1. Factors Producing Tension Between Communities and Police
    2. A Vortex of Forces
    1. The Past Two Years
    2. The Past Half Century
    3. A New Epidemic?
    4. Alternate Explanations for the 2015–2016 Trend
    5. The Danger of Predictions
    1. Case Studies: Cynicism, Withdrawal, and the Persistence of Homicides
    2. Lessons and Next Steps



On September 7, 2016, four of the nation’s newspapers of record weighed in on the connected crises in crime and policing. The New York Times revealed the tensions between the Mayor’s office in Chicago and several community and professional groups over a plan to overhaul Chicago’s police disciplinary board—a plan developed in the wake of the shooting of an unarmed teenager, Laquan McDonald, and the release of a video of that killing. 1 Richard A. Oppel Jr., Chicago’s Plan to Toughen Oversight of Police Is Under Fire, N.Y. Times (Sept. 7, 2016), http://nyti.ms/2cq020E (on file with the Columbia Law Review) [hereinafter Oppel, Plan Under Fire]. The Wall Street Journal related a vigorous defense of New York City’s “broken windows” policing strategy—a strategy that has been a recurring source of tension between the city’s minority com­munities and the police, and whose value in crime reduction has been sharply debated. 2 Pervaiz Shallwani & Mara Gay, NYPD Defends Its Policing Strategy, Wall St. J. (Sept. 7, 2016, 8:45 pm), http://on.wsj.com/2cHe15O (on file with the Columbia Law Review). The Washington Post’s lead editorial decried Chicago’s “deadliest month in two decades” that left city officials at odds over the basic elements of public safety policy and strategic responses to the seemingly intractable violence. 3 Editorial, Chicago’s Violence Is a ‘Police Issue’—at Least in Part, Wash. Post (Sept. 7, 2016), http://wpo.st/1A-x1 [http://perma.cc/3YPD-YCYX]. The Chicago Tribune weighed in as well. In it, Chicago Police Superintendent Eddie Johnson declared that the crisis is “not a police issue, it’s a society issue,” pointing to the need for stricter gun controls. 4 Jeremy Gorner, Peter Nickeas, Elvia Malagon & Alexandra Chachkevitch, Top Cop After Violent Holiday Weekend: ‘It’s Not a Police Issue, It’s a Society Issue,’ Chi. Trib. (Sept. 6, 2016, 10:54 pm), http://www.chicagotribune.com/news/local/breaking/ct-chicago-violence-labor-day-tuesday-20160906-story.html (on file with the Columbia Law Review).

I. Exploring the Tension Between Police and Communities

Violence and homicide are both community and police issues in every city; they are also political and managerial challenges for mayors and local government. Yet shared concerns can be sites of contestation, not collaboration. Consider Chicago, where homicides jumped from 423 in 2014 to 762 in 2016, 5 Amanda Wills, Sergio Hernandez & Marlena Baldacci, 762 Murders. 12 Months. 1 American City., CNN (Jan. 2, 2017), http://www.cnn.it/2ivRtDS [http://perma.cc/BSK8-6D2V]. while nonfatal shootings also spiked from 2,599 6 Jeremy Gorner, Chicago Ends 2014 with Fewer Homicides, but Shooting Victims Up 14 Percent, Chi. Trib. (Jan. 3, 2015, 9:50 am), http://www.chicagotribune.com/
news/ct-chicago-crime-year-end-met-20150101-story.html (on file with the Columbia Law Review).
to 3,550 7 Crimesider Staff, Chicago Saw More 2016 Murders than NYC, LA Combined, CBS News (Jan. 2, 2017, 1:45 pm), http://www.cbsnews.com/news/chicago-murders-shootings-2016-more-than-new-york-city-los-angeles-combined/ [http://perma.cc/29JK-MBJH]. during the same period. Chicago’s spike in shootings and homicides became a political question in the 2016 presidential election. 8 See, e.g., John Kass, Chicago Violence Permeates Presidential Election, Townhall (Sept. 30, 2016, 12:01 am), http://townhall.com/columnists/johnkass/2016/09/30/
chicago-violence-permeates-presidential-election-n2225196 [http://perma.cc/9QEQ-TNPN]. President Donald Trump recently threatened federal intervention unless Chicago’s city leaders can control its “horrible carnage.” See Theodore Schleifer & Joshua Berlinger, Trump Says He’ll Send in Feds if Chicago Doesn’t Fix ‘Horrible Carnage Going On,’ CNN, http://cnn.it/2j3amBE [http://perma.cc/UT5X-VDQG] (last updated Jan. 25, 2017, 5:07 pm).
The city’s turmoil over its devastating homicide and shooting rates, and reports of a police pullback in the midst of this crisis, 9 Max Kapustin et al., Univ. of Chi. Crime Lab, Gun Violence in Chicago, 2016, at 24 (2017), http://urbanlabs.uchicago.edu/attachments/store/2435a5d4658e2ca19f4f225
b810ce0dbdb9231cbdb8d702e784087469ee3/UChicagoCrimeLab+Gun+Violence+in+Chicago+2016.pdf [http://perma.cc/68UA-3QRP] (noting an 80% decrease in police stops “a few months before rates of gun violence in Chicago began to increase”).
are framed by deep community distrust of the police in the wake of fatal police shootings, 10 See Police Accountability Task Force, Recommendations for Reform: Restoring Trust Between the Chicago Police and the Communities They Serve 1 (2016), http://chicagopatf.org/wp-content/uploads/2016/04/PATF_Final_Report_4_13_16-1.pdf [http://perma.cc/ZM6K-TF8E] (explaining the purpose of the task force is to provide recommendations to help rebuild trust with policing infrastructure); see also U.S. Dep’t of Justice, Investigation of the Chicago Police Department 15 (2017) [hereinafter U.S. Dep’t of Justice, Investigation of the Chicago Police Department], http://www.justice.gov/
opa/file/925846/download [http://perma.cc/838M-A2P9] (“The pattern or practice of unreasonable force [which includes fatal shootings], coupled with the recurrence of unaddressed racially discriminatory conduct by officers further erodes community trust and police effectiveness.”); Rob Arthur & Jeff Asher, Gun Violence Spiked—and Arrests Declined—in Chicago Right After the Laquan McDonald Video Release, FiveThirtyEight (Apr. 11, 2016, 3:53 pm), http://fivethirtyeight.com/features/gun-violence-spiked-and-arrests-declined-in-chicago-right-after-the-laquan-mcdonald-video-release/ [http://perma.cc/57NC-KV4F]; Miles Bryan, Gun Deaths in Chicago Reach Startling Number as Year Closes, NPR News (Dec. 28, 2016, 4:37 am), http://www.npr.org/2016/
12/28/506505382/gun-deaths-in-chicago-reach-startling-number-as-year-closes [http://perma.cc/54SA-YA3v]; John Byrne, Emanuel Blames Chicago Crime Uptick on Officers Second-Guessing Themselves, Chi. Trib. (Oct. 13, 2015, 6:30 am), http://www.
chicagotribune.com/news/local/politics/ct-emanuel-fetal-police-met-20151012-story.html (on file with the Columbia Law Review).
as well as by strong resistance from police unions to com­munity and city efforts to strengthen civilian oversight of police and increase discipline of police officers who violate law and policy. 11 See, e.g., Dan Hinkel, Chicago Police Set New Advisory Guidelines on Discipline, but Language Criticized by DOJ Appears, Chi. Trib. (Feb. 1, 2017, 9:14 pm), http://www.chicagotribune.com/news/local/breaking/ct-chicago-police-discipline-matrix-met-20170201-story.html (on file with the Columbia Law Review); see also Jeremy Gorner, Nearly 2 Dozen Chicago Cops Disciplined for Faulty Dashboard Cameras, Chi. Trib. (Jan. 7, 2016, 6:56 am), http://www.chicagotribune.com/news/opinion/
editorials/ct-chicago-police-disciplined-met-20160106-story.html (on file with the Columbia Law Review).
For example, just a few days after August ended with sixty-eight homicides in the city, Chicago’s police union, the Chicago Fraternal Order of Police, urged its officers not to work overtime shifts on the 2016 Labor Day holiday weekend. 12 Jeremy Gorner, Chicago Police Union Calls for Officers Not to Work Overtime on Labor Day Weekend, Chi. Trib. (Aug. 31, 2016, 8:48 pm), http://www.chicagotribune.
com/news/local/breaking/ct-chicago-police-labor-day-overtime-fop-met-20160831-story.html (on file with the Columbia Law Review).
The timing of this pointed expression of police resistance and defiance in the midst of a crisis of violence suggests the depth of the social and political conflict between the community, the police, and the city government. And while it is hard to link this resistance to police conduct on the street, there was indeed evidence of a police pullback in the form of reduced street stops. 13 Kapustin et al., supra note 9, at 24.

Chicago is not alone in its crisis and in its conflicts between citizens and police. Baltimore was the scene of three days of demonstrations—sometimes violent—following Freddie Gray’s arrest, his death while in police custody, and his funeral. 14 Paul Schwartzman, At Gray’s Funeral, Outrage over the Way in Which He Died, Wash. Post (Apr. 27, 2015), http://wpo.st/UCZW2 [http://perma.cc/E3PZ-46AW]. Gray’s death occurred in 2015, a year when Baltimore’s 344 homicides set a modern record 15 Kevin Rector, National Homicide Spike in 2015, Reflected in New FBI Data, Remains Daily Reality in Baltimore, Balt. Sun (Sept. 26, 2016, 6:20 pm), http://www.
baltimoresun.com/news/maryland/crime/bs-md-homicide-data-20160926-story.html (on file with the Columbia Law Review).
before tapering off in 2016 to 318. That total made 2016 the second-deadliest year on record. 16 Kevin Rector, In 2016, Baltimore’s Second-Deadliest Year on Record, Bullets Claimed Targets and Bystanders Alike, Balt. Sun (Jan. 2, 2017, 6:10 am), http://www.
baltimoresun.com/news/maryland/crime/bs-md-ci-homicides-2016-20170102-story.html (on file with the Columbia Law Review).
Reports from Baltimore suggested a police pullback, 17 Richard A. Oppel Jr., West Baltimore’s Police Presence Drops, and Murders Soar, N.Y. Times (June 12, 2015), http://nyti.ms/2jBxJkv (on file with the Columbia Law Review). similar to Chicago, following the indictments of six officers who were participants in Gray’s arrest and eventual death in custody. 18 Justin Fenton, Freddie Gray Case: Judge Allows Malicious Prosecution Lawsuit Against Mosby to Proceed, Balt. Sun (Jan. 6, 2017, 7:03 pm), http://www.baltimoresun.
com/news/maryland/freddie-gray/bs-md-ci-mosby-lawsuit-to-proceed-20170106-story.htm (on file with the Columbia Law Review).
The rancor between police and city government peaked after the acquittal of the first four of the officers tried 19 Jess Bidgood, Another Baltimore Police Officer Acquitted in Freddie Gray Case, N.Y. Times (July 18, 2016), http://nyti.ms/2orZCfx (on file with the Columbia Law Review). and was followed by lawsuits by the officers against the State Attorney, alleging defamation and malicious prosecution among other claims. 20 Nero v. Mosby, No. MJB-16-1288, 2017 WL 386537, at *4 (D. Md. Jan. 27, 2017) (granting in part and denying in part the motion to dismiss plaintiffs’ claims). The demands by the police in Baltimore go well beyond civil damages: Officers and their union representatives have called for the disbarment of the elected former prosecutor, a sign of the depth of the enmity between police and government as well as between police and citizens. 21 Ryan M. McDermott, Marilyn Mosby Never Should Have Charged Officers in Freddie Gray Death, Lawyers Say, Wash. Times (July 27, 2016), http://www.
washingtontimes.com/news/2016/jul/27/marilyn-mosby-never-should-have-charged-officers-i/ [http://perma.cc/XA7M-FLG9].

A. Factors Producing Tension Between Communities and Police

Three factors drive tensions between citizens, rank-and-file police, and big-city mayors in these and other cities: police officers’ resentment of both political and community pushback, rank and file hostility to legal interventions and reform, and increased police killings and use of force, particularly in communities of color.

First, in the wake of events such as the Freddie Gray homicide or the revelations of the police-shooting death of Laquan McDonald in Chicago, 22 Nausheen Husain, Laquan McDonald Timeline: The Shooting, the Video and the Fallout, Chi. Trib. (Jan. 13, 2017), http://www.chicagotribune.com/news/
laquanmcdonald/ct-graphics-laquan-mcdonald-officers-fired-timeline-htmlstory.html (on file with the Columbia Law Review).
criticism of the police in cities across the country sparked acrimony among police officers—not only toward local politicians who were critical of police but also toward their own managers, who tended to join in that criticism. 23 See infra notes 31–46 and accompanying text (providing examples of such viewpoints); cf. Rich Morin et al., Pew Research Ctr., Behind the Badge: Amid Protests and Calls for Reform, How Police View Their Jobs, Key Issues and Recent Fatal Encounters Between Blacks and Police 63 (2017), http://assets.pewresearch.org/wp-content/
uploads/sites/3/2017/01/06171402/Police-Report_FINAL_web.pdf [http://perma.cc/
UD6S-JP2H] (observing differing levels of sympathy administrators and lower-ranking officers have toward protesters).
A Pew Research Center survey of police officers shows that both the protests against the police and an increase in the number of fatal attacks on police in recent years 24 Nat’l Law Enf’t Officers Mem’l Fund, Preliminary 2016 Law Enforcement Officer Fatalities Report 1 fig.1 (2016), http://www.nleomf.org/assets/pdfs/reports/Preliminary-2016-EOY-Officer-Fatalities-Report.pdf [http://perma.cc/VX7J-Y32J] (charting a recent upward trend in officer fatalities). have made them worry more about their personal safety. 25 See Morin et al., supra note 23, at 5 (“About nine-in-ten officers (93%) say their colleagues worry more about their personal safety . . . .”). Police officer fatalities rose from 123 in 2015 to 135 in 2016, a 10% increase. 26 See Nat’l Law Enf’t Officers Mem’l Fund, supra note 24, at 1. But this is a lower death toll than the 177 officers killed in 2011 and the 192 officers killed in 2007. 27 Officer Deaths by Year: Year-by-Year Breakdown of Law Enforcement Deaths Throughout U.S. History, Nat’l Law Enf’t Officers Mem’l Fund, http://www.nleomf.org/
facts/officer-fatalities-data/year.html (on file with the Columbia Law Review) (last updated Mar. 24, 2016).
Despite the decline in killings of police, a small number of recent premeditated ambushes likely have an outsized effect on police fears. 28 Christopher Ingraham, Ambush Killings of Police Officers Has Hit a 10-Year High, Wash. Post: Wonkblog (Nov. 21, 2016), http://wapo.st/2fyiBjk?tid=ss_tw-bottom [http://perma.cc/3KNG-53Z8]; Joe Mozingo, The Cases Might Be Rare, but These Are the Killers Cops Fear Most, L.A. Times (Oct. 12, 2016, 9:50 pm), http://www.latimes.com/
nation/la-na-cop-killers-snap-story.html (on file with the Columbia Law Review); see also George Fachner & Zoë Thorkildsen, Cmty. Oriented Policing Servs., Ambushes of Police: Environment, Incident Dynamics, and the Aftermath of Surprise Attacks Against Law Enforcement 71 (2015), http://ric-zai-inc.com/Publications/cops-p340-pub.pdf [http://perma.cc/83KK-7SH6] (finding violent crime and violence against the police correlate with ambushes against police at the jurisdictional level and progressive hiring practices were associated with lower numbers of ambushes).
The Pew survey also shows that the anecdotal reports of police pullbacks in Chicago and Baltimore reflect a widely shared view among police: 72% say officers in their department are now less willing to stop and question suspicious persons. 29 Morin et al., supra note 23, at 4. Overall, more than eight in ten (86%) say police work is harder today as a result of high-profile incidents of officer shootings of citizens and officer fatalities. 30 Id. at 15.

Second, beyond local politics and tensions, sentinel events in Chicago and Baltimore led to federal interventions that reified criticism of the police and increased police resentment of such interventions. Within a year of the trials of the officers in the Freddie Gray homicide and five months after the Chicago police union urged defiance over Labor Day weekend, both Baltimore 31 See Consent Decree, United States v. Police Dep’t of Balt., No. 1:17-cv-00099-JKB (D. Md. Jan. 12, 2017), http://www.justice.gov/crt/case-document/file/925036/
download [http://perma.cc/URB7-XLUK]; see also Camila Domonoske, Baltimore, DOJ Reach Agreement on Consent Decree for Baltimore Police, NPR (Jan. 12, 2017, 10:56 am), http://www.npr.org/sections/thetwo-way/2017/01/12/509479934/baltimore-doj-reach-agreement-on-consent-decree-for-baltimore-police [http://perma.cc/TAC3-FRMQ] (reporting Baltimore agreed to the consent decree to “institute sweeping reforms in Baltimore’s police department”).
and Chicago 32 Justice Department Announces Findings of Investigation into Chicago Police Department, U.S. Dep’t of Justice (Jan. 13, 2017), http://www.justice.gov/opa/pr/justice-department-announces-findings-investigation-chicago-police-department [http://perma.cc/9W5F-D4SC]. completed negotiations with the Department of Justice (DOJ) to develop and imple­ment consent decrees to undertake reforms in policy and practice—reforms that promise structural changes, remediation of constitutional violations, and healing of breaches between police and community. 33 The future implementation of these agreements is uncertain. In response to questions about the Chicago consent decree during his confirmation hearing for Attorney General, Senator Jeff Sessions refused to commit to honoring the agreement. See Senate Comm. on the Judiciary, Nomination of Jeff Sessions to Be Attorney General of the United States, Questions for the Record, Questions from Senator Durbin 4 (2017), http://www.judiciary.senate.gov/imo/media/doc/Sessions%20Responses%20to%20Durbin%20QFRs.pdf [http://perma.cc/L3HJ-NPFJ] (“[T]here are concerns with the impact of using consent decrees for policy purposes and that caution should be used in these cases.”). In Baltimore, the DOJ has sought a postponement of a hearing to finalize the details of the consent decree and delay implementation of the agreement to allow for time to brief the presidential administration on the proposed agreement. See Kevin Rector, Consent Decree Hearing Postponed to Give Justice Department Time to Brief Trump Administration on Agreement, Balt. Sun (Jan. 20, 2017, 6:52 pm), http://www.
baltimoresun.com/news/maryland/baltimore-city/bs-md-ci-consent-decree-delay-20170120-story.html (on file with the Columbia Law Review).
Chicago and Baltimore are two of twenty-five investigations that the Special Litigation Section of the Civil Rights Division (CRD) of the DOJ conducted under 42 U.S.C. § 14141 since 2009. 34 See Special Litigation Section Cases and Matters, U.S. Dep’t of Justice, http://www.justice.gov/crt/special-litigation-section-cases-and-matters0#police [http://perma.cc/V99M-FLK7] (last visited Feb. 4, 2017). The CRD opened and is currently enforcing nineteen agreements—including fourteen consent decrees and one postjudgment order—in counties and state agencies that are under federal court supervision to implement such agreements. 35 Id. Since the inception of “pattern and practice” interventions in the 1990s, a total of forty police agencies have entered into either stipulated settlements or consent decrees, committing local police to a series of court-supervised structural and policy reforms. 36 See Civil Rights Div., U.S. Dep’t of Justice, The Civil Rights Division’s Pattern and Practice Police Reform Work: 1994–Present 3 (2017), http://www.justice.gov/
crt/file/922421/download [http://perma.cc/DPT3-TDBJ]; see also Goldie Taylor, More than 20 U.S. Cities Are Currently Under a DOJ Consent Decree, but Do They Really Work?, Blue Nation Rev. (May 27, 2015), http://archives.bluenationreview.com/more-than-20-u-s-cities-are-currently-under-a-doj-consent-decree-but-do-they-really-work/ [http://perma.cc/WZ99-RX4S].
These agreements between local entities and the DOJ require extensive changes in policy and practice—and in some cases structural changes in functions including recruitment and discipline—and a cessation of the consti­tutional violations that precipitated the interventions. 37 See Rachel A. Harmon, Promoting Civil Rights Through Proactive Policing Reform, 62 Stan. L. Rev. 1, 12 (2009) (noting the 1992 “[Rodney] King incident highlighted the inadequacy of existing solutions for achieving [police] reform”); Stephen Rushin, Federal Enforcement of Police Reform, 82 Fordham L. Rev. 3189, 3244–47 (2014) (listing DOJ structural police reform investigations and negotiated settlements between DOJ and police departments); see also Samuel Walker, The New Paradigm of Police Accountability: The U.S. Justice Department “Pattern or Practice” Suits in Context, 22 St. Louis U. Pub. L. Rev. 3, 13 (2003) (expounding on the role of litigation and the civil rights movement in bringing on police reform). See generally Debra Livingston, Police Reform and the Department of Justice: An Essay on Accountability, 2 Buff. Crim. L. Rev. 815, 850 (1999) (discussing the mechanisms of 42 U.S.C. § 14141(a)–(b) (1994)).

Police officers often view DOJ interventions and other efforts aiming to enhance police accountability and constitutional compliance as a condemnation of their police work. 38 Morin et al., supra note 23, at 5 (reporting police officers view protests as motivated by “anti-police bias”). These reform efforts can provoke resentment among the rank and file (and union leadership) and some bitterness at the labeling of their actions as extralegal or tinged with racism. 39 Simone Weichselbaum, The Problems with Policing the Police, TIME, http://time.com/police-shootings-justice-department-civil-rights-investigations/ [http://perma.cc/U46M-VWZX] (last visited Mar. 5, 2017) (describing efforts by local forces to “fight back” against DOJ action); see also Morin et al., supra note 23, at 81–83 (reporting that police officers blame anti-police bias and “long-standing animosity” toward police for recent protests). In some places, police have pushed back. Officers in Seattle sought court intervention to stop the implementation of new use of force guidelines that were created through the 2012 DOJ consent decree. 40 Mahoney v. Holder, 62 F. Supp. 3d 1215, 1217–18 (W.D. Wash. 2014). In Boston, no officers answered the Commissioner’s call for volunteers to conduct a pilot program to test the use of body-worn cameras. 41 Denise Lavoie, Volunteers Lacking for Boston Police Body Camera Program, WBUR News (Aug. 14, 2016), http://www.wbur.org/news/2016/08/14/boston-police-body-camera-volunteers [http://perma.cc/27CQ-M9Z9]. When the Commissioner ordered 100 officers to participate, 42 Zeninjor Enwemeka, Ahead of Ruling on Pilot Program, Boston Police Commanders Say They Will Wear Body Cameras, WBUR News (Sept. 8, 2016), http://www.wbur.org/news/2016/09/08/boston-police-command-staff-body-cameras [http://perma.cc/9DMJ-TVUF]. the Boston Police Patrolmen’s Association (BPPA) tried to halt the program (unsuccessfully). 43 Bos. Police Patrolmen’s Ass’n v. City of Boston, No. 16-2670-B, at 1 (Mass. Super. Ct. Sept. 9, 2016), http://d279m997dpfwgl.cloudfront.net/wp/2016/09/
20160912093831348.pdf [http://perma.cc/HJ5V-7Q4T]. The court denied the motion for injunctive relief, citing to the binding obligations of a Memorandum of Understanding between the BPPA and Boston to implement a voluntary Body Worn Camera experiment. Id. at 19. The court also rejected BPPA’s collective bargaining argument. Id. at 14.

