Why Black Lives Matter Should Objectively Focus on Black-on-Black Crime

These are random images for flavor with no relevance to the essay’s substantive content. All rights reserved to the creators. If you believe I have used an image without legal permission, please inform me.

Argument in brief:

The protest movement Black Lives Matter — as well as policy makers, social scientists and journalists interested in race and criminal justice — should focus on trying to reduce the black violent crime rate to a level comparable to other American ethnic groups. (As opposed to focusing on reducing the rate at which blacks are shot by police to that of other American ethnic groups.) They should do so for two reasons:

  1. If the African-American homicide rate was comparable to the white or Asian homicide rate, around 5,000 fewer African-Americans would die every year. If the African-American police shooting victimization rate was comparable to the white one, around 150–200 fewer African-Americans would die every year. Therefore, the major, quantifiable, disproportionate violent threat to black lives is black on black crime, not police shootings.
  2. Different groups of Americans — races, genders and age cohorts — are killed by the police at different rates because they commit crime at different rates. A very large part, and perhaps the entirety of, the difference in police shootings of whites and blacks is explained by differences in violent crime rates. Reducing black on black crime would likely also reduce police shootings of blacks.

I will also address many common objections to this line of argument, and demonstrate that they are incorrect.

Accordingly, I see no reason why an ordinal ranking of the problems facing American society in general or African-American society in particular should put reduction of the racial disparity in police shootings ahead of reduction in the racial disparity in violent crime rates.

Part I: the double incorrectness of focusing on police shootings over black-on-black crime

These images are random flavoring, with no relevance to the actual essay.

The movement Black Lives Matter is supposedly dedicated to saving black lives. It attempts to go about this by protesting police shootings of blacks. This is puzzling for two reasons:

Firstly, because something on the order of twenty times as many African-Americans are killed by black on black crime as by police shootings every year. According to the FBI’s Uniform Crime Reports, from 2010–2014 the annual number of African-American homicide victims has ranged from about 6,000 to about 6,500. According to the Guardian’s database on police shootings, around 300 African-Americans were killed in police shootings in 2015, and around 170 have been killed in 2016 so far; according to the Washington Post’s database, around 260 African-Americans were killed in police shootings in 2015 and around 150 have been killed in 2016 so far. In other words, for every black life terminated by a police officer, there are twenty other black lives terminated by black on black crime. Furthermore, the disproportion relative to the African-American population share is greater in ordinary homicides, in which African-Americans make up around 50% of the victims, than police shootings, in which African-Americans make up around 30% of the victims. If black lives do indeed matter, why doesn’t “Black Lives Matter” try to ameliorate the objectively most disproportionate threat to black lives?

To give this a human face, imagine if, while walking home after robbing a convenience store, Michael Brown had gotten into and altercation and was shot to death by another young black man (rather than a police officer.) This probably would not have attracted the attention of the New York Times, CNN, the Washington Post and Vox. This probably would not have led to nationwide protests. This probably would not have led to rioting and looting necessitating the deployment of the National Guard. Brown’s death wouldn’t have become a tragedy; it would have become a statistic. Yet, paradoxically, the risk of such a death is over an order of magnitude greater for the Michael Browns of the world than the risk of dying in a police shooting.

It’s helpful to consider what would happen if the black homicide rate and the black police shooting victimization rate converged to their white counterparts. The black homicide rate, according to the CDC, has hovered at around 20/100,000 persons per year since the 1990s, whereas the white and Asian homicide rates have remained around 2.5/100,ooo persons per year. If the black homicide rate was reduced to the white or Asian level, over 5,000 black lives per year would be saved. According to the Guardian’s database, the black police shooting rate was about 7.5/1,000,000 persons in 2015 (note the different scales of measurement) and the white police shooting rate was about 3/1,000,000 persons per year. Therefore, if the black police shooting rate was reduced to the white police shooting rate, about 200 black lives would be saved per year.

