H Heuristics · Digital Report № 2026-23 · October 2026
A framework for navigating cascading risks — an interactive dashboard
A cascade loses height and gains length as it travels. The 2021–23 shock rose 1,358 per cent in European gas, 69 per cent in world food prices, and then peaked in household food inflation nine months after world prices did. Resilience is a matter of acting at the right stage, on the right clock.
Cascading risk is the defining mechanism of a polycrisis: a shock in one system travels along the links between systems and arrives somewhere else, later and in a different form. This dashboard measures one such cascade directly rather than describing it. Using monthly World Bank commodity prices, the FAO Food Price Index and FAOSTAT food consumer price indices for 183 countries, it follows the 2021–23 shock along a single chain — European natural gas, fertilizer, grains, world food prices and domestic food inflation — and organises the evidence into a framework of four stages (trigger, transmission, amplification and impact), each matched to a resilience function: diversify, decouple, buffer and absorb.
The chain shows a regular shape. Measured from the 2019 average to the 2021–23 peak, European gas rose 1,358 per cent, the World Bank fertilizer index 261 per cent, the grains index 90 per cent and the FAO Food Price Index 69 per cent: amplitude fell at every link. Duration moved the other way. Gas stayed above half its peak rise for 4 months, fertilizer for 14, grains for 20 and world food prices for 21. The domestic leg came last and lasted longest: world food prices peaked in March 2022, but the median country's food inflation peaked in February 2023 at 13.3 per cent, the median country peaked nine months after the world price, and 87 per cent of countries peaked later than the world price. In December 2022, 69 per cent of countries had food inflation above 10 per cent, against about 12 per cent in 2019.
How hard the cascade hit a country depended less on how much food it imported than on what amplified the shock at home. Across 175 countries, the depreciation of the currency against the dollar between 2021 and 2023 had a Spearman rank correlation of 0.42 with the country's peak food inflation, against 0.13 for the share of food in merchandise imports; reserves cover (−0.28) and inflation before the shock (0.36) also mattered. Countries whose currency fell by more than a quarter saw a median peak of 53 per cent, against 13 per cent where it moved by less than 5 per cent. These are associations, not causal estimates, but they locate the amplifier in the financial channel rather than in trade exposure alone.
The same chain is now carrying a second shock. After the war that began on 28 February 2026 disrupted shipping through the Strait of Hormuz, urea rose 82 per cent in two months and the fertilizer index 44 per cent; both had returned to their February levels by July, while grains were 20 per cent higher in September and still rising and world food prices 8 per cent higher. Domestic food price data do not yet cover the period. On the 2021–23 clock, the domestic leg is still ahead, which means the levers that work at that stage — reserves, debt space, credible monetary policy and social protection — are the ones whose window is open now. The dashboard is descriptive: its timings are measured from one episode and are a guide to sequencing, not a forecast.
A polycrisis does its damage through cascades: shocks that travel from one system into another and arrive later, in a different form. This dashboard follows one cascade from end to end and asks where resilience can intervene.
Most discussion of systemic risk stops at the observation that everything is connected. That is true and not very useful. What a finance ministry, a central bank or a development agency needs to know is narrower: when a shock starts in one system, which way will it travel, how fast, how much of its force will survive each step, and what at home will make it worse? Those are measurable questions, and the shock that ran from European gas markets to household food budgets in 2021–23 is the best-documented answer the decade has produced.
This dashboard follows that shock along one chain, month by month, using public data: the European natural gas price, the World Bank's fertilizer and grains price indices, the FAO Food Price Index and food consumer prices in 183 countries. It then tests which national conditions amplified the shock, follows the 2026 shock from the Strait of Hormuz along the same chain, and turns what the data show into a framework of four stages and four resilience functions. Every figure is computed from the sources listed at the end, and the amplifier explorer lets readers test the central association for themselves.
The argument in brief
A cascade loses height and gains length as it travels: every link in the chain damped the price shock and stretched it out, so that by the time it reached households it was smaller in world-market terms but slower, longer-lasting and more unequal. How hard it landed was set less by exposure than by amplifiers at home, above all a falling currency. Resilience is therefore a question of sequencing: each stage has its own function and its own window, and the 2026 shock now sits between the global stages and the domestic one.
The distinction matters because the usual vocabulary of resilience blurs it. Buffers, redundancy, diversification and safety nets are all good things, but they work at different points in a cascade and on different clocks. Diversifying an energy supply takes years and has to be done before a shock; keeping trade open matters in the first weeks; reserves and fiscal space matter in the months when the shock reaches the exchange rate; and social protection matters most when it reaches the household budget, which in 2022 was close to a year after world prices peaked.
The sections that follow set out the framework (section 1), measure the shape of the cascade (2), follow it into domestic food prices (3), test what amplified it (4), place the 2026 shock on the same clock (5) and translate the evidence into four resilience functions with their levers and windows (6). The method section states the choices that could change the results, and the sources section lists every dataset used.
