Author: John Mason, the founder of Risk Insights, a specialist advisory practice focused on systemic risk and complex systems. His work examines the structure of risk networks to make hidden vulnerabilities visible.
Most organisations are good at identifying risks. Frameworks exist, ownership is assigned, and analysis is conducted with discipline. But for all the individual analysis, a different set of questions remains unanswered:
- What depends on this risk?
- Where do effects from separate risks accumulate?
- How does the system behave when multiple stresses arrive simultaneously?
These are relational questions. They require knowledge of the connections between risks, and individual analysis – however rigorous – does not produce that insight.
What the structure reveals
Risk network modelling starts from a different space. Risks become nodes in the model. The relationships between them – directionality, dependencies, influences – form the connections that define the structure. Then you analyse what emerges. And the structural vulnerabilities reveal themselves:
- Transmission paths through which stress propagates.
- Convergence points where consequences from multiple sources accumulate on a single node.
- Amplification points where modest inputs generate disproportionate downstream effects.
- Feedback loops where cause and consequence become indistinguishable.
These are properties of the network itself. They do not exist in any individual risk. They exist in the connections between them and are made visible when the structure is modelled.
What structural position means for intervention
Not all risks occupy the same structural role, and the role determines what addressing a risk means in practice. Certain risks function as sources, driving effects outward across the system. Some function as transmission mechanisms, connecting separate domains. Others are convergence points, where consequences from multiple directions accumulate. And some are bridges – connections between clusters that determine whether a disruption stays contained or spreads across the network.
Each role invites a different response. Intervening at a transmission point can interrupt cascading effects before they propagate. Intervening at a convergence point addresses symptoms that will persist as long as upstream drivers continue to feed in. Targeting a source can generate benefits across multiple downstream risks simultaneously. These distinctions are not made explicit when risks are assessed individually, yet they exist as properties of position within the network.
Example
A combined authority oversees housing, transport, and economic development across a region. Each policy area monitors its own risks and reports them through separate channels.
Housing identifies a growing shortage of affordable homes in the city centre. Economic development reports persistently high unemployment in two outer districts. Transport highlights funding pressure on peripheral bus services. Each assessment is accurate within its own domain.
Model the structure, and those bus routes turn out to be network bridges that connect the housing cluster and labour cluster. A shortage of affordable housing in the city centre is pushing lower-income workers toward outlying towns. The bus routes are how they reach employment. Cut bus funding, and the housing crisis that pushed workers outward worsens employment opportunities – a risk that crosses departmental boundaries but would not register in their individual risk briefings.
Structure does not wait for consensus
This matters most in the environments where current approaches struggle most: complex, multi-stakeholder systems where objectives diverge. In environments where government bodies, international organisations, private sector actors, and communities operate with different and often conflicting priorities, the value of structural analysis is that it does not require agreement. Stress propagates through structural relationships regardless of intention or consensus. The network does not arbitrate between competing priorities. It maps the structural reality within which all actors operate – whoever ‘owns’ the risks, whatever the policy context, however contested the environment. The question therefore shifts from “do we agree on the risks?” to “what does the structure we all operate within actually do under stress?” The pattern of connections has properties that shape how the system behaves. Those properties exist. The critical question is whether they are found through analysis or discovered through impact.