January 21, 2017

Is the Criminal Justice System Racist?

Most everyone knows that Blacks are arrested more often than whites. Many take this to mean that Blacks commit crime at a higher rate than Whites do. But many other people believe high black arrest rates reflect the fact that the criminal justice system is biased against Blacks. However, there is significant evidence to the contrary. The proportion of blacks who are incarcerated strongly reflects the proportion of crime victims who claim that they were the victim of an crime done by a black. Furthermore, government data shows that Blacks get in trouble for breaking the rules in school far more often than whites do, and this can be thought of as a young age analog to crime. Some believe that there is an iron clad case against the justice system when it comes to drug crime. But, as we shall see, such arguments are based on self report data and do not take into account racial differences in honesty or differences in the ways that Black and White drug users use drugs. After this, the argument that police are killing Blacks at unfairly high rates will be addressed. Finally, we will see that, when the proper controls are introduced, Blacks get the same punishments as Whites do when they commit the same crime.

Arrest Rates and Victimization Reports

One of the strongest arguments in favor of the validity of official arrest rates is the degree to which they correspond with victimization reports. To be more specific, the degree to which the Uniform Crime Report matches the National Crime Victimization Survey is striking.

The National Crime Victimization Survey (NCVS) is a survey carried out yearly by the Department of Justice in which a random sample of approximately 90,000 households and 160,000 individuals are asked about their experience with crime over the last 6 months. Participants are asked if they have been the victim of a violent crime in the last 6 months. If they have then they are asked to answer various questions about the crime and the perpetrator of said crime. These bi-yearly interviews are combined on a yearly basis.

The Uniform Crime Report (UCR) is a aggregation of data sent to the FBI every year by police stations all around the country. Not all police stations send in this data, but the UCR  manages to get information for police stations which have jurisdiction over 277 million Americans ( aprox 94% of the total population). The data the FBI compiles includes information on the demographics of who is arrested every year (Crime in the United States 2014, The Nation’s Two Crime Measures).

As can be seen below, there two metrics both suggest that Blacks account for basically the same amount of violent crime:


Source: Last 2015

Black Misbehavior at School 

Further strengthening the case that Blacks really do commit more crime than Whites is the fact that Blacks get in trouble at school far more often than Whites do. Just consider the following statistics:

  • A 2014 report issued by the department of education found that black preschoolers have an above average rate of suspension. This was especially pronounced among repeat offenders: though Blacks make up about 18% of preschools they make up almost 50% of preschoolers that have been suspended more than once (The Associated Press 2014).
  • Vega 2014 reported on a document from the US Department of Education which showed that in 2011 and 2012 Black females accounted for 12% of elementary school suspensions whereas white girls only accounted for 2% of school suspensions. (Most people suspended were boys.)
  • Lewin 2012 reported on a Department of Education report which analysed data from over 72,000 schools and found that Blacks made up 18% of the students population, but 35% of those to who had been suspended once, 45% of those who had been suspended more than once, and 39% of those who were expelled.
  • Skiba et al 2002 found that Black middle schoolers were more likely than White middle schoolers to be suspended even after controlling for differences in socio-economic status.

Thus, it can safely be said that Blacks are more likely than Whites to get in trouble in school.

Some will argue that this is because teachers exhibit the same racial biases that police do. However, there is evidence that is not the case. Specifically, Wright et al 2014 found that if you compare Blacks and Whites with an equal number of previous behavioral problems you find that both groups are equally likely to be suspended. Further more, Skiba et al. 2002 found that Blacks and Whites were equally likely to be suspended once they were sent to the principal’s office.

Thus, the proponent of the idea that Blacks do not commit more crime than whites would, at this point, have to adopt the point of view that crime victims, teachers, and police, all just so happen to exhibit the same biases about Blacks. That seems pretty far fetched.

Drug Crime

Some will argue that the police may very well not be bias in general against Blacks, but they certainly are when it comes to drugs. To substantiate this claim, they will cite surveys showing that Blacks are no more likely than Whites to use drugs or, as many such studies suggest, that White are actually more likely than Blacks to use drugs. The problems with this argument are two fold: first, it ignores racial differences in honestly, and second, it ignores relevant differences between Black and White drug users.

This may come off as harsh, but the first thing that has to be addressed here is that Blacks are more likely than Whites to lie about using drugs. How do we know? Well, criminologists sometimes conduct studies in which they run biological tests on people’s hair, blood, urine, etc., to test what drugs they have recently taken and then compare that to what drugs they claim they have recently taken. Such studies consistently find that Blacks are more likely than whites to lie and claim that they have not used a drug that they really have used (Page et al. 2009Falk et al. 1992Feucht, Stephens, and Walker, 1994, and Fedrich and Johnson 2005). In fact, as reviewed in Ellis, Beaver, and Wrights Handbook of Crime Correlates, most studies based on self reported criminal history suggest that Blacks are not more likely than Whites to commit crime in general. And we’ve already seen that’s false.


