The Color of Crime, 2016 Revised Edition

By Edwin S. Rubenstein, M.A., New Century Foundation

Major Findings

  • The evidence suggests that if there is police racial bias in arrests it is negligible. Victim and witness surveys show that police arrest violent criminals in close proportion to the rates at which criminals of different races commit violent crimes.
  • There are dramatic race differences in crime rates. Asians have the lowest rates, followed by whites, and then Hispanics. Blacks have notably high crime rates. This pattern holds true for virtually all crime categories and for virtually all age groups.
  • In 2013, a black was six times more likely than a non-black to commit murder, and 12 times more likely to murder someone of another race than to be murdered by someone of another race.
  • In 2013, of the approximately 660,000 crimes of interracial violence that involved blacks and whites, blacks were the perpetrators 85 percent of the time. This meant a black person was 27 times more likely to attack a white person than vice versa. A Hispanic was eight times more likely to attack a white person than vice versa.
  • In 2014 in New York City, a black was 31 times more likely than a white to be arrested for murder, and a Hispanic was 12.4 times more likely. For the crime of “shooting” — defined as firing a bullet that hits someone — a black was 98.4 times more likely than a white to be arrested, and a Hispanic was 23.6 times more likely.
  • If New York City were all white, the murder rate would drop by 91 percent, the robbery rate by 81 percent, and the shootings rate by 97 percent.
  • In an all-white Chicago, murder would decline 90 percent, rape by 81 percent, and robbery by 90 percent.
  • In 2015, a black person was 2.45 times more likely than a white person to be shot and killed by the police. A Hispanic person was 1.21 times more likely. These figures are well within what would be expected given race differences in crime rates and likelihood to resist arrest.
  • In 2015, police killings of blacks accounted for approximately 4 percent of homicides of blacks. Police killings of unarmed blacks accounted for approximately 0.6 percent of homicides of blacks. The overwhelming majority of black homicide victims (93 percent from 1980 to 2008) were killed by blacks.
  • Both violent and non-violent crime has been declining in the United States since a high in 1993. 2015 saw a disturbing rise in murder in major American cities that some observers associated with “depolicing” in response to intense media and public scrutiny of police activity.

The past two years have seen unprecedented concern about racial bias in law enforcement. Deaths of young black men at the hands of the police led to serious rioting in Ferguson, Missouri, and in Baltimore. These and other deaths gave rise to the Black Lives Matter movement, which has carried out hundreds of demonstrations across the country and even in Canada. It is widely assumed that the police and the courts are strongly biased — certainly against blacks, and probably against Hispanics.

This problem cannot be fully understood by concentrating on a few cases, no matter how disturbing they may first appear. There were an estimated 11,300,000 arrests in the United States in 2013, the overwhelming majority of which were carried out properly. It is only in a larger context that we can draw conclusions about systemic police bias or misbehavior. This larger context is characterized by two fundamental factors. The first is that different racial groups commit crime at strikingly different rates, and have done so for many years. The second is that crime, overall, has declined dramatically over the last 20 years. Only after considering these points is it possible to draw well-founded conclusions about police bias.

In 2005, the New Century Foundation published “The Color of Crime,” a study of the relationship between crime, race, and ethnicity in the United States. The study was based on published government statistics and found that blacks were seven times more likely to commit murder and eight times more likely to commit robbery than people of other races, while Asians had consistently low crime rates. Hispanics appeared to be committing violent crime at roughly three times the white rate, but this conclusion was tentative because official statistics often failed to distinguish between whites and Hispanics.

The 2005 study also found that blacks were seven times more likely than whites to be in prison and Hispanics were three times more likely. It also concluded that high black arrest and imprisonment rates — often cited as evidence of a racist criminal justice system — were explained by the black share of offenders.

There has been a very important change since 2005: Crime is down. This is clearly indicated by the broadest measure of criminality in the United States, which is the annual National Crime Victimization Survey (NCVS). In 2013, 90,630 households and 160,040 people were interviewed for the NCVS about their experiences as crime victims — whether reported to the police or not. A 20-year compilation of the survey’s findings indicates that both the number and rate of violent victimizations have declined steadily, albeit unevenly, for at least two decades (see Figure 1).


Violent crime includes rape or sexual assault, robbery, simple or aggravated assault, and domestic violence — but not murder. Total violent victimizations in 2013 (the most recent year for NCVS data) were about one-third their 1994 level, which was a record high; the total number declined from 17.1 million in 1994 to 6.1 million in 2013.

This drop reflects an even steeper decline in the rate of violent crime (violent crimes per 1,000 people 12 years of age or older) — from 79.8 in 1994 to 23.2 in 2013.

A second widely cited measure of crime, the FBI’s Uniform Crime Reports (UCR), confirms that violent crime is in a decades-long decline (see Figure 2). The FBI’s statistics are based on crimes reported to the police, and therefore do not include all crimes, many of which are not reported. Also, some local law enforcement agencies do not submit data for the UCR. For both these reasons, the number of violent victimizations recorded by the UCR — 1.16 million in 2013 — is a fraction of that year’s NCVS figure of 6.1 million. The decline in violent crimes as reported by the UCR — a 37 percent reduction from 1994 to 2013 — is significantly less than the 64 percent drop found by the NCVS over the same period.


The reason for this difference may be that as the actual number of crimes drops, victims are more likely to report violence to the police because it is less routine. In 1994, 40.9 percent of victims told the NCVS that they had filed a police report. In 2013, the figure was 44.3 percent.

While violent crime is unquestionably down since the last “Color of Crime” report, the share of non-white victims is up (see Table 1).

Table 1. Violent victimizations by race of victims, 2002 and 2013
2002 2013 % change
White 5,432,632 3,832,527 -29.5%
Black 1,023,828 815,061 -20.4%
Hispanic 808,355 1,015,672 25.6%
Other (a) 159,736 174,309 9.1%
Two or more races (b) 288,854
Total 7,424,551 6,126,423 -17.5%
Percent of total
White 73.2% 62.6% -14.5%
Black 13.8% 13.3% -3.5%
Hispanic 10.9% 16.6% 52.3%
Other (a) 2.2% 2.8% 32.2%
Two or more races (b) 4.7%
Total 100.0% 100.0% 0.0%
Rate per 1,000 people 12 and older
White 32.6 22.2 -31.9%
Black 36.1 25.1 -30.5%
Hispanic 29.9 24.8 -17.1%
Other (a) (c) 31.0 (c) 26.8 -13.5%
Two or more races (b) 90.3
Total 32.1 23.2 -27.7%
(a) Includes Asians, Native Hawaiians, other Pacific Islanders, American Indians, and Alaska Natives. (b) Category did not exist in 2002. (c) Weighted average of victimization rates for Asian/Native Hawaiian/other Pacific Islander and American Indian/Alaska Native. Data Sources: NCVS (victimizations); Bureau of Justice Statistics, “Criminal Victimization, 2013,” Table 9 (2013 victimization rate); “Criminal Victimization, 2011,” Table 5 (2002 victimization rate); Bureau of Justice Statistics, “Criminal Victimization, 2013,” Table 5 (victims, 2013 and 2004).

From 2002 to 2013, the number of violent victimizations suffered by whites and blacks fell by 29.5 percent and 20.4 percent, respectively, and the white share of total violent victimizations declined from 73.2 percent to 62.6 percent. (In this report, “white,” “black,” and “Asian” always mean “non-Hispanic.”)

Over the same period, Hispanic victimizations rose by 25.6 percent, while the “Other” category (mainly Asians) saw a 9.1 percent rise. Victimization rates for both groups declined — thought not as rapidly as for whites or blacks. The rise in victimizations was the result of a rapid increase in the numbers of these groups. From 2002 to 2013, the Hispanic population age 12 and over, for example, grew 48.6 percent while the corresponding white population grew by only 1.6 percent.

The “Two or more races” category did not exist in 2002 for the NCVS,1 and the high victimization rates for this group probably reflect its small sample size: The rate more than doubled from 2012 to 2013.