Secondly, because a key — quite possibly the main — reason African-Americans are killed in encounters with the police at higher rates than other American ethnic groups is that African-Americans commit violent crimes at higher rates than other American ethnic groups. This is a controversial claim in some people’s eyes, but many analogous claims would seem banal. For instance, Vox emphasizes that, according to their analysis of 2012 data, 31% of police shooting victims were black, despite blacks being 13% of the U.S. population, and the BLM affiliated Campaign Zero’s 2015 police violence report noted African-Americans were 20% of the population of the jurisdictions studied, but 41% of the victims of police violence. In both cases, the authors suggest/imply that this is due in at least some considerable part (though not necessarily entirely) to irrational, unnecessary anti-black bias.

(Update: a Vox feature on police shootings notes the age and gender divides as being largely explained by differing criminality, and the racial divide as being to some ambiguous though non-absolute degree explained by it. I’m very glad that Vox is now being [more] forthright about the empirical facts, but I still think their writers and readers likely fail to extrapolate the logical conclusions from them about the issue that I do here.)

However, few would say that the fact that men are about 95% of the victims of police shootings despite being about 50% of the population is evidence of deep anti-male bias in police culture. The obvious, uncontroversial explanation for this disparity is that men are much more likely to be violent criminals than women. Indeed, according to the Guardian’s data, 546 white men were killed by police in 2015, compared to 12 black females. (Which is apparently enough for Bustle to declare: “police kill black women all the time.”) Adjusting for population, this indicates that white men are killed by the police at about 9x the rate of black women — much higher than the 2.5x rate difference between blacks and whites overall. Does this mean that police departments across America are suffering from an epidemic of anti-white male, pro-black female racial and gender bias? Or does it just mean that white men are more likely to commit violent crimes than black women?

Likewise, the victims of police shootings do not “look like America” in terms of age. According to the Washington Post’s database, under 2% of the victims of police shootings were under age 18, whereas around 26% of the U.S. population is under 18. People ages 18–24 are less than 10% of the U.S. population, but people ages 18–29 are 33% of all victims of police shootings. People ages 45+ are about 34% of the U.S. population, but about 27% of all victims of police shootings. (And, presumably, within that demographic people closer to age 45 are victimized at non-trivially higher rates.) Are these age disparities the result of widespread police ageism, or the result of differences in crime rates by age?

Also, if you believe that blacks are killed in altercations with police at a higher rate than whites due to police racism, why is it that different groups of people of color are killed at very different rates, despite being equally non-white? According to the Guardian’s database, African-Americans are killed by police at a rate of 7.66/1m people annually, much higher than the Hispanic rate of 3.45/1m people, itself much higher than the Asian rate of 1.34/1m people (which is less than half the white rate.) Isn’t it just the craziest random happenstance that the exact same ordering of violent crime rates (African-American>Hispanic>Asian) is replicated in the ordering of police shooting victimization rates? Furthermore, does the fact that the police kill whites at slightly over 2x the rate of Asians mean that, in addition to a culture of anti-black bias, the police have a culture of anti-white bias?

Finally, it’s worth noting that though it seems highly likely that there are officers who kill civilians (civilians who are white, black, Hispanic and Asian) without justification who should be subject to criminal sanction, in probably at least a considerable majority of cases the reason the legal system finds officers innocent of wrongdoing is because they actually are. Consider the case of Michael Brown. Remember that Black Lives Matter activists don’t go around drawing attention to random police shootings; they (like any other political movement) focus on the ones that they think will best showcase their narrative about American society. Black Lives Matter chose Michael Brown’s death to be a key battleground in their battle with the American status quo. The resulting furor over Brown’s death resulted in massive protests, riots and national media attention. We were told that Brown was a “gentle giant” who pleaded “hands up, don’t shoot!”, that officer Wilson’s account of how the situation unfolded was literally unbelievable according to Ezra Klein and also racist according to Jamelle Bouie, and that Brown was martyred on the Catherine Wheel of America’s white supremacist institutions, as Emmett Till before him was.

