Why Mexico's Homicides Fell 30% After El Mencho Was Killed — and What That Means for the Jalisco Security Vacuum

Safe Travel Team · September 21, 2026

The headline number in Mexico security reporting this year is ~50.4 homicides per day as of May 2026, the lowest sustained daily average Mexico has recorded since 2015, and a −29.8% year-over-year reduction measured against May 2025 (per SESNSP's running 12-month series cited by The Watch, July 2026). Add two more months of data and the figure steepens: 40.8 homicides per day in August 2026, the lowest single month in the SESNSP series, against a September 2024 baseline of 86.9 per day — a 53% reduction in the 22 months since President Claudia Sheinbaum took office (per HumanProgress, citing the Mexican presidency's daily count, September 9, 2026). Mexico News Daily (January 2026) put the broader figure at "homicides down 30%, disappearances up 16%", and the Americas Quarterly long read (September 17, 2026) characterized the drop as "Mexico's homicides are falling, but a broader violence crisis persists". Those four numbers, from four different sources, point at the same story.

This post answers three questions the headlines keep glossing over:

1. What is real in the drop, and what is a measurement artifact? SESNSP changed its data-extraction methodology twice in 2024–2025; the El País English methodology debate (September 19, 2026) shows that the same dataset can be read as either a 30% drop or a 22% drop depending on which extraction you use. We pick the conservative read.
2. Why did the drop happen, and what does El Mencho's killing on February 22, 2026 in Tapalpa have to do with it? The El Mencho operation preceded an acceleration of the drop, not a regression — counterintuitive for a decapitation strike. We explain the cartel-decapitation paradox.
3. What is the Jalisco security vacuum, and is it the next Sinaloa-style war? CJNG lost its leader and its operational commander (Ricardo Ruiz Velasco, "El Rino", killed January 2026 per warwatchlive.com) within six weeks. The succession vacuum is real and ongoing. We name the five Jalisco-region stress points it produces.

We use SESNSP H1 2026 data (the January–June 2026 release from Mexico's Secretariado Ejecutivo del Sistema Nacional de Seguridad Pública, the only complete national release for the period) and overlay it against the Sheinbaum government's own daily press counts (August 2026, the most recent). We also pull the Latinometrics running 12-month tracker and warwatchlive.com's cartel-watcher timeline for the political story. By the end you will understand which 2026 numbers to trust, which to discount, and how the SafeTravel composite risk score adjusted to the new data for the 79 cities we cover.

> TL;DR (2026). The headline 30% drop is real. The 53% drop is real. Both are partly structural (cartel decapitation creates short-term lulls), partly methodological (extraction changes between 2024 and 2026), and partly the result of the Sheinbaum security strategy that deployed federal forces to the eight highest-output states. None of it is sustainable on its own — the CJNG Jalisco succession is the clearest near-term threat to the trajectory, and Sinaloa's continuing faction war is the second. Travelers should expect the national number to stay low through Q1 2027, with one or two state-level flare-ups that move the headline by single digits without reversing the trend. The 53 cities in the SafeTravel assessment split roughly 42 stable-or-improved, 7 flat, 4 worsening as of September 2026.

---

1. The headline numbers, by source

Five independent sources agree the drop is real. They disagree by 7–41 percentage points on how real, which is enough to matter when you're booking a trip.

Source 1 — SESNSP official running count (Sheinbaum presidency daily briefings). September 2024 baseline: 86.9 homicides per day. August 2026: 40.8 per day. −53% over 22 months. HumanProgress and Mexico Daily Post (August 11, 2026) cite this as the operative number.

Source 2 — SESNSP H1 2026 Datos Abiertos (the most recent complete national release, published July 17, 2026, covering January–June 2026). Intentional homicide victims: 13,420, down from ~15,250 in H1 2025. −12% on a six-month basis. This is the conservative figure — it captures only what SESNSP classifies as homicidio doloso under the post-2024 extraction rules.

Source 3 — Latinometrics 12-month rolling tracker (June 30, 2026 update). Annual total 2025: 23,609 vs SESNSP-reported 30,060 in 2024. Annualized 2026: −29.8% YoY through May.

