Miranda Luby was only 36 years old when her tinnitus arrived in the middle of the night, without warning and for no clear reason

One night two years ago, I was lying in bed when I began to hear a high-pitched, electronic hum. The scene that followed would be comical if it wasn’t so traumatic. I roamed around the house in my PJs, frantically pressing my ears against various electrical appliances. It took half an hour to realise the sound was coming from inside my own head.

I’d heard ringing in my ears before, rising then falling, like a one-note crescendo of an orchestra. As a kid, I used to believe the myth it was the sound of dying ear cells, playing a swan song for a frequency I’d never hear again. This was different. This time, it didn’t stop.

Dozens of cities have canceled their contracts with Flock. Some are switching to other companies, but privacy experts say they're no better than Flock.

(Image credit: Chip Somodevilla)

A few things are certain in life: death, taxes, LeBron James playing professional basketball — and Americans’ belief that crime is getting worse. Since Gallup began asking about national crime in 1989, a majority of respondents said they thought crime in the US was higher than the year before in every survey year except 2000, 2001, and 2025. Even in that last year, 49 percent still said more crime was happening. It’s perhaps the archetypal example of negativity bias, the mental habit of giving more weight to bad news than good.

Last week, the FBI put some numbers on just how big that bias is. The American murder rate — which peaked in 1980 at 10.2 per 100,000 — fell to an estimated 4.1 per 100,000 people in 2025, matching the rates recorded in 1955 and 1956 for the lowest ever measured in the country. The violent crime rate (murder, rape, robbery and aggravated assault) fell 9.7 percent, the largest single drop since J. Edgar Hoover was running the bureau in 1936. 

That much is known. But what’s not known for sure about the crime drop is almost everything else. Adam Gelb, who runs the Council on Criminal Justice, says that there is no single or easy explanation for the historic lows, while Jeff Asher, the analyst who spotted this decline before nearly everyone else, agrees. For every leading candidate — federal pandemic spending, some normalization after the 2020 social collapse, better violence prevention, even the decline of drinking — there is a corresponding hole. 

But there is one American city where researchers can point to a specific program, measure what it did, and put real numbers on it. And it just happens to be the city that spent 50 years — and five award-winning seasons on HBO — as America’s shorthand for murder.   

What Baltimore did

Baltimore recorded more than 300 homicides a year for most of the last decade, and more than 200 for most of the past half-century. In one month, July 2015, 45 people were killed — an astounding figure for a city of just around 600,000 people at the time. 

But Baltimore finished 2025 with 133 homicides, the fewest since 1977 and about 60 percent fewer than in 2020. Through the first eight months of this year, the city has seen 60 murders, and the police commissioner’s target is to stay under 100. 

What caused this? Three things at once.

Baltimore Mayor Brandon Scott’s Group Violence Reduction Strategy identifies the few people most likely to shoot or be shot — two groups that heavily overlap — and sends people to their doors with two messages: you are on our radar, and here is housing and job training. Separately, the longer-running Safe Streets program puts outreach workers with real street credibility between disputes, aiming to interrupt them before they become shootings.

That’s the harm reduction element. There’s also a tough-on-crime component. The state’s attorney office says it has secured 2,129 convictions since 2023 across several gun-related violent-crime categories, with over 1,700 of those defendants incarcerated. 

Meanwhile, Baltimore’s homicide clearance rate reached 60 percent by mid-2026. Its nonfatal-shooting clearance rate was 54 percent, up more than 12 percentage points from a year earlier. Criminologists have long held that the certainty of being caught deters criminals more than the severity of the sentence, and more successful shooting investigations are a common thread where violence is falling around the country. 

How do you tell whether a program worked, or whether a city just got lucky? Drug trials get a control group; cities don’t. Baltimore got one almost by accident: the strategy launched in the city’s Western District in 2022 and nowhere else, which left the rest of the city as a comparison.

