Police Shootings Are Said to Be "Disproportionate" for Certain Groups . . . But Disproportionate to What?

The evidence suggests they are not disproportionate to the deadly threat posed.

Above: a Black Lives Matter protest in Paris. How many of these people do you think know the statistics set forth in this email?

The Washington Post database of police shootings explains that one of the core beliefs behind the creation of the database is the “fact” that police are shooting black people (and Hispanics) at a “disproportionate” rate.

Although half of the people shot and killed by police are White, Black Americans are shot at a disproportionate rate. They account for less than 13 percent of the U.S. population, but are killed by police at more than twice the rate of White Americans. Hispanic Americans are also killed by police at a disproportionate rate.

But why does the Post define what is “disproportionate” by reference to the percentage of black (or Hispanic) Americans in the population at large? Wouldn’t it make more sense to ask what percentage of the population that poses a deadly threat to police is black or Hispanic?

That is the more relevant question — and there is a way to assess those percentages for any given population.

First, let me clearly state my thesis:

  • If police shootings are a response to deadly threats, rather than motivated by bias, then the total percentage of people (armed or not) of any population shot and killed by police should roughly correspond to the total percentage of people (armed or not) in that same population who pose a deadly threat to police.

Whenever some subset of a specified population — blacks, whites, men, women, you name it — regularly poses an imminent deadly threat to police officers, the subset posing a deadly threat will obviously include both a) people who kill police officers, and b) people who are killed by police officers. If a) is higher in any given population, you would expect to see b) higher for that same population as well.

So, if police shootings generally reflect an unbiased response to deadly threats, then for any given population, the percentage of those who kill police officers will roughly correspond to the percentage of those who are killed by police officers. If, by contrast, police shootings are largely motivated by bias, then for the population against whom police officers are allegedly biased, you will see a far higher percentage killed by police officers than the percentage who kill police officers.

Note that this analysis has nothing to do with the percentage of such people in the population at large. Males are about half the population, but why would you need to know that to know whether police are shooting them for sexist reasons? Similarly, why would you care what percentage of the population is purple if you are trying to determine whether police are shooting them because of their purple skin?

What if I told you that roughly 34% of unarmed victims of fatal police shootings are black — but 37% of known killers of police are black? Wouldn’t you conclude that police officers are killing unarmed black folks at a slightly lower rate than is justified by the actual threat black people are posing to police officers?

I think you would.

And, as you may have guessed, that is the reality in which we live . . . as I am about to show you.

A brief note: the analysis below first discusses statistics related to shootings of “unarmed” people. In the past, I have taken pains to show that a suspect’s status as “unarmed” does not compel a conclusion that a police shooting of that suspect was unjustified. Similarly, a suspect’s status as “armed” does not require one to conclude that a police shooting was justified. Nevertheless, I have focused in the first instance on unarmed “victims” of police shootings, because those shootings are generally thought (rightly or wrongly) to be the most questionable. It is generally easier for people to accept the premise that “armed” people pose a deadly threat to police than to accept that “unarmed” people do.

However, at the end of this analysis, I also analyze all the statistics for all police shooting “victims” — whether armed or unarmed. As we will see, those statistics are even less favorable for the thesis of the Black Lives Matter movement.

FBI CRIME REPORTING STATISTICS

The FBI’s Uniform Crime Reporting (UCR) Program tracks all kinds of statistics relating to incidents where law enforcement officers are killed. The most recent year for which such statistics are available is 2019 (2020 is not yet available). Table 42 tracks characteristics of known offenders in incidents where law enforcement officers were killed. The data go back to 2010.

According to the table, from 2010-2019, there were 537 known offenders in situations where law enforcement officers were feloniously killed. Of that number, 199 (the highlighted number) of the killers were black, meaning 37% of known killers of police are black.

Meanwhile, my analysis of the Washington Post police shootings database (see the end of this email for details) shows that 136 of 402 “unarmed” people fatally shot by police have been black. That means 34% (136/402) of unarmed police shooting victims are black.

