We should pause when the profession whose job is to model risk for a living tells us that its own tools are unable to assess the biggest risks ahead. Last month, the Institute and Faculty of Actuaries (IFoA) published a report on nature risk, stating that the standard quantitative models they rely on to price risk and advise decision-makers cannot adequately capture the dynamics driving systemic disruption.
Food systems, alongside pandemics, are among the report’s two near-term areas examined in detail. The food section covers the chronic pressures already underway, such as concentration of grain production in a small number of countries, breaches of the planetary boundaries that agriculture depends on, and rising food insecurity in countries like the UK, alongside the acute risks of shocks from extreme weather, geopolitical conflict, pathogens, and cyber attacks. The report’s central argument is that these chronic and acute risks interact in ways that present a systemic threat to food security, and that these interactions are precisely what standard quantitative models cannot account for.
In this article, I dig into what this means for how we assess food system risk in the UK and what we can do about it.
The models can’t account for many food system risks
The IFoA report joins a chorus of recent warnings about UK food system fragility I’ve written about before: Tim Lang’s Just in Case report on a lack of contingency planning, the anonymous warning from senior food industry executives that the sector’s resilience is more assumed than tested, Defra’s national security assessment on ecosystem collapse and related risks to food supply, and the National Emergency Briefing’s direct warning to a Westminster audience.
The actuaries’ contribution to this conversation adds a methodological concern, reporting that the professional tools used to assess the scale and severity of these risks are not built for the job. Those tools work by looking at what has happened before and using that data to predict what is likely to come next. They assess the kinds of changes that stay within predictable boundaries, such as prices drifting up or down and gradual shifts in demand.
But food system disruption tends not to behave this way. Familiar conditions can hold true for years before tipping into something else entirely, and can take more than one form.
Some shifts are abrupt. For example, the M&S cyberattack last year showed how a single breach can take payment, ordering, and stock systems offline overnight, leaving empty shelves and disrupted supplies for weeks afterward. A heatwave that damages harvests across several major grain-exporting countries in the same season can push UK prices beyond what many households can afford.
Other shifts build slowly and then cross a threshold that marks an abrupt acceleration in change. The report identifies pollinator collapse as one such case: bees and other insects underpin around three-quarters of global crop production, yet in the UK, they have lost on average a quarter of their habitat since 1980. A model built on historical data might assess that with an expectation of continual, gradual decline. Yet it is possible that at one point, pollination services will fail entirely, with a staggering effect on crop yields. In this case, what matters is when the threshold is crossed, rather than relatively smaller changes in any single year.
Shocks rarely arrive in isolation, either. For example, a geopolitical crisis can create openings for a cyber attack while our attention is diverted, and bad weather here in the UK can coincide with weather-related problems in the countries we import from. Each event also makes the next harder to absorb, and enough pressure of this kind can push the system past a tipping point – a threshold that once crossed offers no straightforward way back. We can see then that the limit lies in the model itself; more data or fine-tuning would not make it easier to identify tipping points or compound risks.
Exposing this limitation in risk assessment changes the shape of the conversation around food system risks and food security. Plenty of people have been worried about the UK food supply for many years; now the profession whose models would tell decision-makers how much to worry is reporting that its models cannot tell them. When regulators and standard-setters ask for quantitative proof before acting on systemic risk, and the risks in question cannot be quantified in the way they are asking, the requirement itself becomes a barrier to action. We need a change in our approach.
Different methods are needed to assess the interactions that matter
There is a source for the information that’s needed: assessing UK food system risk in a way that captures how the system actually behaves can be achieved by drawing on the knowledge of the people who work inside it. Farmers, retailers, civil servants, food bank coordinators, climate scientists, and supply chain analysts, among many others, collectively hold a working understanding of where the pressures lie, how a shock would propagate, and which interventions could mitigate the risks.
