clock menu more-arrow no yes

The truth about violent crime in American cities, explained in 11 charts

Trump speaks of “anarchy and mayhem” in cities. But understanding violent crime is much more complicated than simply looking at numbers.

A Minneapolis police officer unrolls caution tape at a crime scene on June 16.
| Stephen Maturen/Getty Images

Next week, President Donald Trump and former Vice President Joe Biden will face each other for the first time on the debate stage. Some of the most pressing problems of our time will be front and center: the coronavirus pandemic, the Supreme Court vacancy, and the fight for racial justice. So far, the candidates’ discussion of justice issues has focused less on how to address America’s longstanding inequity and more on how cities are facing a violent crime surge in a time of unrest — and who is to blame.

Trump and his supporters have repeatedly spoken of bringing “law and order” to Democrat-run cities that are full of “anarchy and mayhem,” even though racial justice protests around the country this summer have been mostly peaceful. Biden, on the other hand, has mostly skirted talk of unrest, emphasizing that the crime rate dropped while he was the vice president and that a surge of murders happened under Trump’s watch.

Wading through these mixed messages of what’s happening in cities, it’s hard to tell just what the data says. Most types of crime decreased this summer, while serious violent crimes — such as aggravated assault and murder — increased, according to an analysis of crime rates in 27 major US cities by the Council on Criminal Justice, a criminal justice think tank. A preliminary crime report published by the FBI earlier this month shows similar trends nationwide.

To make sense of what this all means, the Marshall Project and Vox have parsed findings from January to June, as well as decades prior for comparison, of not just crime data but media reports, public opinion polls, and stats on policing and jail populations. Politicians and pundits are pointing fingers at what they believe caused the increase in violent crime rates: the protests against police violence, movements to defund the police, and efforts to release people from overcrowded jails and prisons ravaged by the coronavirus. But the data available thus far does not support that these are the culprits.

Understanding what drives crime rates is tricky because there’s no single cause or answer. This is especially true in the pandemic, which has introduced unfamiliar patterns. What is known, however, is that sensational media reports and misleading statements from politicians can blow the degree of violence out of proportion and make the public believe that crime is increasing, even when it isn’t.

As the country gears up for the presidential election — and the messaging of politicians and the media that comes with it — here are 11 data visualizations, along with analysis, that can help think through what the summer’s crime trends mean and how to move forward.

Violent crime was up in early summer; nonviolent and property crime was down

Beginning in late March, cities across the country saw a decrease in most types of crime, including burglary, theft, robbery, and drug crimes, according to the Council on Criminal Justice report.

In 27 major cities across the United States, reports of most types of offenses — burglary, larceny, robbery and drug crimes — dropped significantly, compared to the cities’ combined average crime rate of the past three years.

Richard Rosenfeld, a criminology professor at the University of Missouri St. Louis who authored the report, said that cities’ shutdowns beginning in March largely drove the decreases this summer. More people staying at home meant fewer houses were broken into; fewer people going out at night meant fewer opportunities for theft and robbery, for example.

But for some of the most violent crimes, such as shootings, aggravated assault, and murders, the number of incidents in the cities we examined have increased in the pandemic. Compared with a three-year average between 2017 and 2019, homicides increased 25 percent between April and June.

Data included in the Council on Criminal Justice’s report stops at the end of June, and doesn’t include cities like Portland, Oregon, and Kenosha, Wisconsin, where protest tensions rose and shootings occurred, by a counterprotester and a vigilante, respectively, in August. Or in Louisville, Kentucky, where two police officers were shot on Wednesday following a grand jury’s decision not to charge any officers for killing Breonna Taylor. That said, some reports show violent crime continued at elevated rates in July and August and property crime rates have gone down.

Reports of homicides, aggravated assaults and shootings, which are rarer but more devastating to the victims and their families, rose in 2020 compared to the past three years.

David Abrams, a law and public policy professor at the University of Pennsylvania, has examined major cities’ public crime data since the beginning of the pandemic. He publishes real-time crime trends on City Crime Stats, an online data portal that allows viewers to explore how specific types of crime changed in each city.

While the data portal shows similar trends in upticks of murder and decreases in other crimes, pinpointing the exact factors that drive up murders is much more complicated than understanding what caused the decrease in crimes like burglaries, Abrams said.

One of the main reasons: The motivation behind burglaries or larceny is often money, whereas the motivation behind murders and shootings is more varied, he said.