In New York City, police unions asserted that court-ordered mea­sures aimed at accountability violate their collective bargaining rights. They used this argument in an unsuccessful motion to intervene and overturn the settlement and order resulting from Floyd v. City of New York, a 2013 opinion finding Fourth and Fourteenth Amendment violations in the New York City Police Department’s (NYPD’s) “stop-and-frisk” pro­gram. 44 See 770 F.3d 1051, 1055–56 (2d Cir. 2014) (providing background on police unions’ motions to intervene). See generally Floyd v. City of New York (Floyd I), 959 F. Supp. 2d 540, 562 (S.D.N.Y. 2013) (finding the city liable for violating the Fourth and Fourteenth Amendments through its stop-and-frisk program), aff’d, 770 F.3d 1051; Tracey L. Meares, Programming Errors: Understanding the Constitutionality of Stop-and-Frisk as a Program, Not an Incident, 82 U. Chi. L. Rev. 159, 162–63 (2015) [hereinafter Meares, Programming Errors] (explaining stop-and-frisk policies are best understood as top-down programs). The Second Circuit denied the intervenor motion, citing the city’s right to voluntarily dismiss its own appeal of the case while also reinforcing the boundaries of management prerogatives in the collective bargaining issues. 45 Floyd, 770 F.3d at 1058–61. Indeed, the panel expressed “serious reservations about the prospect of allowing a public-sector union to encroach upon a duly-elected government’s discretion to settle a dispute against it.” 46 Id. at 1060. These episodes of police resistance—if not defiance—reveal tensions not just between communities and police but also between police and elected officials in the executive branches over police discretion in everyday tactical decisions and, more generally, over the autonomy of police to craft strategies for controlling crime.

Third, the connection between race, aggressive police tactics, and police killings has widened the breach in trust between citizens and police—especially Black citizens and police. 47 See Carl Bialik, Why Are So Many Black Americans Killed by Police?, FiveThirtyEight (July 21, 2016, 1:24 pm), http://fivethirtyeight.com/features/why-are-so-many-black-americans-killed-by-police/ [http://perma.cc/T567-NAB9] (exploring why “Black Americans are more than twice as likely as white Americans to be killed by police officers”). Police killings, often highly publicized and increasingly visible on citizen or police videos, have sparked intense anger in Black communities. 48 Daniel Funke & Tina Susman, From Ferguson to Baton Rouge: Deaths of Black Men and Women at the Hands of Police, L.A. Times (July 12, 2016, 3:45 pm), http://www.latimes.com/nation/la-na-police-deaths-20160707-snap-htmlstory.html (on file with the Columbia Law Review) (detailing encounters of Black men and women who died following encounters with police between 1999 and 2016, noting the existence of recordings in some cases and describing community responses); see also Carol E. Lee, Obama Says Police Killings of Two Black Men Should Trouble All Americans, Wall. St. J. (July 7, 2016, 8:47 pm), http://www.wsj.com/articles/two-police-killings-of-black-men-part-of-pattern-of-racial-disparity-obama-says-1467919272 (on file with the Columbia Law Review) (recounting President Obama’s statements highlighting the feelings of injustice that communities of color have felt following police shootings). The Washington Post counted 963 persons shot and killed by police in 2016, a 5% decline from the 991 killings in 2015. 49 Fatal Force, Wash. Post, http://wapo.st/police-shootings-2016 [http://perma.cc/
Y9U6-BNNX] (last visited Feb. 4, 2017). Comprehensive and reliable data on nonfatal police shootings are nonexistent, although some cities maintain databases that track police shootings. See, e.g., BPD Officer Involved Use of Force, Open Balt. (Feb. 2, 2017), http://data.baltimorecity.gov/Public-Safety/BPD-Officer-Involved-Use-Of-Force/3w4d-kckv [http://perma.cc/QBJ3-LF8J].
Other crowd-sourced websites provide slightly different counts, which differ from the accounts of newspapers in small ways owing to differences in inclusion rules, the availability of details of the incidents, and the time windows for each data source. 50 See Franklin E. Zimring, How Many Killings by Police, 2016 U. Chi. Legal F. 691, 701–04 [hereinafter Zimring, How Many Killings] (explaining the shortcomings of various reporting methodologies); Joscha Legewie & Jeffrey Fagan, Group Threat, Police Officer Diversity and the Deadly Use of Police Force 61 tbl.A1 (Columbia Law Sch. Pub. Law & Legal Theory Working Paper Grp., Working Paper No. 14-512, 2016), http://ssrn.com/
abstract=2778692 (on file with the Columbia Law Review) (comparing data sources on police killings). For example, the website http://www.killedbypolice.net/ reported that 1,151 people were killed by police in 2016, nearly 200 more deaths than were reported by the Washington Post for the same year. See Fatal Force, supra note 49; Killed by Police 2017, Killedbypolice.net, http://www.killedbypolice.net/ [http://perma.cc/GR7T-UQEN] (last visited Feb. 4, 2017). This website reported 776 police killings of citizens in 2013 and 1,111 in the following year. Killed by Police 2013, Killedbypolice.net, http://www.killedbypolice.
net/kbp2013.html [http://perma.cc/5UG2-HTYL] (last visited Feb. 26, 2017); Killed by Police 2014, Killedbypolice.net, http://www.killedbypolice.net/kbp2014.html [http://
perma.cc/RKH7-FCS2] (last visited Feb. 26, 2017). The one-year disparity no doubt reflects increased attention to police killings and more detailed reporting on incidents. See, e.g., Eliot C. McLaughlin, We’re Not Seeing More Police Shootings, Just More News Coverage, CNN (Apr. 21, 2015), http://cnn.it/1GbjQN9 [http://perma.cc/AAG9-27QG].
However, all agree that official government records, which rely on voluntary reporting by police agencies, substantially undercount police shootings. 51 See Zimring, How Many Killings, supra note 50, at 703 (asserting media reporting of police killings is more comprehensive than police data that “report only the police department’s version of events”); see also Franklin E. Zimring & Brittany Arsiniega, Trends in Killings of and by Police: A Preliminary Analysis, 13 Ohio St. J. Crim. L. 247, 251 (2015) (“[T]he published count of killings by police is incomplete, unaudited, and varies between sources; estimates say that these killings are underreported by as much as 20%.”). For an in-depth discussion of the limitations of the federal data on police killings, see Daniel Bier, By The Numbers: How Often Do Cops Kill?, Skeptical Libertarian (July 13, 2016), http://blog.skepticallibertarian.com/2016/07/13/by-the-numbers-how-often-do-cops-kill/ [http://perma.cc/WXT7-VP99]. In this graph, Daniel Bier compares crowd-sourced datasets from the Guardian, the Washington Post, and the website “Fatal Encounters” with statistics from the Bureau of Justice Statics and the FBI. Id.

Figure 1: Police Justifiable Homicides and Police Officers Killed in the Line of Duty, 2000 to 2015


Figure 1 52 . For data on police justifiable homicides see, FBI, Expanded Homicide Data Table 14, 2006–2010, FBI: Unif. Crime Reporting, http://ucr.fbi.gov/crime-in-the-u.s/2010/crime-in-the-u.s.-2010/tables/10shrtbl14.xls [http://perma.cc/67BG-N8VY] (last visited Mar. 26, 2017); FBI, Expanded Homicide Data Table 14, 2011–2015, FBI: Unif. Crime Reporting, http://ucr.fbi.gov/crime-in-the-u.s/2015/crime-in-the-u.s.-2015/tables/expanded_homicide_data_table_14_justifiable_homicide_by_weapon_law_enforcement_2011-2015.xls [http://perma.cc/749U-ZQXT] (last visited Mar. 26, 2017); FBI, Table 2.16–Justifiable Homicide by Weapon, LEOs, 1996–2000, FBI: Unif. Crime Reporting, http://ucr.fbi.gov/crime-in-the-u.s/2000 (on file with the Columbia Law Review) (last visited Mar. 26, 2017); FBI, Uniform Crime Reporting Program (UCR) Expanded Homicide Data Table 13, 2001–2006, FBI: Unif. Crime Reporting, http://ucr.fbi.gov/crime-in-the-u.s/2005 [http://perma.cc/YKU5-ANUN] (on file with the Columbia Law Review) (last visited Mar. 26, 2017). For data on police officers killed in the line of duty, see FBI, Table 1: Law Enforcement Officers Feloniously Killed, 2000–2009, FBI: Unif. Crime Reporting, http://ucr.fbi.gov/leoka/2009 [http://perma.cc/VF27-CY6E] (last visited Mar. 26, 2017); FBI, Table 1: Law Enforcement Officers Feloniously Killed, 2006–2015, FBI: Unif. Crime Reporting, http://ucr.fbi.gov/leoka/2015/
tables/table_1_leos_fk_region_geographic_division_and_state_2006-2015.xls [http://
perma.cc/ZGL9-K5RA] (last visited Mar. 26, 2017).
shows the trends since 2000 in police killings of civilians and killings of police in the line of duty. If both of these types of killings are a source of tension and conflict between police and communities, then there is an asymmetry in the reactions to them. Figure 1 shows justifiable homicides by police in order to provide a longer trend line in police killings. These are widely regarded as an undercount of police killings, including both justifiable homicides and other police killings of citizens. 53 See Zimring & Arsiniega, supra note 51, at 247 (describing problems with accuracy of data regarding killings by police officers). The trend lines in Figure 1 are moving in different directions. Killings of police vary from year to year, but the overall trend line shows a decline since 2000. 54 See supra Figure 1. However, killings by police have been on the rise since 2000. 55 See supra Figure 1. The increase in police killings of citizens, the spread of citizen and police videos of these killings, and the publicity these videos receive in the media are flashpoints of anger and resentment, partic­ularly in minority communities. The emergence of active advocacy groups and widespread protests have increased attention to police killings and widened the breach between citizens and police in minority communities. 56 See McLaughlin, supra note 50 (describing protests against police killings); see also About the Black Lives Matter Network, Black Lives Matter, http://blacklivesmatter.com/about (on file with the Columbia Law Review) (last visited Feb. 27, 2017) (describing the mission and work of the Black Lives Matter movement).

It is hard to reconcile the downward trend in killings of police—and of homicides nationally 57 See infra Figure 4. and in most cities 58 See infra Figure 3. —with the rise in killings by police. This disconnect is not lost on advocacy groups such as Black Lives Matter, and it heightens distrust of police. For their part, police responses to these trends seem disproportionate, although ambush killings may play an outsized role in their fears.

The available data on police shootings of civilians show connections between police killings and race—trends that confirm what is common knowledge for many Black citizens and communities. 59 See, e.g., Gregory B. Hladky, Lawmakers: Cases of Police Bias, Abuse Create Deep Distrust, Hartford Courant (Apr. 2, 2015, 6:45 pm), http://www.courant.com/politics/hc-police-minority-community-distrust-20150402-story.html (on file with the Columbia Law Review) (reporting lawmakers’ descriptions of “the distrust and fear of police within Connecticut’s black and Hispanic communities and their doubts about the pace of reform”); see also President’s Task Force on 21st Century Policing, Final Report of the President’s Task Force on 21st Century Policing 9–11 (2015), http://cops.usdoj.gov/
pdf/taskforce/TaskForce_FinalReport.pdf [http://perma.cc/L5WG-MT4A] (discussing the discrepancy between White and non-White communities’ trust of the police).
Professors Joscha Legewie and Jeffrey Fagan suggest that police killings reflect a broader context of racial conflict and racial threat in cities, and show that police killings of Blacks from 2014 to 2016 were a function more of racial disparities in population than racial metrics of crime or violence. 60 Legewie & Fagan, supra note 50, at 34 (describing how the study’s data “reveal a race-specific pattern in the ways in which the police respond to crime”). Other research, as well as Legewie and Fagan’s study, demonstrates that the absence of racial diversity among police contributes to racial gaps in police killings of Black citizens. 61 Id. at 34–35 (finding “evidence that a high proportion of black officers might not necessarily reduce the number of officer-involved killings directly, but mitigates the role of group threat and thereby eases tensions between the police and African-American communities”); see also Sean Nicholson-Crotty, Jill Nicholson-Crotty & Sergio Fernandez, Will More Black Cops Matter? Officer Race and Police-Involved Homicides of Black Citizens, 77 Pub. Admin. Rev. (forthcoming 2017) (manuscript at 22) (on file with the Columbia Law Review) (finding that increasing the percentage of Black officers will likely not decrease violence unless the police department has reached a “critical mass” of Black officers). Using data from 2005 to 2012 in sixteen states, a research team at the Harvard School of Public Health found stark racial and ethnic disparities of eight to one in police killings. 62 Catherine Barber et al., Homicides by Police: Comparing Counts from the National Violent Death Reporting System, Vital Statistics, and Supplementary Homicide Reports, 106 Am. J. Pub. Health 922, 924 (2016) (citing the rate of “legal intervention homicides” as 0.48 per 100,000 population for Blacks compared to 0.06 for Asians). Another study compared police killings of citizens to police contacts through street stops, traffic stops, and arrests and showed that racial disparities in police killings in 2012 were explained by large racial differences in exposure to police and rates of police contact. 63 Ted R. Miller et al., Perils of Police Action: A Cautionary Tale from Police Data Sets, 23 Inj. Prevention 27, 29 (2017) [hereinafter Miller et al., Perils of Police Action] (finding the ratio of people law enforcement killed or medically treated for injury by law enforcement is “surprisingly consistent by race/ethnicity” and “[b]lacks have high arrest and stop rates, and per capita are much more likely than whites to die at the hands of police” (citations omitted)).

Research by Professor Roland Fryer examining police use of force in Houston, one of the nation’s largest cities, shows a nearly 50% greater incidence of use of force by police in encounters with Black or Latino persons but no disparity by race in shootings by police. 64 Roland G. Fryer, Jr., An Empirical Analysis of Racial Differences in Police Use of Force 3–7 (Nat’l Bureau of Econ. Research, Working Paper No. 22399, 2016), www.nber.org/papers/w22399.pdf (on file with the Columbia Law Review). Justin Feldman’s subsequent analyses of Professor Fryer’s Houston results showed that, in fact, Blacks were nearly five times more likely to be shot relative to Whites and Latinos were nearly twice as likely to be shot relative to Whites. 65 Justin Feldman, Roland Fryer Is Wrong: There Is Racial Bias in Shootings by Police, Scholars at Harv.: Justin Feldman Blog (July 12, 2016), http://scholar.harvard.
edu/jfeldman/blog/roland-fryer-wrong-there-racial-bias-shootings-police [http://
Professor Fryer searched for evidence of racial bias in police shootings in Houston, using statistical models to identify intentional bias. He found none. Feldman’s analyses of the same data examined statistical discrim­ination 66 Statistical discrimination reflects differences in the rates of an event by race, after controlling for race-specific and plausible nonrace factors that might explain such differences. Racial bias looks for evidence of intent to discriminate, independent of evidence of racial disparities. See id. (explaining the difference between statistical discrimination and racial bias in the context of discriminatory policing). —or disparate treatment of Black and Latino suspects by police in their use of force—and showed large racial disparities. 67 Id. (finding “black people were over 5 times as likely to be shot relative to whites [and] [l]atinos were roughly twice as likely to be shot versus whites”). Overall, the evidence of racially disparate police enforcement across cities reinforces longstanding beliefs among Black citizens about disparate treatment at the hands of the police and helps spread a narrative of an uneven bur­den that Black citizens bear in police–citizen encounters. 68 Charles R. Epp, Steven Maynard-Moody & Donald P. Haider-Markel, Pulled Over: How Police Stops Define Race and Citizenship 9–13 (2014) (discussing the development and institutionalization of race-based policing practices); see also Rod K. Brunson, “Police Don’t Like Black People”: African-American Young Men’s Accumulated Police Experiences, 6 Criminology & Pub. Pol’y 71, 95 (2007) (“The combination of frequent involuntary police contact, coupled with what study participants considered poor treatment during such encounters, contributed to an accumulated body of unfavorable experiences that collectively shaped young men’s views of police.”); Rod K. Brunson & Ronald Weitzer, Police Relations with Black and White Youths in Different Urban Neighborhoods, 44 Urb. Aff. Rev. 858, 864–65 (2009) (observing “racial and neighborhood differences” between Black and White neighborhoods such that White youth “had a less troubled relationship with and more positive views of the police than Black youth” and “police treatment of residents appeared to be less problematic in the White neighborhood”); Tom R. Tyler, Policing in Black and White: Ethnic Group Differences in Trust and Confidence in the Police, 8 Police Q. 322, 323–24 (2005) (examining racial discrepancies in community trust in the police); Patricia Y. Warren, The Continuing Significance of Race: An Analysis Across Two Levels of Policing, 91 Soc. Sci. Q. 1025, 1026 (2010) (“Perceptions of racial profiling are used to illustrate the link between more widespread racialized stories about police behavior and individual attitudes toward police.”); Melissa Healy, Blacks Are More Likely to Be Killed by Police, but That’s Because They’re More Likely to Be Stopped, Study Says, L.A. Times (July 25, 2016, 3:00 pm), http://www.latimes.com/science/sciencenow/la-sci-sn-cops-race-injury-20160725-snap-story.html (on file with the Columbia Law Review) (examining the Miller et al. study, see Miller et al., Perils of Police Action, supra note 63).

B. A Vortex of Forces

The swirling factors addressed above—spiking homicide rates in some cities, police distrust of citizens generally, the protests of Black citizens in particular, police resistance to court oversight to remedy constitutional violations, an expressed reluctance by police to engage with crime suspects, police shootings and killings of Black citizens, distrust of police, and withdrawal from police–citizen cooperation in communities hardest hit by violence—describe a vortex of forces that can potentially form a political crisis. The depth of the gap in trust between citizens and police is alarming and seemingly intractable.

Even as shootings by police have led to widespread anger and political activism among minority citizens, shootings of police have stirred up deep anger and feelings of alienation among police. Officers are skeptical of the motivations and goals of the citizen protests: Far more officers attribute the protests to anti-police bias than to sincere concerns about police accountability. 69 Morin et al., supra note 23, at 5 (“Two-thirds of officers (68%) say the demonstrations are motivated to a great extent by anti-police bias; only 10% in a separate question say protesters are similarly motivated by a genuine desire to hold police accountable for their actions.”). Most police see no systemic factors in their profession that lead to police killings. 70 Id. (“Some two-thirds [of police officers surveyed] characterize the fatal encounters that prompted the demonstrations as isolated incidents and not signs of broader problems between police and the black community––a view that stands in sharp contrast with the assessment of the general public.”). This clashes with data from general population surveys in which the majority of respondents view police shootings and killings of citizens as symptomatic of deeper problems in that profession. 71 Id. (“In a separate Pew Research Center survey of U.S. adults, 60% say these incidents are symptoms of a deeper problem.”). And there are differences by race in these perceptions not only among the general public but also within police departments between Black and White officers. These different perceptions are not only on policing matters but also on the contexts of racial inequality broadly in American life. 72 See id. at 6–7 (“[V]irtually all white officers (92%) but only 29% of their black colleagues say that the country has made the changes needed to assure equal rights for blacks.”). That police would pull back from their duties 73 This process is often referred to as “de-policing.” in the face of criticism and serious public safety concerns in several cities highlights the deep cynicism and despair animating this crisis. 74 One commentator, Radley Balko, views police self-reporting of a pullback to be a petulant response that “insults the profession generally as well as those officers who continue to perform their jobs at high levels of professionalism . . . . [Balko] imagine[s] that doctors, soldiers, lawyers and members of just about every other profession would feel the same way.” Radley Balko, Why Are Police Groups and Their Advocates Advancing a Theory that Makes Police Officers Look Terrible?, Wash. Post (Jan. 23, 2017), http://wpo.st/It1X2 [http://perma.cc/ST2A-6RK7]. Despite police officers’ rational fear of ambush, their views do not reflect the actual risk of job-related injuries—police are not represented amongst the industry sectors at the highest risk for fatal workplace injuries. Cf. U.S. Bureau of Labor Statistics, Current Population Survey, Census of Fatal Occupational Injuries, Fatal Work Injuries by Adjusted Industry 14 (2015), http://www.bls.gov/iif/oshwc/cfoi/cfch0014.pdf [http://perma.cc/5YNM-N89P] (“Private construction had the highest count of fatal injuries in 2015, but the private agriculture, forestry, fishing and hunting sector had the highest fatal work injury rate.”). But we also could consider their actions to be a model of a rational, not petulant, actor: the possibility that employees—such as police officers—with an underspecified job description and a high degree of discretion might well prefer inaction for which they are unlikely to be blamed as opposed to action for which they might be held responsible.

These intersecting conflicts are the subject of this Mini-Symposium. Our purpose is to identify social, historical, and institutional contexts to understand the current set of crises. In this Essay, we create a historical context of episodic periods of violence epidemics over the past fifty years and provide perspective on previous eras to inform thinking about the present. We also discuss the dynamics of the cynicism and distrust that have created the current crisis of legitimacy between citizens and police and describe the connections between the separate sources of tension. We avoid proscriptions but urge exploration of the dynamics underlying these tensions as a first step to remediate distrust and build cooperation as a pathway to public safety.

Two additional essays contribute to the Mini-Symposium. Professor John Donohue explores the causal influences on crime in 2015 and points to declining prison populations, declining police numbers, and growing markets for illegal drugs as promising candidates. 75 2015 saw a large jump in the murder rate, even as that jump was from historically low levels and as property crime remained quite low. See John J. Donohue, Comey, Trump, and the Puzzling Pattern of Crime in 2015 and Beyond, 117 Colum. L. Rev. 1297, 1321 fig.8 (2017). He goes on to suggest that de-policing, to the extent that it is occurring, does not explain the rising homicide rates in cities nationwide. Professor Tracey Meares suggests that even as we try to understand troubling increases in violent crime in a number of cities, we need to promote the building of trust between the police and the communities they serve. 76 Tracey L. Meares, The Path Forward: Improving the Dynamics of Community–Police Relationships to Achieve Effective Law Enforcement Policies, 117 Colum. L. Rev. 1355 (2017) [hereinafter Meares, The Path Forward]. We are also grateful for the encouragement of FBI Director James Comey, who has helped put many of the issues discussed here in the national spotlight and has pushed for the analysis that, we hope, this symposium contributes to. 77 James B. Comey, Introduction, 117 Colum. L. Rev. 1233 (2017) [hereinafter Comey, Introduction].

