clock menu more-arrow no yes

Murders are up. Crime is not. What’s going on?

The US is in its most violent period in the 21st century. There are some things we know — and others we don’t.

Police cars on a dark street behind a “Police line, do not cross” tape.
Police gather at the scene where two New York City police officers were shot in a confrontation on June 3, 2020.
Spencer Platt/Getty Images

Last year, the US saw the biggest increase in the murder rate in decades. The estimated total number of homicides rose to levels not seen since the late 1990s, even as the overall crime rate declined. So far, the spike has continued into 2021: Murders are up nearly 15 percent so far this year compared to the same period last year, based on data from US cities collected by crime analyst Jeff Asher.

That’s what we know. What we don’t really know yet is why.

The rise in murders, along with a rise in shootings, is reflected not just in the statistics but in real events. Over the weekend, the sound of gunfire sent baseball players and fans into a panic at the Nationals stadium in Washington, DC. The night before, DC’s mayor condemned a drive-by shooting that killed a 6-year-old. Meanwhile, a total of 12 mass shootings from coast to coast left 11 dead and 49 injured in the US.

Year-to-year fluctuations in crime and violence can and do happen. But the size of the murder spike has led to broader national attention. The increase is now part of an ideological proxy war — leading to conflicting opinions even within political parties on what to do about the increase in murders, and plenty of finger-pointing over whether the pandemic, protests over police, or guns are to blame.

We don’t really know, with certainty, what’s behind the rise. All three of those factors likely played a role. And there may even be some unknown factor that researchers won’t notice for years; the theory that higher levels of lead in the environment caused higher crime and violence from the 1960s to 1990s took decades to get widespread national attention.

“There’s a constellation of things going on that make it extraordinarily difficult to be able to pin down a specific explanation,” Morgan Williams, an economist at Barnard College who studies crime and the criminal justice system, told me.

The increase in murder appears to be a uniquely American phenomenon. While murder rates rose in some developed countries last year, like Canada and Germany, the increases are far below the double-digit spikes America is seeing. That’s especially notable because the United States already had a higher baseline of murders, after controlling for population. Despite claims that Democratic mayors or progressive criminal justice policies are driving the increase, it also appears indifferent to the political party in charge: As Asher and criminal justice expert John Pfaff have shown, murder rates increased in cities run by Democrats and Republicans, progressive and not.

The good news is there is a lot more agreement among experts about how to bring down the spike than there is about what caused it. But the best evidence suggests stopping murders in the short term will require more and better, though not necessarily more aggressive, policing — a controversial proposal on the left.

“I know people don’t want to hear this, and I empathize with that,” Anna Harvey, a public safety expert at New York University, told me. “But at least as far as the research evidence goes, for short-term responses to increases in homicides, the evidence is strongest for the police-based solutions.”

The stakes are very high. Nearly 21,000 people were murdered in America in 2020, based on preliminary data. Another increase of 10 percent or more could mean thousands more dead in 2021.

The murder spike is real

The United States saw huge spikes in murders and other crimes from the 1970s to 1990s. But starting in the mid-1990s, the country hit a period of peace: Through 2014, rates of murder, violence, and other kinds of crime declined by more than half. In 2015 and 2016, the murder rate increased but quickly leveled off, then dropped again.

Then, in 2020, murders rose by an estimated 25 percent. As Princeton sociologist Pat Sharkey previously told me, “Last year was clearly the most violent year of the [21st] century so far.”

Chart showing based on preliminary FBI data, the US’s murder rate increased by 25 percent or more in 2020. That amounts to more than 20,000 murders in a year for the first time since 1995, up from about 16,000 in 2019.

The data is preliminary; final official numbers for 2020 will be out later this year. But the findings have been backed by multiple sources, including the FBI, Asher, separate reports from the Council on Criminal Justice, and the University of Pennsylvania–run website City Crime Stats.

A consistent finding in these analyses: The spike is truly national, showing up in every region of the country and most of the cities with available data.

A chart depicting murders in each region of the US.

Some other kinds of crime also increased, according to this early data, including shootings, aggravated assaults, and car thefts. Still, violent crime in general went up at much lower rates, if at all, compared to murders, and overall crime declined, driven in part by a drop in the majority of property crimes.

The split between murder rates and crime rates might seem odd, but there’s good reason to believe the divergence is genuine and not an artifact of underreporting. There were fewer opportunities to commit property crimes last year with businesses shut and people staying home. The one type of property crime that did increase, car theft, is often committed as part of a larger, more serious crime. It’s what criminologists call a “keystone offense”: stealing a car to use in a drive-by shooting, for example, so perpetrators can’t be easily identified.

Based on Asher’s analysis of major US cities, the murder spike has continued into 2021 but likely decelerated. There also seems to be more variation: More cities, including Chicago, are reporting a decrease or at least no increase in murders so far this year. “It’s become a little more concentrated,” Sharkey said.

A leveling off of murder rates is better than a continued increase, but it isn’t great. It means people are still being killed at the highest level this century. Still, it could indicate the spike may be a brief aberration rather than the start of a new trend — a repeat of the 2015-2016 spike.

We still don’t know the causes

The closest to a consensus I’ve been able to find in talking to experts about the cause of the murder spike: It’s complicated.

Experts have rejected some possibilities. Given that murders rose in both Democrat- and Republican-run cities, as well as places that adopted criminal justice reforms and those that didn’t, partisanship and criminal justice reforms don’t seem to be a cause.