Many factors might play into these increases: A 60 percent surge in gun purchases can be followed by more shootings; trapping domestic violence survivors and abusers under the same roof during the quarantine may cause more assaults and murders; and Covid-19 has made police outreach work even more difficult. The pandemic has also turned families and support systems upside down — unemployment is high, schools and many summer programs have closed, and people, especially from low-income communities and communities of color, have faced illness and death in their families from Covid-19, making routines and structures impossible to maintain.

Dorothy Johnson-Speight, a community organizer in Philadelphia, said she is especially troubled by how many shootings and violent crimes involved young people this summer.

She noted that not only have schools closed, but so have most youth programs that can give young people a sense of structure and belonging. Johnson-Speight, who founded the violence prevention group Mothers in Charge after her son was killed in 2001 over a parking dispute, believes many of the shootings in Philadelphia this year involved people who are under the age of 18, though official police figures are not available. A recent example was a 16-year-old shot dead on September 21, with an 18-year-old and a 12-year-old shot on the same day.

“The anxiety and pain and grief are on steroids because of what’s happening with Covid,” Johnson-Speight said. “People have no way of seeing things getting better, and there is nothing at the end of the tunnel. What I hear from parents that lost one or two or three children is, ‘What’s going to happen next? Will my other children suffer the same thing?’”

While the pandemic brings much uncertainty, there is one thing that may lead to a drop in crime: the weather. Historical trends show that the violent crime rate often increases in the summer, reaches its peak in the fall, and drops to the lowest point in winter — as temperatures decrease and people retreat indoors again.

Crime increased after protests against police violence … briefly

Following the police killing of George Floyd, a 46-year-old Black man, in Minneapolis in May, protests against police violence and systemic racism quickly spread across the country, from major cities to historically conservative, majority-white towns — more so perhaps than any civil rights protests in the nation’s history. However, with the protests came news coverage focused on riots, lootings, and scenes of chaos, despite an estimated 93 percent of protests being peaceful.

President Donald Trump has said little about the police violence against George Floyd, Breonna Taylor, Jacob Blake, and other Black Americans, but has spoken consistently of “law and order.” In July, with Black Lives Matter protests still happening in major cities, Trump sent in federal law enforcement agents to nine cities led by Democratic mayors to stop what the president called “shootings, killings, murders and heinous crimes of violence,” whether or not any of those things were happening in those places.

“This bloodshed must end,” Trump said during official remarks in July. “This bloodshed will end.”

The implication was that the protests had caused the rise in violence, or “bloodshed” — but was that true?

The nationwide protests kicked off in late May, when homicides remained low. There was an increase in mid-June, but the Council on Criminal Justice’s data does not break down where the murders happened in each city, which makes it difficult to analyze protests’ direct impact on violent crime.

What is known is that Black Lives Matter demonstrations have been mostly peaceful. Researchers at the Armed Conflict Location and Event Data Project analyzed more than 7,750 demonstrations from 2,400 locations between May and August, and found that less than 7 percent of the protest were violent, which the researchers define as where “demonstrators themselves engage in violently disruptive and/or destructive acts targeting other individuals, property, businesses, other rioting groups or armed actors.” This can range from vandalism and looting to clashing with the police, a much wider net than police’s definition of “violent crime,” which include rape and sexual assault, robbery, assault, and murder.

If anything, aggressive and militarized government response has made demonstrations more violent, researchers concluded. For example, before Trump deployed the federal task force to Portland, Oregon, 17 percent of the demonstrations were violent; after federal law enforcement agents entered Portland, the share of violent demonstrations more than doubled, to 42 percent. Criminologists have warned that sending in federal law enforcement officers, like border patrol agents or Bureau of Prisons guards, with no training or knowledge on local issues can do more harm than good.

Another unintended consequence of escalating federal involvement in policing protests is that it hinders people’s trust in the police. Even before this summer, victims of violent crime said some of the most common reasons that stopped them from going to the police were they “dealt with it another way,” “fear of reprisal or getting offender in trouble,” and “police would not or could not help.” An increasing distrust in police may lead to more vigilantism and more unreported crimes.

Also, violent crimes are rare enough that small changes in absolute numbers can lead to large statistical swings, and that’s especially true for the most serious kind of violent crimes like murders.