II. Understanding the Trends

The recent increases in crime and violence create several challenges for police executives, community leaders, and governance. Crime is complicated and quite local in its impacts; it is city-specific and concentrated in particular neighborhoods within cities. 78 Ruth D. Peterson & Lauren J. Krivo, Divergent Social Worlds: Neighborhood Crime and the Racial–Spatial Divide 7–8 (2010) (describing the importance of neighborhood and area-specific crime studies, as opposed to general city-wide crime studies). While one city may be shaken by a homicide spike, its overall number and rate of homicides may be quite small compared to those of others. Explanations to date have sought to link the recent rise in homicides to a “Ferguson Effect”: Police retreat from “proactive policing” in the wake of height­ened scrutiny of incidents of excessive force and lethal police violence by local residents and the press. 79 James B. Comey, Dir., FBI, Remarks Delivered at the University of Chicago Law School: Law Enforcement and the Communities We Serve: Bending the Lines Toward Safety and Justice (Oct. 23, 2015), http://www.fbi.gov/news/speeches/law-enforcement-and-the-communities-we-serve-bending-the-lines-toward-safety-and-justice [http://
perma.cc/KYL2-JW8Z] [hereinafter Comey, Remarks at University of Chicago]; Heather Mac Donald, The Ferguson Effect, Wash. Post: Volokh Conspiracy (July 20, 2016), http://wpo.st/JkPX2 [http://perma.cc/FG7B-UNJB]; see also Richard Rosenfeld, Nat’l Inst. of Justice, Documenting and Explaining the 2015 Homicide Rise: Research Directions 18–23 (2016), http://www.ncjrs.gov/pdffiles1/nij/249895.pdf [http://
perma.cc/VE6P-74D9] (acknowledging several different potential definitions of the “Ferguson Effect” and exploring each in depth).
However, this provocative claim of a strong causal effect lacks systematic evidence and a careful counter­factual analysis. 80 Neil Gross, Is There a ‘Ferguson Effect’?, N.Y. Times (Sept. 30, 2016), http://nyti.ms/2jKMOAh (on file with the Columbia Law Review) (discussing the “Ferguson Effect” and its origins and highlighting the emerging scholarly empirical literature); Eric Lichtblau, FBI Director Says ‘Viral Video Effect’ Blunts Police Work, N.Y. Times (May 11, 2016), http://nyti.ms/2jKKnO9 (on file with the Columbia Law Review); Ames Grawert & James Cullen, What the Data Tell Us About Crime and the ‘Ferguson Effect,’ Brennan Ctr. for Justice Blog (Mar. 7, 2016), http://www.brennancenter.org/
blog/what-data-tell-us-about-crime-and-%E2%80%98ferguson-effect%E2%80%99 [http://perma.cc/94CW-6XJE] (describing and defining the “Ferguson Effect” but drawing attention to a study that did not find the Effect to be causing an increase in crime). For an alternate explanation, see Matthew Desmond & Andrew V. Papachristos, Why Don’t You Just Call the Cops?, N.Y. Times (Sept. 30, 2016), http://nyti.ms/2jKCGY2 (on file with the Columbia Law Review) (proposing increases in crime result from members of the public being less likely to call the police for help in the wake of highly publicized incidents of police brutality rather than police officers being less active as a result of public scrutiny as proponents of the “Ferguson Effect” suggest).
At the same time, other descriptions have blurred our understanding of the rise in homicides by placing homicide in a heterogeneous category of violent crime, which may mask the sharp edges of the homicide spikes in some cities and the issues that the spikes raise concerning neighborhoods and policing and local governing. 81 Matthew Friedman, Ames C. Grawert & James Cullen, Brennan Ctr. for Justice, Crime in 2016: A Preliminary Analysis 11 (2016), http://www.brennancenter.org/sites/
default/files/publications/Crime_2016_Preliminary_Analysis.pdf [http://perma.cc/
PM7Q-98WQ] (comparing change in crime rates and murder rates and noting trends that “indicate[] there may be a problem with murder in . . . specific cities, not that there is a national trend of rising crime”).
Because the evidence is scant on these causal claims, we need a richer and more disciplined understanding to know exactly what is occurring.

Developing this richer understanding requires a more granular perspective of crime changes in communities. The first step is to measure and describe recent trends in ways that are helpful to each stakeholder group and to provide a basis of evidence from which cities can shape their responses.

The second step is to give tempered and realistic perspectives on the details of the changes in crime and violence, a task that Professor Donohue has begun. 82 See Donohue, supra note 75. What is not known—and it is an essential question—is whether the increases of the past two years signal the beginning of a longer-term trend in violent crime or whether the current increase in a few cities is a temporary disruption in the longer-term trends. If there is a longer-term trend in the works, it is essential to identify which cities are in the midst of it and which might realize an increase at some point in the future. These questions apply as much across neighborhoods within cities as they do across cities.

The third step is to understand the contexts and circumstances within cities that explain why city-specific homicide and other violent crime rates have remained stable or increased. These are processes and social structures that are at work at the city level or across larger entities such as counties and regions. Police officials, local political actors, and neighborhood organizations and city planners recognize that homicides tend to cluster in a small number of neighborhoods and that those troubled places are likely to remain among the most troubled areas in their cities over time. Researchers have called these places “poverty traps” 83 Samuel Bowels, Steven N. Durlauf & Karla Hoff, Poverty Traps 2 (2011); Robert J. Sampson & Jeffery D. Morenoff, Durable Inequality: Spatial Dynamics, Social Processes, and the Persistence of Poverty in Chicago Neighborhoods, in Poverty Traps, supra, at 176, 201 (2011) (demonstrating “communities displaying elevated levels of legal cynicism and lower levels of collective efficacy were those that had been exposed to poverty in 1970 and increases in poverty from 1970 to 1990”). and use this concept to explain not only why there are outbreaks of violence in specific places but also why these are places where the dynamics of delegitimation and de-policing take place in tandem and why citizen withdrawal has taken place contemporaneously with police withdrawal. 84 See, e.g., David S. Kirk & Andrew V. Papachristos, Cultural Mechanisms and the Persistence of Neighborhood Violence, 116 Am. J. Soc. 1190, 1228 (2011) (finding “cultural frames have a constraining influence” and “cynicism constrains choice if individuals presume that the law is unavailable or unresponsive to their needs, thus pushing individuals to engage in their own brand of social control” such as violence).

A. The Past Two Years

Across the nation’s thirty largest cities, crime is as likely to have increased in the past two years as it is to have remained stable or declined. Table 1 shows the changes in overall crime and violent crime for these cities in descending order of population for the two-year period ending in late December 2016. Averaging across the thirty cities, the rate of violent crime increased by more than four times the rate of increase for overall crime. Eight cities reported declines in violent crime rates, and twelve cities reported declines in overall crime rates. Nine of the thirty cities did not provide data.

Table 1: Crime in the Thirty Largest U.S. Cities, 2015 to 2016 85 The data for this table was compiled by the Brennan Center for Justice using crime data obtained directly from police departments in these cities. See Friedman et al., supra note 81, at 4 tbl.1, 15 (providing data and explaining the methodology).

2015 2016 2015 2016
Crime Rate per 100,000 Crime Rate per 100,000 % Change in Crime Rate Violent Crime Rate per 100,000 Violent Crime Rate per 100,000 % Change in Violent Crime Rate
New York City 1,245.6 1,266.3 1.70% 470.2 469.7 -0.10%
Los Angeles 2,900.9 3,052.3 5.20% 575.4 652.1 13.30%
Chicago 3,947.6 4,305.7 9.10% 996 1,157.60 16.20%
Houston 5,332.2 5,337.9 0.10% 924.8 1,008.50 9.10%
Philadelphia 3,878.2 3,629.2 -6.40% 938.5 884.4 -5.80%
San Antonio 5,521.4 6,805.5 23.30% 513.3 783.1 52.50%
San Diego 2,446.0 2,137.8 -12.60% 359 283.5 -21.00%
Dallas 4,765.7 4,563.6 -4.20% 742.4 817.9 10.20%
San Jose 2,678.8 2,460.5 -8.10% 288.8 316.7 9.70%
Austin 4,028.2 3,712.7 -7.80% 319.4 351.2 10.00%
Jacksonville NA NA NA NA NA NA
San Francisco 8,856.9 7,719.0 -12.80% 2,056.20 1,942.10 -5.50%
Indianapolis NA NA NA NA NA NA
Columbus NA NA NA NA NA NA
Fort Worth NA NA NA NA NA NA
Charlotte 4,753.8 5,587.0 17.50% 702.7 860.6 22.50%
Seattle 3,797.4 3,578.1 -5.80% 528.1 516.8 -2.10%
Denver 4,127.8 4,287.8 3.90% 585.4 589.2 0.60%
Detroit 6,633.9 6,405.4 -3.40% 1,865.90 1,734.40 -7.00%
Washington, D.C. 5,449.1 5,084.4 -6.70% 896.4 866.9 -3.30%
Boston 3,009.6 2,856.5 -5.10% 684.3 639.9 -6.50%
Nashville 3,905.5 3,850.0 -1.40% 785.2 795.7 1.30%
Portland NA NA NA NA NA NA
Oklahoma City 4,371.0 4,891.6 11.90% 687.4 734.5 6.90%
Baltimore 6,446.3 6,049.7 -6.20% 1,576.90 1,668.10 5.80%
Louisville 6,160.4 6,403.6 3.90% 2,233.00 2,407.10 7.80%
Average 1.30% 5.50%

The largest increase in violent crime was in San Antonio (52.5%), followed by Charlotte (22.5%). 86 Charlotte data may be skewed by two nights of protests and demonstrations with property destruction following the September 20, 2016, police shooting of Keith Lamont Scott. See Richard Fausset & Alan Blinder, Man Is Shot in Charlotte as Unrest Stretches to Second Night, N.Y. Times (Sept. 21, 2016), http://nyti.ms/2jPjR3C (on file with the Columbia Law Review). The largest decrease in violent crime was in San Diego (-21.0%), followed by Detroit (-7.0%). The largest increases in overall crime rates were in San Antonio (23.3%) and Charlotte (17.5%). 87 But see supra note 86 (noting potentially skewed data based on protests). The largest decreases in overall crime rates were observed in San Francisco (-12.8%) and San Diego (-12.6%).

Table 2: Murder in the Thirty Largest U.S. Cities, 2015 to 2016 88    Id. at 6 tbl.2.

2015 Total Murders 2016 Total Murders % Change in Murders 2015 Murder Rate per 100,000 2016 Murder Rate per 100,000 % Change in Murder Rate
New York City 352 359 2.10% 4.1 4.2 1.20%
Los Angeles 283 276 -2.60% 7.1 6.9 -3.30%
Chicago 493 727 47.40% 18.1 26.6 47.10%
Houston 303 345 13.90% 13.3 14.9 12.10%
Philadelphia 273 293 7.50% 17.4 18.6 6.80%
San Antonio 94 144 52.90% 6.4 9.6 50.00%
San Diego 37 48 30.40% 2.6 3.4 28.90%
Dallas 170 204 20.20% 13 15.4 18.30%
San Jose 30 51 70.60% 2.9 4.8 66.70%
Austin 24 52 115.40% 2.5 5.2 106.30%
Jacksonville NA NA NA NA NA NA
San Francisco 52 56 8.30% 6.3 6.8 8.20%
Indianapolis NA NA NA NA NA NA
Columbus NA NA NA NA NA NA
Fort Worth NA NA NA NA NA NA
Charlotte 60 56 -7.40% 7.3 6.8 -7.60%
Seattle 24 26 8.30% 3.5 3.7 4.90%
Denver 54 56 3.10% 7.9 7.9 0.30%
Detroit 300 293 -2.50% 42.5 41.5 -2.40%
Washington, D.C. 162 144 -10.90% 24.1 21 -12.70%
Boston 39 49 26.70% 5.8 7.3 24.50%
Nashville 63 85 34.50% 9.6 12.7 32.40%
Portland NA NA NA NA NA NA
Oklahoma City 74 95 28.60% 11.7 14.7 25.90%
Las Vegas 134 168 25.30% 9 11.3 25.10%
Baltimore 343 309 -9.90% 55.2 49.8 -9.70%
Louisville 87 111 27.30% 12.9 16.4 26.80%
Average 14.40% 13.10%

As salient as violent crime is, it is a heterogeneous category that includes murder, robbery, assaults, and a small set of other crimes. 89 FBI, 2016 January–June Preliminary Semi-Annual Uniform Crime Report, FBI: Unif. Crime Reporting, http://ucr.fbi.gov/crime-in-the-u.s/2016/preliminary-semiannual-uniform-crime-report-januaryjune-2016/tables/table-1 [http://perma.cc/F2GU-WTPD] (last visited Feb. 3, 2017) (defining violent crimes as “murder, rape . . . robbery, and aggravated assault”). Much of the attention to crime spikes in the past two years has focused on shootings and murders. Murder is the most politically important crime, attracting greater media and political attention than other crimes. Table 2 shows the changes in murders and murder rates for 2015 to 2016 for the thirty largest U.S. cities. The murder rate decreased in the past year in four cities, and the number of murders decreased in five cities. 90 Eight cities did not report data for this analysis. The rest of the twenty-two cities providing data showed increases in these areas in each of the one-year periods. The overall murder count rose by 14.4%, while the murder rate rose by 13.1%. The largest increase for both time periods was in Austin, where the change in the murder count was 115.4% and the change in the murder rate was 106.3%. The largest decrease for both time periods was in Washington, D.C., where the change in the murder count was -10.9% and the change in the murder rate was -12.7%.

Together, Tables 1 and 2 show only a weak linkage between recent trends in increasing (or decreasing) overall crime rates, including violent crime rates, and increases in murder rates. A few cities showed substantial increases in both overall crime and murder: San Antonio, Chicago, and Oklahoma City. Many other cities had large increases in their murder rates but not their overall crime rates: San Jose, San Diego, Nashville, and Boston, among others. Even limiting the comparison to violent crime and murder, the trends seem to suggest a distinct pattern for murder. Murder, then, is different. Its greater salience as a crime is one reason why there has been a sharp focus on murder and consid­erable attention to whether the current rise in homicides portends the end of an era of decline and the onset of a period of instability in murder.

But conclusions based on short-term trends can mask long-term trends with richer explanatory powers. Figures 2 and 3 show longer-term trends for murder in several cities. Figure 2 shows trends for three large cities. The data start in 1985, one year before a nationwide increase in homicides concurrent with the epidemic of crack cocaine in American cities, 91 Alfred Blumstein, Youth Violence, Guns, and the Illicit-Drug Industry, 86 J. Crim. L. & Criminology 10, 11 (1995) (“Both the murder and the robbery rates peaked in about 1980, declined through the early 1980s, and then climbed again during the late 1980s with the intensification of the crack epidemic and the ‘war on drugs.’”); Jeff Grogger & Michael Willis, The Emergence of Crack Cocaine and the Rise in Urban Crime Rates, 82 Rev. Econ. & Stat. 519, 528 (2000) (“Using two different sources of information to date the emergence of crack in 27 metropolitan areas across the U.S., our analysis shows that the arrival of crack cocaine led crime to rise substantially in the late 1980s and early 1990s.”). and continue to 2016, more than twenty years into a nationwide decline in murders—-a decline felt in both large and small cities, as well as in suburban and rural areas. 92 See Alfred Blumstein & Richard Rosenfeld, Explaining Recent Trends in U.S. Homicide Rates, 88 J. Crim. L. & Criminology 1175, 1175 (1998) (“[The homicide rate] climbed a full 24% to a peak of 9.8 in 1991, and has been declining markedly since then . . . .”); Alfred Blumstein, Frederick P. Rivara & Richard Rosenfeld, The Rise and Decline of Homicide—and Why, 21 Ann. Rev. Pub. Health 505, 506–07 & fig.1 (2000) (“The reasons for the rise in homicide in the latter half of the 1980s, peaking in the early 1990s, and the subsequent equally dramatic decline span both criminology and public health.”); Jacqueline Cohen et al., The Role of Drug Markets and Gangs in Local Homicide Rates, 2 Homicide Stud. 241, 241–42 (1998) (discussing recent shifts in homicide rates); Steven D. Levitt, Understanding Why Crime Fell in the 1990s: Four Factors that Explain the Decline and Six that Do Not, 18 J. Econ. Persp. 163, 163 (2004) (“Homicide rates plunged 43 percent from the peak in 1991 to 2001, reaching the lowest levels in 35 years.”); see also Franklin E. Zimring, The Great American Crime Decline 3 (2007) (“The data show a very substantial and nationwide drop, across all categories of serious crime, steadily progressing through the decade.”); Michael Tonry, Why Crime Rates Are Falling Throughout the Western World, 43 Crime & Just. 1, 1–2 (2014) (recognizing and exploring causes for parallel dropping crime rates among English-speaking countries and western Europe). Murders in these three cities declined steadily from their similar 1990–1995 peak rates through 2014. Chicago’s murder rate stabilized at a higher rate than the rates in New York City and Los Angeles (L.A.). In L.A., a sharp decline from 1993 to 1999 was followed by an increase through 2002, when the decline resumed. After nearly a decade of the second decline, the murder rate in L.A. rose slightly from 2014 to 2016. In New York City, the murder rate continued its durable secular decline through 2014. After a very small one-year increase in 2015, the decline in New York City resumed in 2016.

Figure 2: Murders per 100,000 Persons in the Three Largest U.S. Cities, 1985 to 2016 93 The data used to produce this figure are from the FBI’s Uniform Crime Reporting (UCR) program. The FBI’s UCR data from 1975 to 2015 were compiled in one location by the Marshall Project. See Gabriel Dance, Tom Meagher & Emily Hopkins, The Marshall Project, Data for Examining Uniform Crime Reporting Data for 68 Major Cities, GitHub, http://github.com/themarshallproject/city-crime [http://perma.cc/V79G-JP3Q] [hereinafter Dance et al., UCR Data for 68 Cities] (last visited Mar. 27, 2017).


Figure 2 has two important lessons. The first is that not all crime or murder declines are alike. Even in the midst of a robust long-term period of decline, there are temporary increases. Across these three large cities, the timing, duration, and magnitude of the spikes within the longer-term trends vary. These secular declines in fact contain short-term anomalies that may have city-specific explanations.

The second lesson is that cities may have similarities in slopes or rates of decline for long periods but also have level differences over time. Chicago’s decline was more short-lived, but its period of relative stability—with short-term increases and declines—lasted nearly a decade, from 2004 to 2014, before the current spike began. What also stands out from Figure 2 is the magnitude of the two-year spike in Chicago beginning in 2015. This extraordinary spike in Chicago’s murder rate, does have analogs in other cities, as we see in Figure 3.

Figure 3: Murders per 100,000 Population in Ten U.S. Cities, 1985 to 2016 94 Id.


Figure 3 expands the analysis in Figure 2 to include seven additional cities. The city-specific trends reveal that most cities follow a similar pattern of increases and declines over time. 95 Gabriel Dance & Tom Meagher, Crime in Context: Violent Crime Is Up in Some Places, but Is It Really a Trend?, Marshall Project (Aug. 18, 2016), http://www.themarshallproject.org/2016/08/18/crime-in-context#.6HgY9cx0 [http://perma.cc/C66P-A527] (finding sixty-eight cities could fit into four patterns of crime); see also Dance et al., UCR Data for 68 Cities, supra note 93 (providing the table of data used to analyze crime in sixty-eight cities over the forty-year period presented in Dance & Meagher, supra). Several patterns emerge across these city-specific trends. First, Chicago is not alone in a steep increase in homicide rates starting in 2015. Murder rates also increased sharply in St. Louis, Baltimore, Milwaukee, and New Orleans. Indeed, it is worth noting that many of the cities with the highest murder rate in 2016—including St. Louis, 59.3, and New Orleans, 44.5—are not among the thirty largest cities. 96 See Jeff Asher, U.S. Cities Experienced Another Big Rise in Murder in 2016, FiveThirtyEight (Jan. 9, 2017, 2:32 pm), http://fivethirtyeight.com/features/u-s-cities-experienced-another-big-rise-in-murder-in-2016/?ex_cid=story-twitter [http://perma.cc/L5QW-LLKB]. The below chart uses murder rate data from this article.
U.S. Cities with the Highest 2016 Murder Rates
City 2016 Murder Rate (per 100K)
St. Louis 59.3
Baltimore 51.2
Detroit 44.9
New Orleans 44.5
Cleveland 35
Newark, N.J. 33
Memphis, Tenn. 31.9
Chicago 27.9
Kansas City, Mo. 26.4
Atlanta 23.9

Thus, while Chicago’s size makes it account for a disproportionate share of the total increase in homicides during this time, its per capita increase is in line with these other cities that suffered increases as well.Thus, while Chicago’s size makes it account for a disproportionate share of the total increase in homicides during this time, its per capita increase is in line with these other cities that suffered increases as well.

Second, the trends of several cities show continued stability in their murder rates, and any increases in murder rates since 2012 were modest. These cities include Houston, New York City, L.A., and Philadelphia. These four cities also had long-term declines in murder rates dating to the early 1990s. Thus, the long-term trend is one of stability for these places after a lengthy period of decline. Baltimore and St. Louis also had sharp declines in their murder rates followed by short-term sharp spikes of one to two years.Second, the trends of several cities show continued stability in their murder rates, and any increases in murder rates since 2012 were modest. These cities include Houston, New York City, L.A., and Philadelphia. These four cities also had long-term declines in murder rates dating to the early 1990s. Thus, the long-term trend is one of stability for these places after a lengthy period of decline. Baltimore and St. Louis also had sharp declines in their murder rates followed by short-term sharp spikes of one to two years.Second, the trends of several cities show continued stability in their murder rates, and any increases in murder rates since 2012 were modest. These cities include Houston, New York City, L.A., and Philadelphia. These four cities also had long-term declines in murder rates dating to the early 1990s. Thus, the long-term trend is one of stability for these places after a lengthy period of decline. Baltimore and St. Louis also had sharp declines in their murder rates followed by short-term sharp spikes of one to two years.Second, the trends of several cities show continued stability in their murder rates, and any increases in murder rates since 2012 were modest. These cities include Houston, New York City, L.A., and Philadelphia. These four cities also had long-term declines in murder rates dating to the early 1990s. Thus, the long-term trend is one of stability for these places after a lengthy period of decline. Baltimore and St. Louis also had sharp declines in their murder rates followed by short-term sharp spikes of one to two years.