And then the Department of Justice conducted an exhaustive investigation into the circumstances of Brown’s death, analyzing the relevant forensic, physical and testimonial evidence. The inquiry concluded that the evidence supported officer Wilson’s account of the events, that officer Wilson acted in self-defense, that there were no grounds to press charges on and that the “hands up, don’t shoot” meme was completely inaccurate.

And consider the circumstances leading up to the altercation resulting in Brown’s death. Was Wilson in Brown’s area because he was a racist KKK member out on a joy ride to harass random innocent African-Americans? No, he was in the vicinity because he had been notified a robbery of a convenience store had occurred, and for some reason the suspect in the robbery looked exactly like Michael Brown. Did Wilson approach Brown by yelling racial epithets and cursing at him? No, he asked Brown if he’d consider walking on the sidewalk instead of the middle of the street. Did Wilson coldly execute Brown by shooting him in the back? No, Brown attacked Wilson and attempted to take his firearm. Is there any reason to think that the result of Brown’s decisions would have been different if Brown was white, Hispanic or Asian?

It is a tragedy that Michael Brown is dead. The sum of an 18 year old’s transgressions should not be death. But people, like Michael Brown, who die at the hands of the police are generally not civilly discussing Victorian poetry over tea and crumpets with a friend before they die. As a perusal of the Guardian’s data shows, around 80% of the victims are armed; the victims also highly disproportionately tend to be violent criminals, domestic abusers and/or resisting arrest. This is not to deny that officers who kill civilians unjustifiably should face legal sanction, just as anyone else would, or to dispute the injustice of a particular case. It is rather to leaven the aggregate statistics about police shootings with relevant context; the circumstances of, say, Alton Sterling or Philando Castile’s deaths are not necessarily the median cases of police shootings. And again, there are considerably more white, Hispanic and Asian victims of police shootings — whether justified or not —than black ones (though the black per capita rate is significantly higher, as previously noted.)

The question of racial bias in police shootings has seen some quantitative analysis already: on the one hand, a widely reported study by Harvard economist Roland Fryer found no police bias in shootings (based on data from Dallas), and on the other hand UC Davis anthropologist Cody T. Ross found significant racial bias (even accounting for differences in crime rates) in police shootings. I would conjecture that the difference may be explained in part by the fact that Fryer had access to data about encounter rates, whereas the U.S. Police Shooting Database that Ross draws on did not (as he notes in the paper.) I agree with those who think that differences police shooting rates are not entirely mechanically produced by differing crime rates; the African-American share of victims of police shootings, around 30%, is, after all, considerably lower than the African-American shares of prisoners (37.7%), homicide perpetrators (around 50%) and cop killers (around 44%.) I think differences in police departments’ culture, training, tactics, weaponry and, yes, level of racial bias probably play some role independent of crime rates in rates of police shootings. However, differences in crime rates between groups — based on race, gender, age, region, educational attainment etc. — will necessarily play at least some role in differences in police shooting rates. Disproportions in the composition of criminal suspects will lead to disproportions in police stops; every stop carries some non-zero chance of ending in a shooting, so syllogistically, even if the police are completely race/gender/age-neutral, the composition of those who are killed in police shootings will be “biased” if crime rates are disproportionate.

To put it in simpler terms, can anyone really believe that if the African-American violent crime rate in, say, Chicago made the astronomical fall necessary necessary to reach parity with the white Chicagoan rate tomorrow, it would have no significant impact on the rate at which African-Americans would be killed in encounters with the police in Chicago over the following year?

Finally, though racism/racial prejudice morally should not exist under any circumstances, it likely will exist to some extent given observable average racial differences in behavior. In this case, the empirical fact that African-Americans commit violent crimes more often than other American ethnic groups is probably part of why people hold the stereotype that African-Americans are more likely to commit violent crimes than other American ethnic groups. The best way to combat this stereotype would be to reduce the rate at which African-Americans commit violent crimes to one comparable to whites, Hispanics or Asians.