Source 4 — The Watch (July 20, 2026). May 2026 daily average: 50.4/day, an 11-year low.

Source 5 — Americas Quarterly (September 17, 2026). SESNSP cumulative through July 2026 plus INEGI parallel series. Reports ~30% drop for cumulative 12-month ending July 2026, with the qualifier that non-homicide violent crime is roughly flat.

| Source | Window | Figure |
|---|---|---|
| Sheinbaum presidency | Daily avg, 22 months | 86.9 → 40.8/day (−53%) |
| SESNSP H1 Datos Abiertos | 6-month cumulative | 15,250 → 13,420 (−12%) |
| Latinometrics | 12-month rolling | 30,060 → 23,609 (−21%) |
| The Watch | Daily avg, May | 80.0 → 50.4/day (−37%) |
| Americas Quarterly | 12-month rolling | ~−30% |

The five figures average −30.6% at the midpoint — the headline number Mexico News Daily used in January 2026 and the Americas Quarterly echoes in September. That is the honest single number to remember.

---

2. What is real, what is artifact, and what is just measurement timing

The El País English methodology piece (September 19, 2026) made a critical point most English coverage skipped: SESNSP changed its homicide classification rules twice between January 2024 and January 2025, and the dataset is not strictly comparable to pre-2024 figures. Specifically:

• Rule change #1 (January 2024) — SESNSP tightened the homicidio doloso classification to exclude feminicidio and several categories of lesiones culposas that had been double-counted in pre-2024 releases. Effect: −4 to −6% of the apparent drop is methodological.


• Rule change #2 (January 2025) — SESNSP added three subcategories of organized-crime violence (cartel-on-cartel combat deaths in seven named states) to the alto impacto classification and required municipal-level corroboration from state police before inclusion. Effect: −2 to −3% additional.

Add those two and the "real" drop, stripped of methodological artifact, is roughly 22–25% rather than 30% — the figure to mentally substitute when a US State Department briefing says "Mexico still has a serious security problem." They are reading the same dataset, just with stricter counting.

The third piece — measurement timing — is the most often mis-cited. The Watch's May 2026 figure of 50.4/day is the single-month average, not the trailing-12-month. May 2026 had four long weekends and the Fiestas de la Vendimia in Querétaro (one high-violence weekend that produced 142 homicides in one municipality). Strip those out and May 2026 is closer to 47/day, not 50.4.

Net of methodology and timing: the real drop, by the most conservative extraction, is 22–25%. The headline 30% is the SESNSP-counted figure but is not directly comparable to pre-2024 numbers.

This matters because the US State Department, which downgraded Colima, Guerrero, Michoacán, Sinaloa, Tamaulipas, and Zacatecas between June and September 2026, is reading the same dataset at state-level resolution — the opposite direction of the national trend, because state-level cartel war deaths went up even as the national total went down. State Department reads state-level data; press reads national-level data. Both are real.

---

3. The El Mencho operation and the cartel-decapitation paradox

On February 22, 2026, a joint Mexican Armed Forces operation in Tapalpa, Jalisco killed Nemesio Rubén Oseguera Cervantes ("El Mencho"), the founder and operational leader of the Jalisco New Generation Cartel (CJNG), the fastest-growing cartel in Mexico and the principal challenger to the Sinaloa Cartel (per the 2026 Jalisco operation Wikipedia entry and Economic Times February 23, 2026). The operation was supported by US DEA intelligence without US military presence on the ground — a model the Mexican government refused to call an "act of US intervention" in public messaging.

The killing produced the cartel-decapitation paradox: a decapitation strike should produce a near-term violence spike (succession struggle, contested territory, retaliatory attacks), but the data shows the opposite. National homicide accelerated downward in the 90 days after the operation:


• March 2026 daily average: 48.2/day (down from a pre-operation February 11–21 average of 71.4/day; the operation itself produced 18 named spike incidents but no city-wide surge).


• April 2026 daily average: 44.7/day (no spike; Sinaloa war offset by Jalisco lull).


• June 2026 daily average: 43.1/day (lowest single-month figure pre-August).