A third lesson in Figure 3 is from the clustering of six cities with similar homicide rates and similar long-term trends. Despite the increases in murder rates in Chicago and Milwaukee, the longer-term trends for this group suggest periods of decline marked by periods of short-term increases. The other four cities have overall higher murder rates over the past twenty-two years and more volatility in those murder rates. New Orleans in particular has very large increases and declines over time, even if we discount the sharp spike in the two years immediately after Hurricane Katrina. 97 Michael Leitner et al., The Impact of Hurricane Katrina on Reported Crimes in Louisiana: A Spatial and Temporal Analysis, 63 Prof. Geographer 244, 258 (2011) (finding “murder and aggravated assault rates even exceeding prestorm averages by the end of December 2007” in New Orleans); Mark VanLandingham, Letters, 2007 Murder Rates in New Orleans, Louisiana, 98 Am. J. Pub. Health 776, 776 (2008) (“The murder rate for New Orleans remains stubbornly elevated at extraordinarily high levels since Hurricane Katrina and has yet to begin a decline to even prestorm levels, which were very high in the first place.”); Mark VanLandingham, Murder Rates in New Orleans, La, 2004–2006, 97 Am. J. Pub. Health 1614, 1615 (2007) (“Compared with rates in cities of comparable size, murder rates in New Orleans have been substantially higher since at least 2004, and the disparity has been worsening since then.”); Sean P. Varano et al., A Tale of Three Cities: Crime and Displacement After Hurricane Katrina, 38 J. Crim. Just. 42, 48 (2010) (noting “[i]t seems safe, however, to conclude that the Katrina-related diaspora did not correspond with a wide-scale dramatic increase in crime”).

One important way to distinguish and assign meaning to the short-term spikes in Chicago and Milwaukee compared to the spikes in St. Louis and Baltimore is to develop an understanding of longer-term stability in murder rates and how city-specific factors contribute to that stability. Given these differences in the longer-term trends, it would not be unreasonable to predict that city-specific factors contribute to the longer-term stability in murder rates. Also, given the longer-term trends of increase followed by decline over a longer time window, it would not be unreasonable to predict that the current spikes will be of shorter duration and that a mean reversion to the long-term trend could begin within a year or two. Indeed, there is no evidence to suggest that the current increases would depart from the longer-term trend over the past three decades. In the cities with more volatile murder rates over the past twenty years, such predictions would be far more uncertain and prone to error without a much deeper understanding of the city-specific factors: the neighborhood contexts where spikes are concentrated, changes in the socio-economic landscape that may have preceded or accompanied these spikes, changes in the predicate crimes (such as robberies) that often have parallel trends with murder, and factors specific to the social organization of crime and its opportunities.

Figure 4: Murders per 100,000 Population, United States, 1960 to 2016 98 The data used to produce this figure are from the FBI’s UCR report. See FBI, Table 1: Crime in the United States by Volume and Rate per 100,000 Inhabitants, 1996–2015, FBI: Unif. Crime Reporting, http://ucr.fbi.gov/crime-in-the-u.s/2015/crime-in-the-u.s.-2015/tables/table-1 [http://perma.cc/2EZ2-FM5W] (last visited Mar. 27, 2017). Data from 1960 to 2014 are available from the FBI’s Uniform Crime Reporting Statistics Data Tool. See FBI, Estimated Murder Rate from 2009–2014, Unif. Crime Reporting Statistics, http://www.ucrdatatool.gov/Search/Crime/State/TrendsInOneVar.cfm (To create the table using this tool, choose “United States – Total” for option a, “murder rate” for option b, and enter the range from 1960 to 2014 for option c, then click “get table.”) (on file with the Columbia Law Review) (last visited Mar. 27, 2017). The rate for 2016 was estimated by the Brennan Center in Friedman et al., supra note 81, at 1 (noting that the murder rate was projected to rise by 13.1% from 2015–2016).


Efforts to explain these short-term murder increases have been expansive but with little clear empirical or analytic basis. One explanation centers on an increasing tendency among young males to settle disputes with lethal violence—violence that is made easier by ready access to guns. 99 See Julie Bosman & Mitch Smith, As Chicago Murder Rate Spikes, Many Fear Violence Has Become Normalized, N.Y. Times (Dec. 28, 2016), http://nyti.ms/2kqLFuY (on file with the Columbia Law Review) (noting criminology professor Arthur Lirigio’s assertion that “in addition to the proliferation of guns and deepening poverty, personal disputes among gang members have led to more shootings”). Another narrative centers on increasing tensions among highly localized gangs protecting small territories whose boundaries are easily transgressed, provoking attacks that are cast as territorial defense. 100 See id. (quoting Chicago Police Superintendent Eddie Johnson on the proliferation of shootings that were targeted attacks by gang members). Still other explanations suggest an inability of police to solve murders and arrest the killers, leaving them free to kill again. 101 Catherine Lee, The Value of Death: Multiple Regression and Event History Analyses of Homicide Clearance in Los Angeles County, 33 J. Crim. Just. 527, 530–33 (2003) (showing that homicide clearance rates are lower in L.A. in murders of non-White victims and that clearances via arrests are faster for White-victim homicides, combining to leave the killers of non-White victims free to kill again).

Murder rates across American cities are heterogeneous. Disputes over drug sales, domestic homicides, and felony murders are all part of the story. Some murders are of innocent bystanders who are fatally wounded by stray bullets that are intended for others. Nevertheless, there are some constants in the demography, epidemiology, and social contexts of murder. We turn to those next.

B. The Past Half Century

Over the past five decades, there have been three distinct homicide epidemics in the United States. 102 See supra Figure 4. The first started in the mid-1960s and lasted for nearly a decade before a short-term decline through the end of the 1970s. The second began in the late 1970s, had a relatively short duration, and was followed by a precipitous decline in the murder rate through 1985. The third coincided with the crack epidemic that started in the late 1980s and continued through the early 1990s. 103 Grogger & Willis, supra note 91, at 528 (finding “the effect of the emergence of crack was substantial” and “in the absence of crack cocaine, the crime rate in 1991 would have remained below its previous peak in the early 1980s”). Although the onset and peak of each epidemic varied by city in its duration and (with minor variations) its intensity, all major cities assumed the same contours of recurring and cascading homicide epidemics. Figure 4 shows the national trend. City-specific trends for the period 1975 to 2015 reveal that most cities followed a similar pattern over time. 104 See supra note 95 and accompanying text.

There is no shortage of explanations for each of the three homicide epidemics in Figure 4. Explanations include increasing racial residential segregation, 105 Douglas S. Massey & Nancy A. Denton, American Apartheid: Segregation and the Making of the Underclass 8 (1993) (“Residential segregation is the institutional apparatus that supports other racially discriminatory processes and binds them together into a coherent and uniquely effective system of racial subordination.”). increasing poverty and the effects of deindustrialization on U.S. cities, 106 Robert J. Sampson & William Julius Wilson, Toward a Theory of Race, Crime, and Urban Inequality, in Crime and Inequality 37, 53 (John Hagan & Ruth D. Peterson eds., 1995) (suggesting “roots of urban violence” were possibly the result of a cohort of youth that “spent its childhood in the context of a rapidly changing urban environment unlike that of any previous point in U.S. history” when “the concentration of urban poverty and other social dislocations began increasing sharply”); see also Elijah Anderson, Code of the Street: Decency, Violence, and the Moral Life of the Inner City 108 (1999) (noting when jobs disappear “the underground economy of drugs and crime often emerges to pick up the slack”); Douglas S. Massey, Categorically Unequal: The American Stratification System 205–07 (2007) (linking extreme inequality and spatial concentration of poverty to the adaptation of cultural mechanisms that sustain violence). the proliferation of guns and increased firearm availability through both legal and illegal channels, 107 See Alfred Blumstein & Daniel Cork, Linking Gun Availability to Youth Gun Violence, Law & Contemp. Probs., Winter 1996, at 5, 5 (“By examining time trends in age-specific arrest rates for homicide (gun homicide compared to non-gun homicide) and similar trends in drug-related arrest rates (juveniles compared to adults), the role of gun availability . . . is identified as a probable cause of these homicide trends.”); Matthew Miller et al., Rates of Household Firearm Ownership and Homicide Across US Regions and States, 1988–1997, 92 Am. J. Pub. Health 1988, 1992 (2002) (finding “across US regions and states, and for virtually every age group, higher rates of household firearm ownership were associated with higher rates of homicide”); see also Philip J. Cook & Jens Ludwig, The Social Costs of Gun Ownership, 90 J. Pub. Econ. 379, 380 (2006) (finding “the social cost of an additional household acquiring a handgun depends on the rate of violence and the existing prevalence of guns, but under a wide range of assumptions is greater than $100 per year”). and the spread of street gangs in U.S. cities. 108 Jacqueline Cohen & George Tita, Diffusion in Homicide: Exploring a General Method for Detecting Spatial Diffusion Processes, 15 J. Quantitative Criminology 451, 453 (1999) (“Features of the recent homicide epidemic . . . suggest that certain social processes—notably, illegal drug markets and gang rivalries—may be important for explaining the pattern and mechanisms of the spread of homicides.”). Each of these explanations has some empirical and theoretical traction for specific places and specific eras. But common factors that can be plausibly linked to the temporal and spatial patterns in these homicide epidemics can provide a perspective, if not an explanation, that unifies these three tightly linked epidemics. 109 See, e.g., Robert M. Lawless, Jennifer K. Robbennolt & Thomas S. Ulen, Empirical Methods in Law 23–24 (2d ed. 2016) (discussing the conditions to establish a causal relationship between two factors). This Essay proposes that the link is a series of recurring and cumulative epidemics within cities of illicit drug use and illegal distribution, and that these epidemic patterns are shared across cities.

Each of these three periods of increasing and then decreasing murder rates experienced concurrent epidemics of illicit drugs. The first period from the mid-1960s to the late-1970s, with its increasing and then decreasing murder rate, was concurrent with the onset and decline of a heroin epidemic. The heroin epidemic of the 1960s and 1970s hit most of the large and small cities in the United States. 110 Leon G. Hunt, Recent Spread of Heroin Use in the United States, 64 Am. J. Pub. Health 16, 16 (Supp. 1974) (“Aggregate treatment program data seem to show that the peak period of new heroin use occurred everywhere . . . about 1968, but when local data are corrected for addicts’ delay in entering treatment, this synchronous national ‘epidemic’ becomes a sequence of local peaks ranging from 1967 to the present.”); Bruce D. Johnson et al., Drug Abuse in the Inner City: Impact on Hard-Drug Users and the Community, 13 Crime & Just. 9, 14–15 (1990) [hereinafter Johnson et al., Drug Abuse in the Inner City] (describing the rise of heroin use in inner cities from 1965 to 1973). Heroin addiction was linked to violence through distinct mechanisms. The first was acquisitive violence—users and addicts committed crimes to get access to funds to buy drugs. 111 Paul J. Goldstein, The Drugs/Violence Nexus: A Tripartite Conceptual Framework, 39 J. Drug Issues 493, 495–96 (1985) (“The economically compulsive model suggests that some drug users engage in economically oriented violent crime, e.g., robbery, in order to support costly drug use.”). Robberies of other drug users, drug sellers, or individuals unrelated to such transactions were the primary income-generating mechanisms. Robbery rates soared—the linkage between robbery and homicide was even analyzed by a presidential commission on crime in the late 1960s. 112 See President’s Comm’n on Law Enf’t & Admin. of Justice, The Challenge of Crime in a Free Society 18–19 (1967), http://www.ncjrs.gov/pdffiles1/nij/42.pdf [http://perma.cc/9JAN-3YEZ]. The second mechanism linking heroin to violence was competition and conflicts between drug sellers, and the third (though rare) was violence resulting from the psychoactive effects of drugs. 113 Goldstein, supra note 111, at 503 (“Drugs and violence are seen as being related in three possible ways: the psychopharmacological, the economically compulsive, and the systemic.”); see also Paul J. Goldstein et al., Crack and Homicide in New York City, 1988: A Conceptually Based Event Analysis, 16 Contemp. Drug Probs. 651, 684 (1989) (identifying the “crack distribution ‘system’ as the primary source of crack-related homicidal violence” and noting few instances of “psychopharmacological or economic compulsive homicides involving crack”).

The mid-twentieth century heroin epidemic subsided (though it never disappeared), but within a short time, a cocaine epidemic took hold—an epidemic concurrent with the second period of the increase and decrease of the murder rate. 114 Andrew Golub, Bruce D. Johnson & Eloise Dunlap, Subcultural Evolution and Illicit Drug Use, 13 Addiction Res. & Theory 217, 222 (2005) (explaining four distinct phases to drug eras and referencing their application to the “[c]rack era of the late 1980s and early 1990s”); Bruce D. Johnson et al., The Rise and Decline of Hard Drugs, Drug Markets, and Violence in Inner-City New York, in The Crime Drop in America 164, 168 (Alfred Blumstein & Joel Wallman eds., 2000) (referencing the “Crack/Cocaine Era” of the 1980s). The marketing of cocaine in this era differed extensively from the heroin trade in the previous decade. 115 Eloise Dunlap & Bruce D. Johnson, The Setting for the Crack Era: Macro Forces, Micro Consequences (1960–1992), 24 J. Psychoactive Drugs 307, 318–20 (1992). Retail heroin sales were less frequent due to the long duration of a heroin high and a tightly controlled illicit market. In turn, the volume of sales was far lower, and sales were more often done furtively; open-air markets were atypical. 116 Id. at 318–19. Cocaine marketing, on the other hand, flourished in outdoor open-air bazaars. Cocaine’s psychotropic effects were different: Cocaine was a stimulant, in contrast to heroin, so users could repurchase often and maintain their high over time while doing other activities. 117 See Johnson et al., Drug Abuse in the Inner City, supra note 110, at 15–16 (noting “cocaine gained the reputation of a ‘status drug’ that was relatively innocuous”). In turn, the markets that sold cocaine in the street served a wider population. 118 See id. at 23–26 (discussing the range of clientele for cocaine dealers); see also Jerome H. Skolnick et al., The Social Structure of Street Drug Dealing, 9 Am. J. Police 1, 18–20 (1990) (describing marketing arrangements of cocaine deals). Cocaine highs were shorter lasting, and users were more likely to make several purchases of smaller quantities in a shorter period of time.

The emergence and visibility of drug markets in this era attracted police attention in the form of undercover buy-and-bust operations that resulted in a high volume of arrests. 119 Lynn Zimmer, Proactive Policing Against Street-Level Drug Trafficking, 9 Am. J. Police 43, 47–48 (1990) (describing New York City’s Operation Pressure Point as “an aggressive, proactive strategy with both crime-fighting and order-maintenance goals”). One perverse effect of street-level enforcement was the disruption of drug-selling organizations, creating market instability and encouraging conflict and violence as new sellers moved to occupy the vacuums created by drug enforcement. 120 See Goldstein, supra note 111, at 497 (describing such behavior as an example of “systemic violence”); see also Jeffrey Fagan & Ko-lin Chin, Violence as Regulation and Social Control in the Distribution of Crack, in Drugs and Violence: Causes, Correlates, and Consequences 8, 24–35 (Mario de la Rosa et al. eds., 1990), http://archives.drugabuse.
gov/pdf/monographs/103.pdf [http://perma.cc/KV4T-5XK8] (discussing the connection between violence and drug selling); Johnson et al., Drug Abuse in the Inner City, supra note 110, at 39 (describing the limited impact of police arrests of crack-user dealers).
This conflict resulted in what Professor Paul Goldstein has called “systemic violence” within the illicit drug trade. 121 Goldstein, supra note 111, at 497–501.

As Figure 4 shows, there was a drop in homicide and violence at the end of this epidemic period over a five-year interval ending in 1985, as these markets and the population of users stabilized. In any epidemic, there is a period of adoption and spread among susceptible persons, but at some point, anyone who is motivated to try a drug has tried and either continued its use or set it aside. 122 Roberta Ference, Diffusion Theory and Drug Use, 96 Addiction 165, 170 (2001) (finding considerable support for “the application of the diffusion model in a variety of planned interventions relating to drug use”); S.A. Patten & J.A. Arboleda-Florez, Epidemic Theory and Group Violence, 39 Soc. Psychiatry & Psychiatric Epidemiology 853, 853, 855 (2004) (describing epidemic diseases and their application to “psychiatric epidemiological research” to explain “events that might otherwise be difficult to explain”). As the supply of new users declined, so too did the epidemic, and in turn, drug markets declined in size and activity. While homicide levels remained higher than twenty years earlier, the decline over that five-year period was precipitous.

The final period of increase and decrease of the murder rate coincided with the spread of crack cocaine—smokable cocaine—across U.S. cities starting in the mid-1980s. Street drug markets revived, expanded, and became more active. 123 See Johnson et al., Drug Abuse in the Inner City, supra note 110, at 16–17. Crack was sold in smaller quantities, with vials of two-to-four rocks costing about $20 to $30, and since the high was short lived, users continued to make purchases over a sustained period of time. 124 Craig Reinarman et al., The Contingent Call of the Pipe: Bingeing and Addiction Among Heavy Cocaine Smokers, in Crack in America: Demon Drugs and Social Justice 77, 83 (Craig Reinarman & Harry Levine eds., 1997) (describing the short nature of the high and the “strong desire to repeat it”). Young men were attracted to the drug trade in large part because of the promise of incomes far above what they could earn in conventional work. 125 See Philippe Bourgois, In Search of Respect: Selling Crack in El Barrio 115 (2d ed. 2003) (observing “straightforward refusal to be exploited in the legal labor market pushes [dealers] into the crack economy”); Jeffrey Fagan, Drug Selling and Licit Income in Distressed Neighborhoods: The Economic Lives of Street-Level Drug Users and Dealers, in Drugs, Crime, and Social Isolation 99, 100–01 (Adele Harrell & George E. Peterson eds., 1992) [hereinafter Fagan, Drug Selling and Licit Income] (noting the “profitability of drug dealing obviously affects the feasibility of different strategies for steering inner-city youth back to the legal labor market”). Crack cocaine was a very strong stimulant, and sexual stimulation was one of its effects. This led to the transformation of sex markets as crack was marketed together with sex, and women users often exchanged sex for drugs. 126 Lawrence J. Ouellet et al., Crack Cocaine and the Transformation of Prostitution in Three Chicago Neighborhoods, in Crack Pipe as Pimp: An Ethnographic Investigation of Sex-for-Crack Exchanges 69, 85–92 (Mitchell S. Ratner ed., 1993) (describing how the “financial demands of repeated cocaine use” drove prostitution). Competition among crack-distribution organizations––which also sold other drugs, including cocaine and heroin-–-was often fierce and violence was common. 127 See Bruce D. Johnson et al., Emerging Models of Crack Distribution, in Drugs and Crime: A Reader 56, 65 (T. Mieczkowski ed., 1991) (“[M]ost vertically controlled crack groups must be able and willing to resort to ruthless violence to instill fear and respect among competitors and workers alike.”). Street gangs entered the crack trade, also selling cocaine and heroin, and their competition was another context in which violence often erupted. 128 Blumstein, supra note 91, at 26–28 (discussing the impact of the drug industry’s violence on the community and its potential to stimulate violence).

Crack spread across cities, and its arrival in each city was tightly linked with a surge in violence, including homicides. 129 Grogger & Willis, supra note 91, at 523. In many in­stances, drug sales occurred in open-air markets located in the poorest, highest-crime, and most heavily minority-population neighborhoods of the cities. 130 Fagan & Chin, supra note 120, at 9–10 (describing the emergence of crack markets); see also Ric Curtis, Crack, Cocaine and Heroin: Drug Eras in Williamsburg, Brooklyn, 1960–2000, 11 Addiction Res. & Theory 47, 56–57 (2003) (discussing the crack market in Willamsburg, Brooklyn); Johnson et al., Drug Abuse in the Inner City, supra note 110, at 18–22 (describing structure of crack-selling industries). These also were the places where access to legal work, both socially and spatially, was severely limited, and the local formation of relatively lucrative drug markets incentivized drug selling for those shut out of legal work. 131 Fagan, Drug Selling and Licit Income, supra note 125, at 99–101 (discussing the influence of the drug trade on employment during 1980s crack-cocaine epidemic). For others, drug selling offered a measure of dignity compared to the noxious conditions of low-wage hourly legal work. 132 See Bourgois, supra note 125, at 98–99 (discussing crack dealers’ stories of “rejecting what they considered to be intolerable working conditions at entry-level legal jobs” and their dependence on the “crack economy for economic survival and personal dignity”). Put another way, crack markets opened in the context of the very same poverty traps where earlier homicide epidemics were the most salient.

The systemic violence that Professor Goldstein described spread in these cities: Conflicts between dealers and between dealers and cus­tomers created a context in which violence was the most reliable form of resolving disputes. 133 See Fagan & Chin, supra note 120, at 8–9 (noting “[v]iolence in drug dealing can be viewed as an extension of behaviors that are associated with efficiency and success in legitimate businesses”); Johnson et al., Drug Abuse in the Inner City, supra note 110, at 35–38. Dealers themselves were often targets for ordinary robberies because they carried large amounts of cash and worked on street corners with little protection from police or others in their drug-selling clique. 134 Fagan & Chin, supra note 120, at 8–9 (noting “in the absence of legal recourse for legal activities, [street-level] disputes are likely to be settled either by economic reprisal or violence”); Johnson et al., Drug Abuse in the Inner City, supra note 110, at 27 (“In addition to evading police, sellers and suppliers must now defend their ‘businesses’ against robbers, competitors, and ordinary citizens who will call police.”). Violence also was used by drug organizations to maintain discipline and ensure loyalty so that lower-level retail sellers would be deterred from skimming profits from the wholesalers. 135 Fagan, Drug Selling and Licit Income, supra note 125, at 117 (noting that “violence can be used as a regulatory process or a strategy for organizational maintenance in an economic activity that falls outside of legal or formal economic control”); Fagan & Chin, supra note 120, at 8–9 (observing “in the absence of legal recourse for legal activities, [street-level] disputes are likely to be settled either by economic reprisal or violence”).

As in earlier drug epidemics, the crack-cocaine epidemic eventually peaked, reaching its apex in the early 1990s, and then began to recede as the supply of new users was exhausted and older users ceased using crack. Extraordinary law enforcement resources were allocated to con­trolling crack markets, including federal as well as local enforcement. 136 Steven R. Belenko, Crack and the Evolution of Anti-Drug Policy 115, 122–24 (1993) (describing the “intensive law enforcement response” to the growth of crack offenses at federal, state, and local levels); David W. Rasmussen & Bruce L. Benson, Rationalizing Drug Policy Under Federalism, 30 Fla. St. U. L. Rev. 679, 714 (2003) (describing the growth of law enforcement measures at all levels to address drug crime). This combination of intensive criminal justice intervention and the natural cessation of the crack epidemic in the early 1990s also contributed to the downward trend in murder and other forms of criminal violence. Without the impetus of emerging and volatile drug markets—a major animating feature of the previous three consecutive epidemics—the next two decades saw a decline in murders across the country. While there were city-specific trends for short-term disruptions in the pattern of decline, the absence of these animating contexts helped keep the disruptions short.

C. A New Epidemic?

The current upward trend, even for the 2015–2016 period, seems to coincide with an increase in heroin use and abuse of other forms of opioids across the nation. 137 Injury Prevention & Control: Opioid Overdose, Understanding the Epidemic, Ctrs. for Disease Control & Prevention, http://www.cdc.gov/drugoverdose/epidemic/ [http://perma.cc/7KEQ-S4R4] (last updated Dec. 16, 2016); Dan Nolan & Chris Amico, Chasing Heroin: How Bad Is the Opioid Epidemic?, PBS: Frontline (Feb. 23, 2016), http://www.pbs.org/wgbh/frontline/article/how-bad-is-the-opioid-epidemic/ [http://perma.cc/735P-3S5F]. But facts challenge the notion that another homicide epidemic is underway or that the increase in homicide is fueled by yet another era of drug addiction and violence in its distribution networks.