As I said above: if 34% of unarmed victims of fatal police shootings are black, but 37% of known killers of police officers are black, these numbers strongly tend to disprove the assertion that police shootings are motivated by racism. Instead, these figures tend to corroborate the assertion that police shootings are generally a response to a suspect who poses a deadly threat to the officer.

These percentages (37% of known cop killers are black; 34% of unarmed victims of fatal police shootings are black) remain constant even when the analysis is limited to the time period where the two databases fully overlap: 2015-2019. I undertook that analysis (see below) to address the possibility that statistics for different years are meaningfully different — due to factors such as changes in public policy, or the recent increased public focus on police shootings. It turns out that the percentages don’t change.

I did a breakdown by sex as well. This breakdown shows that 93% of “unarmed” people killed by police since 2015 have been male. But males are a little less than 50% of the U.S. population. Is this evidence of systemic sexism against males?!? No. Not when you consider that, according to FBI statistics, fully 97.3% of known cop killers are male.

Again: comparing the statistics on unarmed males shot by police to the percentage of males in the population at large is not the relevant question. The question is: what percentage of people who pose a deadly threat to police is male? And, based on the fact that 97% of people who manage to kill police officers are male, it is easy to understand why a similarly high percentage of unarmed police shooting victims are male (93%).

I did a similar analysis of shootings of unarmed white suspects. As you will see if you slog through the math at the end of this newsletter, this was an especially difficult calculation, for reasons I will not belabor here but will explain in the extended entry. What you need to know here is that, to be able to compare the two databases, I had to perform calculations for non-Hispanic whites — both for those who killed police officers, and for those who were killed by the police while unarmed. It turns out that roughly 42% of unarmed police shooting victims are non-Hispanic whites, while the percentage of non-Hispanic white killers of police officers is somewhere between 42% and 43.3%. Again, the numbers are roughly in sync and dispel any inference of bias in police shootings based on skin color.

Having undertaken the complicated analysis of disentangling white Hispanics from non-Hispanic whites, I also performed a calculation of shootings of unarmed Hispanics. The quote at the beginning of this post, from the Washington Post database, claims thatHispanic Americans are also killed by police at a disproportionate rate.” This quote is not supported by the data. I found was that 18.4% of unarmed people shot by police from 2015 to the present were Hispanic . . . while Hispanics or Latinos form about 18.5% percent of the population, according to Census figures. Meanwhile, about 14.3% of known killers of police have been Hispanic — slightly lower than the percentage of unarmed Hispanic police shooting victims, but not much.

As I said, the above numbers relate to shootings of unarmed individuals from various populations — but “unarmed” does not equate to “posed no danger” any more than “armed” equates to “posed a deadly threat.” So, out of a pigheaded and OCD tendency towards completeness, I tortured myself further by compiling all of the above statistics for the total numbers of people shot by police, whether armed or not.

When you look at all victims of police shootings, and not just unarmed victims of police shootings, you notice something very interesting. For most populations, the percentages line up pretty well with those populations’ share of cop killers. For example:

  • 95.5% of the victims of fatal police shootings are male, compared to 97.3% of known cop killers who are male.

  • 46% of the victims of fatal police shootings are non-Hispanic whites, compared to 42-43% of known cop killers who are non-Hispanic whites.

  • 16.8% of the victims of fatal police shootings are Hispanic, compared to 14.3% of known cop killers who are Hispanic.

The numbers line up pretty closely . . . with one fairly glaring exception: the population of black people:

  • Only 24% of the victims of fatal police shootings are black, compared to fully 37% of known cop killers who are black.

The statistics seem to show that black people are far less likely to be fatally shot than one would predict based on their percentage of known felonious killers of police. The numbers only approach (but still do not reach) that latter percentage when you look only at statistics relating to unarmed victims of fatal police shootings.

Remarkable, isn’t it?

I have not seen an analysis like this anywhere else. Robert C.J. Parry did a similar analysis back in 2016, when the Washington Post database first came online — and his piece gave me the idea to conduct this analysis. But I have seen nothing like it since. That is not to say that nobody else has ever done such an analysis. All I’m saying is: if such an analysis exists, I have not run across it. (Otherwise I would not have wasted all this time doing it myself!)