Four stages a cascade passes through, and the resilience function that works at each. The rest of the dashboard measures every stage with data.
The framework begins from a simple observation about how cascades work. A shock starts somewhere specific — a pipeline, a strait, a harvest, a balance sheet. It then travels along links that exist for good reasons in normal times: natural gas is the main feedstock for nitrogen fertilizer, fertilizer is a large cost in growing grain, and grain feeds both people and livestock. At the border it meets national conditions — the exchange rate, the debt stock, the credibility of the central bank — that can dampen it or magnify it. And it ends in a household budget, where its impact depends on how much of income goes on food and what protection exists when that share rises.
Each of those four stages — trigger, transmission, amplification and impact — has its own resilience function. At the trigger, the task is to diversify, so that no single source can deliver a large shock. Along the transmission chain, the task is to decouple: to stop links from tightening by keeping trade open, holding stocks and using inputs more efficiently. At the amplification stage, the task is to buffer, with reserves, fiscal space and monetary credibility that keep a world price shock from becoming a currency crisis. At the point of impact, the task is to absorb, through social protection that reaches households before they cut meals or sell assets.
The framework is deliberately narrow. It does not try to list every risk or every connection; it asks of any cascade where it is now and which function can still act. That makes it usable in real time, which is the test that matters in a polycrisis, where several cascades can run at once. It also makes clear that resilience is not a single stock that a country has or lacks. A country can be well diversified and poorly buffered, or well buffered and unable to reach the households who need help, and the 2021–23 data show that such gaps decided outcomes.
The framework at a glance
| Stage | What happens | 2021–23 measurement | Function | Indicators to watch |
|---|---|---|---|---|
| 1Trigger | A supply shock in one system: energy, a trade route, a harvest, a financial market. | European gas peaked at 14.6 times its 2019 average (August 2022); Brent rose 25% in the four months after February 2022. | Diversify | Import concentration by source and route; storage relative to the largest plausible disruption. |
| 2Transmission | The shock passes along input–output links, damped and delayed at each step. | Fertilizer index +261%, grains +90%, FAO Food Price Index +69% from 2019 averages; durations of 14, 20 and 21 months. | Decouple | Share of traded calories under export restrictions; stocks-to-use ratios; fertilizer prices relative to grain prices. |
| 3Amplification | National conditions magnify or damp the world price shock as it crosses the border. | Peak food inflation correlates with depreciation (ρ = 0.42), prior inflation (0.36) and reserves cover (−0.28), barely with food-import share (0.13). | Buffer | Exchange rate; reserves in months of imports; debt-distress rating; inflation expectations. |
| 4Impact | Households absorb what is left, through spending, savings, assets and meals. | Median food inflation peaked at 13.3% in February 2023; 69% of countries were above 10% in December 2022. | Absorb | Food CPI; social protection coverage and capacity to scale; food share of household spending. |
Two features of the table carry the argument of the dashboard. First, the measurements in its third column are not on one scale, and they should not be: a cascade changes units as it travels, from dollars per million British thermal units to dollars per tonne, to an index, to a household inflation rate. The question at each stage is therefore not how big the shock is in the abstract, but how much of it survived the previous link, how long it lasted and what changed its size. Sections 2 to 4 answer those questions in turn.
Second, the indicators in the last column are public, and nearly all are monthly or quarterly. A ministry does not need a bespoke model to know where a cascade is: the World Bank's commodity series, FAO's food price data and national consumer price indices show it with a lag of weeks. The binding constraint on resilience is rarely a lack of warning. It is that the functions which act late in a cascade — buffers and social protection — have to be built early, before the warning arrives.
The global stages of one chain, each indexed to its 2019 average. Every link lowered the peak and stretched the episode.
The chart plots the global stages of the chain on a common scale: each series is divided by its 2019 average, so 100 means "as in 2019" and 200 means "twice as high". The vertical scale is logarithmic, so equal distances are equal proportional changes, which lets a gas price that rose more than fourteenfold share a chart with a food index that rose by about two-thirds.
Click a series in the legend to hide or show it. The two shaded bands mark the months after the two triggers this dashboard follows: the invasion of Ukraine in February 2022 and the war that began at the end of February 2026. The two charts beneath the main panel summarise the 2021–23 episode for each stage: how far it rose, and for how long it stayed high.
Four stages of the chain, monthly, January 2019 – September 2026
Natural gas is the European price in US$/mmbtu; fertilizer and grains are World Bank price indices (2010 = 100); world food is the FAO Food Price Index (2014–16 = 100). All are rebased here to their own 2019 averages. Shaded: the twelve months from each trigger.
