FBI Reports Violent Crime Fell 9.3% and Murder Fell 18.1% in 2025; White House and Criminologists Differ Over the Cause
The FBI's annual report, released August 14, 2026, records the largest one-year drops in both categories since its national estimates began in 1936; the administration credits its policies, while many researchers point to a decline that started in 2022.
A 9.3% Drop, and a Fight Over Who Made It Happen
The FBI released its 2025 crime report on Friday, August 14, 2026. The numbers are large. Violent crime fell an estimated 9.3% from 2024. Murder and nonnegligent manslaughter fell 18.1%[1]. Both are the biggest one-year drops the FBI has ever recorded in the national estimates it has published since 1936[1][3].
The murder rate landed at 4.1 per 100,000 people. FBI Director Kash Patel wrote that this ties 1955 and 1956 for the lowest rate since the FBI's estimates began[1]. Robbery fell 18.5%, rape fell 7.6%, aggravated assault fell 7.2%, and property crime fell 12.4%[1][3]. Nobody disputes any of this. The report draws on more than 17,000 law enforcement agencies, over 87% of the agencies in the FBI's program, covering communities with almost 329 million people[1]. That kind of coverage is why both sides treat the number itself as solid.
So here is the collision. Everyone agrees the drop happened. Almost nobody agrees on why. And the reason that gap exists is not spin — it is math. No one can rerun 2025 without the policies in question and see what would have happened instead. That missing comparison is the entire argument.
Whose Number Is It, Anyway
The White House moved fast. Spokeswoman Lauren Bis said, "President Trump promised to make America safe again and the data proves he has done just that"[3]. Press Secretary Karoline Leavitt called it the biggest murder drop ever recorded and tied it to the administration's law-and-order push, including National Guard deployments to Washington, D.C. and Memphis[4][6]. Some coverage went further, describing the murder rate as the lowest since 1900 — a bigger claim than the one the FBI itself made[4].
That gap matters. The FBI's own language is narrower: tied with 1955 and 1956, within its own data series that started in 1936[1]. "Lowest since 1900" is a different, larger comparison drawn from outside the FBI's own numbers. Fox News led with the 1900 figure and the phrase "under Trump administration," which does causal work without ever having to prove it[4]. The Washington Examiner called it vindication, giving the FBI's own superlative top billing[8].
The administration's strongest piece of evidence is not the national number. It is Washington, D.C., where homicides fell 41% in the first half of 2026 compared with the same period in 2025, after federal forces arrived[11]. That is a real, measured, local drop, and it is the case the White House keeps returning to.
The Trend Line Nobody on Either Side Disputes
Here is the second half of the collision. Murder has fallen every year since it peaked in 2022 — under President Biden[3][5]. Data analyst Jeff Asher calls 2025 the third straight year of record-setting declines, not a sudden break from what came before[3]. The Washington Post's headline put the rebuttal first: "Trump touts new stats on low crime. But crime has been falling for years"[5]. CNN's version added a note of doubt with "personal success," inviting skepticism before the reader saw the evidence[6].
Adam Gelb, who runs the nonprofit Council on Criminal Justice, put it plainly: the drop is historic, but what is driving it is genuinely unclear[6]. His reasoning is the strongest version of the researchers' case. Murder fell in cities with National Guard troops and in cities with none at all. It fell in cities led by both parties, under very different policing strategies and different local economies[6]. If one policy explained the whole thing, you would expect the drop to track that policy. It does not.
There's also a scale problem that neither side disputes once you look at the numbers. Washington, D.C. and Memphis together are a small share of total U.S. homicides. Even a dramatic percentage drop in those two cities cannot arithmetically produce an 18.1% decline nationwide[1][12]. Whatever drove most of the national number happened somewhere the federal government never deployed troops.
What "Tied With 1955" Actually Means, and Why the FBI Won't Say More
It helps to understand how this number gets built at all. The FBI does not count every crime in the country itself. Local police departments report their numbers voluntarily, and the FBI estimates a national total from what it receives. That is why the 17,000-agency, 329-million-person coverage matters so much this year — a bigger sample makes a bigger claim more trustworthy[1].
A rate, like 4.1 per 100,000, also depends on population. If a city's population grows, its crime rate can fall even if the raw number of crimes stays the same. This is one small reason researchers are careful about which comparison they use, and why the FBI itself sticks to comparing 2025 only against its own series going back to 1936[1].
That carefulness shows in Director Patel's own wording. His claim, tied with 1955 and 1956, refers only to the FBI's data series. It's a real and large number, but it's not the same as the White House's public "lowest since 1900" framing, which reaches outside that series for a longer, bigger comparison[1][4]. The FBI's release does not say what caused the decline at all. It just reports it[1].