That’s what a team of economists and criminologists used. They measured the Western District against Baltimore neighborhoods that hadn’t started the program yet, and over 18 months the number of people shot there, fatally or not, ran about 30 percent below that stand-in. Killings alone fell by a similar share, but deaths are rare enough that chance couldn’t be ruled out. Police made no more arrests than before. Arrests for serious violent crime rose 81 percent.

A second, looser test set the whole city against a composite of other big American cities and put Baltimore’s murder rate about 25 percent below it. The authors are careful with that one: too much changed in Baltimore at once to say how much it the strategy was responsible for.

Murder is contagious

Baltimore’s success is built on two ideas, and they came from two very different places. 

The first came from a doctor named Gary Slutkin, who spent the 1980s and early 1990s fighting tuberculosis in San Francisco, cholera in Somali refugee camps, and AIDS across Africa. He came home to Chicago in 1995, a year the city recorded over 800 homicides. 

When Slutkin examined the city’s shooting maps, they looked familiar. Clustered in space, clustered in time, and spreading case to case — murder resembled the outbreaks he used to fight. Slutkin reasoned that the way to treat violence was to treat it like cholera: find the people most likely to transmit. Interrupt the transmission. Change the conditions that let it spread. Slutkin built the model into a program called CeaseFire, now Cure Violence Global, and its Baltimore descendant is Safe Streets.  

The second idea came from a criminologist. David Kennedy, working in Boston in the 1990s, found that 61 gangs with about 1,300 members — under 1 percent of the city’s young people — were tied to at least 60 percent of its youth homicides. Get those people in a room with police, prosecutors, and community elders, tell them plainly what comes next, and the shooting slows down.

That approach is called focused deterrence. Baltimore’s Group Violence Reduction Strategy is a version of it. To see how it works over a longer period of time, take a trip across the Atlantic to Glasgow in Scotland, which in the early 2000s was considered the murder capital of Europe.

In 2005, in the midst of that wave of violence, Glasgow’s Strathclyde Police established a Violence Reduction Unit that operated with a basic thesis: evidence shows that violence can be prevented. Among other interventions, the unit sent trauma surgeons into schools to describe what a blade does to a face and stationed outreach workers at hospital bedsides to reach victims in the vital hours after an attack, the moment when retaliation can still be stopped. 

In 2024-25, Scotland recorded 45 homicide victims, the fewest since comparable records began in 1976 and down from 96 two decades earlier. Every case was solved.

Pull the handle

So we have clues to why crime has dropped so much, but not clear conclusions. As successful as Baltimore’s group violence strategy has proved, a 2023 study of a similar street-outreach effort in Boston found no effect whatsoever. And the latest evaluation of Baltimore’s own Safe Streets — 11 sites, each measured against a stand-in version of the same neighborhood without the program — produced numbers that sound terrific but settle very little: youth homicides down 42 percent and youth shootings down 21 percent, but neither is strong enough to rule out chance. And while some sites saw youth homicide fall by as much as 100 percent, two sites saw it rise by 42 and 89 percent. 

Some experts, like the UC Irvine criminologist Emily Owens, argue the crime decline has been so widespread and similar — big cities and small, red and blue governments — that the true cause is some larger social change rather than what any one city is doing. Nor can we be sure the fall will continue. Billions in federal pandemic relief money has supported many of the programs that have worked well in places like Baltimore, and that money is already running out. But we can identify one reason why it happened, and learn from that success. 

As it happens, Gary Slutkin was not the first doctor to look at a map of the dead and see a pattern in it. In September 1854, a cholera outbreak was killing hundreds of people in London’s Soho district. A doctor named John Snow had mapped the outbreak and saw the dead were clustered disproportionately close to the district’s Broad Street water pump. Snow had a theory that water was connected to the disease and convinced a Soho parish board to remove the handle of the pump. 

Nobody in Soho at that moment could have told you what caused cholera. It would only be 29 years later that someone would definitely identify the responsible pathogen, and in doing so, explain why Snow’s intervention worked. But the explanation could wait. The dying stopped that month in Soho, just as it is stopping in Baltimore and around the US today. And that’s what counts.

A version of this story originally appeared in the Good News newsletter. Sign up here!