Police are solving a lot fewer murder cases than they used to

Murder rates have fallen by half over the past 20 years. So why are many murders still going unsolved?

The homicide rates in many US cities have jumped sharply in 2015: an average increase of nearly 15 percent among 25 of the nation's 30 biggest cities. That's inspired a lot of head-scratching and hand-wringing about what might have caused the spike — and in particular, whether distrust between police and the communities they're supposed to protect has made it easier for criminals to operate with impunity.

This debate isn't necessarily premature, but it's definitely too early to know for sure what's going on. We won't even know if this year's homicide spike is a statistical blip or the beginning of an enduring trend for another few years. And even if the trend continues, we might not know why homicides are going up for more than a few years — after all, we still don't know exactly why violent crime has been going down for the past 25 years, and we've had since the late 1990s to figure that out.

But it's impossible to understand homicide and policing in America if you don't understand this: A lot of murders in America don't get solved. Typically, about three in five murder cases end in an arrest. Despite significantly fewer violent crimes — including murder — than there were a quarter century ago, the percentage of violent crimes getting solved hasn't really changed.

Here's a reminder of how much violent crime has dropped in the US since the early 1990s:

During that period, as you can tell from the later portions of this chart, murder rates have similarly dropped:

But here's what's happened to the percentage of crimes that result in someone's arrest over that time period:

Police solved more murders in 1995 than even happened in 2014

As you can see from the chart above, murder cases are still way more likely to end in an arrest than other forms of violent crime. (And that goes double for other forms of crime, like burglary.) So it's not exactly like police officers are assiduously tracking down every petty thief in town while letting murderers roam free.

But since police departments have about half as many murders to solve as they did in the early 1990s, you would think they would have been able to make arrests in more of the murders that do happen. After all, according to these statistics, police solved more murders in 1995 than even happened in 2014. Instead, police have only been able to solve about 60 to 65 percent of the murders they find out about — regardless of how many murders that is — over the past two decades.

This is especially weird given how much attention gets paid to homicide rates — especially in major cities. Indeed, the debate over the homicide spike has been going on since the middle of this year — before the final statistics for 2015 were even in. And while, on the whole, major cities aren't that much worse at solving murders than other police departments are, some cities' clearance rates are extremely low.

In Baltimore, for example, only 45 percent of murder cases ended in an arrest in 2014 (and in 2015, it's looking more like 31 percent). That actually means the Baltimore Police Department has gotten a lot worse at solving murder cases than it was in the early 1990s — 20 years ago, its clearance rate was 70 percent.

Solving murder cases requires community trust

Just like homicide rates themselves, homicide clearance rates are extremely complicated, and it's irresponsible to say that there is one major reason why they haven't gone up over the past 20 years.

But at the level of trying to solve each individual murder case, homicide detectives often have the same complaint: They know who did it, or they know that people in the community know who did it, but they can't get anyone to talk to police or testify in court.

A multi-part investigation in the Baltimore Sun at the end of 2015 followed homicide detectives trying to solve a single case. The difficulty in getting witnesses to come forward was first and foremost — and police officers who spoke to the Sun attributed it to distrust between residents and police:

An anonymous tipster had dialed 911 to tell police where the gunman had ditched the weapon.

Sgt. Sean Jones, the detectives’ supervisor, counted it a small victory.

"That tells me we have at least one citizen that gives a s——."

A couple of decades ago — the last time the city saw so much killing — Baltimore’s homicide unit closed more than 70 percent of its cases. Veterans talk of returning to the office from a crime scene to find a fistful of tips waiting for them.

But the widening gulf between police and the community since then has made witness cooperation a rarity.

This isn't just about intangible feelings; there's evidence that the factors that make it easier to solve a homicide are tied to community trust. A 2000 study commissioned by the National Criminal Justice Reference Service found that among the most important factors in predicting whether a homicide would be cleared were "valuable information" provided by a witness at the crime scene, and subsequent detective interviews of friends, neighbors, and acquaintances of the victim as well as other witnesses. It's harder to do that when the people you're supposed to be interviewing don't trust you.

The Baltimore police blame the distrust on the killing of Freddie Gray in police custody earlier this year, which was followed by protests and unrest (and, at the end of 2015, by the beginning of the high-profile trials of the officers involved in Gray's death). On the face of it, that sounds a lot like the "Ferguson effect": the idea that crime is up because police officers are afraid to do their jobs in the face of public scrutiny. But the notion that witnesses are unwilling to come forward because they don't trust police isn't new. Jill Leovy makes this point in her 2015 book Ghettoside, which is drawn from her reporting on Los Angeles's homicide department over the past decade.

In Leovy's telling, residents are partly hostile to the police because they think police are overly aggressive. But they're partly hostile because they don't trust police to protect them. The two are combined: the sentiment that police are more interested in making petty "broken windows" type busts than in solving homicides.

Frankly, the fact that homicide clearance rates have been flat, even though police have only had half as many murders to solve, would seem to prove those residents right. In reality, of course, it's a vicious cycle: When residents don't trust police, homicides can't get solved; when police don't solve homicides, residents don't trust police.

The "Ferguson effect" is often talked about in terms of blame — in particular, blaming citizens for having insufficient respect for police. But when the problem runs this deep, "who started it?" isn't a particularly useful question. The fact of the matter is a lot of people who kill other people in America don't get arrested for it — and that is interconnected to deeper problems between police and the communities they're supposed to serve.