For example, homicides in 20 cities tracked in Rosenfeld’s report increased by more than 50 percent around the last week of June, which is an alarming trend compared to the past three years. However, looking at the raw numbers, homicides increased from roughly 70 homicides per week to 101 per week, or fewer than one additional death in each city every day. Most of the increase took place in Chicago.

And then there is another historical trend: While the trauma and loss that accompany each murder cannot be measured by numbers, the level of violence in American cities does not come close to the level of violence during the 1990s, where nearly every 30 in 100,000 people were killed. In recent years, it’s been about 10 in 100,000.

The police killing of an unarmed Black man in Minneapolis on May 25 sparked protests across the country. Compared with the past three years, in those first days of unrest through June, reports of commercial burglaries surged before returning to normal levels just as quickly.

In all, criminologists say it’s difficult to draw any conclusions between protests and violent crimes — especially during a time when the US coronavirus death toll surpassed 100,000, the country was experiencing an unprecedented level of unemployment, and coronavirus-related precautions restricted police’s ability to solve crimes.

That said, some more common crimes associated with protests, such as burglary, can perhaps shed more insight on the impact of protests on crime. Commercial burglary — or breaking into a business establishment — is typically associated with what is commonly called looting. Among all types of crimes tracked in the Council on Criminal Justice report, commercial burglary had the most significant spike in the beginning of June, when police violence protests began to spread.

Within one week, the number of commercial burglaries in major US cities jumped from nearly 5,000 to almost 10,000. But the number of incidents dropped just as quickly in the following week, back to below-normal levels.

Compared with the past three years, most cities saw a brief spike in commercial burglary reports through June, shortly after protests against police violence broke out. In some cities, such as Denver, San Francisco and Chicago, commercial burglaries were already on the rise before protests began.

The evidence suggests that significant looting was confined to the first wave of protests. But there could be another explanation: Active police enforcement — or an emphasis on enforcing specific crimes — can swing crime rates up and down.

Crime trends are affected by police enforcement

Something to know about crime trends: They are shaped by police action and inaction. Crime trends reflect crime reports collected by law enforcement agencies. Crime reports are created when law enforcement responds to calls or uses tactics such as traffic stops or stop-and-frisk.

While the Supreme Court ruled that it’s illegal to stop and frisk someone simply for living in a “high crime area,” research still shows people in predominantly Black and Hispanic neighborhoods are searched a lot more frequently. Even though most people who are stopped are innocent, their interactions with the police can have lasting effects, including feeling discouraged to report a crime to the police themselves.

New York City is a good example of the power of police-initiated actions. When the city began to shut down in April, the number of drug crimes plummeted. Then it began to steadily increase through April and May, as people emerged from lockdown and police officers began patrolling again, getting close to pre-pandemic levels. And when the protests sparked by Floyd’s death spread across the city in late May and early June, the number of drug crimes again dropped overnight.

Reports of drug crimes in New York City had large swings this year compared with trends in the past three years, which experts and public defenders believe is a direct reflection of circumstances.

It’s unlikely that drug crime data represents how the number of people consuming and selling drugs changed over this summer, said Alice Fontier, managing director of the Neighborhood Defender Service of Harlem, a public defender’s office.

What the data shows, Fontier said, is how the New York Police Department deployed its officers throughout the summer. When the pandemic first hit, the department was pulling back on drug searches, partially because many officers were under quarantine. Their practice began to return to normal until protests against police violence broke out, when many of the department’s officers shifted to crowd control instead, Fontier said.

NYPD did not respond to multiple inquiries by The Marshall Project, but during an interview with the Police Executive Research Forum, NYPD Commissioner Dermot Shea said he found the narrative of police pulling back because of protests “offensive.”

New York’s trend in drug crimes is similar to what the data shows in many other cities, including Chicago, Philadelphia, and Memphis.

Many cities, such as Boston, Louisville and Memphis, demonstrated similar patterns in drug crimes as seen in New York City: a swift drop when the pandemic started, followed by a steady increase in April and May, only to drop again after the protests.

When a significant number of officers are under quarantine for Covid-19, or when police departments shift resources from making drug busts to responding to protests in riot gear, crime trends change accordingly.

“What we see and experience over time is that the number of drug arrests is directly correlated to the amount of focus and resources the NYPD puts into these cases,” Fontier told The Marshall Project.

