Tuesday 16 December 2025 was the day the World Bank Tokyo office hosted Online Morning Seminar #212, Fragile and Conflict Affected Situations: Intertwined Crises, Multiple Vulnerabilities, with Jeetendra Khadan and Peter Selcuk, two of the report’s three authors (seminar page). The invitation opens with a hard claim: conflict and instability are driving extreme poverty up faster in 39 economies than anywhere else (press release). This article is a briefing built around that claim. It takes the World Bank’s June 2025 assessment, reads it against everything else that landed between January and November 2025 on African hunger, aid, input prices, climate finance and farm level evidence, adds the 2026 follow up data, and asks what it all means for a farmer cooperative in Kenya.
It follows our earlier pieces, including who pays when the rain fails. It connects to AgriKonnekt, which provides Digital MRV and technical assistance that connect smallholder cooperatives to climate finance, and to Kilimora’s Gender Lens work.
The argument in one paragraph. During 2025 three things happened at once. Fragility deepened: the World Bank counted 421 million people on less than $3 a day in conflict affected economies, more than in the rest of the developing world combined. Emergency money contracted: the World Food Programme (WFP) expected 40% less funding than in 2024, and refugee rations in Kenya fell to 28% of a full ration. And Africa wrote itself a new ten year farm plan in Kampala while the evidence on what actually moves smallholder outcomes (risk sharing, physical delivery, early cash, readable data) kept accumulating. The gap between the first two and the third is where cooperatives and their financiers now work.
World Bank, 27 June 2025; Global Report on Food Crises 2026 as reported by AFP and Action Against Hunger
2025 at a glance
The table sets the World Bank report in its year. Each row is a dated publication or decision, and the last column says why it matters for farm finance.
| Date | What happened | Why it matters |
|---|---|---|
| 9 to 11 Jan | African Union summit in Kampala adopts the Kampala CAADP Declaration and a ten year strategy (2026 to 2035). Targets include 45% more agrifood output by 2035 and $100 billion of investment. AU, ILRI | Sets the continental yardstick a cooperative can report against. |
| 20 Jan | US executive order “Reevaluating and realigning United States foreign aid” halts USAID funding, a key partner in Kenya’s refugee camps. The Standard | Start of the emergency aid squeeze. |
| 28 Feb | Kenya’s IPC short rains assessment: about 2.2 million people in Crisis or worse, 266,000 of them in Emergency. IFRC/ReliefWeb | Drought after failed rains from late 2024. |
| 7 Mar | WFP warns that food and nutrition aid for 2 million people in the Central Sahel and Nigeria will be suspended in April without funding. WAM | Aid cuts reach West Africa within weeks. |
| 10 to 12 Apr | CILSS regional meeting in Lomé: 34.5 million people in Crisis or worse in the Sahel, West Africa and Cameroon, rising to about 49.4 million in June, August. CILSS | Lean season peak. |
| 16 May | Global Report on Food Crises 2025: 295.3 million people in 53 countries in acute food insecurity in 2024, the sixth annual rise. FSIN | Baseline for the year. |
| 20 May | WFP tells Kenya it will cut refugee rations to 28% of a full ration from June and end cash assistance. Capital FM | Lowest assistance ever given to refugees in Kenya. |
| 27 Jun | World Bank publishes Fragile and Conflict Affected Situations (Global Economic Prospects, chapter 4). World Bank | The report this briefing reads. |
| 28 Jul | SOFI 2025: global hunger down to 8.2% of people, but 307 million Africans hungry. UN joint release | Africa diverges from the world. |
| Jul, Aug | Kenya’s long rains assessment: 1.8 million in Crisis or worse (July, September), 2.1 million projected to January 2026. NDMA | Good rains, thin recovery. |
| 31 Aug, 5 Sep | Africa Food Systems Forum in Dakar; Africa Food Systems Report 2025 launched. AGRA | Finance and the “hidden middle” take centre stage. |
| 7 Oct | GEMs consortium of 29 development banks releases default and recovery statistics for 1994 to 2024. GEMs release | Risk data that was missing finally public. |
| Oct | Global Hunger Index 2025: progress stalled since 2016. Alliance2015 | 2030 target slipping. |
| 29 Oct | World Bank Commodity Markets Outlook: fertiliser prices up about 21% in 2025 while food prices fall. World Bank blog | Farmers squeezed from both sides. |
| Late Oct | UNEP Adaptation Gap Report 2025: developing countries need $310 to 365 billion a year by 2035; flows were $26 billion in 2023. UNEP | Adaptation money is the scarce kind. |
| 3 to 4 Nov | IPC confirms famine in El Fasher and Kadugli, Sudan. IPC | Second famine confirmation in Sudan in under a year. |
| 11 Nov | FAO, WFP Hunger Hotspots: 16 hotspots to May 2026; Kenya listed. SDG2 Advocacy Hub | Early warning for the next season. |
| 18 Nov | WFP 2026 Global Outlook: 318 million in Crisis or worse, funding for about 110 million. Al Jazeera | Triage becomes policy. |
| Nov | COP30 opens with a $1.4 billion Gates Foundation commitment to smallholder adaptation. AgTech Navigator | Philanthropy steps toward the gap. |
| 16 Dec | World Bank Tokyo seminar #212 on the FCS report. World Bank Tokyo | The occasion for this briefing. |
Ten findings
- Fragility is concentrated in Africa. More than 70% of the people living through conflict and instability are African (World Bank chief economist Indermit Gill, quoted by Down To Earth), and half of the 39 economies have held fragile status for 15 years or more (chapter 4).
