During the first week of November 2025, two stories sat side by side in the World Bank’s newsletters. Hurricane Melissa, the first Category 5 storm to make landfall in Jamaica in recorded history, triggered a full US$150 million catastrophe bond payout. A few pages away, a new report counted the jobs that climate resilience could protect. Between them lies the question this article asks: when drought, flood or storm hits, who pays, how fast, and on what proof? We read the newsletters beside UNEP, the Climate Policy Initiative, IFPRI and others, and then ask what it means for the smallholder cooperatives Kilimora serves through AgriKonnekt.
Ten findings
- The world economy is smooth; the shocks are local. Global GDP per capita grew about 2% a year from 1994 to 2024, yet one hurricane cost Jamaica an estimated US$6 to 7 billion in infrastructure damage.
- Insurance paid, but covered little. The bond and a regional facility paid about US$242 million, roughly 3.5% to 4% of that damage on our arithmetic.
- Adaptation finance is about 8% of the need. International public flows were US$26 billion in 2023 against needs of US$310 to 365 billion a year by 2035.
- Resilience is a jobs policy. The World Bank estimates adaptation could create or protect the equivalent of about 150 million jobs by 2050.
- Smallholders sit at the end of the queue. Small scale agrifood systems got 0.8% of tracked climate finance in 2019/20, while supplying about 35% of the world’s food.
- Kenya has a public insurance programme at scale. It had covered 1.6 million farmers by early 2025, yet about 97% of African smallholders are estimated to lack insurance.
- Plot level proof is the missing input. Area and weather indices cannot say whether this farmer’s plot recovered, which is where 10 m satellite checks and field sensors fit.
- Projects must outlast their funding. The evidence sessions at COP30 asked whether adaptation results are “built to last”, the same test AgriKonnekt sets for member owned practice.
- Disasters are more frequent, and farms bear the cost. FAO counts about 400 disasters a year over the past two decades, up from about 100 a year in the 1970s, and US$3.8 trillion of lost crop and livestock output over 30 years.
- Most of the loss is uninsured. Swiss Re puts 2024 economic losses from disasters at US$318 billion, of which 57%, or US$181 billion, was not covered.
The smooth line and the sharp shock
Our World in Data’s November 2025 Data Insight makes a point that is easy to miss. Global GDP per capita rose from just over US$11,000 in 1994 to around US$21,000 in 2024 (international dollars at 2021 prices), at a remarkably steady rate of roughly 2% a year, interrupted only by 2008 to 09 and 2020. Aggregates hide where the damage lands.
Global GDP per capita, international dollars at 2021 prices. Drawn as a constant growth path between the two endpoints OWID reports, so it smooths over the 2008 to 09 and 2020 dips
Hurricane Melissa shows the other side. Jamaica’s prime minister put infrastructure damage at US$6 to 7 billion. The World Bank confirmed that pre agreed parametric triggers, based on the storm’s central pressure and path, were met, so the 2024 bond paid its full US$150 million. The Caribbean Catastrophe Risk Insurance Facility added US$70.8 million for wind and surge and US$21.1 million for excess rainfall. Money arrived fast, which is the point of parametric cover. But the bond is designed for the largest events: a year earlier Hurricane Beryl caused heavy losses yet narrowly missed its trigger, which reignited debate about basis risk, the mismatch between losses felt and thresholds crossed. Jamaica went back to market in April 2026 for a new US$150 million bond to replace the one Melissa used.
Jamaica, Hurricane Melissa: reported insurance payouts against the top of the prime minister’s damage estimate. Payouts are US$150M + 70.8M + 21.1M = US$241.9M, 3.5% of US$7B
The protection gap before the storm
Jamaica is not an outlier. Swiss Re Institute’s sigma 1/2025 report, published in May 2025, put global economic losses from disaster events in 2024 at US$318 billion, of which 57% was uninsured. That is a protection gap of US$181 billion, up from US$177 billion in 2023. The same report says the United States accounted for nearly 80% of global insured natural catastrophe losses, so insurance is concentrated where cover is already deep. Our reading is that the global insured share says little about a Kenyan smallholder.
Global economic losses from disasters in 2024, US$ billion, total US$318 billion. The insured share is our subtraction (318 minus 181)
What disasters have been taking from the field
FAO’s 2023 flagship report, The Impact of Disasters on Agriculture and Food Security, gave the first global estimate of what disasters cost crop and livestock production: about US$3.8 trillion over the 30 years to 2021, an average of US$123 billion a year, or about 5% of annual global agricultural GDP. Disaster events rose from around 100 a year in the 1970s to around 400 a year in the past two decades.
