Mathematician Prieto-Curiel models path to curb Mexico cartels

TL;DR

Mathematician Rafael Prieto-Curiel models cartel dynamics to show recruitment drives violence in Mexico.
Cartels employ 160,000-185,000 people and recruit 350-370 members weekly to replace losses.
Stopping recruitment shrinks groups and cuts homicides, unlike arrests which often increase violence.

The story at a glance

This profile covers Rafael Prieto-Curiel, a Mexican mathematician at Vienna's Complexity Science Hub, whose models challenge common anti-cartel strategies. Key figures include Prieto-Curiel, former president AMLO who dismissed his 2023 study, and his former boss in Mexico City police. It's reported now after Prieto-Curiel won a major science award in 2024 for his work. Mexico has seen over 30,000 killings yearly since 2018 amid high cartel integration.

Key points

In a 2023 study, Prieto-Curiel and colleagues estimated cartels as Mexico's fifth-largest employer with 160,000-185,000 members, needing 350-370 weekly recruits via TikTok, games, and force; AMLO called it "false" without evidence.
His award-winning model simulated 150 criminal groups with alliances and rivalries; thousands of runs showed only curbing recruitment reliably reduced manpower and killings.
Arrests or killing leaders typically create splinter groups and more violence, with doubled prosecutions still projecting higher 2027 deaths than now.
Born in 1987 in Mexico City, Prieto-Curiel studied at ITAM, worked briefly in finance, then led strategic analysis at Mexico City's C5 center.
There, his crime prediction model boosted daily arrests from 1 to 120 and cut response times from 17-20 minutes to 4 by focusing cameras on high-risk areas.
At UCL, he modeled how fear of crime spreads socially and stays high even if actual crime drops, matching Mexico's persistent insecurity.
Now he consults OECD and World Bank; halving recruitment could significantly lower casualties, but full violence drop to 2012 levels would take three years.

Details and context

Prieto-Curiel started at C5 when Mexico City had just 8,000 cameras for 80,000 blocks, monitored poorly by dozens of officers on multiple screens. His three-year data model let operators preempt crimes in hotspots, proving predictive policing's value despite not always boosting public safety feelings.

His UCL simulation used virtual agents in neighborhoods; fear rose from personal or shared stories but faded slowly, explaining why insecurity persists despite crime ups and downs—like Mexico's steady high perceived danger.

The cartel model highlights their societal embed: even optimistic no-recruitment scenarios need years to unwind damage, as peace has fallen 14% since 2015 per Mexico Peace Index, with organized crime killings nearly tripling.

Key quotes

Alejandro Herrera Bonilla, former boss: “at the beginning of the program, we caught one criminal every day,” but after a year of running Prieto-Curiel’s model, they were stopping about 120 suspects each day, and decreasing response time from 17–20 minutes to just four.[[1]](https://mexiconewsdaily.com/business/who-is-rafael-prieto-curiel-the-mathematician-saving-mexico-through-numbers/)[[2]](https://mexiconewsdaily.com/business/who-is-rafael-prieto-curiel-the-mathematician-saving-mexico-through-numbers)
Prieto-Curiel: driven by “the love of science and for the love of his country.”

Why it matters

Cartels' scale as major employers shows violence is structural, not just about kingpins, challenging Mexico's heavy focus on arrests and military ops. For Mexicans, it means policies targeting youth recruitment could save lives faster than current tactics, which simulations link to more deaths. Watch if Sheinbaum's government shifts toward anti-recruitment efforts, though structural changes face big hurdles.