About . . . . . . Classes . . . . . . Books . . . . . . Vita . . . . . . . Links. . . . . . Blog

by Peter Moskos

September 26, 2016

Spin This: The biggest murder increase in 45 years

Murder is up. Who knew? (I've been saying so since last October.) Eventually, we're all going to have to accept this (not in a moral sense but in a statistical sense). The accepted liberal reaction to this increase seems to be "it's not a big deal" and "Don't freak out." Let's not get "hysterical." Let's talk about "gun control." (In the early 1990s, by the way, it was all about "drug treatment." That didn't happen either. And crime went down.)

What I really do not understand is why the Left is willing to concede crime prevention to the Right. (I bet Trump won't be downplaying this in tonight's debate.)

False argument #1: The best violence-reduction strategy is a job-production strategy.

It sounds nice, but I say bullsh*t. As if unemployed people just can't help but shoot each other.

Do not get me wrong: Poverty is bad. But it just so happens that 2015, the year with the big murder increase, also saw the biggest decrease in poverty since 1991. 3.5 million people rose out of poverty last year. That's great news. It really is. (Full report & summary in the NYT.)

But we still hear this from people like this St. Louis alderman:
How do we use that [crime] data to elevate the consciousness of our community? How do we use that data to provide the opportunity for people to get meaningful jobs, with livable wages?
No. I mean, yes! Please, work on that, too. But the question from these data is how the hell we get police back into policing and crime prevention. Sure, it sucks when dad loses his job, but consider how much worse it is for dad to get killed coming home from work. (I would even say that you can't have a real job-production strategy until you achieve violence reduction. Who the hell is going to open a business where you will get robbed and workers get mugged walking to their car?)

The Guardian goes on to summarize the Brennan Center's position:
Last year’s national murder increase was not a uniform trend, but a sum of contradictory changes in cities across the country. Early analyses of the 2015 murder increase suggested much of it might be driven by murder spikes in just 10 large cities.
(Now I see how clever the Brennan Center was to put out their paper last week, so it becomes cited immediately to put things "in context.")

False argument #2: It's just happening a few cities.

No. It's not.

Homicide (and almost all violent crime) is up in every grouping of towns and cities (such as "under 10,000" and "over 1,000,000"). Period. Now that doesn't mean it's up in every city. But what a weird and nonsensical standard. Sure, if we remove all the places where crime is up, crime wouldn't be up. But that's we have fancy statistical concepts like "overall," "in general," and "trend."

Even if we were to remove the 6 cities with the largest increase -- and I don't know why we would -- but just to see if the problem is isolated in a few cities, let's take out Baltimore, Chicago, Milwaukee, Washington, Cleveland, and Houston (collectively those cities saw about 420 more murders in 2015) -- even without these cities the rest of America would still have 600 more murders and the biggest homicide increase in 25 years. That's how bad these just released numbers are.

Now we can say that violence in concentrated in certain neighborhoods. That's true. But we've long known this. Indeed, as you can tell from looking out your window, there aren't armed marauders outside your castle gates. What matters, or at least should matter, is that more American are being murdered. I find it distasteful (particularly when it comes from the Left) to say "most people" don't have to worry about crime because the "average person" is still safe. The fact that violence disproportionately affects a subset of Americans may indeed mean it's not a "national crime wave," but it is all the more reason to care.

False argument #3: It might just be a statistical blip.

But it's not. I mean, it could be a statistical blip.... If it were just one or two percent. But it's up 11 percent. The last time we saw an 11 percent one-year increase in murder was 1971. That's exactly my entire lifetime. And that was in the middle of eight-year run when homicides doubled from ten to twenty thousand. This "blip" was literally the deaths of 1,600 more Americans. The number of people killed went up from 14,164 in 2014 to 15,696 in 2015. That one-year increase negated 5 years of homicide decline.

If you think this increase in murder "no cause for alarm" and people who care are "overreacting," to you, I respectfully say "go to hell." We worked too hard to get to where we are (or were) with lower crime. And a "don't-overreact" reaction does not help. And it may lead exactly to the right-wing law-and-order backlash you so fear. (But on the flipside, to those who don't really care but will use these deaths to make some racist point about "black-on-black" crime and "those people," I say with all my heart, "no really, to hell with you, too!")

Why I care (and why you should, too):

Among academics, it's quite uncool to blame criminals for crime or give police credit for crime prevention. But then how many statisticians who use the UCR Homicide Supplement can point to a specific row and say, "Yeah, I handled that one."

Too many who say they're for "justice" never really have to think about the injustice of just even one real murder victim (one not shot by police). But then maybe I care because I was a Baltimore cop. Every single cop can tell you a story about a dead person. Why? Because they care. Granted, some cops do care more than others, but you can't police and not care.

