Never has a debate gone better for me than the one where I stated in fall 2023 that the "post-Covid crime spike has started to decline significantly, expect it to decline further as we go forward." I admit, I would never have predicted that by 2025, we'd be in the following place:
Murder Rate in 2025 - Potentially the lowest in recorded history (records begin in the early 1960s so we can't say "ever").
Property Crime Rate in 2025 - Also potentially the lowest in recorded history
Violent Crime Rate in 2025 - Likely at its nadir since the late 1960s
Article on the same below.
______
Some thoughts on crime, data, and crime data.
It's Not Just Murder That's Likely At Historic Lows
The US is on track to record the lowest violent crime rate since 1968 and lowest property crime rate ever.
Jeff Asher
Jun 23
Early crime data for 2025 — when considered in context of the nation’s crime trends in 2023 and 2024 — points to new historic lows being reached this year. The available data from 2024 and the first third of of 2025 suggests a strong possibility that the United States will report the lowest murder rate ever recorded, the lowest property rate ever recorded, and the lowest violent crime rate since 1968.
Let me explain.
The Real-Time Crime Index (RTCI) launched last September and has since provided monthly assessments of US crime trends. The sample has grown from between 290 and 300 agencies in the first few months to more than 400 in last month’s sample.
The most recent sample covers data through April 2025 and provides a glimpse into US crime trends that won’t be formally reported by the FBI until September or October 2026. Using the RTCI data to evaluate national trends brings out five main caveats/reminders that people should consider at all times:
First, the RTCI’s most recent sample covers nearly 380 agencies making up more than 95 million people nationwide (though not ever agency reports every crime type). History has shown that a sample of this size is reasonably accurate at predicting the year-end percent change even this early in the year so long as nothing historic like COVID and the summer of 2020 occurs (knocks on wood).
Second, the margins are not particularly close on this analysis, meaning that the bottom line should hold true even if the RTCI percent change is off by more than usual.
Third, not every crime gets reported to police. This is clearly seen in the National Crime Victimization Survey, and a drop in reported crime is just that.
Fourth, the counts being reported by agencies are not complete. Agencies may add a handful of offenses as the rest of the year goes on, so an 8 percent decline right now may actually be a 7 percent decline through the same timeframe when the data is finally done being reported. That’s a large part of the rationale for being conservative when estimating the crime decline this early in the year.
And fifth, there’s still time left for something historic to occur that changes the trajectory of this year’s crime trends, but time is running out. Yes there are six months left in the year, but nearly half the crime that is going to occur this year has already occurred.
On to the meat of the analysis.
The first major question to address is the reliability of data from a large sample of agencies through April in predicting the year-end national change. In an ideal world, a 6 percent drop in robbery through April in the RTCI sample would accurately predict a 6 percent drop nationally when the FBI releases this year’s data in 14 months.
Sadly, we do not live in an ideal world (as evidenced by only the one Saints Super Bowl). The challenge is further complicated by the FBI’s switch to NIBRS in 2021 giving us unreliable estimates of the nation’s crime totals that year. Unreliable 2021 data means guesstimating the change in 2022 is also a challenge (if you don’t have a confident estimate for 2021 then you can’t say how the 2022 figures relate).
Reviewing the other, more solid, years affirms that the RTCI through April accurately (albeit imperfectly) identifies national trends. Usually the RTCI is off by around 2 to 3 percentage points from the national year end percent change reported by the FBI, so a 3 percent drop in the RTCI at this point is best read as being about even or a small decline.
Take 2023, for example. Murder was down 10.5 percent in this month’s RTCI sample through April 2023, violent crime was down 2.4 percent and property crime was even. Nationally, these measures ended the year down 11.2 percent, 3 percent, and 2 percent respectively in the FBI’s annual report.
That’s really good predictive power!
So, if this sample of 380 to 400 agencies is a decently good predictor of the national trend, what does it mean for 2024 and 2025? I’ve already shown how the early numbers add up to possibly the lowest murder rate ever recorded in 2025, but what about violent and property crime? Well, to figure out where 2025 may end up we have to guess as to the FBI’s estimate for 2024.
First, let’s assume that the US population grew by about 1.6 million in both 2024 and 2025 matching the FBI’s estimated population growth from 2022 to 2023. Let’s be conservative and say that the violent crime count dropped by 1 percent in 2024 (the RTCI sample has it at -3.3 percent) and property crime fell by 4 percent (the RTCI sample has it at 7.3 percent). On top of that let’s stay conservative and predict violent crime will fall by 6 percent from 2024 to 2025 and property crime by 8 percent (the RTCI sample has it at -11 percent and -13 percent respectively in the RTCI through April).
Calculating the rates for 2024 and 2025 from those predicted percent changes and adding it to the nation’s crime rates since 1960 gives us the below graphs:
US violent and property crime rates are at or near historic lows
US Violent & Property Crimes Per 100k, 1960 - 2025
Under these assumptions, the 2025 reported violent crime rate of 335 per 100k would be the lowest since 1969 while the reported property crime rate of 1,677 would be the lowest since… ever. The FBI began publishing property crime with a similar methodology in 1960 (prior to that they only reported theft of $50 or more as far as I can tell).
What gives me confidence to publish this analysis in June 2025 is that the margins for the conclusion aren’t particularly tight. The US is on track to record a substantially lower violent crime rate than every year since the moon landing. And the nation’s reported property crime rate should be much lower than any property crime rate ever recorded even if you're conservative in just how much you think it’s falling.
Does that mean that every city is seeing large declines, that there is not room for continued improvement, or that every crime type is at a historic low everywhere? Of course not. But it does provide context for how frequent crime is reported in the United States and hopefully incentivizes a better understanding of why these lows are being reached.
A 1 percent drop in violent crime each of the last two years would result in a violent crime rate nearly 3 percent below the 1970 rate. A 4 percent drop in property crime in 2024 and 6 percent drop in 2025 would result in the lowest property crime rate ever recorded. And it's worth reiterating that the RTCI sample points to much larger drops occurring so far in both 2024 and 2025.
So, the caveats are that the year is not quite halfway through, not all crimes get reported to law enforcement, we’re likely dealing with a bit of underreporting making the drops seem slightly larger than they are, not all cities and counties are seeing declines, and our measurement is imprecise to begin with.
And with all that in mind, it’s definitely plausible — likely even — that the US in 2025 will report the lowest murder and property crime rates ever recorded to go along with the lowest violent crime in more than 55 years. For confirmation of this assessment, however, we will have to wait another 16 months.