- Growth has reversed. Since 2020, per capita GDP in these economies has shrunk by 1.8% a year while it grew 2.9% elsewhere in the developing world (AP via Ahram).
- The farm is the labour market. Agriculture is just over 20% of output and about 40% of employment in these economies, roughly twice the shares in other developing economies (chapter 4).
- Hunger follows conflict. About 18% of people in these economies, around 200 million, faced acute food insecurity in 2024, against 1% elsewhere. Conflict was the main driver of acute hunger for 147.4 million people in 19 countries in 2025 (2026 GRFC via Action Against Hunger).
- Africa is moving the wrong way on chronic hunger. Global hunger fell to 8.2% in 2024, but in Africa 307 million people, over 20% of the population, were hungry (SOFI 2025).
- The aid cushion is thinner. WFP expected $6.4 billion in 2025 against $10 billion in 2024, and humanitarian food aid in crisis countries fell 59% between 2022 and 2025 (Invezz on WFP; Action Against Hunger via allAfrica).
- Inputs got dearer while crops got cheaper. Fertiliser prices were projected up about 21% in 2025 while food prices were projected down 6.1% (World Bank; Techjuice on CMO).
- Adaptation money is not reaching farms. Climate finance to small scale agrifood systems was 0.8% of the total, while developing countries need $310 to 365 billion a year for adaptation by 2035 against $26 billion flowing in 2023 (Climate Policy Initiative via Green Queen; UNEP).
- Capital follows readable data. Development banks’ own default statistics, published in October 2025, show infrastructure lending to private borrowers defaulting at 3.6%, below the 4.4% for non financial firms (PPIAF, GI Hub).
- Delivery design decides results. Insurance lifted adoption of a new product by 50% among Kenyan retailers, in person training was completed by 71% against 22% for an app in Ethiopia, and early cash improved food security after floods (Killeen; Cassidy, Ebrahim and Ubfal; Harris School).
The pressure: fragility is already a farm problem
The money tells the story before any adjective does. The World Bank counts 39 economies as facing conflict or instability, 21 of them in active conflict, home to more than one billion people. Sub Saharan Africa accounts for about half of the economies and 70% of their population. Extreme poverty runs at nearly 40%, against 6% in other developing economies, and the stall is old: the rate fell by close to 20 percentage points in the two decades to the mid-2010s, then stopped (chapter 4). The income gap is stark too. In 2000 average GDP per capita in these economies was a little under half that of other developing economies; by 2024 it was less than a quarter.
World Bank, Fragile and Conflict Affected Situations, 27 June 2025. Hunger shares are for 2024 (about 18% against about 1%).
Why this lands on farms. On average agriculture accounts for just over 20% of output and 40% of employment in fragile economies, about twice the shares elsewhere, so a shock to farming is a shock to the labour market. The report also puts a number on it: a 1% rise in conflict related fatalities per million people cuts per capita GDP by about 3.7% after five years in these economies, with a 7% cumulative loss in industry value added and a 2% loss in agricultural value added. Conflict also raises food insecurity by disrupting local production, imports, transport and the markets that connect them (chapter 4).
Share of output and of employment, 2022, %. World Bank chapter 4: fragile economies’ shares are “about twice” those of other developing economies; the comparator bars are derived as half.
Jobs sit underneath these numbers. Between 2000 and 2022 the working age population of fragile economies grew by 270 million, but employment grew by only 140 million, so the share of working age people in work slid from just under 60% to under 55%. Roughly 90% of the world’s refugees and displaced people come from these economies. Private credit is about one third of the level in other developing economies, and it has barely moved since 2010. Where there is no labour market and no credit, a cooperative is often the only institution a farm household can lean on.
Millions, fragile and conflict affected economies, 2000 to 2022. World Bank, Global Economic Prospects June 2025, chapter 4.
People on under $3 a day in fragile and conflict affected economies, millions. World Bank, 2025; 2030 is a projection. These economies are projected to hold nearly 60% of the world’s extreme poor by 2030, up from about 50% in 2024.
Debt tightens the squeeze. About 70% of these economies are in or at high risk of debt distress, up from around 40% a decade ago, and just over three quarters of their external public debt is owed to official lenders. Output is almost 13% below pre pandemic projections, about three times the shortfall elsewhere, and even if growth returned to its 2010 to 19 pace, output in 2030 would still be about 9% below trend (chapter 4). The authors add a warning that matters for aid policy: if external assistance becomes harder to access, financing pressures in these economies are likely to worsen.
Share of fragile and conflict affected economies in debt distress or at high risk of it, %. World Bank chapter 4: “around 40% a decade ago”, “about 70%” as of mid-2025.
The report is not only grim. Working age people are expected to reach about 60% of the population by 2040 and to exceed the share in other developing economies by about 2055. Some resource rich economies stand to gain from critical minerals demand. But the dividend is conditional, and the conditions are ones finance can influence: economies with stronger governance, better human development, deeper financial markets and greater readiness for climate disasters lose less from conflict. The cost of waiting is steep, because after a high intensity conflict GDP per capita typically falls about 20% over five years.
The hunger: four counts, one direction
Hunger is counted in several ways, and 2025 produced a new reading from each. They measure different things, so they should not be added or compared like for like, but they point the same way for Africa.
Millions of people. Different measures and coverage; read each bar on its own terms. SOFI counts chronic undernourishment; the others count acute food insecurity (IPC Phase 3 or worse).