I wasn't a cop for long (less than 2 years in total), and even I lost track of how many victims I dealt with. But a few do stand out. And this isn't even getting into my cop friends who were shot, killed, nearly killed, had to kill somebody, or carry physical and emotional wounds for life.

I remember the stare of a young black man at the same track we ran around while in the academy. His backpack made me think he was a good kid, on his way home from school. He was shot, perhaps after being robbed. We made long eye contact, even though he was dead.

I remember the guy with a gunshot to the head one 321 Post. He was still alive when I got to him. But he clearly a goner, with blood and brain dripping from the hole in his head. His sisters were wailing while he died.

How many Harvard PhD students have the intimate experience of sorted through a victims' clothes? Clothes that are literally dripping with blood and yet still reeking of body odor. You're trying to go through everything, looking for pockets, for any sign of identification of the life that used to be. And then there are the death notifications.

Think of all those deaths. Last year there were 133 more murders in Baltimore than there were in 2014. [This year the numbers are down slightly compared to 2015, and the chutzpah of some people to herald Baltimore's "crime drop" is shocking.] Take a moment and picture all those dead bodies, almost all shot young black men and teenagers. Visually stack them up like cordwood if you wish, or lay them all head-to-toe. It's real human carnage.

If you took all the Baltimore murder victims from just last year and laid them head-to-toe where the Ravens play football, that line of dead bloody bodies could score six endzone-to-endzone touchdowns. And the increase in violence last year happened all after April 27th. All it took was one man's in-custody death coupled with anti-police protests, bad leadership, a riot, and a politician's horrible choice to press criminal charges against six police officers in the matter of Freddie Gray's death. (All charges ended up being dropped after multiple trials without a single conviction on any charge.)

This is actually one time I don't care about the historical perspective. Less than the 1990's crack-crazy murder rate is not good enough. We got down to a homicide rate like Canada (about 1/4 of ours), and maybe I'll be satisfied. We can start caring now. Or we can start caring after a few more thousand people are needless killed. And if you think I'm over-reacting, consider that you might be under-reacting.


Kevin Jackson said...

Hi Peter:

I don't want to speculate why we had an increase in murders, and certainly, any increase is a bad thing.

But the increase needs to be put in perspective. Is what's below from ABC a bad way to look at the data?

"Cities across the country suffered an uptick in violent crime last year, including a nearly 11 percent jump in murders from the year before, according to new statistics compiled by the FBI.

There were 1,197,704 violent crimes committed around the nation last year -- a 3.9 percent increase from 2014, according to the FBI's Uniform Crime Report. However, last year's statistics were still slightly lower than in 2011, and more than 16 percent below the 2006 level, the FBI said today."

The implication from some is that crime is skyrocketing, which doesn't seem quite right. It is definitely increasing, but it's still at very low levels.

I don't see why it's bad to acknowledge both that crime is increasing and that it is still at low levels overall. I see you say you don't care about "historical perspective," but why not?

Peter Moskos said...

Because the historical perspective in America is too high. Crime is not now at a low overall level. We have a crazy rate of violent crime. We get down to the level of Canada, and we can talk. And here's the thing: there's no reason we couldn't get that low.

And yes, even if one does buy the "low by historic US levels" (which I do not), it's the magnitude of the increase (which is still continuing) that is shocking. It's not just up. It's up, crazy up!

[And I strongly believe but cannot prove that all other crimes are increasingly under-reported, for all the same reasons crime is up.]

David Madden said...

Violent crime has to be underreported. There are roughly 800k cops in America. Assume half of them are in patrol and of that 400k, 100k are supervisors who don't take reports. That leaves each patrol officer to take the initial report for 4 violent crimes per year. I work in a quiet suburb and I take way more than that.
Admittedly, this is all anecdotal.

Peter Moskos said...

You know, usually I think why I don't like keeping this blog (it does take a lot of time and writing and pays me nothing). But one thing I really do love about this blog are the comments. This may be the only website in the world where the comment section is actually intelligent, educational, and even-tempered than the actual posts! Thank you, all.

Andy D said...

David *I* work in a generally REALLY quiet rural area and even I take more than 4 per year...and I can think of a dozen or more per year where no report gets taken because the victim refuses to provide information and denies that a crime occurred. In my agency there isn't a lot of pressure to down-play or under-report crimes. But in bigger agencies that pressure is quite real. I think Peter definitely has the right idea trust homicide stats and not trusting the others very much.

Kevin Jackson said...

Hi Peter:

I guess this is where being subjective and looking at the data in different ways comes into play.

So, in order just to keep my own thoughts organized:

1. Certainly any crime is unwelcome, and we should focus on how to keep it down. But comparing the U.S. to Canada seems a bit off, if only because the original point crime in one year over crime in another year in America. To say that it's "crazy" is certainly defensible, but compared to where it used to be, it looks like we are still quite safe.

