Violent crime up second year in a row

This isn’t good news.

It’s not “carnage,” and it’s not a reason to panic or adopt cruel, stupid policies, but it really isn’t good news.

Yes, year-to-year homicide rates are statistically noisy, especially at the city level, because homicide is a relatively rare event. But an increase of roughly 20% over two years isn’t just statistical noise. And though there are indeed dramatic increases for identifiable local reasons in places such as Chicago, that’s not what’s driving this train: the 2016 increase showed up in small tows as well as big cities. [Update: Robert VerBruggen makes a key observation: white-on-white homicide rose more in percentage terms than black-on-black homicide.]

But the real reason to be concerned isn’t that homicide has gone up for two years in a row; it’s that it was flat in the two previous years, pretty clearly breaking the 20-year downtrend starting in 1994. Yes, even after two bad years, we’re much better off than we used to be. But even at 5 homicides per 100,000 - about half the 1994 rate - our current rate is three times the Canadian rate, five times the UK rate, and ten times the Norwegian rate. That’s nothing to write home about. Getting the homicide rate moving back down ought to count as an important national goal.

But how?

Better policing clearly matters. New York and L.A. homicides are still trending down; New York in particular is having a spectacularly good year, after deciding three years ago to radically cut back on stop-question-frisk. (If you still think more incarceration is what we need, note that NYC incarceration rates - the Rikers population plus the people we send upstate) are down about 40% from their peak.) Watching one of the NYPD’s periodic CompStat meetings, with police leadership reviewing every recent homicide in the target precinct and asking “What could we have done to prevent that one?” gives you a sense of what “Black lives matter” means as a strategy rather than a hashtag.

But better policing isn’t something Amazon Prime will deliver in two days; it’s the product of years, or even decades, of hard and not-very-politically-rewarding work. And better policing means - not only for reasons of justice but on a purely operational level - lawful, respectful policing that maintains its perceived legitimacy in high-crime neighborhoods. The National Network for Safe Communities has been doing the hard work of forging alliances between police management and community leadership around effective violence-reduction strategies for those neighborhoods; listen to NNSC and not Jeff Sessions if you’re interested in doing something about violence rather than using it as a political weapon. (Videotapes of the NNSC annual conference at the link.)

Yes, the “Ferguson effect” - increased homicide as a result of both decreased police activity and decreased community cooperation with the police after a well-publicized fatal police shooting of a civilian - is probably real, and a significant contributor to upsurges in Chicago, Baltimore, and St. Louis. You’re welcome to blame Black Lives Matter activists if that makes you feel better, or to urge the people using that slogan to dial back the anti-police rhetoric if you think you have any influence.

But if there’s a real Ferguson effect, that’s also a good reason to change police training and tactics to reduce the number of civilians killed by officers. No one has a good count of those events - which itself suggests that departments are less than fully focused on keeping those numbers down - but estimates seem to center on 1000-1200 per year. The comparable figure for Germany is nine. No, that’s not a misprint for “ninety”: German cops kill about nine civilians per year. Yes, U.S. civilians are more violent and better-armed, and of course the German population is only about a quarter of ours, but still and all there doesn’t seem to be any good reason our rate of police killings of civilians should be 25 times the German rate. In particular, Frank Zimring has found, in studying the converse problem of cops killed by civilians, that 96% of those deaths involve firearms, strongly suggesting that shooting unarmed civilians, or those armed only with knives or clubs, is not really essential to protecting police.

Of course there are lots of ways of reducing homicide that don’t involve policing: some inside the criminal justice system, others (early-childhood programs, lead removal) outside it. But most of those take time.

If you’d like to reduce the homicide rate for next year, you have precisely one proven option: raising alcohol taxes. More than half of all homicides involve alcohol, on one side or both, and higher taxes demonstrably cut back on heavy drinking. Phil Cook estimates that a 10% increase in the price of alcohol - which would result from doubling the current federal alcohol tax of about 10 cents per drink - would cut back on violent crime by about 3%. The statistical-noise problem makes it harder to estimate the effect on homicide using Cook’s studies of state-level tax changes, but there’s no good reason to think homicide is more or less responsive to changes in heavy drinking than violent crime in general. So a good horseback guess is that doubling the federal tax now would lead to about 500 fewer homicides next year.

Since All Lives Matter, why don’t we do it?