First, the onset of this opioid epidemic predated the recent rise in homicides. In previous eras, the onset of the epidemics of drugs and homicide were closely correlated in time and space in urban centers through the late 1990s. 138 See, e.g., President’s Comm’n on Law Enf’t & Admin. of Justice, supra note 112, at 222 (beginning the exploration between drug abuse and violent crimes in the mid-1960s); Grogger & Willis, supra note 91, at 528 (discussing the correlation between drug prevalence and violent crime in the crack epidemic). The opioid epidemic, mainly in rural America, began approximately in 2000, 139 Laxmaiah Manchikanti et al., Opioid Epidemic in the United States, 15 Pain Physician ES9, ES29–ES30 figs.15 & 16 (2012) (showing a surge in opioid analgesic overdose deaths and treatment admissions beginning in 1999 and a rate that rose monotonically through the end of the following decade). as homicides were well into their twenty-year decline. From 2000 to 2014, the rate of drug overdose deaths increased by 137%—in 2014, about 60% of those excess deaths from drug overdoses were due to opioid use. 140 Rose A. Rudd et al., Increases in Opioid Overdose Deaths—United States, 2000–2014, 64 Morbidity & Mortality Wkly. Rep. 1378, 1378–79 (2016). In 2009, the death rate from opioids surpassed the death rate from motor vehicle accidents. 141 Leonard J. Paulozzi, Prescription Drug Overdoses: A Review, 43 J. Safety Res. 283, 283, 284 fig.1 (2012). Deaths from prescription opioid pain medications rose by nearly 400% from 1999 to 2014. 142 Nat’l Ctr. for Health Statistics, Ctrs. for Disease Control & Prevention, Number and Age-Adjusted Rates of Drug-Poisoning Deaths Involving Opioid Analgesics and Heroin: United States, 2000–2014 (2016), http://www.cdc.gov/nchs/data/health_
policy/AADR_drug_poisoning_involving_OA_Heroin_US_2000-2014.pdf [http://perma.
However, Figure 3 shows that the homicide rate nationally did not begin to rise until 2015 143 See Richard Rosenfeld, supra note 79, at 6 (showing homicide rate increases in large cities beginning in 2015). in some cities and later in several others.

Second, unlike the drug epidemics from the 1960s to the 1990s, opioid addiction is more likely to affect Whites than other racial or ethnic groups and to be centered outside urban areas. 144 Haeyoun Park & Matthew Bloch, How the Epidemic of Drug Overdose Ripples Across America, N.Y. Times (Jan. 19, 2016), http://nyti.ms/2jVUlKb (on file with the Columbia Law Review) (visually showing the spread of heroin overdose deaths from a small set of rural areas in 2003 to much of the western United States and Appalachian areas by 2014). Heroin and opioid mortality rates in 2014 were more than two times greater for Whites than Blacks or Latinos. 145 See Nolan & Amico, supra note 137 (“By 2014, whites and Native Americans were dying at double or triple the rates of African-Americans and Latinos, according to the CDC.”). The Centers for Disease Control and Prevention has not yet published heroin or opioid addiction overdose mortality rates for 2015 or 2016. This is in contrast to the 1960s, when Blacks and Latinos were far more likely to die from heroin than Whites. 146 See Hunt, supra note 110, at 23 (explaining heroin addicts in treatment in 1974 were predominantly Black men); Johnson et al., Drug Abuse in the Inner City, supra note 110, at 14 (explaining heroin use “exploded during the period 1965–73” among minorities but that since that time “much lower proportions of blacks have been initiating heroin use”); see also Theodore J. Cicero et al., The Changing Face of Heroin Use in the United States: A Retrospective Analysis over the Past 50 Years, 71 JAMA Psychiatry 821, 825 (2014) (“Although minority groups were predominant [heroin] users in the 1960s and 1970s, nearly 90% of respondents who began use in the last decade were white.”). This shift in the demography of drug abuse relocates the drug–homicide connection away from the populations and communities most at risk for homicide mortality. 147 Peterson & Krivo, supra note 78, at 13 (noting “the rate of homicide offending and victimization was six to seven and a half times higher for African Americans than for whites each year from 2000 to 2005”).

Third, the spatial distribution of the opioid epidemic suggests a diffusion of heroin and opioid use away from the nation’s urban centers since 2000. 148 See Katherine M. Keyes et al., Understanding the Rural–Urban Differences in Nonmedical Prescription Opioid Use and Abuse in the United States, 104 Am. J. Pub. Health at e52, e52 (2014) (recognizing and seeking to explain “the excess burden of prescription opioid misuse in rural compared with urban areas”); see also Luke Runyon, Why Is the Opioid Epidemic Hitting Rural America Particularly Hard?, Harvest Pub. Media (Dec. 22, 2016), http://harvestpublicmedia.org/article/why-opioid-epidemic-hitting-rural-america-particularly-hard [http://perma.cc/7S9N-T538] (discussing the influx of opioids in rural areas). For example, heroin and opioid overdoses in California in 2013 also were highest in the state’s rural counties, including Shasta, Placer, and El Dorado counties. 149 Phillip Reese-Presse, See Where California’s Heroin, Opioid Problems Are Worst, Sacramento Bee (Aug. 17, 2015), http://www.sacbee.com/site-services/databases/
article31324532.html [http://perma.cc/M6UL-YZTE].
In general, county-level rates of overdose deaths from heroin and prescription opioids remained higher in rural areas than in the urban centers. 150 Park & Bloch, supra note 144. Even without an acute opioid epidemic in cities, homicide rates in cities have remained higher than in rural or suburban areas for the past fifty years, 151 Erica L. Smith & Alexia Cooper, Bureau of Justice Statistics, Homicide in the U.S. Known to Law Enforcement, 2011, at 11 figs.19 & 20 (2013), http://www.bjs.gov/content/
pub/pdf/hus11.pdf (on file with the Columbia Law Review) (showing elevated homicide rates in large cities with populations greater than 500,000).
particularly in places where racial residential segregation also remains concentrated. 152 John R. Logan & Brian J. Stults, US2010 Project, The Persistence of Segregation in the Metropolis: New Findings from the 2010 Census 21 (2011), www.s4.brown.edu/
us2010/Data/Report/report2.pdf [http://perma.cc/QEB9-6MUK]; see also Massey & Denton, supra note 105, at 223 (reviewing census data and reaffirming the prominence of racial segregation in urban areas).
Explanations for the diffusion include higher rates of opioid pre­scriptions in rural areas, an outmigration of young adults from cities and suburbs to rural areas, denser kinship and social networks in rural areas that may facilitate drug diversion from legal to illegal channels, and economic stressors that heighten vulnerability to drug use generally. 153 Keyes et al., supra note 148, at e54 (discussing factors related to the spread of heroin use from urban to rural areas starting in the late 1990s). Users may also minimize the perceived harm of pharmaceutical opioids, owing to their oral ingestion that avoids the stigma and health complications of injection. 154 See, e.g., Don C. Des Jarlais et al., The Transition from Injection to Non-Injection Drug Use: Long-Term Outcomes Among Heroin and Cocaine Users in New York City, 102 Addiction 778, 782 (2007) (identifying self-reported reasons users stopped injecting, including issues of stigmata and self-image).

In cities, heroin deaths have also increased since 2010, but the increase is not skewed by race. St. Louis, for example, is one of the cities with a sharp spike in murders in the past five years. 155 See supra Figure 3. And while the heroin overdose rate rose in St. Louis from 2010 to 2014, the rate of heroin overdose deaths was slightly higher for Whites (10.7 per 100,000 persons) than Blacks (10.0 per 100,000 persons). 156 St. Louis Cty. Dep’t of Pub. Health, Drug-Poisoning Deaths Involving Heroin, St. Louis County, Missouri 2 (2016), http://www.stlouisco.com/Portals/8/docs/Health/
Health%20Data/Heroin%20Deaths%20Profile%2022416%20FINAL.pdf [http://perma.cc/TV23-MB9N].
The rate increased from 2013 to 2014 nearly identically for Black and White persons at the same time that homicides were spiking. 157 Id. at 4. Yet Black persons were more likely to be homicide victims during much of this period, despite the racial balance in heroin overdose deaths. 158 Alex Ihnen, Understanding St. Louis Homicides, 2005–2012, Next STL (Jan. 5, 2013), http://nextstl.com/2013/01/understanding-st-louis-homicides-2005-2012 [http://
perma.cc/9RMV-K7RY] (reporting “[o]f the 567 homicides from 2008 to 2011, for which the race of the victim is available in the SLMPD annual reports, 502 are listed as black, while 64 were white”).

Overall, the current two-year increase in homicide rates in cities may be driven or animated by factors that differ from the commonalities of the drug epidemics that coincided with the previous three cascading homicide epidemics. The evidence, then, points away from the concurrent-epidemic explanation for a homicide spike in the past two years.

D. Alternate Explanations for the 2015–2016 Trend

The concurrent-epidemic explanation for past homicide epidemics has its critics: Some researchers have minimized the drug-epidemic explanation and pointed to different drivers of the 1980s–1990s homicide epidemic. Their work might provide alternate explanations for the 2015–2016 trends. For example, Professors Philip Cook and John Laub cite the convergence of larger birth cohorts with increased gun availability to explain the spike in homicide in the previous era. 159 Philip J. Cook & John H. Laub, The Unprecedented Epidemic in Youth Violence, 24 Crime & Just. 27, 29, 58–59 (1998) [hereinafter, Cook & Laub, Unprecedented Epidemic] (noting negative correlation between youth population and rates of violent crime). However, in a later article, Professors Cook and Laub outright reject the birth-cohort explanation. Philip J. Cook & John H. Laub, After the Epidemic: Recent Trends in Youth Violence in the United States, 29 Crime & Just. 1, 23 (2002) [hereinafter Cook & Laub, After the Epidemic] (rejecting cohort effects as an explanation for the epidemic and noting contradictions in the drug–homicide nexus that mitigate but do not eliminate its explanatory power). This explanation seems to comport with the contours of the current epidemic, particularly in its concentration in cities and the elevated rates of victimization among young Black and Latino males. Professors Cook and Laub also note an increasing concentration in homicides among minority males in the past epidemic, a generally downward age trend in homicide offending, an increasing concentration of homicides in structurally disadvantaged urban areas, and an increasing share of homicides committed with guns. 160 See Cook & Laub, Unprecedented Epidemic, supra note 159, at 29 (connecting the availability of guns with rates of violence among youths). Again, these contours seem to reflect the recent spike in homicides.

Several studies also cited increased gun availability as an explanation for the previous epidemic. 161 See, e.g., Franklin E. Zimring & Gordon Hawkins, Crime Is Not the Problem: Lethal Violence in America 106–23 (1997) (discussing the role of guns in increasing death from violence); see also Anthony A. Braga, Serious Youth Gun Offenders and the Epidemic of Youth Violence in Boston, 19 J. Quantitative Criminology 33, 51–52 (2003) (finding that in Boston “violence that was highly concentrated among serious youth offenders who carried high capacity semiautomatic pistols and used their guns in an increasingly deadly way”). Professors Fagan, Franklin Zimring and June Kim characterized the homicide epidemic from 1986 to 1995 as a gun violence epidemic—rather than a homicide or violence epidemic—that afflicted young minority males disproportionately. 162 See, e.g., Jeffrey Fagan, Franklin E. Zimring & June Kim, Declining Homicide in New York City: A Tale of Two Trends, 88 J. Crim. L. & Criminology 1277, 1289, 1291–96 (1998). They showed that from 1985 to 1996, nearly all the increase and decline in homicide, assault, and other injury violence was the result of gun violence and that non–gun violence rates remained unchanged. 163 Id. at 1289–90. Professors Fagan and Garth Davies showed a similar pattern of rising and falling gun violence rates across neighborhoods from 1986 to 2000, characterizing some neighborhoods as experiencing an acute contagious epidemic, while others experienced rates that rose more slowly and declined more slowly, in a pattern suggesting a “slow epidemic.” 164 Jeffrey Fagan & Garth Davies, The Natural History of Neighborhood Violence, 20 J. Contemp. Crim. Just. 127, 137–42 (2004); see also Jeffrey Fagan, Crime and Neighborhood Change, in Nat’l Research Council, Understanding Crime Trends: Workshop Report 81, 82–83 (Arthur S. Goldberger & Richard Rosenfeld eds., 2008) (establishing a framework for understanding how “changing neighborhood contexts influence crime”). In the places with more acute epidemic patterns, neighborhood social structure suggested an ecological explanation for the differences in violence and injury epi­demics across neighborhoods over time. 165 See Fagan & Davies, supra note 164, at 138 (finding “neighborhood ecology is a significant predictor of total homicide, gun homicide, and homicides with African American victims” and “[i]t is not surprising that neighborhood disadvantage predicts homicide rates over time”). Today, there are similar patterns, with spatial concentration of homicides in cities including Chicago 166 Daniel Kay Hertz, The Debate over Crime Rates Is Ignoring the Metric that Matters the Most: ‘Murder Inequality,’ Trace (July 25, 2016), http://www.thetrace.org/
2016/07/crime-rates-american-cities-murder-inequality/ [http://perma.cc/7HHW-NDR8] (showing 2015 homicides by Chicago neighborhood).
and Milwaukee. 167 Id. (showing 2015 homicides by Milwaukee neighborhood).

Professor Alfred Blumstein cited increased firearm availability nationally as a driver of the 1985–1993 homicide epidemic, linking firearms with expanding drug-distribution networks that were centered (though not exclusively) in street-gang conflicts. 168 Blumstein, supra note 91, at 23–26. Others also focused on street gangs to explain the urban concentration of homicide in that era, but these studies also shifted attention from the gangs themselves to social structural changes that were predictive of gang formation and spread starting in the mid-1980s. 169 See, e.g., S.E. Costanza & Ronald Helms, Street Gangs and Aggregate Homicides: An Analysis of Effects During the 1990s Violent Crime Peak, 16 Homicide Stud. 280, 282 (2012) (establishing quantitative support for the assertion “that activities associated with expanding markets in drugs and guns, at least some of which are facilitated through the activities of local gangs, very likely contribute to aggregate patterns of lethal violence across urban landscapes”). Again, these factors are reflected in the current patterns in Chicago but not necessarily in other cities (e.g., Baltimore, St. Louis) that also experienced recent homicide spikes. 170 See Ford Fessenden & Haeyoun Park, Chicago’s Murder Problem, N.Y. Times (May 27, 2016), http://nyti.ms/2jR2kKO (on file with the Columbia Law Review) (discussing the increasing role gangs and guns play in Chicago’s violence). It is important to remember that Chicago’s two-year increase is an outlier in its magnitude due to the size of the city: Figure 3 shows that Chicago did experience an increase in homicides far greater than the nation’s two other largest cities, but Figure 4 shows that several midsize cities also experienced sharp homicide increases during the same period—including increases in the per capita homicide rates—and that their homicide rates exceeded Chicago’s during the same period. 171 Compare supra Figure 3, with supra Figure 4. Yet Chicago has gained political, media, and institutional notoriety when other cities have not. 172 See, e.g., Andrew Blake, ATF Looks to Expand Federal Presence in Chicago Amid Violent Crime Surge, Wash. Times (Feb. 4, 2017), http://www.washingtontimes.com/
shareaholic&utm_medium=twitter&utm_source=socialnetwork [http://perma.cc/62JX-9P6Y]; Monica Davey & Niraj Chokshi, Trump Threatens Federal Intervention in Chicago, Citing ‘Carnage,’ N.Y. Times (Jan. 25, 2017), http://nyti.ms/2kqK3FG (on file with the Columbia Law Review).
Chicago accounted for 272 of the 606 “excess” homicides in 2016 compared to 2015, or 44.8% of the homicide increases in sixty-three of the nation’s largest cities. 173 See Major Cities Chiefs Ass’n, Violent Crime Survey—Totals: Year End Comparison, January 1 to December 31, 2016 and 2015, at 1–2 (2017), http://www.
majorcitieschiefs.com/pdf/news/violent_crime_data_20162015_year_end_report.pdf [http://perma.cc/Z42X-7YQE].
Excluding Chicago and excluding the forty-nine deaths in the Pulse Night Club mass shooting in Orlando, Florida, 174 Ralph Ellis et al., Orlando Shooting: 49 Killed, Shooter Pledged ISIS Allegiance, CNN (June 13, 2016), http://cnn.it/1XgEczk [http://perma.cc/M8BJ-4NKK]. the increase across these sixty-three cities was 4.6%. 175 See Major Cities Chiefs Ass’n, supra note 173, at 1–3. This is a sizable increase but not the sort of systemic increase that forecasts an epidemic.

E. The Danger of Predictions

The ingredients of the homicide spikes of the past fifty years are either not present today or perhaps are less salient. We are in a period of uncertainty as to what happens this year and beyond. Certainly, the scope of major homicide increases, such as in Chicago, demands attention. But so too do homicide spikes in several other large (e.g., Detroit) and midsize (e.g., St. Louis, Baltimore, Milwaukee, Nashville) cities. The increases in 2016—Chicago notwithstanding—were concentrated in smaller cities, while homicides declined in several others. Figure 4 shows that some cities with large homicide increases in 2015 experienced homicide declines in 2016. Others stabilized. Whether these are random walks in a new period of instability in homicide rates, temporary blips in a longer historical trend of declining homicide rates, or harbingers of a new homicide epidemic is impossible to say. 176 ee, e.g., David F. Greenberg, Time Series Analysis of Crime Rates, 17 J. Quantitative Criminology 291, 300–301 (2001) (noting it “is unrealistic to suppose, however, that the same generating process responsible for temporal changes in crime will continue unchanged forever”).

Prediction is a risky business, and forecasting another homicide epidemic would be unwise and unfounded. Professors Cook and Laub conclude that the combination of volatility in violence rates and the difficulty of testing counterfactuals for each of the overlapping plausible explanations may require both long time horizons and complex research designs to understand the ebb and flow of violence epidemics, the differences across cities, the persons and neighborhoods affected, and the variation in duration and intensity from one place to the next and over different epochs. 177 Cook & Laub, After the Epidemic, supra note 159, at 30–31 (making such conclusions in the context of examining youth violence). We may have a better idea of how epidemics might launch, but those predictions are fraught with error. And once started, there are even greater risks of error in predicting their duration or severity. Figure 4 alone shows the number of momentary bumps in homicide rates that might have appeared to be the start of a new epidemic but that in fact were simply random perturbations in an overall pattern of decline.

Mistakes can have disastrous consequences: Predictions in the mid-1990s of a new epidemic of predatory violence by a new generation of “superpredators”—even as violence declined after a decade-long epidemic—were widely cited to justify significant and harsh changes in the state and federal law, 178 See, e.g., William J. Bennett, John J. Dilulio, Jr. & John P. Walters, Body Count 13–14 (1996) (explaining in 1996 that Americans have become “desensitized” to crime and that the dropping crime rates of the early 1990s might be “the lull before the coming crime storm”); see also John J. Dilulio Jr., The Coming of the Super-Predators, Wkly. Standard (Nov. 27, 1995), http://www.weeklystandard.com/the-coming-of-the-super-predators/article/8160#! [http://perma.cc/M3L5-4QWC]. The “superpredator” myth was translated into a forecast of the archetypical “symbolic assailant.” Jerome Skolnick identified this archetype as a police trope for Black males based on his research with police in the 1960s at the peak of one of the three homicide epidemics in the United States. See Jerome H. Skolnick, Justice Without Trial: Law Enforcement in a Democratic Society 45–48 (1966) (describing the “symbolic assailant” as “persons who use gesture, language, and attire that the policeman has come to recognize as a prelude to violence”). and led to profound but unnecessary changes in law and policy regarding youth violence. 179 In 1995 alone, nearly half the states changed their laws to facilitate the criminal prosecution of minors who were charged with any felony crime, including both violent and property crimes. Donna M. Bishop, Juvenile Offenders in the Adult Criminal System, 27 Crime & Just. 81, 84 (2000); Lara A. Bazelon, Note, Exploding the Superpredator Myth: Why Infancy Is the Preadolescent’s Best Defense in Juvenile Court, 75 N.Y.U. L. Rev. 159, 178 & n.101 (2000); see also Cook & Laub, After the Epidemic, supra note 159, at 20 (“Explanations that attributed rising violence rates to the character of the youths . . . proved influential with legislators, who in most states responded to the epidemic with more punitive policies for juvenile crime.”); Jeffrey Fagan, Juvenile Crime and Criminal Justice: Resolving Border Disputes, 18 Future Child. 81, 83 (2008) [hereinafter Fagan, Border Disputes] (finding “evidence shows that policies promoting transfer of adolescents from juvenile to criminal court” enacted in response to rising juvenile crime rates in the 1970s and 1980s “fail to deter crime among sanctioned juveniles and may even worsen public safety risks”). See generally Barry C. Feld, Bad Kids: Race and the Transformation of the Juvenile Court (1999) (examining the evolution of juvenile courts and the problematic shift toward rejecting the recognition of a need for separate kinds of treatment of minors in the juvenile justice system); Franklin E. Zimring, American Youth Violence: A Cautionary Tale, 42 Crime & Just. 265 (2013) (examining faulty assumptions that caused exaggerated warnings of future surges in youth violence). Soon enough, the sustained decline in homicide starting in the mid-1990s proved that prediction to be tragically flawed, leaving a statutory legacy with the perverse effect of increasing criminality among young offenders 180 Fagan, Border Disputes, supra note 179, at 84 (examining failure of transfer policies); Robert Hahn et al., Ctrs. for Disease Control & Prevention, Effects on Violence of Laws and Policies Facilitating the Transfer of Youth from the Juvenile to the Adult Justice System: A Report on Recommendations of the Task Force on Community Preventive Services, Morbidity & Mortality Wkly. Rep.: Recommendations & Rep., Nov. 30, 2007, at 1, 9, http://www.cdc.gov/mmwr/pdf/rr/rr5609.pdf [http://perma.cc/Q9Y2-Y36V] (“To the extent that transfer policies are implemented to reduce violent or other criminal behavior, available evidence indicates that they do more harm than good.”); Lonn Lanza-Kaduce et al., Juvenile Offenders and Adult Felony Recidivism: The Impact of Transfer, 28 J. Crime & Just. 59, 60 (2005) (“Contrary to the expectations of the proponents of ‘get-tough’ policies, previous research generally found that transfer is related to higher rather than lower rates of recidivism.”); Franklin E. Zimring & Stephen Rushin, Did Changes in Juvenile Sanctions Reduce Juvenile Crime Rates? A Natural Experiment, 11 Ohio St. J. Crim. L. 57, 69 (2013) (producing statistical analysis revealing “changes in juvenile homicide rates were highly similar to changes in young adult homicide rates, despite the national wave of legislation targeted at juvenile offenders,” which “undercuts” the causal claim between the decrease and such statutory changes). while aggravating racial disparities. 181 See Feld, supra note 179, at 5 (noting “punitive juvenile justice policies impose harsh sanctions disproportionately on minority youths and foster the growing procedural and substantive convergence between juvenile and criminal courts”); Donna M. Bishop & Charles E. Frazier, Race Effects in Juvenile Justice Decision-Making: Findings of a Statewide Analysis, 86 J. Crim. L. & Criminology 392, 412 (1996) (“The quantitative analyses demonstrate appreciable effects of race on delinquency case processing that disadvantage minority offenders.”).