Source: Wright, Beaver, and Ellis, 2009, page 25

As was noted previously, these arguments also ignore important differences between Black and White drug users. What kind of differences? Well, to quote one study comparing Black and White drug users: “African Americans are nearly twice as likely to buy outdoors (0.31 versus 0.14), three times more likely to buy from a stranger (0.30 versus 0.09), and significantly more likely to buy away from their homes (0.61 versus 0.48).” Ramchand , Pacula, and Iguchi MY 2006. Similarly, a  report issued by the Justice Department found that Black drug users use drugs more often than White drug users, use more dangerous drugs than White drug users, and are more likely to use drugs in areas with high crime rates (Lagan 1995). All 6 of these differences will make Black drug users more likely get arrested than White drug users. Given all this, there is no good reason to suppose that Black drug arrest rates reflect racism. It is far more likely that they reflect Black drug user’s dishonestly as well as the reckless manner in which they use drugs.

Police Brutality

Recently, it has become particularly fashionable to claim that Blacks are often un-fairly killed by police. The media has dedicated tremendous amounts of attention to specific supposed instances of this phenomena such as Treyvon Martin, Micheal Brown, and Freddie Gray.

This obsession with colorful (and often dishonest) anecdotes is not coincidental. A quick look at the relevant statistics quickly dispels the police war on Blacks myth. As we saw above, both NVCS and UCR data show that Blacks account for about 1/3 of rape and assault offenders and over half of robbery offenders. UCR data also suggests that Blacks account for roughly half of murder offenders, 38% of all violent criminals, and 29% of all persons arrested (2014, Crime in the United States, Table 43). Given this, if we had a just police force which only killed criminals who posed a serious danger to society, and if such criminals of both races were equally likely to be killed by police, then we would expect somewhere between 29% and 38% of those killed by police to be Black.

There are various sources on police killings. Which one is best is a matter of controversy. But it doesn’t really matter, because they all show basically the same thing. To begin with, the Harvard economist Sendhil Mullainathan analyzed UCR data and found that 32% of those killed by police were black (Mullainathan 2015). Similarly, the sociologist Peter Moskos analyzed data from the website Killedbypolice.net, which claims to be “The most accurate, most comprehensive and always up-to-date list of people killed by U.S. law enforcement officers.” The site basically aggregates all news stories in the country about someone being killed by a cop. The site is supposed to offer a rigorous alternative to biased and lazy statistics released by the government. Using this data set, Moskos found that 30% of those killed by police in 2013-2015 were Black (Moskos 2015). A third source we can use is the CDC’s Compressed Mortality Database. This database’s focus isn’t on crime but, rather, the causes of the causes of American deaths. However, one such cause is being killed by law enforcement (excluding legal execution). Using this data, we can see that the CDC estimates Blacks to have been 27% of those killed by police between 1999 and 2014 (Compressed Mortality Database). Thus, across multiple data sources, we see that, if anything, Blacks make up a lower proportion of those killed by police than what we would expect given their crime rates. In light of this, there is no justifiable reason to supposing that the police are running around unjustly killing Blacks.


Unfair Sentencing

Another common argument purporting to show that the justice system is racist states that Blacks get longer sentences than Whites even when they commit the same crimes.  This is true, but is fully explained by factors other than racism.

When a criminal is sentenced for a crime there are more relevant factors than the crime he just committed. Other variables, such as how he presents himself in the courtroom, and the likelihood that he will commit another crime in the future, also play a role. If we hold these things constant, we see that Blacks and Whites gen the same sentences for the same crimes.

This was the finding of Beaver et al. 2013. In this study, researchers compared criminal’s sentencing time after controlling for their verbal IQ and their self reported history of violence. They found that holding these variables constant completely eliminated the racial cap in sentencing.


This is a long post, so a summary may be helpful. First, we saw that the Blacks being arrested more often than Whites is probably not explained by racism. This was evidenced by the correspondence between arrest records, victimization surveys, and rates of misbehavior in school. After that, we saw that Black drug arrest rates can be accounted for both by the greater dishonestly of Black drug users on surveys that ask about drug use, and by the differences in behavior between Black and White drug users which make Black drug users more likely to get arrested. Then we looked at police killings and showed that Blacks are killed by police at roughly the rate you would expect given their representation among criminals in general. Finally, we saw that Blacks get longer sentences than Whites for reasons completely unrelated to racism. Thus, in conclusion, Blacks really do commit crime at the rate that arrest statistics suggest and, more generally, the American criminal justice system is not biased against blacks.


Facebook Comments
  • Emil Kirkegaard

    “This was the finding of Beaver et al. 2013.
    In this study, researchers compared criminal’s sentencing time after
    controlling for their verbal IQ and their self reported history of
    violence. They found that holding these variables constant completely
    eliminated the racial cap in sentencing.”

    That’s not correct. The table is 5. The effect of race was .664 when only adjusting for age, and .374 after adjusting for the others (betas). Neither has a p value below .05, but the adjusted value is certainly not 0. I.e. the difference is perhaps not real to begin with and was reduced by about 44%.

    • You are correct, I should not have described it that way. I will change it within the week, as I was planning on significantly re-writing this article (not for this reason) anyhow.