July 22, 2013

Is violent crime actually falling?

You've probably noticed that Chicago's weekend shooting wrap-up articles tend to have headlines with ratios of dead to wounded like this from the long Fourth of July weekend:
Chicago Shootings: 12 Killed, At Least 62 Wounded In Gun Violence Over Long Holiday Weekend

These days in Chicago, a city with fine trauma care centers, it's not uncommon for only 15 or 20 percent of gunshot victims to die, at least according to my scan of weekend wrap-up headlines.

From the WSJ in 2012:
At the same time, medical data and other surveys in the U.S. show a rising number of serious injuries from assaults with guns and knives. The estimated number of people wounded seriously enough by gunshots to require a hospital stay, rather than treatment and release, rose 47% to 30,759 in 2011 from 20,844 in 2001, according to data from the Centers for Disease Control and Prevention's National Electronic Injury Surveillance System-All Injury Program.

Now, this statistic might not be the gold standard of serious violent crime, either. Perhaps guns improved enough to to do more serious damage requiring more than treatment and release. What about better ambulances keeping victims alive long enough to get to the hospital rather than the morgue? Still, this stat raises questions about the assumption that violent crime is steadily falling.
The CDC estimates showed the number of people injured in serious stabbings rose to 23,550 from 22,047 over the same period. 
Mortality rates of gunshot victims, meanwhile, have fallen, according to research performed for The Wall Street Journal by the Howard-Hopkins Surgical Outcomes Research Center, a joint venture between Howard University and Johns Hopkins University. In 2010, 13.96% of U.S. shooting victims died, almost two percentage points lower than in 2007. 

66 comments:

Moldbug said...

Questions? It doesn't raise questions. It answers the question.

What is our hard data? Our hard data is not that "crime is falling," but that crime reports are falling. It's a fact that crime reports are falling. It's an inference to reason from the fact to that crime reports are falling, to the theory that crime reports are falling because crime is falling.

The curious thing about the people who espouse this theory is that they often express it (see recent Economist article) as a "mystery." Indeed it's mysterious. Most crime is committed by the usual criminal classes; these classes are neither becoming less numerous, nor less criminal.

The gunshot wound data, which is not ganked, reveals the answer. Crime reports are falling not because people are nicer and shoot each other less, but because fewer people are reporting crime. To become a statistical report a crime has to pass through a layer of reporters, from victim to cop, each of whom can decide to not-report and leave no statistical trace, each of whose motivations and incentives are changing in favor of not-reporting.

It's typical of 20C thinking, with its ascription of papal infallibility to anything that (a) is a number and (b) is printed in the NYT, to make this inferential leap, totally unsubstantiated and palpably dubious, from number of reports to number of crimes...

Anonymous said...

I hope you saw this NYT story. What is it about?

http://www.nytimes.com/2013/07/22/business/in-climbing-income-ladder-location-matters.html?hp&_r=0

Mencius Moldbug has a post somewhere about how the violent crime stats in NYC these days are probably fudged by the cops. If you go by what the ER's in NY claim instead of what the city claims, violent crime in NYC has increased over the past 20 years.

Anonymous said...

I asked my father, an ER doctor whether it was worse to get stabbed or shot. He said stabbings were worse since usually gunshot victims only have one hole punched through them but stabbing victims tend to have 20 or more stab wounds. The wars have really contributed quite a bit to trauma medicine and emergency life support. If you can keep a soldier alive who has had 3 limbs blown away a gunshot is childs play.

David M. said...

The crime stats are always shown per 100,000 of population. A much more valuable crime stat would be all violent crimes per x-number of males aged 15-35. Some sort of consideration for the percent of males in prison would also be necessary to really determine whether violence as a social phenomenon is actually decreasing or increasing.

Contra Pinker, if you look at violent crime rates over the 20th century, and then consider that society in the early 1900's had relatively few people living past the age of 65, and a much more pyramid-shaped age structure, not to mention poorer medical care, less effective policing, and a much smaller prison population, the contention that we are becoming less violent by nature doesn't seem all that convincing.

Anonymous said...

Just because doctors and paramedics are getting better at keeping us alive, doesn't mean the crime rate is going down.

Anonymous said...

"What about better ambulances keeping victims alive long enough to get to the hospital rather than the morgue?

Seems to me that rising violent crime would be unlikely to be accompanied by an improvement in emergency medical services.

Jefferson said...

I never knew how racially segregated the city of Chicago is.

Zip codes in Chicago range from 100 percent Black to less than 1 percent Black.

http://zipatlas.com/us/il/chicago/zip-code-comparison/percentage-black-population.htm

Chicago is a very Balkanized city. Within Chicago you have zip codes that racially resemble Whitopias as well as zip codes that racially resemble Haiti.

JeremiahJohnbalaya said...

http://www.nytimes.com/2013/07/22/business/in-climbing-income-ladder-location-matters.html?hp&_r=0

Trust NY Times to bury all the details in fluff. And why are they only concerned with moving from the bottom to the top quintiles? Wouldn't it be interesting to know about moving up and out of the bottom to other (ie. BETTER) quintiles? Or is it too important to bury the fact (which Sowell raises) that that happens all the time.

Anonymous said...

I put it to my girlfriend, whose hospital is THE trauma place in Boston. She's the ICU case manager, and she currently has 3 shooting victims, 2 of whom would have been corpses 10 years ago, though the prognosis on 1 of them isn't good (9mm to head).

To further prove Moldbug's point, only one of them has been reported on.

