Since 2013, synthetic opioid overdoses have surged in the US, claiming over 70,000 lives each year. This crisis has had international spillovers, reigniting drug-related violence in Mexico, where a new wave of cartel warfare erupted after 2013. This column shows that the geography of this second wave is closely tied to fentanyl trafficking routes. Exploiting quasi-natural variation in port access and inland transportation networks, it estimates that exposure to fentanyl trafficking increased homicide rates by up to 20 per 100,000 residents. Fentanyl has grown into a vector of transnational violence; addressing it requires both domestic action and international coordination.
The US opioid crisis, and particularly the rise of fentanyl, has profoundly reshaped global criminal dynamics, yet its international violent repercussions remain underexplored. As Aklin et al. (2023) document, fentanyl has disrupted supply chains and introduced new strategic incentives for transnational criminal organisations. Prior research shows how shocks to the US drug market reverberate in Mexico: cocaine shortages have intensified cartel violence (Castillo et al. 2020), and the decline of the heroin market has altered incentives in opium-producing regions (De Haro 2025). Some studies also examine how drug-related violence affects labour markets and firm behaviour (e.g. Utar 2021). Yet, surprisingly, the most consequential transformation – the explosion of fentanyl consumption since 2013 – has not been systematically linked to cartel violence in Mexico.
Our paper (López Cruz and Torrens 2025) fills this gap, showing how fentanyl trafficking triggered a second, geographically distinct wave of homicides, driven by an external demand shock originating in the US. Figure 1 shows a two-wave trajectory of drug-related violence in Mexico. The first wave was sparked by the Mexican war on drugs, launched in December 2006 by newly elected President Felipe Calderón. Framed as an urgent effort to reassert state control over territories dominated by drug trafficking organisations, the campaign led to the rapid militarisation of public security. Violence escalated sharply, with homicides peaking around 2011 before beginning to decline.
The second wave, beginning in 2013, coincides with the onset of the fentanyl epidemic in the US. As synthetic opioid overdoses surged north of the border, homicide rates in Mexico rose again – this time in regions that had previously remained relatively calm. The simultaneous spike in fentanyl-related deaths in the US and renewed violence in Mexico points to a new, externally driven escalation.
Figure 1 Trends in homicides in Mexico and synthetic opioid overdoses in the US
How trafficking routes were reshaped
Unlike the Mexican war on drugs, which was initiated by internal policy choices, the fentanyl shock was externally generated. US demand for synthetic opioids rose sharply after 2013, and Mexican drug trafficking organisations – particularly those based along the Pacific coast – adapted quickly. As Hamilton (2021) and others have noted, Chinese precursor chemicals began flowing through Pacific ports in Mexico, fuelling a shift in trafficking strategies. Instead of vying for control over traditional south-north heroin corridors – long contested since the Mexican war on drugs – Pacific drug trafficking organisations pivoted eastward, establishing new west-east routes toward the Gulf.
This reorientation reshaped Mexico’s criminal landscape. Violence surged along new fentanyl corridors, while traditional heroin routes saw a decline in homicides. The strategic redirection of trafficking operations – documented in Figure 2, which shows the locations of poppy fields, major ports, and key crossing points – had profound implications for the geography of violence.
Figure 2 Key geographic features relevant to drug trafficking in Mexico
Two empirical strategies: Ports and routes
To identify the causal effect of fentanyl trafficking on violence, we leverage two quasi-natural experiments: one at the ports and one along inland routes.
First, we exploit a geographic distinction in port exposure. Pacific deep-water ports, capable of handling international maritime trade, are directly connected to Chinese shipping routes and were disproportionately targeted by fentanyl-smuggling drug trafficking organisations. In contrast, Gulf coast ports – although also open to international trade – were largely bypassed due to their greater distance from Asia and higher transoceanic shipping costs. To isolate the fentanyl effect, we compare homicide trends in areas hosting these strategic ports across coasts. Using difference-in-differences (DiD), synthetic DiD, and triple DiD approaches, we consistently find that exposure to fentanyl trafficking increased homicide rates by approximately 15 to 18 per 100,000 inhabitants.