While the black victimization rate exceeded that of whites and Hispanics in both 2002 and 2013, the gap between the black and Hispanic rates narrowed dramatically — from 6.2 victims per 1,000 people (36.1 – 29.9) in 2002 to just 0.3 (25.1 – 24.8) in 2013. If this trend continues, the Hispanic victimization rate will soon exceed the black rate.

In 2002, whites were 9 percent more likely to be victims of a crime than Hispanics. By 2013, these groups had changed places, with the white victimization rate 10.5 percent lower than the Hispanic rate. This may be due to illegal immigration of Hispanics. A disproportionate number of such immigrants are young men, who are the group most likely to commit crimes, and they may also be vulnerable as victims.

Property crimes such as burglary and motor vehicle theft also appear to be in long-term decline, falling from an estimated 35.1 million cases in 1993 to 16.8 million in 2013 (see Figure 3).


As with violent crimes, property crime victims are increasingly non-white (see Table 2).

Table 2. Property crime victimizations by race of head of household, 2002 and 2013
2002 2013 % change
White 13,108,165 10,491,279 -20.0%
Black 2,534,714 2,447,316 -3.4%
Hispanic 2,344,423 2,657,590 13.4%
Other (a) 567,016 1,177,902 107.7%
Total 18,554,318 16,774,087 -9.6%
Percent of total
White 70.6% 62.5% -11.5%
Black 13.7% 14.6% 6.8%
Hispanic 12.6% 15.8% 25.4%
Other (a) 3.1% 7.0% 129.8%
Total 100.0% 100.0% 0.0%
Rate per 1,000 households (b)
White 168.1 130.6 -22.3%
Black 191.0 161.9 -15.2%
Hispanic 185.3 139.4 -24.7%
Other (a) 88.7 89.6 1.0%
Total 168.2 131.4 -21.9%
Note: Property crimes include household burglary, theft, and motor vehicle theft.

  1. Includes households headed by Asians, Native Hawaiians, Pacific Islanders and, in 2013, persons of two or more races.
  2. Assumes the average household of each group contains the same number of persons age 12 and older.

Data sources: National Crime Victimization Survey, (number); Bureau of Justice Statistics, “Criminal Victimization, 2013,” Appendix Table 1.

The overall property crime rate has declined since the last “Color of Crime” — from 168.2 victimizations per 1,000 households in 2002 to 131.4 per 1,000 in 2013. The number of property crimes suffered by white households fell 20 percent from 2002 to 2013, and their share of such crimes dropped from 70.6 percent to 62.5 percent. Both whites and blacks suffered fewer property crimes in 2013 than in 2002.

Race of Offender

It is surprisingly difficult to arrive at a definitive picture of the races of offenders. The National Crime Victimization Survey categorizes crime victims by race and Hispanic ethnicity, but until recently, it did not consider Hispanics a separate offender category; it usually called them “white” or “other race.” Furthermore, beginning in 2009, the year the Obama administration took office, the NCVS stopped publishing information on race of offender, even though it continued to gather the data. In 2015, the Department of Justice finally released a partial set of offender-race information (see page 13 below).

The Uniform Crime Reports program, which is the basis of the FBI’s national tabulation of arrests, includes Hispanics in the “white” category. Arrest and incarceration rates by race — to the extent they are even available — must often serve as imperfect indicators of actual offense rates by race.

As we will see in greater detail below, blacks are arrested at much higher rates than any other racial group.2 It is common to argue that these high rates are the result of racial bias, and that bias continues through every stage of criminal processing: indictment, plea bargain, trial, sentencing, parole, etc. In 2008, then-senator Barack Obama asserted that blacks and whites “are arrested at very different rates, are convicted at very different rates, [and] receive very different sentences … for the same crime.” This view is echoed by the media but is not supported by either the scholarly literature or by government statistics.

Police, in particular, are often accused of racial bias, but is it really plausible that they arrest blacks they know are innocent but ignore white criminals? A 2008 summary of earlier research compared the races of offenders as identified by victims to the races of perpetrators arrested by the police and found that “the odds of arrest for whites were 22 percent higher for robbery, 13 percent higher for aggravated assault, and 9 percent higher for simple assault than they were for blacks, whereas there were no differences for forcible rape.”

A 2015 study of American men based on the National Longitudinal Study of Adolescent Health found that controlling for IQ and lifetime records of violence completely accounted for racial differences in arrest rates.

Fortunately, there is an excellent database that throws light directly on the question of racial bias in arrests: the National Incident-Based Reporting System (NIBRS). In 2013, 6,328 law enforcement agencies covering approximately 29 percent of the US population reported crime to the FBI using NIBRS categories, which include races of reported offenders as well as races of persons arrested. It is reasonable to assume that both the racial mix of this massive sample and the behavior of police officers are representative of the entire United States.

Unfortunately, NIBRS does not distinguish between whites and Hispanics, which means blacks are the only racial group for which we have consistent information. However, blacks are the group most frequently said to be victims of police bias, so if the police treat them fairly it is probably safe to conclude they treat other groups fairly.

Figure 4 compares the percentages of criminals that victims say were black to the percentages of arrested suspects who were black. If police are arresting a larger proportion of blacks than the proportion of criminals victims say were black, it would be evidence of bias.


For most crimes, blacks make up a larger percentage of reported offenders than they do of those arrested. In only seven of the 22 NIBRS crime categories did blacks account for a larger share of arrests: homicides, counterfeiting/forgery, embezzlement, fraud, stolen property offenses, drug offenses, and gambling. Interestingly, these are crimes for which there may be no witnesses — such as embezzlement or stolen property — or are “victimless” crimes, such as drug offenses and gambling. The racial identification of suspects in these cases may not be reliable.

In crimes that involve direct contact with victims and in which race of offender can therefore be clearly identified, black arrest rates are below reported offender rates. For example, blacks were identified as 73 percent of robbery offenders but accounted for only 59 percent of robbery arrests.

When crimes from all categories are aggregated, black offenders were 14 percent less likely than non-blacks offenders to be arrested. This suggests that police do not show anti-black bias, but make arrests that closely match the proportions at which people of different races commit crime.

NIBRS data come disproportionately from smaller police departments. In 2013, only 10 percent of the population covered by the system lived in cities of 250,000 and greater. What do arrest statistics show for large metropolitan areas?

New York City, for example, does not participate in NIBRS but it records the races of arrested offenders, and consistently distinguishes between whites and Hispanics. In 2014, 374 people were arrested for murder. Their races were as follows:

White: 2.9 percent

Black: 61.8 percent

Hispanic: 31.8 percent

Asian: 2.7 percent

Other: 0.8 percent

Police take murder very seriously and investigate all cases carefully. Press and judicial system scrutiny are high. Arrest rates for murder therefore track actual crime rates more closely than for any other crime. Murder is probably the crime for which it would be most difficult for police to make “biased” arrests even if they wanted.

Given a population (page B1 of report) that was 32.8 percent white, 22.6 percent black, 28.9 percent Hispanic, and 13.0 percent Asian, a black was 31 times more likely than a white to be arrested for murder, a Hispanic was 12.4 times more likely than a white, and an Asian was twice as likely. These multiples and those for other crimes appear as graphs on the next page. A “shooting” is discharge of a firearm in which a bullet strikes a person.



There is another way to express these disparities. If New York City had been all white in 2014 — and the additional whites committed crimes at the same rates as the city’s actual white residents — there would have been 32 murder arrests instead of 374, 1,844 robbery arrests instead of 10,163, and 16 arrests for shootings rather than 503. These figures would reflect reductions in these crimes of no less than 91, 81, and 97 percent, respectively.

There are race differences in crime rates throughout the United States, but the differences are particularly sharp in New York and other major cities. This is probably because whites who live in urban centers are often relatively wealthy whereas blacks and Hispanics who live in cities are relatively poor.

In the graphs on the following page, the most serious offenses are displayed above, with the less serious offenses below (except for firearms violations, which are serious crimes). Where possible, the graphs are arranged to depict the less serious version of the same crime directly below the more serious version. Misdemeanor sex crimes, for example, do not rise to the level of rape, and include forcible touching and sexual misconduct. Grand larceny is theft of anything with a value greater than $1,000 and includes auto theft, while petty larceny is theft of anything less valuable. Felonious assault includes attack with a deadly weapon, whereas misdemeanor assault includes pushing and spitting. Misdemeanor criminal mischief includes such crimes as cemetery desecration and calling in false fire alarms.