SOFI 2025; GRFC 2025; GRFC 2026 via Action Against Hunger; WFP 2026 Global Outlook; World Bank chapter 4
Chronic hunger (SOFI 2025, published 28 July). An estimated 8.2% of the world, about 673 million people, went hungry in 2024, down from 8.5% in 2023 and 8.7% in 2022. The improvement is in Asia and Latin America. In Africa 307 million people, more than 20% of the population, were hungry, and hunger has risen by 113 million there since 2015. By 2030 a projected 512 million people will still be chronically undernourished, almost 60% of them in Africa. The cost of a healthy diet is the transmission belt: the number of people unable to afford one fell globally to 2.6 billion, but in low income countries it rose from 464 million in 2019 to 545 million in 2024, and food price inflation in low income countries peaked at 30% in May 2023 (UN joint release; WFP).
Prevalence of undernourishment, % of population, 2024. SOFI 2025 regional results as summarised by the Food Security Portal. Central Africa “more than 30%”, Eastern Africa “nearly 26%”, Northern Africa “nearly 11%”.
Acute hunger (Global Report on Food Crises). The 2025 edition, released on 16 May, counted 295.3 million people, or 22.6% of the population analysed, in high acute food insecurity in 53 countries and territories in 2024, the sixth consecutive annual increase. Weather extremes, mainly El Niño droughts and floods, pushed 96 million people in 18 countries into crisis, with Southern Africa and the Horn among the worst hit; economic shocks were the main driver for 59.4 million people in 15 countries; and 95.8 million displaced people, 75% of them inside their own countries, lived in the crisis countries (FSIN key messages; Rural 21). The 2026 edition, published on 24 April 2026, counts 266 million people in 47 countries for 2025, more than double the 105 million (11%) of 2016. It cautions that the fall from the 2023 peak of 282 million partly reflects reduced country coverage and fewer surveys, not genuine improvement (World Hunger Education Service). Conflict was the main driver for 147.4 million people in 19 countries, and for the first time in the report’s history famine was confirmed in two places in one year, in parts of Gaza and Sudan (Aaj News).
People in high acute food insecurity, 2025, millions. Global Report on Food Crises 2026 via Action Against Hunger. “Other 37 countries” is derived (266 minus 177).
The forward view. WFP’s 2026 Global Outlook, released on 18 November 2025, expects 318 million people to face Crisis or worse in 2026, more than double the figure for 2019, 41 million of them in Emergency or worse, with more than two thirds of acute food insecurity attributable to conflict driven crises (Al Jazeera). The FAO, WFP Hunger Hotspots report of 11 November found acute hunger likely to worsen in 16 countries and territories between November 2025 and May 2026, with conflict the primary driver in 14 of them. Sudan, Palestine, South Sudan, Yemen, Mali and Haiti were of highest concern. Kenya, Burkina Faso, Chad and the Rohingya in Bangladesh were also listed as hotspots (SDG2 Advocacy Hub; WFP USA).
| Country | Concern level | People in acute hunger | Main drivers named |
|---|---|---|---|
| Sudan | Highest | 24.6M | Conflict, economic shocks, floods, dry conditions |
| South Sudan | Highest | 7.7M | Conflict, economic shocks, floods, displacement |
| Mali | Highest | 1.5M | Conflict and insecurity |
| DR Congo | Very high | 27.7M | Conflict, floods, displacement |
| Somalia | Very high | 4.4M | Conflict, dry conditions |
| Kenya | Hotspot | 2.1M | Economic shocks, dry conditions |
FAO, WFP Hunger Hotspots, November 2025 to May 2026 outlook, as tabulated by Security Management.
The Global Hunger Index (October 2025). The global score is 18.3, barely moved from 19.0 in 2016, hunger is serious or alarming in 42 countries, and at least 56 countries will not reach even low hunger by 2030. At the pace seen since 2016, low hunger worldwide would not arrive until 2137. Conflict fuelled 20 acute food crises affecting nearly 140 million people in the past year. In Africa south of the Sahara, hunger rose in ten countries since 2016, and the undernourishment indicator reached extremely alarming levels in six, Kenya among them. The index also names what works: Mozambique, Rwanda, Somalia, Togo and Uganda each cut their scores by more than five points since 2016 (GHI synopsis; People in Need; Down To Earth).
Why the numbers differ
The World Bank’s poverty figures use a $3 a day line and relate to 2025. SOFI counts chronic undernourishment, a measure of habitual calorie shortfall, while the Global Report on Food Crises and WFP count acute food insecurity on the IPC scale, a snapshot of crisis level hunger in analysed areas. The World Bank’s 200 million for fragile economies is drawn from the Food Security Information Network’s 2024 data and covers a different set of countries from the 53 in the 2025 report. Coverage also moves year to year, which is why a falling headline can coexist with a worsening reality.
Sudan and the Horn: where hunger is worst, and where it eases
Sudan is the world’s largest hunger crisis and the clearest test of what conflict does to a food system. The IPC analysis published on 3 November 2025 found 21.2 million people, 45% of the population, in Crisis or worse in September 2025, including 6.3 million in Emergency and 375,000 in Catastrophe. The independent Famine Review Committee confirmed famine in El Fasher and Kadugli, the second confirmation in Sudan in less than a year, and flagged 20 more areas at risk. Screening found global acute malnutrition between 38% and 75% in El Fasher and 29% in Kadugli. Nearly 9.6 million people were displaced inside the country (IPC special snapshot; allAfrica).
People in IPC phases, millions. IPC Special Snapshot, 3 November 2025. Phase 3+ includes the Phase 4 and Phase 5 groups.
The same analysis carries the most useful sentence in the year’s hunger literature. The national total slipped by about 6% because conditions stabilised and access improved in central Sudan, where fighting receded, and FAO, WFP and UNICEF summarised it as hunger and malnutrition easing where conflict subsides (FAO, WFP and UNICEF release, 4 November 2025). Returns of around 2.6 million people to their areas of origin were part of that picture. The same release stressed that seeds, tools and livestock are lifelines for Sudanese farmers and herders, and that restoring access and local production is essential to saving lives. That is a food systems claim, and it is the one a cooperative can act on.