Part II: pre-empting common objections

I’m not the first person to offer such a perspective, and this question has already been litigated considerably. I think it’s worthwhile to address some of the most common objections to the view I’ve briefly elaborated before proceeding. The Black Lives Matter movement has previously been criticized by observers such as journalist Heather Mac Donald, linguist John McWhorter and former New York City mayor Rudolph Giuliani for focusing on police shootings rather than (or in conjunction with) black-on-black crime, particularly homicide. In response, outlets such as Vox (1,2,3–number 3 is the single stupidest article that I have ever read on Vox dot com, incidentally), USA Today, the Atlantic, the Nation, the New York Times and MTV’s Youtube news channel have run pieces defending a focus on police shootings over black on black crime.

The counterarguments offered are remarkably similar, even though they’re voiced by many different authors in many different publications. They are all incredibly fallacious, and universally fail to rebut the charge that black on black crime is a more pressing issue than disproportionate police shootings of blacks. Let’s look at some of the central claims:

“There’s nothing unusual about ‘black on black crime’; it’s basically the same thing as ‘white on white crime’.”

This one appears in virtually every contrary argument: the USA Today piece, the MTV video, this New York Times opinion piece, this Huffington Post article, and most horrifically of all, in this moronic attempt at satire penned by Matt Yglesias at Vox. The argument is that because 93% of black homicides have black perpetrators and 84% of white ones have white perpetrators, “black on black crime” is no different from “white on white crime.”MTV’s Franchesca Ramsey goes a step further and openly makes the (demonstrably false) claim “Black people are not more violent or more likely to commit crimes [than members of other American ethnic groups].

This is at worst purely incorrect and at best ridiculously misleading. In fairness, part of the problem here is that the meme “black on black crime” on its face isn’t the best description of what people like Giuliani and Mac Donald are drawing attention to. A more accurate term for the phenomenon they’re trying to describe would be “unusually high black criminality”. African-Americans commit (and are victims of) homicide, as well as many other violent crimes, at many times the rates of American whites, Hispanics and Asians. This, not the banal fact that most black homicides are intra-racial, is the primary complaint “but what about black on black crime?” encapsulates. Contemporary black on black crime is so much higher than contemporary white on white (or Asian on Asian, etc.) crime that it merits being addressed as a unique analytical category. Let’s look at racial homicide rates graphically and put some numbers on them:

Source: Center for Disease Control report on homicides.
Source: RCA analysis of CDC data.
Source: the Rise and Fall of Violent Crime in America, by Barry Latzer
Source: the Rise and Fall of Violent Crime in America, by Barry Latzer. Note that even as the as the aggregate homicide rate fell dramatically in the 1990s, the black/white ratio saw only a fairly small change.
Source: FiveThirtyEight article

These graphs demonstrate a phenomenon you can easily observe for yourself by looking at the homicide statistics: American ethnic groups do not commit, or become victims of, homicide at the same rate. Since the great crime decline of the 1990s, the African-American homicide rate has been about 20/100,000 people per year, the Hispanic rate has been about 5–7/100,000 people per year, the white rate has been about 2.5/100,000 people per year, and the Asian rate has been around, and recently slightly lower than, the white rate. Thus, when Matt Yglesias says (sort of with his tongue in his cheek) “white on white murder is out of control” he is flatly incorrect; black on black homicide currently happens at around 7-8x the rate of white on white homicide. The FBI’s Uniform Crime Report for 2013, for example, finds that 31.1% of homicide offenders were white, 38% were black, 1.8% were “other” and the race of 29.1% of offenders was not known. Despite African-Americans being less than 14% of the U.S. population and non-Hispanic whites being around 62%, the absolute number of homicide victims and perpetrators for the racial groups is comparable. (And indeed, the number of African-American homicide victims and perpetrators sometimes actually exceed their white equivalents, despite the vast difference in sheer population.)