Silver said...

"Perhaps guns improved enough to to do more serious damage requiring more than treatment and release. What about better ambulances keeping victims alive long enough to get to the hospital rather than the morgue?"

Or perhaps the medical community became more sensitive to gunshot wounds and began to hospitalize more of the marginal cases.

"Still, this stat raises questions about the assumption that violent crime is steadily falling."

With respect to the topic of this post, firearm violence, the Bureau of Justice Statistics report Firearm Violence 1993-2011, based on data from the National Crime Victimization Survey contained the following:

Total Non-fatal Firearms Victimizations [not equivalent to 'shootings']:

2001-563,100
2002-540,000
2003-467,000
2004-456,500
2005-503,500
2006-614,400
2007-554,800
2008-317,300
2009-410,000
2010-415,000
2011-467,300

(Standard Error roughly 50,000 each year.)

That's down from over 1,500,000 in 1993.

Number of firearm homicides largely unchanged from 2001-2011, mirroring the slight uptick mid-decade then fall of non-fatal firearm victimizations.

Some rather interesting racial data in Appendix Table 10. Hispanic rate of non-fatal firearm violence consistently less than 2.0 times the white rate over last decade. Not perfect, but not 'nightmare scenario' levels either.

(Note to Mr. Moldbug: this is survey data, not voluntary reporting. Maybe they learned a few things in the 20th century after all.)

Thin-Skinned Masta-Beta said...

The Soviet journalists covering jubilant fulfillment USSR 5 year plans seem to have been reassigned to the US crime stats beat.

Dave Pinsen said...

"Mencius Moldbug has a post somewhere about how the violent crime stats in NYC these days are probably fudged by the cops. If you go by what the ER's in NY claim instead of what the city claims, violent crime in NYC has increased over the past 20 years."

What percentage of reported gun shot wounds represent accidents, shootings by police, or gang-on-gang violence which the victims report as accidents? It's pretty clear that crime has declined in NYC, at least in the parts you would want to visit. Decades ago, everyone in New York had either been mugged or knew someone who had. That affected behavior in obvious ways. Fear was palpable and people walked at DefCon 1.

Today, it's not uncommon to see young women walking or jogging by themselves late at night in Manhattan. They don't look worried about crime at all.

More objectively and quantitatively, consider real estate prices. Would they be booming in cities like New York or San Francisco if crime there were on the rise?

agnostic said...

Anyway, the shift toward more deadly types of assault weapons actually corroborates the story about falling crime rates.

Predilection toward violence is a spectrum trait, and the population is distributed along the spectrum. The distribution has moved toward the less-violent end over the past 20 years. This is the fall in reported crime rates, in surveyed experiences with crime, in reported fear of walking alone in your area, etc.

Murderers are still out there -- but these days they represent a much higher percentile (or z-score, however you want to think of it), of a much less violent population.

In the good old days, the average American 20 year-old might have started a fist-fight in a club, bar, or party at least once during a typical year. Or however often, it doesn't matter.

Now, that average American 20 year-old has probably never gotten into a fight in his entire life, and will not for the foreseeable future. When and if their tempers flare up, they just bottle it up and redirect somewhere else -- spawn-camping in some video game, watching a voyeuristic revenge movie, posting agro comments on the internet, or whatever. But not in a violent way.

To commit murder in such a non-violent society puts you so far out into the right tail of the distribution. You're the type of guy who'd only use a knife, gun, etc. Murderers are more thuggish than before, because the middle range of the distribution used to be fairly violent itself.

This applies to the secular decline in homicide rates as well. In Chaucer's day, it was common for two young rowdy dudes to get into a quarrel at the tavern, and one stick the other with a knife to defend his honor in public.

Now that that sort of thing is so uncommon, only the truly thuggish fraction of the population will lie beyond the threshold for committing murder, and will use more gruesome methods.

Time for historian types to start thinking harder about probability.

josh said...

In addition to blacks being pushed out of major cities, people have adjusted their lifestyles to avoid blacks. The blacks on TV can't hurt them.

Orthodox said...

I don't know enough about gun history, but my general sense is that the military and police and probably individuals (following their lead) have become more efficient in their use of scarce resources. That is to say, they have moved towards the caliber gun with the required stopping power and do not "waste" resources.

However, the average round of the same caliber probably is more deadly today. So the question is whether the most common calibers have gone down in size, or whether they have stayed consistent.

agnostic said...

Official homicide rate statistics also CLAIM TO show a decline from a peak in 1933 through a trough that lasted through 1958.

But then we all know that the world of Leave It to Beaver and Father Knows Best was every bit as violent as it had been in the days of Pretty Boy Floyd, Dillinger, Babyface Nelson, and Bonnie and Clyde. (All tracked down and killed in the single year of 1934 -- it was a busy year for G-men).

The apparent decline was just due to the coppers under Fiorello La Guardia cooking the books.

The past 20 years is not the only extended period where crime / homicide rates have fallen. If you don't know what the other periods are, and what they do and do not share, you can't hope to figure out what moves crime up or down.

Perhaps because this topic touches so many taboo areas -- death, violence, crime, sex, authority, community, revenge, justice, etc. -- almost nobody is willing to take an honest look at the big sweep of things and try to learn what history has to teach us, showing humility.

Instead, everyone has their own impassioned lecture to give on all of those taboo areas, that trying to get them to shut that off in order to talk about changes in crime rates, seems hopeless.

Anonymous said...

