Police violence, racial bias & the Black Lives Matter protests
What proponents and critics of the BLM movement are each right about
George Floyd’s death gave rise to the largest nationwide protests in the history of the United States. Across the country, and even across the world, people were outraged by the video of a white police officer kneeling on a black man’s neck for almost 8 minutes as he begged for his life and cried out that he couldn’t breathe. The incident gave renewed life to the Black Lives Matter movement, which for years has been bringing attention to the disproportionate violence black people suffer at the hands of the police.
But critics of BLM claim that when you look at the studies and the data, there is actually little to no evidence of racially motivated violence by the police. They argue that the entire BLM movement is built on a false narrative, and that proponents are acting on an outrage sparked by a handful of isolated but heavily publicized events.
Proponents of BLM argue that critics are cherry-picking their studies and that the data shows plenty of bias. They argue that critics are deliberately avoiding the obvious conclusion and that many of them are doing so to fight change and maintain the status quo.
There is no shortage of articles and podcasts from both proponents and critics arguing for their points of view. There’s also no shortage of Twitter spats on the topic. What is rare is a civil, long-form conversation between two people on opposite sides of this debate: in a series of letters, Matthew Thornton and Ali Rizvi exchanged a number of civil arguments around the topics of police violence, racial bias, systemic racism, the protests, and others. In just under 13,500 words, each passionately argued for their own side and picked apart the others’ arguments as best they could. They agreed on some things and very much disagreed on others.
My synthesis of Matthew and Ali’s viewpoints is below. I used their letter exchange as a jumping off point but also dug into the evidence they cited as needed. I tried to answer three separate questions:
What role does race play in police shootings and other fatal police encounters?
What role does race play in the criminal justice system, outside of police shootings?
How supportive should we be of the protests?
What role does race play in police shootings and other fatal police encounters?
To begin, let’s establish two points that are relatively uncontroversial:
The majority of those who are shot and killed by police are white, not black. Black people make up about 25% of police shooting victims year after year.
Black people are significantly more likely to be shot and killed by police than white people. They make up 13% of the general population, so are 3.5 times more at risk than white people.
Next, let’s note that there are two ways this disparity could have arisen: by black people having a disproportionate number of encounters with the police, or by black people dying at the hands of police in a higher percentage of encounters.
The data points to the disparity being almost entirely driven by the former. Fryer (2016) found that while black people and Hispanics are 50% likelier than white people to experience some sort of force in their interactions with the police, there were no racial disparities in shootings across encounters, a finding he’s referred to as “the most surprising result of my career”. He didn’t find a racial disparity in the raw data of police shootings and he didn’t find one when contextual factors were taken into account either.
Fryer’s work and its frequent citing by skeptics of the BLM movement have been criticized for a few reasons:
Fryer’s paper was never officially peer-reviewed. (Though, Daniel Engber points out that this officiality doesn’t mean much and Fryer’s work was reviewed very thoroughly by a number of his peers.)
Fryer’s data relies almost entirely on police reports written by the involved officers themselves. A recently released NYC health department report, a report that was buried for nearly three years, found that more than half of all deaths related to legal interventions between 2010 and 2015 weren’t reported and that racial disparities in the newly uncovered deaths were significant. In this case, most of the deaths had little or nothing to do with the arresting officers and were issues of death coding. But this incident shows how the police could be under-reporting the deaths of certain racial groups and this wouldn’t show up in the data.
Because of collider bias, Fryer’s finding can’t be used to say there are no racial disparities in police shootings as a whole, but often is. It can only be used to say that blacks are not any likelier than whites to be shot and killed in any particular encounter.
When it comes to the role of racial bias in the police deciding to shoot, the data appears mixed. For instance, Nix et al. (2017)’s analysis of the 2015 police shooting data found that black shooting victims were more than twice as likely to be unarmed than white shooting victims, and that minority groups were significantly less likely than white people to have been attacking the officer or other civilians when shot. However, James et al.’s experiments from putting 80 police officers into “strikingly real deadly force simulators” found that officers were slower to shoot armed black suspects than armed white ones, and that they were less likely to shoot unarmed black suspects than unarmed white ones.
Ultimately, there does not seem to be any data or studies that directly contradict Fryer’s finding that there are no racial disparities in shootings across police encounters. Other studies like Miller et al. have also found the same thing.
Bullets aren’t the only thing the police can use to kill civilians, but police shootings are what get talked about the most and what most studies focus on. There don’t seem to be any claims that the racial data on police shootings doesn’t generalize to all fatalities; shootings account for about 90% of civilians deaths at the hands of the police anyway (The Guardian).
