In the five years after the demise of George Floyd on May 25, 2020, what did we learn from the racial reckoning?

I guess we learned not to do it again.

Or so I hope.

Hence, let’s take a look at the data on the issue that was said to be the cause of the storm: Is there racial bias in police killings?

It hasn’t been talked about much since Trump’s reelection, but law enforcement’s use of deadly force was an obsessive topic during the great awokening. Here’s Google Ngram’s graph of all the times “police shootings” appeared in American books from 1900 to 2022:

Yet all the hubbub was an almost complete flop at reducing fatal police shootings, which have gone up every single year since 2016:

In early June, three political scientists, Tom S. Clark, Adam N. Glynn, and Michael Leo Owens, will release a book called Deadly Force: Police Shootings in Urban America that argues that their new database shows that cops really are biased against young black men.

I haven’t read the unpublished book, but I did compare their findings as teased in a U. of Chicago press release to the data that’s publicly available on The Washington Post’s Pulitzer-winning “Fatal Force” website of all police shootings in the country, which the Post started right after Black Lives Matter emerged over Ferguson in 2014.

I’m not sure why the professors are reinventing the wheel with their own database, but it can’t hurt to get a second perspective. (One shortcoming of the Post’s database is that it doesn’t include non-firearm police killings, although I’ve never seen anybody argue that the extra data would matter much.)

The book’s PR write-up points out that, unsurprisingly, many police departments are not exactly forthcoming about shootings:

…About 100 departments [out of the 300 biggest departments in the country] refused to provide their records, sometimes pointing to obscure, vague provisions in the open records law. Some cities claimed not to have information on the whereabouts of shootings or the date, and still others didn’t have information on the subject of the shooting. Some departments consider the fact of a police shooting to be part of a personnel record and withheld information about the officer in certain cases.

Even for less politicized types of data like traffic fatalities, America’s federalism makes our data much slower and less certain than in well-organized countries like Japan. That’s why private organizations like the Post have built their own databases.

“One of the key pieces of information we wanted to know was whether a suspect was armed, and we learned that information in only about three out of every 10 shootings,” Clark says.

Interestingly, The Washington Post’s database of fatal police shootings reports that in 2018–2023, 85.2 percent of police shootings were of armed suspects and another 3.0 percent were of suspects with replicas of weapons, for a total of 88.2 percent in which the cops at least claimed the dead man (95.3 percent of police shootings are of men) had a real or fake weapon.

The dead man was unarmed in 4.9 percent of the cases, and another 6.9 percent of fatal shootings were supposedly “Undetermined” or “Unknown.”

I’d be interested in the discrepancy between the professor’s figure and the Post’s.

In the movies, dirty cops are always planting guns on people they’ve killed. How often that happens in real life is hard to say. The anti-police Marshall Project lists 29 news stories about cops planting evidence of all kinds since 2017, but eight of the links are about four Baltimore cops convicted of planting a BB gun on some poor guy one of them had run over with his police car.

Thinking about the logistics of planting guns, I really doubt that many cops drive around every day with a spare unregistered gun to plant if they happen to shoot somebody who turns out not to have been a threat. In the Baltimore case, for instance, four policemen got involved before one who owned a BB gun could be found.

Of course, the term “unarmed” can cover a fair amount of ground. I suppose NFL lineman-size Michael Brown of Ferguson was technically unarmed when he tried to wrestle the policeman’s gun away. And Trayvon Martin was unarmed when he was pounding George Zimmerman’s head into the street with his fists. Still, both survivors appear to have had genuine reasons for fearing for their lives.

Does “unarmed” mean driving a car? It does sometimes in the Washington Post database. A decade and a half ago I did some amateur sleuthing regarding a law enforcement shooting in a parking lot near the corner of Laurel Canyon and Ventura in the San Fernando Valley, across the street from the bookstore I frequent. The police initially blamed it on all the drug dealing going on in the parking lot, which set off my doubts. I’ve been in that parking lot approaching 1,000 times (literally) and never have witnessed any lawbreaking other than bad driving.

And that’s what eventually the story turned out to be. A bunch of plainclothes cops “debriefing” in the parking lot saw a guy they thought was behaving suspiciously and started hassling him, but because they were out of uniform the hasslee thought the hasslers were criminals. A fight broke out, and then—follow me closely here—another guy driving around the parking lot decided he’d better get the hell away from this bizarre brawl, and he hit one of the DEA agents, whose partner immediately shot the driver. And then the shooter cop screamed at the dying young man, “Why’d you make me do that?”

A week later, I ran into the mother of the dead driver who was in the parking lot trying to figure out what happened. I told her to sue. She was eventually awarded $3 million. Because the dead man was white, the story received negligible media coverage until the civil trial verdict.

In Texas and California, for example, the states compel every police department to share police shooting information with the state attorney general: It must be reported routinely and in a consistently formatted manner.

That’s not too much to ask of police departments: When you shoot somebody, you ought to fill out a lot of paperwork.

