Contextualizing the Largest Decline in Murder Ever Recorded

Murder is down so far in 2023, how does it relate to previous years?

A piece I had been working on for a while finally came out this week in The Atlantic detailing the drop in murder occurring nationally so far this year. Longtime Jeff-alytics readers should not be surprised as I’ve written on this subject a few times, but you should all read the piece regardless!

One thing that I was not able to include in the piece is greater context about what the potential 2023 drop in murders (-12.2 percent YTD as of early June) could mean in terms of historic national murder trends.

First off, let’s start with some assumptions:

• Let’s assume that our big city sample will continue to overstate the national trend, so the national trend will be a slightly smaller increase if big cities are up and a slightly smaller decrease if big cities are down. This one has been pretty uniformly true for this entire century.

• Let’s assume that the FBI’s estimate of murder rising 4.3 percent in 2021 is accurate. It matches what the big city data sample shows (roughly a 6 percent increase) and comes from a decent sized NIBRS sample that should be reasonably accurate.

• Let’s assume that murder fell 4 percent in 2022. Murder was down 5 percent in our sample of 93 cities with data for 2021 and 2022 which works out to a 2-4 percent drop nationally. Let’s be optimistic!

• Let’s assume that murder falls 10 percent in 2023. That’s still a big IF. The current trend may or may not hold up, but if the big city sample is -12 percent at the end of the year then a 10 percent decline nationally would be par for the course. We won’t know for sure until NIBRS data is released — likely weeks before the 2024 Presidential election which will be somewhat amusing — so this methodology is the best way of making an informed guesstimate in something approaching real-time.

If all of these assumptions are true, then we get the below picture of murders nationally from 1960 to present.

Well down from 2021, still well above 2019 — to say nothing of 2014’s historic low murder rate. And here’s the same graph but murder rate per 100k because there is a tendency to get yelled at when not producing both count and rate.

In The Atlantic piece I make the claim that a double-digit decline in murder would be the largest decline ever recorded. The veracity of this claim can be seen in the below graph showing the percent change each year since 1961 which is the first year that the FBI produced a national murder count.

As you can see, double-digit increases have occurred occasionally, including the massive 29 percent increase in 2020, but the 9 percent decline in 1996 is the largest so far. There has never been a double-digit percent decline in murders recorded nationally from one year to the next. The same is true if you look at the change in the number of murders from one year to the next.

A 10 percent decline in murder in 2023 would mean over 2,000 fewer people murdered this year relative to the year before, the first time that feat would have been accomplished on record.

The year is not even halfway over, and there is no guarantee that the current trend will continue for the rest of the year. If it does continue though then there will be ample reason to believe that the US has experienced one of the largest declines in murder ever recorded.