The regional cut for the six IGAD countries (Djibouti, Kenya, Somalia, South Sudan, Sudan and Uganda) shows how fast the Horn’s problem has grown. The number of people in high acute food insecurity there rose from 13.9 million in 2016 to 41.7 million in the 2025 edition, published on 16 September 2025, with Sudan alone at 24.6 million (Nairobi Business Monthly). The 2026 edition, launched in Nairobi on 24 September 2026, restates 2025 at 43.3 million and finds 40.3 million, 34% of the population analysed, in 2026. The fall is due only to Sudan and Uganda improving; conditions worsened in Djibouti, Kenya, Somalia and South Sudan after poor October, December 2025 rains hit crops, pasture and water. The region holds nearly 14.5 million internally displaced people and 4.5 million refugees and asylum seekers (Kenya News Agency; allAfrica).
People in high acute food insecurity, IGAD region, millions. IGAD Regional Focus of the Global Report on Food Crises; the 2025 value is as restated in the 2026 edition (the 2025 edition reported 41.7 million).
Sudan’s 19.5 million covers February to May 2026; “other five” is derived (40.3 minus 19.5). Source: IGAD regional focus, 2026.
Kenya in 2025: good rains, thin recovery, then the short rains failed
Kenya’s 2025 is a compact version of the continental story. The drought that followed failed rains from late 2024 left about 2.2 million people in Crisis or worse in February and March, 266,000 of them in Emergency, mainly in Turkana, Mandera, Garissa, Wajir and Marsabit (IFRC DREF report). The March, May long rains were near or above average, all arid and semi arid (ASAL) counties had returned to the normal drought phase by June, and the July long rains assessment found 1.8 million people in Crisis or worse, with 180,000 in Emergency in Baringo, Mandera, Marsabit and Turkana (REACH; WFP Kenya country brief). The relief was real but shallow. Humanitarian researchers noted that the cumulative effects of several failed seasons still undermined livelihoods and food availability, and the assessment projected 2.1 million people in Crisis or worse by January 2026 because short rains were forecast below average (IFRC). By December the National Drought Management Authority (NDMA) had nine of the 23 ASAL counties in Alert and Mandera in Alarm, with Mandera, Marsabit and Turkana in IPC Phase 3 (NDMA bulletin as reported). The 2026 regional report later confirmed the October, December rains had failed.
People in IPC Phase 3 or worse, millions. February, March 2025 (short rains assessment), July, September 2025 (long rains assessment), projection October 2025 to January 2026.
The refugee camps show what an aid cut looks like on the ground. Kenya hosted 837,907 registered refugees and asylum seekers in March 2025, 475,000 of them in Dadaab and 302,000 in Kakuma (The Standard). WFP supplied 60% of the minimum food basket in 2023, 45% in 2024 and 40% in early 2025; on 20 May it announced a cut to 28% from June and an end to cash assistance for the 720,000 refugees it supports, the lowest level ever (Capital FM). In Kakuma, ration cuts triggered protests in March in which four people were shot (The Standard). The Associated Press reported in June that rations at Kakuma had been halved, after the Trump administration paused support to WFP in March (AP via PBS; Africanews).
%. WFP figures as reported by The Standard (2023, 2024, early 2025) and Capital FM (planned from June 2025).
West Africa and the Sahel: the same squeeze, earlier in the year
West Africa’s lean season arrived with the aid cuts. In March 2025 the regional Cadre Harmonisé analysis found 34.5 million people in Crisis or worse across the Sahel, West Africa and Cameroon, projected to rise to about 49.4 million between June and August if measures were not taken (CILSS). In Nigeria the February, March analysis projected 30.6 million people in crisis or worse for June, August, down 2.5 million from the 33.1 million projected in November 2024 (Leadership). WFP warned in March that funding shortfalls would force it to suspend assistance to 2 million people in the Central Sahel and Nigeria in April (WAM). The pressure continued: Nigeria’s analysis for June, August 2026 projects 34.7 million people in acute food and nutrition insecurity (Leadership).
People in Crisis or worse (CH Phase 3+), millions. Sahel, West Africa and Cameroon, March 2025 analysis; Nigeria, February, March 2025 analysis, projected June, August.
Conflict is the main driver, and its mechanism is the one the World Bank describes: markets, roads and fields stop working. The CILSS review noted that conflict continues to disrupt markets, erode livelihoods and displace millions across 15 countries, and that around 9.1 million people were forcibly displaced across the region at the start of 2025 (FSIN regional report 2025).
The money: emergency aid is thinning, inputs are dearer, adaptation finance is not arriving
Emergency aid
WFP expected $6.4 billion in 2025, 40% below the $10 billion of 2024, with the United States, historically its largest donor, cutting sharply and other donors reducing too (Invezz). For 2026 it aims to reach 110 million of the most vulnerable at a cost of $13 billion and expects to receive close to half of that (WFP via allAfrica). Across the whole food crisis system, Action Against Hunger, working from the 2026 Global Report on Food Crises and the Hunger Hotspots reports, finds humanitarian food aid fell 59% between 2022 and 2025, back to levels last seen in 2016 and 2017, while official development assistance fell 23% in 2025 (allAfrica; 3BL Media). The conclusion is not a plea for more of the same aid. If emergency money is thinning, durable answers must come from local food systems that produce, store and trade through shocks, and from private and blended capital that can be shown where to go.
