Did the homicide decline change racial disparities in firearm homicide victimization?
Does the cyclical nature of violence epidemics effect racial disparities in victimization?
A couple of years ago, I had the good fortune to write a short article with Alex Piquero for JAMA, which reported how racial disparities in firearms homicide victimization changed during the pandemic. That analysis examined 2018-2022 and captured a period where homicide spiked. Now, we have data through 2025, which includes both the period when homicide spiked and the current, rapid decline. Here’s an updated graph showing homicide declines, where the magnitude of the current decline is self-evident.
The key finding from the JAMA paper was that among 15-19 year-olds, the ratio of Black firearms homicide victimization to White firearms homicide victimization peaked at 27:1 in 2021. You can read the methods and other findings from the original paper here, and I have used an identical approach in this analysis.1
Expanding the analysis through 2025, we find that some of the key conclusions remain.
This figure shows firearms homicide victimization rates per 100,000 people. Let me orient you to these graphics, since the next two will follow the same format. These are bar graphs grouped by five-year age bands. Each color in the bar represents the value for a specific year, beginning with 2018 and ending with 2025. The top left panel shows the overall rate. The top right panel shows white (not Hispanic or Latino). The bottom left panel is Black (not Hispanic or Latino). And the bottom-right is Hispanic or Latino.
The highest firearms victimization rates overall and within groups were in 2021.
By year and age group, White (not Hispanic or Latino) rates peaked in 2021 for 30-34 at 4.3 per 100,000. Black (not Hispanic or Latino) rates peaked in 2021 for 20-24 at 81.2 per 100,000. Hispanic or Latino rates peaked in 2021 for 20-24 at 13.6 per 100,000.
By 2025, those rates had declined precipitously for all ages and races. One key research question is whether the racial disparities declined at similar rates.
This figure shows the rates by year and five-year age groups. As noted earlier, the ratio of Black-White firearms homicide victimization peaks in 2021 for 15-19 year-olds, at 27:1. This is a substantial increase from 2018, when the ratio was 18.1:1. Notably, the homicide decline that reduced overall firearms homicide below the 2018 level did not reduce the black-white firearms homicide victimization as much. In 2025, that ratio was 21.1:1.
Zooming in, though, it becomes apparent that the reduction in the disparity ratio among 15-24 year-olds is a bit unusual: older age groups show a flatter trajectory.
Looking into the data, it appears that the 15-24 Black-White ratio shows a much larger decline than among 25-49 year-olds (under age 15 and over age 50 firearms homicide victimization is much more unusual).
Interpretation
These analyses are meaningful in multiple dimensions, but I want to focus on two here.
With respect to the ratio of black-white disparity in firearm homicide victimization, there is clearly a narrowing of the gap among the highest risk groups. It might not be clear to all why this is somewhat unexpected. All else being equal, it would be reasonable to expect that declines would be equally distributed, that there would be fewer homicide victims, but the ratio would stay the same.
For a ratio to change, either the numerator or the denominator must change more than the other component. In this case, the numerator is the Black non-Hispanic firearm homicide victimization rate, and the denominator is the White non-Hispanic firearm homicide victimization rate.
Suppose the Black-to-White rate ratio was 5:1 in one year and 3:1 a few years later. That means Black victims were still experiencing firearm homicide at a higher rate than White victims, but the size of the disparity had narrowed. If overall firearm homicide victimization declined during that period, a falling ratio means that the Black victimization rate declined more rapidly, in relative terms, than the White victimization rate.
The clearest pattern in these data is not simply that firearm homicide victimization has declined since the pandemic peak, but that the decline has been especially large among Black non-Hispanic youth and young adults. Because those groups experienced the highest victimization rates during the surge, even a substantial decline still leaves rates far above those of White non-Hispanic youth and young adults. The result is a partial narrowing of the Black-to-White victimization ratio, especially among people ages 15 to 34. This is encouraging, but the disparity remains very large.
That leads me to once again call for more macrocriminology research. Since there are comparatively few studies that investigate crime cycles and national statistics, we don’t know why the Black-White firearms homicide ratio grew so large, and we don’t know why it is declining.
I have argued here that macrocriminology is an interesting science to explore because there are all sorts of crime-related phenomena that occur at the national level and can best be understood as such. Obviously, over the last six years, we have experienced two major crime cycles: a rapid peak and a rapid decline. Now, of course, this does not, by itself, prove that crime is routinely cyclical. But it is certainly evidence that crime is occasionally cyclical.
Here, I think the aggregate data suggests some possible explanations. This obviously needs testing.
The data do suggest that the worst of the homicide victimization fell disproportionately among young Black men, 15-24. The disparity ratio widened amid the explosion in violence in 2020 and 2021. That suggests that the cause was something relatively unique to this age group.
Routine Activities Theory is a fine place to start looking for answers, and where I think the answers are. School and business closures disproportionately impacted young people, especially those who lived in places with a history of trauma and violence. This left young people with serious disputes in close proximity to each other, and with few pro-social options, which was a recipe for violence. Others have found that those dislocations from schools were long-lasting and were a key predictor of the surge in violence.
The reversal we have seen is from a slow reengagement with formal and pro-social institutions, largely driven by large federal investments in local governments.
All of this leaves open the question of what can be done to drive down the firearms homicide victimization rate among young Black men. For the current decline to be sustainable, that question needs an answer.
