You might call it a new golden age. America’s economy is strong, overdose deaths are falling and crime rates are down. For the second consecutive year murders in America have plummeted. The surge in violence in 2020, which was the deadliest year in over two decades, may now seem like a distant memory to some. Yet for criminologists and policymakers the question of what caused that spike in the first place remains unanswered.

A popular theory, advanced prominently by Heather Mac Donald of the Manhattan Institute, a think-tank, is that it was caused by a “George Floyd effect”. The theory is as follows: after the murder of Floyd by police officers in Minneapolis in 2020, police lost trust in high-crime communities and among African-Americans, leading to lower clearance rates for murders. When people think they will not get caught, they commit more crime. Another version of the Floyd effect thesis holds that police officers, beset by rising public hostility, deliberately pulled back from high-crime neighbourhoods, for fear of being prosecuted for doing their jobs. Either way, protests against police brutality lead directly to more murders, a bitter unintended consequence for the protesters and, perhaps, evidence of the kind of soft liberalism from big-city Democrats that Donald Trump was elected to expunge.

A recent report from the Brookings Institution, a think-tank, advances an alternative theory. Rohit Acharya and Rhett Morris, the report’s authors, argue that the rise in murders began in April 2020, about six weeks before the murder of Mr Floyd. They contend that high unemployment and school closures in poor neighbourhoods, both brought about by covid-19 and the policy response to it, left teenage boys idle. This, not the Floyd effect, was responsible for the murder spike. This would suggest an awful trade-off: those early lockdowns saved many lives, but they also may have resulted in more murders.

Using weekly national homicide data, Messrs Acharya and Morris show that throughout the summer of 2020 murders rose 30% compared with the summer of 2019. Crucially, they do not find an inflection point around the end of May, when Mr Floyd was killed. Across the six weeks preceding his death national weekly murders increased by around 17 murders per week, a rate 70% greater than the same period in 2019. And during the six weeks following his death, murders rose at a similar rate.

What, then, caused this increase? The authors theorise that the economic circumstances of the pandemic are to blame. Criminologists concur that, in general, poverty correlates with crime rates. In Atlanta, 65% of all homicides occur in neighbourhoods where at least 30% of the population lives below the poverty line. Nearly every big American city displays this trend. Poorer neighbourhoods were also disproportionately affected by the pandemic: job losses and high-school dropout rates were far higher. Cities with a greater share of young men living in these conditions saw larger increases in homicides in 2020.

Juveniles typically commit few murders. Though roughly half of murders go unsolved and not all jurisdictions report the age of the murderer, the available data suggest that fewer than 10% of homicides are committed by those under the age of 18. Yet between 2019 and 2020 juveniles accounted for an estimated 15-20% of the overall surge. That seems consistent with the idea that closed schools and idle teenage boys are a big part of the story.

Criminologists tend to be wary of single explanations. “It’s very difficult to come up with a definitive conclusion about what happened in 2020, because so many things changed at the same time”, says Aaron Chalfin, a criminologist at the University of Pennsylvania who was not involved with the Brookings report. He notes that, in the past, unemployment rates have not correlated with murder rates, although that does not necessarily mean no such relationship arose during the pandemic. And teasing out the interactions between variables is trickier still. Were school closures in poorer neighbourhoods responsible for juveniles committing more murders, or was it school closures plus fewer police officers patrolling the streets?

The research, says Neil Gross, a professor of criminology at Colby College (who was also not involved in the study), suggests that the nature of social ties in poor areas matters. Crime is often lower where “people know their neighbours and can look out on the street for errant teenagers and contact their parents”, says Mr Gross. That suggests yet another potential suspect: such neighbourhood watchers were locked down at home. ■

Stay on top of American politics with The US in brief, our daily newsletter with fast analysis of the most important political news, and Checks and Balance, a weekly note from our Lexington columnist that examines the state of American democracy and the issues that matter to voters.