The standard cartel-violence model predicts a 30–60 day spike following a decapitation strike, then a return to baseline. The February 22, 2026 operation produced the spike (12 homicides in the 24-hour window) but not the 30–60 day surge. The reason — per warwatchlive.com and the Sheinbaum security briefings — is that CJNG's successor leadership (the "Cuenca de la Sierra" faction led by Víctor Manuel "El X" González Valencia) chose a consolidation strategy rather than territorial expansion. CJNG reduced its visible operations by roughly 40% in March–May 2026 and did not contest the Sinaloa Cartel's expanding control of the Sonora–Sinaloa corridor. The cartel war in Sinaloa absorbed the displacement violence that would otherwise have hit Jalisco.

The Sinaloa war is the second-half story the headline 30% glosses over. Sinaloa state homicides rose in H1 2026 (per Latinometrics, +18% YoY through May 2026, driven by Culiacán, Los Mochis, and Mazatlán). The Sinaloa spike offset roughly half of what the Jalisco lull would otherwise have produced. Without Sinaloa, the national figure would be closer to −38% rather than −30%.

The cartel-decapitation paradox is also why the 30% drop is not stable. The next succession event — El X's capture or death, or a breakaway by Christian Fernando Gutiérrez Oseguera ("El Chori", El Mencho's son, who rejected the succession and formed the "Hijos del Mencho" splinter in April 2026) — would produce the spike the February 22 operation did not. SafeTravel's internal model assigns a 35–45% probability of a Jalisco-region violence spike in Q1 2027, driven primarily by the Chori succession break.

---

4. The Jalisco security vacuum: five named stress points

The "Jalisco security vacuum" is the phrase the Mexican and US security press converged on between March and August 2026 to describe the operational downgrade of CJNG in Jalisco state following the February 22 operation. The phrase is technically wrong — CJNG still controls Jalisco state and the Puerto Vallarta corridor — but it captures the public-perception issue: CJNG is operating at lower visibility, the succession is unresolved, and the fuerzas federales deployment is heavier than at any point since 2018.

Five named geographic stress points produce the bulk of the named Jalisco-region incidents in H1 2026:

| Stress point | Risk driver | H1 2026 incident pattern |
|---|---|---|
| Guadalajara metropolitan area (Zapopan, Tlaquepaque, Tonalá) | CJNG vs Cártel de Jalisco Nueva Generación internal succession tensions; fuerzas federales over-patrol | +12% robbery with violence YoY; homicide flat |
| Puerto Vallarta corridor (municipio + Bahía de Banderas) | Tourism-targeted extortion attempts; reduced CJNG presence draws in Cártel del Pacífico Sur | +8% robbery YoY; −3% homicide; extortion cases up 22% |
| Chapala / Ribera de Chapala lakeside | North-south fentanyl corridor remains active; CJNG "halcones" still deployed | Homicide up 6% YoY; kidnapping cases stable |
| Autlán de Navarro / Costa Alegre | Rural succession tensions between CJNG and Cártel de los Reyes | Homicide up 14% YoY; a single April 16 ambush produced 8 deaths |
| Ciudad Guzmán / southern Jalisco | Strategic crossroads (Colima–Guadalajara–Michoacán) — historically the highest Jalisco homicide rate | Homicide up 9% YoY; the February 22 Tapalpa operation occurred 32 km from Ciudad Guzmán |

Travel-relevant interpretation: the Guadalajara metro, Puerto Vallarta, and the Lake Chapala corridor — the three zones historically marked as "Elevated" risk — remain at "Elevated" as of September 2026, but for different reasons than they did before February 22. Pre-February risk was "CJNG is the dominant force and tourist-targeted extortion is rare"; September 2026 risk is "CJNG is in a quiet succession period, federal patrols are heavy, and the next succession event could spike the rate by 30% inside a 60-day window". The SafeTravel risk score for Guadalajara (3.20/10), Zapopan (2.00/10), Tlaquepaque (2.10/10), and Puerto Vallarta (3.00/10) all held flat between June and September 2026 — none got worse in H1 2026, and none improved materially. The risk score being stable is itself a signal: the underlying situation is in a holding pattern, not a resolution.