Frankly, I find that to be an astounding lacuna in the Big Media reporting on this story.

How many times have you heard a TV talking head say that black males are “disproportionately” shot by the police? How many times have you read an online op-ed that argues that police shootings are usually motivated by racism?

And how many of those talking heads or op-eds have even mentioned the issue of whether the group in question (usually blacks, black males, or young black males) poses a disproportionate danger to police on a daily basis? Who even thinks to ask the question?

I suggest you bookmark this piece. (If you’re getting it by email, you can click on the title at the top of the email to access a bookmark-suitable version on the Web.) From now on, when people make these arguments, you’ll have the facts at your fingertips.

P.S. If you’re interested in the nitty-gritty of the math, and you would like to check my arithmetic, I have set forth my calculations below. Normal people can stop reading here; obsessives and fact-checkers can find details of my calculations below the line.


THE MATH

As stated, I am using the 2019 statistics from the FBI’s Uniform Crime Reporting (UCR) Program tracks all kinds of statistics relating to incidents where law enforcement officers are killed. I have already provided above the math for the percentage of unarmed black Americans killed by police: 199/537, or 37%.

Meanwhile, for the time period covered by the Washington Post database (2015 to present), we can click on the filters and find 402 “unarmed” people shot and killed by police during that time period:

Of that number, 136 were black:

That is about 34%.

As I note in the main piece, the databases overlap. The FBI’s Uniform Crime Reporting (UCR) Program goes back to 2010, further than the Washington Post database of police shootings — and the former ends in 2019, while the latter is kept more or less up to date even today. As a result, as noted above, I decided to apply a more granular analysis, assuring that I would be comparing apples to apples. So, I compared the narrower time period in which the two tables fully overlap: namely, the time period of 2015 to 2019. The tables do not allow automatic tabulation in all instances, so I had to do much of the arithmetic myself.

As you will see, the percentages really don’t change, remaining constant at 37% of known killers of police officers being black, and 34% of unarmed victims of fatal police shootings being black.

Killings of police by known offenders, 2015-2019

  • Killings of police by all known offenders, 2015-2019 (from the table above): 37 + 56 + 44 + 56 + 49 = 242

  • Killings of police by known black offenders, 2015-2019 (from the table above): 18 + 17 + 16 + 23 + 15 = 89

  • TOTAL: 89/242 = 36.8%

  • 36.8% of known killers of police officers were black, 2015-2019

Unarmed people killed by police, 2015-2019

  • All unarmed people killed by police, 2015-2019 : 95 + 58 + 71 + 58 + 54 = 336

  • Unarmed black people killed by police, 2015-2019 (from the Washington Post database of police shootings): 38 + 19 + 22 + 23 + 12 = 114

  • Total: 114/336 = 34%

  • 34% of unarmed people killed by police were black, 2015-2019

(For the remaining sets of statistics in this email, I am returning to the databases as a whole, and not merely the years 2015-2019.)

Males

I decided to do a breakdown by sex as well. This breakdown shows that, of 402 “unarmed” people killed by police since they started keeping the database, 374 were male. 374/402 = 93%. Meanwhile, 523 of 537 known attackers were male. 523/537 = 97.3%.

Whites — Specifically, Non-Hispanic Whites

If you consult the Washington Post database for shootings of unarmed whites, you find 168 white “unarmed” police shooting “victims” (168/402 = 41.8%).

How does this stack up to known white killers of police officers? At first glance, if you consulted the FBI figures in table 42 linked above, you might believe that the correct figures are as follows: 303 of 537 known killers of police officers have been white (303/537 = 56.4%). Which would lead you to a seemingly startling (but as we will see, misleading) conclusion: 56.4% of cop killers are white, but only 41.8% of unarmed police shooting victims are white. Whoa! Now we’re really starting to see a racial disparity, right? Maybe the BLM people were right after all! Cops are biased in favor of white people when they shoot to kill!

