AM), https://www.baltimoresun.com/news/crime/bs-md-ci-cr-king-murray-sentence-reduction-20210111-yb7ovx2q3vgy7ewyohfsanwgi4-story.html [https://perma.cc/G5TZ-BYG2]. hurt people severely,59×59. See, e.g., Christina Carrega, 5 Years After Eric Garner’s Death, A Look Back at the Case and the Movement It Sparked, ABC News (July 16, 2019, 5:42 AM), https://abcnews.go.com/US/years-eric-garners-death-back-case-movement-sparked/story?id=63847094 [https://perma.cc/749R-WNKS]. and hurt particularly vulnerable communities.60×60. See sources cited supra note 9. Taking for granted that police reform will not eliminate such injustice, this framing makes room for both abolition and taxation.61×61. Interestingly, policing looks worse than the death penalty on this picture. The police probably kill many more innocent people than the death penalty does. See Execution List 2021, Death Penalty Info. Ctr. (Feb. 10, 2022), https://deathpenaltyinfo.org/executions/2021 [https://perma.cc/2RQ5-VXSE] (listing the eleven people executed in the United States in 2021). And it is probably worse to kill innocent people than guilty people.

Next, deontological abolition doesn’t need to be absolutist. Consider whether a mayor may kill one person to save one million. Philosophers have long thought that perhaps they should.62×62. See, e.g., Michael S. Moore, Torture and the Balance of Evils, 23 Isr. L. Rev. 280, 328 (1989). “Threshold” and “ratio” deontology handle this problem.63×63. See, e.g., Larry Alexander, Deontology at the Threshold, 37 San Diego L. Rev. 893, 899 (2000). On the first view, deontological commitments act like “a dam holding back a lake,” binding actors except in true catastrophes.64×64. Id. On the second, deontological commitments always represent magnitudes.65×65. See id. Killing may be five, ten, or fifteen times worse than letting die. These positions have supporters and critics.66×66. Compare, e.g., Sunstein & Vermeule, supra note 41, at 719, with Alexander, supra note 63, at 911–12. But because most people comfortably believe that a surgeon can kill one patient to save one million but not five, abolitionists may believe the same.

The deontological abolitionist argument is now complete: when the state does unjust violence, it is much worse than when the state merely allows it.67×67. This represents the ratio view. On a threshold view, the argument would be that state-done harm is forbidden, unless it would prevent a specified amount of state-allowed harm. This Note will use the ratio view from here on. For that reason, there should be a strong presumption against state institutions that do unjust harm, even if they offset private harm. Like the death penalty, policing is one such institution.

B. Deontology and the State

Professors Cass Sunstein and Adrian Vermeule have argued that the doing/allowing distinction is either incoherent or unwise as applied to the state. In countering their argument, this section demonstrates that deontology ought to bind the state, if it binds anyone.

Though Sunstein and Vermeule write in terms of acting and omitting rather than doing and allowing, and focus on the death penalty, the argument they challenge is the one developed here. Sunstein and Vermeule “doubt that the distinction between state actions and state omissions can bear the moral weight given to it.”68×68. Sunstein & Vermeule, supra note 41, at 707 (“Whatever its value as a moral concept where individuals are concerned, the act/omission distinction misfires in the general setting of government regulation.”). The crux of their doubt seems to be that a state is unique in (1) being forced to act, and (2) being forced to be a “but-for”69×69. An act is a “but-for” cause of an outcome if the outcome would not have occurred but for the act. cause of bad outcomes: “Suppose that government officials face a choice between two (and only two) packages of policies for reducing the murder rate” — one with the death penalty and one without it.70×70. Sunstein & Vermeule, supra note 41, at 722. The package without the death penalty will “ensure a significant increase in the number of deaths. In this setting, it is hard to make sense of the claim that capital punishment involves intentional government ‘action’ in some morally distinctive way.”71×71. Id.

Sunstein and Vermeule miss the mark.72×72. This Note is not the first to argue as much. See, e.g., Carol S. Steiker, No, Capital Punishment Is Not Morally Required: Deterrence, Deontology, and the Death Penalty, 58 Stan. L. Rev. 751, 755 (2005); Adam Omar Hosein, Doing, Allowing, and the State, 33 Law & Phil. 235, 242–45 (2013); Eric Blumenson, Killing in Good Conscience: What’s Wrong with Sunstein and Vermeule’s Lesser Evil Argument for Capital Punishment and Other Human Rights Violations?, 10 New Crim. L. Rev. 210, 212 (2007). The state is unique neither in needing to act nor in needing to cause deaths.73×73. See Hosein, supra note 72, at 245, 248. Consider Sarah, the surgeon from the Transplant Problem. Sarah will act (or “choose between policies”) no matter what. Her action will either be to care for her original patient and allow the five new patients to die, or to kill her original patient and save the five new patients. Sarah’s decision will also cause future deaths, no matter what she chooses. Sunstein and Vermeule’s two arguments for state uniqueness are therefore wrong. If only the doing/allowing distinction can explain what Sarah should do (that is, not kill), and the state is no different than Sarah, then the distinction explains what the state should do too (that is, not kill — through the death penalty or the police).