The risks of a false prediction stand in tension with the demands for public safety: Police executives do not have the luxury of waiting to see whether homicide spikes will end with a reversion to the mean. At risk on both sides of this equation is the trust of communities in criminal legal institutions, especially the police. And with trust comes legitimacy and partnership with legal actors in the coproduction of security. 182 Tom Tyler & Yuen J. Ho, Trust in the Law: Encouraging Cooperation with the Police and the Courts 65–67 (2002) (asserting “trust in particular individuals that develops by virtue of their actions . . . lead[s] to inferences that the decisions of the authorities should be accepted”); Tom R. Tyler et al., Street Stops and Police Legitimacy: Teachable Moments in Young Urban Men’s Legal Socialization, 11 J. Empirical Legal Stud. 751, 775 (2014) [hereinafter Tyler et al., Teachable Moments] (demonstrating “higher legitimacy is related to lower levels of criminal behavior and also demonstrat[ing] that cooperation with the police is greater when legitimacy is high”). Just as fair and legal treatment promotes trust, so too does effective policing in ensuring public safety. These are the immediate challenges for police, communities, and government. We turn next to a framework and principles to meet these challenges.

III. The Ecology of Policing, Cynicism, and Violence

Even as crime rates declined for two decades, tensions between citizens and police rose, especially among African Americans 183 Ronald Weitzer & Steven A. Tuch, Race and Policing in America: Conflict and Reform 70–88 (2006) (presenting evidence of perceptions of racially biased policing and the impact on minority communities); Lawrence D. Bobo & Devon Johnson, A Taste for Punishment: Black and White Americans’ Views on the Death Penalty and the War on Drugs, 1 Du Bois Rev. 151, 170 (2004) (noting changes in the criminal justice system have taken a “heavily disproportionate toll on Black communities, especially low-income Black communities”); David A. Harris, The Stories, the Statistics, and the Law: Why Driving While Black Matters, 84 Minn. L. Rev. 265, 269 (1999) (establishing the problem of “driving while black”). and, to a lesser extent, among Latinos. 184 Mark Hugo Lopez & Gretchen Livingston, Pew Hispanic Research Ctr., Hispanics and the Criminal Justice System: Low Confidence, High Exposure 1 (2009), http://www.
criminal_justice_system_2009.pdf [http://perma.cc/4MUK-SP7C] (“At a time when Latinos are interacting more than ever with police, courts and prisons, their confidence in the U.S. criminal justice system is closer to the low levels expressed by blacks than to the high levels expressed by whites.”).
In the 1990s, at the outset of the homicide and general crime decline, much of the distrust was focused on racial profiling by police. 185 See generally R. Richard Banks, Beyond Profiling: Race, Policing and the Drug War, 56 Stan. L. Rev. 571 (2003) (noting the existing “focus on racial profiling” and arguing that “policy analyses should consider the race-related consequences of the drug war, without regard to whether officers engage in racial profiling”); Harris, supra note 183, at 277–88 (surveying the statistical analysis of racial distribution of traffic stops in New Jersey, Maryland, and Ohio); Anthony Thompson, Stopping the Usual Suspects: Race and the Fourth Amendment, 74 N.Y.U. L. Rev. 956, 956–59 (1999) (noting “[r]ecent studies support” that “the police target people of color, particularly African Americans, for stops and frisks”). Over time, minority citizens’ distrust of police expanded to include police use of force 186 Steven A. Tuch & Ronald Weitzer, Racial Differences in Attitudes Toward Police, 61 Pub. Opinion Q. 642, 647 (1997) (showing stronger negative reactions by Blacks compared to Whites in responses to highly publicized incidents of police use of force). and, later, everyday policing of disorder. 187 Kirk & Papachristos, supra note 84, at 1198 (noting in “socially and economically disadvantaged neighborhoods, people come to understand that the [police] will offer them little in the way of security, either economic or personal”); Tyler et al., Teachable Moments, supra note 182, at 771 (noting “adverse effects of an increasing number of police stops on perceptions of fairness and legality, both in terms of views about personal experience and with general judgments about police behavior”). But as these tensions became more salient in minority communities, police were less likely to see their use of force as unjustified. 188 David Weisburd et al., Nat’l Inst. of Justice, Police Attitudes Toward Abuse of Authority: Findings from a National Study 5 (2000) http://www.ncjrs.gov/pdffiles1/
nij/181312.pdf [http://perma.cc/5W2U-QL7C]; see also Morin et al., supra note 23, at 75–76.
Even as citizen distrust of police differed along racial lines, there still was broad support for use of force in intimate police cultures and subcultures. 189 See, e.g., Anthony J. Micucci & Ian M. Gomme, American Police and Subcultural Support for the Use of Excessive Force, 33 J. Crim. Just. 487, 489–90 (2005) (discussing “informal standards” that “help to explain how and why police violence is extended from its legitimate to its illegitimate use”). Today, that support is racialized within police departments: White and Latino police officers tend to view fatal encounters between Blacks and police as isolated incidents in contrast to the views of a majority of Black officers. 190 Morin et al., supra note 23, at 16–17.

In the past decade, these tensions were exacerbated by highly publicized police shootings of citizens—shootings that were often captured on video and replayed on television and the Internet. Citizens expressed deep anger over these incidents, often in political demonstrations. Police pushed back politically to those demonstrations, and there were signs of a police withdrawal in the face of criticism and citizen anger. 191 ee supra section I.A. These exchanges only magnified the distrust between citizens and police. These incidents tapped into Black citizens’ longstanding distrust of the police that was evident for well over a decade before the current increase in homicides and remains acute today. 192 See Malcolm D. Holmes & Brad W. Smith, Race and Police Brutality: Roots of an Urban Dilemma 2–6 (2008) (citing research from the 1960s to the present on distrust of the police among African Americans); Devon Johnson & Joseph B. Kuhns, Striking Out: Race and Support for Police Use of Force, 26 Just. Q. 592, 597 (2009) (citing research reporting race is strongest predictor of support for use of force and that Whites are more likely to support use of force than Blacks); Jason Sunshine & Tom R. Tyler, The Role of Procedural Justice and Legitimacy in Shaping Public Support for Policing, 37 Law & Soc’y Rev. 513, 515 (2003) (citing studies revealing minorities hold “much more negative attitudes about the police” and “have lower trust and confidence in institutions of social control”); Brian L. Thompson & James Daniel Lee, Jr., Who Cares if Police Become Violent? Explaining Approval of Police Use of Force Using a National Sample, 74 Soc. Inquiry 381, 383–84 (2004) (citing studies finding minorities “generally disapprove of police violence” and “feel the criminal justice system is biased”); Ronald Weitzer & Steven A. Tuch, Race and Perceptions of Police Misconduct, 51 Soc. Probs. 305, 305 (2004) (“[R]ace is one of the most salient predictors of attitudes toward the police and other criminal justice institutions: blacks are more likely than whites to express dissatisfaction with various aspects of policing.”). Anger and distrust for police grew alongside not just incidents of police violence toward citizens but police programs that encouraged investigative stops and arrests and summonses for minor crimes and violations—tactics that disproportionately affected Black and Latino citizens. 193 Jeffrey Fagan et al., Stops and Stares: Street Stops, Surveillance and Race in the New Policing, 43 Fordham Urb. L.J. (forthcoming 2017) (manuscript at 21–22) (on file with the Columbia Law Review) (discussing the disproportionate racial impact of increased investigative stops, searches, and arrests); see also Jeffrey Fagan & Tracey L. Meares, Punishment, Deterrence and Social Control: The Paradox of Punishment in Minority Communities, 6 Ohio St. J. Crim. L. 173, 219–20 (2008) (describing the rise in order-maintenance policing and its disproportionate effect on Black and Latino citizens); Tyler et al., Teachable Moments, supra note 182, at 755–56 (discussing the rise of Terry stops and “proactive policing” and identifying the practices’ disproportionate effect on minorities).

The distrust today is a two-way street. Police regard protests of high-visibility police shootings of citizens as motivated less by the perceived injustice of those shootings and more by bias against police. 194 Morin et al., supra note 23, at 17–18 (finding 92% of police officers believe that bias against the police is at least partially motivating these protests). Only about one in three officers thinks protesters are sincerely motivated by a desire to hold officers accountable. 195 Id. This may result in what this Essay previously touched on—signs of a police pullback or withdrawal in the face of citizen and political criticism, extensive video surveillance and publicity of contested police actions, and even killings of police. 196 See supra Part I. In other words, it could lead to de-policing. In Chicago, this was evident in the form of sharply reduced Terry stops following the release of a police video of the shooting of Laquan McDonald, an unarmed teenager who was fleeing from police. 197 Kapustin et al., supra note 9, at 24 & n.24 (suggesting the video footage of the police shooting of Laquan McDonald was partially responsible for an 80% decline in Terry stops in Chicago). In Baltimore, arrests for less-serious felonies and misdemeanors declined substantially in the period between the August 2014 demonstrations over the police killing of Michael Brown in Ferguson, Missouri, and the April 2015 death of Freddie Gray in Baltimore. 198 tephen L. Morgan & Joel A. Pally, Ferguson, Gray, and Davis: An Analysis of Recorded Crime Incidents and Arrests in Baltimore City, March 2010 Through December 2015, at 48–49 (2016), http://socweb.soc.jhu.edu/faculty/morgan/papers/
MorganPally2016.pdf [http://perma.cc/NCQ4-QLDR].
At the same time, homicides in Baltimore increased sharply during the police slowdown, 199 Id. at 67. but then declined in 2016 as arrests resumed. 200 Stephen L. Morgan, Fall 2016 Update for Ferguson, Gray, and Davis: An Analysis of Recorded Crime Incidents and Arrests in Baltimore City, March 2010 through December 2015 October 2016, at 1 (2016), http://socweb.soc.jhu.edu/faculty/
morgan/papers/MorganPally2016FallUpdate.pdf [http://perma.cc/5QUB-HTD6].
While police officers describe a general withdrawal by police, they offer few specifics as to their own or their colleagues’ actions. 201 Morin et al., supra note 23, at 4. But see id. at 45 (finding more officers worry about colleagues spending too much time diagnosing before acting on a situation). Clearly, there is evidence of a pullback in Chicago and Baltimore, but to date, there is no support beyond anecdotes for claims of a general withdrawal. 202 See Mac Donald, supra note 79 (offering anecdotal evidence to support the conclusion that “proactive policing is down”); Rosenfeld, supra note 79 at 18–23 (arguing more empirical evidence is needed to analyze the legitimacy of a “Ferguson effect” that has caused police to supposedly “disengage from their duties”).

Although it is unclear whether any increase in the homicide rate can be attributed to a police pullback, the pullback is part of a loosely coupled system of distrust and resentment between citizens and police that entwines violence, cynicism, and public safety into a complex and tangled ecology. For us, the language of a “chill wind” adopted by FBI Director Comey 203 See Comey, Remarks at University of Chicago, supra note 79 (describing a “chill wind blowing through American law enforcement”). is meant to capture the connections in these mechanisms and their consequences for urban homicide. We turn next to how this tangled ecology plays out in the enduring homicide problems of several U.S. cities.

A. Case Studies: Cynicism, Withdrawal, and the Persistence of Homicides

In several cities, people living in neighborhoods where homicide is a recurring reality share a strikingly similar narrative about policing, community, and murder. In this narrative, murders are not uncommon but remain unsolved, citizens experience policing as detached from serious crime and aimed at the wrong behaviors and the wrong people, they see police as indifferent or disrespectful if not abusive, and they are unwilling to cooperate in murder investigations by the police, whom they view as an “occupation force.” These interlocking tensions reinforce the social and economic isolation of places that already have the features of a “poverty trap.” 204 See Bowels et al., supra note 83, at 3–5 (describing “poverty traps”). Residents feel not only unsafe but unfairly treated.

In this web of recurring social interactions, these perceptions—a variety of police insults and recurring episodes of murder and other violence––engender frustration and anger and, in turn, withdrawal from cooperation with the police. 205 See Monica C. Bell, Police Reform & the Dismantling of Legal Estrangement, 126 Yale L.J. (forthcoming 2017) (manuscript at 7–8) (on file with the Columbia Law Review) (explaining the concept of “legal estrangement”); Matthew Desmond, Andrew V. Papachristos & David S. Kirk, Police Violence and Citizen Crime Reporting in the Black Community, 81 Am. Soc. Rev. 857, 858 (2016) [hereinafter Desmond et al., Police Violence and Citizen Crime Reporting] (describing “legal cynicism” and the resulting withdrawal from the police); David S. Kirk & Mauri Matsuda, Legal Cynicism, Collective Efficacy, and the Ecology of Arrest, 49 Criminology 443, 444 (2011) (explaining “legal cynicism” as a “cultural frame in which the law and the agents of its enforcement are viewed as illegitimate, unresponsive, and ill equipped to ensure public safety”); Robert J. Sampson & Dawn Jeglum Bartusch, Legal Cynicism and (Subcultural?) Tolerance of Deviance: The Neighborhood Context of Racial Differences, 32 Law & Soc’y Rev. 777, 783 (1998) (“[I]nner-city contexts of racial segregation and concentrated disadvantage, where inability to influence the structures of power that constrain lives is greatest, also breed cynicism and perceptions of legal injustice.”). Citizens see the police as indifferent to their concerns, raising dignitarian issues that deepen the sense of isolation and powerlessness before the police. 206 I. Bennet Capers, Policing, Race, and Place, 44 Harv. C.R.-C.L. L. Rev. 43, 68–69 (2009) (stating similar claims in terms of “public shaming”). And as we have noted, criticisms of the police—who also feel isolated by very salient incidents of police deaths—foster resentments that match the resentments of those living in conditions of simultaneous over- and underpolicing.

1. Los Angeles. — In April 2016, L.A. police commanders deployed a surge in South L.A. to counter an uptick in the homicide rate in the area. 207 Cindy Chang, The LAPD’s Biggest Conundrum: How to Suppress Crime Without Alienating South L.A.’s Black Residents, L.A. Times (Feb. 4, 2017), http://www.latimes.com/local/california/la-me-lapd-crime-trust-20170203-story.html (on file with the Columbia Law Review) [hereinafter Chang, Biggest Conundrum]. Nearly half the violent gun crimes in the city take place in South L.A., and the homicide rate had been rising citywide for the past three years. 208 See id.; see also supra Figures 3, 4. Officers from the elite Los Angeles Police Department Metropolitan Division flooded the area—a predominantly Black and Latino part of the city—and intensified their stop activity to search for weapons, scofflaws, or fugitives. 209 Chang, Biggest Conundrum, supra note 207; see also Jeffrey Fagan & Elliott Ash, New Policing, New Segregation: From Ferguson to New York, 105 Geo. L.J. Online 16 (forthcoming 2017) (manuscript at 4–5) (on file with the Columbia Law Review) (discussing “broken windows” strategy); Christopher Slobogin, Government Dragnets, 73 Law & Contemp. Probs., Summer 2010, at 107, 108–10 (discussing police dragnets). They stopped drivers for infractions such as tinted windows or broken taillights as a pretext to search for guns or identify persons who were wanted on outstanding warrants. 210 Chang, Biggest Conundrum, supra note 207. The surge netted 300 illegal firearms in six months. 211 Id.; see also Cindy Chang, LAPD Reassigns Elite Cops After Killings Fall in South L.A. but Crime Jumps in Other Parts of L.A., L.A. Times (Oct. 1, 2016), http://www.latimes.com/local/lanow/la-me-lapd-crime-redeploy-20161001-snap-story.html (on file with the Columbia Law Review) (discussing the results of the surge).

While the homicide rate declined in South L.A. during the surge, police tactics alienated Black residents. A 2016 survey conducted for the Los Angeles Police Commission—after the homicide rate in South L.A. had declined to the same levels as a year earlier—showed deep distrust and anger toward the police among African American residents, leading to a “‘profoundly serious disconnect’” between the LAPD and the city’s Black citizens. 212 Kate Mather & Cindy Chang, LAPD Watchdog Takes a Long Look into Allegations of Racial Profiling, L.A. Times (Nov. 15, 2016), http://fw.to/WjQB4oY (on file with the Columbia Law Review). Black respondents to the survey complained of repeated stops, disrespect from police, and the types of “petty indignit[ies]” and baseless police stops that the Terry court feared 213 See Terry v. Ohio, 392 U.S. 1, 16–17 (1968); see also Bell, supra note 205, at 56–57 (noting “police cases more than others send messages about social inclusion” and that programs like the systematic stop-and-frisks at issue in Terry can result in “legal estrangement”); Josh Bowers, Probable Cause, Constitutional Reasonableness, and the Unrecognized Point of a “Pointless Indignity,” 66 Stan. L. Rev. 987, 1008 (2014) (noting the Terry Court’s fear that many frisks are “wholly unrelated to a desire to prosecute for crime” (alteration in original) (quoting Terry, 392 U.S. at 13)); Capers, supra note 206, at 64 (noting Terry’s implicit acknowledgement of the disproportionate impact of its decision on disadvantaged communities); William J. Stuntz, Terry’s Impossibility, 72 St. John’s L. Rev. 1213, 1218 (1998) (describing four distinct harms—privacy harm, targeting harm, discrimination, and physical harm—suffered by victims of unreasonable police stops). The reactions of Blacks in South L.A. to being stopped mirrors the sense of racial targeting that Professor Bennett Capers discusses in his research into race-based public shaming. See Capers, supra note 206, at 68–69. but whose pretextual basis the Court sanctioned in United States v. Whren. 214 See Chang, Biggest Conundrum, supra note 207; see also Whren v. United States, 517 U.S. 806, 812 (1996) (noting the Supreme Court has “never held, outside of inventory search or administrative inspection . . . that an officer’s motive invalidates objectively justifiable behavior under the Fourth Amendment”).

The other source of alienation and cynicism of the residents in South L.A. was the inability of police to “clear” a homicide, that is, to make an arrest in that case. 215 See Jill Leovy, Ghettoside: A True Story of Murder in America 48–49 (2015) (discussing clearance rates in L.A. County and South L.A.). Clearance rates in L.A. citywide were below the national average of 62.5% for the period 2009 to 2012 and averaged 38% for the period from 1994 to 2006. 216 See Cal. Dep’t of Justice, Crimes and Clearance, OpenJustice, http://openjustice.doj.ca.gov/crime-statistics/crimes-clearances (on file with the Columbia Law Review) (last visited Mar. 2, 2017) (providing data for clearance rates in L.A. County through 2015). Current data show that homicide clearance rates improved from 60.2% in 2012 to 68.1% in 2015. 217 Id. The most recent data are from 2015. Id.; see also L.A. Police Comm’n, Investigation of the Los Angeles Police Department’s Homicide Clearance Rates 4 (2015), http://www.lapdpolicecom.lacity.org/072115/BPC_15-0214.pdf [http://perma.cc/5UFT-4J38]. Still, homicide clearance rates vary by police district and seem to have remained stubbornly low for South L.A. for over two decades. 218 See Chang, Biggest Conundrum, supra note 207; see also Leovy, supra note 215, at 48–49; Tammy Rinehart Kochel, Applying Police Legitimacy, Cooperation, and Collective Security Hypotheses to Explain Collective Efficacy and Violence Across Neighborhoods, Int’l J. Comp. & Applied Crim. Just., 2017, at 1, 14 (finding that police competence in maintaining public safety in neighborhoods encourages citizens to cooperate with police and engage in their own efforts independent of police to “co-produce social order”); Nicholas Peterson, Neighborhood Context and Unsolved Murders: The Social Ecology of Homicide Investigations, 27 Policing & Society 372, 383 (2015) (showing that, for all homicides in Los Angeles between 1990 and 1994, the odds of clearing a homicide decline by 50% for each 1% increase in the Black population of a neighborhood, and decline by 31% for each 1% increase in the Latino population of a neighborhood). Residents there interpreted the low clearance rate as indifference to Black homicides, one part of the L.A. Black community’s belief that police favor White crime victims. 219 See Mather & Chang, supra note 212 (describing stark racial disparities in L.A. residents’ perceptions of the police force’s trustworthiness). The distrust and anger of citizens in Los Angeles toward the police has deepened even further this year. In a 2017 survey by Loyola Marymount University, researchers reported that nearly 7 out of 10 respondents between ages eighteen and twenty-nine predict that a repeat of the 1992 Rodney King riots will happen within the next five years, an increase of 33% from the 2012 survey. Fernando Guerra, Brianne Gilbert & Berto Solis, LA Riots 25 Years Later: 2017 Los Angeles Public Opinion Survey Report 3 (2017), http://lmu.app.box.com/s/
it0e448kjmhz2jg1s39eht0g8ye63yhr (on file with the Columbia Law Review).

The low homicide clearance rate “left killers at large and families of murder victims without answers.” 220 Mike Reicher, LAPD Closed Homicide Cases Without Bringing Killers to Justice, Analysis Shows, L.A. Daily News (Jan. 24, 2015), http://www.dailynews.com/general-news/20150124/lapd-closed-homicide-cases-without-bringing-killers-to-justice-analysis-shows [http://perma.cc/PW7F-AAD5]. In this setting, a dragnet of intrusive stops—simply harassment to many residents of South L.A. 221 See Chang, Biggest Conundrum, supra note 207 (depicting South L.A. residents’ perception of police stops as harassment). —produced little offset in justice and led to alienation and cynicism that, along with fear of retribution from local crime groups and offenders, deterred residents in the neighborhood from helping police to solve murders. 222 See Leovy, supra note 215, at 74–86 (discussing reluctance of witnesses to testify and witness intimidation). Furthermore, more experienced and capable detectives sought out “better” assignments in wealthier neighborhoods, avoiding the Black neighborhoods most in need, which compounded and reinforced the dynamics of unsolved murders and cynicism among local residents. 223 Id. at 26.