Index, 2022 = 100. Action Against Hunger, The 10 Largest Global Food Crises, using Global Report on Food Crises 2026 data.
US$ billions. 2024 received and 2025 projected (WFP); 2026 plan; 2026 expected funding is “close to half” of the plan and is plotted at 6.5 (derived).
Inputs
The World Bank’s October 2025 Commodity Markets Outlook set up a squeeze for farmers. Food commodity prices were projected to fall 6.1% in 2025 and energy prices 12%, but fertiliser prices were projected to rise 21%, with urea up 30% and DAP up 26%. Rising input costs and trade restrictions, including Chinese export limits on nitrogen and phosphate and EU tariffs on Russian and Belarusian fertiliser, drove the rise, and affordability indexes remained above pre-2022 levels for all three major fertilisers (World Bank; Techjuice; Miller Magazine). Kenya’s press picked up the implication: fertiliser was the only major commodity group defying the easing trend, which threatened production costs (Eastleigh Voice). The pressure carried into 2026, when the World Bank said the Strait of Hormuz carried nearly a third of global fertiliser trade and urea prices jumped sharply as flows were disrupted (Milling MEA; gCaptain).
Projected change in 2025 price index, %. World Bank Commodity Markets Outlook, October 2025. Red = prices a farmer pays rise; green = prices fall.
Adaptation money
UNEP’s Adaptation Gap Report 2025, released ahead of COP30, puts the adaptation finance needed in developing countries at $310 billion a year by 2035 on modelled costs, and $365 billion when extrapolated from national plans. International public adaptation finance was $26 billion in 2023, down from $28 billion, so needs are 12 to 14 times current flows. The Glasgow goal of doubling to about $40 billion by 2025 will be missed, and the $300 billion a year Baku goal covers mitigation as well. Private adaptation finance is about $5 billion a year and could reach $50 billion with policy support and blended finance (UNEP; Mail & Guardian).
The farm level picture is starker. Climate Policy Initiative found that small scale agrifood systems received $5.53 billion, 0.8% of the $660.2 billion tracked, in 2019 to 20, a 44% fall from 2017 to 18, even though smallholders supply about 35% of the world’s food. Public sources provided 96% of it, and sub Saharan Africa received the largest regional share at 34% ($1.86 billion). The unmet financing need is put at $170 billion a year for smallholder farms and $106 billion for agri SMEs (Green Queen on CPI). At COP30’s opening the Gates Foundation committed $1.4 billion to smallholder adaptation, including AIM for Scale, which aims to give 100 million farmers AI based weather forecasts by 2030 (AgTech Navigator). That helps, and it also shows the scale: $1.4 billion is under 1% of the $170 billion annual smallholder financing gap.
US$ billions per year. UNEP Adaptation Gap Report 2025 (needs by 2035, flows in 2023, private potential) and Climate Policy Initiative (smallholder finance).
The African answer: Kampala
The Kampala CAADP Declaration, adopted by 55 AU member states on 9 to 11 January 2025 and in force from 1 January 2026, replaces the Malabo agenda, whose 2025 targets no member state was on track to meet as of 2023. It sets 22 targets under six commitments, including a 45% rise in agrifood output by 2035, $100 billion of investment, tripling intra African trade in agrifood products, a 50% cut in post harvest loss and 35% of agrifood GDP from locally processed food (AU; AKADEMIYA2063; WOAH). It also shifts from agricultural growth to a wider agrifood systems approach. The Africa Food Systems Report 2025 shows why: agricultural output has grown 4.3% a year since 2000, the fastest in the world, yet one in three African children under five is stunted, and Africa’s food import bill exceeds $100 billion a year (AGRA). Output alone has not been enough.
What the evidence says works
Four studies published in 2025 carry most of the weight of this section. None was run on a smallholder farm, so each transfers by analogy, and the table at the end of the section says how far. What they share is a point that the aid squeeze makes urgent: the cheapest gains come from how support is designed, not from how much of it there is.
Insurance beats advice, even for stocking shelves. Grady Killeen ran two field experiments with over 1,200 Kenyan retailers on a new, effective motorcycle helmet made at a $3 million factory near Nairobi. Although more than half of nearby shops thought helmets would sell, under 6% stocked them two years after the factory opened. In the first experiment a temporary returns policy lifted adoption from 6.8% to 16.4%. In the second, 350 small retailers in western Kenya were offered stock with or without an insurance contract; insurance raised stocking by 50%, 10 percentage points, with the largest effects among shops whose owners were most risk averse. The study’s own reading is that capital was not the binding constraint, risk was (World Bank blog; Georgetown GUI2DE). Replace helmets with a new seed variety or biofertiliser and this is the agrodealer’s dilemma. Dealers do not refuse good products out of ignorance; they refuse because the downside lands on the household. Consignment and return terms are a cheaper fix than another workshop.
% of retailers stocking helmets. Returns policy: Killeen’s first experiment (6.8% to 16.4%). Insurance: +10 percentage points, a 50% rise, which implies about 20% and 30% (derived).
Training works only if people finish it. Cassidy, Ebrahim and Ubfal ran a trial with 2,000 women entrepreneurs in Addis Ababa, chosen to give digital training its best chance. Take up was over 75% for both routes, but 22% earned the certificate on the app against 71% in the classroom, and neither route moved business practices or performance much six months on. In person delivery acted as a commitment device, and in person onboarding was essential even to get people onto the app. The classroom cost about 84% more per enrolled student, but far more per person who completed the app course (World Bank blog; Policy Research Working Paper 10992). For smallholder women facing time and mobility limits, a group meeting at a fixed place and hour is the product, and the app is the reminder. The paper also warns that business training alone offers limited benefits, which points to what comes with it: capital, inputs and markets.