There's a timing wrinkle, too. The report covers all of calendar year 2025. Trump took office on January 20, 2025, so roughly three weeks of that year predate his term, and policy effects on street crime typically take time to show up. That makes 2025 a genuinely hard year to hand to either administration.
Two Local Numbers, Both Real, Pointing Different Ways
The D.C. story looks different depending on which measure you pick. Homicides there fell 41% in the first half of 2026 versus a year earlier — a number the administration cites often[11]. But a separate Newsweek analysis of the full year found that overall violent crime in D.C. rose 3% over the same period, driven by a 41% jump in assaults with a dangerous weapon[12]. Homicides down sharply, other violent crime up: both numbers come from real data, and neither one is the whole picture on its own.
There's a cost side to this fight too. The Center for American Progress, a liberal think tank, estimated the National Guard deployments could cost $1.7 billion if they continue through the end of 2026, and concluded they had not measurably reduced urban crime[10]. The D.C. deployment has since been extended through the end of Trump's term[6][17]. Mayors and police departments in deployment cities argue their own officers, prosecutors and local violence-prevention programs deserve credit for arrests made and programs run on the ground — work that predates any federal involvement[10].
Coverage split largely along these same lines. UPI's wire report, picked up by outlets including Breitbart, read close to a straight data recap with no causal claim in the headline at all[7]. NBC News led with the record decrease but placed the pre-Trump trend line high in the piece, shaping how readers would weigh what followed[3]. CBS News described the rate as a "20-year low" — a smaller claim than the FBI's own 1955-and-1956 comparison in the same report, whether from a different measure or simply understating the finding[9].
What's Actually Being Fought Over
Strip away the specific numbers and the dispute is about something else: credit. For the White House, a historic drop is the strongest available argument for the deployments it has already made and extended[4][6]. For researchers and city officials, conceding that argument means accepting a precedent for federal troops in American cities well beyond this one year[10].
None of that changes the material facts on the ground. Fewer Americans were murdered in 2025 than in almost any year the FBI has ever measured, and the decline shows up broadly, across cities of every political stripe[1][6]. What remains open is the question nobody currently has the data to answer: how much of a decline already years in motion belongs to any single policy, in any single year.
Summary
The FBI released its 2025 national crime report on Friday, August 14, 2026. It shows violent crime fell an estimated 9.3% from 2024, and murder and nonnegligent manslaughter fell an estimated 18.1%[1]. Both are the largest one-year drops since the FBI began making these national estimates in 1936[1][3]. The murder rate came in at 4.1 per 100,000 people. FBI Director Kash Patel wrote that this ties 1955 and 1956 for the lowest rate since the FBI's estimates began[1].
Nobody in the argument disputes those numbers. The fight is over what caused them. The White House credits President Trump. Spokeswoman Lauren Bis said, "President Trump promised to make America safe again and the data proves he has done just that"[3]. Press secretary Karoline Leavitt tied the drop to the administration's law-and-order push, including National Guard deployments to Washington, D.C. and Memphis[4]. The administration also described the murder rate as the lowest since 1900 — a stronger claim than the FBI's own[4].
Many criminologists push back on the causal link, not the data. Murder has fallen every year since it peaked in 2022, which was under President Biden[5][6]. Data analyst Jeff Asher calls 2025 the third straight year of record-setting declines[3]. Adam Gelb, who runs the Council on Criminal Justice, said the drop is historic but what is driving it is unclear, because rates fell in cities with very different policing strategies, economies and political leadership[6].
So the real dispute is a counterfactual one: how much of a national decline already in motion can be assigned to a president who took office in January 2025? The administration's best case is that it sped up an existing trend and points to Washington, D.C., where homicides fell 41% in the first half of 2026 after federal forces arrived[11]. Critics' best case is that the same drop shows up in cities with no federal presence at all, and that D.C.'s other violent-crime numbers were mixed[10][12].
The Event
On Friday, August 14, 2026, the FBI released its 2025 Reported Crimes in the Nation statistics. The report estimates that violent crime fell 9.3% and that murder and nonnegligent manslaughter fell 18.1% compared with 2024[1]. It puts the 2025 murder rate at 4.1 per 100,000 inhabitants[1]. President Trump publicly credited his administration's policies for the decline the same day, and the White House press office repeated that claim[4][6].
Undisputed Facts
- The FBI's 2025 report estimates violent crime fell 9.3% and murder and nonnegligent manslaughter fell 18.1% from 2024[1].