2. New York City. — New York City has become a safe city. In 2016, the number of shootings dropped by 12% to 998; this was the first time that shootings fell below 1,000 since shooting data were compiled. 224 Jen Kirby, New York City Had a Record-Low Crime Rate in 2016—But that’s Not the Story in Other Cities, N.Y. Mag. (Jan. 4, 2017), http://nymag.com/daily/
intelligencer/2017/01/new-york-city-had-record-low-crime-rate-in-2016.html [http://
Murders dropped from 352 to 335 in the same period, reversing a one-year increase in 2015. Overall, the total crime rate dropped by 4%, with a yearly reduction in almost every major category of crime, including major felonies. 225 NYC Crime Falls to Lowest Levels in Decades: NYPD, NBC N.Y. (Jan. 4, 2017), http://www.nbcnewyork.com/news/local/NYC-Crime-Falls-to-Lowest-Levels-in-Decades-NYPD-409660785.html [http://perma.cc/JW4Q-P7K4] (quoting NYPD statistics). This downward trend continues a decline that began in 1992 and has continued nearly uninterrupted for twenty-five years. 226 Franklin E. Zimring, The City that Became Safe: New York’s Lessons for Urban Crime and Its Control 5–7 (2012); see also supra Figures 3, 4. Still, crime rose by 2.4% in the NYPD Housing Bureau, 227 NYC Crime Falls to Lowest Levels in Decades: NYPD, supra note 225. The Housing Bureau is the division of the New York City Police Department that patrols New York City Public Housing Authority developments. See Housing Bureau, NYPD, http://www.nyc.gov/html/nypd/html/housing_bureau/housing.shtml [http://perma.cc/7EK7-V6QD] (last visited Feb. 11, 2017). which is significant because it is mainly in the city’s housing projects that a complex ecology of policing, poverty, cynicism, and violence churns. 228 Jeffrey Fagan et al., Crime in Public Housing: Clarifying Research Issues, Nat’l Inst. Just. J., Mar. 1998, at 2, 2–3, http://www.ncjrs.gov/pdffiles/jr000235.pdf [http://perma.cc/KF94-3NLR] (noting much contemporary public housing today “suffers from poor design, deteriorating structures, inadequate funding, a large concentration of poor people, isolation from social services, high crime rates, and cumbersome management”); Adam Carlis, Note, The Illegality of Vertical Patrols, 109 Colum. L. Rev. 2002, 2002–03 (2009) (noting an “intractable fear of the projects” has led to aggressive policing strategies within those communities).

As in L.A., the highest homicide area in New York City, the South Bronx, has had the fewest detectives assigned per violent crime. 229 Benjamin Mueller & Al Baker, Rift Between Officers and Residents as Killings Persist in South Bronx, N.Y. Times (Dec. 31, 2016), http://nyti.ms/2jVye66 (on file with the Columbia Law Review) [hereinafter Mueller & Baker, Rift Between Officers and Residents]. Among the NYPD’s seventy-seven precincts, nine of the fourteen lowest-staffed detective squads are in the Bronx. Id. The underpolicing in this police precinct is common to the rest of the Bronx, an area with a population of 1,455,444 in 2015 230 Quick Facts, U.S. Census Bureau, Quick Facts, http://www.census.gov/
quickfacts/table/PST045216/36005,36 [http://perma.cc/D9B2-W85T] (last visited Feb. 11, 2017).
and the city’s highest per capita murder rate. 231 See Major Crime in New York City, 2009–2015, Newsday, http://data.newsday.
com/long-island/data/crime/new-york-city-crime-rate/ [http://perma.cc/NA9K-3F4J] (last updated Aug. 31, 2016). Although there were fifty-five more murders in Brooklyn (Kings County) in 2015 (146) than in the Bronx (91), id., the population in Brooklyn is 2,636,735, nearly twice that of the Bronx. See Kings County, NY, Census Rep., http://censusreporter.org/profiles/05000US36047-kings-county-ny/ [http://perma.cc/
MPH9-YACL] (last visited Feb. 11, 2017).
Journalists covering the Bronx quote citizens’ reports that their crime complaints are ignored by the police, their 911 calls are unanswered for hours, and some 911 calls end up with an arrest of the caller. Those arrests in turn can lead to detention for hours in a “fetid holding cell” in a police precinct for a minor offense. 232 Mueller & Baker, Rift Between Officers and Residents, supra note 229.

At the same time, aggressive policing tactics—stops and frisks, misdemeanor arrests, and summons activities—that characterized the NYPD’s policing model for over two decades have been intensively applied in the Bronx, particularly in the special “Impact Zones” that include public housing developments. 233 See John MacDonald, Jeffrey Fagan & Amanda Geller, The Effects of Local Police Surges on Crime and Arrests in New York City, PLOS One 1–3 (2016), http://journals.
plos.org/plosone/article/file?id=10.1371/journal.pone.0157223&type=printable [http://
perma.cc/UF7U-Y8TD]; see also Fagan & Davies, supra note 164, at 142 (warning that focusing on specific neighborhoods may “mask potentially important cultural dynamics from the broader social environment that bear on the natural history of violence”); Carlis, supra note 228, at 2009–11 (describing how “finding an area to be ‘high crime’” is used to justify more aggressive policing tactics).
Courts have cited constitutional problems in those practices, ordered remedies, and awarded damages. 234 New York City recently agreed to terminate the practice of vertical patrols and trespass abatement in both public housing and subsidized housing, stipulating to Fourth and Fourteenth Amendment violations in this practice. See Stinson v. City of New York, 282 F.R.D. 360, 383–84 (S.D.N.Y. 2012) (granting class certification for a § 1983 claim against the city and police officers for unconstitutional summonses); Stipulation of Settlement and Order at 7–11, Ligon v. City of New York, 12 Civ. 2274 (AT) (S.D.N.Y. Feb. 2, 2017), http://www.nyclu.org/sites/default/files/field_documents/ligon_settlement.pdf (on file with the Columbia Law Review) [hereinafter Ligon Stipulation] (describing parties’ mutual agreement that enforcement of New York City’s trespass law must conform with constitutional principles); Settlement Stipulation of Settlement at 6–9, Davis v. City of New York, 10 Civ. 0699 (SAS) (S.D.N.Y. Jan. 7, 2015), http://www.naacpldf.org/files/case_
issue/Davis%20v.%20City%20of%20New%20York-Stipulation%20of%20Settlement.pdf [http://perma/cc/2CP3-STNC] (agreeing to changes in the NYPD’s trespass enforcement policies in New York City public housing); see also Floyd I, 959 F. Supp. 2d 540, 658–67 (S.D.N.Y. 2013) (finding New York City liable for violating plaintiffs’ Fourth and Fourteenth Amendment rights with the stop-and-frisk policy). New York City agreed to settle the Stinson case and pay $75 million to plaintiffs. Benjamin Weiser, New York City to Pay Up to $75 Million over Dismissed Summonses, N.Y. Times (Jan. 23, 2017), http://nyti.ms/2jR44V5 (on file with the Columbia Law Review) (describing settlement for over 900,000 summonses, issued primarily in minority communities, found to be without legal justification).
The courts have found that few of those sanctions—stops, arrests, summonses—result in convictions or enforced summonses, as noted in the settlement in of those lawsuits. 235 See Floyd v. City of New York (Floyd II), 959 F. Supp. 2d 691, 692 (S.D.N.Y. 2013) (recognizing need for court monitoring of summonses); Stinson, 282 F.R.D. at 383–84 (upholding certification of a § 1983 class action alleging a policy of facially invalid summonses); Ligon Stipulation, supra note 234, at 7–11 (describing how the NYPD will change its enforcement policies in and around buildings in the Trespass Affidavit Program); see also Issa Kohler-Hausmann, Managerial Justice and Mass Misdemeanors, 66 Stan. L. Rev. 611, 690–91 (2014) (explaining the “transition from nonconviction dispositions to convictions, for the overwhelming majority of misdemeanor cases, [is] largely a factor of the temporal proximity of arrests” and “likelihood of arrest is a function of not only an individual’s conduct, but also of policing practices”). Those that lead to arraignment often take years to resolve and burden defendants with repeated court appearances, lost days of work, childcare complications, days in detention, and a minor criminal record. 236 William Glaberson, Faltering Courts, Mired in Delays, N.Y. Times (Apr. 13, 2013), http://nyti.ms/ZmOE6s (on file with the Columbia Law Review). The South Bronx precinct, known as the 4-0 (the 40th precinct), had the second-highest number of civilian complaints for officer misconduct (verbal abuse, physical abuse, and inaction in the face of citizen crime reports) in 2016. 237 See Complaint Activity Map, N.Y.C. Civilian Complaint Review Bd., http://www1.nyc.gov/site/ccrb/policy/complaint-activity-map.page [http://perma.cc/56PQ-G8SU] (last visited Feb. 11, 2017) (displaying an interactive map showing 111 complaints received from the 40th precinct in 2016). The Bronx also has the five precincts with the highest number of personal injury court claims alleging police misconduct adjusted for the local crime rate. 238 See Office of the N.Y.C. Comptroller, CLAIMSTAT: Protecting Citizens and Saving Taxpayer Dollars 2 (2014), http://comptroller.nyc.gov/wp-content/uploads/documents/
ClaimStat.pdf [http://perma.cc/92RF-MS55]; Marc G. Peters & Philip K. Eure, N.Y.C. Dep’t of Investigations, Using Data from Lawsuits and Legal Claims Involving NYPD to Improve Policing 2 (2015), http://www1.nyc.gov/assets/oignypd/downloads/pdf/2015-04-20-litigation-data-report.pdf [http://perma.cc/MX9J-PGJP].

A troubled court system that compounds the harms of petty police enforcement, underpolicing of murder and violent crime, perceived injustices of overpolicing of minor crimes with only tenuous connections to major crimes, and perceived harassment and fear of violence has bred cynicism toward police, which leads Bronx citizens to withdraw from cooperation with police to solve crimes. 239 Mueller & Baker, Rift Between Officers and Residents, supra note 229 (noting “[d]etectives’ phones barely ring with tips”). As in L.A., citizens fear retribution if they cooperate with police, and police are unable to provide protection to break that particular chain; residents actively intervene to discourage others from reporting crimes or cooperating with the police. 240 Id. (detailing how fear of retribution stops citizens from reporting serious crimes). They carry the resentments of the policing model as a negative incentive—multiplied by fear of retaliation—and choose not to help the police. For example, when a young Black mother was killed by a stray bullet in 2016, Bronx detectives had no responses to a reward of $2,500 posted for tips, a fraction of the reward offered for leads in the killing of a White jogger in a Queens precinct earlier in the year. 241 Benjamin Mueller & Al Baker, A Mother Is Shot Dead on a Playground, and a Sea of Witnesses Goes Silent, N.Y. Times (Oct. 7, 2016), http://nyti.ms/2k82BWG (on file with the Columbia Law Review) [hereinafter Mueller & Baker, Mother Shot Dead].

In New York City, as in L.A., perceived police injustices and indifference, persistent and unresolved cases of violence, and residents’ cynicism and withdrawal are all implicated in the neighborhoods where murder persists even as the rest of the two cities become safer. This is not to say that the persistence of violent and drug crimes, often carried out by local gangs tied to specific housing developments, are not part of that ecology. 242 Beth Kormanik, Threat from Gangs Heightens Anxiety in a Brooklyn Neighborhood, N.Y. Times, (Jan. 20, 2012), http://nyti.ms/2m6bzJJ (on file with the Columbia Law Review) (discussing gang issues in a Brooklyn neighborhood and relationship to crime); Mueller & Baker, Mother Shot Dead, supra note 241 (describing the murder of Jessica White as a terrible byproduct of a dispute between two drug dealers in her public housing development). But the disconnect between citizens and police removes what might otherwise be a social process to undo the cycle.

There are signs of progress. To its credit—perhaps spurred into action by a New York Times series highlighting the inadequacy of investigative resources in the South Bronx—the NYPD recently announced plans to substantially reinforce the ranks of homicide investigators there. 243 See Benjamin Mueller & Al Baker, Police to Strengthen Force in New York’s Most Violent Borough, N.Y. Times (Feb. 8, 2017), http://nyti.ms/2kQztI2 (on file with the Columbia Law Review) (noting the NYPD “plans to send 75 new investigators to the Bronx to address a steep and longstanding shortage of detectives in the department’s most violent and underserved borough”).

3. Chicago. — The homicide spike in Chicago since 2014—unprecedented in the city’s last half century 244 See supra Figures 3 & 4. —has coincided with revelations of serious institutional problems with the police department. Both a citizen task force in Chicago and the DOJ CRD’s recent investigations of the Chicago Police Department (CPD) revealed systemic problems of excessive force, deadly force, inadequate systems of accountability, unlawful detentions, and the absence of transparency in disciplinary practices for officer misconduct. As in New York and L.A., important commands in the city were understaffed. The two reports also note low morale among officers, complaints about insufficient training and support, and sharp criticism by the public. 245 Police Accountability Task Force, supra note 10, at 6–13 (describing findings); U.S. Dep’t of Justice, Investigation of the Chicago Police Department, supra note 10, at 7–12 (summarizing the prevalence of these problems in the investigation into Chicago’s police practices). In addition to the Labor Day refusal-to-work incident, 246 See supra notes 11–12 and accompanying text. there are indicators of a pullback in proactive enforcement—street stops—by CPD officers in Chicago’s highest-crime neighbor­hoods. 247 See, e.g., Arthur & Asher, supra note 10 (claiming new measures making paperwork more burdensome discourage police from engaging in “proactive policing”); Bryan, supra note 10 (noting that police are more cautious after the national outrage); Byrne, supra note 10 (noting Mayor Rahm Emanuel believes that Chicago police officers are more hesitant to engage in assertive patrol practices after the release of the video of Laquan McDonald shooting incidents). A recent report by the Crime Lab at the University of Chicago showed an 80% drop in street stops (including investigative stops and Terry stops), from 50,000 stops per month to 10,000 stops per month, in the period from November 2015 to January 2016. 248 Kapustin et al., supra note 9, at 24.

The drop in stops preceded the sharp spike in homicides in Chicago by two months. Factors proposed for the cause of the decrease in stops included: the outcry by community and political leaders following the release of the video of the shooting of Laquan McDonald; 249 See Mark Konkol, Kelly Bauer & Tanveer Ali, Laquan McDonald Video Shows Police Shooting Him 16 Times, DNA Info (Nov. 24, 2015, 4:18 pm), http://www.dnainfo.
com/chicago/20151124/archer-heights/laquan-mcdonald-video-shows-police-shooting-him-16-times [http://perma.cc/3LF6-MGW4]; see also Oppel, Plan Under Fire, supra note 1 (describing the Mayor’s plan in response to increased violence after the release of the video of Laquan McDonald’s death).
the city’s agreement with the ACLU of Northern Illinois requiring documentation of each stop; 250 See Jeremy Gorner, ACLU, Chicago Police Agree to Changes on Controversial Street Stops, Chi. Trib. (Aug. 7, 2015, 6:01 am), http://www.chicagotribune.com/news/ct-chicago-police-aclu-street-stops-met-20150807-story.html (on file with the Columbia Law Review). and a new state law, SB 1304, requiring police officers to issue a receipt to each person they stop. 251 725 Ill. Comp. Stat. Ann. 5/107-14 (West 2016); see also Police Reform Bill Passes the Illinois General Assembly, Cmty. Renewal Soc’y (June 3, 2015), http://www.
communityrenewalsociety.org/police-reform-bill-passes-illinois-general-assembly [http://
perma.cc/2SPD-2KR7] (interpreting the bill). In addition to requiring officers to issue receipts after street stops, SB 1304 set guidelines for body-worn cameras, required the creation of a database for police officer professional conduct, mandated independent investigations of officer-involved deaths, prohibited chokeholds, and implemented several other reforms to strengthen investigations and oversight of CPD. See Police Reform Bill Passes the Illinois General Assembly, supra.
Although Chicago experienced a sharp decline in stops before its homicide spike, homicides in New York City did not increase following a similar decline in stops from 2012 through 2015. In fact, homicides there continued to decline in 2016 after a small uptick in 2015. 252 Kirby, supra note 224 (noting New York City’s unique declining crime rate); NYC Crime Falls to Lowest Levels in Decades: NYPD, supra note 225. Perhaps there is some optimal number of stops in each city, but calculating that figure is indeed a challenge.

Chicago, then, seems to follow the same pattern as New York City’s South Bronx neighborhood and South L.A.: Citizens who witness shoot­ings and murders refuse to cooperate with police. 253 Monica Davey, In Chicago, Bodies Pile Up at an Intersection of ‘Depression and Rage,’ N.Y. Times (Dec. 9, 2016), http://nyti.ms/2k4UY3U (on file with the Columbia Law Review) (detailing uncooperative witnesses after a major shooting). Citizen trust in the police is low, and it is lowest in the South and West Side neighbor­hoods—places that are both the most heavily policed and the most in need of police protection and criminal investigations. 254 Monica Davey & Giovanni Russonello, In Deeply Divided Chicago, Most Agree: City Is Off Course, N.Y. Times (May 6, 2016), http://nyti.ms/2kV84of (on file with the Columbia Law Review). At the same time that homicides go unsolved, citizens arrested for petty offenses spend long periods in jail awaiting court dates and then plead guilty to misdemeanors that lead to spells on probation, creating an even deeper wedge between citizens and police. 255 See Shaila Dewan & Andrew W. Lehren, After a Crime, the Price of a Second Chance, N.Y. Times (Dec. 12, 2016), http://nyti.ms/2mdUJF4 (on file with the Columbia Law Review). This distrust breeds further violence by breaking the already threadbare ties between citizens and police, which then puts both citizens and police at some risk. Through dozens of interviews with adolescents and young adults on the West Side, Professor Craig Futterman and his colleagues captured the cycle of violence and distrust:

It is not often enough noted that law enforcement itself is among the major casualties of the lack of police accountability . . . . Distrust of the police sets in motion a cycle with devastating effects for our children’s safety, sometimes with grave consequences. Because kids do not trust police, they will not call the police when they’re in trouble. Because people will not cooperate with the police, the police cannot solve and prevent crimes. Because the police cannot solve and prevent crimes, residents are yet less likely to go to the police and more likely to look for resolution or restitution outside the law. In the absence of meaningful accountability, the cycle goes on and on, reducing urban neighborhoods to something akin to failed states. 256 Craig Futterman et al., “They Have All the Power”: Youth/Police Encounters on Chicago’s South Side 23 (Univ. of Chi., Pub. Law & Legal Theory Working Paper No. 573, 2016), http://chicagounbound.uchicago.edu/cgi/viewcontent.cgi?article=2022&context=public_law_and_legal_theory (on file with the Columbia Law Review).

Much of the violence in in these neighborhoods is common and deadly, often stemming from gang conflicts. 257 Ben Green, Thibaut Horel & Andrew V. Papachristos, Modeling Contagion Through Social Networks to Explain and Predict Gunshot Violence in Chicago, 2006 to 2014, 177 JAMA Internal Med. at e1, e4 (2017), http://jamanetwork.com/journals/
jamainternalmedicine/fullarticle/2594804 [http://perma.cc/26BH-3HH9] (showing that 52.3% of the subjects of gun violence in Chicago between 2006 and 2013 were gang members, increasing the penetration of gang members into violence networks and creating a contagion vector of gangs).
Yet there are few exits from gang life and little incentive to exit even when one is available. 258 See Andrew V. Papachristos, Murder by Structure: Dominance Relations and the Social Structure of Gang Homicide, 115 Am. J Soc. 74, 75 (2009) [hereinafter Papachristos, Murder by Structure] (describing the prevalence and causes of gang violence in impoverished Chicago neighborhoods); see also John Eligon, Bored, Broke and Armed: Clues to Chicago’s Gang Violence, N.Y. Times (Dec. 22, 2016), http://nyti.ms/2jHCKsf (on file with the Columbia Law Review). Interpersonal disputes are often the spark for gang violence and killings. 259 See Papachristos, Murder by Structure, supra note 258, at 75–76 (arguing individual killings spark contagion-like spread due to norms of reciprocity); Eligon, supra note 258 (describing how gang members sometimes lash out in inexplicable bursts of violence over non-gang-related issues, such as disputes for attention). But gangs in Chicago also are complex organizations and heterogeneous in their makeup, with unemployed drug dealers and working men in the same groups. 260 See Eligon, supra note 258. Income from drug sales is central to the economic lives of some in the gang, though certainly not for all. 261 See id.; see also Paul E. Bellair & Thomas L. McNulty, Gang Membership, Drug Selling, and Violence in Neighborhood Context, 26 Just. Q. 644, 649 (2009) (describing economic and social incentives in disadvantaged neighborhoods of seeking gang membership and selling drugs); Steven D. Levitt & Sudhir Alldi Venkatesh, An Economic Analysis of a Drug-Selling Gang’s Finances, 115 Q.J. Econ. 755, 755–56 (2000) (reviewing literature and providing an economic analysis of the nature and causes of drug selling by gangs). While similar disputes and conflicts between local drug-dealing groups often spark murders in New York City, 262 Mueller & Baker, Mother Shot Dead, supra note 241; Benjamin Weiser, Sweep in Bronx Tackles Decades of Gang Chaos, N.Y. Times (Apr. 27, 2016), http://nyti.ms/
2ky1OTc (on file with the Columbia Law Review).
gangs there are more concentrated in public housing developments, 263 Kormanik, supra note 242. while in Chicago, gangs spread across the South and West Sides. Accordingly, there are fewer gang conflicts in New York City simply by virtue of the smaller number of gangs. 264 Benjamin Mueller & Al Baker, Drop in Gang Violence Drove New York Shootings Below 1,000 in 2016, N.Y. Times (Jan. 3, 2017), http://nyti.ms/2kqPAZ1 (on file with the Columbia Law Review).

One other factor likely driving Chicago’s spike in homicides is guns. Chicago’s gun homicide rate in 2016 was 25.1 per 100,000 persons, nearly twice the rate of 14.7 in Philadelphia, the next-highest large city measured. 265 Kapustin et al., supra note 9, at 9. At the same time, its non-gun homicide rate of 2.7 per 100,000 persons was comparable to the rates in Houston and Philadelphia, among the other largest cities in the United States. 266 Id. Only New York City and L.A. had lower non-gun homicide rates among the five largest U.S. cities. Id. Gun availability in Chicago, as evidenced by gun seizures, is very high: Approximately 8,300 guns were seized in 2016. 267 2016 Chicago: 762 Murders, 3,500 Shootings, 8,300 Guns Confiscated, WOWT News (Jan. 1, 2017), http://www.wowt.com/content/news/2016-Chicago-762-murders-more-than-3500-shootings–409173925.html [http://perma.cc/C2SK-JE39]. This is an increase of more than 30% from the 6,521 illegal guns seized by mid-December of the previous year and far above the 6,815 and 6,429 guns seized in 2013 and 2014, respectively. 268 Sarah Kollmorgen, Chicago Criminals’ Favorite Gunmakers: A Visual Ranking, Trace (Jan. 6, 2016), http://www.thetrace.org/2016/01/chicago-crime-guns-chart/ [http://perma.cc/U2JX-3MBM]. One gun per hour was seized in Chicago in the first six months of 2016; 100 guns were seized on Father’s Day weekend alone. 269 Chicago Cops Have Seized a Gun per Hour This Year, CBS Chi. (June 21, 2016), http://chicago.cbslocal.com/2016/06/21/chicago-cops-have-seized-a-gun-per-hour-this-year/ [http://perma.cc/Q8BM-YHEL]. Most of the guns were seized in the South and West Sides, the areas of the city with the highest homicide rates. The volume of these seizures is startling in light of the pullback in policing and in Terry stops in Chicago—which has occurred even as the CPD has highlighted the connection between gun availability and murder. 270 Office of the Mayor, City of Chi. & Chi. Police Dep’t, Tracing the Guns: The Impact of Illegal Guns on Violence in Chicago 7 (2014), http://www.cityofchicago.org/
dam/city/depts/mayor/PressRoom/PressReleases/2014/May/05.27.14TracingGuns.pdf [http://perma.cc/5MMW-UY3N].