Sunstein and Vermeule next argue that “[e]ven if the distinction between acts and omissions can be rendered intelligible in regulatory settings” — which it can — “its moral relevance is obscure.”74×74. Sunstein & Vermeule, supra note 41, at 709. To them, the rationales for deontology “lose their force where government is the moral agent.”75×75. Id. at 724. They think these rationales are that those who do harm are more vicious than those who allow it, that imposing an individual duty to aid would interfere with freedom, and that outcomes improve when people use a doing/allowing heuristic.76×76. Id. at 725. These are interesting points, but only a consequentialist could think they are what motivates deontology.

Rather, deontology is usually motivated by respect for individual persons, or a belief that negative duties are stronger than positive ones.77×77. See, e.g., supra p. 1163. There is no reason for these concerns to apply differently to the state. This explains why most people think Foot’s judge shouldn’t falsely convict, and why nobody thinks they should do so without batting an eye (except, maybe, for Sunstein and Vermeule). It further explains why establishing tort liability for a city’s failure to protect is nearly impossible.78×78. See, e.g., Weiner v. Metro. Transp. Auth., 433 N.E.2d 124, 127 (N.Y. 1982). And indeed, even Sunstein and Vermeule open their paper by calling the position that deontology binds states “intuitive.”79×79. Sunstein & Vermeule, supra note 41, at 707. What may be less intuitive is that this point challenges policing the same way it does the death penalty.

C. Why Not Deontological Abolition?

At least two serious problems remain for deontological abolition. Police misconduct may not be state action, and not every “ratio” or “threshold” justifies abolition. First, though it initially seems obvious that police violence is “done” by the state, the reality is less clear.80×80. This argument comes from a forthcoming book by Professors Christopher Lewis and Adaner Usmani. See Christopher Lewis & Adaner Usmani, What’s Wrong with Mass Incarceration? (forthcoming). When a mayor decides to expand their city’s police force, they do not will the resulting police violence. Instead, the violence flows from individuals whom the mayor can regulate but not control. In this way, a mayor looks similar to any government regulator choosing what to allow. Expanding the police may be an allowing versus allowing choice, not a doing versus allowing choice.81×81. Police violence here clearly departs from the death penalty, which is deliberately done by the state.

This Note considers two rebuttals here, though it is open to more. First, the badge of the state might imbue state agents with the citizenry’s collective responsibility. When the public entrusts officers to act in its name, it could take their actions as its own. Second, the United States may implicitly direct officers to commit injustice. Cities give officers tools to harm (guns); promote them for doing things that are often unjust (making arrests);82×82. See generally Shaun Ossei-Owusu, Police Quotas, 96 N.Y.U. L. Rev. 529 (2021). and then routinely fail to hold them accountable (criminally, civilly, or professionally) when they misbehave.83×83. See infra p. 1173.

Interestingly, these rebuttals imply that police violence could become less “done” by the state in certain circumstances: (1) if the public rejected the police as its agents and began to see officers as normal individuals acting in a regulated capacity; or (2) if cities radically improved police incentives and accountability. Note that the first scenario would raise a separate question about permitting only certain citizens to use violence. This Note cannot settle exactly where a mayor’s behavior falls between doing and allowing. Similarly, does an individual do the violence of their hitman? What about their irresponsible bodyguard? This Note has aimed to show that abolitionists face a difficult question.

If abolitionists can handle the first critique, a second relates to their deontological ratio or threshold. Because most people are not absolutists, the doing/allowing distinction entails only that one ingredient in the policymaker’s soup should be the doing/allowing multiplier. That multiplier represents how many times worse it is to do injustice than to allow it. To be precise, a policymaker should remove any given officer for whom D × P > C — where D is the doing/allowing multiplier, P is the amount of police injustice that the officer would likely create, and C is the amount of crime injustice that removing the officer would likely allow. To illustrate, if doing harm is two times worse than allowing harm (if the multiplier, D, is two), then officers should be released unless or until releasing the next officer would cause twice as much private violence (C) as it would reduce police violence (P).

Whether this formula endorses abolition depends on the values of the three variables. This Note will not provide them. Defining the proper doing/allowing multiplier is quite difficult, even for those with deontological intuitions. Providing a concrete number would improperly anchor readers, and the most truth-conducive step forward is to allow discourse to fill in the space. Regarding the empirical variables, too, this Note will not stake a firm claim. The first reason for this decision is that this Note aims to provide an adaptable framework. The second is that existing studies are insufficient for confident assertions. Though one above-referenced study finds that an average officer prevents 0.06 to 0.10 homicides,84×84. Chalfin et al., supra note 12, at 11. no static estimate can capture officers’ likely diminishing marginal effect on crime.85×85. That is, changing the number of officers probably has a greater effect on crime when there are fewer total officers. Another complicating factor is that C and P may look different in different cities. Yet another is that police reforms might decrease P. And perhaps most importantly, these variables need to import opportunity costs. Because hiring an officer takes money away from other programs with (at least minimal) anticrime effects, determining P and C is not easy.