2. Isn't population worth mentioning? We have tens of millions of more people since the 1980s and 1990s.

3. As for the magnitude of the increase, it certainly does seem like a lot. But this can go in a bunch of ways. Consider that if you have 10 murders per year, then have 12 the next. You can say murders went up by 20 percent, but of course, the overall increase isn't huge.

To be clear, I am not accusing you nor anyone else of anything bad. I'm just saying that small base numbers and increases can lead to figures that sound worse to most people than they actually are.

4. I'm curious as to why you don't buy the notion that crime is at relative lows historically. Is it only because you think crime is under-reported?

As John Pfaff pointed out, crime rates could have been underestimated in the 1960s, too, so perhaps they are lower overall.

5. "[And I strongly believe but cannot prove that all other crimes are increasingly under-reported, for all the same reasons crime is up.]"

Can you explain this a little more?


Overall, I think America is still safe, based on the statistics we have (and even if crime is under-reported, which it might be). I'm not trying to deny it's bad that murders and other crime is up, nor argue anything like it's okay that it's concentrated in one area over another.

I don't doubt that we can get crime down to levels like they have in other countries.

I just worry that losing perspective on the overall picture, which again indicates that it's still safe in America, allows demagogues to gain power and pursue policies that won't do anything helpful and might make things worse.

Peter Moskos said...

#3: The increase in murder wasn't big because we have a low number of murders. The increase in murder was big because a lot more people were murdered. You're talking statistical hypotheticals. But I fail to see why 1,500 more murdered people don't matter. We can debate whether it's the biggest increase since 1990 (by number) or 1971 (by rate), but at some level, who cares which year it's the biggest increase since. It's a long time ago no matter which you choose.

#2: And yes the US is bigger, and that's why generally rates are better than raw numbers.

#4: It's not that I don't buy it, I don't understand why it matters. If there a malaria outbreak tomorrow, we wouldn't say it doesn't matter because malaria is at historic lows in the US. To say "still near historic lows" implies A) that there's some more "natural" higher rate of violence in the US (A mean to regress to). And B) that violence has bottomed and can't get any lower. I don't buy either of those. And B was said throughout the 1990s by those who said -- incorrectly -- that police didn't matter and to reduce crime we would first have to fix society. They're still at it, in effect.

Your last point is what worries me. Demagogues don't gain power if we address and deal with our problems honestly. What feeds the beast (along with craziness, ignorance, and hate) are lies, coverups, when denial of reality. Not only is it morally wrong to downplay the rise in murder, it's politically wrong. There is nothing demagogic or conservative about saying, "crime is up and we're concerned and here's how we're going to police our most violent neighborhoods." It's the reaction and response that matters. Yes, systemic racism in the criminal justice actually is a problem. But it is neither the cause nor solution to rising crime right now. And if Hillary's only or best response to, "what are you going to about rising crime?" Is "address systemic racism in the criminal justice system," we're doomed.

Peter Moskos said...

#5: People are less likely to report crime if there's more crime. People are less likely to call police to report crime if they don't like or fear the police. People are less likely to report crime if cops take forever to show up. And cops are less likely to find crime if they're not being proactive.

Yes, crime stats are dodgy. I don't like them, particularly over decades. Murder stats are much less dodgy.

If murder is up, crime is up. I guarantee it. But in theory the NCVS should eventually tell us if under-reporting is up. But like all poll data, the NCVS ain't what it used to be.

Jack Catchem said...

The murder rate is also carefully tracked by criminologists. They are aware of the propensity for police departments to have fun with numbers depending on the politics of the day. The one rate that is the most difficult to fudge is the murder rate. Especially if the murder rate continues to climb dramatically, it is time to raise an eyebrow at the CompStat madness.

Case in poin: an overheard joke while walking through the detectives room back in a former department, "What do you call a robbery with no suspect in custody? A theft!"

This is despicable. A crime report is a piece of intelligence, let the crimes be what they are so we can plot a response based on good data.

EA5 said...

Have you read Randolph Roth's American Homicide? I haven't read it but Balko mentioned the Roth's 4 factors to murder rates in one of his articles a couple days ago so I was considering checking it out.

Peter Moskos said...

David, that "back of the envelope" math is really interesting. Curious. Perhaps it is the difference between Part 1 and Part 2 crimes?

David Madden said...

I was just thinking of Part 1 violent felonies per UCR guidelines. Aggravated assaults are not exactly an everyday thing, but a bar fight where someone gets knocked out or gets hit with a bottle or a pool cue do happen. "Real" robberies are rare, but bullshit ones where you can't prove the bullshittery still get coded as robberies. Throw in the occasional rape(that sounds terrible, but you know what I mean) and your average patrol cop just has to handle more than four violent crimes per year. I have no evidence to support this, I'm just speaking from what I see.