In a high-level analysis of the data, the researchers looked at the predictors of how many shootings happen in a city each year. The major finding in this instance is that as violent crime goes up, police shootings become more common; however, as police departments get larger, there are fewer shootings.

“[In this case] our inference is largely that police officers have the resources and the capacity to be able to deal with confrontational situations without resorting to lethal force,” Clark says.

A criminal trying to get away would be more likely to shoot a lone cop than two cops because the one you don’t shoot first will try to shoot you.

The more cops the better is also true with people having a psychotic break. A very large friend of mine, a former college football lineman, got dangerously manic, so the well-staffed Chicago Police Department sent four cops to jump on him at once. The four were able to wrestle him into a straitjacket without having to shoot him, tase him, or even bounce their truncheons off his skull.

Good job, officers. Thanks.

Looking within cities, the team explored where shootings were more likely to happen. The neighborhoods with the highest rates of shootings had more poverty and more violent crime. They also find that as violent crime goes up, police shootings increase—but they found this to be particularly true in neighborhoods with larger Black and Hispanic populations. Police respond to crime more violently in neighborhoods with larger Black and Hispanic populations.

Might have something to do with Hispanics and, especially, blacks being more shooty on average. In 2018–2023, according to the CDC, blacks died by firearm homicide thirteen times as often per capita as whites. You might think that it would be hard to work at the U. of Chicago on the South Side without being aware of this fact. But sophisticated Americans seem to be able to turn off their cognizance whenever they choose.

Next, they looked at the individuals who were shot. They found that a city’s Black population is far overrepresented relative to their size among the city’s population, relative to their size among the criminal suspect population, and relative to their size among the violent criminal suspect population.

Relative to their size among the illegal-handgun-shooting population?

There’s a common bit of verbal sleight of hand using the term “violent criminal” to minimize exactly how much more likely young blacks are to pull out a piece and start shooting: “Violent criminals” include a lot of bar brawlers and wife beaters of all races, but shooters tend to be very black. At the peak of the racial reckoning in 2021, the FBI reported that 60.4 percent of known murder offenders were black, and the CDC reported that 55.0 percent of homicide victims were black.

Most people being shot are armed (mostly with guns), but there are also many unarmed shootings.

White men, regardless of age, are more than 95% likely to be armed if shot.

In the Washington Post database for the years 2018–2023 (a span of six years I chose for the convenience of comparing it to CDC mortality data), 88.58 percent of the dead white men are listed as armed or having a replica weapon, as are 88.64 percent of the dead black men. That’s just about the closest racial equivalence I’ve ever seen in a half century of looking at stats.

A 50-year-old Black man is more than 92% likely to be armed.

The Washington Post lets you look at age 45 and up, for which I see 90.0 percent of older black men were armed (or with a replica) and 90.2 percent of older white men.

But an 18-year-old Black man is only 77% likely to be armed.

This wouldn’t surprise me if it replicated because young African American males tend to have a culture (as exemplified in countless hip-hop videos) of trying to look like they could be carrying even when they aren’t.

But does it replicate? The Post lets you look at Under 18 or 18 to 29, so I’ll look at 29 and Under: 88.8 percent of younger black males were armed or carrying a replica weapon versus 87.9 percent of younger white males.

“Another way of putting it is, if someone is unarmed and shot, they are most likely to be a young Black man,” Clark says.

…“We use the data to estimate how much racial bias there is in the decision to shoot, and we estimate that between 10 and 30% of all Black people shot would not have been shot had they been white,” Clark says.

Perhaps, although I’d like to hear why that hasn’t been replicated in the Post database (which is not paywalled). Think about it: If there was slam dunk evidence of racial bias in this prizewinning database, wouldn’t you have heard about it by now?

A question of apparently less interest to the researchers is: Why do whites get shot and killed by the cops so much more than blacks do relative to their homicide rates?

We can look at the CDC’s numbers for two causes of death: homicides and legal interventions (i.e., police shootings). (The CDC’s numbers for legal intervention killings are only about 70 percent as big as The Washington Post’s numbers for police killings. I don’t know why. I’ll use the larger numbers from the Post.)

The CDC doesn’t track who committed the homicides, but we can get a feel for the differences from looking at the race of victims of homicides and legal interventions. (I’ll lump the two categories together from now on and just call them violent deaths.)

In 2018–2023, the CDC reports, 63,255 black males died violently. The Washington Post reports 1,461 black males died at the hands of law enforcement, or 2.3 percent of all black male deaths by homicide or legal intervention.

In contrast, 25,370 white males died by homicide or legal intervention, according to the CDC, with 2,663 being killed by the police, or 10.5 percent.

So why do the cops kill 4.5 times as great a percentage of white male victims of violent death than of black male victims of violent death?

I don’t think it’s racist. I think it’s just more likely that whites (and Hispanics) engage in suicide by cop more than do blacks.

But nobody much cares.