Rather than provide definitive answers, this framework can assist policymakers in making morally consistent choices about policing. If a mayor estimates their city is near a police “equilibrium” — where hiring another officer would stop as much injustice as it would create, or P = C — they should remove at least some officers from the street, given the harm multiplier. The presence of equilibrium seems a fair enough assumption, if cities are more or less rational and do not yet respect the doing/allowing distinction. Even if cities are irrational, assuming equilibrium makes sense if cities are as likely to have too many as too few police for the purpose of minimizing violence. If anything, police political power86×86. Cf., e.g., Stephen Rushin, Police Union Contracts, 66 Duke L.J. 1191, 1201–38 (2017) (discussing police union power). might suggest P > C for many cities already. Though, again, policy should be shaped in light of city-specific and deontology-conscious details beyond this Note’s scope. Finally, to be transparent, this analysis implies that a policymaker should hire any officer for whom D × P > C. But there is no good reason to believe most cities are biased against policing.87×87. A critic may think the United States’s relatively small police force suggests the opposite. However, empirical variables might differ in other countries, other countries might not be deontological enough, and other countries might just have too many police. And there is no situation in which this argument supports hiring more officers than a rational city should absent the doing/allowing multiplier. In sum, deontological intuitions lend support to something like police abolition, though not without barriers.

III. A Consequentialist Argument for Police Abolition: The Police Pain Multiplier

Of course, one can also dismiss deontological abolition by dismissing the doing/allowing distinction in the first place. “Consequentialism” is defined by the conviction that an agent should always act to create the best outcomes, even if that requires them to do harm.88×88. This is an incredible oversimplification. For examples of consequentialism’s nuance, see the essays in Consequentialism and Its Critics (Samuel Scheffler ed., 1988); and Walter Sinnott-Armstrong, Consequentialism, in Stanford Encyclopedia of Philosophy (Edward N. Zalta & Uri Nodelman eds., Winter 2022 ed.), https://plato.stanford.edu/entries/consequentialism [https://perma.cc/QHZ2-4ACD]. Note that Sunstein and Vermeule go both one step further and one step less far. They argue that deontology applied to the state is incoherent, not just wrong, while also refusing to argue that deontology is wrong at all in the interpersonal sphere. See Sunstein & Vermeule, supra note 41, at 720–21. Can abolitionists convince consequentialists on their own terms?

A consequentialist might turn to abolition if police injustice is more harmful than private injustice, all else equal. In the language used above, there could be a multiplier lurking behind every police injustice — a “police pain multiplier.” If an injustice committed by an officer is worse than one committed by a civilian, then of course policymakers should weigh a change in police violence more heavily than a change in private violence. Indeed, they should use a formula that looks quite similar to the deontologist’s formula above, just with a different kind of multiplier. And if that police pain multiplier is large enough, the formula would support abolition. The question at this argument’s heart is whether policing really does cause special pain. This Part argues that it does, because of its unique emotional impact.

A. On Policing’s Inequity

A keen reader might note two other consequentialist arguments alive in the public discourse: (1) that police violence is special because it is racist (or classist, or transphobic) in application;89×89. See, e.g., Kaba, supra note 1 (“Everywhere, they have suppressed marginalized populations . . . .”); Akbar, supra note 16, at 416 (“The rebellions in Ferguson and Baltimore brought to the center of public discourse the violence that police, prosecutors, and the courts exercise in and against Black communities every day.”). and (2) that police violence is special because of its racist (or classist, or transphobic) source.90×90. See, e.g., Kaba, supra note 1 (“Policing in the South emerged from the slave patrols . . . .”); Gruber, supra note 11, at 874 (“Policing Did Not Originate to Fight Crime.”); Dorothy E. Roberts, The Supreme Court, 2018 Term — Foreword: Abolition Constitutionalism, 133 Harv. L. Rev. 1, 20–29 (2019). These arguments struggle not because police violence is defensible, but because private violence isn’t either. This section will focus on racism, but its points hold for other inequities too. As an initial point, these arguments fall under the consequentialist umbrella. The first position relies on the premise that group inequity is an intrinsically bad consequence, like pain. On this view, a mayor should sometimes allow slightly more pain in order to make that pain more equitably dispersed. The second position relies on the premise that the persistence of a system founded on oppression is an additional bad consequence over and above inequity itself.91×91. Similar arguments to these abound for the death penalty, see Race and the Death Penalty, ACLU, https://www.aclu.org/other/race-and-death-penalty [https://perma.cc/4MZG-827U], but this Note will not analyze them. Also, some might point to policing’s racist legacy as an explanation for why its practices are currently oppressive, rather than a reason to abolish it. That makes sense, but is not at issue here. Still others may identify the duty to end racist systems as deontological in nature. Insofar as such a duty could sanction an increase in racist private violence, it is implausible.