Housing policies like section 8 are moving the violent poor out of cities to small towns, for example:

More Questions About Crime In Wilkes-Barre After Shooting

http://wnep.com/2013/07/19/more-questions-about-crime-in-wilkes-barre-after-shooting/

http://en.wikipedia.org/wiki/Wilkes-Barre,_Pennsylvania

Art Deco said...

I am looking at the CDC data and they report non-fatal injuries from firearms at 22.11 per 100,000 in 2001 and 23.64 per 100,000 for 2011, so I am not seeing this 50% increase alluded to in the other thread.

The wounded hospitalization rate increased from 9.44 to 11.62 per 100,000 during that time period. That could also be driven by changes in hospitalization protocols.

Assault related injuries (non fatal) from firearms would appear to have increased from 14.4 per 100,000 to 17.85 per 100,000 per the CDC. (FBI reports a decline in aggravated assault).

(non-fatal) Assault injuries from all causes have declined per CDC from 623 per 100,000 to 568 per 100,000.

Assault injuries (all causes) requiring hospitalization seem to have increased from 27 to 38 per 100,000. (Again, the FBI reports a decline in aggravated assaults).

This would appear to be local to what's been going on with firearms. One has to ask what has been driving hospitalization, worse injuries or changes in the reaction to injuries?

Art Deco said...

Just because doctors and paramedics are getting better at keeping us alive, doesn't mean the crime rate is going down.

Again, rape, aggravated assault, and tracked property crimes have also seen declines in their frequency (the latter by 27% over 20 years).

Anonymous said...

"Seems to me that rising violent crime would be unlikely to be accompanied by an improvement in emergency medical services."

Modern EMS systems and Emergency Medicine as a specialty only date back to the 1970's. Congress only passed the Trauma Care System Planning & Development Act in 1990.

There have been plenty of technical advances since then that have been cheap enough for even poor, violent cities to implement - such as AED's, administering emergency meds via intraosseous infusion, better intubation techniques, hypothermia for cardiac arrest, etc.

And cities like Philly, Baltimore, and Seattle serve as civilian and military training centers for emergency medicine and trauma care, so they get federal money for research, which helps.

Jehu said...

I did a few posts on my blog a while back on trauma medicine and the homicide rate. One of the things I noticed is that the conversion rate of aggravated assault to successful homicide stopped falling much after 2000. One of my strong suspicions is that fudging of the denominator (the number of aggravated assaults) has increased an awful lot since then. We have a fair number of stories in the media about such fudging in Wisconsin and New York. One might make a first pass at estimating HOW much fudging we're talking about by assuming that the murder:aggravated assault trendline continued in its historic pattern and that the delta represents the fudge. But that's pretty crude.

Anonymous said...

On a lighter note:

It's noticed that Mr Sailer did not comment on Phil MIckelson's amazing 66 in final round for his first ever British Open. Maybe ol' lefty doesnt look the modern super fit athletic part but getting the job done should count for something.

All Phil would need is to win the US Open and he's in the same elite company of 5 others who have won the 4 majors at least once.

A great victory should be given the respect even if he's not the media's favorite. Standing by a significant other who had cancer is also to be highly commended.

A great victory, Phil! God Speed to winning the US Open next season.

As for Eldrick the Erkel-lite, TICK, TICK, TICK goes the father-time clock.

jody said...

"I asked my father, an ER doctor whether it was worse to get stabbed or shot. He said stabbings were worse"

this is correct for year 2013 ER treatment. as long as you make it to the ER, knife attack survivors are usually in worse shape than shooting victims.

this is because most people are shot by handguns, and 32, 380, and 9mm FMJ bullets, which is what most criminals shoot people with, do mainly just poke holes in you. if you get shot once or twice with them, you probably won't die if you don't get hit in the brain or heart.

i'm not saying 40 and 45 are any more effective - with FMJs they just poke holes in you too. i'm just saying, street criminals do most of the shootings and they overwhelmingly use 380 and 9mm, and load with FMJ, since that's cheapest and easiest to get.

the other biggest category of shootings are suburban and rural people hit with 22 long rifle, which is the least dangerous small arm caliber to be hit with, and is also unlikely to kill you. even out of a rifle, 22 LR is not moving at rifle velocity.

once you start talking about handguns of 9mm caliber and up loaded with year 2013 hollowpoints, things get more dangerous. 380 and below, even with hollowpoints, are less dangerous due to lower velocity and mass, and have trouble penetrating you and hitting stuff that will kill you right away.

shotguns are somewhere in the middle, depending on what they are loaded with. projectiles out of shotguns also do not fly at rifle velocities, so here the size of the pellets is what matters. bird shot can definitely kill you but is a lot less likely to kill you than buck shot. number 1, number 0, or number 00 buck is what usually kills with multiple hits.

then when you step up to rifles, things get a lot more dangerous. once bullets are moving at rifle velocities, 2000 feet per second or higher, the wounds they cause are very destructive and much harder for any doctor to save a gunshot victim.

one surgeon's opinion:

http://www.youtube.com/watch?v=tku8YI68-JA

but the gun community has been debating this stuff for 30 or 40 years. if you research, you'll find lots of references to fackler's research versus marshall and sanow. doctor gary k roberts is probably the current leading authority on this stuff.

Anononymous said...

"wounded seriously enough by gunshots to require a hospital stay, rather than treatment and release , rose 47%"


Why make the distinction? More sophistry to lower the statistics. You get shot with a bullet. You bleed all over the sidewalk. You go to the hospital in an ambulance. They pull the bullet out with forceps and stitch up the hole. Ahh ... but you didn't need to stay overnight, so we're not going to count it.