Releasing low-risk people from jails and prisons didn’t drive up crime rates

As Covid-19 began to spread across the country in April, jails and prison soon became hot spots for the outbreak. It didn’t come as a surprise. Overcrowding in prison and jails means some facilities have people sleeping on the ground, and in most facilities, even basic Centers for Disease Control and Prevention guidelines such as hand-washing with soap or covering your mouth when you sneeze are virtually impossible to follow.

At the beginning of the pandemic, some jails moved to cut down their populations, releasing people who were incarcerated for pretrial detention or who were almost finished with their misdemeanor sentences. And some prisons, which incarcerate people who are convicted, followed suit.

Public backlash came just as quickly. Some victims of crimes were upset about the early releases, and police departments claimed that coronavirus-related jail releases drove the spike in violent crime.

Data contradicts this narrative. A recent study by the American Civil Liberties Union shows that in 28 major US cities that saw a decrease in jail population between March and May, all but one (Denver) also saw decreases in the most serious type of crimes this summer.

At the beginning of the pandemic, some cities began to release pretrial detainees from jails to decrease their risk of getting the coronavirus. While critics warned these mass releases would lead to a spike in crime, an ACLU report found that most cities with a decrease in jail population also saw fewer reported crimes overall from March through May, compared to the previous year.

At the beginning of the pandemic, San Francisco District Attorney Chesa Boudin said he heard warnings that releasing people from jail, or arresting fewer people, would lead to more crimes, and that the price of keeping jail inmates safe from Covid-19 is too high.

Yet there was no crime surge. Between March and May, San Francisco’s jail population dropped by more than 40 percent. Its crime rate also dropped sharply compared to the same period in 2019. Both trends, Boudin said, were “unprecedented.”

“If fewer people are incarcerated, then more people will be able to keep steady jobs, safe housing, and get the mental health help they need,” Boudin said. “That all leads to fewer crimes.”

In Denver, the only city that saw an increase in crime and the largest decrease in jail population (by almost 800 people), the trend is short term, and it’s hard to read too much into the numbers, Denver Police Chief Paul Pazen told The Marshall Project.

While virtually all types of crimes have gone down in Denver, Pazen said commercial burglaries drove up the crime rates — as businesses closed during the pandemic, his department saw commercial burglaries more than double this summer. More than 60 percent of the stores were broken into by people who are homeless, Pazen said.

It’s too early to tell if “defund” efforts have impacted crime rates

After the death of George Floyd, “defunding” the police, or moving money from police spending to social services, became central to the police reform conversation.

A few cities have started the defunding process, but it’s too early to affect recent trends. For example, Minneapolis City Council members vowed to disband the police department following Floyd’s death, but their effort is facing major setbacks. In New York City, the nearly $1 billion cut to its police budget took effect on July 1, after murders and shootings were already rising in the city.

How “defund” policies affect crime remains to be seen.

What is clear is the coronavirus is likely to cause the first major drop in police spending in decades, spending that has increased from $220 to $280 per resident from 2000 to 2017, even when violent crime decreased by more than 20 percent during the same time.

Even though as crime has trended down, police spending has consistently grown for decades.

Nearly half of 258 police chiefs and sheriffs who responded to a recent survey said they are expecting or already receiving budget cuts in the coming year, according to the Police Executive Research Forum. Most of the cuts range between 5 and 10 percent.

Many police chiefs who responded to the survey warn that unintended consequences may come out of the budget cuts. Hiring freezes, for example, will mean fewer patrols, longer response time, and less proactive actions from the police department. The domino effect, they warn, will eventually lead to a spike in the crime rate.

Richard Auxier, a senior policy associate in the Urban-Brookings Tax Policy Center, said payroll costs often take up 60 to 70 percent of the police budget, meaning that things like hiring freezes, pay cuts, and layoffs are likely first steps.

But it’s too early to say how those budget cuts will affect crime rates. And even if more policing leads to less crime, activists warn that it carries collateral consequences, such as more arrests and the general harassment of minority communities, that other approaches don’t have. The more important part of the “defund the police” conversation should be about how we should spend the money instead, Auxier said.

For example, a 2018 study shows that one-quarter of people who died in police shootings showed signs of mental illness, and the recent police suffocation of Daniel Prude has reignited talk about how mental health professionals are better suited to handle these interactions than police.