These arguments do not entail abolition. Concerning unequal application, about 25% of 2019’s police firearm killings victimized Black people, who comprise only 14.2% of the population.92×92. Police Shootings Database 2015–2022, Wash. Post (Nov. 17, 2022), https://www.washingtonpost.com/graphics/investigations/police-shootings-database/ [https://perma.cc/Z7H9-NBKH]; Race and Ethnicity in the United States: 2010 Census and 2020 Census, U.S. Census Bureau (Aug. 18, 2021), https://www.census.gov/library/visualizations/interactive/race-and-ethnicity-in-the-united-state-2010-and-2020-census.html [https://perma.cc/V5G2-7CD4]. Yet 54% of 2019’s total murders victimized Black people.93×93. Expanded Homicide Data Table 2, Fed. Bureau of Investigation: Unif. Crime Reporting Program (2019), https://ucr.fbi.gov/crime-in-the-u.s/2019/crime-in-the-u.s.-2019/tables/expanded-homicide-data-table-2.xls [https://perma.cc/E2JJ-KCZV]. If all police killings are unjust, and the police stop murders, these data suggest that taking police off the street would make unjust killing rates less equitable. Total police killings may not be a strong proxy for unjust police killings. But even the racial disparity in unarmed victims of police killings — 32% of whom are Black94×94. Police Shootings Database 2015–2022, supra note 92. 32% reflects the data from 2015 through 2022. In 2019 alone, 22% of unarmed victims of police shootings were Black. Id.  — is lower than the racial disparity in murder victims. This Note cannot exhaustively compare the many racial inequity tradeoffs between police and private violence. For example, police disproportionately target Black people for traffic stops.95×95. See Frank R. Baumgartner, Derek A. Epp & Kelsey Shoub, Suspect Citizens: What 20 Million Traffic Stops Tell Us About Policing and Race 76 (2018). But Black people are also disproportionately likely to die by traffic accident.96×96. See Nat’l Highway Traffic Safety Admin., U.S. Dep’t of Transp., Early Estimates of Motor Vehicle Traffic Fatalities and Fatality Rate by Sub-Categories in 2020, at 8 (2021). It is therefore difficult to know whether racial equity in outcomes would improve or suffer from removing traffic cops. The point is that abolitionists cannot assume removing officers will decrease racial disparities.

The source of violence’s inequity is also racist, whether in policing or in crime. Many have documented that modern policing finds its roots in slave patrols and Jim Crow.97×97. See, e.g., Vitale, supra note 2, at 34–49; K.B. Turner, David Giacopassi & Margaret Vandiver, Ignoring the Past: Coverage of Slavery and Slave Patrols in Criminal Justice Texts, 17 J. Crim. Just. Educ. 181, 184–86 (2006). Crime is similar. For one, the majority of hate crimes are racially motivated.98×98. See Hate Crime Statistics, U.S. Dep’t of Just., https://www.justice.gov/hatecrimes/hate-crime-statistics [https://perma.cc/3ST5-JAK4]. Moreover, the following three facts overlap to sustain oppression through crime: (1) because of current and former racist practices like redlining, a disproportionate number of Black Americans experience disadvantage and live in majority-Black communities;99×99. See generally, e.g., Richard Rothstein, The Color of Law (2017); Calvin Schermerhorn, Why the Racial Wealth Gap Persists, More than 150 Years After Emancipation, Wash. Post (June 19, 2019, 6:00 AM), https://www.washingtonpost.com/outlook/2019/06/19/why-racial-wealth-gap-persists-more-than-years-after-emancipation [https://perma.cc/LQZ9-YFBV]; Scott Winship, Richard V. Reeves & Katherine Guyot, The Inheritance of Black Poverty: It’s All About the Men, Brookings (Mar. 22, 2018), https://www.brookings.edu/research/the-inheritance-of-black-poverty-its-all-about-the-men [https://perma.cc/YD4M-CAA5]. (2) disadvantage definitively leads to criminal behavior;100×100. See, e.g., Jeffery T. Ulmer, Casey T. Harris & Darrell Steffensmeier, Racial and Ethnic Disparities in Structural Disadvantage and Crime: White, Black, and Hispanic Comparisons, 93 Soc. Sci. Q. 799, 813 (2012); Ruth D. Peterson & Lauren J. Krivo, Macrostructural Analyses of Race, Ethnicity, and Violent Crime: Recent Lessons and New Directions for Research, 31 Ann. Rev. Socio. 331, 347 (2005). and (3) crime is generally local and intraracial.101×101. E.g., Rachel E. Morgan & Alexandra Thompson, Bureau of Just. Stat., U.S. Dep’t of Just., Criminal Victimization, 2020 — Supplemental Statistical Tables 1 (2022). There is, then, an obvious line from racist policies, to localized Black disadvantage, to criminal behavior, to Black victims of crime.102×102. See, e.g., Richard Powell, Redlining, Neighborhood Decline, and Violence: How Discriminatory Government Policies Created Violent American Inner Cities (Sept. 2021) (Ph.D. dissertation, City University of New York) (on file with the Harvard Law School Library). Race-specific trauma may also manifest in criminality after being filtered through a still-racist world.103×103. See Monnica T. Williams, Angela M. Haeny & Samantha C. Holmes, Posttraumatic Stress Disorder and Racial Trauma, 32 PTSD Rsch. Q., no. 1, 2021, at 1, 2. That dynamic, too, would have racist roots and would disproportionately affect Black victims. Consider the traffic accident example above. Black Americans are more likely to die by traffic accidents in part because Black communities often lack road infrastructure.104×104. See Governors Highway Safety Ass’n, An Analysis of Traffic Fatalities by Race and Ethnicity 19 (2021). Any given instance of reckless driving in a Black community is more likely to end in manslaughter, purely because of whom the government chooses to protect.