Actually, no . . . because of the confusing way the FBI keeps crimes statistics relating to race and ethnicity. The FBI statistics count “Hispanic or Latino” as an “ethnicity” but have no category for “Hispanic” or “Latino” in their racial category . . . which means that the Hispanic/Latino ethnicity is largely folded into the “white” racial category. The way this is done for FBI crime statistics is described in detail at this link. The Washington Post database, by contrast, separates “Hispanic” into its own category, meaning that when they use the term “white” they mean “non-Hispanic whites” which, according to Census figures, are about 60% of the population.

So if you want to compare apples to apples, you have to figure out what percentage of killings of police by known offenders are committed by non-Hispanic whites. We know from the table that whites committed 303 of the killings, but it also says that 77 of the killings were committed by Hispanics. But of what race? Well, about 2.5% of Hispanics nationwide identify as black. Meanwhile:

The respondents in the "some other race" category are reclassified as white by the Census Bureau in its official estimates of race. This means that more than 90% of all Hispanic or Latino Americans are counted as "white" in certain statistics of the US government.

So of the 77 killings of police officers committed by Hispanics, if “more than 90%” were committed by someone the U.S. government considers “white,” that number is at least 70. That’s the number we will use, although it could be higher. To learn how many of these killings were committed by non-Hispanic whites, you have to subtract the figure we are using to represent Hispanic killers of police officers (70) from the figure representing all “white” killers of police officers (303). That figure is 303-70 = 233. The percentage in that calculation would be 233/537 = 43.3% of known cop killers who are non-white Hispanics. (If all 77 of the Hispanic killers listed in the FBI database were white, the total number of non-Hispanic white killers of police officers would be 303-77 = 226, and the percentage would be 226/537 = 42%.)

Now that you know that the percentage of non-Hispanic white killers of police officers is somewhere between 42% and 43.3%, according to our calculations, it seems entirely unsurprising to learn that roughly 42% of unarmed police shooting victims are non-Hispanic whites. As with virtually all of this statistics, when you analyze the data correctly, and focus on the correct issues, the numbers line up very well.

Hispanics

The Washington Post quote that opened this email said “Hispanic Americans are also killed by police at a disproportionate rate.” I’m not sure how they support this claim. According to their own database, 74 of 402 unarmed people shot by police from 2015 to the present were Hispanic. That’s 74/402, or 18.4% of the population. Hispanics or Latinos form about 18.5% percent of the population, according to Census figures. How is that disproportionate? Meanwhile, 77 of 537 known killers of police officers have been Hispanic, according to FBI figures. That number is 14.3% (77/537) — slightly lower than the percentage of unarmed Hispanics shot by police, but not terribly far off.

Computing Statistics for All Fatal Police Shooting “Victims” — Not Just the “Unarmed”

Here are statistics for the total numbers of people shot by police, whether armed or not. Here, the denominator is much larger than the 402 unarmed victims: 6303 people have been listed in the database as killed by the police since 2015. Here are the results in brief:

  • Black victims of fatal police shootings (unarmed and armed): 1512/6303 = 24%

    • Compared to percentage of known cop killers who are black: 37%

  • Male victims of fatal police shootings (unarmed and armed): 6023/6303 = 95.5%

    • Compared to percentage of known cop killers who are male: 97.3%

  • Non-Hispanic white victims of fatal police shootings (unarmed and armed): 2890/6303 = 46%

    • Compared to percentage of known cop killers who are non-Hispanic whites: 42-43%

  • Hispanic victims of fatal police shootings (unarmed and armed): 1058/6303 = 16.8%

    • Compared to percentage of known cop killers who are Hispanic: 14.3%

The numbers line up pretty closely, with one fairly glaring exception discussed earlier in the email: black people, who are only 24% of all shooting victims — despite being 37% of known cop killers. That’s a pretty wide gap — by far the widest gap discussed in this email.

The statistics seem to show that black people are far less likely to be fatally shot than one would predict based on their percentage of known felonious killers of police.

Does Big Media ever tell you that?