Disaster events a year worldwide. FAO, The Impact of Disasters on Agriculture and Food Security, 2023
The burden is uneven. Low income and lower middle income countries lost 10 to 15% of their agricultural GDP, small island developing states nearly 7%, and Africa nearly 8% of agricultural added value against 4% for Asia. Drought is the sharpest risk for farms: data from post disaster needs assessments show that over 65% of drought losses fell on agriculture, against about 20% for floods, storms, cyclones and volcanic activity.
Disaster losses in crops and livestock as a share of agricultural value added, 1991 to 2021. The lightest bar marks the top of a reported 10 to 15% range
The losses are of the staples people eat. Cereals lost 69 million tonnes a year on average, equal to France’s entire 2021 cereal harvest, and fruit and vegetables and sugar crops each approached 40 million tonnes. Meat, dairy and eggs lost 16 million tonnes. FAO also says smallholders farming under rain fed conditions bear the brunt, and that the real total is probably higher because fisheries, aquaculture and forestry losses are not systematically counted.
Average annual crop and livestock production lost to disasters, million tonnes, 1991 to 2021. FAO, 2023. Bars for fruit and vegetables and for sugar crops show “approaching 40”
The adaptation gap, in one chart
UNEP’s Adaptation Gap Report 2025, released on 29 October ahead of COP30, puts developing countries’ adaptation needs at US$310 billion a year by 2035 on modelled costs, rising to US$365 billion on country plans. International public adaptation finance was US$26 billion in 2023, down from US$28 billion in 2022, so the gap is 12 to 14 times current flows. About 59% of flows are debt, which raises affordability concerns. The private sector contributes about US$5 billion a year and could realistically reach US$50 billion with targeted policy and blended finance.
Adaptation finance for developing countries, US$ billion a year. Needs are for 2035; flows are 2023. UNEP Adaptation Gap Report 2025
The same report notes progress on planning: 172 countries have at least one national adaptation policy, strategy or plan, and countries report more than 1,600 implemented actions, mostly in biodiversity, agriculture, water and infrastructure. Plans, then, are not the missing piece. Money that reaches the field, with proof it was used well, is.
Jobs: the case that persuades finance ministers
The World Bank’s Jobs in a Changing Climate report, published on 7 November 2025 and covering 93 economies, translates resilience into employment. In 49 countries, climate impacts could cost the equivalent of 43 million “more and better paid” jobs by 2050, while adaptation could create or protect the equivalent of 25 million. Simple extrapolation to all low and middle income countries gives 260 million lost and 149 million created or protected, the “150 million” in the headlines. These are modelled equivalents, not forecasts of headcount.
Equivalent “more and better paid” jobs by 2050, million. Left: 49 countries with country reports. Right: extrapolated to all low and middle income countries. World Bank, 2025
The Bank’s companion numbers point at the labour market Africa is about to face. Africa’s Pulse says the continent’s working age population will grow by more than 624 million over 25 years, about 25 million a year on our arithmetic, and the newsletter adds that 33 million people work in the global forest sector. For Kilimora the link is concrete: adaptation work in agroforestry, soil carbon and farm records is also a source of local jobs, from farmer educators to MRV data collectors.
The smallholder’s share: one square in a hundred
If resilience is a jobs policy, smallholders are where the jobs are. Yet the Climate Policy Initiative and the ClimateShot Investor Coalition found that small scale agrifood systems received US$5.53 billion, 0.8% of the US$660.2 billion of tracked climate finance in 2019/20, down 44% from US$9.85 billion (1.7%) two years earlier. They supply roughly 35% of the world’s food. The report estimates unmet needs at US$170 billion a year for smallholder farms and US$106 billion for agri SMEs, and found that 96% of the money was public. Sub Saharan Africa received the largest regional share, 34% or US$1.86 billion.
Each square is 1% of US$660.2 billion tracked in 2019/20. The highlighted square is small scale agrifood systems at 0.8%. Climate Policy Initiative and ClimateShot, via IFAD
Kilimora’s own framing differs in measure but not in message: AgriKonnekt reports that smallholders look after 80% of arable land in sub Saharan Africa and grow 70% of its food, yet less than 3% of global agricultural investment reaches them. The two statistics measure different things, climate finance and agricultural investment, so we do not combine them. Both say the same thing about where the money is not.