In sum, Black people are more likely to be victimized by both police and private violence, and for racist reasons in both settings. That conclusion reflects a stain on American society, not a defense of policing. But it also shows why abolition does not follow from policing’s racism. To be sure, this argument should motivate policy agendas that can decrease the racial disparities in both police and private violence.105×105. For a comprehensive argument on this point, see Lewis & Usmani, supra note 80. Sadly, abolition might just change the balance between the two.

B. Police Violence and Emotional Trauma

Still, police violence causes more pain than private violence does. Policing may be uniquely damaging on a number of grounds, but at least one is that it causes special emotional trauma.106×106. This Note uses the term “emotional” to describe a range of nontangible conditions, including psychological ones. There are a number of reasons to think this is true. To begin, multiple studies show that police violence causes incredible distress for both victims and their communities, especially marginalized communities. Police victimization is clearly associated with depression,107×107. See, e.g., J.E. DeVylder, H.Y. Oh, B. Nam, T.L. Sharpe, M. Lehmann & B.G. Link, Prevalence, Demographic Variation and Psychological Correlates of Exposure to Police Victimisation in Four US Cities, 26 Epidemiology & Psychiatric Scis. 466, 466 (2017); Jordan E. DeVylder, Hyun-Jin Jun, Lisa Fedina, Daniel Coleman, Deidre Anglin, Courtney Cogburn, Bruce Link & Richard P. Barth, Association of Exposure to Police Violence with Prevalence of Mental Health Symptoms Among Urban Residents in the United States, JAMA Network Open, Nov. 2018, at 1, 1. subjects of stop-and-frisk policing report increased anxiety,108×108. See, e.g., Amanda Geller, Jeffrey Fagan, Tom Tyler & Bruce G. Link, Aggressive Policing and the Mental Health of Young Urban Men, 104 Am. J. Pub. Health 2321, 2324 (2014). and rates of local Black suicide109×109. See, e.g., Ilias Kyriopoulos, Sotiris Vandoros & Ichiro Kawachi, Police Killings and Suicide Among Black Americans, 305 Soc. Sci. & Med., July 2022, at 1, 3 (2022). and mental distress110×110. See, e.g., Jacob Bor, Atheendar S. Venkataramani, David R. Williams & Alexander C. Tsai, Police Killings and Their Spillover Effects on the Mental Health of Black Americans: A Population-Based, Quasi-experimental Study, 392 The Lancet 302, 302 (2018). both rise when police kill Black victims.111×111. See also Jocelyn R. Smith Lee & Michael A. Robinson, “That’s My Number One Fear in Life. It’s the Police”: Examining Young Black Men’s Exposures to Trauma and Loss Resulting from Police Violence and Police Killings, 45 J. Black Psych. 143 (2019) (collecting anecdotes supporting this conclusion); Jordan DeVylder, Lisa Fedina & Bruce Link, Impact of Police Violence on Mental Health: A Theoretical Framework, 110 Am. J. Pub. Health 1704, 1706 (2020) (collecting studies with similar conclusions).

Harrowing personal stories illustrate the trauma of even observing police violence, especially for Black Americans.112×112. See, e.g., Minnyvonne Burke, For Black People, The Mental Effects of Police Killings Linger, NBC News (Nov. 4, 2021, 4:30 AM), https://www.nbcnews.com/news/us-news/black-people-mental-effects-police-killings-linger-n1281353 [https://perma.cc/D5UC-Y4C2]; Kenya Downs, When Black Death Goes Viral, It Can Trigger PTSD-Like Trauma, PBS (July 22, 2016, 8:04 PM), https://www.pbs.org/newshour/nation/black-pain-gone-viral-racism-graphic-videos-can-create-ptsd-like-trauma [https://perma.cc/W96A-L6X6]. One witness of George Floyd’s murder wrote, a full year after Floyd’s death, that “I’m not who I used to be. . . . I used to shake so bad at night my mom had to rock me to sleep. . . . Having panic and anxiety attacks every time I seen a police car, not knowing who to trust . . . . I hold that weight.”113×113. Darnella Frazier, Facebook (May 25, 2021), https://www.facebook.com/darnellareallprettymarie/posts/1727632277437871 [https://perma.cc/KWX6-6R2H]; see also Burke, supra note 112. Of viral videos showing Black victimhood, activist April Reign writes: “To watch how they died, again and again, to hear their last words as they ask what they did wrong or plead for their life, leaves me feeling powerless.”114×114. April Reign, Why I Will Not Share the Video of Alton Sterling’s Death, Wash. Post (July 6, 2016, 5:46 PM), https://www.washingtonpost.com/posteverything/wp/2016/07/06/why-i-will-not-share-the-video-of-alton-sterlings-death [https://perma.cc/6CXD-N5BM].

