Murder, which grew rarer for 20 years, is on the rise again. But by how much? In 2015, the number of murders increased by 11% nationwide. During 2016, an escalation of gang violence in Chicago left 764 people dead in a city where 485 had been killed a year before. A dispute ensued over whether the Windy City was simply an isolated example or a barometer of a wider problem. National statistics for 2016 will not be released for eight months, but to get an early sense of the answer The Economist has gathered murder statistics for 2016 for the 50 cities with the most murders. These places contain 15% of the country’s population and around 36% of murder victims. Our numbers show that, in 2016, murders increased in 34 of the cities we tracked. Three cities experienced a spike in deaths sharper than the 58% suffered by Chicago. Since cities tend to reflect the country as a whole, this suggests that the murder rate is rising at its fastest pace since the early 1970s.
Today’s violence needs to be set in context. Despite the recent uptick, the murder rate in our 50 cities was lower in 2016 than it was in 2007, and for the 26 years before that. Criminologists disagree about why murder became less common. What they do agree on is that the improvement has been uneven. Newark, just ten miles from New York city, has a murder rate that is nine times higher than its neighbour’s. And unlike New York, where murder is at just 15% of its 1990 peak, in Newark the rate has barely budged.
After the riots in Ferguson, Missouri, in 2014, Heather Mac Donald, author of “The War on Cops”, offered a simple explanation for the rise in murder. The riots were a response to the killing of Michael Brown, a black man, by a police officer. The “Ferguson effect”, Mrs Mac Donald argued, occurred when police officers retreat from cities when relations with the people they serve became bitter, causing crime to go up nationwide. Murders and shootings did increase by 57% in St Louis, a city close to Ferguson, in the two years after Brown’s death. Similarly, when Freddie Gray died in the custody of Baltimore police in April 2015, murders and shootings in the city increased 70% during the year that followed. But we find little evidence for a broader “Ferguson effect” in the rest of the country. Among our 50 cities, data show that in the four months immediately after Brown’s death there was no change in the arrest rate for murder, and just a five percentage-point fall in the arrest rate for gun assaults. This does not look like a widespread retreat by the country’s police forces.
A stronger message from the individual murder records from the FBI for 50 cities is that the quality of police work, the availability of (usually illegal) guns and the chances of getting caught all matter a lot. This is partly because the motivation for murder is changing. Gang-related killings have steadily increased over the past 35 years, from just one in 100 murders in 1980 to nearly one in ten in 2015. Drug-related murders—which are likely to have some gang-related element—have increased in the past two years, after falling for two decades. In 2015 they accounted for one in 25 murders in big cities.
When murder rose in the late 1980s, the age of both victims and perpetrators dropped by some three years. Unlike then, the recent spate of killings has seen the age of both victims and offenders continue to rise. The average killer in 2015 was four years older than in the 1990s. Gang-related murderers are, on average, six years older now than they were then.
This is where variation in law enforcement comes in. Gang murders involving guns are particularly hard to solve. After 20 years of stability, the murder-clearance rate—where a murder is solved because an arrest is made—fell suddenly in 2013 from 60% to 55% in 2015. There are stark differences by race. In 1980, 56% of murder victims in our dataset were black. In 2015, 68% were. In the early 1980s, police solved around 65% of murders regardless of the race or sex of the victim. Among black women and white men, that percentage has changed little. Among black men it fell to under 55% in 2012, and has since dropped to just 47% in 2015. People are more likely to kill if they think they will not get caught, and unsolved killings can set off a cycle of revenge.
Largely thanks to DNA evidence, police are increasingly capable of solving murders when the victim is attacked with hands, bats or knives. In these cases, clearance rates have increased from 70% to 78% in the past dozen years. Against this trend, when the victim is killed by a gunshot, a suspect is arrested just half the time. In the 1980s the arrest rate for gun-related murders was higher, at 65%.
Taken together our evidence suggests that police should focus their efforts on tackling gang-related murders where a black man is killed with a gun. Bill Bratton, who has led police forces in Boston, Los Angeles and New York, likens the policing of cities to a doctor treating a patient. While some cities may only require a check-up and a few sessions of therapy once in a while, others need invasive surgery. Among our 50 cities, gun use has increased from 65% to 80% of all murders since 1980. But that number varies enormously by city. Guns were responsible for 60% of murders in New York and 85% of those in Chicago between 2010 and 2015. Whereas New York and Chicago have made similar rates of progress in reducing murders in which a gun is not used, Chicago’s gun-murder rate is five times New York’s.