“Killing drug lords gets headlines, but complexity analysis suggests they are the wrong people to target to bring down a cartel”.
New Scientist put together an interesting piece called ‘Destroying drug cartels, the mathematical way’. The article focussed on the recent death of Lazcano the leader of the Mexican cartel Los Zetas and on how mathematics, modeling and complexity theory are being used to fight drug wars like these.
My research looked at similar drug fueled conflicts and we found that the mathematical signatures that defined the violence in places like Colombia looked very similar to the signatures in more traditional type insurgencies like Iraq. The way that drug cartels organise themselves, evolve and compete with each other is very similar to the dynamics seen within insurgent groups in places like Iraq and Afghanistan. This is because there are only a few effective ways to organise a force against a conventional army/paramilitary unit – and either the groups evolve to find this solution or they die trying. The upside of this though is that strategies employed in Iraq to break apart insurgencies can be used to inflict damage to the Mexican drug ecosystem.
The key parts of the insurgent dynamics involve coalescence (possible over large distances and biased towards larger groups) and fragmentation that breaks the groups up into thousands of different pieces. This continued process of coalescence and fragmentation allows many different strategies to be tried out by the drug cartels and for the best ones to propagate. Removing the head of the organisation makes it more likely for the group to fragment and the process of innovation continues as the people and resources are redistributed to the most promising emerging groups. Indeed since 2006, 25 of Mexico’s 37 drug cartel leaders have been killed — whilst at the same time violence has continued with over 60,000 people killed during that same period.
So we have an interesting dynamic at play in the Mexican drug war. If you kill the heads of the cartels — you continue to provide the fuel to the insurgency that allows them to rapidly innovate in violence (we are now seeing RPGs on the ground and IEDs in the streets). But if you don’t ramp up the military/government offensive then you are left with drug cartels running parts of the country. Standard military theory here suggests that if you increase forces on the ground you get a corresponding monotonic decrease in the expected length of the conflict. But with the ecosystem model for insurgents, you have a tipping point. Below this tipping point adding troops increases the conflict — but get over this tipping point and you rapidly end the violence (see paper here). That tipping point comes when your forces have a strength of 13:1. Which from the sounds of things on the ground in Mexico is a long way away from where things are. One solution for this type of ecosystem is then is to target the mid sized groups, leave the large groups alone and effectively decouple the process of coalescence.
Using Quid software we are able to collect documents about the Mexican drug violence from multiple information streams, extract out the key entities and relationships and start to construct a virtual representation of the drug cartel ecosystem. This analysis allows us to start to understand the complexity of the system and we can see how the evolution of the different players is shaping the landscape.