Conflict Archive · Research Essay
Mexico’s Drug War Since 2006: Cartel Fragmentation, State Force, and Organized Violence
Mexico’s organized violence includes cartel rivalries, state operations and civilian targeting, but homicide and criminal violence do not automatically meet UCDP armed-conflict criteria.
Mexico’s Drug War Since 2006: Cartel Fragmentation, State Force, and Organized Violence
Mexico’s escalation of military operations against drug-trafficking organizations in 2006 marked a major phase in the country’s modern security crisis. Since then, organized violence has been shaped by federal and state security strategies, competition among criminal organizations, fragmentation of larger groups, and local struggles over trafficking routes, extortion and territorial control. The term “drug war” is widely used, but it does not automatically mean that every violent incident qualifies as an armed conflict under a particular dataset’s definition.
From major organizations to fragmented networks
Mexico’s criminal landscape has repeatedly changed. Large organizations have split after leadership arrests, internal disputes and enforcement pressure. Regional factions may compete over trafficking routes, ports, border crossings, local markets and extortion economies. Names and alliances can shift quickly, and a group’s claimed affiliation does not always prove stable command or control.
Violence can involve criminal organizations fighting one another, attacks against police or military forces, and killings of civilians or public officials. These categories are related but should not be merged without evidence. A homicide associated with organized crime is not automatically a battle death, and an incident involving a cartel label may be misattributed or contested.
State strategy and public security
Federal deployments and operations against major organizations have altered the balance of power in different regions. Arrests or killings of senior figures can weaken a group but may also trigger succession disputes and fragmentation. Local police, prosecutors, military units and state governments operate under different capacities and constraints, creating uneven security outcomes across the country.
Assessing these strategies requires more than comparing annual homicide totals. Homicide statistics include varied motives and circumstances; missing-person records, discoveries of clandestine graves, extortion, forced displacement and attacks on journalists or candidates measure other aspects of the crisis. Each indicator needs its own definition and should not be treated as a direct substitute for conflict-event data.
Civilian consequences
Communities may face extortion, recruitment, threats, disappearances, forced displacement and restrictions on daily movement. Journalists and local officials investigating criminal activity can be at particular risk. Some residents adapt their routines to avoid checkpoints or contested routes; businesses may pay protection demands or close. These consequences are often poorly represented by fatality statistics alone.
Reliable attribution is difficult where groups use shifting names, where witnesses fear retaliation and where official investigations are incomplete. The archive should distinguish confirmed facts, allegations and analytical inference. It should also avoid stigmatizing entire towns or communities as criminal because armed groups operate nearby.
Dataset classification and limits
For this archive, Mexico is a case of organized armed violence requiring explicit classification, not a blanket assertion that all cartel activity constitutes a conventional war. UCDP’s inclusion criteria and actor definitions should be applied to specific organized conflicts and events. State-based, non-state and one-sided violence categories capture different relationships; homicide and disappearance datasets provide complementary but non-equivalent evidence.
Any casualty series must state whether it counts intentional homicide, conflict-related deaths, battle-related deaths or another measure. Figures from different definitions should not be added together. The archive should also separate the national overview from regional theatres where evidence supports a narrower chronology.
Historical dataset coverage through 2025 should be reconciled with source-specific records. Events in 2026 remain provisional until date, location, actor and classification have been checked. The absence of a qualifying UCDP conflict record would not imply an absence of severe criminal violence or civilian harm.
Sources and further reading
- Mexico’s National Institute of Statistics and Geography (INEGI), homicide statistics.
- Mexico’s National Search Commission and official missing-person resources.
- International Crisis Group reporting on criminal governance and regional violence in Mexico.
- Uppsala Conflict Data Program datasets, definitions and codebooks.