Alternative programs are not new, and they’ve been proven to create a safer community. In Eugene, Oregon, a 30-year-old program has been successful at reducing police interactions with people who are in crisis, dispatching medics and mental health professionals to respond to 911 calls that are not about crime — like mental illness, homelessness, or addiction. In 2019, they responded to 20 percent of all 911 calls in the town, costing a fraction of the price of traditional police interventions. Cities like Olympia, Washington, and Denver have also adopted similar programs.

The way we see crime is politicized and influenced by news sources

So is violent crime out of control? That can depend on whom you ask — and which cable news station they watch.

For example, this summer, Fox News has spent more time covering violent crime than CNN and MSNBC combined, according to an analysis of data compiled by the Stanford Cable TV News Analyzer.

As protests for racial justice began to spread across the country this summer, a data analysis of major cable news network transcripts initially showed a spike in coverage of crime and shootings. Then, as protests proved to be largely peaceful, and MSNBC started spending less airtime on these topics, while Fox News continued with its coverage.

Since the police killing of George Floyd, Fox News has leaned into a narrative of looting and property destruction, filling its segments with headlines like “Portland Plagued by Violent Clashes, Riots” and “Businesses Experience Worst Looting in Decades.”

While CNN and MSNBC’s coverage of violence and crime also spiked after the Floyd protests took off in May, it has dropped significantly since then.

In the 2000s, cable and local TV news became more popular, contributing to a shift in public opinion on crime. Before the early 2000s, more and more people believed there were fewer crimes in the United States, according to Gallup polling data, which matched the truth — that crime rates were decreasing. However, that trend was completely reversed in 2001, and not much has changed since: As crime continues to decrease, more people believe the opposite is true — that crime is up.

Beginning in the early 2000s, more people began to believe there are more crimes happening in the country, despite data showing otherwise.

Dan Romer, research director of the Annenberg Public Policy Center, said the rise of police television shows, like NCIS and CSI, and how much airtime local TV news gives to violent crime has fed the discrepancy.

Romer, who studies media and its social impact, said producers at local TV news stations face daily pressure to fill the evening report with different beats, like sports, local government, news, and crime — and the idea is to capture viewers’ attention.

“No matter what is going on, there’s going to be a crime in the news region of the news station,” Romer said. “It can be hit-and-run, it can be shooting — the crime news hole stays consistent over time. Stations get that’s an attention-getter. The crime rates could be changing dramatically, but they wouldn’t know it.”

Bias in reporting and story selection can also plague how crime is portrayed in local TV news, Romer said. For example, there has historically been emphasis on stories where the suspect is Black and the victim is white, even though Black men are more likely to be victims of violent crimes. This sways public opinion, too.

“People talked about media literacy and teaching it to children,” Romer said. “People need to know even though they see a lot of violence on local news or hero movies, it’s not necessarily what the world is like.”

This extends to how politicians paint America. Americans disagree on a lot of things, but a recent poll by Monmouth University shows that Republicans, Democrats, and independent voters all agree that maintaining law and order is a major problem in the country right now. What they disagree about is the root cause of the problem, let alone who is best positioned to solve the problem.

For example, while 24 percent of people believe the actions of protesters are fully justified, just as many people believe they are not justified at all. The split on whether Trump or Biden can solve the problem is similarly even. The disagreements often fall along party lines, which may also be influenced by where people get their news.

Abrams says the news — as well as politicians — won’t give you the full story when it comes to crime, though. Parsing data is more than just reporting the numbers.

“If there is a bad weekend with a lot of shootings, people want to know what happened, and rightfully so,” Abrams said. “But to really understand how crime has changed, let’s look at the week, the month, the year, the decade. Crime has gone way, way down from the peaks in the ’80s and ’90s. Even the highest spikes in a few cities over the summer are small blips in comparison.”

Earlier this week, the Commission on Presidential Debates announced six topics moderator Chris Wallace has selected for the first debate on September 29, including “Race and Violence in Our Cities.” This framing, that the two are interlinked, is the problematic narrative that Romer warned about. It is also a near guarantee that racial justice protests and the violent crime streak this summer will be focal points of the debate. Understanding the nuance and context of crime rates is crucial for evaluating each candidate’s story of what the unrest and division in this country is really about.

This article was published in partnership with the Marshall Project, a nonprofit news organization covering the US criminal justice system.

Future Perfect

Why you’re not getting a delta-specific booster yet

Explainers

Murders are spiking. Police should be part of the solution.

Explainers

We can’t just run away from wildfires

View all stories in Explainers