Reported Crime May Be Falling At A Historic Rate In The US

How does 2025 stack up compared to other years?

Have you listened to last week’s Jeff-alytics podcast yet? It’s a great conversation with Laura Arnold, co-founder of Arnold Ventures so please give it a listen and subscribe for future episodes! Now, onto the newsletter.

Murder is falling at the fastest pace ever recorded in 2025. Data from the Real-Time Crime Index shows murder down around 19 percent through August in a sample of more than 550 agencies covering nearly 115 million people nationwide.

And it's not just murder that is down a lot as seen in the latest data on reported crime from the RTCI:

We have more or less complete national estimates from the FBI back to 1960 and the largest decline in murder recorded so far was last year’s 14.9 percent drop which may get revised to be slightly smaller. Before that it was a 9.3 percent drop in 2023. Before that it was a 9.1 percent reduction in 1996.

We won’t know exactly where this year will end up and how well the RTCI will predict the FBI’s estimates, but the gap between 15 percent and 20 percent is large enough to feel reasonably confident in a third straight year of record murder decline this year.

But what about the drop in other crimes?

There’s some uncertainty give that the year isn’t over and the RTCI is probably overestimating the declines by at least a little bit due to underreporting. The RTCI can give us the direction of crime trends with a lot of certainty, but the exact change is still a bit up in the air.

We can safely assume that reported crime is down a lot pretty much everywhere though, even if the exact numbers may bounce by a few percentage points around in the coming months. So, even if they're not set in stone, how would the declines in crimes other than murder as seen in the RTCI compare in a historical sense?

As I said, we have relatively decent national estimate data back to 1960 from the FBI. The FBI has been publishing the Uniform Crime Report since 1930, but reporting gets progressively more incomplete the farther back you go. Going back to 1960 shows just how much crime is falling in 2025 (per the RTCI) compared to the last 65 years.

Overall violent and property crime are both on track to have the largest one-year declines ever recorded. The drop in reported violent crime is being driven by enormous drops in murder and robbery. I’ve written about the former a lot, and a large drops in robberies makes sense given that carjackings — a subset of robberies — are falling a lot too.

Overall property crime is falling driven by similalrly enormous drops in burglary and auto theft. Theft is down the second most ever recorded — per the RTCI — eclipsed only by the drop in 2020 when everything was closed.

Crime is not guaranteed to decline at a record clip across every crime type in 2025, but we are likely seeing very large drops in every measured category of crime. Simply put, the changes shown in the RTCI point to drops in reported crime that are very large by historical standards and may feature multiple records broken when all is tallied and reported.

But what about if we try and go back before 1960? The FBI has scanned PDFs of each Uniform Crime Report dating back to 1930. As I mentioned, the reporting gets less complete the earlier you go. From nearly complete in the early 1960s to around 80 percent of the country covered in the 1955 Uniform Crime Report…

To city-only reports in the 1940s covering increasingly smaller population groups as you approach 1930.

Going back before 1960 is somewhat silly from a data comparability perspective. The percent changes from 1932 to 1938, for example, are drawn from a sample of just 73 cities of 100,000 or more that reported data. And overall violent and property crime counts aren’t really available.

Still, the changes we are seeing in 2025 so far hold up as extremely large by historical standards with the understanding that the pre-1960 stuff isn't exactly apples to apples.

The only larger declines were in theft in 1943 and 2020 and aggravated assaults in 1932 when reporting was far from complete.

The US will likely see the largest one-year drop in murder ever recorded in 2025. Again. Will the same be true of overall reported violent and property crime? The evidence so far certainly seems to suggest it, but the exact degree of this year's drops won't be known for certain for a few more months.