Three plausible explanations, none of which exclude the others, have come up repeatedly:

1) The Covid-19 pandemic: The coronavirus was a massive force in 2020, and may have also affected murder rates. The pandemic shut down programs that likely safeguard Americans from violence, including policing, social services, and community-led efforts. It left some people, particularly teen boys and young men, with more free time to stew over interpersonal conflict as workplaces and schools shut down. And it fed a general sense of chaos and despair throughout the year, perhaps amplifying perceptions that desperate times can call for desperate measures.

But much of the world also struggled with Covid-19, from Mexico to Canada to much of Europe, and didn’t see double-digit percent increases in murders last year. That suggests the virus can’t be the sole cause.

2) The US protests over police brutality: Beginning with the Ferguson, Missouri, protests in 2014, protests over Black men and boys killed by police — Michael Brown in Ferguson, Tamir Rice in Cleveland, Freddie Gray in Baltimore, and others — swept through cities, followed by a rise in murder and sometimes other violence. Several sides in the ensuing debate claimed a “Ferguson effect,” although there’s been little empirical research to date on the issue.

One theory held that officers, afraid of getting caught in the next viral moment that leads to protests, backed off from proactive policing. On the other side, the public could have lost trust in police and been less likely to cooperate as witnesses or informers, making it harder to close cases, make arrests, and get dangerous people off the streets. A greater sense that the criminal justice system can’t be trusted also could have led people to take matters, violently, into their own hands.

The same thing could have led to the 2020-2021 spike, following the mass protests over the murder of George Floyd by a Minneapolis police officer. “It’s so much like the conversation we were having five or six years ago,” Richard Rosenfeld, a criminologist at the University of Missouri St. Louis, told me.

3) America’s gun problem: The US has the most number of guns in civilian hands, and the last year saw a huge spike in the number of firearms purchased by Americans. The research is clear here: More guns mean more gun violence — and more deadly violence, because the presence of a gun allows just about any conflict, from public arguments to domestic abuse, to escalate.

A recent study found that the number of gun purchases in a state didn’t correlate with more gun violence. But the data in that study ended in July 2020, before much of the country saw the biggest murder increases.

Even if new gun purchases weren’t to blame, it’s possible existing guns are: Asher found evidence that more people were carrying guns last year, leading to more police finding guns in the course of an arrest. So perhaps it’s not so much that people bought new firearms but that they started carrying the arsenal of weapons they already had.

Perhaps the best explanation: All of these factors played a role.

There are many ways all these explanations could have interacted. As one example: Covid-19 and protests both fueled a sense that the social fabric was unraveling, and more people — particularly in the worst-off neighborhoods — felt they had to fend for themselves. They equipped themselves with guns to act on their own if they felt a threat. And this made any given conflict more likely to escalate to deadly violence.

Ultimately, though, there are too many unknowns to draw hard conclusions.

We do know about the solutions

Despite the lack of clarity about what caused the spike in violence, there’s a lot of evidence for solutions that could bring down the trend — regardless of what may be the cause.

The best research for controlling crime and violence quickly indicates the police must play a role. This doesn’t have to mean punitive practices like stop-and-frisk and arrests over low-level offenses, which have drawn the ire of protesters over the years. (In fact, the experts I spoke to were clear it shouldn’t.) But there’s strong evidence that more police lead to fewer homicides, and solid research backs strategies like hot spot policing and problem-oriented policing.

These strategies tend to be more focused, like hot spot policing’s heightened surveillance of very specific high-crime blocks. Or they tend to be more planned: Problem-oriented policing requires formal evaluations of a problem and solutions, and calls for bringing in community partners to make sure the issue is addressed at its root. It’s a shift from dragnet efforts in which officers target entire neighborhoods to stop or arrest as many people as possible.

In fact, these approaches can actually reduce overall incarceration. For example, the evidence for hot spot policing suggests that officers’ mere presence deters crime, since people are less likely to do illegal things in front of a cop. Police don’t have to do anything — just stand there and watch. And fewer crimes committed means fewer arrests.

“It’s not like you’re putting more police in the streets and they’re making a bunch of arrests,” Harvey said. “There are cost-effective ways of policing violent crime that are more effective and not creating the collateral consequences that some of the other strategies create.”

Earlier this year, President Joe Biden and Democrats in Congress approved an economic relief package that includes funding for local and state governments to hire more police for “community policing.” A challenge for progressive politicians, though, is that many of them have wanted to pull back from relying so much on police or cut funding to departments altogether.

Some alternatives to policing, such as community-based violence interrupters, haven’t panned out well. Programs with “violence interrupters,” in which community members intervene to stop a conflict from escalating, have drawn praise, a documentary, and now federal funds from the Biden administration — but the evidence for this approach is weak and mixed.

Other alternatives have fared better. There’s good evidence for providing summer jobs programs, raising the age to drop out of school, greening vacant lots, installing more streetlights, providing more drug addiction treatment, implementing better gun control, and raising the alcohol tax, among other ideas.

The problem, experts told me, is that even the effective non-police strategies tend to take time to work. Police can be active on a high-crime block in minutes, but it can require years to lift up people and neighborhoods, economically and otherwise, and address root causes of crime that these alternatives are supposed to target. They aren’t all designed to reduce the number of murders quickly.

“It doesn’t mean police are a panacea for these things,” Williams said. “But it does mean we should be very careful about throwing around interventions that we don’t necessarily know come with any important benefits or costs.”

Sign up for the newsletter Sign up for The Weeds

Get our essential policy newsletter delivered Fridays.