Height: peak rise from the 2019 average
Length: months above half the peak rise
The first thing the clock shows is attenuation. Between its 2019 average and its August 2022 peak, the European gas price rose by 1,358 per cent. The fertilizer index, for which gas is the main input into nitrogen products, rose by 261 per cent to its April 2022 peak. Grains rose by 90 per cent and the FAO Food Price Index by 69 per cent. Each step absorbed part of the shock: fertilizer is only one cost of growing grain, grain is only one ingredient of the food basket, and at every step substitution, stocks and contracts took some of the strain.
The second is that duration moves the other way. The gas price stayed above half of its peak rise for only four months, from July to October 2022. The fertilizer index stayed there for 14 months, grains for 20 and world food prices for 21. A shock that is sharp and short at the source becomes milder but more persistent downstream, because the slower parts of the system — planting decisions, contract renewals, retail pricing — each pass it on with their own delay. Persistence is what households and public budgets feel.
The third is that the chain had more than one starting point. World food and grain prices began rising in mid-2021, before the gas spike, as demand recovered from the pandemic, energy costs climbed and several harvests fell short; the invasion of Ukraine in February 2022 then hit grain exports directly at the same time as it drove gas higher. That is typical of a polycrisis: cascades from different sources merge on the same links. For the framework, it means the trigger stage is often plural, and the transmission stage is where separate shocks become one.
World food prices peaked in March 2022. Food inflation in most countries peaked months later, and stayed high for longer.
The domestic leg is where a cascade becomes a livelihood problem, and it is the leg most often missed by analyses that stop at world prices. To measure it, the dashboard computes year-on-year food inflation for every country with official monthly consumer price data in FAOSTAT and follows the distribution across countries month by month.
Two kinds of measure are shown: the share of countries with food inflation above 5 per cent and above 10 per cent, and the median country's rate. The dashed vertical line marks March 2022, the month the FAO Food Price Index peaked. Country coverage is close to 190 a month through 2024 and thins during 2025, so the series end in December 2025.
Share of countries with high food inflation, monthly, 2019–2025
Food inflation in the median country
When countries peaked
The share of countries with food inflation above 10 per cent averaged about 12 per cent in 2019. It passed 45 per cent in April 2022 and peaked at 69 per cent in December 2022, nine months after world prices did; at that point 89 per cent of countries had food inflation above 5 per cent. The median country's rate peaked in February 2023 at 13.3 per cent, more than five times its 2019 average of 2.4 per cent, and stayed above half of that peak for 18 months, from March 2022 to August 2023.
Peak timing varied, but most countries came late. Of 183 countries with enough data, 87 per cent reached their peak food inflation after world prices peaked, 70 per cent six months or more later, and about 30 per cent a year or more later; the median lag was nine months. A late peak is what the framework predicts: domestic prices respond to world prices through import contracts, transport and processing margins, wages and the exchange rate, each of which adjusts with its own delay.
By December 2025 the distribution had returned close to its pre-shock shape: 27 per cent of reporting countries had food inflation above 5 per cent and 8 per cent above 10 per cent, with a median of 3.0 per cent. In the aggregate, the episode is over. The price level, however, does not come back down: food inflation falling to 3 per cent means prices rising more slowly from a much higher base, which is why households in many countries still experience food as expensive long after the inflation statistics have normalised.
Why the same world price shock produced 13 per cent food inflation in one country and 50 per cent in another.
If a cascade were only a matter of exposure, the countries that import the most food would have suffered most. The data say otherwise. The explorer plots each country's peak food inflation in 2022–23 against four candidate amplifiers measured before or during the shock: the change in its exchange rate against the dollar from 2021 to 2023, the share of food in its merchandise imports in 2021, its reserves in months of imports in 2021, and its food inflation in 2019.
Choose an amplifier to redraw the chart; the rank correlation and the number of countries update with it. Peak inflation is on a logarithmic scale because a handful of countries — Lebanon, Venezuela, Zimbabwe and Argentina among them — reached rates of 250 to 490 per cent. Search for a country, or click a point, to highlight it and see its values; income groups can be hidden from the legend.
Spearman ρ —Countries —
Median peak food inflation, by currency change
How strongly each amplifier tracks the peak
The currency is the strongest of the four. Across 175 countries, depreciation against the dollar between 2021 and 2023 has a rank correlation of 0.42 with peak food inflation. Where the currency moved by less than 5 per cent, the median peak was 13.3 per cent (65 countries); where it fell by 5 to 25 per cent, 17.0 per cent (91); and where it fell by more than 25 per cent, 53.1 per cent (19). The mechanism is direct: world food is priced in dollars, so a falling currency raises the local price of every imported calorie — and of fuel and fertilizer — at the moment world prices are already high.