Examination of city level data is also informative: in New York City as of 2015, African-Americans are about 25% of the city’s population, and about 60% of murder victims/suspects/arrestees are African American. Whites are 44% of the city’s population, and about 6% of murder suspects/arrestees are white; Hispanics are about 27% of the city’s population, and about 30% of murder suspects/arrestees are Hispanic; Asians are about 12% of the city’s population, and about 2% of murder suspects/arrestees are Asian. In Chicago in 2015, African-Americans were about 77% of murder victims, and about 32% of the city’s population; Hispanics were about 18% of murder victims and 28% of the city’s population; and non-Hispanic whites were less than 5% of homicide victims and about 32% of the city’s population. Similar disparities are present in Los Angeles and many other large multi-racial American cities. (Ron Unz has done an interesting, more comprehensive analysis of race and crime in American cities and has come to similar conclusions.)

The upshot of all this is that progressives are frustratingly uninformed, obtuse, or both, when they attempt to draw an equivalency between white homicide and black homicide due to the fact that both are largely intra-racial. They are arguing against an extremely weak (mis)representation of their opponents’ arguments. People (such as Heather Mac Donald in her recent book the War on Cops) who ask “but what about black on black crime?” are not arguing that black murders should diversify their targets or that we should expect that no African-American ever should commit murder; they are arguing that as a society we should reduce the African-American homicide rate to one closer to those of other American ethnic groups. Therefore, attempting to deflect the argument “isn’t black on black crime a more serious problem than police shootings of blacks?” with the claim that there is no difference between white on white (or Hispanic on Hispanic or Asian on Asian) crime and black on black crime is substantially misguided.

A police officer shooting a civilian is of more concern than a civilian shooting a civilian/police officers who murder civilians are not convicted”

This argument is ubiquitous — understandably so, because it’s a valid (though incomplete) one. It’s made in a Ta-Nehisi Coates piece, this USA Today article, this Fusion article, this MTV video, this NBC exchange between Rudy Giuliani and Michael Eric Dyson and this Nation article, to name just a few.

To understand the problem with this argument, consider this passage from the aforementioned USA Today article:

With the movement’s attention comes a familiar refrain: Why doesn’t Black Lives Matter focus on “black-on-black” crime?
It’s a question asked, in various forms, from Facebook to cable networks to comments on this site. The answer, one writer says, is Black Lives Matter isn’t solely focused on the loss of black lives but also on a lack of justice.
“When a civilian has committed a violent crime, they’re generally arrested, tried and then convicted,” Franchesca Ramsey, a writer and activist who discusses race, explains in the MTV series Decoded (which you can watch here in full).
“Conversely, there’s a lot of evidence that it’s very rare to secure an indictment against a police officer for excessive force. And an indictment is just a trial; it isn’t even a conviction.”
“Black Lives Matter isn’t just about the loss of life, which is always terrible. It’s about the lack of consequences when black lives are taken at the hands of police.”

Now consider the following scenario: on Earth Prime, 10 million African-Americans are murdered by civilians every year, and 10 are murdered by police officers. All the perpetrators of the 10 million civilian homicides are convicted every year; the 10 police officers are not. Which is more important in this scenario, reducing black on black crime or police shootings of blacks?

Obviously, in this exaggerated reductio ad absurdum, no matter what your view on the real world debate is, you’d say reducing the rate of black on black crime. But read the very representative USA Today passage again — would literally anything in its argument about why focusing on police shootings over black on black crime have changed on Earth Prime? It would still be “very rare to secure an indictment against a police officer for excessive force” and when a civilian commits a violent crime, they’d still be “arrested, tried and then convicted.”

In other words, the argument that because a case of police murder is more important than civilian murder, we should necessarily focus on the former is an argument that requires any police murder to be not only more serious than civilian murder, but literally innumerably more serious.

