The Government Office for Science has published the technical annex behind its Global Supply Chains foresight work, setting out how firm level modelling was used to identify vulnerabilities that may be invisible in direct supplier data. Added on 24 September, Annex B explains how disruption can move through interconnected production networks and how systemic risk was assessed across firms and countries.
The main foresight report was published on 15 June, so the new development is the release of the modelling methodology rather than the wider study itself. The annex describes an agent based model in which individual firms are treated as separate actors with their own production characteristics, supplier relationships and positions within the wider network. It is designed to explore how shocks propagate, not to predict the precise outcome of a future disruption.
That distinction is important because conventional concentration measures generally capture direct exposure. A manufacturer may appear to have several first tier suppliers while those businesses depend on the same specialist processor further upstream. A disruption at that shared node can then affect several apparently separate supply routes at once. The modelling is intended to expose those indirect dependencies and the feedback effects created when shortages pass through multiple tiers.
The framework links firms through observed input and output relationships using transaction data from the Government’s Global Supply Chain Intelligence Programme. The underlying information combines international trade documentation and company data to map who supplies whom, where firms and facilities are located and how products move through the network. Transactions from 2020 to 2024 are used, with older links given less weight while remaining available to capture intermittent relationships.
To compare the consequences of disruption, the model calculates an Economic Systemic Risk Index. A global measure estimates the percentage of total output at risk across the network, while a UK measure focuses on output at risk among UK based firms. A company does not need to be especially large to score highly if it occupies a bottleneck, supplies an input that is difficult to replace or sits inside a dependency shared by many downstream businesses.
The production model assumes that essential inputs are required in fixed combinations, so an excess of one material cannot automatically compensate for the loss of another. Firms can switch to alternative suppliers where suitable capacity is represented elsewhere in the network, but the model deliberately limits some forms of substitution. That conservative treatment is intended to avoid understating vulnerability when a severe disruption removes an important source of supply.
The annex is equally explicit about the model’s limits. Inventories and buffer stocks are not represented directly, which means a business holding several weeks of material may appear to feel a disruption earlier than it would in practice. Transport delays, production lead times and detailed physical logistics are also outside the framework, while ports, shipping routes and transport capacity are not modelled as a complete physical network.
Those constraints mean the results are best read as structured stress tests rather than forecasts of trade flows, production losses or company behaviour. The Government Office for Science also states that the wider report is not a statement of government policy. The modelling is intended to compare relative vulnerability and resilience across firms, countries and scenarios, not to predict that a named company, region or supplier will fail.
The methodology adds detail to a problem already visible in industrial purchasing. Recent evidence from UK manufacturers shows businesses spending more to protect production continuity, including through alternative sourcing and higher inventory. Those measures can reduce exposure to some disruptions, but they are difficult to target when critical dependencies sit several tiers beyond a manufacturer’s direct commercial relationships.
That problem becomes harder as production combines specialist materials, electronics, software enabled components and internationally distributed processing steps. The main foresight report notes that around half of UK production depends on intermediate inputs produced abroad, while climate pressures, geopolitical fragmentation and transport disruption can affect common nodes used by several sectors. A supplier that appears marginal in direct spend can therefore carry much greater operational importance than its invoice value suggests.
The new annex gives analysts a clearer view of how those dependencies were translated into a model and where the conclusions should be treated cautiously. It does not turn supply chain resilience into a single score that procurement teams can apply mechanically. Instead, it shows why firm size, import share and direct supplier concentration are incomplete measures when production networks contain shared upstream dependencies.
The practical challenge remains visibility. A first tier supplier list may be well understood while the processors, material producers and logistics dependencies behind it remain opaque. The modelling published on 24 September provides a way to stress test that hidden structure, but the resulting risk signals still need to be combined with inventory data, supplier knowledge and operational judgement before they become sourcing or investment decisions.
