How did homicide and violent-crime rates vary across major U.S. cities between 2019 and 2025?
Executive summary
Homicide and violent-crime rates across major U.S. cities fell substantially from the pandemic-era peaks through 2025, with analyses showing a 14% decline in homicides in the first half of 2025 versus the same period in 2019 and steep year‑over‑year drops between 2024 and 2025 in many places [1] [2]. Yet the progress is uneven: a minority of cities drove much of the national improvement, more than half of sampled cities still had homicide levels above pre‑2020 rates, and discrepancies across data sources and jurisdictions complicate definitive city‑by‑city comparisons [2] [1] [3].
1. National arc: from 2020 surge to multi‑year decline
After a surge in homicides during 2020–2021, crime trends reversed beginning in late 2022 and accelerated into 2024–2025, with CCJ finding 6% fewer homicides in its 2024 city sample compared with 2019 and a 14% reduction in the first half of 2025 versus the first half of 2019 [4] [1]. Independent trackers and preliminary FBI figures echoed sharp recent declines—preliminary FBI data showed an 18% drop in homicides between Sept. 2024 and Aug. 2025 and the Real‑Time Crime Index recorded roughly a 20% drop in murders in 2025 versus 2024—underscoring that the downward momentum is visible across multiple datasets [5] [6].
2. City-level winners: big drops concentrated in specific places
Some large and historically violent cities posted especially large improvements: New York reported steep declines in 2025, with shootings and homicides down sharply compared with 2024 and 2021 highs, while CCJ and other aggregators highlight very large percentage declines in places such as Baltimore and St. Louis when comparing 2024/2025 to 2019 baselines [5] [7] [8]. Analysts caution, however, that a small number of high‑homicide cities often dominate national patterns—CCJ notes that the late‑2022 decline was driven by big drops in a few cities, meaning national percentages can mask local variation [2].
3. Uneven recovery: many cities still above pre‑pandemic levels
Despite overall declines, CCJ and Stateline report that more than half of their sampled cities remain above pre‑2020 homicide rates, and only about 38% of their sample were below pre‑2020 levels when looking across rolling 12‑month periods in earlier CCJ reporting—an indication that the recovery is far from uniform and that headline national declines coexist with persistent local hotspots [2] [3].
4. Violent-crime composition shifted even as totals fell
Beyond counts, analysts highlight a troubling shift in lethality: declines in non‑fatal violent incidents accompanied increases in the share of violent incidents that became lethal, meaning fewer violent events overall but a higher probability a violent incident resulted in death [2]. CCJ’s mid‑year 2025 analysis reported fewer aggravated assaults, robberies, and sexual assaults compared with 2019, yet flagged rising lethality as a policy concern [1] [2].
5. Data caveats and competing narratives
Comparing cities across 2019–2025 is complicated by differences in which jurisdictions report timely data, the sampling frames used by researchers, and reliance on preliminary local or FBI releases; CCJ’s findings derive from a selected set of 40+ cities with available data and the FBI’s formal annual report lags into 2026, leaving room for revisions and alternative interpretations [4] [9]. Media coverage emphasizing dramatic local declines (e.g., New York) can feed optimistic narratives, while victim accounts and household surveys cited by some outlets suggest underreporting and divergent lived experience in certain communities [5] [10].
6. Bottom line for policy and public understanding
The best available mid‑2025 snapshots show a meaningful rollback of the pandemic‑era homicide spike across many large cities and a broader drop in violent crime categories, but the gains are concentrated, not universal, and offset by increases in the lethality of violence and continued high rates in a subset of cities—facts that argue for targeted local interventions, continued data transparency, and cautious interpretation of national headlines [1] [2] [3].