Steve Sailer said...

My personal impression is that property crime is way down, with the exception of smartphone theft.

Anononymous said...

"Is violent crime actually falling?"

You would hope so, with all the money we spent on jail construction. A family of 4 pays $800 in tax per year for them.

black men in jail or prison:
1980: 143,000
2008: 846,000 = 5.9x

In 2011, 7.5% of black males ages 25 to 39 were imprisoned.

Whiskey said...

We know that there has been deliberate under-reporting of homicides and other serious crimes for years in NYC, in Philly, in DC, and Chicago, often in response to Compstat like evaluations and the like. This just in, demand performance to a metric and managers will game the system.

And we also know that there have been short term gains in certain areas of high-profile cities. Manhattan from say, Central Park southwards has gotten safer -- but has the Bronx, or Queens, or other unglamorous boroughs? Property values should indicate that movement, and from out here in LA my observation is mixed.

But the thing about crime (and we are really talking about Blacks and Hispanics) is that it can move, quickly dominate a city (Detroit circa 1967-73, Chicago likely now), and profoundly impacts the willingness of those with lower resources to stay in a city. Even in the darkest days, Manhattan had really, really rich people insulated from crime. But those lacking the money for security buildings, car service, bodyguards and the like, tend to flee -- and take their tax money and investments with them.

Look at LA. There has been a lot of effort to make Downtown a happening place, with Staples Center and all that. But the constant grinding crime of Skid Row not that far off, and the upper reaches of South Central, and the MS-13 stomping grounds of Westlake Village off to the West, and East LA, make that a non-starter. Its nothing like say, Chicago's Lakeside district.

On the other hand, Santa Monica positively impacts Venice, which has seen pockets of gentrification despite constant street crime, because Santa Monica is isolated to an extent by the 10 Freeway and the ocean from incremental crime. South of Venice, there's Marina Del Rey and Playa Del Rey and the airport/Dockweiler State beach (and Hyperion Sewage Treatment plant). Not much street crime or activities that generate street crime: drug sales, gangs, etc.

This has made Venice a gentrification project since the 1980's. My bet is it will go full gentrification sooner or later because beachside property is worth a lot of money if cleaned up. And Downtown will always have Skid Row, MS-13, and South Central bordering it.

[It seems massively foolish to have connected via Blue Line and Downtown Subways, South Central LA to Hollywood, its an expressway to stop-and-rob tourists there.]

Moldbug said...

agnostic, I find your lack of faith disturbing. You're exactly the fool Carlyle describes in his essay on statistics, which you've never read and never will, who offers us "with asseveration and protrusive importunity," your basket of gravel as gold:

http://www.stat.ucla.edu/history/chartism.pdf

To recap: I point out that you've made a completely unscientific and unsupported assertion, that the ratio of crimes to crime reports has remained constant across the last few decades. Our host shows us significant data, much less likely to be manipulated (see under: Goodhart's law), which conflicts with your unsubstantiated assumption.

And our host has no special expertise in the area, but you can go to any cop blog, any time of day, and see what they think of crime statistics. Do you? I know you don't.

And in 500 words of bloviation, what argument do you offer? That in America in the 20C, crime declined previously. So it's quite possible that, later, it declined again. Sample size: 1. Neighbor, please! If this is the scientific mind at work, give me witchcraft and astrology.

You have a bunch of numbers - official numbers; Numbers, really, collected by the government; and from this autistic digital skeleton you (and the Economist, and everyone else) has constructed a Story. In your mind, this Story is reality itself - just as St. Trayvon is real to millions of Americans. But if your Numbers are a giant sack of shit? The Soviets, too, had a Numbers and a Story.

I love your "average American." You might as well talk about an "average African," an "average human," an "average galactic lifeform." This ridiculous category exists in your mind for one and only one reason: that Numbers in this category are collected. You have the table, you can run the query!

America is not in any way a homogeneous entity. I have never met an "Average American." I haven't even met an "Average White American" or an "Average Black American." I am reasonably familiar with the "Average Hipster-White American," the "Average Ghetto-Black American," etc, etc, etc. I will accept these as reasonable patterns about which we can have a reasonably specific discussion.

We can agree, for example, that Manhattan (Lower Manhattan) has much less street crime than it had 20 years ago, because it has many fewer AGBAs and consists almost entirely of AHWAs - whose microscopic violence level has perhaps diminished to the femtoscopic. We can even agree that the violence level of AGWAs, Average Guido-White Americans (meaning guineas, Bohunks and the like) has diminished considerably over the last 50 years.

(We can also agree that the AGBA is pretty much what he was 20 years ago, and any numbers suggesting that he became 3x less violent between 1993 and 2001, but did not change at all between 2001 and 2011, belong under the desk with Khrushchev's grain production statistics.)

These are actual historical phenomena. Your "Average American" is just a bowl of data mush. The government may have to serve this mush to you. You don't need to serve it to us. And we don't need to eat it.

The Department of Information doesn't sort its files by AGBA, AHWA, AGWA? Too bad - then you don't have any useful data on these categories. You'll have to actually think. "Almost nobody is willing to take an honest look at the big sweep of things and try to learn what history has to teach us, showing humility." Down, dog, and kennel!

Anonymous said...

"It's typical of 20C thinking, with its ascription of papal infallibility to anything that (a) is a number and (b) is printed in the NYT, to make this inferential leap, totally unsubstantiated and palpably dubious, from number of reports to number of crimes..."