From sovereign bonds to the single plot
Parametric logic is being pushed down from governments to farmers. ACRE Africa, formerly Kilimo Salama, sells index insurance backed by weather stations and satellite data, with payouts triggered automatically and a mobile first, agent led distribution model. In one programme it covered 43,000 smallholders alongside training. The government’s Kenya Agricultural Insurance Programme, launched in 2016 and analysed by IFPRI, had by the start of 2025 trained over 3.5 million farmers and insured 1.6 million across 38 counties, mainly through area yield and weather indices. In 2025 the sector also saw blockchain based parametric cover for Kenyan farmers, from IBISA with Ensuro and One Acre Fund.
Kenya Agricultural Insurance Programme, farmers, million, at the start of 2025 (IFPRI). Share insured is our division
The gap is still large. The African Development Bank estimates that about 97% of African smallholders lack insurance, and premiums stay unaffordable without subsidy. A further difficulty is basis risk, the same one that haunted Jamaica. An area index can pay a farmer whose crop survived and miss the neighbour whose crop failed. What narrows it is plot level evidence, which is what AgriKonnekt builds for a different purpose. Field sensors read soil moisture, soil carbon and pH, and Sentinel-2 imagery at 10 m resolution checks vegetation and moisture trends on every plot, so each record is cross checked and time stamped. Kilimora is not an insurer, and we do not claim its records have been used in a policy. But the data trail it creates, owned by the cooperative, is the kind of input that could make cover cheaper and fairer to price.
The payment end is already in place. AgriKonnekt pays farmers directly to their mobile accounts within 24 hours once funders release money, and lessons run over USSD on any phone with no data plan. In a drought or flood, speed matters as much as size: a payout after the planting window is a loan repayment, not a recovery.
What Kenya’s herders already show
Kenya has used satellites to trigger insurance payouts for longer than the crop programme has existed. Index based livestock insurance (IBLI), developed by ILRI with Cornell and UC Davis, began in northern Kenya in 2010. It uses a satellite vegetation index (NDVI) as a proxy for forage, and pays when forage runs short, before animals die. In 2015 the government launched the Kenya Livestock Insurance Program (KLIP), buying drought cover from private insurers for vulnerable pastoralists. In its first season it bought about 5,000 fully subsidised covers, each for five tropical livestock units, roughly five cows or ten goats.
The early evidence is encouraging. ILRI’s 2014 research brief linked IBLI to a 36% drop in distress sales of livestock during drought. A VoxDev summary of later research finds that satellite triggered payouts reduced drought induced conflict by around 25%, with cover reaching around 80% of eligible areas by 2019. FAO adds the economic case for acting before a shock: each US$1 spent on anticipatory action can return up to US$7 in benefits and avoided agricultural losses.
Benefits and avoided agricultural losses for each US$1 spent on anticipatory action, up to the figure shown. FAO, 2023
Two cautions apply. These are rangeland results for herders, not for crop plots, and an index measured over a wide area still carries basis risk. The lesson for cooperatives is that the trigger, the payment and the proof can be built from satellite and mobile data. What is still missing for crop farmers is a record fine enough to see their own plot.
Trust, procurement and guarantees
Money that moves fast still needs institutions people believe in. The World Bank’s governance update argues that public procurement in Western and Central Africa is where commitments become roads that survive rainy seasons and clinics that open on time. It also flags the implementation gap in Malawi, which has passed transparency laws but struggles to put beneficial ownership transparency into practice, and the long court delays and high litigation costs that Liberia is working through. The lesson for climate funds is blunt: a rule that exists on paper protects no one without a record that can be checked.
Risk sharing instruments help. The World Bank Group’s guarantee arm, MIGA, reported a record US$9.5 billion in guarantees across 44 projects in fiscal 2025, and its platform now offers project based loan guarantees that help governments attract private investment. AgriKonnekt works on the same principle at cooperative scale: partners such as AUDA-NEPAD, FAO, AGRA and MicroEnergy Credits help derisk smallholder led projects, because lenders see thin data and unrated groups. Verified records and a cooperative that carries accountability turn “unbankable” into something a guarantor can price.