Women led firms, Addis Ababa, % of invited firms. Cassidy, Ebrahim and Ubfal (World Bank working paper 10992 and blog).
Cash before the flood is worth more than cash after. Researchers randomised villages in Bangladesh and Nepal to receive WFP’s forecast based cash within days of a flood trigger, or the same cash after the water had receded. Early recipients scored 1.1 points higher (2.4%) on the food consumption score and were 4.2 percentage points more likely to reach acceptable food security, driven by more households eating meat, and had better mental well being. In Bangladesh they also borrowed less and saved more. After late recipients were paid, the groups looked similar on food security, so early money bought welfare during the crisis, and did not fade (Harris School; Becker Friedman Institute; Policy Research Working Paper 11244). This is the closest thing in the literature to unemployment cover for a farm household, and it is exactly what a forecast trigger makes possible.
Smallholders overestimate the exit. Adrien Dautheville surveyed 900 young market traders in Dar es Salaam. 65% were looking for salaried work while running their businesses. They put the chance of a salaried job within a year at 48%, against an actual transition rate of about 1%, and expected a 70% chance of still operating a year on, when 95% were still open after five months. Active job seekers invested less time and capital and earned lower profits. Telling them the real transition rate shifted beliefs and made them more willing to accept a job at their current earnings, but the effect faded within five months and did not change investment (World Bank blog; IGC policy brief). The lesson for a value chain programme is two sided: do not plan labour and input demand on the assumption that farmers are about to leave, and do not assume that information alone will make them invest.
%, Dar es Salaam retail markets. Dautheville, World Bank Development Impact blog and IGC policy brief, 2025. Business survival is measured after five months.
Mongolia mirrors our drylands. The 2021 to 22 dzud destroyed herders’ feed, and the World Bank’s Livestock Commercialization Project switched on a contingent emergency component that delivered feed, veterinary services and shelters, and one off cash to almost 180,000 herder households. The project also backed local companies building feed hubs that produce, store and distribute concentrate feed through harsh winters, trained nearly 3,000 veterinary professionals, and helped deliver 6.5 million vaccine doses, resulting in zero active cases of the targeted disease by April 2022 (World Bank). The next dzud came in 2023 to 24: the Red Cross reports that over 8 million animals, 12.5% of the national herd, died and 180,000 herder households were severely affected (IFRC). A cash trial in 2021 found recipients bought feed, health care and food and avoided high interest loans, with preliminary evidence of lower livestock deaths (People in Need CashEval). Replace dzud with a failed long rains season in Turkana or Marsabit and the logic travels: feed, vets and early cash, financed before the winter, not after it. The December 2025 Mongolia newsletter adds two financing stories: Golomt Bank’s $160 million package, led by IFC with Proparco and BlueOrchard and Mongolia’s first blue finance deal, targets water management and small businesses, themes close to Maji na Miti’s work; and the Fourth Energy Sector Project, aimed at reliable power for about 200,000 people, is a reminder that cold rooms and mills fail without a steady grid (World Bank Mongolia newsletter, December 2025).
| Study | Setting | Finding | Farm translation |
|---|---|---|---|
| Killeen | Kenyan retailers | Insurance +50% stocking; returns policy 6.8% to 16.4% | Return and consignment terms for agrodealers |
| Cassidy et al. | Ethiopian women entrepreneurs | 71% finish in person; 22% finish app | Group meetings first; digital as reminder |
| Adusumalli et al. | Bangladesh, Nepal floods | Early cash: +4.2 points acceptable food security | Forecast triggered advances for drought and flood |
| Dautheville | Tanzanian market traders | Expect 48% job exit, actual about 1% | Plan on farmers staying; information alone will not shift investment |
Making it investable: the data problem
A cooperative’s difficulty is rarely poor farming. Funders cannot trust what they cannot see, and checking each plot by hand costs more than the money on offer. AgriKonnekt exists to make that check cheap, with sensors, satellites and trained cooperative staff, so that plot data becomes the pipeline. October 2025 supplied the best evidence that data changes the price of risk. The GEMs consortium, 29 development banks and finance institutions set up by the European Investment Bank and IFC, released default and recovery statistics for 1994 to 2024. Loans to more than 10,000 private borrowers in 169 countries defaulted at an average of 3.54%, similar to private firms in advanced economies, and loans to public entities at 2.61%, with recovery averaging 88% for state owned borrowers. For infrastructure, MDB and DFI lending to private counterparts defaulted at 3.6%, below the 4.4% for non financial firms (GEMs press release; PPIAF; CGD). The Center for Global Development called it a data dividend, because the lack of such data had inflated the price of risk.
Average default rates, %, development bank and DFI lending, 1994 to 2024. GEMs Consortium, October 2025; infrastructure and non financial figures from PPIAF, GI Hub.
Agriculture has the same gap: a smallholder is priced at the worst case because nobody holds a repayment record. Verified data, kept honestly over several seasons, is how a cooperative earns a better price on its next loan. The Dakar forum put the same instinct into institutions. AGRA and the Africa Enterprise Challenge Fund launched Africa100, targeting 100 anchor firms in 12 countries in the “hidden middle” that link farmers to markets, and AGRA agreed with Ethiopia, Malawi, Rwanda, Sierra Leone and Tanzania to build “investible flagships”, national pipelines of bankable projects (AGRA). UNEP adds that private adaptation finance of about $5 billion could reach $50 billion with blended finance that de risks investors. Each of these treats the pipeline, not the farmer’s intent, as the thing to build.