Second, we analyse the inland diffusion of violence by modelling how drug trafficking organisations adapted their operations in response to the fentanyl shock. We construct a travel-cost index that incorporates both road distance and cartel territorial control – capturing the relative ease with which a drug trafficking organisation can move drugs from one location to another. Using this index, we solve for the optimal trafficking routes from key fentanyl entry points (Pacific ports) to US border crossings (see Figure 3 for an iconic example). The resulting simulated network reflects the most efficient west-east corridors drug trafficking organisations would likely have used to avoid rival territories and minimise operational risks.
Figure 3 Example of fastest and optimal routes from the Pacific Coast to the US border
Based on this model, we classify municipalities into three groups:
- Those affected by the 2007 Mexican war on drugs, located along traditional south-north heroin routes
- Those affected by fentanyl, situated along the newly identified optimal west-east corridors
- Doubly treated municipalities positioned at the intersection of both route systems
Using a difference-in-differences framework adapted for staggered treatment timing, we compare homicide trajectories across these groups. The results show that fentanyl-affected municipalities experienced a sharp and distinct increase in violence only after 2013, while areas affected by the war on drugs had already stabilised. Doubly treated municipalities display two clear waves of violence, consistent with sequential exposure to both shocks.
Results and magnitude
Our findings consistently point to a large and robust fentanyl effect. The introduction of fentanyl-related trafficking routes increased homicide rates by roughly 20 per 100,000 inhabitants – an effect more than twice the national average increase in the same period. In comparative terms, we estimate that the impact of fentanyl was at least 1.75 times greater than that of the Mexican war on drugs.
These magnitudes are substantial. Homicide increases of this scale can disrupt local economies, reduce educational attainment, and trigger displacement and migration. In many regions, fentanyl trafficking has reproduced – and in some cases exceeded – the destabilising effects of the original 2007 government crackdown.
Credibility of the identification strategy
What allows us to identify the fentanyl effect so cleanly is the exogenous origin of the shock. The reconfiguration of drug trafficking organisations’ strategies was not driven by Mexican law enforcement but by shifts in global demand and supply chains. This quasi-random nature of the treatment, combined with geographic variation in exposure, creates a credible identification strategy. We further confirm the robustness of our results through placebo tests and spatial econometric techniques that account for spillovers.
This approach builds on the logic of prior work, such as Dell (2015), who examined how political turnover diverted drug routes and reshaped violence patterns; Galiani et al. (2016, 2018), who studied how security reallocations affected criminal displacement; and López Cruz and Torrens (2023), who explored criminal diversification and spatial diffusion of violence in Mexico. But unlike those studies, we show that an international demand shock – rather than domestic policies – can significantly reshape criminal incentives and spatial conflict patterns.
Implications for policy
Our findings highlight the international ramifications of US drug demand and the unintended consequences of logistical adaptation by transnational criminal organisations. The rise of fentanyl created new profit opportunities for Pacific-based drug trafficking organisations, prompting a reconfiguration of trafficking routes and the ignition of new territorial disputes. This reorganisation reignited violence in areas stabilised after the Mexican war on drugs and extended conflict into regions previously untouched by drug-related violence.
The scale of the resulting escalation was large enough to reverse the national decline in homicide rates that had followed the 2011 peak of the Mexican war on drugs. From a broader perspective, both the war on drugs and the fentanyl shock triggered a similar pattern: cartels adjusted their logistics, rivalries intensified, and violence surged until a new equilibrium emerged. In this sense, externally and internally induced shocks – whether government crackdowns or shifts in global drug demand – can be equally destabilising when they alter the strategic calculus of criminal organisations. Understanding these dynamics is essential for designing interventions that limit the incentives for reoptimisation or deter escalations over newly valuable routes.
Three key lessons for policymakers emerge:
- Drug policy is regional. Demand-side shocks in one country can destabilise others, particularly in criminal ecosystems where actors adapt quickly.
- Geography shapes conflict. Strategic infrastructure – such as deep-water ports and inland road corridors – plays a critical role in determining which regions bear the brunt of violence.
- Violence is not only supply-driven. Even without intensified enforcement, shifts in trafficking incentives can produce sharp surges in conflict.
In sum, fentanyl has grown into more than a public health crisis: it is a vector of transnational violence. Addressing its effects will require both domestic action and international coordination that acknowledge how deeply intertwined drug markets, logistics, and violence have become.
Source: cepr.org
