Record Murder Drop Stuns America

The United States just experienced the sharpest one-year decline in homicide on record, and the best evidence indicates the drop was broad-based, durable across much of the country, and large enough to push the national murder rate to a modern-era low.

The Short Version

  • Homicides fell roughly one-fifth in 2025, the largest recorded one-year drop; multiple datasets converge on an 18–21% decline.
  • Early federal estimates and independent city data point to the lowest national murder rate in decades, plausibly since the early 20th century.
  • The decline was widespread across cities and regions, continuing decreases that began in 2023 and 2024.
  • No single policy explains the shift; improvements reflect many forces, from policing practice to pandemic unwind and offender-victim dynamics.

What the data shows: the scale and breadth of the decline

Several independent streams of evidence tell the same story. The Council on Criminal Justice (CCJ), analyzing 35–40 large cities, reported a 21% year-over-year decline in homicides from 2024 to 2025, representing 922 fewer killings in its sample—a historically steep drop for a single year. Crime analyst Jeff Asher’s Real-Time Crime Index, built from hundreds of local agencies, likewise pointed to a near-20% nationwide decline by year end, signaling the largest one-year decrease in modern records. On the federal side, preliminary FBI figures pegged murder and non-negligent manslaughter down about 18%, reinforcing the same directional magnitude.

Put simply: multiple, methodologically distinct sources—city panels, near-real-time aggregators, and early-release federal estimates—converge on the same outcome. That convergence matters because homicide is a low-base, high-variance offense; it can swing sharply in one city without implying a national shift. In 2025, however, reductions appeared across major metros in both red and blue states, echoing the broader pattern identified by NPR’s expert roundups: the decline was not narrowly regional or partisan.

How we got here: from pandemic spike to rapid normalization

To understand why a 20% drop is plausible, recall the arc. Homicides spiked during the pandemic era amid social disruption, thinning police-citizen contact, strained courts, and volatile local illegal gun markets. Starting in 2023, murders declined nationally—first by double digits—then continued falling into 2024 and accelerated in 2025. CCJ’s series frames the three-year stretch as an average annual decline of mid-teens percentage points, with 2025 the steepest leg down in the sequence.

Several plausible mechanisms likely compounded. Police agencies restored proactive activity and investigative capacity that had lapsed during COVID constraints; courts cleared backlogs; community groups and city-led initiatives re-engaged hot spots; and some of the unique pandemic-era frictions—like volatile routine activity patterns and school closures—unwound. None of this is monocausal, but the timing matches the operational realities practitioners saw on the ground. The New York Times encapsulated the synthesis succinctly: a confluence of factors, not a single lever, sits behind the staggering drop.

Is 2025 truly a “historic low” in murder?

Two different “lows” get conflated: the rate and the one-year change. On the latter, the evidence is clear—the U.S. likely recorded the largest single-year decline in murder in the modern statistical era, with reputable datasets clustering around 18–21%. On the former, early reads argue that the 2025 homicide rate was the lowest observed since at least the early-1960s FBI era and plausibly since circa 1900, once you calibrate for population growth; national sampling through mid-2026 still showed murder levels tracking at or below that floor. Several outlets and analysts, drawing on both FBI and CCJ, framed 2025 as a 125-year low or close to it, with the appropriate caveat that final federal tables lag and historical pre-1960 comparisons rely on reconstructions rather than continuous federal series.

For a careful reader, the right takeaway is this: by the best available measures, the rate in 2025 fell back below late-2010s levels and likely reached a modern-era nadir, consistent across independent sources. Claims of “lowest since 1900” are analytically defensible as a shorthand, provided one understands they rest on blending finalized modern FBI rates with reputable historical reconstructions and large-city proxies.

Attribution: what changed, who gets credit, and what experts agree on

Political actors inevitably argue credit when crime turns; this cycle is no exception. The White House emphasized a return to “law and order,” highlighting federal enforcement and support for local policing, and cited FBI early estimates to underscore the scale of the drop. Independent institutions confirm the drop itself—CCJ’s city panels, NPR’s surveys of experts, and national analysts’ indices—while cautioning against single-cause explanations. Asher’s national read and CCJ’s multi-city data depict a geographically broad decline that did not depend on one party’s leadership in any particular locale.

What do subject-matter experts tend to emphasize? Three families of drivers recur: place and people (where violence concentrates and which groups are at highest risk), police and prosecution (deterrence, incapacitation, clearance rates, and focused deterrence), and social/market conditions (illegal firearm supply, drug markets, and post-pandemic routine activity). The CCJ synthesis and allied analyses point to improvements across these fronts in 2025, with no single lever sufficient to explain a national swing of this size on its own.

How the measurement works—and why multiple lenses matter

Murder is counted more reliably than other crimes, but there are still measurement nuances. National FBI figures aggregate local agency submissions via the Uniform Crime Reporting/NIBRS systems and typically publish with a lag; early releases provide directional insight but are not the last word. Large-city panels (like CCJ’s 35–40 cities) update more quickly and signal movement among jurisdictions that account for a disproportionate share of national homicides. Real-time indices stitch together hundreds of departments to estimate national trends faster than the FBI can finalize them.

This is why convergence is key: when city panels, near-real-time aggregators, and preliminary federal estimates align on the same ballpark decline, you can have high confidence in the core finding even before the FBI’s final annual table. In 2025 they did—landing between an 18% and 21% reduction—an extraordinarily steep, nationally meaningful swing.

What it means next: sustaining gains and watching the right indicators

Big declines can fade if underlying risks reassert themselves. The durable version of this progress requires three things. First, keep homicide clearance and shooting clearance rates moving up; deterrence depends on credible, swift consequences, particularly for the small networks where violence clusters. Second, stay focused on place-based prevention—people and locations with the highest risk—using tools like street outreach, group-violence intervention, and problem-oriented policing rather than broad, unfocused sweeps. Third, track leading indicators: nonfatal shootings (often an upstream signal for homicide), gun thefts, domestic-violence homicides, and juvenile victimization/offending patterns. Early 2026 sampling suggested murders were still running nearly one-fifth lower than the prior year’s early months, an encouraging sign for continuity.

The lesson from 2025 is not just that murder can fall quickly; it is that coordinated local practice, supported by functional courts and targeted federal partnerships, can bend even volatile crime metrics down at national scale. The country just demonstrated it. The task now is to turn an extraordinary year into a durable baseline.

Sources:

facebook.com, thetrace.org, factcheck.org, whitehouse.gov, usatoday.com, theguardian.com, cnn.com, bbc.com, brennancenter.org

In The Spotlight

© 2026 Conservative Insider. All Rights Reserved.