Thread by @JohnFPfaff

It's hard to care right now for a lot of reasons, but:

The 2019 Uniform Crime Reports just posted.

The takeaway:

• Murder up baaaarely--by 0.3% (51 cases)
• Rape (-3%) and robbery (-5%) down
• Agg assault up--by 1%

• All prop crimes down, tot pop down 4%.

Of course, these stats are:

• Pre-COVID
• Pre-lockdowns (formal and unofficial)
• Pre-Floyd protests, pre-"Defund"-going-national

They are a picture of a past that has little bearing, it seems, on the present.

I tried to see if any improvements happened with clearance rates (percent of reported crimes ending in arrest), but no data there yet.

Arrests are down about 2% from last year.

Serious violence arrests down 5%
Serious property arrests down 8%
Drug arrests down 6%

Marijuana's share of drug arrests fell as well:

Trafficking: 3.3% to 2.9%
Possession: 36.8% to 32.1%

(Taking UCR stats out to the tenths place is pretty preposterous, though. Trafficking is flat. Possession may have dropped, at least likely didn't rise.)

So this is interesting: that small increase in homicide was driven by the South. All other regions saw declines in murder, but that slight national uptick (or the absence of a real decline) is due to a jump in the South.

If you break it out by Southern states, looks like a lot had increases, so not just one state that is driving that.

But Mississippi saw a 55% (!!) increase in homicide last year. What is going on in Mississippi?

(Yes, baserates a bit, but that seems like more than that.)

Of the ~86% of all murder for which there is additional data, 74% of them were committed with a gun. Unsurprising, likely about the same as in previous years. Just... putting that out there.

ucr.fbi.gov/crime-in-the-u…

Jumping back to arrests, this is a striking statistics: total arrests* are down 21% since 2010.

Juvenile arrests have been cut by more than HALF. That's a real change, and that's a real, meaningful improvement.

* Here's a good example of how our crime stats are so much fun.

How many arrests were there is 2019?

10M?
6.9M?
7.6M?

Reflects the inconsistency in reporting certain topics.

But keep that in mind when ppl stress small changes: there is SO much noise here.

But to go back to that juvenile arrest statistic. It's good, but... we are who we are.

In 2019, we arrested over 2,500 kids UNDER THE AGE OF TEN.

Only 177 for serious violence, and 736 for other assaults.

That leaves over 1,600 kids UNDER TEN arrested for NON-VIOLENCE.

We arrested nearly half a million kids under 18, and over 150,000 kids under the age of 15.

Most of these arrests would simply be impossible anywhere else.

We have a column for "under 10." It's so routine that it is part of our formal reporting practices.

Who did we arrest by race, compared to their percent of the population?

White: 69% of arrests, 76% of the pop*
Black: 27% of arrests, 13% of the pop (that's 2x)
Hispanic**: 19% of arrests, 19% of the pop.

* I'm using White, both H and Non-H here, since the arrest data doesn't separate them out. Guessing % White-Hispanic in arrests is bigger than in the pop as a whole.

** Hispanic crime data is notoriously worse than the rest of it (note the even-smaller arrest #). So... be wary.

And now to jump directly into the always-ugliest part of this data: race and homicide.

Numbers seem similar to past yrs:

Most murders are intraracial for both whites and Blacks, but slightly higher for Blacks.

For cross-race homicides, rate is higher for Black offenders.

BUT!

Before the predictable racist takes emerge:

Most murders involve ppl who know each other, and our social circles, like our nbhds, remains highly segregated.

But what about stranger cases?

Well, whites remain more segregated than Blacks, and there are more whites out there. So?

So, if a white person is "looking" to commit a random murder, they are disproportionately more likely to run into another white person than a Black person when compared to the chance that Black person "looking" to commit a random murder will run into a white person.

In other words, the offender-victim disparity in cross-race homicides is to some very real degree a reflection of the nature of racial segregation.

Given numbers and segregation patterns, it is simply "harder" for a Black person to kill a white person than vice versa.

But, of course, the fundamental story is that almost all homicides are intraracial. White-on-white and Black-on-Black.

Because homicide tends to be a more "personal" crime, and our social circles, social lives, dating circles, etc., are all highly segregated.

But in the end, what do we do with any of this, by which I mean the entire 2019 UCR?

It seems like the historical record of a bygone era.

The upheavals of 2020--COVID and crim justice protests/reactions alike--make it seem silly to draw inferences from this for today.

NYC's CompStat data suggests that the generally calm nature of 2019 had been extending into Q1 of 2020, and then everything jumps around: murders go up a bit, shootings more (tho dropping now), robberies down.

2020 is its own thing. 2021? Who knows at this point?

But... at least here, at the end of September 2020, we have a picture of 2019 (seriously, labor economists? I never EVER want to hear you complain about delays in your data, like, never).

Remember 2019? When we flew places? And other pedestrians didn't scare you? This was then.

Oh, clearance data is finally up (though the graph is broken, a fitting metaphor for our clearance rates).

Nothing has changed really at all. Clearances are low--under half of all reported violence ends in an arrest, and under 1 in 5 property crimes--and stable over two decades.