If 6,000 African-Americans were murdered by police officers every year, and 300 were murdered by African-American civilians, I would say that reducing the rate at which African-Americans are killed by police officers is obviously more important than reducing black on black crime. I would be very puzzled if someone were to tell me that they thought society ought to focus on reducing black on black crime instead of police shootings of blacks. We live in a world where 6,000 African-Americans are killed by civilians every year, and 300 are killed by police, however. So I find myself puzzled instead by intelligent, educated people telling me that we ought to focus as a society on reducing the rate of police shootings of blacks instead of black on black crime.

And, as I’ll elaborate later on in the essay, the relatively high rate of black on black crime isn’t disconnected from the relatively high rate of police shootings of blacks; the latter is in no small part a consequence of the former.

(Ironically, Ta-Nehisi Coates has used logic almost identical to my own here in a different context — the aftermath of the assassination of police officers by anti-cop radicals.)

“High rates of black criminality are caused by poverty and structural disadvantage”

This is again an omnipresent argument: in this USA Today piece, this Vox piece and this MTV video, for starters. Because African-Americans suffer from poverty, unemployment and “structural racism” at disproportionate rates, they also commit violent crimes at higher rates, therefore demonizing African-Americans as responsible for societal problems is wrong.

The thing is, this is… a complete non-sequitur. In order for arguments like mine to be correct, it doesn’t actually matter why blacks, whites, Hispanics and Asians commit violent crimes at different rates. It could be the result of differences in poverty rates, cultures, the kind of dirt people live above, water fluoridation, lead paint exposure or anything else you can imagine, as long as it’s something society can change. (Scott Alexander makes a somewhat similar point here.)

As it happens, I think the evidence suggests strictly economic factors have not entirely explained changes in the aggregate crime rate or differences in crime rates (and particularly homicide rates) between groups (though they certainly play some role.) For some more on this, you can take a look at Barry Latzer’s fantastic new book or this long form post by the data blogger Random Critical Analysis. But again, this isn’t strictly relevant to my argument here.

Conclusion: broader issues in the American discourse about race and criminal justice

I’ve made a specific, narrow, argument in this essay: that disproportionate black on black crime is an objectively greater problem than, and part of the cause of, disproportionate police shootings of blacks. (Especially for a movement calling itself “Black Lives Matter.”) But, venturing into the realm of irresponsible, un(der)sourced generalizations, I think the the analytical flaws in the arguments I’ve tried to address here reoccur in many conversations progressives have about race and criminal justice. Progressives really like to talk about ways that (in their views) the criminal justice system treats people of different races (especially white and black people, sometimes Hispanics, almost never Asians) differently for the exact same behavior. In a perfect example, Vox dot com editor-in-chief (and very milquetoast liberal) Ezra Klein wrote an article straightforwardly headlined: “America’s criminal justice system is racist.” (And also Michelle Alexander’s extremely popular book “the New Jim Crow” is predicated on this notion.) Some on the new “social justice” (or “politically correct”) left go even further and actually deny the empirical reality of racial differences in crime rates, such as popular (at least, popular among people my age) social justice bloggers the Love Life of an Asian Guy and Kat Blaque. (Amusingly, in a follow up post, when asked for sources the proprietor of the former blog made the claim that asking for sources is racist.)

But I think the evidence (as reviewed by Heather Mac Donald here and Scott Alexander here) militates against the strong form of that conclusion. The reason there are racial disparities in the criminal justice system is mostly because the system treats groups of people who behave differently on average equally; not because the system treats groups of people who behave similarly on average unequally. Group differences in behavior, in a big picture sense, are what we need to change as a society. American society with a perfectly non-discriminatory criminal justice system would be better, but it wouldn’t be all that much better than society currently. American society without racial differences in crime would be much, much better (particularly for non-Asian racial minorities) than society currently. If progressives want to convince people that living in a multiracial, multicultural society is a good thing (and I think it is), they should think harder and talk more about equalizing racial crime rates (hopefully by lowering the higher ones.)