Donald Trump talks a lot about murder rates. He is wrong that homicides are the highest in 47 years; in fact, the national rate in America is still near the historic lows of 2014. But in the past two years, the rate has begun rising. The increase has been led by a handful of cities — Baltimore, Chicago, St Louis, Milwaukee, Birmingham and Memphis — where murder is near record highs. In St Louis, the rate is now more than 60 per 100,000 people, twice the level of notoriously violent South Africa.

If that trend persists, the United States will be in trouble. It needs to be nipped in the bud. But how? Most people who study the epic decline in crime that took place in the 1990s have no definitive idea why it happened. Mass incarceration probably had a fairly small effect. The theory that legalised abortion was behind the decline, made famous by the book Freakonomics, was probably incorrect. A third explanation has to do with the rise and fall of drug epidemics, which fuel violence between gangs and other narcotics sellers. And a fourth idea is that reductions in lead pollution reduced violent tendencies.

The answer may be found in New York and Los Angeles, two large cities once famous for their high crime rates, but which are now among the nation’s safest. Some credit the so-called broken windows policing strategy, where police target minor crimes to create an atmosphere of public order. Others believe that it was the elimination of open-air drug markets that did the trick.

The explanations are endless. But I want to suggest one that gets little attention. This is the leviathan theory of violence, as popularised by psychologist Steve Pinker in his 2011 book The Better Angels of Our Nature: Why Violence Has Declined.

The word “leviathan” hearkens back to the book of that title by the 17th-century philosopher Thomas Hobbes, who theorised that government’s monopoly on the legitimate use of force is necessary to pacify an inherently violent state of nature. Pinker adds insights from modern psychology and game theory to tell a story where violence is the result of a bad equilibrium created by the lack of good government.

Pinker posits that in an anarchic society, everyone has to defend himself and his family on his own. The only way to deter attacks is to establish a reputation for aggression and willingness to resort to violence. This can be done only by inflicting violence on other people — a bunch of tough guys run around assaulting and killing people to show how brutal they are, keeping themselves and their families and friends safe from opportunistic attacks. But that pre-emptive violence just adds to the overall level of danger in society, forcing other people to respond in kind and establishing an equilibrium where everyone is violent.

According to Pinker, government can break this cycle. By protecting people from attacks, the police and courts allow normal citizens to stop being pre-emptively violent, which leads to further drops in violence. The bad equilibrium of anarchy gives way to a good equilibrium where people are peaceful and the Government does not have to do anywhere near as much policing.

This theory fits our anecdotal experience of violent people. Gangsters are generally obsessed with reputation and respect. Honour culture is a well-established phenomenon in the psychology literature. Far from being the super-predators of which Hillary Clinton once warned, most violent people tend to be super-defenders — young men struggling in vain to carve out a sphere of safety in a world of anarchy, but in the end simply contributing to that anarchy.

This theory could help to explain why murder, but not property crime, has risen in the past two years. A rash of highly publicised police killings, starting with Michael Brown in Ferguson, Missouri, in 2014, has made its way around the news and social media.

Those high-profile killings, which many attribute to police racism, have probably decreased the black community’s trust of the police, especially in the cities where the deaths occurred. Police shootings have been shown to lead to fewer 911 calls in predominantly black neighbourhoods. After all, why call the cops when the cops may just shoot an innocent bystander — or even the person who made the call?

There is also circumstantial evidence that some police spend less effort solving murders when the victims are black.

The policy implication is clear: police departments in cities such as St Louis and Baltimore with high murder rates should make a greater effort to protect vulnerable residents, especially black people, from violence. Community policing strategies, where police attempt to integrate into and befriend a local community, would probably help.

Programmes such as Operation Ceasefire, which try to identify and protect potential victims before they are attacked, should be expanded. Cities should hire more detectives to solve murders, especially in poor and black neighbourhoods. And police forces involved in high-profile shootings of black citizens should make open and honest efforts to show the community that they are not infected by racist attitudes, as well as bringing culpable officers to justice.

If the US is going to stem the rise in murder, it is going to have to be proactive and smart. Convincing people that the Government is there to protect instead of persecute them is a good place to start.

• Noah Smith, a former assistant professor of finance at Stony Brook University, is a Bloomberg View columnist