Crime Is Plummeting. Half of America Doesn't Believe It. Both Sides Have a Point.

The United States is in the middle of what may be the fastest recorded decline in murder. The FBI says violent crime hit its lowest rate since 1969. Property crime is at levels not seen since 1961. By almost every official metric, America is getting dramatically safer.

And yet, when Gallup asked Americans in 2025 whether crime had increased in the past year, 49% said yes [1]. Retailers claim organized theft is an existential crisis. Police departments say they cannot hire enough officers. The National Crime Victimization Survey—the federal government's other major crime measure—paints a far murkier picture than the FBI's rosy numbers.

So who is right? The answer is more complicated than either side of the political aisle wants to admit.

The Official Numbers: A Historic Decline

The FBI's 2024 annual report, released in August 2025, delivered striking figures. National violent crime decreased an estimated 4.5% from 2023, producing a rate of 359.1 per 100,000 residents—the lowest since 1969 [2]. Murder and non-negligent manslaughter plunged 14.9%. Robbery fell 8.9%. Aggravated assault dropped 3.0%. Rape declined 5.2%. Property crime fell 8.1%, to its lowest rate since 1961 [3].

Source: FBI UCR / Jeff Asher Real-Time Crime Index

Data as of Mar 20, 2026CSV

The decline was not concentrated in a few cities. Crime analyst Jeff Asher, whose Real-Time Crime Index tracks 570 law enforcement agencies covering roughly 115 million people, found that crime fell in every major category across every population group in 2024 [4]. Early 2025 data suggests even steeper drops: a roughly 20% decline in murder, which would be the largest single-year decrease ever recorded [4].

The Council on Criminal Justice's mid-year 2025 report, covering 42 cities, found 11 of 13 offense categories declining. Homicides fell 17% compared to the first half of 2024. Robberies dropped 20%. Gun assaults plummeted 21%. Motor vehicle theft declined 25% [5].

Source: Council on Criminal Justice

Data as of Mar 20, 2026CSV

Multiple major cities reported homicide counts at historic lows. New Orleans logged its fewest murders since 1970. Detroit, Baltimore, Philadelphia, Oakland, and San Francisco all hit levels not seen since the 1960s [4]. New York City recorded record-low shootings.

Putting It in Historical Context: The 1990s Peak and the 2014 Low

To understand where current crime rates sit, you need the long view. The U.S. homicide rate peaked at 10.7 per 100,000 in 1991 during the crack epidemic [6]. It then fell steadily to a low of 4.4 per 100,000 in 2014 [6]. The COVID-era spike pushed it back up to roughly 6.8 in 2020—alarming, but still well below the 1990s peak.

By 2024, the murder rate had fallen back to approximately 5.0 per 100,000 [4]. Asher's projections suggest a 2025 rate of around 4.0—which would be the lowest in modern American history, surpassing the 2014 low and approaching levels last seen in the early 1960s [4].

The violent crime rate tells a similar story. In 1991, it was 758 per 100,000. By 2014, it had fallen to 372. The 2024 figure of 359 represents a new floor [2]. Property crime rates have declined even more dramatically: the 2024 rate is roughly one-third of the early 1990s peak.

The bottom line: America is far safer than it was 30 years ago, and by most measures safer than it has been at any point since the mid-20th century.

The Other Dataset: Why the NCVS Tells a Different Story

But the FBI's Uniform Crime Reports are not the only federal crime measure—and the other one is far less reassuring.

The National Crime Victimization Survey, administered by the Bureau of Justice Statistics, surveys approximately 240,000 people per year about whether they have been crime victims, regardless of whether they reported those crimes to police. In theory, the NCVS captures the "dark figure" of unreported crime that the FBI misses entirely.

The two sources have diverged dramatically in recent years. The most striking example: the FBI reported that violent crime dropped 2% from 2021 to 2022. The NCVS, covering the same period, showed violent victimization surging 75% [7]. That is not a rounding error. It is a chasm.

The Council on Criminal Justice has documented this divergence, noting that from 2000 to 2019, FBI serious violent crime rates and NCVS rates of serious violence reported to police tracked each other closely. Since 2020, they have separated [7].

Several factors explain the gap. The UCR only captures crimes reported to law enforcement; the NCVS includes unreported incidents. The two systems define crimes differently and cover different populations—the NCVS excludes homeless and institutionalized persons and only surveys people aged 12 and older. Timing differences between incident-year and interview-year data create additional noise.

But there is a more troubling possibility: the gap may reflect falling reporting rates. If fewer victims are calling the police—due to distrust, futility, or simply the hassle—then the FBI's numbers would improve even if actual victimization stayed flat or rose. The NCVS's 2024 results found that neither property nor violent crime were statistically different from 2022 and 2023 levels, contradicting the FBI's reported declines [8].

The 2024 NCVS results carry their own caveats. The survey was redesigned for the first time since 1992, cutting the sample roughly in half and significantly widening margins of error [8]. Still, the persistent divergence between the two federal sources means that triumphant claims about "historic" crime drops deserve an asterisk.

The NIBRS Crater: When a Third of America Disappeared from the Data

Any serious analysis of recent crime trends must reckon with the FBI's catastrophic data transition.