Especially when there are politically correct reasons for reality to be submerged.


.

"Seems to me that rising violent crime would be unlikely to be accompanied by an improvement in emergency medical services."

In the context of political correctness that is exactly what you'd expect.

.

"Again, rape, aggravated assault, and tracked property crimes have also seen declines in their frequency (the latter by 27% over 20 years)."

Crime has been falling among certain populations and increasing among others. The actual reality is the balance between these two trends which varies from place to place. The perceived reality is something different.

.

"My personal impression is that property crime is way down, with the exception of smartphone theft."

Criminals switched to that for obvious cost/benefit reasons.

(A lot of it is thefts from kids also which adds to under-reporting.)

Silver said...

"I did a few posts on my blog a while back on trauma medicine and the homicide rate. One of the things I noticed is that the conversion rate of aggravated assault to successful homicide stopped falling much after 2000."

Your argument assumes all aggravated assault is attempted homicide. That assumption is untenable.

Art Deco said...


What is our hard data? Our hard data is not that "crime is falling," but that crime reports are falling.


Moldbug,

The Bureau of Justice Statistics makes use of victimization surveys. They have surveys of households and of certain trades like funeral directors. One recent metric published (in May 2013) was a series of descriptive statistics for "criminal victimizatoins" using firearms (that is, brandishing and injuries done). They find a 75% decline over a period of 18 years, from 7.3 per 1000 in 1993 to 1.8 per 1000 in 2011. About 3/4 of incidents involve brandishing. The survey data incorporate observations consistent with the aforementioned CDC data, that does show a modes increase in non-fatal firearm injuries in the latter half of this period (2001-11).

Anononymous said...

Moldbug:
"Average American."

The average American has 0.98 testicles.

Dahinda said...

I always wonder why Chicago gets such a bad rap. Sure there are shootings in Chicago but in terms of violent crime Chicago never makes it into the top ten in most lists. For example http://www.businessinsider.com/most-dangerous-cities-in-america-2013-6#1-flint-mich-25
For some reason the MSM has glommed on to the "Chicago is the shooting capitol of the country" narrative, probably begun from the traditional New York animus toward Chicago combined with some sort of strategy of gun control nuts, and now Chicago is known for shootings. I visited Chicago last week and was all over the North and West Sides and never heard a shot. My wife and I were amazed at how alive and cleaned up the city was and how many neighborhoods that were run down 25 years ago were now as alive as Naperville or Arlington Heights!

Anonymous said...

Medical advances mask epidemic of violence by cutting murder rate
Roger Dobson
BMJ 2002;325:615 (21 September)
bmj.com
ncbi.nlm.nih.gov

...Without this technology, we estimate there would be no less than 50,000 and as many as 115,000 homicides annually instead of an actual 15,000 to 20,000...

The aggravated assault rate was, by 1997, almost 750% higher than the baseline figure. The team also described the dramatic overall decrease in trauma mortality in the second half of the 20th century...

The period of greatest change came between 1972 and 1977, on the heels of the US involvement in the Vietnam war, which triggered big advances in trauma care.

beowulf said...

"Again, rape, aggravated assault, and tracked property crimes have also seen declines in their frequency (the latter by 27% over 20 years)."

These numbers can be fudged by cops, homicide numbers can't be because a doctor must sign the death certificate which in turn is sent to Uncle Sam (after all, the Social Security Death Index isn't self-reported).

The one caveat is doctors can only tag bodies they can bag. People who go missing and are never found Jimmy Hoffa-style aren't counted in homicide stats.

Anonymous said...

We can agree, for example, that Manhattan (Lower Manhattan) has much less street crime than it had 20 years ago, because it has many fewer AGBAs and consists almost entirely of AHWAs - whose microscopic violence level has perhaps diminished to the femtoscopic. We can even agree that the violence level of AGWAs, Average Guido-White Americans (meaning guineas, Bohunks and the like) has diminished considerably over the last 50 years.

These inferences about the violence levels of AHWAs, AGWAs, AGBAs, etc., are statistical inferences based on data about them. So I'm not sure why you dismiss statistics altogether.

Maguro said...

Obviously the authorities have incentives to fudge the crime numbers these days, but didn't they have those same incentives 20, 30, 40 years ago? Any reason to believe that they're fudging the numbers more now than they did back then?

Moldbug said...

The Bureau of Justice Statistics makes use of victimization surveys. They have surveys of households and of certain trades like funeral directors. One recent metric published (in May 2013) was a series of descriptive statistics for "criminal victimizatoins" using firearms (that is, brandishing and injuries done). They find a 75% decline over a period of 18 years, from 7.3 per 1000 in 1993 to 1.8 per 1000 in 2011. About 3/4 of incidents involve brandishing.

You've constructed, I suppose, some kind of magical distinction between a report and a survey. Which leaves the former easily manipulated, but confirms the usual papal infallibility on the latter.

I find these numbers wholly incredible. It does not particularly impress me that you got them from the high and holy U.S. Government - which as we know has neither the motive, the opportunity, nor the propensity to gank a number. Oh, wait, it does. I find the hospital gunshot wound statistics entirely credible, using much the same reasoning.

Or do you have some scenario under which they could both be right? AGBAs, perhaps, are drawin' less but shootin' more? The mind boggles.

Moldbug said...

Dahinda, spend some quality time at Second City Cop.