- Those are the largest one-year decreases in both categories since the FBI introduced its data-based national estimates in 1936[1][3].
- Robbery fell an estimated 18.5%, rape 7.6% and aggravated assault 7.2%; property crime fell an estimated 12.4%[1][3].
- The 2025 murder rate was 4.1 per 100,000 inhabitants, which FBI Director Kash Patel wrote is tied with 1955 and 1956 for the lowest since FBI estimations began[1].
- The report draws on submissions from more than 17,000 law enforcement agencies, more than 87% of agencies in the FBI's program, covering communities with almost 329 million residents[1].
- The FBI published a preliminary 'First Look' at the same 2025 data on May 13, 2026, covering about 96% of the population[2][13].
- Violent crime and murder rates have declined every year since a spike in 2022, before Trump's return to office in January 2025[3][5].
- Trump deployed National Guard troops to Washington, D.C. in August 2025 and to Memphis in September 2025; the D.C. deployment has since been extended through the end of his term[6][17].
- In Washington, D.C., homicides fell 41% in the first half of 2026 compared with the first half of 2025[11].
The Pressure
Strip away the moralizing and blame. What structural realities persist regardless of which narrative wins?
- Credit is the scarce resource, not data
- Both sides accept the FBI numbers. What is being fought over is attribution, and attribution is worth more than the number to whoever wins it. For the White House it justifies troop deployments and helps in November 2026[4][6]. For mayors and researchers, conceding it hands over a precedent for federal forces in cities[10].
- How the FBI number is built
- The FBI does not count every crime in America directly. Local police departments report voluntarily. The FBI then estimates national totals from that pool. Coverage matters enormously: the 2025 report drew on more than 17,000 agencies covering almost 329 million residents, which is why it is treated as reliable[1]. In 2021 a switch in reporting systems left large gaps, which is part of why arguments about 'crime is up/down' got so confused for several years[13]. A national rate is per 100,000 people, so a growing population can lower the rate even if the raw count holds steady.
- The counterfactual problem
- No one can observe the 2025 that would have happened without federal deployments. That is the whole dispute in one sentence. The administration's evidence is a before-and-after in specific cities[11]. The critics' evidence is that the same decline shows up in cities with no federal presence[6]. Both are real; neither is proof.
- Scale mismatch
- Washington, D.C. and Memphis together are a small fraction of U.S. homicides. Even a large percentage drop in those two cities cannot arithmetically produce an 18.1% national decline[1][12]. Whatever caused most of the national drop happened in places with no federal deployment.
- Timing of the data window
- The FBI report covers calendar year 2025. Trump was inaugurated on January 20, 2025, so about three weeks of the year predate his term. Policy effects on street crime also lag. This makes 2025 a genuinely ambiguous year to assign to either administration.
Material realityFewer Americans were murdered in 2025 than in almost any year the FBI has measured. The rate of 4.1 per 100,000 ties the mid-1950s in the FBI's own series[1]. This is the fourth consecutive year of decline from the 2022 peak, and murder had already fallen sharply in 2023 and 2024[3][5]. The decline is broad — it shows up in cities run by both parties, with and without federal troops[6]. National Guard troops remain in Washington, D.C. under an order extending through January 2029, at a deployment cost one liberal think tank puts at $1.7 billion across cities[10][17]. D.C.'s own numbers are mixed: homicides down 41% in the first half of 2026, but one analysis found overall violent crime up 3% on a 40% rise in assaults with a dangerous weapon[11][12]. None of that changes if the credit argument is settled tomorrow.
Narrative as a weaponThree groups are actively shaping how you read this. The White House wants you to read a number as a verdict on a policy — its message discipline is visible in the upgrade from the FBI's "tied with 1955 and 1956" to "lowest since 1900," a bigger claim drawn from a different estimate[1][4][6]. Left-leaning outlets want you to read it as a trend the president walked into, and they do that by putting the rebuttal in the headline, which risks making a genuinely historic drop sound like a disputed talking point[5][6]. Criminologists want you to read it as a puzzle nobody has solved, which is their honest professional position and also conveniently keeps them out of the political fight[3][6]. The safest reading: the numbers are strong and are not seriously contested; the causal story is not established by anyone, in either direction.
How Each Side Sees It
Each major actor’s view — how it frames things, its underlying incentive, and how it’s materially affected. Tap a side to read it.
Frames it asTheir argument is about deterrence and signal. A president does not have to write state criminal codes to change behavior on a street. He can flood a city with federal agents, restart aggressive federal gun and drug prosecutions, and make it publicly clear that arrests will be made. That changes what an armed 19-year-old expects to happen to him tonight. They also argue that trends do not continue on their own — a decline can stall, and it did not, which they treat as evidence the push worked. Their sharpest single number is Washington, D.C.: homicides down 41% in the first half of 2026 after federal forces arrived[11]. Press secretary Karoline Leavitt has framed it as the biggest murder drop ever recorded, credited to a law-and-order push[4]. Spokeswoman Lauren Bis: "President Trump promised to make America safe again and the data proves he has done just that"[3].
WhyPublic safety was a core 2024 campaign promise and the stated legal basis for deploying troops to American cities[6][17]. A historic FBI number is the strongest available proof that the deployments were justified — and the strongest defense against lawsuits and criticism that they were not[10].
Impact on themPolitically large. Ahead of the November 2026 midterms, the administration can point to a federal dataset rather than an opinion poll. The D.C. deployment is now set to run through the end of his term, a commitment of federal money and personnel that this data helps defend[17].
Frames it asTheir objection is methodological, not partisan. To claim a policy caused a change, you need a comparison — what would have happened without it. Murder started falling in 2023 and has fallen every year since[3][5]. It fell in cities that got federal troops and in cities that got none, under Democratic and Republican mayors, in growing and shrinking local economies. Adam Gelb of the Council on Criminal Justice says the drop is historic but the cause is genuinely unclear[6]. The Council lists candidate factors: changes in how police and prosecutors operate, new technology, and broad social shifts after the pandemic[3]. Analyst Jeff Asher notes 2025 was the third straight record-setting year — a pattern, not a break[3]. Their deeper point: if you credit a president for a decline he inherited, you have also agreed to blame the next one for an increase he inherits.
WhyProfessional credibility rests on not over-claiming causation. Several of these analysts also publicly resisted claims that crime was soaring in 2023 and 2024, and are applying the same standard now[5][16].
Impact on themTheir datasets are cited by both parties, so being seen as a partisan referee would cost them access and influence. The Council on Criminal Justice is a nonprofit funded largely by philanthropic foundations and works with members from both parties; AH Datalytics is a private data-analytics firm[3][6].
Frames it asThe Bureau's position is that its job is to publish the count, not the explanation. The release states the percentage changes and the historical comparison and stops there[1]. Director Kash Patel's own line is carefully bounded — 4.1 per 100,000 is "tied with 1955 and 1956 for the lowest rate since FBI estimations began," which is a claim about the FBI's own series, not about all of American history[1]. The agency also emphasizes coverage: more than 17,000 agencies and nearly 329 million residents, which is why it treats the 2025 estimate as solid[1]. It released an unusually early preview in May 2026 to get the numbers out faster[2].
WhyThe FBI's crime data is only useful if both parties trust it. Overstating a finding under a politically appointed director would damage the series itself.
Impact on themThe Bureau sits between its own careful wording and an administration that has restated the figure more strongly in public[1][4]. That gap is now part of the story.
Frames it asThey argue the credit is being taken for work they did. Local departments made the arrests, ran the violence-interruption programs and installed the cameras. They also argue the federal deployments were expensive and unfocused: the Center for American Progress, a liberal think tank, put the cost of the National Guard deployments at $1.7 billion and concluded they did not reduce urban crime[10]. On D.C. specifically, they point out the picture is mixed — one year-out analysis found overall violent crime up 3%, driven by a 41% rise in assaults with a dangerous weapon, even as homicides fell sharply[12]. And they note the constitutional stakes: sending troops into American cities is a precedent that outlasts any one crime statistic.
WhyMayors need credit for safety gains to win reelection and to protect local control of policing. Some are also in active litigation or political conflict with the administration over the deployments[10].
Impact on themCities keep the federal presence and its costs regardless of who wins the argument. Falling crime also weakens the legal rationale of an ongoing emergency, which cuts in their favor.
Like this article?