In 2021, the FBI retired its legacy Summary Reporting System and required all agencies to submit data through the National Incident-Based Reporting System. Thousands of departments were not ready. More than 6,000 law enforcement agencies—covering roughly 35% of the U.S. population—failed to submit data [9]. That included the nation's two largest cities, New York and Los Angeles, and most agencies in California, New York, Illinois, Pennsylvania, and Florida.

Source: FBI / The Marshall Project

Data as of Mar 20, 2026CSV

The result was a statistical crater. For decades, the FBI had achieved roughly 95% population coverage. In 2021, that figure collapsed. The FBI was forced to develop new estimation procedures to fill the gap, and the reliability of year-over-year comparisons between 2020 and 2021-2022 data remains questionable [9].

Coverage has improved significantly since then—reaching approximately 85% by the end of 2023, including 9 of the 10 largest U.S. cities [9]. But the transition created a period where national crime data was, in the assessment of The Marshall Project, "too unreliable to tell" what was actually happening [9].

Asher has flagged a related problem: the FBI's revision patterns. When the bureau releases updated figures, the revisions have grown larger. Murder in 2023, for example, was initially reported as down roughly 11% but was later revised to approximately 9.5% [4]. The FBI's 2024 murder estimate may undercount actual homicides by around 700 [4]. These are not signs of a broken system, but they are signs of a system still recovering from a self-inflicted wound.

Clearance Rates: Solving Fewer Crimes Than Ever

Even if crime is falling, police are solving a smaller share of the crimes that do occur—and this trend predates the pandemic.

The national homicide clearance rate has dropped from 72% in 1980 to 61.4% in 2024 [10]. Rape clearance rates fell from 49% to 27% over the same period. Aggravated assault clearance dropped from 59% to 49% [10]. Robbery, always the hardest violent crime to clear, sits at roughly 30%.

This matters for the crime debate in two ways. First, falling clearance rates may reflect rising crime that goes undetected or unsolved at higher rates. If police are unable to investigate and close cases, some crimes simply vanish from the pipeline—they are reported but never result in an arrest, which can discourage future reporting. Second, low clearance rates undermine deterrence. If a robber knows the odds of arrest are under one in three, the calculus shifts.

The staffing crisis is a major driver. As of early 2025, sworn police staffing was up slightly from the prior year but still 5.2% below 2020 levels [11]. New York City is short over 3,000 officers. Chicago is down more than 1,300. Los Angeles, more than 1,000 [11]. Sixty-five percent of agencies report having to reduce services or eliminate specialized units due to shortages—up from 25% in 2019 [11].

Departments are responding by lowering hiring standards. Pennsylvania State Police dropped college requirements. Oklahoma lowered the minimum age from 23 to 21. California is considering eliminating marijuana use as a disqualifier. Mississippi has relaxed some criminal history restrictions [11]. Whether this addresses the staffing crisis without creating new problems remains to be seen.

The Retail Theft Debate: Moral Panic or Genuine Crisis?

Few crime topics have generated more heat and less light than organized retail theft.

The National Retail Federation's 2023 survey found that retailers tracking ORC reported incidents rising 57% from 2022 to 2023 [12]. Its 2024 report documented an 18% increase in shoplifting incidents [12]. The industry estimates $45 billion in annual losses to shoplifting.

But the NRF's credibility on this topic took a serious hit in December 2023, when it retracted a widely cited claim that "nearly half" of the industry's $94.5 billion in shrinkage was attributable to organized retail crime [13]. The claim, it turned out, was a "mistaken inference" linking unrelated data sources. The NRF's own shrink survey showed all external theft—not just organized crime—accounted for just 37% of losses, with the rest attributable to employee theft, administrative errors, and other factors [13].

The retraction matters because the inflated figure had been deployed aggressively in Congressional testimony and media appearances to justify new legislation and tougher enforcement. Several major retailers, including Target, had cited theft as a driver of store closures—only to later walk back or qualify those claims.

This does not mean retail theft is fabricated. Shoplifting is a real problem, and the rise of online resale platforms has created new fencing opportunities. But the gap between industry rhetoric and verifiable data is wide. Few national chains have disclosed specific inventory shrinkage figures in SEC filings that would allow independent verification of their claims. The NRF itself ceased publication of its longstanding annual shrink report in 2024, making trend comparisons even harder [13].

The Perception Gap: Why Americans Feel Unsafe When Crime Is Falling

Gallup's 2025 polling found that 49% of Americans described crime as "extremely" or "very" serious nationally—down from recent peaks but still remarkably high given the data [1]. Only 12% said the same about crime in their own area. This national-versus-local gap has persisted for decades and points to a perception driven more by media and politics than personal experience.

The partisan dimension is stark. In 2024, 90% of Republicans said national crime was rising. By 2025—with a Republican president in the White House—that figure dropped 36 points [1]. Democrats moved in the opposite direction, becoming more likely to report rising crime after losing the presidency. People report feeling safer walking at night and perceiving lower local crime when their party controls the White House [1].

