THE DECADE PRINT · 2015–2026

Everything falls. Extortion doesn't.

Eleven years of Mexico's municipal crime records, on one scale.

Chart: index of investigation files opened, January–July of each year, 2015 = 100. Extortion rises to 176 (completed) and 204 (including attempts) while homicide, business robbery, cargo robbery, vehicle theft and kidnapping all fall.
Investigation files opened, Jan–Jul of each year, indexed to 2015 = 100. Source: SESNSP municipal crime database, July 2026 cut, processed by CORRIDOR.

What we count

Business-risk basket

Homicide, extortion, business robbery, cargo robbery, vehicle theft and kidnapping — tracked monthly at municipal level.

2,510 municipalities

Every indicator joined on the INEGI municipal key, so security and economic data can be read on the same map.

Floors, not measurements

These are files opened, not crimes committed. Most extortion is never reported — and we say so on every page.

Method note

The two universes of 2026

In January 2026 Mexico's SESNSP split extortion, homicide and feminicide into completed and attempted categories. Against Jan–Jul 2025, completed extortion reads −3.6%; completed plus attempted reads +11.7%. Homicide falls −28.8% completed-only, or −17.5% on the basis comparable with the historical series. Neither reading is wrong — they measure different universes. Our rule: we publish both, we say which is which, and we never chart the two definitions as one line.

A provider who hands you one clean extortion number for 2026 picked a universe without telling you.

Coming next

Coming in Issue No. 2 — Announced investment vs. delivered jobs

Investment announcements make headlines. Payrolls make economies. Starting with Issue No. 2, CORRIDOR tracks announcements above US$100M against formal IMSS employment in the receiving municipality — 6, 12 and 24 months after the press release.

We publish the gap and its distribution, not verdicts: municipal payroll data measures the whole municipality, and legitimate projects take time. What nobody should have to guess is the difference.

Get Issue No. 1