Almost without exception, the black/white and Hispanic/white arrest multiples are lower for the less serious crimes. Whatever else this difference may mean, it is strong evidence that the police are not making biased arrests. Police have broad discretion as to whether they will arrest someone for forcible touching, shoplifting, or setting off a false fire alarm.

If racist police wanted to vent prejudices on non-whites, these are the crimes for which they could most easily do so. They can walk away if someone complains he was spat on, and if they are racist they can walk away if the spitter is white but make an arrest if the spitter is black. Police cannot walk away if someone is lying on the sidewalk bleeding from a knife wound. They must try to make an arrest, whatever the race of the suspect. The graphs that show the lowest non-white/white arrest multiples are for crimes in which police have the greatest arrest discretion and are therefore strong evidence that New York City police are not biased in their arrest patterns.

Like New York, Chicago keeps detailed annual statistics on major crimes. Until 2010, it published the race of offenders, but after the election of Mayor Rahm Emanuel, it stopped releasing that information. The 2010 report (page 35) shows that, like New York City, there are stark racial differences. Arrests for murder were as follows: whites — 8, blacks — 190, Hispanics — 48, Asians — 1.

The racial mix of Chicago’s population in 2010, as reported on page 25 of the police department report (whites — 31.7 percent, blacks — 32.4 percent, Hispanics — 28.9 percent, Asians — 5.4 percent) meant that a Chicago black was 24 times more likely than a Chicago white to be arrested for murder, and a Hispanic was 6.7 times more likely.

Table 3 shows the multiples for the white arrest rates for a variety of crimes. Sharp racial differences appear not only for crimes of violence but also for property crimes, such as burglary and auto theft. The Chicago data fit the national pattern: Blacks have, by far, the highest arrest rates, followed by Hispanics. Asians have the lowest arrest rates.

Table 3. Multiples of the White Arrest Rate Chicago, 2010
Black Hispanic Asian
Murder 23.8 6.7 0.75
Robbery 27.3 3.9 0.69
Sexual Assault 10.4 4.9 0.23
Aggravated Assault 10.3 3.5 0.72
Burglary 10.8 3.1 0.24
Larceny 4.9 1.2 0.50
Auto Theft 19.9 4.5 0.77
Narcotics 11.5 2.5 0.38

If the same calculation is done as with New York City to arrive at crime rates in a theoretical all-white Chicago, murder would decline by 90.2 percent, rape by 80.8 percent, and robbery by 90.2 percent.

Chicago police also collected information on the sex of arrested criminals. Most people understand that men are more violent and dangerous than women, and this is reflected in the city’s arrest statistics. In 2010, men were 12.8 times more likely than women to be arrested for murder and 19.4 times more likely to be arrested for robbery, compared to the black/white multiples for these crimes of 23.8 and 27.3, respectively. This means that although men in Chicago are more dangerous than women, by comparison, blacks are even more dangerous when compared to whites. Similar calculations for New York City are not possible because the NYPD does not release arrests by sex.

Other American cities release crime statistics but not always for the same categories. Milwaukee records races of suspects in both homicides and “non-fatal shootings.” In 2014 (the most recent year available), blacks were 12 times more likely to be murder suspects than whites, and Hispanics were four times more likely. For non-fatal shootings, blacks were 25 times more likely than whites to be suspects, and Hispanics were 7.6 times more likely.

Pittsburgh releases arrest statistics, which follow the same pattern. In 2012 (the most recent year available), blacks were 26.6 times more likely than whites to be arrested for murder. The multiples for robbery, rape, and aggravated assault were 9.8, 7.5, and 5.6, respectively. The Hispanic population was so small (2.7 percent) that comparisons were not meaningful.

St. Louis, Missouri, keeps track of homicide suspects by race (page 41 of report). In 2013, 96 were black, one was white, one was Hispanic, and one was Asian. Murder victims are not a cross section of the population. In 2013, 82.5 percent had a criminal record (page 40 of report).

A larger geographical territory, such as California, gives a broader picture of racial differences in arrest rates, and one in which racial disparities are not nearly so stark. California is one of the few states that treat whites, blacks, and Hispanics separately, so there can be no confusion about how many Hispanics are being counted as whites. The NIBRS data, together with academic studies of policing that find little racial bias in arrests, suggest that arrest figures are probably realistic indices of the different rates at which people of different races commit crime (see Table 4).

Table 4. Felony arrests by offense and race in California, 2002 and 2013
All Violent Crimes All Property Crimes
2002 2013 % chg. 2002 2013 % chg.
Number Number
White 40,309 30,415 -24.5% 47,728 36,672 -23.2%
Black 29,230 23,683 -19.0% 29,693 22,660 -23.7%
Hispanic 54,016 42,635 -21.1% 50,935 41,304 -18.9%
Other 8,504 6,380 -25.0% 9,786 6,359 -35.0%
Total 132,059 103,113 -21.9% 138,142 106,995 -22.5%
Percent of total Percent of total
White 30.5% 29.5% -3.4% 34.5% 34.3% -0.8%
Black 22.1% 23.0% 3.8% 21.5% 21.2% -1.5%
Hispanic 40.9% 41.3% 1.1% 36.9% 38.6% 4.7%
Other 6.4% 6.2% -3.9% 7.1% 5.9% -16.1%
Total 100.0% 100.0% 0.0% 100.0% 100.0% 0.0%
Arrest rate (per 100,000 persons) Arrest rate (per 100,000 persons)
White 256 204 -20.1% 303 246 -18.6%
Black 1,392 1,093 -21.5% 1,414 1,046 -26.1%
Hispanic 453 290 -36.1% 428 281 -34.4%
Other 180 97 -46.1% 208 97 -53.3%
Total 383 269 -29.8% 401 279 -30.3%
Multiple of white rate Multiple of white rate
White 1.00 1.00 0.0% 1.00 1.00 0.0%
Black 5.44 5.35 -1.7% 4.67 4.24 -10.0%
Hispanic 1.77 1.42 -20.0% 1.41 1.14 -24.0%
Other 0.71 0.48 -32.6% 0.69 0.39 -74.3%
Total 1.50 1.32 -12.1% 1.32 1.13 -16.8%
Data sources: CA Dept. of Justice, “Crime in California,” 2002 and 2013, Table 31 (arrests); American Community Survey 2013 (2013 population); CA Dept. of Finance, CA Current Pop. Survey Report: March 2002 Data.

Consistent with national data, arrest rates of blacks, Hispanics, and “others” have all declined relative to that of whites over the past decade (see the last figures in the table: “Multiple of white rate”). Still, in 2013, California blacks were 5.35 times more likely than whites to be arrested for violent crimes, and 4.24 times more likely to be arrested for property crimes. The corresponding figures for Hispanics were 1.42 and 1.14.

“Others,” who are mostly Asians, appear to be a model group. Violent-crime arrest rates are less than half those for whites, and property crime arrest rates are 60 percent lower.3

The “Total” figure indicates the multiple of the white arrest rate for the total population. Black and Hispanic arrest rates raise the multiple while arrest rates for “Others” lower it.

While black and Hispanic arrest rates have declined relative to white rates, very high black arrest rates are still the rule for most crimes (see Table 5). In 2013, the black arrest rate multiple (compared to whites) ranged from a low of 1.56 for “dangerous drugs” offenses to a high of 13.39 for robbery. Black arrest rate multiples rose for burglary, forgery, kidnaping, arson, and, especially, dangerous drugs, for which the black multiple more than doubled since 2002. Dangerous drugs are methamphetamine, phencyclidine, and barbiturates. The sharp increase, matched with a very sharp decline in arrests for “narcotics,” probably reflects a shift in “War on Drugs” enforcement policy towards these drugs rather than marijuana.