Most of Chicago’s increase in homicides was in gun homicides, which rose from 261 in 2015 to 687 in 2016. 271 Kapustin et al., supra note 9, at 9–10. While gang disputes are a major source of gun homicides, gun robberies accounted for the major increase in robberies in 2016. 272 Id. Chicago, in other words, like New York City in the 1980s, suffers not only from concentrated poverty and legal cynicism among both citizens and police but also from a gun violence epidemic. 273 Bourgois, supra note 125, at 115–19 (showing cynicism among drug dealers toward law and legal work from the 1980s into the early 1990s); Fagan, Zimring & Kim, supra note 162, at 1289–90 (detailing an epidemic in New York City that was exclusively guns); Paul Jargowsky, Ghetto Poverty Among Blacks in the 1980s, 13 J. Pol’y Analysis & Mgmt. 288, 288–305 (1994) (discussing concentrated poverty in major cities including Chicago and New York City in the 1980s).

Figure 5 shows the density of Federal Firearm Licensees and gun manufacturers within twenty miles of the Chicago city border as of January, 2016, about one year after the onset of the rise in Chicago gun homicides in 2015. The Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF) issues firearms licenses (FFLs) to regulate the distribution and manufacture of firearms. 274 ee Firearms, ATF: Bureau of Alcohol, Tobacco, Firearms and Explosives, http://www.atf.gov/firearms [http://perma.cc/S2FR-QC9D] (last visited Apr. 13, 2017). The Bureau also conducts firearms licensee qualifications and compliance inspections. Figure 5 shows that there were dozens of FFL licensees selling firearms within that perimeter, including several on the city border adjacent to the South and West Side areas where homicides are concentrated. Another cluster of dealers is concentrated within a few miles of the Indiana state line on Chicago’s southern boundary.

Whether Chicago’s crime guns are entering the city from these sources cannot be fully determined from DOJ gun trace data, since only a portion of crime guns are recovered and submitted for trace. But based on what is known, signs do point to crime guns entering the city from those dealers. One dealer, located near the southeast side of the city, sold over 1,500 crime guns in a five-year period from 2009 to 2013. 275 Office of the Mayor, City of Chi. & Chi. Police Dep’t, supra note 270, at 7. Four nearby dealers accounted for 20% of the recovered crime guns in Chicago during this period. 276 Id. at 6. Together, dealers in Illinois and neighboring Indiana accounted for 11,874 recovered crime guns during this period, nearly 60% of the total. 277 See id. at 5 (showing 8,050 guns recovered in this period originally sold in Illinois and 3,824 guns originally sold in Indiana, totaling 11,874). If Chicago is suffering from a gun homicide epidemic, these sources—arrayed in a pattern that may not be unique to Chicago 278 un Sellers in the United States, OREGONLIVE: The Oregonian, http://
projects.oregonlive.com/guns/dealers [http://perma.cc/RTH5-USQZ] (last visited Apr. 19, 2016) (providing an interactive map displaying gun dealerships by county across the United States).
—might be thought of as one of the “pumps” that is infecting the neighborhoods with the instruments of murder.

Figure 5: Chicago Area Gun Dealers, January 2016 279 This graph was creating using ATF’s data for FFL holders in Indiana and Illinois for January 2016. See Downloadable Lists of Federal Firearm Licensees (FFLs), ATF: Bureau of Alochol, Tobacco, Firearms and Explosives, http://www.atf.gov/firearms/
listing-federal-firearms-licensees-ffls-2016 [http://perma.cc/NPR9-SEFL] (last updated Jan. 30, 2017) (data on file with the Columbia Law Review). The addresses were geocoded using World Geocoding Service through ArcGIS Mapping with ArcMap. See ArcGIS, ESRI, http://www.arcgis.com/
home/index.html (last visited Apr. 14, 2017). Map boundaries for Chicago were downloaded from the City of Chicago. See City of Chicago, Data Portal: Boundaries – City, http://data.cityofchicago.org/Facilities-Geographic-Boundaries/Boundaries-City/ewy2-6yfk (data on file with the Columbia Law Review). Illinois boundary maps (shape files) were obtained from Illinois State Boundary, Illinois Geospatial Data Clearinghouse http://clearinghouse.isgs.illinois.edu/data/reference/illinois-state-boundary (data on file with the Columbia Law Review). Indiana boundaries were downloaded from TIGER/Line Shapefile, 2013, State, Indiana, Current County Subdivision State-based, Data.gov, http://catalog.data.gov/dataset/tiger-line-shapefile-2013-state-indiana-current-county-subdivision-state-based (data on file with the Columbia Law Review).


In Chicago, the gun supply interacted with the social and economic contexts of the South and West Sides to animate the homicide epidemic of these past two years. Wherever the guns are sold, they seem to land in the poorest places and the places most susceptible to violence. The loss of faith in the police in the highest-homicide areas extends to a broader loss of faith in many of the city’s institutions, from schools to courts to the police. 280 See, e.g., Denisa R. Superville & Stephen Sawchuk, In Chicago School Funding Battles, Human Costs Are Steep, EdWeek (May 17, 2016), http://www.edweek.org/
ew/articles/2016/05/18/in-chicago-school-funding-battles-human-costs.html (on file with the Columbia Law Review) (noting parents’ frustration with the “tiered education system where students on the North Side and in selective-enrollment schools get needed resources while those on the West and South sides ‘get the crumbs’”).
Police abuse and impunity go hand in hand to create deeply felt cynicism among the residents. 281 N.Y. Times & Henry J. Kaiser Family Found., Poll of Chicago: April 21 – May 3, 2016, at 7 (2016), http://assets.documentcloud.org/documents/2824824/Chicago-Trn-Final.pdf [http://perma.cc/S3VM-9V5E] (noting only 54% of those polled believe calling the police when something illegal is happening will make the situation better). Accompanying this despair is severe economic deprivation and isolation. 282 See Robert J. Sampson, Great American City 46 (2012) (noting factors of social inequality are often associated with geographic racial isolation in Chicago neighborhoods); Robert J. Sampson, Stephen W. Raudenbush & Felton Earls, Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy, 277 Science 918, 918–19 (1997) (proposing an explanation for the link between higher rates of violence in neighborhoods and low socioeconomic status and residential instability); Scott J. South & Kyle D. Crowder, Neighborhood Effects on Family Formation: Concentrated Poverty and Beyond, 64 Am. Soc. Rev. 113, 115 (1999) (discussing concentration effects from social isolation of underserved neighborhoods); see also Donna L. Franklin, Susan E. Smith & William E.P. McMiller, Correlates of Marital Status Among African American Mothers in Chicago Neighborhoods of Concentrated Poverty, 57 J. Marriage & Fam. 141, 150 (1995) (noting the youngest and most economically disadvantaged African American mothers were also the most residentially isolated group). Like New York City and L.A. neighborhoods, Chicago’s poorest and most violent neighborhoods also exemplify the concept of the poverty trap. 283 Sampson & Morenoff, supra note 83, at 198–201 (describing the persistence of poverty and the durability of its social effects in Chicago neighborhoods). The robust science showing the connections between neighborhoods and socioeconomic deprivation can be applied in multiples to Chicago: The extremes of its violence spike in the past two years mirror the extremes and the breadth of its concentrated poverty and segregation that has persisted over decades.

B. Lessons and Next Steps

In this symposium, Professor Meares echoes the President’s Task Force on 21st Century Policing and concludes that police need to commit more to building trust within the communities they serve. 284 Meares, The Path Forward, supra note 76. She is hardly alone in this. In an October 2015 speech in Chicago, FBI Director Comey spoke of “two lines: one line is law enforcement and the other line is the folks we serve and protect, especially in communities of color.” 285 Comey, Remarks at University of Chicago, supra note 79. He worried that “those two lines are arcing away from each other, at an increasing rate.” 286 Id.

This Essay contends that this extends to trust for homicide investigations. L.A., New York City and Chicago are not alone in experiencing low clearance rates owing in part to citizen distrust and withdrawal. The Baltimore Sun opined on the centrality of clearance rates for public safety:

Clearance rates in the United States have been in decline for years, but the national average is still better than 20 points higher than Baltimore’s rate . . . . Cases that aren’t cleared by police are too often cleared by the streets, leading to the type of reciprocal killings that plague Baltimore. 287 See Editorial, Chicago’s Homicide Crisis Gets National Attention but Baltimore’s Is Much Worse, Balt. Sun (Jan. 3, 2017), http://www.baltimoresun.com/news/opinion/editorial/bs-ed-chicago-baltimore-20170103-story.html (on file with the Columbia Law Review).

Clearance rates matter in reassuring people that police are dedi­cated to their safety and that they can deliver on promises of security. This is especially important in an era of proactive policing, when there are implicit tradeoffs between aggressive policing and dignity. 288 Tyler et al., Teachable Moments, supra note 182; Jeffrey Fagan, Tom R. Tyler & Tracey L. Meares, Street Stops and Police Legitimacy in New York 210 (Columbia Univ., Pub. Law Research Paper No. 14-514, 2016), http://ssrn.com/abstract=2795175 (on file with the Columbia Law Review) (“Targeting harm, then, encompasses both an innocence harm plus the procedural indignity of being targeted as criminal with what appears to the person stopped to be more of a hunch or an assumption than a reasonable basis.”). As these three case studies show, the trust and cooperation of citizens are essential to reducing murder by leveraging citizen cooperation in homicide investigations. Nearly fifteen years ago, as homicide rates were plunging in most cities, Professors Janice Plunket and Richard Lundman suggested that “the significantly lower clearance rates in Black census tracts and integrated census tracts are a function of less trust and less cooperation and information from citizens.” 289 Janice L. Puckett & Richard J. Lundman, Factors Affecting Homicide Clearances: Multivariate Analysis of a More Complete Conceptual Framework, 40 J. Res. Crime & Delinq. 171, 185 (2003). They noted, “[w]hen people are reluctant to talk to homicide detectives, when they are uneasy about telling homicide detectives what they saw, what they know, and what they suspect, the necessary result is lower clearance rates.” 290 Id.; see also Susan Clampet-Lundquist, Patrick J. Carr & Maria J. Kefalas, The Sliding Scale of Snitching: A Qualitative Examination of Snitching in Three Philadelphia Communities, 30 Soc. F. 265, 267–70 (2015) (describing the reasons witnesses or victims of crimes choose not to contact or cooperate with police).

All three case studies point toward the importance for police to shift their focus from petty crimes, which have become the core of the new policing, toward stronger efforts in homicide investigations. The lesson is that the most appropriate strategic response to homicide spikes, if not epidemics, may have less to do with the number of police than with how those police are deployed. 291 See Steven Mello, Police and Crime: Evidence from COPS 2.0, at 27 (2016) (unpublished manuscript), http://ssrn.com/abstract=2812701 (on file with the Columbia Law Review) (reporting that increases in the number of police appear to reduce murders by a small fraction, between 0.3–0.4 per officer); see also Aaron Chalfin & Justin McCrary, Are US Cities Underpoliced? Theory and Evidence 35 (2013), http://eml.berkeley.edu/
~jmccrary/chalfin_mccrary2013.pdf [http://perma.cc/W4VT-ZVM5] (discussing a series of measurement-error-corrected estimates and suggesting increases in police help to reduce violent crime, such as murder). In fact, the 2016 homicide clearance rate by the Baltimore Police Department was 38%, well below the national average of 62%. Adam May, Baltimore County Police Clearing Homicides at Unprecedented Rate, WBAL-TV (Dec. 29, 2016), http://www.wbaltv.com/article/baltimore-county-police-clearing-homicides-at-unprecedented-rate/8544960 [http://perma.cc/9FSS-WVCZ]. In contrast to the low clearance rate in the City of Baltimore, the clearance rate in the surrounding areas of Baltimore County was 82.4%. Id.

Homicide and nonfatal shootings in small social and physical spaces are contagious. 292 See, e.g., Green et al., supra note 257, at e6 (reporting that gunshot violence follows an process of social contagion that is transmitted through social interactions, with an influence that is greater than the vulnerabilities to violence from living in conditions of social isolation and economic deprivation); Andrew V. Papachristos, Christopher Wildeman & Elizabeth Roberto, Tragic but Not Random: The Social Contagion of Nonfatal Gunshot Injuries, 125 Soc. Sci. & Med. 139, 140 (2014) (finding “nonfatal gunshot injuries are far more concentrated than previously thought” and “the probability of victimization is strongly associated with network exposure to nonfatal gunshot victims,” leading to the inference that gun violence spreads through processes of social contagion . . . concentrated in risky networks and are associated with specific behaviors”). Many of the processes assumed to reflect the features of a place, a structural view, may actually be the product of social interactions that take place in those dense social spaces. This suggests a process of social contagion in which behaviors, attitudes, and beliefs are spread through social contacts: exchanges of status, conflict between or within social networks, or reciprocal perceptions of risk. 293 Green et al., supra note 257, at e2.

Guns appear to be the “host” in homicide epidemics. 294 K. Christoffel, Firearm Injuries: Epidemic Then, Endemic Now, 97 Am. J. Pub. Health 626, 626 (2007) (tracing the spread of gun injuries as an epidemic followed by a decline to a stable pattern over time); see also Cohen & Tita, supra note 108, at 457 (noting that youth participating in crack markets “increasingly carry guns and use them to settle interpersonal disputes, thereby spreading gun violence more broadly among the youth population”); Daniel Cork, Examining Space–Time Interaction in City-Level Homicide Data: Crack Markets and the Diffusion of Guns Among Youth, 15 J. Quantitative Criminology 379, 383 (1999) (explaining that firearms are “tool[s] of the trade” in crack markets because of the need for dealers to resort to self-help); Fagan & Davies, supra note 164, at 142 (“The role of guns in homicide may explain why gun homicides are more likely to show a contagious pattern than violence, generally.”); Jeffrey Fagan, Deanna L. Wilkinson & Garth Davies, The Social Contagion of Violence, in Cambridge Handbook of Violent Behavior 688, 689 (2007) (“Although disease spreads through a host and agent . . . social contagion involves the mutual influence of individuals within social networks who turn to each other for cues and behavioral tools that reflect the contingencies of specific situations . . . .”); Elizabeth Griffiths & Jorge M. Chavez, Communities, Street Guns and Homicide Trajectories in Chicago, 1980–1995: Merging Methods for Examining Homicide Trends Across Space and Time, 42 Criminology 941, 946 (2004) (noting “changing patterns of firearm violence” and focusing on the “subsequent impact on changing rates of both total and neighborhood-specific homicide”); David Hemenway et al., Gun Carrying Among Adolescents, Law & Contemp. Probs., Winter 1996, at 39, 47–48 (finding “carrying firearms makes other students feel less safe, which increases the likelihood that they will in turn carry guns” and concluding “results of contagion modeling suggest that small initial changes in gun carrying can have multiplicative effects”); Papachristos, Murder by Structure, supra note 258, at 75–76 (“Individual murders between gangs create an institutionalized network of group conflict—sustained patterns of interaction—net of any individual’s participation or motive. Within this network, murders spread through a process of social contagion as gangs respond to threats by evaluating the highly visible actions of others in their local networks.”); Papachristos et al., Social Networks and the Risk of Gunshot Injury, 89 J. Urb. Health 992, 993 (2012) (noting that “the normative conditions surrounding gun use are transmitted through processes of peer influence, especially among young men with criminal histories”); Deanna L. Wilkinson & Jeffrey Fagan, The Role of Firearms in Violence Scripts: The Dynamics of Gun Events Among Adolescent Males, Law & Contemp. Probs., Winter 1996, at 55, 57 (“[T]he presence of firearms contributes to increasingly alarming homicide rates among inner-city adolescents.”); Deanna L. Wilkinson & Jeffrey Fagan, A Theory of Violent Events, in The Process and Structure of Crime: Criminal Events and Crime Analysis 169, 178–80 (Robert F. Meier & Leslie W. Kennedy eds., 2001) (exploring the “situational contexts of violent events” and asserting “[v]iolent events appear to be situated transactions that reflect the interactions of the parties and the features of both the social and situational contexts”). Disrupting the retaliatory killings in dense social networks by removing murderers is a logical step in reducing murder risk. Providing exit paths through a combination of incentives and deterrence has also proven valuable in reducing homicide rates among former gun offenders. 295 See Andrew V. Papachristos, Tracey L. Meares & Jeffrey Fagan, Attention Felons: Evaluating Project Safe Neighborhoods in Chicago, 4 J. Empirical Legal Stud. 223, 224 (2007) (discussing how Chicago’s Project Safe Neighborhoods program reduced homicide rates by more than 35%); Andrew V. Papachristos, Tracey L. Meares & Jeffrey Fagan, Why Do Criminals Obey the Law? The Influence of Legitimacy and Social Networks on Active Gun Offenders, 102 J. Crim. L. & Criminology 397, 399–400 (2012) (using studies conducted among active gun offenders to conclude that individuals in this group differed in perspective about compliance with the law); Danielle Wallace et al., Desistance and Legitimacy: The Impact of Offender Notification Meetings on Recidivism Among High Risk Offenders, 33 J.Q. 1237, 1239 (2015) (“Research supports the efficacy of crime prevention programs which incorporate procedural justice and legitimacy into a focused deterrence approach.”).

Perceived injustices, including both “petty indignities” and egre­gious acts of police violence, can disincentivize citizens from cooperating with police. 296 See Bell, supra note 205, at 5 (“The empirical evidence shows that these feelings of distrust manifest themselves in reduced likelihood among these groups to obey law enforcement officers’ commands, and to cooperate with officers to help them solve crimes.”); Mark Berg et al., Cynical Streets: Neighborhood Social Processes and Perceptions of Criminal Injustice, 54 Criminology 520, 521–22 (2016) (arguing that “negative or unfavorable police-citizen encounters are more likely to occur in neighborhoods with greater levels of cynicism”); Desmond et al., Police Violence and Citizen Crime Reporting, supra note 205, at 870 (finding “publicized cases of police violence against unarmed black men have a clear and significant impact on citizen crime reporting”); Eric A. Stewart et al., Neighborhood Racial Context and Perceptions of Police Based Racial Discrimination Among Black Youth, 47 Criminology 487, 849–50 (2009) (recognizing that “neighborhood social context has been shown to be integral to shaping race relations, police behavior, and police-citizen encounters”). And the risks of those perceived injustices appear to be greater under the aggressive and proactive tactics of the “new policing.” 297 See Jeffrey Fagan, Dignity is the New Legitimacy, in The New Criminal Justice Thinking 308, 313 (Sharon Dolovich & Alexandra Natapoff eds., 2017) (noting that “front-end encounters with the police under the ‘new policing’ are fertile grounds for petty or pointless indignities or gratuitous humiliations”); Jonathan Simon, The Second Coming of Dignity, in The New Criminal Justice Thinking, supra, 275, 281 (noting that the “severity of criminal justice treatment[] has created a gulf between the communities harmed and humiliated by criminal justice encounters and those that feel protected and respected by those encounters”); Bowers, supra note 213, at 992–95 (“In the age of order-maintenance policing . . . it is inevitable that some not insignificant number of low-level, full-custodial arrests reflects very bad choices (or worse).”); Jeffrey Fagan, Indignities of Order Maintenance 44 (2016) (unpublished manuscript) (on file with the Columbia Law Review) (noting that “indignities, including those that carry a racial component, can undermine the ties of citizens to law and the moral and social norms that underlie them”). Indeed, the risks of policing indignities in a community depend quite a bit on the city’s policing model—a model over which a community likely has little influence. 298 See Charles F. Sabel & William H. Simon, The Duty for Responsible Administration and the Problem of Police Accountability, 33 Yale J. on Reg. 165, 210–11 (2016) (finding “post-bureaucratic organization takes markedly different forms with correspondingly different implications for accountability”). But dignitarian insults are only one consequence of the new policing. An equally important cost is the loss of effectiveness in crime control. Two empirical studies of New York City’s aggressive policing model show few returns to crime control from saturated investigative stops. One study employed an experimental design to show that while a high rate of stops based on an expansive and subjective definition of reasonable suspicion had minimal effects on crime, a higher rate of stops based narrowly on factors approximating a probable cause standard substantially reduced crime. 299 See MacDonald, Fagan & Geller, supra note 233. A second test used a longitudinal (panel) design and reached the same conclusion: Stops based on probable cause significantly reduce crime. 300 See Jeffrey Fagan, Terry’s Original Sin, 2016 U. Chi. Legal F. 43, 46 (“The analyses show significant reduction in crime in neighborhoods . . . with greater numbers of probable cause stops, and ratios of probable cause stops to other stops.”). Both studies argue for a shift in police practice away from inflated notions of suspicion and toward a more specific and narrow notion of “suspicion” based on probable cause. In New York City, the setting for these studies, the same expansive policing tactics generally did not recover weapons or contraband or lead to arrests of active offenders for felony crimes. 301 See, e.g., Floyd I, 959 F. Supp. 2d 540, 573 (S.D.N.Y. 2013) (using statistical data to demonstrate the unlikelihood of discovering weapons or contraband); see also Meares, Programming Errors, supra note 44, at 162–63 (2015) (explaining the theory that stop and frisk carried out “en masse as a program” is inconsistent with Terry’s authorization for such stops in the case of individual incidents).

The new policing logic focuses on the wrong behaviors and externalizes inefficiency costs to impacted citizens in the form of pointless stops and a range of perceived injustices. The turn to an investigative model can reverse the distraction of police from minor crimes and infractions and restore focus on violence and murder. In other words, if police do not “sweat the little stuff” but rather “go after the hard stuff,” they will reap dividends in the trust and cooperation of the people most affected by that violence.


All too many cities experienced recent increases in homicides. This Essay suggests that, given existing information, there is no reason to believe that these increases presage a new homicide epidemic and a return to the era of elevated homicide rates that persisted in many U.S. cities over three decades through the mid-1990s. These spikes could be momentary upticks in the long-term decline that has been ongoing for over two decades; they could also signal a new era of unpredictable and random surges and declines in homicide across cities during an otherwise stable pattern over time.

Homicide spikes appear to be more volatile in smaller cities, with the exception of the recent severe homicide outbreak in Chicago. Closer examination of the conditions in neighborhoods in three large cities shows the intersection of policing tactics and social contexts in which the concurrent withdrawal of citizens and police from the co-production of security has profound consequences for the control of violence. This Essay argues for a shift in policing tactics from order maintenance and proactive police contacts—with their potential to produce injustices and indignities—to a focus on investigations of homicide, with the promise both of bringing offenders to justice and of restoring trust in the police.