Peter Moskos said...

764,449 aggravated assault does seem awfully low. Suspiciously low, in fact. Especially since it's supposed to be crimes reported to police (as opposed to the charge that is prosecuted or pled out). Certainly cities (and thus departments) would have it in their interests to under-report crime stats to the FBI...

I'm going in... this could be a hellish rabbit hole.

Peter Moskos said...

Well spot checking Baltimore for Agg Assault, using Baltimore City Open Data with UCR for 2014, it matches close enough (ie: very close). Save with arson, burglary, and well, everything. Now it's possible that Baltimore reports its data correctly. As I do trust Balto open data
The data gets in there pretty quickly.

Baltimore had about 4250 Agg Assaults in 2014. Let's say 1,000 cops on patrol taking reports. Comes out to 4.25 per officer per year. Which does sound low, but I think is correct.

So how's this for a hypothesis: Perhaps we're confusing "calls cops go to" with "incidents." Five cops show up a bar fight and each of those 5 officers remembers the agg assault and adds it to their mental tally. But only one cop writes the report (and add to that that the UCR counts incidents better than victims, so multi-victim scenes tend to get interpreted as one incident.)

In other words, maybe you're remembering things you actually did see rather than reports you wrote as the primary? If you want to email me, I'd be happy to tell you the official number of crimes in your county (that's how it's broken down in what I'm looking at) in 2014. You can see if it jibes with reality.

Kyle said...

Something irrelevant: http://www.nytimes.com/video/opinion/100000004677716/supreme-court-v-the-american-voter.html

The road of Trump's presidency has already paved, just hope that he crashes.

Alex Elkins said...

An obvious caveat, curiously unmentioned here (afaik), is base rates. Yes, murder spiked (the increase is disturbing and cause for concern), but an eleven-point increase from a low starting position is less disturbing than an eleven-point increase when murders are already high.

Aagh! Murders jumped 300 percent. Last year, we had one. This year, we had 4.

What do we think of this line of attack? Is it irrelevant, beside the point? Genuinely curious.

Peter Moskos said...

I think it's crazy. Because last year we didn't 4 murders. We had 15,700. It's statistically ignorant at best. But coming from those who know better? Disingenuous and dishonest.

And people are making a hypothetic statistical arguments that don't even apply in this case! Had this increase (+1,532) happened when homicide was highest (1991), it would still be a 6 percent increase... and that would still be the biggest increase since, well, 1990.

Basically people are saying, it depends on how you measure it. But it doesn't. Let's say we conceded the point and didn't use rate. In absolute numbers the increase is still the biggest since 1990. So what exactly are we quibbling about? The point of saying "biggest incease since" isn't really about the year but to give notice that what is going on is *not* business as usual.

I'm trying to think of a analogy. How many of these arguments make sense?

A kid grows a foot in a month to 7 feet tall and people respond with, "well, it's not really a growth spirt because he was already tall."

Or maybe observing that a football field is big. 100 yards long. And somebody comes back with, "well, yeah, '100' if you use these "yards" no other country uses. It's really only 91.44 meters."

Homicide decreased 20 percent. "Well, it's not a big deal because homicide is low?" Of course not.

Global warming? A 1 degree rise in temperature is less of a big deal because last year was already hot, the warmest on record?

Test score rise drastically... but don't give the principle and teachers any credit because the school had been really crappy?

Cancer? Well, your tumor is growing and not in remission, but no need to worry becuase it's not as large as it was before chemo a few years ago?

There's simply no way you look at this data and pretend it's not a big deal. Any attempt to do so is simply trying to obfuscate reality.

Vidoqo said...

Not sure where my comment went. No matter - I extended it into a post on my blog here: http://supervidoqo.blogspot.com/2016/10/a-teacher-in-hood.html

Again, thanks for the excellent post.

Alex Elkins said...

Thanks, Peter. I often wonder about this divide, between crime skeptics and crime realists (guess which side I'm talking about), and whether it boils down to dishonesty or ignorance or...

I tend to think this divide reflects 1) personal/geographic distance from the problem of violent crime and 2) different priorities regarding the costs of enforcement versus non-enforcement. (I borrow this latter point from John Pfaff.) It's unreal that the 2015 crime stats were so hyped. I think that fact alone speaks to point 2 above. For the last few years, the public has become more aware of the fiscal and human costs of over- and under-policing and mass incarceration.

Of course, the divide also likely speaks to a more basic rift, which returns us again to a fundamental question posed often in this space: do police matter in reducing crime and maintaining order?

Otis Blue said...

One curious aspect of this whole situation is that it provides proof positive that "White Privilege" does indeed exist.