A major actuarial lesson in Lloyd's 2026 geopolitical-risk material is that losses should not be isolated to political violence or political-risk policies. A single escalation can create connected exposures in marine, cyber, aviation, energy and trade. This is an accumulation problem, not a list of independent class stresses.
One event, several mechanisms
Restrictions on oil-carrying vessels can disrupt supply chains. Attacks on infrastructure can create cyber and energy losses. Flight re-routing can accumulate ground exposures at regional airports. Trade restrictions can affect credit, cargo, business interruption and contingent business interruption. The question for the model is how these mechanisms depend on one another.
Build a class-to-mechanism map
Start with a causal map: event, operational channel, exposure, policy response and loss timing. Then map the same exposure to every class that could respond. This identifies double counting, hidden exclusions and missing second-order losses. A class-level model remains useful, but it must feed an enterprise accumulation view.
Testing and management action
Scenario severity should be realistic and severe, with assumptions recorded before results are reviewed. Test policy wording, cancellation notices, outwards protection and correlated market effects. Management needs to know which limits, underwriting controls, liquidity resources and reinsurance actions remain available if several classes move together.
This is the type of cross-risk reasoning I use as CEO of Quantica Risk Modelling: the objective is not to forecast headlines, but to expose dependencies early enough for an insurer to act. Source: Lloyd's Capital Briefing 2026.