Maybe private violence has similar effects,115×115. See, e.g., All. for Safety & Just., Crime Survivors Speak 4. but there are also strong theories for why police violence is different.116×116. See DeVylder et al., supra note 111, at 1705–08 (giving eight varyingly compelling reasons that police violence is unique). In the prison context, Professor Allegra McLeod seems to make similar arguments about “dehumanization.” See McLeod, supra note 30, at 1173–84. First, victims of police violence have limited paths for recourse.117×117. See DeVylder et al., supra note 111, at 1707. Officers are notoriously hard to convict for myriad reasons, including prosecutorial discretion, the “blue wall of silence,” and pro-police bias in jury pools.118×118. See, e.g., Kate Levine, How We Prosecute the Police, 104 Geo. L.J. 745, 762–67 (2016); Vitale, supra note 2, at 20–23; German Lopez, Police Officers Are Prosecuted for Murder in Less Than 2 Percent of Fatal Shootings, Vox (Apr. 2, 2021, 11:30 AM), https://www.vox.com/21497089/derek-chauvin-george-floyd-trial-police-prosecutions-black-lives-matter [https://perma.cc/BDF7-GTSY]. Similarly, qualified immunity prevents courts from holding officers civilly liable for constitutional violations unless they violate “clearly established” law — an astonishingly high standard.119×119. See, e.g., Jamison v. McClendon, 476 F. Supp. 3d 386, 404–08 (S.D. Miss. 2020) (documenting the shameful mark of qualified immunity on American law); Nathaniel Sobel, What Is Qualified Immunity, And What Does It Have to Do with Police Reform?, Lawfare (June 6, 2020, 12:16 PM), https://www.lawfareblog.com/what-qualified-immunity-and-what-does-it-have-do-police-reform [https://perma.cc/883C-QQHY]. Even aside from justice in court, police unions have made it nearly impossible to hold officers professionally liable.120×120. See, e.g., Rushin, supra note 86, at 1201–38. More barriers to police accountability abound — including doctrines of Article III standing and class certification.121×121. See generally Sunita Patel, Jumping Hurdles to Sue the Police, 104 Minn. L. Rev. 2257 (2020).

For many, violence is more traumatizing when experienced without a path to recourse.122×122. Cf. Kainaz Amaria & Fabiola Cineas, The Complicated Relief of the Chauvin Verdict, Vox (Apr. 20, 2021, 9:53 PM), https://www.vox.com/2021/4/20/22394618/derek-chauvin-murder-verdict [https://perma.cc/JDT3-Q4MM]. One boy surveyed in Baltimore called the police “scary” for exactly this reason: “[A]t any time they can do whatever. They can hop up on you and just beat you up for no reason.”123×123. Smith Lee & Robinson, supra note 111, at 167 (surveying young boys). Work in the context of domestic violence further demonstrates the emotional importance of receiving recourse,124×124. See Mary P. Koss, The RESTORE Program of Restorative Justice for Sex Crimes: Vision, Process, and Outcomes, 29 J. Interpersonal Violence 1623, 1641–42 (2013) (finding that most victims valued “mak[ing] the responsible person accountable” and were less traumatized when the offender was made accountable); Alexandra Topping, “No Support”: Domestic Abuse Survivors on Feeling Ignored by Police, The Guardian (Sept. 17, 2021, 1:04 PM), https://www.theguardian.com/society/2021/sep/17/no-support-domestic-abuse-victims-on-being-ignored-by-police [https://perma.cc/F57E-5CJ2] (providing anecdotes about recourse’s importance); cf. Ann L. Coker, Paige H. Smith, Martie P. Thompson, Robert E. McKeown, Lesa Bethea & Keith E. Davis, Social Support Protects Against the Negative Effects of Partner Violence on Mental Health, 11 J. Women’s Health & Gender-Based Med. 465, 472 (2002) (finding that victims of abuse who receive support are at lower risk of suicide). as does a dominant theory of tort law’s foundations.125×125. See John C.P. Goldberg & Benjamin C. Zipursky, Recognizing Wrongs 113–14 (2020) (“[T]ort law itself . . . is a particularly salient, visible, and deeply rooted embodiment of the principle of civil recourse.” Id. at 114.). And though this Note focuses mostly on emotional pain, the absence of legal recourse might also prevent police violence victims from receiving the financial support available to other crime victims.126×126. This is a preliminary hypothesis. Compare Janet Nguyen, Protesters Injured by Police Face Big Medical Bills, Marketplace (May 26, 2021), https://www.marketplace.org/2021/05/26/protesters-injured-by-police-face-big-medical-bills [https://perma.cc/3UPB-42K9], with State Crime Victims Compensation, Benefits.gov, https://www.benefits.gov/benefit/4416 [https://perma.cc/URX3-SX9V]. But see Cheryl Corley, Police Settlements: How the Cost of Misconduct Impacts Cities and Taxpayers, NPR (Sept. 19, 2020, 7:00 AM), https://www.npr.org/2020/09/19/914170214/police-settlements-how-the-cost-of-misconduct-impacts-cities-and-taxpayers [https://perma.cc/B9XB-BTPH].

