Federal statistics on violent crime by race come from two very different sources, and they tell overlapping but distinct stories. FBI arrest records show who police take into custody. The Bureau of Justice Statistics’ National Crime Victimization Survey (NCVS) shows who violence happens to and whom victims identify as the offender. Read together, both datasets point to the same core findings: Black Americans are arrested and victimized at rates higher than their share of the population, most violent crime occurs between people of the same race, and the racial gaps shrink dramatically once poverty and neighborhood are taken into account.
What Arrest Data Shows
The FBI’s 2019 arrest tables are the last complete nationwide picture before the agency’s shift to the National Incident-Based Reporting System disrupted collection. For that year, White individuals accounted for 44.7% of robbery arrests and Black individuals for 52.7%. For aggravated assault, White individuals accounted for 61.8% and Black individuals for 33.2%.1Federal Bureau of Investigation. Crime in the U.S. 2019 – Table 43 On homicide, where the offender’s race was known, 55.9% were Black or African American and 41.1% were White.2Federal Bureau of Investigation. FBI – Expanded Homicide
Those shares need population context. Census Bureau estimates for 2024 put non-Hispanic White Americans at about 57.5% of the U.S. population and Black Americans at about 13.7%.3U.S. Census Bureau. United States QuickFacts Black Americans are arrested for violent offenses at rates substantially higher than their share of the population; White Americans are arrested at rates somewhat below theirs. The raw percentages don’t explain why, and as the sections below describe, treating arrest shares as a direct measure of who commits crime overstates what the numbers can support.
Violent Victimization Rates by Race
The NCVS interviews roughly 240,000 people across 150,000 households each year about violence they experienced, whether or not they called police.4Bureau of Justice Statistics. National Crime Victimization Survey The 2023 survey found the following victimization rates per 1,000 persons age 12 or older:5Bureau of Justice Statistics. Criminal Victimization, 2023
- Black (non-Hispanic): 26.9
- White (non-Hispanic): 22.5
- Hispanic: 21.3
- Asian, Native Hawaiian, or Pacific Islander: 10.7
- Other, including American Indian, Alaska Native, and multiracial individuals: 50.4
The “Other” rate is roughly double the national average. Sample sizes for this group are small, which reduces precision, but the pattern is persistent. BJS data covering 2017 through 2021 showed a similar rate of 54.8 per 1,000 for the same group.6Bureau of Justice Statistics. Violent Victimization by Race or Hispanic Origin, 2008-2021
When simple assault is set aside and only serious violence is counted, the Black victimization rate rises to 12.3 per 1,000 and the White rate to 8.3 per 1,000, a wider gap than the overall numbers suggest.5Bureau of Justice Statistics. Criminal Victimization, 2023 Racial disparity in violent crime shows up on the victim side at least as strongly as on the arrest side.
Most Violent Crime Is Intraracial
One of the most consistent findings in federal crime data is that violence overwhelmingly occurs between people of the same race. In the 2023 NCVS, among violent incidents where the victim could identify the offender’s race:5Bureau of Justice Statistics. Criminal Victimization, 2023
- White victims: about 1.95 million of 2.93 million known-offender incidents involved a White offender (roughly 66%). Black offenders accounted for about 13%.
- Black victims: about 487,000 of 670,000 known-offender incidents involved a Black offender (roughly 73%). White offenders accounted for about 18%.
- Hispanic victims: about 275,000 of 906,000 known-offender incidents involved a Hispanic offender (roughly 30%), with White offenders accounting for about 42% and Black offenders about 21%.
The lower intraracial share for Hispanic victims reflects demographic overlap: Hispanic communities often share neighborhoods with non-Hispanic White and Black populations, which raises the odds of cross-group contact. Earlier BJS data from 2020 showed the same basic pattern, with 69% of violence against White victims and 66% of violence against Black victims committed by someone of the same race.7Bureau of Justice Statistics. Criminal Victimization, 2020 – Supplemental Statistical Tables
People live near, work with, and socialize with others of the same racial background. Violence grows out of proximity, personal relationships, and neighborhood conflicts far more often than it crosses racial lines. Interracial violence exists, but it is the minority of incidents in every racial group measured.