If the police aren’t more likely to kill a black person than a white person in any given encounter, this implies that the racial disparity in shooting deaths is driven almost entirely by the number of encounters. Here’s how physicist Neil deGrasse Tyson put it:
Some studies show that the risk of death for an unarmed person at the hands of the police is approximately the same no matter the demographics of who gets arrested. Okay. But if your demographic gets stopped ten times more than others, then your demographic will die at ten times the rate.
So why do black people have more encounters with the police? Broadly speaking, there are two proposed explanations for this: rates of offending and racial bias.
The first explanation is that this discrepancy is the result of the police trying to stop violent crimes. In 2018, black Americans made up 53% of known homicide offenders and accounted for 60% of robberies; the numbers are similar for assaults and other violent crimes. These numbers are relatively stable over time: from 1980 to 2008, black people committed 52.5% of all homicides. Since 90% of all homicides are intraracial, the reliability of the homicide offender data can be attested to by the similar percentage of homicide victims of each race. The reliability of the offender data for other crimes can be attested to by anonymous victim reports in the National Crime Victimization survey.
Skeptics of the BLM movement argue that this first explanation can fully, or almost fully, account for the racial discrepancy in police encounters.
But other evidence suggests otherwise. The most common source of police encounters are traffic stops, where black people are 40% more likely than white people to get pulled over. This disparity lessens significantly at nighttime, i.e. when the driver’s race can’t easily be identified. This shows racial bias in who the police are deciding to pull over, a bias that’s difficult to square with the hypothesis that the increased stops are just due to the police trying to prevent violent crimes.
So how much of the discrepancy can be explained by differences in the racial disparities in rates of offending, as opposed to racial bias?
Fryer (2016) found that black people were about 4 times more likely to be stopped relative to their proportional population, but that this bias starts disappearing when considering the crime rates of each race. Because crime rates differ from crime to crime, there are multiple ways of doing this analysis: in one analysis, nearly 2/3rds of the bias disappeared; in another, almost all the bias disappeared; in another, the bias went the other way. Fryer summarized his results as follows:
Put simply, if one assumes police simply stop whomever they want for no particular reason, there seem to be large racial differences. If one assumes they are trying to prevent violent crimes, then evidence for bias is exceedingly small.
Ross (2015) found that the racial bias observed in police shootings could not be explained as a response to local-level crime rates. He compared county-level racial bias in shootings with the local crime rates and found no relationship between the two. He also found no relationship between the racial bias in shootings in an area with any of the area’s race-specific crime rates. Unfortunately, the paper doesn’t seem to have looked for a relationship between racial bias in shootings and the crime rate disparities between races, which would have been the primary analysis of interest.
Johnson et al. did do this analysis and found that the each race’s rate of offending in an area strongly predicted shooting victims of that race: “[A]s the proportion of violent crime committed by black civilians increased, a person fatally shot was more likely to be black”. (This study has been criticized and since been retracted because it also found that the officer’s race was not predictive in predicting the race of shooting victims, and used that to claim that there was no evidence of anti-black or anti-Hispanic disparities across shootings, a claim that can’t be supported by that finding alone (collider bias again). This criticism and retraction appear to have nothing to do with the section determining that rates of offending were predictive, however.)
These studies provide evidence that crime rate disparities between races are an important factor in the increased encounters black people have with the police. The aforementioned traffic stop study, however, shows that racial bias is also a prominent factor.
What role does race play in the criminal justice system outside of police shootings?
As mentioned before, Fryer (2016) did find racial bias in the police’s use of force. Blacks and Hispanics were more than 50% more likely to encounter some kind of force in their interactions with the police. While adding controls for context and civilian behavior reduces this disparity, it does not fully explain it either.
But that 50% number is when the dataset is based off police reports. When the data is from civilians who are asked about their experiences with the police, this number shoots up to 350%. This holds true for every level of nonlethal force, from officers putting their hands on civilians to officers striking them with batons. Controlling for relevant variables could reduce about 2/3rds of the bias, but that would still mean black people are more than twice as likely to endure police force.
Even black people who were recorded as compliant by police were 21% more likely to experience force than white people who were recorded as compliant. While compliance makes people of all races less likely to experience force, this benefit was larger for whites than blacks. Wilson et al. found that people perceive young black men as more physically threatening than young white men and this leads them to justify hypothetical uses of force against black crime suspects.
It isn’t just police brutality either. In Los Angeles, 24% of black drivers are searched in comparison to 16% of Latinos and 5% of whites. But the chance of one of these searches finding drugs, weapons or other contraband is highest for whites at 20%, with black people and Latinos possessing contraband in 16-17% of cases.