Police violence is also unique because it constitutes institutional betrayal. Institutional betrayal occurs when an institution commits “violations of trust and dependency.”127×127. Carly Parnitzke Smith & Jennifer J. Freyd, Institutional Betrayal, 69 Am. Psych. 575, 577 (2014). When a powerful institution betrays someone — just like when a friend or family member does — it causes “anxiety, sexual dysfunction, and other trauma-related outcomes.”128×128. Id. at 578 (reviewing dozens of studies). Because citizens trust the police and/or rely on them for basic safety, police misconduct is prototypical institutional betrayal.129×129. See Melissa L. Barnes, Collective Ongoing Betrayal Trauma: Gendered and Racialized Police Violence Toward the Black Community 93 (Sept. 2022) (Ph.D. dissertation, University of Oregon) (on file with the Harvard Law School Library) (finding that subjects felt institutional betrayal after reading vignettes about police abuse). The police are especially primed to betray because they often fail to respond to abuse, and an institution’s failure to respond to abuse within it is a frequent cause of betrayal. See Smith & Freyd, supra note 127, at 582. For instance, when the witness of George Floyd’s murder didn’t know “who to trust,” she was likely feeling betrayal’s effects.130×130. Frazier, supra note 113; see also Burke, supra note 112. And because dependency alone creates the conditions for those effects, they can linger long after trust is lost. Meanwhile, only a small minority of private violence satisfies betrayal’s criteria. Admittedly, underpolicing might also cause victims to feel institutional betrayal. But it likely hurts more to be attacked by a trusted party than to suspect that the trusted party could have prevented an attack.131×131. This point looks like it requires the doing/allowing distinction, but it stands even if the distinction has no moral force. It capitalizes on victims feeling that the distinction has force.

A number of other psychological or emotional factors could also distinguish police injustice. Police violence may feel more racist than private violence (even if, as above, it likely isn’t), causing special pain rooted in intergenerational harm.132×132. Cf. Thema Bryant-Davis, Tyonna Adams, Adriana Alejandre & Anthea A. Gray, The Trauma Lens of Police Violence Against Racial and Ethnic Minorities, 73 J. Soc. Issues 852, 861 (2017); Robert T. Carter, Racism and Psychological and Emotional Injury: Recognizing and Assessing Race-Based Traumatic Stress, 35 Counseling Psych. 13, 88 (2007). Police violence may also create an unrivaled sense of domination, given the police’s incredible power.133×133. See Smith Lee & Robinson, supra note 111, at 157–69 (retelling multiple boys’ experiences of policing as an ominous and pervasive force); DeVylder et al., supra note 111, at 1705 (“The Police Are a Pervasive Presence.”). Next, police violence may be more stigmatizing than private violence.134×134. See DeVylder et al., supra note 111, at 1708 (suggesting this is the case); cf. Diane M. Quinn & Stephenie R. Chaudoir, Living with a Concealable Stigmatized Identity: The Impact of Anticipated Stigma, Centrality, Salience, and Cultural Stigma on Psychological Distress and Health, 97 J. Personality & Soc. Psych. 634, 634 (2009) (showing that anticipated stigma is directly related to negative health outcomes). These dynamics and others should be explored in future work. For now, police violence appears at least somewhat more painful than private violence. Decisionmakers should act accordingly.

C. Why Not Consequentialist Abolition?

As with deontological abolition, though, the consequentialist case is vulnerable. First, perhaps policymakers should challenge the assumptions that lead police violence to cause more emotional pain than private violence. For instance, in accepting that police violence represents institutional betrayal and racism where crime does not, a policymaker implies that state policies and racism are not responsible for crime. But both factors do set the conditions for crime to flourish.135×135. See supra p. 1171. Highlighting each’s role in private violence — by essentially instructing people to feel institutional betrayal and racial subordination when crime befalls them — could increase political will for new social and antiracism programs. That might lead to less harm in the long run. Consequentialists framing policing’s uniqueness should keep it in mind.