Poverty and Geography Change the Picture
Numbers broken down by race alone look very different once income is controlled for. A BJS study covering 2008 through 2012 found that people living at or below the federal poverty level experienced violent victimization at 39.8 per 1,000, more than double the 16.9 per 1,000 rate for high-income households. The pattern held across racial groups: poor White, poor Black, and poor Hispanic residents all faced far more violence than higher-income counterparts of any race.8Bureau of Justice Statistics. Household Poverty and Nonfatal Violent Victimization
The sharpest finding came from comparing people at the same income level in the same setting. Poor urban Black residents experienced violent victimization at 51.3 per 1,000; poor urban White residents, at 56.4 per 1,000. At similar poverty levels in similar neighborhoods, the racial gap in victimization largely disappeared. Poverty also raised the firearm victimization rate to 3.5 per 1,000, compared to 0.8 to 2.5 per 1,000 for people above the poverty line.8Bureau of Justice Statistics. Household Poverty and Nonfatal Violent Victimization
Black Americans are disproportionately concentrated in high-poverty urban neighborhoods because of generations of housing segregation, disinvestment, and unequal access to education and employment. Those conditions correlate with higher crime rates regardless of the race of the residents. Presenting violent crime by race without this context invites readers to treat correlation as causation.
Why Arrest Rates Aren’t the Same as Offense Rates
Arrest data measures police activity, not crime itself. Three factors create a real gap between who commits crime and who gets arrested for it.
Policing is not evenly distributed. Neighborhoods with heavier police presence generate more arrests for the same underlying level of criminal activity. Because many high-poverty, predominantly Black neighborhoods are more heavily policed, residents there face a higher probability of arrest for conduct that might go undetected elsewhere. Research on traffic stops has found that Black drivers are stopped at higher rates than White drivers during daylight, when a driver’s race is visible, and the disparity shrinks after dark.
Arrest is not the same as guilt. A substantial share of arrests do not result in conviction, and studies have found that Black individuals are more likely than White individuals to be arrested without any illegal activity being confirmed. The gap between arrest rates and actual offense rates is clearest in drug enforcement, where surveys show similar usage rates across racial groups despite very different arrest rates.
Recent arrest data also has structural gaps. The FBI started phasing out its old summary reporting system in favor of NIBRS, which captures more detail about each incident, including offender and victim demographics.9Federal Bureau of Investigation. National Incident-Based Reporting System As of late 2024, about 76% of law enforcement agencies, covering roughly 87% of the U.S. population, report NIBRS data, but the five most populous states have a combined agency participation rate of only 48%.10Every CRS Report. Benefits of NIBRS The missing jurisdictions are not demographically random, which can skew the national picture in ways that are hard to quantify.
Arrest data is not useless. It is one of the few large-scale datasets on violent crime, and the patterns it shows are worth understanding. Treating arrest percentages as a precise measure of who commits crime, though, confuses the map for the territory.
How Race Gets Recorded
Police agencies and the NCVS record race in opposite ways, and both methods have limits. Officers filling out incident reports typically record an offender’s race based on official identification or their own visual assessment, which may not match how the person identifies. The NCVS instead asks victims to identify their own race and describe the perceived race of the person who victimized them, which captures the victim’s experience but introduces error from cross-racial identification, especially in brief or traumatic encounters.
Both datasets currently use the older federal race categories, with Hispanic origin asked separately from race. That separation has long complicated interpretation, because a person can be classified as both White and Hispanic in different tables. When reading any crime statistic by race, it is worth checking whether Hispanic individuals are counted within the White category or broken out separately, since the choice can shift the apparent White share noticeably.