These inferences about the violence levels of AHWAs, AGWAs, AGBAs, etc., are statistical inferences based on data about them. So I'm not sure why you dismiss statistics altogether.

Because these patterns are real. When you picture an AGBA or an AHWA, you have a "stereotype" (ie, pattern you're not supposed to recognize) in your head. When you picture an AGBA or an AGWA of 30 years ago, ditto. When we say, "across the last 30 years, AGBAs have become more civilized," we have a meaningful statement that can be assessed intuitively.

You can aggregate the national statistics of Tibet and Morocco into one pseudo-country, Tibocco. Your statements about the "Average Tiboccan" are precisely as valid, mathematically speaking, as agnostic's statements about the "average American." But historically speaking, anyone who talks about the "average Tiboccan" is speaking gibberish. We may argue whether or not the Moroccan or Tibetan data sets need to be broken down further into more cohesive local categories, but no one can argue for making inferences about Tiboccans. Capisce?

Or do you think AHWAs and AGBAs have more in common, becides mere geographic proximity, than Tibetans and Moroccans? Hm, that's not clear to me. To say the least.

But, when you speak of "average Americans," you are in a sense making this untenable argument subliminally. The assumption that all Americans have something meaningful in common is, in a sense, built into your database. It's certainly not a fact-based assumption.

Moldbug said...

Obviously the authorities have incentives to fudge the crime numbers these days, but didn't they have those same incentives 20, 30, 40 years ago? Any reason to believe that they're fudging the numbers more now than they did back then?

Our ancestors had some of the same incentives. They had fewer computers, however, and more scruples. Bureaucratic mendacity is a kind of technology itself, and has its own Moore's law in a sense.

For instance, I wouldn't be so quick to assume that homicide numbers can't be managed. Sure, it's hard. So is heavier-than-air flight...

Art Deco said...


These numbers can be fudged by cops, homicide numbers can't be because a doctor must sign the death certificate which in turn is sent to Uncle Sam (after all, the Social Security Death Index isn't self-reported).


Take it up with the chaps on the previous thread who told me homicide numbers are invalid because of medical advances. (By the way, the homicide rate has declined by half in 20 years, and by more than 3/4ths in New York City).





Power Child said...

So, this means the last remaining defense of the war on drugs that makes any kind of sense (that it provides a handy proxy for cops to keep bad guys off the street, thus making us safer as evidenced by a lowered crime rate) is bogus?

James Kabala said...

"but has the Bronx, or Queens, or other unglamorous boroughs?"

Are the lawyers at the Bronx courthouse still afraid to go out for lunch, as depicted in Steve's favorite book? Do they still move their cars to safer spots after sundown? I don't know about the Bronx specifically, but the pervasive fear of that type felt in many American cities twenty-five years ago seems certainly felt in many fewer today.

Silver said...

"Medical advances mask epidemic of violence by cutting murder rate
Roger Dobson
BMJ 2002;325:615 (21 September)
bmj.com
ncbi.nlm.nih.gov "


It's the same untenable assumption that aggravated assault = attempted homicide. In some cases it might, but not in all, which is what the argument requires.

Silver said...

"Crime has been falling among certain populations and increasing among others. The actual reality is the balance between these two trends which varies from place to place. The perceived reality is something different."

How can you be sure of any of that without recourse to statistics?

Anonymous said...

interesting that Moldbug apparently appears to believe that statistics have been manipulated to show a downwards trend during the "20th century".

But apparently these manipulators forgot to do this through both the rising crime periods of 1900 - 1933 and 1963 - 1993.

What bizarre democracy caused peccadillo must be responsible for this phenomenon. Surely the answer lies in a wedge of Carlyle?

Anonymous said...

San Diego police stats indicate that per capita violent crime in that city has increased 9x from 1950. It would not surprise me if the nation has more or less followed that course.

Eric said...

More objectively and quantitatively, consider real estate prices. Would they be booming in cities like New York or San Francisco if crime there were on the rise?

I think there's a virtuous spiral in effect here. Low-lifes are being priced out of major metropolitan centers, so crime goes down. This makes the area more desirable and thus more expensive, pricing out even more criminals.

Eric said...

However, the average round of the same caliber probably is more deadly today. So the question is whether the most common calibers have gone down in size, or whether they have stayed consistent.

The standard police round used to be .38, which has just about been completely replaced in favor of more powerful 9mm or .40 S&W now.

Legion 6 said...

"In 2010, 13.96% of U.S. shooting victims died, almost two percentage points lower than in 2007."


-2007 was before the crash; maybe people on average are buying (cheaper) low caliber pistols and bullets, which will do less damage, or conserving ammo by poppin a few less caps in punk *ss fools.

I know for certain that the price of ammo has been going up recently because the feds have been buying enough to take Detroit singlehandedly if they wanted.

So maybe its not that far of a stretch- certainly within a reasonable interpretation of the laws of supply and demand.

Anonymous said...

http://www.nytimes.com/2013/07/23/world/americas/prices-fuel-outrage-in-brazil-home-of-the-30-cheese-pizza.html?_r=0


Wow.

Art Deco said...


You've constructed, I suppose, some kind of magical distinction between a report and a survey. Which leaves the former easily manipulated, but confirms the usual papal infallibility on the latter.


No, I have not. I merely note that one is constructed from a census of reports from local law enforcement and the other is a survey of the public. Two different methods and sources, same data.

Give it up buddy. You don't know what your talking about.

