Dynamics and disruption: structural and individual effects of police interventions on two Dutch jihadi networks

Abstrakt

The dynamics of criminal networks over time is considered to be one of the most intriguing aspects of these networks, as the ability to adapt and respond to changes in internal and external environment is crucial for their functioning. Changes may affect both the overall network structure and the tendencies of actors to interact with others, two aspects that must be distinguished because tracking only the structural level may mask variability at the individual level.

This is important considering attempts to disrupt these networks by law enforcement might disrupt the network at the structural level, while triggering contradictory unintended consequences by increasing individual-level connectivity. Longitudinal studies of criminal networks are rare as suitable data is scarce.

We study the structural and individual implications of disruption in criminal networks using a unique longitudinal dataset on two Dutch jihadi networks (n1=57 and n2=26). Both these networks were tracked over two time points - before and after disruption.

At the structural level, both networks seem like mirror opposites - the larger one is becoming sparser, decentralized, with longer distances and slightly increasing transitivity, whereas the smaller network counter-intuitively becomes structurally more cohesive after the disruption. To obtain the actor-level tendencies, we used stochastic actor oriented models (SAOM) to analyse the change between the waves.

The model specification is based on a theory of action which posits that actors in criminal networks under disruption seek security rather than efficiency and try to remain concealed as much as possible. In spite of the differences between the networks, the tendencies of actors are similar in both studied networks with actors being inclined towards triadic closure and translation of pre-existing ties to cooperation ties.

We discuss our findings in the light of their policy implications. We conclude discussing the limitations of SAOMs for networks with large composition change.