This journey started on Twitter, with this:
Gascon is the DA for LA County, so of course he’s going to say stupid, self serving things… But while I can’t get particularly upset with him for it, I can still point out the obvious.
First off, I reject the premise of the question: We do hear about those cities. The media obviously covers them. But if the gripe is why don’t we hear about them as much as we hear about LA, then he answered his own question: Because LA is bigger, the numbers are bigger, and editors have decisions to make.
LA Experienced 355 homicides in 2020, up from 258 in 2019, which was a relatively low water mark over the last decade. Tulsa, on the other hand, experienced 78 homicides in 2020, which is up from the 58 in 2019.
So, really, you’re an editor: What’s the story? The 355 homicides in LA, representing a 100 death increase in raw numbers and a 37% increase year over year or the 78 homicides in Tulsa, representing a 20 death increase in raw numbers and a 34% increase year over year?
But More Than That, Was He Even Correct?
Ish. 355 homicides in a city of 4 million is 8.8 homicides per 100,000. That’s less per capita than Jacksonville at 19.78 (178 in 900,000), as an example, also less than three times the rate. To be fair, he wasn’t off by much.
It is important, however, to contextualize this: Gascon’s point only functions if you focus on the year 2020, and the homicides per capita number. Why? Because the 2021 numbers aren’t going to return the same results… Tulsa’s homicide rate decreased to 62 in 2021, Jacksonville’s decreased from 178 to 128. LA’s increased to 397.
So, really, you’re an editor: What’s the story? The 30% decrease in homicides in Jacksonville, or the 11% increase over the highest homicide year this side of the century in LA?
So I Did What I Do.
I Tweeted. This is almost always a mistake, and I tend to get suck-holed into some absurd argument or another, but it’s just so damn addicting somehow.
You can follow the discussion from there, but my general points were:
The only three American cities that are tracked as part of the 50 worst cities for homicide rates globally are St. Louis, Baltimore, and Detroit. (Four if you count San Juan, take that for what you will.)
Picking a couple of cities with Republican mayors and comparing them to a couple of cities with Democrat mayors with similar stats doesn’t really interact with how most cities with high homicide rates, and most of the cities with the worst homicide rates, are led by Democrats.
The data was old, and newer data would be worse for his argument.
There are decent arguments that someone could make against these points. For two, for instance, you could argue that the homicide rates have more to do with poverty than governance, and that people in poverty tend to vote Democrat. It’s not a bad argument, there’s more to it than that, and I’d be willing to have that conversation. Twitter didn’t make that argument.
I was told, in order:
That all major news organizations talk specifically in terms of crime from the last year the FBI released stats for.
That Gascon made the point that LA was safer than most, if not all Red Counties, and that he was correct.
That crime in Democrat led cities in red states is the fault of Republicans, because they control the laws.
That Gascon could not have reported numbers from 2021 because that year wasn’t done.
That I was missing the point, and the data for counties was significantly different than the data for the cities and states.
With the exception of 2, they’re all facially absurd positions. Some comically so. 2 is still absurd in that that’s absolutely not what Gascon said, and not what was in his link, But the question tickled something in the back of my mind; Because of the reality of the law of large numbers and the relative population density of Republican counties, it could be possible for that to be true, at least in enough red counties to make “most” true. This was the tweet in question:
I pressed him. Again, view the conversation for yourself. In defense of his argument, he presented a DOJ information package about 2020 crime statistics in California.
Spoiler: It doesn’t say what he thought it did.
Frankly, it was so obvious that it didn’t say what he thought it did, that I thought that I had missed something, or that he linked the wrong information. Eventually, I realized he just hadn’t read it and had no capacity to do so. I pressed. “Page 19” he says. Well… Page 19 is the information by county, but the homicide rate per 100,000 is 6.8 in blue LA County and 0 in Red Sierra, so what, exactly, was he looking at? Already obvious is that it’s not true that LA is safer than “all” red counties, not even if you’re only looking at California. He sputtered like a 100 year old combustion engine and eventually shut down.
But I wondered, even if he wasn’t going to make the argument, did “most” red counties in California have a higher homicide rate, per capita, than LA county?
Spoiler: No.
I parsed the data. I grabbed the population data for each county from here. I determined whether a county was blue or red by their 2020 presidential election votes, and the homicide rates from the DOJ packet above. Two things immediately stood out:
The DOJ wasn’t calculating a homicide rate per 100,000 for counties with less than 100,000 people unless there were more than 5 homicides in the county. Because lower population counties tended to vote Republican, this artificially understated homicide rates per 100,000 in Republican counties. This isn’t a bad thing in an of itself, a single homicide in Alpine County (population 1100) would have resulted in a rate of more than 90 homicides per 100,000 people, but for transparencies sake, I manually calculated the rate for all counties. My math for the most part matched the DOJ’s, all within .5 homicides per 100,000 population, and you can check the math by downloading the sheet from below. I assume what differences there are to be the result of a different population number.
For counties with very low net homicide rates, the increase/decrease percentages were going to be extreme. Sutter country, as an example, had one homicide in 2018, and four in 2019. That’s technically a 300% increase. That’s a real number, but for clarity, not all rates are the same. A 300% increase in LA would have been 1800 people.
The Results!
LA County has a population of 9.9 million people, and experienced 677 homicides in 2020. That works out to a rate of 6.81 per 100,000. The DOJ’s number was 6.7. Again, I assume their population number was different than mine, for the sake of argument, I’ll use theirs because it is lower.
There are 23 counties in California that cast more votes for Donald Trump than Joe Biden. Of those 23, the DOJ did not calculate a homicides per 100,000 rate for 16 of them, so I’m forced to use my numbers (which tended to be higher anyway, so this is the generous way of doing it). Five still did not have a single murder in 2020. Eleven more had rates less than LA county, the remaining six had higher rates.
So, no, “most” red counties do not have higher homicide rates than LA county. In reality, very few do.
And that’s the most generous interpretation, not only did I use the lower DOJ number for LA and my slightly higher calculation for the Republican counties, Trinity county had two homicides in their 11,000 person area, and that technically saddled them with a rate of 17.5. Glenn county also had two homicides in their 28,000 slice of heaven, giving them a 6.74, which I counted as technically higher than LA, at 6.70. So take that for what you will. There are only four counties with significantly higher rates than LA, and only one of those, Kern county, had more than ten homicides.
Interestingly, when you add the figures for all of the counties by party, Republican areas do tend to have a slightly higher rate. The difference is marginal, less than 1 per 100,000 in any given year, but it’s there, and the data is what it is.
In mitigation of that, like I said above, not all rates are equal. If I had to choose whether to roll the dice with the 700 annual homicides in LA county, or the 10 or so in King’s county, I’d move to Hanford and feel good about it. It would also be interesting to run these numbers again once the DOJ releases 2021 figures, because it would only take about 30 fewer homicides in Republican counties to tip that scale, and like I said above, LA had a banger, pun intended, of a 2021.
Spreadsheet and Formulas: