Donald Trump promised to end American carnage. Seven years later, he was still campaigning against it, even as crime fell. His record shows how a claim about public safety became an argument for his return to power—and for what he would do once he had it.
· About 13 minutes
THE 47 FILES / TRUMP’S RECORD ON CRIME
01 · The national record
The country Trump described, and the one he inherited.
≈22%Rise in the murder rate, 2016–2020
≈23%Fall in the murder rate, 2020–2024
18.5%Official fall in the murder rate, 2024–2025
At the September 2024 presidential debate, Donald Trump described crime as rising through the roof. He was seeking a return to the White House, and the country he described needed rescuing. Yet murder was falling. By the end of that year, its rate would be about 23% below the level in 2020, his last year in office. When Trump returned in January 2025, he inherited that improvement. A year later, his White House would present the continuing decline as a turnaround he had delivered.
Murder: national rate, 2008–2025.
Offenses per 100,000 residents. Dashed lines mark inauguration years. Calendar years do not align perfectly with time in office.
The gap is deliberate. The FBI warns that 2021 estimates are not comparable because reporting did not reach a statistically acceptable level of national representation. Rape’s definition changed in 2013; comparisons spanning that change are withheld below.
What happened between taking office and leaving it.
To understand what was missing from that account, follow the figures across the changes of president. The table below starts each administration with the year before it took office and ends with its last complete year. It includes both of Obama’s terms, Trump’s first term and Biden’s presidency. For Trump’s return, there is one complete year so far. The comparison shows the conditions each president inherited and left behind. It cannot, by itself, establish how much their policies changed them.
Murder · rates per 100,000 and approximate endpoint change
Administration
Baseline
End year
Change
Obama · first term
2008: 5.4
2012: 4.8
≈-11.1%
Obama · second term
2012: 4.8
2016: 5.4
≈12.5%
Obama · both terms
2008: 5.4
2016: 5.4
≈0.0%
Trump · first term
2016: 5.4
2020: 6.6
≈22.2%
Biden
2020: 6.6
2024: 5.1
≈-22.7%
Trump · current term (one year)
2024: 5.1
2025: 4.1
≈-19.6%
Changes calculated from published rates rounded to one decimal place. They can differ from FBI percentages calculated with unrounded estimates. The headline 2025 decline uses the FBI’s published 18.5% figure.
Inspect all seven categories and annual rates
Year
Murder
Rape
Robbery
Aggravated assault
Burglary
Larceny / theft
Motor vehicle theft
2008
5.4
27.9
145.9
276.9
731.9
2157.1
315.2
2009
5
27.8
133.3
264.3
717
2054.5
259
2010
4.8
27.7
119.1
253
699.7
1993.2
238.6
2011
4.7
26.8
113.8
241.3
700
1964.4
229.7
2012
4.8
26.9
112.9
242.3
670
1947.9
229.8
2013
4.5
36.5
108.9
229
608.5
1896.9
221
2014
4.4
36.9
101.1
229.3
536.3
1814.6
215
2015
4.9
38.8
101.2
237.2
489.2
1760.6
220
2016
5.4
40.6
102.7
248
467.3
1739.4
236.8
2017
5.3
42
95.2
244
424.9
1680
234.6
2018
5
44.2
83.8
241.7
373
1585.1
227.3
2019
5
43.7
78.4
241.5
331.1
1535
216
2020
6.6
40.1
71.2
271.3
306
1379.3
243.8
2021
Not comparable
Not comparable
Not comparable
Not comparable
Not comparable
Not comparable
Not comparable
2022
6.6
43.1
67.2
272.6
272.6
1414.8
284
2023
6
40.8
67.4
267.4
253.2
1364.4
321.9
2024
5.1
39.2
61
257.6
231
1274.1
256
2025
4.1
36.1
49.4
238
193.6
1144.1
197
The 2020 shock
The pandemic overturned his first-term record.
Trump entered office in 2017 promising an immediate end to American carnage. Murder fell between 2017 and 2019, giving him an improvement to point to. Then came 2020. In that pandemic year, the murder rate rose from 5.0 to 6.6 per 100,000 people, leaving it roughly 22% above the level before he took office. The reversal was severe, but it was not uniform: burglary and theft fell, while aggravated assault and vehicle theft rose. The different lines in the chart matter here. They show why neither a claim of universal disorder nor a claim of presidential control can explain everything that happened.
02 · Annual totals versus early signals
Back in office, Trump had a decline to celebrate.
2025: the completed national estimate.
During Trump’s first year back, violent crime and murder did fall sharply. The FBI’s completed 2025 estimates count 1,119,768 violent offenses and 5,245,768 property offenses. Their rates were 327.6 and 1,534.8 per 100,000 residents. Agencies reporting to the FBI covered 96.2% of the population, and the bureau produced estimates for the whole country. Trump could fairly point to a substantial improvement during his first year. The figures did not establish that he had begun it, or how much his policies contributed.
The early figures for 2026 suggest the improvement continued. In the Real-Time Crime Index’s January–July comparison, violent crime fell about 5.4% and murder about 16.3%. Its sample covers 565 agencies and about 113.2 million residents, making it a useful indication of what police departments are recording. It is still a comparison of counts over the same seven months, not a completed national rate for the year. The distinction matters when an early indication becomes the basis for a presidential claim.
Real-Time Crime Index · January–July · downloaded September 28, 2026
Category
2025 count
2026 count
Change
Violent crime
299,114
283,106
-5.4%
Murder
4,027
3,369
-16.3%
Rape
25,905
24,638
-4.9%
Robbery
59,413
51,518
-13.3%
Aggravated assault
209,769
203,581
-2.9%
Property crime
1,313,523
1,190,352
-9.4%
Burglary
162,231
138,523
-14.6%
Larceny / theft
938,091
878,639
-6.3%
Motor vehicle theft
213,201
173,190
-18.8%
Even within a broad decline, people can face quite different risks. The Council on Criminal Justice’s separate mid-year study found homicide, robbery and vehicle theft falling, but domestic violence, sexual assault and shoplifting rising in its smaller samples. A person experiencing one of those crimes would have little reason to feel reassured by a national average. That does not invalidate the wider decline. It is a reason to look more closely at what is falling, what is rising and whose experience the headline leaves out.
For someone who has been assaulted, the incident does not become less real because it never reaches a police report. That is a limitation of every police-based account of crime, including the one above. The National Crime Victimization Survey approaches the question from the other direction, asking people aged 12 and over about nonfatal violence they experienced, whether or not they reported it. It also includes simple assault, which sits outside the FBI’s violent-crime definition. Its rates are per 1,000 people, and it does not measure murder.
Bars show 95% confidence intervals. Legacy survey series, 2008–2024. The 2025 instrument redesign begins a separate series.
In 2024, BJS estimated 23.3 violent victimizations per 1,000 people, compared with 11.2 that were reported to police. Much of the experience captured by the survey would therefore be missing from a reading of police reports alone. BJS found no statistically significant change in the overall rate from 2023 or 2019. The bars on this chart show the uncertainty around each estimate; movements within that uncertainty should not be presented as a definite change. Together, the two systems offer a fuller account than either can provide on its own.
Trump’s account was much simpler than these overlapping records. Soon after taking office in 2017, he claimed murder was at a 47-year high. By 2020, he was claiming a record low. Neither claim was true. Out of office in 2024, he again described a country overwhelmed by crime; challenged with FBI figures, he attacked their credibility. After returning to power, he cited falling crime as evidence of success. Read the statements in sequence and the shift becomes clear: the numbers he used to describe the country changed with the case he was making about himself.
January 20, 2017
“This American carnage stops right here”
A promise his first term did not deliver.
On his first inauguration day, Trump tied the promise of safety directly to his arrival in office. Murder subsequently fell between 2017 and 2019, before surging in 2020. By the end of his term, the rate was about 22% higher than the year before he took office. The pandemic is essential to any explanation of that reversal. It also shows the limits of a promise that a president could bring the violence to an immediate end.
Less than three weeks after his inauguration, Trump told county sheriffs that murder was at a 47-year high. The latest complete year then available was 2015, with 4.9 murders per 100,000 people. In 1980, the rate had been 10.2. A recent increase was real, but Trump had described it as a historic peak when the country had recorded a rate more than twice as high within the period he cited.
By June 2020, Trump was presenting his time in office as a period of record safety. The published annual data then available covered 2018, when the murder rate was 5.0 per 100,000. Four years earlier, under Obama, it had been 4.4. Even before the pandemic-year surge appeared in the annual figures, the record-low claim failed against a recent year in the same series.
At the September 2024 debate, Trump needed voters to see a country that had deteriorated in his absence. His description of soaring crime did not match the national trend. The completed figures for that year put the murder rate at 5.1 per 100,000, down from 6.6 in 2020. Robbery and burglary were down too. Vehicle theft had risen earlier, an important exception to the broader improvement, but it could not sustain his account of a nationwide surge.
Confronted at the debate with FBI figures that challenged his account, Trump questioned the figures and cited cities missing from the reporting. The gaps deserved scrutiny, particularly the serious comparability problem in 2021. But the FBI estimates national totals rather than assuming that a city with missing reports had no crime. Identifying a weakness in collection did not establish that the subsequent declines were fraudulent.
Trump linked his account of rising crime to people entering the United States. The national figures did not show the surge he described, and these tables do not record offenders’ immigration status. They could not establish the explanation he was offering. Individual crimes can be devastating and deserve investigation on their own facts. Using them to describe an entire population, or the direction of crime across the country, requires different evidence.
The emergency rhetoric erased an existing decline.
When Trump announced his intervention in Washington, his description of a city overwhelmed by violence omitted its recent improvement. The US Attorney’s Office had reported a 35% fall in violent crime in 2024, bringing it to its lowest level in more than 30 years. Residents still faced serious crimes. But the decline meant Trump was intervening in a city where violence had already fallen substantially, rather than beginning a recovery from an uninterrupted rise.
A year after Trump returned, the White House credited him with reversing a national violent-crime surge. The further fall during 2025 gave the administration a real achievement during its tenure to discuss. Its account left out the earlier turnaround: murder had already fallen substantially by the end of Biden’s presidency. The difference is consequential. Continuing an improvement and starting it are different claims about a president’s record.
The national end-of-term comparison contradicts the framing.
The White House’s account of Biden’s legacy invited readers to picture a country descending into disorder. Compare 2024 with 2020 and the national record shows lower murder, robbery, burglary and overall violent crime. Vehicle theft was an exception that a fair assessment should include. It did not change the direction of the other measures, which contradicted the administration’s description of what Biden had left behind.
Supported. The decline should be reported honestly.
The FBI reported 9.3% fewer violent crimes and 18.1% fewer murders in 2025. Adjusted for population, the rates fell 9.7% and 18.5%. Those are substantial reductions during Trump’s first year back, and they belong in his record. They establish what happened during that year. Determining which policies produced the improvement, and how much they contributed, requires evidence beyond the fact that he was president when it happened.
On August 11, 2025, Trump announced his intervention in Washington. His description of a capital overwhelmed by violence collided with the decline already documented above. The gap between his description and that record now had consequences beyond a campaign argument. Crime was being used to justify bringing troops into the capital. His case depended on an account of conditions that left out the city’s recent progress.
What happened in the capital after the intervention.
The results since then need the same scrutiny as the justification. Through September 28, 2026, Washington’s preliminary counts recorded fewer homicides than at the same point in 2025, but more assaults with a dangerous weapon and a higher total for violent crime. These are local DC Code categories, which differ from the FBI’s national ones. The mixed picture does not establish the effect of the deployment. It does make a sweeping claim that the intervention solved the city’s crime problem harder to sustain.
DC preliminary year-to-date counts · same cutoff in each year
The administration also had to defend what troops were doing once deployed. On September 2, 2025, Judge Charles Breyer found that military conduct in Los Angeles violated the Posse Comitatus Act, which restricts the military’s role in civilian law enforcement. The ruling, with its delayed injunction, concerned that deployment. It was a district-court finding, not a final prohibition on every use of troops. Even so, it exposed a legal limit that the administration’s public-safety argument could not simply set aside.
Illinois: the Supreme Court denied emergency relief.
The dispute reached the Supreme Court in the Illinois case. On December 23, 2025, the justices declined to pause an order blocking the deployment. Their preliminary analysis questioned the authority claimed by the administration for using the military to execute federal laws. The block remained in place at that emergency stage, without a final decision on the full case. Invoking crime had helped Trump make the political case for intervention; it had not settled whether the intervention was lawful.
Across the record, the distinction is between what happened and what Trump claimed to have made happen. His first term ended with murder sharply higher. The rate fell under Biden and again during Trump’s first year back. Neither trend gives a president sole responsibility for the result. Both have to remain in view when assessing his promises, his claims of credit and his justification for expanding federal power. The sources below allow those accounts to be checked against the same evidence.
Data sources, revisions and survey methods
The charts use historical rates from the FBI’s 2025 report, including revisions. The original transcription retains 2021, but the chart leaves a gap because the FBI warns that year is not comparable. Changes calculated from rounded rates are marked approximate. The downloads include annual rates, the latest national totals, preliminary comparisons and survey uncertainty. Historical national counts and combined rates for entire presidential terms are not included; they would require a separate calculation from the underlying records.
The dates are part of the evidence. Preliminary city totals can be revised, later FBI releases may update earlier estimates, and a redesigned survey can break a comparison. This page preserves the dates and qualifications so readers can see what was measured, what was known and how far a conclusion can reasonably go.