The second, more obvious problem is: Even if police violence is more painful than private violence, is it so much more painful that cities should choose abolition? To bring back a familiar formula, a policymaker should remove any officer if the injustice the officer is likely to commit (P), multiplied by the police pain multiplier (R), is greater than the crime injustice the officer is likely to prevent (C). That is, if R × P > C for the officer. The same constraints as above hide P and C.136×136. See supra section II.C, pp. 1167–69. And while this Note will not give R a concrete value in the interest of preventing anchoring, it will provide a few thoughts on point.

For one, private violence causes emotional pain, too.137×137. See, e.g., Patrick J. Fowler, Carolyn J. Tompsett, Jordan M. Braciszewski, Angela J. Jacques-Tiura & Boris B. Baltes, Community Violence: A Meta-analysis on the Effect of Exposure and Mental Health Outcomes of Children and Adolescents, 21 Dev. & Psychopathology 227, 227–28 (2009). Private violence traumatizes victims,138×138. See, e.g., James K. Hill, Victims’ Response to Trauma and Implications for Interventions: A Selected Review and Synthesis of the Literature, at iv (2003). and community violence exposure causes depression and anxiety in both children and adults.139×139. See, e.g., Sharon F. Lambert, Rhonda C. Boyd, Nicole L. Cammack & Nicholas S. Ialongo, Relationship Proximity to Victims of Witnessed Community Violence: Associations with Adolescent Internalizing and Externalizing Behaviors, 82 Am. J. Orthopsychiatry 1, 1–2, 5 (2012); Cheryl Clark, Louise Ryan, Ichiro Kawachi, Marina J. Canner, Lisa Berkman & Rosalind J. Wright, Witnessing Community Violence in Residential Neighborhoods: A Mental Health Hazard for Urban Women, 85 J. Urb. Health 22, 33 (2007). As such, today’s policymakers must interrogate the abolitionist’s explanatory theories, and further empirical studies would be quite useful. Further, significant elements of pain are nonemotional, including financial pain and physical pain. Financial pain here can emerge from things like theft, missed work, and medical bills. Physical pain is physical pain. Therefore, when determining a police pain multiplier, policymakers must balance emotional against nonemotional pain. Policymakers should also consider whether police violence victims suffer special, nonemotional pain. For instance, as above, does a lack of legal recourse worsen their financial burdens? Or, as Harvard economist Desmond Ang has shown, do police killings decrease local academic performance more than private killings do?140×140. See Desmond Ang, The Effects of Police Violence on Inner-City Students, 136 Q.J. Econ. 115, 162 (2021). This Note has not focused on these factors, but they are perfectly capable of increasing the police pain multiplier.

In sum, like with deontological abolition and regardless of R’s precise value, the consequentialist formula will likely instruct many cities to remove police officers. The formula will reach that result for cities that are near “equilibrium” in their police forces. Again, equilibrium represents a scenario in which adding an officer will create equally as much police violence as it will prevent crime, in which P = C. For example, suppose the police pain multiplier is two. That could result from (1) police violence causing more emotional pain than crime does; (2) emotional pain rivaling nonemotional pain in severity; and/or (3) police violence causing more nonemotional pain than crime does. This Note has given strong reasons to believe (1). It has left (2) and (3) open. A multiplier of two would instruct cities to remove a significant number of officers, unless they are seriously underpoliced (if P is much lower than C). But as illustrated above,141×141. See supra p. 1169. most cities are probably not so underpoliced. Moreover, the police pain multiplier at least tempers an otherwise pro-police conclusion.


This Note’s first aim has been to set an agenda for abolitionists and their allies: show why police injustice is worse than private injustice. By embracing an unconvincing empirical faith, too many abolitionists have ignored that challenge, leaving their movement politically and philosophically wanting. This failure has also too quickly allowed empirical pessimists to assume that police effectiveness entails police justification.

This Note’s second aim has been to explore some approaches to the task, finding two promising, pushing others aside,142×142. See, e.g., supra section III.A, pp. 1169–72 (on policing’s racial inequity). and reserving yet others for future work.143×143. See, e.g., supra p. 1161 (on abolition’s fueling socialism) & p. 1177 (on victim non-emotional pain). The first promising framework sounds in the doing/allowing distinction, the second in the special emotional pain imposed by police violence. If both are compelling, they easily combine: the deontological multiplier applies to police-created pain, which is greater than expected because of the police pain multiplier. In abstract terms, the resulting prescription would be to remove officers whenever D × (R × P) > C.144×144. D is the doing/allowing multiplier. R is the police pain multiplier. P is a given officer’s effect on police injustice. And C is a given officer’s effect on criminal injustice.

How close these arguments come to justifying abolition is, again, a matter of interpretation and future work. As arguments in this framework are raised and refined, abolition’s moral status will become clearer. All justice-concerned people — police officers included — should want that. Perhaps surprisingly, this Note concludes by encouraging research on police alternatives. The animating conflict here is one between stopping crime and stopping police violence. Though difficult, that trade-off is necessary given today’s possibilities. Hopefully tomorrow’s make it irrelevant.