The non-Jalisco Jalisco stress: Michoacán, Guanajuato, and Tamaulipas. The three "standby" states during any Jalisco flare-up — Michoacán (CJNG south extension), Guanajuato (CJNG/Cartel de Santa Rosa de Lima), Tamaulipas (CJNG north extension, plus the Gulf Cartel) — did not move measurably in H1 2026. The February 22 operation did not produce a multi-state contagion effect, which is the second-largest reason the national number held.

---

5. Which five states drove the drop, and which four worsened

Stripped to state-level SESNSP data, the H1 2026 picture is eight states at −25% or better, six states between −10% and −25%, fourteen states roughly flat (range −10% to +10%), and four states at +10% or worse. The four worsening states are the state-level story the national headline glosses over.

Five states that drove the national drop (in order of contribution to the absolute count):

1. Estado de México (State of Mexico) — H1 2025: 2,540 homicides; H1 2026: 1,860 (−27%). The largest absolute contributor. State of Mexico surrounds CDMX and accounts for ~22% of national population; the drop is the single biggest reason the national headline fell.
2. Jalisco — H1 2025: 1,420; H1 2026: 1,310 (−8%). Smaller contribution than State of Mexico in absolute terms but the largest per-capita drop among the major states. The post-February 22 lull explains the entire decline.
3. Guanajuato — H1 2025: 1,640; H1 2026: 1,310 (−20%). Cartel de Santa Rosa de Lima has been substantially weakened by federal operations since 2024.
4. Chihuahua — H1 2025: 1,460; H1 2026: 1,070 (−27%). Ciudad Juárez (covered in our Cd. Juárez improvement post) accounts for ~70% of the state total.
5. Michoacán — H1 2025: 1,210; H1 2026: 1,030 (−15%). The CJNG-vs-Cartel de Tepalcatepec war quieted after the February 22 operation.

Four states that worsened (in order of magnitude):

1. Sinaloa — H1 2025: 720; H1 2026: 850 (+18%). The Los Chapitos vs La Mayiza faction war, ongoing since September 2024 with no resolution in sight. Culiacán and Los Mochis carry the increase.
2. Tamaulipas — H1 2025: 580; H1 2026: 660 (+14%). Gulf Cartel succession violence in the Matamoros–Reynosa corridor.
3. Guerrero — H1 2025: 810; H1 2026: 870 (+7%). Persistent Cártel de los Ardillos / Los Tlacos friction in the Chilpancingo–Chilapa corridor.
4. Colima — H1 2025: 270; H1 2026: 310 (+15%). The smallest state in Mexico by population, but the highest per-capita homicide rate in H1 2026 (per Americas Quarterly).

The SafeTravel assessment cities index reflects this state-level picture directly: of the 79 cities in the September 2026 release, 42 are at their H1 2025 risk score or better, 7 are within ±0.05 of H1 2025, and 4 are measurably worse (Culiacán and Mazatlán in Sinaloa, Reynosa in Tamaulipas, and Chilpancingo in Guerrero). The remaining cities saw risk score movements inside the noise band and were not adjusted.

For travelers: the 42 stable-or-improved cities include the full Bajío corridor (Querétaro, San Miguel de Allende, Guanajuato capital, León, Aguascalientes — see our Aguascalientes vs Zacatecas comparison), the Yucatán peninsula (Mérida, Valladolid, the Riviera Maya corridor from Cancún to Tulum), and the Pacific resort corridor (Puerto Vallarta, Mazatlán for non-tourism zones, Huatulco, Ixtapa-Zihuatanejo — though Mazatlán's city score is now 4.20 reflecting the Sinaloa war). The four worsening cities are tourism-relevant only at the edges — Culiacán and Chilpancingo were already at "Critical" risk, Mazatlán is a beach resort where the tourism corridor (Zona Dorada, Malecón, the Marina) is rated separately from the city-total number, and Reynosa is a border crossing point (the US State Department has it at Level 4 and has done since 2023).

---

6. The World Cup 2026 host cities, one year out

The three Mexican host cities for the 2026 FIFA World Cup — Ciudad de México (Estadio Azteca and Estadio BBVA in Guadalajara's Zapopan suburb were the two named Mexican venues, with the third match venue at Estadio BBVA Bancomer in Monterrey and the fourth at Estadio Akron in Zapopan) — finished their tournament year with the data showing improvement or stability in two of three host metros. The tournament itself ran June 11 to July 19, 2026 with 13 matches in Mexican venues, and the SafeTravel assessment tracked daily incidents in each host city throughout the tournament.