The Bias Ledger average rating 4.6
The same story, as framed by outlets across the spectrum, ordered least to most biased. The bias score (1 = straight, 10 = heavily spun) is an AI assessment of that framing — click an outlet to see its track record. The tell is the word choice or omission that reveals the angle.
| Outlet | Vantage | Bias | How they frame it | The tell |
|---|---|---|---|---|
| UPI | U.S. wire service, center | 2 | "FBI statistics: Violent crime, murders fell by record levels in 2025" | Closest to a straight recap; the attribution fight is compressed to a few lines. That neutrality also under-signals to readers that who-gets-credit is the actual live dispute. Republished by outlets including Breitbart with the wire headline intact. |
| NBC News | U.S. center-left | 3 | "FBI reports record decrease in violent crime in 2025" | Straight data headline. Carries both the White House quote and the pre-Trump trend line, but places the "crime was falling before he took office" line high, which sets the frame. |
| CBS News | U.S. center-left | 4 | "Violent crime plummeted in 2025, with murder rate hitting 20-year low, FBI says" | "20-year low" is a substantially weaker superlative than the FBI's own comparison to 1955 and 1956 in the same report[1]. Whether an error or a different measure, it understates the finding in a story where the size of the drop is the contested asset. |
| Washington Examiner | U.S. right | 5 | "Homicide rate in 2025 lowest in 70 years, FBI says" | Attributes the superlative to the FBI, which is accurate, but frames the story as vindication and gives the timing objection less room than the achievement. |
| CNN | U.S. left-of-center | 5 | "Murder is at lowest rate since the 1950s. Trump touts it as a personal success." | Leads with the accurate FBI framing, then flags the credit claim with "personal success" — a phrase that invites the reader to doubt it before any evidence is presented. |
| The Washington Post | U.S. left-of-center | 6 | "Trump touts new stats on low crime. But crime has been falling for years." | The rebuttal is the headline. The word "touts" characterizes the president's claim before the reader sees the number, and the record-setting size of the 2025 drop is subordinated to the trend argument. |
| Fox News | U.S. right | 7 | "White House says murder rate plummeted to lowest level since 1900 under Trump administration" and "Leavitt claims biggest murder drop ever, credits Trump's law-and-order push" | Uses the White House's "since 1900" figure rather than the FBI's own "tied with 1955 and 1956." The phrase "under Trump administration" does causal work without making a causal claim the piece has to defend. |
References
- FBI Releases 2025 Reported Crimes in the Nation Statistics — Federal Bureau of Investigation · U.S. federal law enforcement agency; primary source, director appointed by the sitting president
- FBI Releases Historic Early Look at Annual Crime Data — Federal Bureau of Investigation · U.S. federal law enforcement agency; primary source
- FBI reports record decrease in violent crime in 2025 — NBC News · U.S. commercial broadcast network; center-left newsroom
- White House says murder rate plummeted to lowest level since 1900 under Trump administration — Fox News · U.S. right; owned by Fox Corporation
- Trump touts new stats on low crime. But crime has been falling for years. — The Washington Post · U.S. left-of-center newsroom; owned by Jeff Bezos
- Murder is at lowest rate since the 1950s. Trump touts it as a personal success. — CNN · U.S. left-of-center newsroom; owned by Warner Bros. Discovery
- FBI statistics: Violent crime, murders fell by record levels in 2025 — UPI · U.S. wire service; owned by News World Communications, founded by the Unification Church
- Homicide rate in 2025 lowest in 70 years, FBI says — Washington Examiner · U.S. right; owned by Clarity Media Group (Philip Anschutz)
- Violent crime plummeted in 2025, with murder rate hitting 20-year low, FBI says — CBS News · U.S. commercial broadcast network; center-left newsroom
- The Trump Administration's $1.7 Billion National Guard Deployments Fail To Reduce Urban Crime — Center for American Progress · U.S. liberal advocacy think tank; closely tied to Democratic Party policy networks
- One Year After Trump Sent National Guard to DC, Homicides Plunge 41%, Overall Crime Stats Mixed — The Daily Signal · U.S. conservative outlet founded by The Heritage Foundation
- Did Trump's DC Takeover Cut Crime? What One Year of Data Shows — Newsweek · U.S. general-interest outlet; center, digital-traffic driven
- Preliminary FBI data shows a sharp drop in violent crime — Stateline · U.S. nonprofit state-policy newsroom, part of States Newsroom; funding skews progressive
- Trump To Tout National Crime Decline Based On 2025 FBI Data — Criminal Justice Journalists · U.S. nonprofit trade association for crime reporters; aggregation newsletter
- Why Crime Rates Are Falling Across the U.S. — TIME · U.S. general-interest magazine; center-left
- Murder rate down 18% in 2025. Who gets to take credit? — Straight Arrow News · U.S. outlet founded by businessman John Solomon's backers; markets itself as unbiased, editorially center-right
- DC National Guard Deployment Extended Through Trump's Term, One Year After Crime Emergency — HNGN · U.S. digital aggregator; low-transparency ownership, treat as secondary
- US homicide rate falls to 75-year low as violent crime declines, FBI says — Fox Television Stations · U.S. local-station network owned by Fox Corporation; right-leaning parent, straighter local copy