Network Model Calibrates Hidden Cocaine Flows Using PCA-Derived Interception Risks

The study constructs an interception-risk metric from three PCA components of trade volume, enforcement intensity, and connectivity, then optimizes flows subject to country supply and demand. Parameters are tuned so modeled routes match known trafficking activity; the resulting map shows substantial traffic on links with minimal observed seizures. This produces quantitative evidence of the waterbed effect where intensified interdiction on one corridor shifts volume to alternate paths.

The approach improves on seizure-only maps by estimating total flows rather than detected fractions alone. It draws on established network interdiction literature and prior displacement studies from the 2010s Andean cocaine surge. Limitations include reliance on containerized trade proxies that may underweight non-container routes and the absence of dynamic adaptation by traffickers.

Future work would strengthen evidence by incorporating longitudinal seizure series and testing predictions against independent enforcement data from 2027 onward. Such validation could directly inform targeted resource allocation rather than blanket interdiction increases.