Mapping food system risk from inside the system
Structured expert elicitation is a way to capture that combined understanding in a form that supports decision-making. For example, a 2024 study drew on 58 experts from academia, industry, government, and civil society to apply this approach to the UK food system. They found that over 40% of those consulted judged that civil unrest arising from a food crisis was at least possible within the next 10 years. Standard quantitative models have not generated these kinds of results, which the IFoA report argues is central evidence of the systemic threat it identifies.
The 2023 study mapped 17 risks to the UK food system, spanning environmental hazards such as extreme weather and ecosystem collapse; geopolitical and economic shocks such as trade restrictions and financial crashes; biological threats including pathogens and pandemics; and infrastructure failures ranging from transport strikes to cyberattacks. It then traced how they might combine to drive the country into crisis.
Rather than generating a probability, it identified the chronic conditions within the system, triggers that might provoke them, and combinations the experts judged most consequential—such as a heavy reliance on imported staple crops that leaves the system exposed if, for example, an extreme weather event abroad reduces supply, compounded by protectionist trade restrictions, together producing a significant shortage. Extreme weather, ecological collapse, and trade restrictions emerged as the top-ranked drivers across both 10- and 50-year horizons.
Turning risk maps into priorities for action
A 2026 extension of that study, which I was part of, translated that map of risks into a set of pathways and used it to rank interventions. At the system-wide level, the highest-ranked changes were longer-termism in policy planning, food system thinking across government, stronger international collaboration, and a forum for diverse voices in food policy. At the pathway-specific level, diversification was considered important across production, supply chains, distribution, and diets, alongside agroecology and regenerative farming for weather resilience, and infrastructure investment to mitigate cyber disruption.
The ranking gives policymakers a basis for action that quantitative methods struggle to produce. As I highlighted in a previous article, most of the disruption pathways described in the earlier study have already been activated in the three years since the research was conducted—much faster even than the decadal horizons it considered. Given the study’s assessment of disruption being uncannily accurate, the intervention pathways are worth taking seriously.
None of this displaces quantitative analysis. Rather, the two are complementary, with structured elicitation mapping the connections between factors that models have to treat in isolation.
What this means for UK food security decisions
Decision-makers addressing UK food security concerns must combine practical methods with traditional risk analysis to produce actionable results. Three categories of these methods stand out:
- Narrative scenarios let policymakers stress-test proposed interventions against extreme-but-plausible disruptions. If a planned change in farm support, supply chain regulation, or food reserves doesn’t hold up under a severe weather event coinciding with sustained trade restrictions, that possibility is worth testing in advance. This generates a robustness check of interventions that hold up across many possible futures—and these are the ones to support.
- Ranked interventions identify what actions could be most impactful. The ranking rests on a structured argument for why these priorities outrank other candidates—a form of prioritization that supports decision-makers and that quantitative methods rarely produce.
- Serious games like BAFR-UK’s Barefood take the same outputs into shared rooms, where people from across the food system work through how cascading and compounding shocks would play out across their decisions. The result is the kind of cross-sector preparedness that traditional planning exercises rarely produce.
There is also a regulatory implication of the IFoA’s findings. Requiring proof in a form that the underlying dynamics cannot generate is itself a policy choice, with consequences for whether systemic food risk gets acted on. Regulators and standard-setters who recognize this are positioned to change what counts as actionable evidence.
Closing the gap: From methods to decisions
For several years, the chorus of warnings about UK food system fragility has been growing. The IFoA’s contribution is to identify, from within the risk profession, that its tools are not built to answer the questions being asked of them. That admission changes the shape of the conversation, from whether to heed the warnings to whether the country has the means to assess what it now faces. The means exist, and they have been applied to the UK food system. The disruption pathways have been mapped, and interventions to address them have been identified. What is now needed is for those who set UK food security policy to draw on what has been developed—and to ask for more.
I write about the future of food and the connections between our food systems, the environment and public health. Sign up for my newsletter.
You can also read more here.