Dependence on food imports, by contrast, is barely associated with the peak (ρ = 0.13). Exposure determines whether a cascade arrives; it does not determine how hard it lands. The other two amplifiers sit in between. Countries with more reserves cover in 2021 tended to have lower peaks (ρ = −0.28), consistent with reserves letting a central bank slow a depreciation; countries with higher food inflation before the shock tended to have higher peaks (ρ = 0.36), consistent with weaker monetary anchors letting a relative-price shock feed into expectations.
These are associations across countries, not causal effects. Currency weakness and high inflation are partly the same macroeconomic story, and several of the highest peaks were in countries already in crisis before 2022. But the pattern is consistent and its policy reading is clear: the amplification stage is mostly financial and institutional, so the buffers that matter most for a food cascade are not food buffers alone. Income groups tell the same story: median peak food inflation was 14.4 per cent in high-income countries and 23.8 per cent in low-income ones.
The Strait of Hormuz shock is travelling down the same chain. Aligned on the month before each trigger, the two episodes can be read on one clock.
The war that began on 28 February 2026 disrupted shipping through the Strait of Hormuz, which before the war carried about 35 per cent of seaborne crude oil trade and, according to the World Bank, nearly a quarter of global urea exports. The chart aligns the 2026 shock with the 2022 one: each series is indexed to its value in February of the year in question, so the lines show what happened after each trigger, month by month. Choose a stage to compare the two.
The four tiles beneath the chart summarise where each stage stood in September 2026, seven months after the trigger and the most recent month in the World Bank and FAO data, with the 2022 reading at the same point for comparison. The domestic stage cannot be shown yet: official food price data for most countries do not extend beyond the first months of 2026.
Urea: 2022 and 2026 episodes, months after the trigger
+64%
Brent crude, September against February 2026; −6% at the same point in 2022
+82%
Urea at its April 2026 peak; by September it was 14% below its February level
+20%
Grains index, September against February 2026 and still rising; +7% at the same point in 2022
+8%
FAO Food Price Index, September against February 2026; −3% at the same point in 2022
The 2026 shock started higher up the chain than the 2022 one and has been more persistent there. Brent crude in September 2026 was 64 per cent above its February level, after a fall in the early summer and a renewed rise; seven months after the 2022 invasion it was 6 per cent below. Fertilizer behaved as it did in 2022, only faster and higher: urea rose 82 per cent in two months and was back below its pre-war level by July, and the fertilizer index close to it. The fertilizer link carried the shock quickly, but briefly.
Grains are where the two episodes now differ most. Seven months after the 2022 trigger the grains index was 7 per cent above its February level and falling; in September 2026 it was 20 per cent above and still rising, with US hard red winter wheat up 35 per cent. World food prices have followed more slowly, up 8 per cent. On the 2021–23 evidence, a rise in world grain and food prices passes into domestic food inflation with a median delay of about nine months, and is magnified where currencies weaken.
That places the 2026 cascade between its global and domestic stages. If it follows the 2022 timing, the domestic leg would not peak before 2027. The amplifiers are already visible: as of 31 March 2026, 9 of the 68 low-income countries covered by the IMF–World Bank debt sustainability framework were in debt distress and 23 more at high risk. This is not a forecast — one episode cannot supply one — but a reading of which resilience functions still have time to act, the subject of the final section.
How the cascade was measured, and the choices that could change the results.
Commodity prices are monthly nominal US-dollar series from the World Bank's Pink Sheet (updated 2 October 2026): the European natural gas price, urea, Brent crude and US hard red winter wheat, and the fertilizer and grains price indices. World food prices are the FAO Food Price Index as published in October 2026; FAO revises recent months. Domestic food inflation is the author's calculation of the year-on-year change in each country's FAOSTAT food consumer price index, using only observations flagged as official or as supplied by an external organisation, and excluding imputed values.
Amplitude is the rise from a series' 2019 average to its 2021–23 peak; duration is the longest run of consecutive months, containing the peak, in which a series stayed above half its peak rise. Peak timing is measured within 2021–24 for the 183 countries with at least 40 monthly observations, and peak magnitude within 2022–23 for 181. Amplifiers come from the World Development Indicators: the official exchange rate (2021 to 2023 change; Croatia excluded because euro adoption in 2023 redenominated its series), food imports as a share of merchandise imports (2021) and reserves in months of imports (2021). FAOSTAT and World Bank country codes are joined through the UN M49 table.
Three limits matter. The timings come from one episode, and the next cascade will have its own speed; the clock is a guide to sequencing, not a forecast. The cross-country results are rank correlations, computed in the browser with average ranks for ties, and they describe association, not cause. And price indices in US dollars are not welfare measures: the household impact depends on the food share of spending and on incomes, which this dashboard does not observe. Related analysis: the polycrisis century measures how often systems are stressed together, and food, energy, water examines the nexus this chain runs through.
Data series were downloaded from the publishers below in October 2026.