Table 5. Multiples of white felony arrest rate in California, 2002 and 2013
Homicide 2002 2013 % change Car theft 2002 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 9.78 8.58 -12.3% Black 5.90 3.82 -35.2%
Hispanic 3.40 2.50 -26.5% Hispanic 1.98 1.50 -24.4%
Other 1.56 0.65 -58.1% Other 0.75 0.39 -47.9%
Total 2.44 1.95 -20.3% Total 1.60 1.25 -22.2%
Forcible rape 2002 2013 % change Forgery 2002 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 7.58 6.36 -16.1% Black 5.03 5.40 7.5%
Hispanic 2.68 2.25 -16.1% Hispanic 1.32 1.18 -10.3%
Other 0.90 0.56 -37.6% Other 0.70 0.57 -18.7%
Total 1.97 1.71 -13.2% Total 1.31 1.25 -5.2%
Robbery 2002 2013 % change Arson 2002 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 15.88 13.39 -15.7% Black 2.05 2.24 9.2%
Hispanic 2.71 1.96 -27.8% Hispanic 0.80 0.74 -6.4%
Other 0.86 0.46 -46.8% Other 0.53 0.27 -48.5%
Total 2.48 1.98 -20.3% Total 0.93 0.85 -8.7%
Assault 2002 2013 % change Drug offenses 2002 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 4.44 4.32 -2.7% Black 4.47 2.32 -48.1%
Hispanic 1.65 1.32 -19.7% Hispanic 1.23 0.97 -21.2%
Other 0.68 0.47 -30.0% Other 0.37 0.29 -21.3%
Total 1.39 1.22 -12.1% Total 1.20 0.94 -21.8%
Kidnapping 2002 2013 % change Narcotics 2002 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 6.73 6.95 3.2% Black 17.11 3.56 -79.2%
Hispanic 2.84 1.94 -31.6% Hispanic 2.11 0.69 -67.4%
Other 0.97 0.60 -38.3% Other 0.49 0.22 -54.5%
Total 1.98 1.63 -17.7% Total 2.29 0.89 -61.1%
Burglary 2002 2013 % change Dangerous drugs 2002 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 4.40 4.66 5.9% Black 0.75 1.56 107.5%
Hispanic 1.40 1.11 -20.8% Hispanic 1.02 1.07 4.4%
Other 0.68 0.38 -44.1% Other 0.33 0.28 -17.1%
Total 1.30 1.14 -12.2% Total 0.90 0.93 3.5%
Theft 2002 2013 % change Driving offenses 2002 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 4.40 3.74 -15.0% Black 1.75 1.73 -1.0%
Hispanic 1.22 1.07 -12.8% Hispanic 1.63 1.29 -20.8%
Other 0.66 0.39 -41.6% Other 0.56 0.47 -15.7%
Total 1.24 1.08 -13.1% Total 1.20 1.06 -11.6%
Data sources: CA Dept. of Justice, “Crime in California,” 2002 and 2013, Table 31 (arrests); American Community Survey 2013 (2013 population); CA Dept. of Finance, CA Current Pop. Survey Report: March 2002 Data (population).

Community Survey 2013 (2013 population); CA Dept. of Finance, CA Current Population Survey Report: March 2002 Data (2002 population).

California Hispanics, on the other hand, were less likely than whites to be arrested for drug offenses, narcotics, and arson. At the other extreme, Hispanics were 2.50 times more likely to be arrested for homicide; for forcible rape, the figure is 2.25. Hispanics are now the single largest group in California. In 2014, there were 14.92 million whites and 14.99 million Hispanics in the state.

On average, Hispanics are younger than whites and blacks. This means there are relatively more Hispanics in the peak -crime ages of 18 to 29. Some analysts have argued that when age distribution is taken into account, Hispanics are no more likely than whites to commit violent crimes.4

It is not possible to test this theory with national arrest data because not all jurisdictions distinguish between Hispanics and whites. However, California distinguishes consistently between whites and Hispanics, and also includes arrest rates by age group. Table 6, which is probably representative of the entire country, deserves careful study.

Table 6. Multiples of white felony arrest rate by race, ethnicity, and age in California, 2013
All violent offenses Kidnapping
White Hispanic Black Other White Hispanic Black Other
All ages 1.000 1.431 5.270 0.500 All ages 1.000 1.962 6.846 0.627
10 to 17 1.000 1.612 9.430 0.510 10 to 17 1.000 1.421 4.144 0.773
18 to 29 1.000 1.417 5.066 0.460 18 to 29 1.000 1.977 6.604 0.648
30 to 39 1.000 1.159 3.996 0.454 30 to 39 1.000 1.479 6.151 0.525
40 to 69 1.000 1.039 4.168 0.470 40 to 69 1.000 1.454 5.598 0.486
70+ 1.000 1.145 3.112 1.093 70+ 1.000 3.404 12.061 0.000
Homicide Weapons offenses
White Hispanic Black Other White Hispanic Black Other
All ages 1.000 2.523 8.460 0.687 All ages 1.000 1.878 5.516 0.417
10 to 17 1.000 8.178 18.374 2.524* 10 to 17 1.000 2.242 4.747 0.567
18 to 29 1.000 2.685 10.089 0.700 18-29 1.000 1.796 5.698 0.432
30 to 39 1.000 1.871 5.525 0.523 30-39 1.000 1.157 4.149 0.310
40 to 69 1.000 1.207 5.171 0.557 40-69 1.000 0.986 3.783 0.270
70+ 1.000 0.425 1.508 0.486 70+ 1.000 0.400 4.966 0.458
Forcible rape Driving under the influence
White Hispanic Black Other White Hispanic Black Other
All ages 1.000 2.271 6.266 0.588 All ages 1.000 1.244 1.659 0.467
10 to 17 1.000 1.105 4.168 0.311 10 to 17 1.000 0.680 0.372 0.000
18 to 29 1.000 2.032 5.653 0.605 18-29 1.000 1.156 1.072 0.528
30 to 39 1.000 2.187 5.167 0.504 30-39 1.000 1.054 1.588 0.405
40 to 69 1.000 2.244 6.567 0.554 40-69 1.000 1.212 2.148 0.344
70+ 1.000 0.000 0.000 1.945 70+ 1.000 1.418 4.020 1.621
Robbery Property offenses
White Hispanic Black Other White Hispanic Black Other
All ages 1.000 1.978 13.197 0.483 All ages 1.000 1.149 4.182 0.413
10 to 17 1.000 2.980 29.807 0.981 10 to 17 1.000 1.416 6.118 0.462
18 to 29 1.000 1.579 10.356 0.421 18 to 29 1.000 0.957 3.431 0.357
30 to 39 1.000 1.065 6.953 0.308 30 to 39 1.000 0.827 2.982 0.353
40 to 69 1.000 1.181 9.034 0.334 40 to 69 1.000 0.904 4.022 0.398
70+ 1.000 2.431 1.723 0.000 70+ 1.000 1.243 5.169 0.803
Assault Burglary
White Hispanic Black Other White Hispanic Black Other
All ages 1.000 1.335 4.257 0.497 All ages 1.000 1.120 4.596 0.401
10 to 17 1.000 1.293 5.049 0.398 10 to 17 1.000 1.440 6.820 0.496
18 to 29 1.000 1.354 4.026 0.459 18 to 29 1.000 0.869 3.745 0.335
30 to 39 1.000 1.145 3.621 0.467 30 to 39 1.000 0.751 3.096 0.330
40 to 69 1.000 1.008 3.783 0.477 40 to 69 1.000 0.862 3.973 0.384
70+ 1.000 1.140 3.183 1.145 70+ 1.000 1.702 3.837 1.414
Motor vehicle theft Theft
White Hispanic Black Other White Hispanic Black Other
All ages 1.000 1.486 3.762 0.409 All ages 1.000 1.077 3.683 0.405
10 to 17 1.000 1.923 4.987 0.447 10 to 17 1.000 1.359 6.196 0.451
18-29 1.000 1.347 2.949 0.355 18 to 29 1.000 0.937 2.884 0.371
30-39 1.000 0.930 2.836 0.355 30 to 39 1.000 0.850 2.684 0.347
40 to 60 1.000 0.940 4.027 0.363 40 to 69 1.000 0.912 4.069 0.386
70+ 1.000 0.851 3.015 1.945 70+ 1.000 0.972 7.322 0.417
Data sources: California Dept. of Justice, “Crime in California,” 2013, Table 33 (arrests by offense, age, race, and ethnicity); California Dept. of Finance, Demographic Research Unit, Report P-3, December 14, 2014 (population by age).

