Untangling the massive increases in homicides since 2020
Coincidence? COVID policies made us crazy? "Floyd effect"?
The last three posts investigated the increased deaths starting in 2020, looking at how many were attributed to COVID and what other causes of deaths have seen major increases starting in 2020. One of causes of deaths that increased significantly, even though not large fraction of total deaths, is homicides (by assault of some kind).
From CDC Wonder, I gathered the data from 2018-2023 on deaths by homicide by state and month. In the charts that follow, we are not including a handful of states with very few homicides as their numbers will not be reliable1. In particular, we exclude AK, DE, IA, MT, NE, SD, UT. Hopefully my readers trust me that this has no impact on the analysis that follows.
Homicides in 2018 & 2019 showed very little trend. Roughly 18K in 2018 and 18.3K in 2019 (1.5% increase). In what follows, we’ll compare the number of homicides in 2020 & beyond vs. the 2018-19 averages. In a way, we are calculating “excess homicides”.
Annual Increases
Starting with annual numbers, we see the following:
The grey bars represent the 2018-19 average of deaths by homicide, the blue are the additional homicides, with a % increase labeled at the top of the bar. Above the 18-19 average, we see 30%, 39%, and 32% increases in 2020, 2021, and 2022 respectively. A total of 18.7K additional homicides occured in those years above what would be expected if those years had the number based on the 18-19 average.
Was this a coincidence or did COVID (lockdowns, etc.) drive people nuts & increased homicides?
This is not a question we can answer definitively, since there were only a few months in 2020 that would not have been impacted by lockdowns, etc. However, to get some clues on this, we can look at how 2020 was trending before we could expect COVID policies to have had an effect vs. after. To do this, we will hypothesize that Q1 (Jan-Mar) was largely unaffected by COVID policies, while Q2-Q4 were impacted. Therefore, we will compare 2020 Q1 vs 2018-19 Q1 average, and 2020 Q2-Q4 to 2018-19 Q2-Q4 average. In addition, including 2021 again as well as 2022 Q1-Q3 (doing this because I am not confident 2022 Q4 reporting is complete…meaning the number 👆 for 2022 may look worse in future) :
While we only have 3 months to use, it would appear there was a major trend shift starting in 2020 Q2 when COVID policies likely starting having an impact on behaviors. While 2020 Q1 looked bad, a 9% increase over 2018-19 Q1 average, it was not nearly as bad as the rest of 2020 and beyond.
But wait, was it the COVID policies that impacted behaviors & homicides, or was it due to the aftermath of George Floyd’s death and subsequent changes in policing, etc.
Another question I can’t definitively answer, but we’ll use a similar method to see if we can generate clues. To do this, I will further split the data for 2020. We will have 2020 Q1 (pre-COVID), 2020 Q2 (COVID policies, before “Floyd impact” expected), and 2020 Q3 & Q4 (possible “Floyd Impact”):
From 1st pass, recognizing limited data, it would appear there were 2 step changes in 2020. First, in Q2 we jump from a 9% increase in homicides to a 27% increase, compared to same periods in 2018-19. Then, 2020 Q3 & Q4 show a 42% increase over same periods of 2018-19. Similar levels of increases persist through at least 2022 Q3.
Obviously homicides have had alarming increases in 2020-22. Was this an expected trend based on 2018, 2019, and Q1 2020. Almost certainly not to the degree we saw. Was this large increase then due to COVID policies making people crazy, or was it due to some effect (changed policing, or otherwise) post George Floyd?
My understanding is that if the number of deaths for a certain combination (e.g. State & month) is below 10, then the number is reported as zero. Therefore, if Montana had 9 homicides in 2018, but 11 in 2020, we would see 0 for 2018 & 11 for 2020, making it appear there had been 11 additional homicides, when in fact there were only 2 additional homicides. To ameliorate this, I am simply removing the states where this could be a potential problem. It does not affect the overall numbers in any significant way.
I've done some analysis re: homicides here:
https://marypatcampbell.substack.com/p/the-geography-of-homicide-states
That's by state.
No, I'm not looking at "Floyd effect" vs lockdown, etc., and I haven't spent a lot of time on homicide vs other causes because it's smaller than ones like drug ODs.
more here:
https://marypatcampbell.substack.com/p/video-us-mortality-trends-2020-2022-9e9
Interesting. But. One of the problems for us "Covid truth seekers" is that, whatever indicator you choose, there will always be counter-example countries (even if you restrict to "the West"). There has not been much of an increase in homicides in Germany, for example - but then, homicides are much rarer around here (the USA has ten times more homicides but only four times more people).