There’s no longer much debate, at least as far as experts are concerned: For two straight years — 2015 and 2016 — the murder rate in the United States went up, significantly. This fact isn’t the result of media hysteria or cherrypicking or the never-ending attempts on the part of some demagogues to churn up fears about violent crime — it is a statistical reality. The basic facts, as summed up in an important National Institute of Justice paper published last year, are that the homicide rate rose from 4.4 to 4.9 homicides per 100,000 Americans from 2014 to 2015, “an 11.4-percent increase and the largest one-year percentage rise in the U.S. homicide rate since 1968.” Then, from 2015 to 2016, there was another, albeit smaller increase — from 4.9 to 5.3 per 100,000, or an increase of 8.2 percent.
But this reality has been misunderstood by a lot of people, partly because it’s difficult to talk about violent crime in anything but a political way — especially given that the homicide rise in question took place as candidate and then president Donald Trump attempted to paint many American cities as blighted death zones — and partly because when it comes to something as complicated as the homicide rate, it takes a lot of time and effort to understand what’s going on. It isn’t unusual for there to be a years-long gap between the discovery of a trend in criminology and anything approaching an adequate explanation of what gave rise to it.
Richard Rosenfeld, the lead author of the NIJ paper and a criminologist at the University of Missouri–St. Louis, is starting to unravel the mystery of America’s recent crime increase, though, and what he’s found so far should cause a lot of people to rethink their views on the subject — and highlights the extent to which white Americans and black Americans are living in very different worlds affected by very different sociological factors. Last week, I spoke with Rosenfeld about criminologists’ ongoing attempts to understand the recent uptick, and three important points jumped out from both our conversation and from the NIJ paper he lead-authored:
Rather than there being a single rise in the homicide rate, there are two: one black and one white.
Most of the conversation about the increasing homicide rate has focused on black perpetrators and victims for a couple reasons. The first is that the increase has been concentrated in a small handful of cities, most notably Baltimore and Chicago, that have large black populations. The second is that some pundits and researchers, most notably Heather Mac Donald of the conservative Manhattan Institute, have pointed to the so-called “Ferguson effect” as an explanation for the rising crime rate. Their thesis is that as a result of protests in Ferguson, Missouri, Baltimore, and elsewhere, police have effectively “stood down” — worried about getting sued or sparking social unrest or otherwise upsetting the delicate relationship between police departments and black communities, these officers aren’t working as aggressively to do the sort of police work that helps keep the murder rate in check.
In reality, though, while the rate of black homicide victimization has gone up, that’s just part of the story — white people, too, are both murdering and getting murdered at strikingly increasing rates as well, as two charts from the paper show:
Rosenfeld and his co-authors explain that increases in the white homicide and homicide-victimization rates are a big part of the story here — “the growth in the non-Hispanic white victimization rate was greater than in any year since the September 11, 2001, terrorist attack.” In fact, according to some calculations Rosenfeld ran for me, the increase in the white murder rate drove a large chunk of the overall increase. “If white homicides had not increased from 2014 to 2016, the 2016 homicide rate would have dropped from 5.3 per 100,000 population [in 2016] to 4.8 per 100,000, 9.4% lower [than the actual rate],” he said in an email, meaning the overall trend would look quite different and less worrying. It’s noteworthy, if not necessarily surprising, that the increase in white homicide has mostly been ignored. (Edit: Because of a silly misunderstanding on my part, this paragraph originally stated, incorrectly, that without the white homicide-rate increase there wouldn’t have been an overall increase. I updated this paragraph to correct the error.)
What’s going on here? As Rosenfeld and his co-authors explain, criminologists believe there to be two distinct types of crime rises. One comes across as the result of gradual, structural factors that have downstream effects on the crime rate. For example, if over the course of ten years a country’s demographic makeup changes, and the percentage of young men between ages 16 and 24 increases, that’s the sort of thing that could lead to an increase in crime (since just about everywhere, the group of humans with the most violent proclivities is young men). Similarly, if the standard of living for a country increases significantly over a decade, that may well cause a reduction in crime.
On the other hand is what social scientists call an “exogenous shock” — some big thing that hits the system like a boulder from seemingly out of nowhere. Because the recent upticks in both the white and black murder rates appear to be so sudden and severe, Rosenfeld and his co-authors believe that they are consistent with the idea that two different exogenous shocks can help explain the current increase in crime: one which hit the white community, and one which hit the black community.