*Very low numbers, which may not be representative: 5 whites, 7 others.

Hispanics of every age group are more likely than whites to be arrested for every type of violent crime. For the highest-crime age group of 18 to 29, this is particularly noticeable for homicide (2.7 times the white rate), rape (2 times) and kidnapping (2 times).

Though the sample size is small, Hispanic teenagers (ages 10 to 17) were 8.2 times more likely than whites of the same age group to be arrested for homicide: In 2013, 74 Hispanic and five white teenagers were arrested for this crime. (There are more than twice as many Hispanics as whites in this age range, which brings down the multiple.) Gang violence is probably a cause.

Assaults are the most common violent crime, and accounted for 80 percent of all arrests for violent crimes in California in 2013. Hispanics of every age group are arrested for assault at higher rates than whites, for an overall multiple of 1.34.

Blacks, however, are arrested at sharply higher multiples than both whites and Hispanics. Robbery is a perennial standout in this respect, with multiples as high as 29.8 in the 10 to 17 age group. There was only one age group, for only one crime, for which blacks were arrested at a lower rate than whites: drunk driving in the 10 to 17 age range. Again, this is a crime for which police have almost complete discretion when deciding to make an arrest, so there would be a high black/white multiple if officers were intent on targeting blacks unfairly.

Blacks are the main cause of urban crime. As Hispanics displace blacks in America’s largest cities and in its largest state, crime rates should remain stable or decline because Hispanics are displacing the group with the highest crime rates.

To return to Table 5, it is notable that the “Other” category of offenders — which is composed largely of Asians — has the lowest arrest rates across the board. This is consistent not only with earlier versions of “The Color of Crime” but with virtually every study conducted on American crime rates. As the number of Asians in a jurisdiction rises, crime rates can be expected to fall.


The criminal justice system begins with arrest, and continues through incarceration and release. Federal and state incarceration statistics are consistent, reliable, and distinguish between whites, blacks, Hispanics, and people of other races. They are therefore the best single nationwide indicator of the racial distribution of criminal activity. Of course, for incarceration rates to be a valid measure of crime rates, the criminal justice system must be free of systematic bias at every stage — from indictment to the decision to grant parole.

There have been many investigations of bias in the justice system, and a summary of the literature is beyond the scope of this report. However, a typical 2009 academic study of the process up through sentencing concludes that:

[U]sing arrests as a marker of violent offending, the overrepresentation of blacks among offenders admitted to state prisons occurs because they commit a disproportionate number of frequently imprisoned (i.e., violent) crimes.

One simple test for bias is to see whether incarceration rates by race generally correspond to arrest rates. Earlier, we noted that if the racial distribution of arrests matches the racial distribution of offenders identified by victims or witnesses, we can conclude that there is little or no systematic police bias. A comparison between arrest and incarceration rates should be a similar test for justice system bias. Incarceration rates consistently higher than arrest rates could suggest post-arrest bias, whereas incarceration consistently lower than arrest rates could suggest unjustified arrests that lead to release or acquittal.

Figure 7 compares the percentage of arrests of blacks (average percentages from 2009 to 2013) to incarcerations of blacks. Total incarcerations are slightly higher than total arrests for some crimes, but for others they are lower. We cannot rule out justice system bias. However, the general concordance between arrest and incarceration rates suggests an absence of serious or systematic racial bias. Statistics that fail to distinguish between whites and Hispanics at the time of arrest mean it is not possible to make a similar comparison for Hispanics.


The decline in arrests noted above is beginning to be reflected in incarceration figures. Department of Justice data show that the total incarcerated population peaked in 2008, and has since declined slightly (see Figure 8).


Between 1993 and 2013, the total incarcerated population rose by about 878,000, or 61 percent, while the US population rose by only 22 percent. Since the last “Color of Crime” (2002 data), the total inmate count is up by 201,000, or 10 percent. The incarceration rate peaked at 788 per 100,000 US residents in 2007 and declined to 730 by 2013. This is practically identical to the rate of 732 for 2002.

Federal prisoner data are no longer available by offense and race.5 However, our primary interest is in violent crime, which is almost always handled in state courts. As shown in Table 7, the state prison population sentenced for violent crimes increased by about 18 percent since the last “Color of Crime” report. (There is an increase because this is a comparison with 2001 figures; a comparison with 2009 figures would show a decrease.)

Table 7. Prisoners Under State Jurisdiction Sentenced for Violent Crimes, 2001 and 2013
2001 2013 % change
White 208,100 223,900 7.6%
Black 267,800 282,100 5.3%
Hispanic 102,600 162,300 58.2%
Other 17,600 36,500 107.4%
Total 596,100 704,800 18.2%
Percent of total
White 34.9% 31.8% -9.0%
Black 44.9% 40.0% -10.9%
Hispanic 17.2% 23.0% 33.9%
Other 3.0% 5.2% 72.6%
Total 100.0% 100.0% 0.0%
Incarceration rate (per 100,000 residents)
White 105 113 7.6%
Black 746 709 -5.0%
Hispanic 265 289 9.1%
Other 120 138 15.0%
Total 207 220 6.3%
  1. Includes Asians, American Indians, Pacific Islanders, and for 2012, persons of two or more races.

Data sources: Bureau of Statistics, “Prisoners in 2014,” Appendix Table 4 (2013); “Prisoners in 2002,” Table 15, page 10 (2001).

Hispanics and “others” were the fastest growing incarcerated categories: up 58 percent and 107 percent, respectively. Part of the rise in “others” reflects a statistical reclassification: Prisoners who identified themselves as being of “two or more” races were counted as “others” in 2013. The multi-race category did not exist in 2001.

Blacks were the only group to see a reduction in incarceration rates, but the reduction was small, and blacks remained far more likely to be incarcerated for violent crimes than any other group. The black violent-crime incarceration rate in 2013 was 6.3 times that of the white rate. This does not necessarily mean that blacks are 6.3 times more likely to commit violent crimes, because prison time depends on the severity of the crime as well as prior record.

Table 8 shows the change in incarceration rate multiples from 2001 (the benchmark year for the last “Color of Crime”) to 2013.

Table 8. Multiples of White Incarceration Rate in State Prisons, 2001 and 2013
All crimes 2001 2013 % change All property crimes 2001 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 7.14 5.27 -26.2% Black 5.00 3.45 -31.0%
Hispanic 2.48 2.06 -16.9% Hispanic 1.63 1.11 -31.9%
Other (a) 0.97 1.36 40.2% Other (a) 0.85 1.21 42.4%
All violent crimes 2001 2013 % change Burglary 2001 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 7.10 6.26 -11.8% Black 4.97 4.11 -17.3%
Hispanic 2.52 2.55 1.2% Hispanic 1.65 1.38 -16.4%
Other (a) 1.14 1.22 7.0% Other (a) 0.91 1.37 50.5%
Murder 2001 2013 % change Larceny 2001 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 8.26 7.53 -8.8% Black 6.44 1.26 -80.4%
Hispanic 2.76 2.55 -7.6% Hispanic 1.79 0.78 -56.4%
Other (a) 0.73 2.18 198.6% Other (a) 1.32 0.11 -91.7%
Manslaughter 2001 2013 % change Car theft 2001 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 5.52 2.78 -49.6% Black 5.36 2.54 -52.6%
Hispanic 2.84 1.61 -43.3% Hispanic 3.11 1.87 -39.9%
Other (a) 1.71 4.72 176.0% Other (a) 0.39 1.07 174.4%
Rape/Sexual Assault 2001 2013 % change Fraud 2001 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 2.77 2.51 -9.4% Black 4.20 2.62 -37.6%
Hispanic 1.19 1.67 40.3% Hispanic 0.93 0.61 -34.4%
Other (a) 0.90 1.02 13.3% Other (a) 0.39 1.45 271.8%
Robbery 2001 2013 % change Other Property Crimes 2001 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 14.74 13.10 -11.1% Black 4.17 2.27 -45.6%
Hispanic 3.93 3.49 -11.2% Hispanic 1.57 0.87 -44.6%
Other (a) 1.54 1.53 -0.6% Other (a) 0.73 2.43 232.9%
Assault 2001 2013 % change Drug Offenses 2001 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 7.17 6.57 -8.4% Black 13.45 5.86 -56.4%
Hispanic 3.35 3.06 -8.7% Hispanic 4.20 2.07 -50.7%
Other (a) 1.59 0.24 -84.9% Other (a) 0.49 2.24 357.1%
Other Violent Crimes 2001 2013 % chang. Public Order (b) 2001 2013 % change
White 1.00 1.00 0.0% White 1.00 1.00 0.0%
Black 4.72 5.78 22.2% Black 4.99 6.07 21.6%
Hispanic 2.06 2.38 15.5% Hispanic 2.35 1.86 -20.9%
Other (a) 1.15 NA NA Other (a) 0.52 0.77 48.1%
  1. Includes Asians, American Indians, Pacific Islanders, and, in 2013, persons of two or more races.
  2. Includes weapons, drunk driving, morals, and other public order offenses.

