Liberty Evidence Engine

Evidence · 2026-08-16

Crime pockets: how concentrated is U.S. homicide in major cities?

National murder totals hide geography. Using 2023 major-city homicide counts against the FBI national murder estimate, see how much of the total sits in the top 5 / 10 / 15 / 25 cities — and what happens if you drop the largest city. Interactive cuts on real counts; not a causal gun-law model.

Download city CSV Methods

Interactive cut · real city counts

What share of the national murder estimate sits in the top cities?

Homicides in cut

—

Share of national estimate

—

National denominator: 18253 (FBI estimated U.S. murder and nonnegligent manslaughter, 2023). Recomputed in your browser from the same CSV you can download — no black box.

Preset concentration cuts

cut cities_in_cut homicides_in_cut share_of_national_murder_estimate_pct approx_population_in_cut note
Top 5 cities in panel 5 2158 11.8 15414199 Chicago, Philadelphia, Memphis, New York City, Houston
Top 10 cities in panel 10 3519 19.3 22104164 + LA, DC, Baltimore, Detroit, Dallas
Top 15 cities in panel 15 4490 24.6 26960012 + Indianapolis, New Orleans, Phoenix, Kansas City, Milwaukee
Top 25 cities in panel 25 5695 31.2 35223228 All cities in this large-city panel
Top 10 excluding Chicago 10 3148 17.2 20908486 Sensitivity: one-city dependence
All 25 cities in panel 25 5695 31.2 35223228 Large cities only - not every US city

Methods & limits

Question. When policy debates quote one national homicide or murder number, how much of that harm is concentrated in a relatively small set of major urban places — and how sensitive is that picture to which cities you include?

  • This page measures geographic concentration, not the effect of any gun law. It is an anti-omitted-variable exhibit: national slogans bury urban structure.
  • City numerators are 2023 calendar-year homicide/murder counts as reported in major-city police year-end figures and widely cited local compilations (city proper / primary agency, not always metro CSA).
  • National denominator is the FBI Crime in the Nation 2023 estimated total for murder and nonnegligent manslaughter (18,253). City PD counts and FBI estimates are related but not definitionally identical; shares are illustrative concentration, not an accounting identity to the last case.
  • Interactive controls recompute from the downloadable city CSV in the browser: choose top N by homicide count; optionally exclude the single largest city in the panel (Chicago) to test one-city dependence.
  • Population figures are approximate recent city estimates for rate context only; ranking for cuts is by homicide count, not rate.
  • We do not estimate shall-issue effects, ownership elasticities, or John Lott-style causal models here. Those are out of scope for this page.
Limitations. Panel is 25 large reporting cities, not a census of all U.S. places. Within-city crime is even more concentrated in a few neighborhoods — this page stops at city totals. Agency definitions, late clears, and justifiable homicide handling can differ. NIBRS/SRS transition affects national series. Do not treat share-of-national as proof that rural America has no violence or that any statewide statute caused the pattern. Tract-level pocket maps and full county panels are future work.

Sources.

  • FBI — Crime in the Nation / UCR murder estimates (2023) — National murder and nonnegligent manslaughter estimate used as denominator.
  • Major-city police year-end crime statistics (2023) — City homicide counts from primary agency YE releases and contemporaneous local reporting (NYPD, LAPD, CPD, etc.). See source_note column in the city CSV.
  • Concentration literature (context) — Crime concentration in small geographic units is a long-standing empirical regularity (e.g. Weisburd law of crime concentration). This page is a transparent national-to-city illustration, not a full neighborhood analysis.