Anonymous said...

http://blogs.the-american-interest.com/wrm/2013/07/23/note-to-paul-krugman-it-took-more-than-markets-to-ruin-detroit/

"The soft bigotry of low expectations meant that neither federal nor state prosecutors intervened until very late as the thieves looted the ruins."

Oh no, not this crap again.
The soft zealotry of high expectations make conservatives believe that we can expect anything better from a city that is 90% black.

Anonymous said...

You can aggregate the national statistics of Tibet and Morocco into one pseudo-country, Tibocco. Your statements about the "Average Tiboccan" are precisely as valid, mathematically speaking, as agnostic's statements about the "average American." But historically speaking, anyone who talks about the "average Tiboccan" is speaking gibberish. We may argue whether or not the Moroccan or Tibetan data sets need to be broken down further into more cohesive local categories, but no one can argue for making inferences about Tiboccans. Capisce?

This analogy is incorrect since in the US we're often talking about violent crime rates in cities i.e. in specific geographic locations with heterogeneous populations but with mobility and without strict segregation.

Dave Pinsen said...

"I think there's a virtuous spiral in effect here. Low-lifes are being priced out of major metropolitan centers, so crime goes down. This makes the area more desirable and thus more expensive, pricing out even more criminals."

The virtuous cycle continues when the criminal-prone move out to the suburbs or exurbs. With everyone driving cars, there's less opportunity for muggings and other crimes. Also, consider the proliferation of cheap consumer goods and credit cards. Why break into someone's house and risk getting shot or going to jail for a flat screen TV when you can just buy one at Walmart or Target on a credit card?

a Newsreader said...

You can aggregate the national statistics of Tibet and Morocco into one pseudo-country, Tibocco. Your statements about the "Average Tiboccan" are precisely as valid, mathematically speaking, as agnostic's statements about the "average American." But historically speaking, anyone who talks about the "average Tiboccan" is speaking gibberish. We may argue whether or not the Moroccan or Tibetan data sets need to be broken down further into more cohesive local categories, but no one can argue for making inferences about Tiboccans. Capisce?

This analogy is incorrect since in the US we're often talking about violent crime rates in cities i.e. in specific geographic locations with heterogeneous populations but with mobility and without strict segregation.


This analogy is indeed correct in that it explains how it is necessary to look more closely at the details (see Simpson's Paradox). That these details are even more confused by migration patterns and the mendacity of the official statistics doesn't help much either.

The point is that the statistics we have aren't very helpful here, and we will need to rely on other methods to figure out what is really happening.

Anonymous said...

This debate is difficult because you're looking at the *net* effect of different movements in opposite directions and lots of politically motivated distortions.

Because of that i would say the only safe data is medical data which where available shows violence going up dramatically.

.

In simple terms this is what has been happening across the western world imo.

Violent crime has been falling in x neighborhoods and increasing in y neighborhoods where x > y so the *net* effect has been falling overall crime.

However the number of neighborhoods where crime is decreasing x, is itself decreasing while the number of neighborhoods where it is increasing y, is increasing so although on the surface overall crime is decreasing there is a disguised underlying trend. When x and y cross over crime will head dramatically upwards.

It's a bit like the way a tsunami starts with the tide going out - the lull before the storm.

Most western countries are in the lull part with the US close to the point where it turns into a storm.

.

At the same time just to further confuse the issue, in cities like New York, LA and DC the y neighborhoods are being displaced to middle-class suburbs outside those cities.

.

Apart from medical data another metric that shows this effect is the proportion of prisoners from immigrant groups.

If you take a country like France or Sweden the x neighborhoods are mostly indigenous while the y neighborhoods are those containing particular immigrant groups with high violent crime rates. A graph of immigrant prison inmates as a percentage of the total would serve as a reasonable proxy for displaying the underlying trend between x vs y neighborhoods.

.

The pro-immigration lobby has to hide the underlying trend else the voters would freak.

Anonymous said...

Dave Pinsen

"The virtuous cycle continues when the criminal-prone move out to the suburbs or exurbs."

No it doesn't.

Moldbug said...


No, I have not. I merely note that one is constructed from a census of reports from local law enforcement and the other is a survey of the public. Two different methods and sources, same data.

First of all, the data isn't the same. Your survey data takes a bizarre knee-bend where it declines enormously from 1993 to 2001, then stays flat from 2001 to 2011. What happened? 9/11?

Second, the sources aren't different at all. The proximate source is the same: bureaucrats who have a vested interest in lower crime numbers. You're telling me it's impossible to gank a survey? You're telling me that if you were working on the NCVS, and you wanted the numbers to come out lower because it would advance your career, you couldn't find a way to help?

The hospital data is much more trustworthy both because it confirms what our eyes tell us (gangbangers have not become less gangbangy), and because it is produced by agents without an institutional interest in any particular outcome.

It's really unfortunate that I have to spell this kind of reasoning out to an apparently intelligent person.

Anonymous said...

"It's the same untenable assumption that aggravated assault = attempted homicide."

It's not that. It's that anything that could end up as a homicide, if it doesn't, will be contained within the category of aggravated assault.

Say 10% of serious woundings end in homicide (intentional or not) and you have 100 serious woundings then they end up categorized as
- 90 aggravated assaults
- 10 homicides

Say the mortality rate drops to 5% for the same number of serious woundings then in theory the stats would change to
- 95 aggravated assaults
- 5 homicides

However in districts where there's a lot of political pressure to fiddle crime stats the drop in homicides can provide cover to reduce the other stats too i.e. a 50% drop in homicides makes it possible to reduce aggravated assaults by 50% without it looking strange, down to for example
- 45 aggravated assaults
- 5 homicides

.