Data sources: Bureau of Justice Statistics, “Prisoners in 2002,” Table 15, page 10 (2001); “Prisoners in 2014,” Appendix Table 4 (2013).

While blacks were incarcerated at rates well above whites (and other groups) in 2013, their relative incarceration rates have come down for most offenses since 2001. In some cases the black/white multiple has narrowed significantly: a sharp 56.4 percent drop in the multiple for drug offenses, 49.6 percent for manslaughter, 80.4 percent for larceny, etc. Hispanic multiples are generally down as well, though the multiple for rape was up 40 percent.

Perhaps the biggest surprise is the surging incarceration multiples for “Other,” a group comprised of Asians, American Indians, Pacific Islanders, and persons of two or more races. This was probably caused by a change in prisoner categorization. American Indians, whose incarceration rates are second only to those of blacks, were treated as a separate category in 2001. Including them in the “Other” category in 2013 undoubtedly pushed up the apparent rates for Asians.

Table 9 shows that despite the historic decline in crime, the number of prisoners incarcerated for murder has increased 4 percent since 2001, and that Hispanics accounted for all of the increase. The number of Hispanics incarcerated for murder rose by about 40 percent, while the number of non-Hispanics fell by 3.2 percent. Although the Hispanic population grew significantly faster than the non-Hispanic population from 2001 to 2013, its incarceration rate for murder fell less than half as much as the non-Hispanic rate over that period.

Table 9. Prisoners sentenced for murder in state prisons, 2001 and 2013
2001 2013 % Change
Hispanic 27,800 39,000 40.3%
non-Hispanic 131,400 127,200 -3.2%
Total 159,200 166,200 4.4%
Percent of total
Hispanic 17.5% 23.5% 34.4%
non-Hispanic 82.5% 76.5% -7.3%
Total 100.0% 100.0% 0.0%
per 100,000 residents
Hispanic 72 70 -2.8%
non-Hispanic 53 48 -9.0%
Total 55 52 -5.5%
Data source: Bureau of Justice Statistics, Prisoners in 2014, Appendix Table 4, (2013); Prisoners in 2002, Table 15, page 10. (2001)

Interracial Crime

As noted above, in 2009, the US Justice Department’s National Crime Victimization Survey (NCVS) stopped publishing data on the races of criminals. Recently, however, figures from the 2013 NCVS were released in which blacks, whites, and Hispanics were, for the first time, treated separately.

From these data it can be calculated, first, that a great deal of violent crime — 40 to 60 percent — is committed within the races. However, when violence crosses racial lines it does so in a very unequal manner. In 2012 and 2013, blacks committed an annual average of 560,600 crimes of violence against whites whereas whites committed only about 99,400 such crimes against blacks. This means blacks were the attackers in 84.9 percent of the violent crimes involving blacks and whites.

Table 10. Violent Victimizations, 2012-2013
Race of victim Annual average
number of
Race of offender, %
White Black Hispanic Other Unknown
White 4,091,971 56 13.7 11.9 10.6 7.8
Black 955,800 10.4 62.2 4.7 15 7.7
Hispanic 995,996 21.7 21.1 38.6 11.6 6.9
Other 440,741 40.3 19.3 10.6 20.3 9.5
Total 6,484,507 42.9 22.4 14.8 12.1 7.8
Data Source: BJS, National Crime Victimization Survey, 2012–2013, Special Tabulation

Some observers argue that the overwhelming preponderance of black-on-white over white-on-black violence suggests that blacks deliberately target whites for violence. Others argue that since there are 4.7 times as many whites as blacks in the population, it is to be expected that black criminals are more likely to encounter white victims than vice versa.

Both positions must be evaluated in light of several considerations. First, blacks who commit assault, robbery, and rape are likely to be members of the underclass, who live in largely black neighborhoods. If they chose victims without regard to race they should be more likely to encounter other blacks rather than whites. Second, black/Hispanic interracial crime fits the same lopsided pattern: Of the 256,074 violent crimes involving those two groups, blacks were perpetrators 82.5 percent of the time. Unlike the nearly five-fold difference in numbers between blacks and whites, there are only about 30 percent more Hispanics than blacks. The high black-aggressor figure suggests that blacks may also deliberately target Hispanics — perhaps even more than they target whites.

The imbalance can be expressed differently: When whites commit violence they target other whites 82.4 percent of the time, blacks 3.6 percent of the time, and Hispanics 7.8 percent of the time. In other words, white violence is directed overwhelmingly at other whites. When blacks commit violence only a minority — 40.9 percent — of their victims are black. Whites are 38.6 percent and Hispanics are 14.5 percent. Hispanic assailants also attack their own group less often than they attack others. Their victims are: Hispanics — 40.1 percent, whites — 50.7 percent, and blacks — 4.7 percent.

Finally, interracial crime can be expressed in terms of the greater or lesser likelihood of a person of one race to commit violence against a member of the other. In 2012/2013, the actual likelihood of attack was extremely low in all cases, but statistically, any given black person was 27 times more likely to attack a white and six times more likely to attack a Hispanic than vice versa. A Hispanic was eight times more likely to attack a white than the reverse.

The Department of Justice keeps national records on murder. In 2013, it reported 5,621 single-offender, single-victim cases in which the race of the murderer was known. Like most federal statistics, there is no clear distinction between whites and Hispanics, so the only meaningful racial categories are black and non-black. Blacks killed 2,698 people — 48 percent of the total — and non-blacks killed 2,923 or 52 percent. Since blacks were just 13.3 percent of the population, it meant a black was six times more likely than a non-black to commit murder. Although most murders are within the same race, blacks were 13.6 times more likely to kill non-blacks than non-blacks were to kill blacks.

Police Killings of Blacks

There has been considerable concern about police killings of blacks, especially after Officer Darren Wilson shot Michael Brown on Aug. 9, 2014, in Ferguson, Missouri. Officer Wilson is white and Brown was black and unarmed. Although an extensive investigation by the Department of Justice found that Officer Wilson acted appropriately in response to Brown’s physical attacks and subsequent threats, the shooting provoked riots and gave impetus to the Black Lives Matter movement, which claims that racist police officers routinely kill unarmed blacks.

This claim has been difficult to verify, since there are no national statistics on police shootings. Moreover, as noted above, broad arrest statistics and academic studies do not support the view that arrests are biased, so there is little reason to think shootings would be biased. The Black Lives Matter movement has not been driven by data but by the deaths of specific people such as Michael Brown, Eric Garner, Tamir Rice, Sandra Bland, and Freddie Gray.

Freddie Gray’s death, due to injuries after his arrest in Baltimore, also provoked riots and was taken up with great energy by Black Lives Matter. Three of the six officers charged in Gray’s death were black — and they were charged with the most serious crimes — but the incident is still cited as an example of police racism.