PPIAF, the World Bank’s infrastructure advisory facility, reports that it has helped mobilise over $30 billion in private capital since inception, with a further $3 billion in the pipeline, across 131 countries. All approved activities are gender informed and 90% carry climate co benefits. That is roads, grids and water, not farms, but the mechanism transfers: capital arrives when someone has built a pipeline an investor can read. India shows the other side. Its National Infrastructure Pipeline lists more than 7,400 projects worth about INR 111 trillion ($1.6 trillion), and traditional funding covers only 83 to 85% of the need. PPIAF is helping the state of Odisha close the rest by recycling assets it already owns (PPIAF, GI Hub newsletter, November 2025).
PPIAF Annual Report 2025, as summarised in the PPIAF, GI Hub newsletter. US$ billions.
PPIAF, ESMAP, AfDB and GIZ are also training officials in francophone Africa on renewable project finance and auction design, with Morocco’s Masen as lead trainer. The lesson for clean cooking and distributed energy is that regulation and institutions come before the money.
Systems before projects: Tunisia
TERI’s Tunisian Insights No. 13 (Fall 2025) reads like a plumbing manual, which we mean as praise. It launches technical assistance on Water Plan 2050, including wastewater reuse studies and training for water and irrigation operators, on public procurement with life cycle costing and fraud prevention, and on labour market policy. The water work is close to what Maji na Miti attempts in drylands: irrigation operators need capacity as much as pipes. The procurement work matters because the weak joint in a cooperative’s purchase of equipment is usually the contract, not the crop. Our own procurement and anti corruption policies hold us to the same standard.
TERI’s Tunisian Insights No. 13, Fall 2025
The legal aid studies found territorial disparities and thin awareness, and the response was to hold workshops in every governorate rather than the capital alone. Tunisia is also designing an unemployment insurance scheme and a Job Loss Fund. Farmers have neither. A farm income shock behaves like a job loss, and the forecast triggered cash described above is closer to unemployment cover for a farmer than anything now on offer.
Who gets the capital
In October 2025 the World Bank’s Africa Gender Innovation Lab and IFC published a systematic review of 27 rigorous studies on women’s access to capital (Abebe, Cassidy and Weis, working paper 11232). It finds that grants can spur entrepreneurship but the effects are mostly short lived and are not experienced by women running subsistence businesses. Loan design and constraints such as collateral and social norms shape the result (Policy Research Working Paper 11232). The older Ghana experiment points the same way: women with initial profits below the median gained almost nothing from extra capital, and only in kind grants raised women’s profits (NBER working paper 17207). This suggests that cooperatives should tier what they offer women: working capital and storage for those with traction, and savings, skills and time saving support for those without. It also fits the Ethiopian training result above, where even a well designed course did little for business practices.
WEESA, the South Asia lab’s community of practice, held a November session titled “Who Gets the Capital?”, arguing that investment flows disproportionately to men and that a household view reveals women’s contributions hidden by firm level data. Its July digest on time use and a World Bank study of informal women led firms in Peru, which links lower sales to home based work and unequal household duties, make the same point: for many women the binding constraint is the day’s 24 hours (WEESA Year in Review, December 2025).
WEESA members, South Asia Gender Innovation Lab, December 2025 digest. “5,000+” is plotted as 5,000.
Three further items suit a cooperative briefing. WEESA’s March session examined SEWA, India’s Self Employed Women’s Association, on what collectivising does for work, rights and amenities for millions of women, which is the cooperative case at scale. UFGE reports that its Moldova work supports reforms widening women’s participation in agricultural chambers and technical assistance for 300 female farmers by 2026, a modest and honest number, and the WE Finance Code now covers 31 pilot countries and more than 430 financial service providers (UFGE holiday newsletter, December 2025). And the Africa Food Systems Forum in Dakar recognised that placing women at the centre of food systems is vital for gender equality, food security and rural growth, awarding the 2025 Africa Food Prize to two women, among them Professor Mary Abukutsa Onyango of Kenya (AGRA, Forum key decisions).
What it means for a Kenyan cooperative
Pulled together, 2025 gives a cooperative in Kenya a planning picture. The numbers below are the ones that should sit on a board table.
| Indicator | 2025 reading | Source |
|---|---|---|
| People in Crisis or worse, ASAL counties | 2.2M (Feb, Mar); 1.8M (Jul, Sep); 2.1M projected to Jan 2026 | IPC Kenya |
| People in Emergency (Phase 4) | 266,000 (Feb, Mar); 180,000 (Jul, Sep) | WFP Kenya |
| Counties on drought alert, December 2025 | Nine of 23 on Alert; Mandera on Alarm | NDMA bulletin |
| Refugee ration, from June 2025 | 28% of a full ration; cash ended | WFP via Capital FM |
| Global fertiliser price index, 2025 | About +21% (urea +30%, DAP +26%) | World Bank |
| FAO, WFP early warning | Kenya listed as a hunger hotspot, Nov 2025, May 2026 | Hunger Hotspots |
| Global Hunger Index 2025 | Kenya among six African countries with an extremely alarming undernourishment indicator | GHI via Down To Earth |
| 2026 regional outlook | Conditions worse in Kenya after poor Oct, Dec 2025 rains | IGAD via KNA |
| Continental targets from 2026 | 45% more agrifood output; 50% less post harvest loss; 35% locally processed | Kampala Declaration |
What we would ask for
These are proposals, not programmes already run. Each comes with a first test so that it can fail cheaply.