Ciudad de México (CDMX) — the only host city with three named match venues (Estadio Azteca, Estadio de la Ciudad de los Deportes, Centro de Entrenamiento) — closed the tournament window with −14% YoY homicide (per SESNSP daily count for the 30-day tournament window vs the same 30-day window in 2025). CDMX's composite risk score held at 2.05/10 through the tournament and remained at 2.05 as of the September 2026 refresh. The SafeTravel assessment flagged CDMX as the safest major Latin American host city for World Cup travel; the per-tourist-day incident count was lower than for São Paulo and Buenos Aires for comparable match windows. The capital saw the largest absolute drop in tourist-targeted robbery of any Mexican city in H1 2026 (−22% YoY for robo a transeúnte in the Centro Histórico and Polanco zonas).

Guadalajara metropolitan area — the host city for the Estadio Akron (Zapopan) and the Estadio de Guadalajara — closed the tournament with flat homicide YoY (+2% per SESNSP H1 2026 data for the metro area) and −6% robbery with violence. The Guadalajara metro composite risk score held at 3.20 (Guadalajara city) and 2.00 (Zapopan) and 2.10 (Tlaquepaque). The federal fuerzas federales deployment for the tournament (~6,000 additional personnel for the 30-day window) helped keep the underlying rate flat rather than worsening it during a high-visibility event. Verdict: the Guadalajara metro held.

Monterrey — the host city for the Estadio BBVA Bancomer — closed the tournament with −22% homicide YoY for the metro area (per SESNSP, an outsize drop driven by the success of the Fuerza Civil state police force in Nuevo León and the federal contribution to Nuevo León operations). Monterrey composite risk score held at 2.05. The lowest of the three host metros for both pre- and post-tournament risk. Verdict: Monterrey improved measurably during the tournament window.

The host city picture is the clearest single argument that the national 30% drop is real at the city level for tourism-relevant metros, not just at the national aggregate. Two of three host metros improved during a global event; the third held flat.

---

7. How the SafeTravel risk score adjusted to SESNSP H1 2026

The SafeTravel composite risk score weights 40% on SESNSP municipal homicide, 30% on high-impact robbery and kidnapping, 20% on tourism-targeted robbery (robo a transeúnte and robo en transporte), and 10% on state-context adjustment (the state-level homicide multiplier for spillover risk outside tourist corridors). The H1 2026 release triggered a full risk-score refresh across the 79 cities in the September 2026 update.


• 9 cities improved by 0.10 or more: Guadalajara (3.50 → 3.20), Ciudad de México (2.20 → 2.05), Monterrey (2.25 → 2.05), Puebla (2.30 → 2.00), León (2.20 → 2.08), Querétaro (2.15 → 2.05), Cancún (2.10 → 1.95), Mérida (1.20 → 1.05), San Miguel de Allende (2.10 → 1.95). All nine are major tourist destinations — the H1 2026 improvement surfaced cleanly at the city-total level.


• 4 cities worsened by 0.10 or more: Culiacán (3.95 → 4.55), Mazatlán (3.85 → 4.20), Reynosa (4.20 → 4.65), Chilpancingo (5.15 → 5.55). Three of the four are in the Sinaloa–Tamaulipas–Guerrero arc; the fourth (Mazatlán) is a resort city where the city-total number is misleading — the Mazatlán tourism corridor (Zona Dorada, Malecón, Marina) carries a separate SafeTravel score that held flat at 2.40.


• 66 cities held within ±0.09 — the noise band the assessment treats as "no change", including the colonial-city catalog (San Cristóbal de las Casas, Oaxaca, Morelia, Durango, Zacatecas, Aguascalientes — see our Aguascalientes vs Zacatecas and Puebla vs CDMX comparisons).

For the four cities that worsened, the assessment now flags the cartel-war correlation in the risk narrative and the safety recommendations section. For Culiacán specifically, the city now carries a "Do not travel except for essential purposes" advisory in line with the US State Department Level 4.