In the absence of government data, the Washington Post investigated every reported case of a fatal shooting by the police during 2015. It found 990 cases, with the following racial distribution of victims:

White: 50.0 percent (495 victims)

Black: 26.1 percent (258)

Hispanic: 17.4 percent (172)

Asian: 1.4 percent (14)

Other/Unknown: 5.2 percent (51)

Given their proportions in the population, a black person was 2.45 times more likely than a white person to be shot and killed by police, a Hispanic was 1.24 times more likely, and an Asian was only one third as likely. It is reasonable to expect people of different races to find themselves in potentially lethal confrontations with the police in proportion to their likelihood to commit violent crime, with blacks most likely and Asians least likely.

As noted in Table 4 above, in California — a large state that keeps consistent statistics on race and ethnicity — blacks are arrested for violent crimes at 5.35 times the white rate, and Hispanics at 1.42 times the white rate. The low likelihood of Asians being killed by police is in keeping with low Asian arrest rates for violent crime. The black and Hispanic multiples for police shooting deaths are well within the arrest multiples — the black multiple is less than half — and certainly do not suggest undisciplined police violence.

Moreover, FBI data show that from 2005 to 2014, blacks accounted for 40 percent of police killings. Since blacks were approximately 13 percent of the population, it meant they were 4.46 times more likely than people of other races to kill a police officer.

In its study, the Post found that men were 22.9 times more likely than women to be shot and killed by officers. No one suggests that law enforcement bias accounts for this huge multiple, which is undoubtedly caused by differences in behavior between men and women. In the case of racial multiples, police bias cannot be ruled out, but it is reasonable to assume that the multiples are explained by race differences in behavior.

The Washington Post noted further that all but 93 of the 990 people fatally shot by police were armed, usually with a firearm or knife. The unarmed victims had the following racial distribution:

White: 34.4 percent (32 victims)

Black: 40.8 percent (38)

Hispanic: 19.4 percent (18)

Asian: 0 percent (0)

Unknown: 5.4 percent (5)

An unarmed black was therefore 5.6 times more likely than an unarmed white to be shot by police, and a Hispanic was 2.6 times more likely. The black multiple is certainly high, though not that much higher than the California violent-arrest multiple of 5.35 noted above.

There is no obvious explanation for why unarmed blacks were shot and killed at a white multiple that was twice that for armed blacks. If police bias is the cause, there is no clear reason why it should be worse in the case of unarmed suspects. The sample size of 93 is small, so random events produce a large effect.

It may be that race differences in how suspects behave when they are arrested explain at least part of the difference. There are no national data, but a five-year study of non-felony arrests in San Francisco found that blacks were 9.6 times more likely than whites (including Hispanics) to be charged with resisting arrest, and whites were 8.6 times more likely than Asians to be so charged. In Chicago, from September 2014 to September 2015, blacks accounted for 77 percent of arrests for obstruction of justice and resisting arrest (page 4 of report), meaning they were 6.8 times more likely than non-blacks to be arrested on these charges. If these findings are typical, they help explain why the arrest of a black non-felony suspect — who would more than likely be unarmed — could escalate into potentially lethal violence.

The Post’s analysis was intended to throw light on police bias but failed to indicate the races of the officers involved in fatal shootings. This would be useful information. A 2015 Department of Justice study (page 3) of police shootings in Philadelphia found racial differences in “threat perception failure,” that is, cases in which an officer shot an unarmed suspect because the officer thought the suspect was armed. Black officers were nearly twice as likely as white officers to shoot an unarmed black (11.4 percent of all shootings by black officers vs. 6.8 percent of all shootings by white officers). The percentage of such errors by Hispanic officers — 16.7 percent — was even higher.

Black officers may be somewhat more prone to error in general. About 12 percent of police officers in the United States are black. Between 2005 and 2015, 16.6 percent of the 54 officers criminally charged for fatally shooting someone while on duty were black.

Homicide is a serious problem for black men. Since at least 2002 and up to 2013 (the latest data available), murder was the leading cause of death for black men, ages 15 to 34. Their murderers are almost always other black men. According to a Department of Justice report, (page 13), from 1980 to 2008, 93 percent of black homicide victims were killed by blacks.

By contrast, the 256 police judicial killings of blacks in 2015 would be only 4.2 percent of the 6,095 blacks who were murdered in 2014 (the most recent year for which national data are available). The 38 unarmed blacks killed by police accounted for just 0.6 percent. Police shootings of unarmed blacks is a very small problem compared to murder in the black community.

Immigrants and Crime

When Donald Trump announced his candidacy for the Republican nomination on June 16, 2015, he said of Mexican immigrants: “They’re bringing drugs. They’re bringing crime. They’re rapists. And some, I assume, are good people.”

This caused much controversy, not least because even professional criminologists have difficulty quantifying immigrant criminality. Prisons can easily determine the race and ethnicity of inmates but determining immigration status is harder because it is generally based on self-reporting. Convicted felons are not eager to admit they are not citizens because they could be deported after they are released. So-called sanctuary cities forbid local police departments even from asking prisoners about immigration status.

Despite the difficulty in determining immigrant crime rates, Hispanics have higher crime rates than the majority white population, and many immigrants are Hispanic. Immigrants are people whom the United States chooses to admit to its territory. If the selection process were perfect, no criminals would be admitted and the immigrant crime rate would be zero. For immigrants to have even low crime rates reflects poorly on immigration policy.

Illegal immigrants are, by definition, not even selected, and there are only partial data on illegal immigrants, the crimes they commit, and where they are from. The Department of Justice’s State Criminal Alien Assistance Program (SCAAP) reimburses prisons and jails for the costs of holding illegal aliens. Prison systems must document the immigration status of inmates to get SCAAP payments — sanctuary jurisdictions choose to forego this subsidy.

In a 2011 report, “Criminal Alien Statistics,” the Government Accountability Office studied the 249,000 illegal aliens for whom SCAAP funds were paid in 2009. It found that this group of aliens had been arrested a total of 1.7 million times — an average of roughly seven arrests per illegal alien inmate — and had been charged with 2.9 million separate offenses, or roughly 12 offenses each.

All told, these criminal aliens accounted for the following numbers of arrests for the following crimes:

Homicides: 25,064

Sex offenses: 69,929

Assaults: 213,047

GAO found that about 66 percent of the SCAAP criminal illegal aliens in state prisons were born in Mexico and another 17 percent were born in the Dominican Republic, Guatemala, Honduras, El Salvador, Cuba, or Jamaica. Local jail inmates were even more heavily Mexican. Seventy percent were born in Mexico, while another 13 percent were from other Latin American countries. The operating costs (prison staff salaries, medical care, food, utilities) of incarcerating these criminal illegal aliens in state prison systems totaled $7 billion for fiscal years 2003 through 2009. This figure obviously does not include the costs of incarcerating illegal aliens in sanctuary jurisdictions because those costs are unknown.

Race and Drug Arrests

It is often claimed that blacks suffer unfairly from the enforcement of drug laws, and it is true that longer sentences for possession of crack cocaine rather than powder fell more heavily on blacks. As a 1997 report from the University of Chicago notes:

From 1965 through the early 1980s, blacks were approximately twice as likely as whites to be arrested for drug-related offenses … Following the federal government’s initiation of the “War on Drugs,” black arrest rates skyrocketed, while white arrest rates increased only slightly. By the end of the 1980s, blacks were more than five times more likely than whites to be arrested for drug-related offenses… [T]hese differences reflect the government’s targeting and enforcement of specific types of drug use and trafficking.6

Even setting aside the question of crack versus powder cocaine, it is often claimed that blacks suffer from police bias in drug arrests. It is common to allege that although blacks and whites use marijuana at roughly the same rate, blacks are nearly four times more likely than whites to be arrested for possession.

The assumption that whites and blacks take drugs at the same rate is based on answers to survey questions. However, there is evidence that blacks are less likely than whites to report illegal drug use. A number of studies have asked subjects about drug use and then checked their answers against the results of urine- or hair-analysis tests.

A 2005 study in the Journal of Urban Health, for example, found that blacks were ten times more likely than whites to lie about cocaine use. Hispanics were five times more likely to lie. There were similar differences in reported use of marijuana. The study concluded that “the results replicate and extend a growing body of research suggesting that African Americans underreport substance use on surveys.” Studies from 2003, 2008, and as long ago as 1994 report similar findings, though one from 2001 found ambiguous results.