| Proposal | First test | Evidence behind it |
|---|---|---|
| 1. Price the risk the retailer carries | Pilot return or consignment terms with a handful of agrodealers on one new input; measure stocking over two seasons | Killeen: returns lifted adoption from 6.8% to 16.4%; insurance +50% |
| 2. Keep training physical | Hold group sessions at a fixed place and hour; use digital for reminders; compare completion with the 71% classroom and 22% app benchmarks | Cassidy et al. |
| 3. Forecast triggered advances | Explore advances through AgriKonnekt triggered by NDMA drought phases or seasonal forecasts, with a funder willing to accept a trigger in place of a damage assessment | Adusumalli et al.: early cash raised acceptable food security by 4.2 points |
| 4. Fodder and water as value chain assets | Back feed hubs and water points in the drylands, starting where NDMA flags Alert or Alarm | Mongolia feed hubs and herder cash |
| 5. Plan for thinner aid | Build blended and blue finance pipelines, not grant dependence | WFP 40% cut; UNEP private adaptation potential of $50B |
| 6. Tier financial offers for women | Offer working capital and storage to women with traction, and savings, skills and time saving support to those without | Abebe, Cassidy and Weis review of 27 studies |
| 7. Log verified plot outcomes | Record plot outcomes cooperative by cooperative so a lender has a credit history to price | GEMs default statistics |
| 8. Group input purchasing | Pool fertiliser orders early through the cooperative to blunt price swings | World Bank: fertiliser +21% in 2025 |
| 9. Report against Kampala targets | Track post harvest loss and local processing alongside yields | Kampala CAADP Declaration |
| 10. Hold procurement to life cycle costing and fraud checks | Apply to equipment purchases and ask partner cooperatives to do the same | TERI Tunisia procurement work |
How each claim is grounded
Findings come in three kinds: direct (stated by the source), derived (our arithmetic on source figures) and inference (our argument by analogy). The table says which each key claim is.
| Claim | Basis |
|---|---|
| Conflict and instability drive poverty and hunger up fastest in Africa | Direct: World Bank report and press release; Global Report on Food Crises |
| Humanitarian aid for food crises has fallen sharply | Direct: Action Against Hunger and FAO, WFP report a 59% fall, 2022 to 2025; WFP projects a 40% fall in its own funding |
| Hunger is rising in Africa while falling globally | Direct: SOFI 2025 |
| Fertiliser rose while food prices fell in 2025 | Direct: World Bank Commodity Markets Outlook, October 2025 |
| Insurance, in person training and early cash would work for Kenyan smallholders | Inference: the trials were in Kenya (shops), Ethiopia, Bangladesh, Nepal and Tanzania, not on farms |
| Verified plot data can lower the price of agricultural credit | Inference: by analogy from GEMs infrastructure credit data |
| Tiered capital for women improves outcomes | Inference from the review of 27 studies and the Ghana trial |
| Kenya’s 2025 food insecurity peaked in the dry season and eased after the long rains | Direct: IPC and NDMA assessments |
Sources
World Bank: Report page; Press release, 27 June 2025; Chapter 4 (conference edition); Executive summary; Working paper 11210; Tokyo seminar #212; Seminar slides; Commodity Markets Outlook, fertiliser.
Hunger and food crises: GRFC 2025 key messages; GRFC 2025, AFP; SOFI 2025; SOFI 2025, Africa; Hunger Hotspots, Nov 2025; IPC Sudan snapshot; WFP 2026 Global Outlook; Global Hunger Index 2025; GRFC 2026, Action Against Hunger; GRFC 2026 on Gaza and Sudan; GRFC 2026 coverage caution; IGAD regional focus 2026; IGAD regional focus 2025; CILSS Lomé press release.
Kenya: NDMA long rains 2025; IFRC drought DREF; WFP Kenya refugee rations; Refugee services and US cuts; AP on Kakuma.
Finance and policy: UNEP Adaptation Gap Report 2025; Climate Policy Initiative via Green Queen; Gates at COP30; AU on Kampala; Kampala at a glance; Africa Food Systems Forum, AGRA; GEMs press release; PPIAF, GI Hub on GEMs.
Evidence: Killeen; Cassidy, Ebrahim and Ubfal; Forecast based cash, PRWP 11244; Dautheville; Women’s access to capital, PRWP 11232; Mongolia Livestock Commercialization Project; IFRC on the Iron Dzud.
Newsletter sources: World Bank Mongolia (December 2025), Interest Bearing Notes (December 2025), PPIAF, GI Hub (November 2025), TERI’s Tunisian Insights No. 13 (Fall 2025), UFGE holiday newsletter and WEESA Year in Review (December 2025). Kilimora: home page; AgriKonnekt; Gender Lens Programming; Procurement Policy; Anti Corruption Policy; Who pays when the rain fails?.
Differences between sources: the World Bank poverty figures use a $3 a day line and relate to 2025, while the food crisis reports count people in acute food insecurity, a different measure; SOFI counts chronic undernourishment; the IGAD total for 2025 is 41.7 million in the 2025 edition and 43.3 million as restated in 2026; Sudan’s 19.5 million covers February to May 2026 only; the 2026 food crisis report cautions that its drop from the 2023 peak partly reflects reduced coverage.
Our own arithmetic: comparator shares for agriculture in other developing economies (half of the fragile economy shares, from “about twice”); the 37 other countries (266 minus 177 million); the five other IGAD countries (40.3 minus 19.5 million); the 41 in the food aid index (100 minus 59); WFP’s 2026 expected funding of about $6.5 billion (“close to half” of $13 billion); the implied 20% and 30% insurance adoption rates (+10 points, +50%); WEESA’s “5,000+” plotted as 5,000. Kilimora statements about AgriKonnekt and its policies are the company’s own and are not independently audited.