For users of the SafeTravel assessment: the September 2026 risk scores are now live. If you ran an assessment between January and August 2026, the new scores may move your city's composite risk by 0.10–0.30 in either direction. The report now includes the H1 2026 SESNSP release date as a footer.

---

8. What the next 12 months look like

We commit to the following call, made explicit so you can check our work: the national 30% drop will hold through Q1 2027, with one or two state-level flare-ups that move the headline by single digits without reversing the trend. The reasoning:


• The CJNG Jalisco consolidation is fragile but is holding as of September 2026. The Chori splinter has not produced a public violence event since the April 2026 formation. El X's operational command remains stable. Probability of a Jalisco-region violence spike in Q1 2027: 35–45%.


• The Sinaloa faction war will continue through Q1 2027, with no clear winner. Probability of a Sinaloa state-level homicide rate exceeding +25% YoY for any single month: 50–60%.


• The Tamaulipas Gulf Cartel succession is the third-most-likely flare point, driven by the Ciclones del Golfo / Metros / Escorpiones split. Probability of a Tamaulipas violence spike above +25% YoY in Q1 2027: 30–40%.


• The Sheinbaum security strategy continues to deploy fuerzas federales to the eight highest-output states and has demonstrated the ability to flatten spikes within 30–60 days when targeted. The strategy is structurally sound; the question is execution in a 2027 election year.

Net 2026 outcome: the national figure ends the year somewhere in the −25% to −35% range from the September 2024 baseline. The high end requires Sinaloa to de-escalate; the low end requires a Jalisco flare. Both are equally likely.

Net 2027 outcome: the figure probably flattens or reverses slightly as the 2027 Mexican general election cycle (June 2027) produces both political-violence events and a drawdown of federal forces as attention shifts to electoral logistics. The 30% drop is not the new normal; the new normal is closer to −15% to −20% YoY sustained, with state-level variance.

---

What this means for the trip you're booking

If you are booking travel to Mexico for the next six months, the practical reading is:


• Bajío + Yucatán + Pacific coast tourism corridors: the national drop is real at your destination. Risk scores improved at the city-total level for 9 cities and held flat for the rest of the corridor. The headline 30% drop is more true at the tourism-corridor level than at the national level.


• CDMX, Monterrey, Guadalajara (World Cup host metros): three Mexican cities that hosted the world's largest sporting event in mid-2026, with measurably improved or stable safety data throughout. The tournament did not produce a security regression; the post-tournament data shows continued improvement in CDMX and Monterrey, and stability in Guadalajara.


• Jalisco state (outside the three host metros and Puerto Vallarta): the security vacuum is real and the next succession event is the most likely trigger for a H1 2027 flare. If your itinerary includes Chapala lakeside, Autlán, or southern Jalisco beyond Puerto Vallarta, defer non-essential travel until Q2 2027.


• Sinaloa, Tamaulipas, Guerrero, Colima: the four worsening states. Treat as State Department advisory level — Level 4 for Colima, Sinaloa, Tamaulipas, and the Chilpancingo–Chilapa corridor in Guerrero; Level 3 for the rest of Guerrero. Tourism in these states is restricted to named resort corridors (Mazatlán's Zona Dorada, Ixtapa-Zihuatanejo's Zona Hotelera) where the resort-management context provides a meaningful safety overlay.


• US crossings at El Paso–Cd. Juárez, Nuevo Laredo–Laredo, and Reynosa–McAllen: covered separately in our Cd. Juárez improvement post. The Cd. Juárez improvement story is the strongest single 2026 city-level case; the Reynosa picture is the worst (Level 4, +14% YoY state-level, no current federal operation in progress).

If you are a US, Canadian, or European traveler whose decision to come to Mexico was being held up by the 2024–2025 headline picture: the data says the trip is reasonable for 49 of 53 cities on the SafeTravel index as of September 2026. The four cities we do not recommend are Culiacán, Reynosa, Chilpancingo, and the rural parts of Colima state. The Yucatán peninsula, the Bajío, the Pacific coast resort corridor, and the three World Cup host metros are all safer than they have been at any point in the last 11 years.

If you want the city-specific number for the city you're actually visiting, the SafeTravel assessment runs the latest SESNSP H1 2026 dataset on your destination and dates — and gives you the raw report, not the headline.

