Criminal networks are known for their ability to reorganise and continue operating even after major law enforcement crackdowns. New research from the University of Amsterdam’s Informatics Institute and Institute for Advanced Study shows why: the networks adapt in ways that reduce their visibility while maintaining, and in some cases increasing, their levels of criminal activity. The study was published in the journal Proceedings of the National Academy of Sciences on 15 December.

The researchers – in collaboration with the Dutch National Police – analysed unique intelligence data on criminal organisations operating in the Netherlands and combined it with computational modelling to understand how such networks respond to disruption. The researchers examined how individuals in these networks take on different roles (e.g. transporter, broker, financier or coordinator) and how their relationships shift when authorities intervene.

Smaller, more organised groups

The results showed how targeted actions, such as arresting key individuals or encouraging them to leave criminal activity, can fragment the network. However, instead of collapsing, the network often reorganises into smaller, more specialised groups. To understand this, the researchers developed a game theoretical model in which individuals weigh the benefits of collaboration against the risks of being detected. The model was able to reproduce the same patterns seen in real intelligence data: when the network is pressured, its structure shifts in a way that reduces visibility while preserving criminal coordination. This outcome is known as criminal opacity amplification, because the network becomes harder to monitor at precisely the moment when enforcement pressure is highest.

Changes to the strategic environment

‘Interventions don’t simply remove criminals from the system. They change the strategic environment,’ says study author Dr Casper van Elteren. ‘When criminal networks adapt, they may become less visible, more decentralised, and in some cases temporarily more active. This means well-intended strategies can unintentionally strengthen the resilience of organised crime.’

The study also shows that there is no single ideal intervention strategy. Actions aimed at individuals with central positions in the network can be effective, but only under specific conditions. Random or poorly timed interventions, by contrast, run the risk of making the problem worse.

Not just who, but how and why

‘These networks behave like complex adaptive systems found in nature. Their members adapt their connections to keep the organisation functioning,’ says co-author Dr Vítor V. Vasconcelos. ‘To disrupt them effectively, we must understand not just who is involved, but how and why they coordinate.’

The researchers suggest that future intervention strategies should move away from single, large-scale actions toward adaptive, ongoing approaches that monitor how networks evolve after disruption.

‘This research demonstrates that complex social problems cannot be solved with simple one-step interventions. Effective disruption must account for how systems reorganise in response to pressure,’ says co-author Dr Mike Lees. ‘This applies not only to criminal networks, but also to challenges such as misinformation networks, financial fraud structures and even ecological management.’