As predicted, the FBI crime stats for 2015 were released Monday. Also as predicted, they show an 11 percent increase in murders from 2014. This will almost certainly be demagogued by Donald Trump at Monday’s debate, as well as by law-and-order groups to make their case on issues such as mass incarceration, police militarization and mandatory-minimum sentencing.

Trump might want to consult with some chiefs of police themselves. Here’s Ronal Serpas, head of Law Enforcement Leaders to  Reduce Crime and Incarceration, a group that includes the current police chief of Los Angeles, as well as former chiefs of police for New York and Washington, D.C.

“Irresponsible claims that crime is out of control are not backed up by the facts, as new data analyses show. Overall crime remains the same as it was last year, though a small number of cities have seen a rise in murder rates. Increases in such cities are troubling and must be addressed immediately. “Police officers continue to work hard every day to keep our country safe, and sensational headlines don’t help us in that mission. False narratives on rising crime create panic and division, instead of giving police and communities an opportunity to work with each other toward an even safer place to live.”

It’s also important to keep some perspective, as this tweet from defense attorney and activist David Menschel points out:

Beginning in about 1994, crime in the United States dropped consistently for 20 years. It was never going to continue dropping forever. That doesn’t mean, however, that a leveling off or even a slight uptick means we’re headed back to the early 1990s.

One of the problems with measuring crime as a rate is that rate increases tend to loom larger as the base number falls. For example, let’s say the murder rate in City X was 20 murders per 100,000 people in 1994. If the rate went up by one murder per 100,000 the following year, that would amount to a 5 percent increase. Now let’s say that City X saw a steep drop in crime over the next 15 years, and dropped to a murder rate of, say, just 5 people per 100,000 in 2009. After bottoming out, the rate then went up by 1 murder per 100,000 the following year. That would make for a 20 percent increase. Measuring crime by its rate in the population will always be infinitely better than using raw numbers. But it isn’t without its limitations.

These problems then get magnified when headlines frame them in ways such as, “U.S. sees largest crime increase in 30 years.” That may be true. But that’s only because (a) crime was falling for 25 or more of those years, and (b) any increase from a historic low is going to look larger than increase from a period when crime was more common. If you wanted to be really devious, you could write a headline like, “City X murder rate jumps 20 percent, five times more than next largest increase of the last 30 years.”

City X isn’t less safe now than it was then. It’s quite a bit more safe. But context and long-term perspective are harder to fit into a headline.

One more thing: One of the more interesting recent studies of murder rates is the book “American Homicide” by Randolph Roth. The book surveys homicide in the United States going back to colonial times. Roth argues that the homicide rate seems to coincide with four sociological phenomena: a loss of government legitimacy in the eyes of the governed, political instability, a loss of camaraderie and shared citizenship among residents of an area caused by political antagonism, and a loss of faith in social order.

Now think about where crime is soaring, places such as Chicago, Baltimore and St. Louis. Chicago has been racked by a steady stream of police scandals. The city’s top prosecutor was recently ousted by voters who saw her as a rubber-stamp for brutal and dishonest cops. And the city is regularly ranked among the most corrupt in the country. Baltimore’s police department was just eviscerated in a blistering Justice Department report that found widespread brutality, corruption and racism. The city’s government has also been plagued by scandal. St. Louis and the county around it has long been weighed down by system in which municipal governments essentially prey on their own citizens — and the poorer the town, the more reliant they are extracting fines from their residents.

Roth’s theory encompasses both the actual legitimacy of government and the perception of government legitimacy. Here, too, we see evidence for the theory in these three cities. The aforementioned problems have persisted in all three cities for decades, and all three have always had comparatively high murder rates. But in all three cities, recent high-profile incidents cast a national spotlight on those problems, only further undermining local faith in the legitimacy of local government. And the homicide rates soared ever higher.

Now consider this: If Roth is right, if crime and homicide are driven by citizens’ belief in the legitimacy of the current government, a feeling of belonging with their fellow citizens, order and political stability, what might the effect on crime and homicide be if, say, local politicians and law enforcement leaders began telling underserved communities that if they criticize police for perceived brutality or discrimination, the police might simply stop serving them? What effect might it have if those same officials announced that that the police already have stopped doing their jobs in some of those communities?

There are, of course, lots of variables that affect the crime rate. Roth’s theory is based on broad observations about what historically was happening as the crime rate waxed and waned. I’m sure you could find cities that don’t perfectly fit the model. But it certainly seems to apply in the places that are driving the country’s overall increase in violent crime.