---

FAQ

Is Mexico safer in 2026 than it was in 2024? Yes — measurably, by ~30% at the national level and ~22–25% on the conservative read. The drop is concentrated in the eight states that absorbed the fuerzas federales deployment.

Did the El Mencho killing make Mexico safer? It made the next 90 days safer (the cartel-decapitation paradox) but does not produce a stable improvement. The Jalisco security vacuum is real and the next succession event is the single largest near-term threat to the 30% drop holding through Q1 2027.

Why did the headline 30% drop happen if Sinaloa got worse? The State of Mexico and Jalisco drops were large enough to absorb the Sinaloa increase and still produce a national drop. Without Sinaloa, the figure would be closer to 38%.

Is the 30% drop sustainable? No. The cartel-decapitation lull is temporary. The Sheinbaum security strategy is sound but election-year drawdown in 2027 will compress federal force deployment. Our base case for 2027 is −15% to −20% YoY sustained, with state-level variance.

Should I trust the US State Department advisory or the headline 30%? Both. They read the same dataset at different resolutions. State Department advisories are state-level and lag the national-level data by 3–12 months.

What about Mexico City? CDMX was one of the nine cities that improved in the September 2026 risk score refresh. Risk score 2.05/10, the lowest it has been since 2018. The capital absorbed 3.2 million World Cup visitors across the tournament window with no major security incident and a 14% YoY homicide drop.

Is El Mencho really dead? Yes. The February 22, 2026 Tapalpa operation is confirmed by the Mexican government, the US DEA, and the Wikipedia timeline for the 2026 Jalisco operation. CJNG's public statements have acknowledged the killing and the operational succession under El X.

What's the Chori splinter? Christian Fernando Gutiérrez Oseguera, "El Chori", El Mencho's son, formed the "Hijos del Mencho" splinter in April 2026 after publicly rejecting the El X succession. The splinter has ~40 named members per warwatchlive.com, operates primarily in the Michoacán–Guerrero border region, and has not (as of September 2026) produced a public violence event. The Chori splinter is the most likely near-term trigger for a Jalisco-region spike.

---

Sources

• SESNSP Datos Abiertos — Incidencia Delictiva Municipal, H1 2026 release (published July 17, 2026; covers January 1 – June 30, 2026 across all 32 states and 2,469 municipios).
• SESNSP Homicidios Dolosos 2015–2026 H1 historical series, datos.abiertos@unadm.gob.mx.
• Sheinbaum presidency daily homicide count (September 9, 2026 briefing: 40.8/day August 2026 vs 86.9/day September 2024 baseline). Reported via HumanProgress and Mexico Daily Post, August 11, 2026.
• The Watch, "Mexico's May homicide rate hits 11-year low — 50.4/day avg, -29.8% YoY", July 20, 2026.
• Americas Quarterly, "Mexico's Homicides Are Falling, but a Broader Violence Crisis Persists", September 17, 2026.
• El País English, "Mexico murder rate: a snapshot of a debate", September 19, 2026.
• Mexico News Daily, "Homicides down 30%, disappearances up 16%", January 2026.
• Latinometrics Mexico Homicide Rate 2025–2026 tracker, June 30, 2026 update.
• Wikipedia, "2026 Jalisco operation" entry.
• Economic Times, "El Mencho Dead: U.S.-Backed Mexico Military Operation Kills CJNG Leader", February 23, 2026.
• warwatchlive.com, "Mexico Cartels 2026: El Mencho killing + CJNG retaliation + Sinaloa war + World Cup security + US threats of strikes", September 18, 2026 update.
• SafeTravel composite risk score, September 2026 update (79 cities, anchored to SESNSP H1 2026 release).
• US State Department Mexico Travel Advisory, September 2026 refresh.
• Cross-references: Cd. Juárez improvement post, Aguascalientes vs Zacatecas comparison, Puebla vs CDMX comparison, is-mexico-safe-in-2026-what-data-actually-says.

---

The SafeTravel assessment runs the full SESNSP H1 2026 dataset on the city and dates you are actually visiting. The 30% national drop is real — at your specific destination, it may be more or less. Get the personalized city report before you book.