"How can you be sure of any of that without recourse to statistics?"

You can't but experience can tell you when and where stats are more likely to be fiddled.

Over the last 30 years violent crime in white neighborhoods has been going down because an aging population has fewer young men (and maybe other reasons as well) so there is no need or incentive to fiddle those stats.

(There is an exception here for the rustbelt underclass where it's been going up again.)

At the same time there has been increasing violent crime among some specific first-generation immigrant groups and pretty much all settled immigrant groups who don't have the ability to get out of the underclass in a high-tech world.

(The last point is critical.)

These increases are politically very incorrect and there has been a lot of effort to disguise it.

For example, every police district in the western world which overlaps with an area of underclass has a sergeant whose job it is to downgrade the local equivalent of aggravated assault to assault.

.

"Obviously the authorities have incentives to fudge the crime numbers these days, but didn't they have those same incentives 20, 30, 40 years ago?"

The number of neighborhoods where there's a reason to fudge the figures has increased as a percentage of the total.

.

"So, this means the last remaining defense of the war on drugs that makes any kind of sense (that it provides a handy proxy for cops to keep bad guys off the street, thus making us safer as evidenced by a lowered crime rate) is bogus?"

I don't think it's entirely bogus as things would be worse otherwise.

However i think there is a big flaw with the current setup in that a lot of criminals fade after their mid 20s which means keeping minor drug dealers in jail past that age probably doesn't help much at all while at the same time by keeping the jails full it leads to a revolving door for more violent criminals.

So i think re-arranging who is in prison by taking into account the mid-20s effect could have a much greater impact.

Silver said...

ArtDeco,
"Take it up with the chaps on the previous thread who told me homicide numbers are invalid because of medical advances. "

Criminologist Eric Monkkonen in his highly-praised "Murder in New York City" guesstimates that some 30% of homicide victims' lives may have been saved 100 years ago had modern medical methods been available, based as much on response time as anything else. He admits he lacks the expertise to make a definitive statement, though.

Anonymous said...

Migration patterns indeed. Boston has a higher murder rate than Tampa. Boston should have priced the riff- raff out of town long ago (highest rents for a 1-bedroom apartment in the country). Is its crime fed by Boston's lack of concealed carry or the its generous welfare benefits, including the $1,400/month Section 8 housing voucher?

Power Child said...

@Anonymous of 7/24/13, 12:14 AM:

I don't think it's [the "Proxy" as a reason for maintaining the War on Drugs] entirely bogus as things would be worse otherwise.

If we legalized drugs, both cartels and the street gangs who retail for them would lose their major revenue stream. (Yes, they have others but none come close to illegal drugs.) Cartels would shrink as they adjusted to those backup revenue streams. In their current incarnation, gangs would have a halflife as their ex-drug dealer members aged (keeping the drug dealer persona alive with them), but then gangs would go back to how they were once upon a time: basically just groups of adolescents who got into mischief. Would that really be worse?

And would it be worse if the 4th through 10th Amendments were once again given more meaning than just as words on paper?

Or if the government didn't have to process (not jail, necessarily, but process) all those harmless white kids who get busted for smoking a little pot or dropping some ecstasy? Etc.

However i think there is a big flaw with the current setup in that a lot of criminals fade after their mid 20s which means keeping minor drug dealers in jail past that age probably doesn't help much at all while at the same time by keeping the jails full it leads to a revolving door for more violent criminals.

So i think re-arranging who is in prison by taking into account the mid-20s effect could have a much greater impact.

In the current setup, people go to jail for minor drug offenses but then leave jail as hardened criminals. Usually more racist too. It's been that way for a long time, too.

Are you suggesting that the penalty for being caught with drugs at age 20 should be different than being caught with drugs at age 30 or 40?

Anonymous said...

"If we legalized drugs...Would that really be worse?"

That's a 3rd option. My original comment was about the use of the war on drugs as an indirect way of hoovering up violent criminals. I said it does partly work.

.

"In the current setup, people go to jail for minor drug offenses but then leave jail as hardened criminals."

Disagree completely. People who are soft going into jail will come out soft and vice versa.

.

"Are you suggesting that the penalty for being caught with drugs at age 20 should be different than being caught with drugs at age 30 or 40?"

Previous convictions.

Anonymous said...

“I always wonder why Chicago gets such a bad rap. Sure there are shootings in Chicago but in terms of violent crime Chicago never makes it into the top ten in most lists.”

Those lists should be viewed with a huge bag of salt.

“Because the Chicago Police Department tallies data differently than police in other cities, the FBI often does not accept its crime statistics. Chicago police officers record all criminal sexual assaults, as opposed to only rape. They count aggravated battery together with the standard category of aggravated assault. As a result, Chicago is often omitted from studies such as Morgan Quitno's annual "Safest/Most Dangerous City" survey, which relies on FBI-collected data.”

http://en.wikipedia.org/wiki/Crime_in_Chicago#Policing

“For some reason the MSM has glommed on to the "Chicago is the shooting capitol of the country" narrative… I visited Chicago last week and was all over the North and West Sides and never heard a shot.”

Shootings are primarily confined to black neighborhoods. Outside of those areas, shootings are not common. However, other crimes such as robbery have become more common in previously safe areas. I am guessing you did not spend much time in predominantly black areas.

Anonymous said...

http://www.theamericanconservative.com/american-pravda-liberal-bias/#more-89180