For the white community, it was the opioid epidemic. The market for opioids has absolutely exploded, and any time an illicit drug market expands, it tends to bring with it a lot of misery. In this case, whites, forming a disproportionate share of the market for opioid use, are bearing the biggest brunt of that misery. “Surging demand for heroin and synthetics attracts more sellers into the market,” explained Rosenfeld in a follow-up article focusing on the rise in the white murder rate. “As the market expands, so does the violence resulting from disputes among market participants over price, purity, quantity, and other terms of exchange. There is no reason to believe that whites would be immune to the systemic violence that almost inevitably accompanies underground commerce in prohibited substances.”
As for the shock that hit the black community, Rosenfeld and his colleagues believe the data are consistent with a “Ferguson effect” — just not the one that’s gotten the most attention.
There are two different “Ferguson effects.”
The Ferguson effect is a thoroughly politicized concept at this point, because it contains an implicit rebuke of the protests that exploded in the wakes of the deaths of Michael Brown, Eric Garner, Tamir Rice, Freddie Gray, and other young black boys and men who have been unjustly killed by police. The thinking goes that all those angry, loud protesters are making it harder for police to do their jobs.
But there’s another way of looking at this: It could be the case that yes, deteriorating relationships between communities and the police are helping drive the increased homicide rate, but that things go in the other direction. That is, some communities have come to view the police with such profound anger and fear that their members are less likely to seek out the assistance of law enforcement, and this is making it easier for people to get away with murder.
Rosenfield does believe there’s a Ferguson effect at work here, but he believes there’s stronger evidence to suggest the second version than the one most tightly associated with Mac Donald, in which police are standing down. Mac Donald’s view, he explained, relies mostly on survey data of police officers who claim that they have sensed their colleagues or their departments pulling back, in a sense, but beyond that there just isn’t much evidence to support her story. “It’s not social science, it’s a polemic,” Rosenfeld said of her work. “Now, she’s a very effective polemicist, and in some respects I think she gets the situation right. She was correct, early on, to call attention to the homicide increase that was occurring across the country in many of the big cities. It wasn’t a media creation. It wasn’t simply limited to a couple select cases — Baltimore, Chicago, St. Louis — in fact it was much more widespread. So in that case she was right to sound the alarm … [But she] supports [her Ferguson effect] argument with very, I think, thin anecdotes — she basically relies on police officers who have told her that their colleagues have been drawing back. Rarely have those officers, by the way, said that they themselves have been drawing back.”
But the version of the Ferguson effect Rosenfeld favors does have some evidence behind it, even if the case isn’t yet bulletproof. For one thing, there’s a solid body of past work in criminology and sociology showing that the extent to which members of a given community believe in the legitimacy of local law enforcement affects how they interact with law enforcement. More specifically, a rather ingenious study lead-authored by Matthew Desmond, a MacArthur-winning sociologist at Harvard (and the author of the truly excellent 2016 book Evicted), lends some solid support to this hypothesis. As I explained in a study write-up, Desmond and his colleagues obtained detailed emergency-call data from the city of Milwaukee, and showed that after a horrific, high-profile event in which a group of police brutally assaulted a young black man, 911 calls appeared to be significantly depressed in black neighborhoods relative to what Desmond and his colleagues’ number-crunching suggests they should have been:
Rosenfeld doesn’t think it’s an open-and-shut case that his preferred version of the Ferguson effect explains the increase in the black homicide rate — but he does think it’s the best available hypothesis for explaining the rise in the black homicide rate. He’s currently conducting some research that will help clarify things further.
People should be concerned, but shouldn’t mistake what’s going on now for the 1970s or 1980s.
The proper reaction to the current crime trend should probably be something like cautious concern. A graph from the NIJ paper nicely puts things in perspective:
It’s clear that in big cities, there has been a very real increase in crime, and that nationally there has been a slight uptick. In both cases, though, it would take many more years like 2015 and 2016 to get us anywhere near where we were even in 1995, a point in time at which the crime rate was already dropping from the peak of the crack epidemic. So it doesn’t make sense to argue that we are in the midst of some scary new crime wave. On average, America is still an extremely safe place relative to the fairly recent past — it just depends a great deal, as it always has, on where you live.