Crime Is Likely Down An Enormous Amount So Far In 2025

Sifting through the evidence early in the year.

I was a guest on Jerry Ratcliffe’s excellent podcast a few years ago and we talked about a variety of topics surrounding crime data and crime trends. At some point, the discussion turned to use of Year-to-Date to evaluate crime trends and Jerry, very vividly, exclaimed “Oh, I have thoughts about Year-to-Date.”

I think about that comment any time I try to write about guesstimating a year’s crime trends — especially before the World Series.

Jerry would probably rightfully tell you that YTD is unreliable, especially early in the year, and YTD frequently reflects something that happened last year rather than telling you what’s happening now. Think of it like a scoreboard of now rather than a forecast of the future.

It usually takes until a month that ends in ‘r’ before YTD is reasonably accurate at a citywide level for knowing whether crime is going to increase or fall in a given year. This is especially true when the tracked changes are reasonably small — a 6 percent drop in crime through September can very easily be an increase when all is said and done.

Where I think using YTD has value though is when collecting data from dozens or hundreds of agencies to communicate a larger trend as it’s developing. YTD can help as a measurement tool to set expectations and understand the direction even if caution is required to account for uncertainty. Having months and months of data from every agency would be ideal, but that isn't usually an option so YTD is frequently the only analytic game in town.

So, with that in mind, I’m going to ignore the voice in my head and write about crime in the United States through the first few months of 2025 because all of the available data points to a very large, across-the-board, decline in crime nationally that is worth wider awareness.

Before I get into the numbers, however, I’m going to attach three disclaimers to this analysis.

First, crime data is often prone to being underreported in the most recent handful of months. This issue gets fixed as the year goes on and won’t be a big deal by the start of football season. Right now, however, this year’s sample is only made up of “the most recent handful of months” which means we are likely undercounting the January and February 2025 totals and comparing them to completed 2024 reporting.

In practice, this means that Agency A will initially report 248 thefts occurred in February 2025, but the final total it eventually reports in a few months is actually 251. Last February’s count of 302 is complete and unchanged, so what was initially reported as an 18 percent drop in theft in Agency A in February will eventually become a 17 percent drop.

This undercounting is not large enough to change the trend across the whole sample given the scope of the decline. It does, however, add uncertainty to whether the 16 percent drop in property crime we’re reporting may actually be a 15 percent drop when underreporting is taken into account.

Second, it is still early in the year and there is lots of time for things to change. There is a reason to generally ignore agencies talking up how well they're doing in January.

The Real-Time Crime Index only has data available through February — so not even all of Q1 is evaluated in the RTCI. Other sources with data through mid-April paint an identical picture of falling crime adding confidence in the overall assessment, but even a perfect understanding of crime trends right now would still leave 8+ months of time left for those trends to change this year. If you're winning a game 31-7 at the end of the first quarter that's great, but it doesn't tell you what the final score will be.

Third, not every crime gets reported to police and just because the national trend is heading one direction does not mean their aren’t outlier cities heading in another direction. Murder is down a ton nationally in the RTCI sample through February but it was up in Little Rock, Arkansas and Houston, Texas. Crime data is imperfect and the best we can hope for is accuracy rather than precision.

With those disclaimers in mind, we can still say that crime is almost certainly down an enormous amount so far this year compared to 2024. This assessment is based on multiple avenues and sources of evaluating crime data.

There are more than 380 agencies covering nearly 100 million people (96.8 million, so close to 100!) in the latest Real-Time Crime Index sample through February 2025. Property crime was down 16 percent in the sample driven by a nearly 30 percent drop in motor vehicle theft while violent crime was down 14 percent thanks to 21 percent drops in murder and robbery.

That’s a large sample, by far our biggest since the RTCI launched in September 2024 with just over 300 cities covering around 70 million people. And the declines in crime are widespread throughout the country as seen in the below map.

Still, any analyst worth their weight should be immediately suspicious of these findings. I spent a lot of words last summer talking about how the FBI was almost certainly badly overstating national crime declines of this magnitude in their quarterly data.

Viewing the RTCI’s rolling over 12 month graph shows these declines aren’t new. Rather they are a continuation of the trend that started in late 2022. Here’s murder:

Violent crime:

And property crime:

Yet my gut reaction to figures like these is to assume it’s a data reporting issue. The proper approach, therefore, is to seek out evidence that confirms or contradicts this gut response.

One reason to feel good about these figures is that a review of more current crime data (mostly through mid-April) from big cities shows an identical pattern. That suggests the declines seen in the RTCI are real and not mostly an underreporting issue.

To show this I grabbed murder and violent crime data that was already aggregated by police departments from the 20 cities with a) the highest number of murders reported in 2023 and b) available data through at least February from police department sources (all but two agencies had data through at least March). I figured that would give a sufficient sample of big city crime to make sure it’s not a data reporting or “it's February” issue.

Murder is down 23 percent in these 20 cities with all but two cities showing a decline. Digging deeper, four of the 20 cities reported their fewest first quarter murders since before 1970 this year (Detroit - 1964, Los Angeles - 1965, Baltimore - 1965, and Philadelphia - 1966).

Violent crime counts in these 20 cities was also down a lot, roughly matching the decline seen in the RTCI. Both of the cities with increasing violent crime so far this year (Atlanta and Birmingham) have reported falling murder.

Next, I found YTD shooting data from 27 cities with available 2025 data through at least the end of March. Shootings were down in 21 of those cities with a 27 percent decline overall across the sample. Another datapoint in favor of the plunging crime being real.

The Gun Violence Archive also mimics the large downward trend. There were 3 percent fewer shooting victims through April 16th in 2025 compared to 2019 and 17 percent fewer victims compared to the same timeframe in 2024.

The GVA — which should not be prone to underreporting — is showing comparable trends to what we’re seeing in the RTCI where murder is down 21 percent compared to 2024 and 3.9 percent compared to 2019. Taking a longer view, fatal and non-fatal shooting victimization over the last 12 months is approaching where it was in 2019 as the downward trend that began in 2023 continues.

The crime trends through the first few months of 2025 are incredibly encouraging and backed up by the RTCI, aggregated data from individual cities, shooting data from individual cities, and the Gun Violence Archive.

In some respects, these findings aren’t all that surprising. Murder almost certainly fell at historic rates in 2023 and 2024, auto theft dropped a ton in 2024 after surging for four straight years, and other kinds of crime were declining in 2024. Crime trends tend to change for complicated reasons and rapid deviations (like with murder in 2020 or auto thefts in 2022) are relatively rare. Crime trends are normally like a large ship in that they take quite a while to turn around.

But these are not normal times. Huge amounts of uncertainty — economic, political, societal — could conceivably lead to large changes in America in the coming months.

All we can say with certainty is that major crime — as counted by the FBI — almost certainly declined quite a lot in the first quarter of 2025. These would be record levels of decline for most crime types if they held up for the full year.

Will this trend persist for the remaining 8+ months of the year? Only time will tell.