There is other evidence that drug use is not the same across racial groups. The US Department of Health and Human Services keeps records of how many people of different races went to emergency rooms because of an acute reaction to illegal drugs. In 2011, the most recent year for which data are available, blacks were 2.8 times more likely than whites to end up in the ER because of marijuana, and seven times more likely because of cocaine. For all drugs, the multiple was 3.5. There is no reason to think these figures reflect anything other than different rates of illegal drug use.

Ever since receiving home rule in 1975, Washington, D.C., has had a black mayor and most of its police chiefs have been black. In any given year, as many as two-thirds of its police officers are black. And yet, in 2010, a black district resident was 8 times more likely than a white resident to be arrested for marijuana possession. It is hard to imagine this was because of police discrimination rather than differences in marijuana use.

Finally, it would be possible for blacks to be no more likely than whites to use drugs but still be arrested more often for using them, even by scrupulously race-neutral police. That is because blacks commit a larger number of other crimes. If someone is arrested for robbery, for example — and in California blacks were more than 13 times more likely than whites to be arrested for robbery — the police search the suspect for drugs. If they find drugs they add a charge of possession in addition to robbery. Higher rates of illegal activity expose blacks to more intense criminal processing.

In any case, since the last “Color of Crime,” the black/white incarceration multiple for drug offenses has declined by nearly 60 percent (see Table 8). This probably reflects the government’s response to prison overcrowding as well as a reorientation of enforcement activity in the War on Drugs rather than a decline in police bias.

Crime and Punishment

High incarceration rates combined with declines in violent crime suggest that incarceration reduces crime. Those who disagree with this conclusion argue that if one criminal is locked up, another will simply take his place. However, criminals do not wait for a competitor to leave the business. Opportunities for violence and theft are virtually unlimited.

A second argument against mass incarceration is based on the theory of diminishing returns. As incarceration rates go higher, police start scraping the bottom of the criminal barrel. The marginal prisoner becomes less and less dangerous compared to those already behind bars.

In fact, the chances of a criminal being caught and sent to prison are very low. Even the JFA Institute, an anti-incarceration advocacy group, estimates that in only 3 percent of violent or property crimes do offenders end up in prison. According to the Bureau of Justice Statistics (BJS), in 2004 only 1.6 percent of burglars were in prison.

Recidivism rates are strong evidence that incarceration prevents crime. The BJS tracked 404,638 state prisoners released in 2005. Within one year, 43.4 percent had been rearrested. This percentage grew with every succeeding year, rising to 76.7 percent at the end of five years. At that point, the released prisoners had been arrested a total of 1,173,000 times, or an average of 2.9 times each. These figures do not include the estimated 12 to 15 crimes a felon commits every year when he is not imprisoned and that do not result in an arrest.7

Many people believe that prisons are filled with non-violent drug offenders. It is true that half of federal prisoners are guilty of drug crimes — almost always trafficking — but they account for only 13 percent of the nation’s prison population, with the rest in state prison. Casual drug use almost never results in a prison term. In 2013, according to the Bureau of Justice Statistics (page 15 of report), only 3.7 percent of state prisoners were convicted only of drug possession, and this was generally the result of a plea bargain to avoid charges of trafficking. Most had long prior records. Only 12.2 percent of state prisoners were convicted even of trafficking. Fully 53.8 percent of state prisoners were guilty of violent crimes, and 18.8 percent had committed property crimes.

America’s high incarceration rate is nothing to be proud of, but any significant reduction in the inmate population is likely to release large numbers of unreformed criminals who will prey on law-abiding people of all races.

A Ferguson Effect?

The year 2015 ended on a worrying note. National figures for all index crimes had not yet been compiled at the time this report was completed, but one crime had clearly reversed its decades-long downward trend: murder. In the most populous American cities, murder increased at the following rates over 2014:

New York City: 5 percent

Los Angeles: 10 percent

Chicago: 13 percent

Houston: 23 percent

Philadelphia: 12 percent

In Baltimore, murders were up a shocking 63 percent and gun violence (including non-fatal shootings) was up 75 percent. Washington, D.C., saw an increase of 54 percent. In Milwaukee, murders were up 65 percent in 2015, and 91 percent of the victims were either black (83 percent) or Hispanic (8 percent). Detroit bucked the trend and saw a decline of 1.3 percent.

There was some speculation that the overall rise was caused because of less active policing, or “de-policing.” Officers may have been unwilling to take risks in crime prevention for fear that even a slight misstep could come under career-ending scrutiny. This was known as “the Ferguson effect.” The officer who shot Michael Brown in Ferguson, Missouri, was forced out of his job because of media coverage and public protest even though a Justice Department study later found that he had acted properly. The rise of the Black Lives Matter movement and harsh anti-police rhetoric also put police under intense scrutiny, which may have led to less aggressive crime fighting.

Baltimore appears to have suffered from the Ferguson effect. There was a sharp rise in gun violence in the months after the riots of April, 2015, during which the police did not arrest looters because they were told not to “engage” with rioters. The riots also led to the indictment in May of the six officers involved in the death of Freddie Gray. Baltimore police felt betrayed by charges that some legal experts considered unreasonable. A number of department veterans predicted that demoralized officers would take fewer risks and that there would be a spike in crime. That is exactly what happened.

Whatever the cause, the extraordinary decline in crime rates since the 1990s may have come to an end.


  1. We present “Two or more races” as a separate victim group in 2013 because the NCVS breaks it out that way. As noted below, however, incarceration figures published by the Bureau of Justice Statistics lump all non-Hispanic prisoners of “Two or more races” into the “Other, non-Hispanic” category. Had we done this in the above table, violent victimization of “Other, non-Hispanics” would have risen by 190 percent since 2002. In fact, there are undoubtedly people who identified themselves as white or black in 2002 who would chose “Two or more races” in 2013. To the extent that this is true, the decline in white and black victimization rates in the table is overstated.
  2. High crime rates among blacks are not limited to the United States. Statistics released by the Metropolitan Police in London, England, show that in 2009/10 blacks accounted for 54 percent of arrests for street crimes, 59 percent for robbery, and 67 percent for gun crimes. Blacks accounted for just over 12 percent of London’s population of 7.5 million.Likewise, according to information that had to be sought through a freedom-of-information request, in 2002, blacks were 8.1 percent of the population of Toronto, Canada, but accounted for 27 percent of all charges for violent crimes.
  3. “Other, non-Hispanic” victims of violent crime are also far less likely to report their victimizations to police: only 12.8 percent of them did so in 2013, according to the NCVS. The average reporting rate for all victims that year was 44.3 percent.
  4. For the best argued case for this view, see Ron Unz, “His-panic,” The American Conservative, March 2010.
  5. The source for federal prison inmates cited in the earlier “Color of Crime,” the Compendium of Federal Justice Statistics, is no longer published. Current federal inmate data are available, in database format, from the National Archive of Criminal Justice Data (NACJD). Ten years ago NACJD files were available to the public; today access is restricted. In any event, the state prison population is about six times larger than the federal prison population. It is not surprising, therefore, that our 2013 incarceration rate multiples for blacks and Hispanics, which are based on state prison inmates alone, are very close to the multiples in the earlier report, which were based on combined state and federal inmates.
  6. Robert J. Sampson and Janet L. Lauritsen, “Racial and Ethnic Disparities in Crime and Criminal Justice in the United States,” Crime and Justice, University of Chicago Press, 1997, page 327.
  7. Paul G. Cassell, “Too Severe? A Defense of the Federal Sentencing Guidelines,” Stanford Law Review, 56 (2004): 1017-48.

About the author

Edwin S. Rubenstein is the president of ESR Research. He has worked as a senior economist at W.R. Grace & Co., and as research director at the Hudson Institute. His articles have appeared in Harvard Business Review, New York Times, Wall Street Journal, and Investor’s Business Daily. He is the author of two books: The Right Data and From the Empire State to the Vampire State: New York in a Downward Transition.

Share This