But dismissing public concern as pure partisanship ignores legitimate reasons people feel less safe. Several high-salience crime types genuinely spiked during and after the pandemic, even as aggregate figures declined. Carjackings surged in many cities, though CCJ data shows them now 24% below 2024 levels and roughly at 2019 levels [5]. Smash-and-grab robberies generated enormous media coverage. Fentanyl overdose deaths—over 70,000 annually at their peak—blurred the line between public health crisis and public safety crisis. Motor vehicle theft, despite recent declines, remains 25% above pre-pandemic 2019 levels in CCJ's city sample [5].

These offenses are disproportionately visible. A carjacking at a gas station or a smash-and-grab captured on security camera has an outsized psychological impact compared to the statistical reality that your overall odds of being a victim have declined.

Progressive Prosecutors: Scapegoats or Contributing Factors?

The political debate over crime has centered heavily on progressive prosecutors—district attorneys who campaigned on reducing incarceration, diverting low-level offenders, and reforming cash bail. Critics, led by the Heritage Foundation and Manhattan Institute, argue these policies directly fueled crime increases [14]. The electoral evidence is mixed: San Francisco's Chesa Boudin was recalled and Los Angeles's George Gascón lost re-election, but Manhattan's Alvin Bragg and Philadelphia's Larry Krasner won by wide margins [14].

The most rigorous academic study on the question, a 2024 analysis published in Criminology & Public Policy, examined crime rates in the 100 largest U.S. counties from 2000 to 2020 using difference-in-differences regression. Its finding: jurisdictions that elected progressive prosecutors had 7% higher total index crime rates, driven by higher property crime. But violent crime rates were not statistically higher [15].

That finding is uncomfortable for both sides. Progressives cannot simply dismiss the property crime effect. But conservatives who blame progressive DAs for murder spikes are not supported by this data. The study also noted significant methodological challenges: many "progressive" prosecutors replaced predecessors who were themselves relatively progressive, making clean comparisons difficult [15].

During Boudin's tenure in San Francisco, motor vehicle theft and larceny increased, but robbery and assault remained stable, and his office actually increased charging rates for serious offenses by 16% overall [14]. Gascón saw violent crime rise in Los Angeles during 2020-2021, as it did nationwide, while misdemeanor declination rates nearly tripled [14]. The pattern suggests local prosecutorial policy is one variable among many, not the master switch its critics claim.

The "Defund" Question: What Actually Happened

The "defund the police" movement of 2020 is frequently blamed for rising crime. The reality is more prosaic. An ABC News analysis of over 100 city and county budgets found that 83% were spending at least 2% more on police in 2022 than in 2019 [16]. Only eight agencies cut funds by more than 2%. At least 14 cities that announced cuts scaled them back before adoption or reversed course within a year [16].

Austin, often cited as a "defund" city, cut about one-third of its police budget for fiscal year 2021—then approved a record $442 million for police by FY2022, with further increases in FY2023 [16]. The narrative of widespread, sustained defunding is largely fiction. What is real is the staffing crisis: officers leaving faster than departments can hire, driven by retirement, low morale, and competition from other sectors.

Who Deserves Credit—and Does It Matter?

President Trump has claimed credit for falling crime rates, as has the Biden administration before him. The honest answer is that neither deserves much. The murder decline began in mid-2023, accelerated through 2024, and continued into 2025—spanning both administrations [4]. Crime is overwhelmingly a local phenomenon shaped by policing strategies, economic conditions, drug market dynamics, demographics, and community factors that no president controls.

The more important question is whether the decline is durable. The CCJ's mid-2025 data offers a cautionary note: while homicides are down 14% from 2019 levels in their city sample, only 38% of cities have homicide rates below their 2018-2019 pre-pandemic baselines [5]. The decline is real but incomplete. Most cities are recovering from the COVID-era spike, not breaking new ground.

What We Know, What We Don't, and Why It Matters

The evidence supports several conclusions simultaneously:

Crime is falling, and substantially. The FBI data, the Real-Time Crime Index, the CCJ city sample, the Gun Violence Archive, and CDC mortality data all point in the same direction. This is not a statistical artifact.

The official numbers likely understate the full picture. The NCVS divergence, falling reporting rates, the NIBRS transition crater, and declining clearance rates all suggest the "true" crime level may be higher than FBI figures indicate—even if the directional trend is genuinely downward.

Specific crimes remain elevated. Motor vehicle theft is still well above 2019 levels. Some cities remain far above pre-pandemic baselines. The aggregate masks real variation.

Public perception is partially irrational and partially rational. Partisan identity is the strongest predictor of whether someone thinks crime is rising. But high-salience offenses, visible disorder, and the fentanyl crisis give people legitimate reasons to feel less safe even as their statistical risk declines.

The data infrastructure is improving but still imperfect. The NIBRS transition was necessary but badly managed. Coverage is recovering. The NCVS redesign introduces its own uncertainties. Anyone claiming precise knowledge of national crime trends between 2020 and 2023 is overstating what the data can support.

The crime debate in America is not really about crime. It is about trust—in institutions, in data, in each other. The numbers say the country is getting safer. Whether Americans believe it depends less on the numbers than on who is telling them.