Police Violence Has Not Declined Despite Black Lives Matter-Era Policy Changes

Meaningful Change Requires a Shift from Punishment to Care

Following the rise of the Black Lives Matter movement, various policy changes were passed to address police violence across the country. However, our new study published in the medical journal JAMA Network Open found that there has been no meaningful change in the national rate of nonfatal injuries caused by police violence between 2004 and 2021. These findings contribute to a growing body of evidence suggesting there has been little progress toward reducing police violence in the U.S. despite decades of widespread protest and policy reform.

The Black Lives Matter movement first mobilized against anti-Black violence in 2013 following the acquittal of George Zimmerman for killing Trayvon Martin. It brought an unprecedented degree of attention to the issue of police violence after the 2020 lynching of George Floyd. These mobilizations inspired a wave of state and local reforms aimed at reducing police violence. New regulations on use of force, such as chokehold bans, were especially widespread. After 2020, many jurisdictions also made efforts to limit the scope of policing by establishing programs that deployed alternative responders, not law enforcement, for mental health crisis response and community violence interruption.

Findings Reveal No Change in Levels of Police Violence

Despite these efforts, the overall level of police violence has remained relatively constant. In our new study, we found that there was no clear change in nonfatal injuries caused by police over the observed period — which included these periods of intense protest. As with any statistical model, our estimates come with some amount of statistical uncertainty. Taking into account our margin for error, the rate of nonfatal injuries could have increased as much as 20% or reduced as much as 30%. 

These findings leave open the possibility that there was no actual change at all. These findings complement evidence from other research, which found that rates of killings by U.S. police have increased in the four years after the murder of George Floyd. In short, if these policy changes were intended to decrease rates of police violence, they do not appear to have succeeded.

In Conclusion

We never should have expected immediate, meaningful improvements. Some of the reforms that were implemented targeted very narrow categories of police behavior that wouldn’t necessarily account for a large proportion of police activity. Others were not funded at a large enough scale to make a sizable impact. For example, while chokeholds are highly dangerous and banning them is appropriate, they accounted for well under one percent of all police killings nationally¹, even before restrictions were widespread. Similarly, most alternative crisis response programs have committed a fraction of the resources that are needed to be effective: many would need to expand by 50- to 100-fold to achieve the service level of CAHOOTS², the long-standing crisis response program in Oregon that has served as a national model. While funding for alternative approaches has stalled, policing continues to absorb more public resources — local government spending on law enforcement increased by 10% between 2012 and 2022 after adjusting for inflation³.

Taken together, these findings underscore the limits of narrow reforms within a system structured around punishment and control. Lasting reductions in police violence will require a broader reorientation of public safety policy — from the harmful status quo of reliance on law enforcement to investments in care, prevention, and community well-being.

Footnotes:

1. Based on our analysis of data from the Associated Press Lethal Restraint dataset for 2012 and 2013. Over this period, an average of 5 police killings per year involved chokeholds. Under the conservative assumption of 1,000 total police killings per year, this translates to 0.5% involving chokeholds.
2. Based on our analysis of program-reported data, which shows that CAHOOTS responds to approximately 90 daily calls per 100,000 population in the jurisdictions that the program serves. As one contrasting example, New York City’s B-HEARD pilot program responded to approximately 0.5 calls per day per 100,000 population in the 2024 fiscal year, which was the third year of the program’s operation. B-HEARD would need to expand by 180-fold to achieve the response rate of CAHOOTS.
3. Based on our analysis of data from the US Census Bureau’s Annual Survey of State and Local Government Finances, which shows that direct expenditures on law enforcement by local governments increased from $108 billion in 2012 to $119 billion in 2022 (both are reported in 2022 dollars).