Black people are also more likely to be charged, incarcerated and receive longer sentences for similar crimes than whites. The racial disparities in sentencing are even present between lighter-skinned, medium-skinned and darker skinned blacks, who receive progressively longer sentences. While light-skinned blacks’ sentences weren’t statistically distinguishable from whites’ after controlling for the legally relevant factors, dark-skinned blacks received sentences nearly 5% longer. Gelbach & Bushway found that judges seemed to “set bail as if the value of blacks’ lost freedom is less than two-thirds the value of whites’ lost freedom”.
Put together, all of this accrues to a criminal justice system that treats black people harsher than it does white people. While some of the raw disparities can be explained by factors like crime rates and socioeconomic status, there is also strong evidence of racial bias.
How supportive should we be of the protests?
An estimated 16 to 25 million people participated in the George Floyd protests, and the overwhelming majority marched and protested peacefully. A central tenet of Black Lives Matter – that racist policing is responsible for black people being killed at disproportionate rates – has little evidence supporting it. The narrative that there is an epidemic of racist white cops murdering innocent black men is especially unsupported by the data; those cases are simply what go viral.
But there is much evidence that black people face worse treatment by the police and in the criminal justice system than white people and that these differences are driven by racial bias. These disparities have downstream effects: higher crime rates may help explain racial disparities in police encounters, but unjust systems and institutions help perpetuate those higher crime rates.
There is a separate but related problem of the police often not being held accountable for bad behavior. The vast majority of police officers who shoot and kill are never convicted. And from Rodney King to George Floyd to Martin Gugino, there are many instances where video cameras and police reports have told very different stories.
Black Lives Matter has been leading the charge to demand both racial justice and police reform. The seriousness of these issues can be attested to by the fact that 74% of Americans, including 53% of Republicans, have expressed support for them. Since the demonstrations began, NASCAR has gotten rid of Confederate flags, the NFL has begun supporting players like Colin Kaepernick taking a knee during the anthem, and many major corporations have shown solidarity.
Several cities have passed, or are at least seriously considering, all sorts of policies that were not on the table just a few weeks before. Greater investment is being made in police reform, criminal justice reform, public education, social work, mental health resources, and more. Some of these policies have already been tried and tested: Camden, NJ successfully dismantled and dramatically restructured its police force in 2012. But others appear misguided, like Portland ending its School Resource Officers (SRO) program, even though SROs have been found to reduce sexual abuse in children, guns in school, and overall violence.
When demanding such sweeping changes to police departments, it is worth putting the death counts of civilians at the hands of the police in perspective.
First, America is a nation of 330 million people, 393 million guns, and 375 million police encounters a year. Of the 990 fatal police shootings in 2015, an attack by the civilian was already in progress in 74% of them and the other 26% weren’t necessarily illegal either. This gives us an upper bound on the scope of the problem: there is less than a one in a million chance of any police encounter leading to a civilian death that isn’t obviously justifiable. George Floyd’s death stands out because it is an aberration; it is not the norm.
Second, there were 16,214 homicides in the US in 2018. When the race of the victim was known, 53.3% of the victims were black, meaning roughly 7,407 black victims. This means that even if the police never killed a black person again, this would only bring the murder rate of black people down by 3%. There are summer weekends in Chicago where over 100 people get shot: this is the level of violence in the country that we ask police officers to risk their lives to bring down.
These figures don’t mean we should accept the current levels of police brutality and unjust killings: a single instance is too many. But they serve as a powerful reminder for why cultivating a sense of distrust and contempt for the police is dangerous. It’s important to treat individuals as individuals and not based on their group identity. “ACAB” may be a popular slogan but the officer who murdered George Floyd does not represent the nearly 700,000 law enforcement officers in the country.
The protests have also involved looting, burning, vandalizing and violence. Many in the media have justified and rationalized these incidents. The actions of the rioters and their defenders are not only deeply immoral, but gifts to the far right and to the re-election prospects of Donald Trump.
But the violence and looting can only be declared a prominent feature of BLM if a significant number of protesters partook in or endorsed it. About 14,000 people have been arrested – even if we assume there were no wrongful arrests and that 10 rioters got away for every arrestee, this still means that 99% of the protesters were peaceful. George Floyd’s family has denounced the violence, as did former president Barack Obama.
Every social movement in history has had its share of opportunists and assholes. Even the most worthwhile movements have odious characters partaking and sometimes even leading. The protests and public uprisings of the past played a key role in making America one of the least racist, misogynistic, and homophobic countries in the world today – and those struggles were riddled with violence and turmoil as well.
Whatever one’s opinions on them, these protests will help shape America for years to come.