The evidence agenda is moving too. COP30’s Evidence for Climate Action programme, which ran 10 to 21 November 2025, put on sessions on anticipatory cash assistance days before floods in Bangladesh and Nepal, on whether adaptation projects are “built to last”, and on forestry finance and geospatial evaluation. We have not seen those results, so we treat them as questions to follow. The durability one matters most to us. AgriKonnekt’s answer is to hand curriculum, records and decisions to member led organisations so practice continues after a project cycle closes.
New tools in the pipeline
The Agritech Digest points to practical supply. AGRO-WELL has launched a funding call for smart farming tools tested on commercial farms to improve labour efficiency, safety and sustainability. Drone start ups are making pest detection, field mapping and targeted spraying affordable for smallholders. Seed Advisory’s report urges a financial system redesign that connects farmers to digital and climate smart finance and moves from crisis driven funding to long term inclusion. And Startupbootcamp’s BOLD2 accelerator offers €25,000 and €100,000 in support to ten food and agritech start ups, with applications closing on 7 December 2025. All of these are summaries from the digest, not independent assessments, but together they suggest that data collection will get cheaper. The scarce thing will remain trusted, checked, farmer owned data.
The verdict, and what we would ask for
Resilience finance fails smallholders in three places: the money is small, it is slow, and it cannot see the plot. The Jamaica case proves that fast, rules based payouts work when the trigger is trusted. Kenya’s insurance programme shows that scale is possible. The CPI figures show that smallholders are still last in line. Three signals would change this view: a rising smallholder share in the next CPI update, published basis risk results from plot level data, and adaptation grants, not loans, reaching cooperatives.
- Raise the smallholder share of adaptation and climate finance with a published target and an annual report of the actual share.
- Prefer grants and concessional finance for smallholder adaptation, as UNEP urges, to avoid adding debt.
- Let cooperatives hold verified plot records and share them with insurers, lenders and funders on the member’s terms.
- Test plot level cross checks, sensors plus 10 m satellite, to measure and reduce basis risk in Kenya’s index insurance.
- Match every payout to a speed target, such as funds in a member’s mobile account within days of a verified trigger.
- Use guarantees and blended finance to make cooperative led projects bankable at the scale of thousands of farms.
- Evaluate whether adaptation projects last after funding ends, and publish what stays in member hands.
- Count jobs by gender, type and duration, so the jobs promised by resilience can be audited.
- Fund anticipatory action and early warning triggers in cooperative areas, since FAO finds up to US$7 of benefit for each US$1 spent.
- Record disaster losses and payouts by farm type and sub sector, as FAO urges, so the protection gap can be measured where smallholders farm.
Sources & notes
SOURCES: World Bank Newsletter, 14 November 2025; World Bank Governance Global Department update; World Bank Group Guarantees monthly update, November 2025; IEG Evidence for Climate Action Week Two programme, 10 to 21 November 2025; AgriTech Digest issue on nextProtein and AGRO-WELL, November 2025; Our World in Data, Data Insight on global GDP per capita, November 2025 (ourworldindata.org). Additional sources: World Bank, Jobs in a Changing Climate and its overview; UNEP, Adaptation Gap Report 2025: Running on Empty, with summaries from SEI and IPS; World Bank, Hurricane Melissa triggers 100% payout; Insurance Business, CCRIF payouts and bond design and basis risk debate; Irie FM, damage estimate; Artemis, replacement bond, April 2026; IFAD, CPI and ClimateShot report on small scale agrifood finance, and WEF, small scale farmers and climate change; IFPRI, index insurance in Kenya; Business Daily, parametric products for farmers; People Daily, ACRE Africa programme; ICMIF, IBISA, Ensuro and One Acre Fund; Kilimora AgriKonnekt, as shown on 5 November 2026. Earlier sources, published before November 2025: FAO, The Impact of Disasters on Agriculture and Food Security 2023, with summaries from IISD and FAO newsroom; Swiss Re Institute, sigma 1/2025, Natural catastrophes, May 2025; ILRI research brief on index based livestock insurance, September 2014; VoxDev on satellite triggered livestock insurance and conflict; UC Davis BASIS on KLIP payouts and the Financial Protection Forum Kenya country profile. Newsletter and third party figures are as published and have not been independently verified. The 2.2% growth rate, the 3.5% to 4% payout share, the 8% finance coverage, 25 million a year, the 46% insured share, the US$276 billion sum, the US$137 billion insured share of 2024 losses, the “quadrupled” disaster count (100 to about 400 a year) and the chart paths are our own arithmetic. Kilimora statements about its systems are the company’s own.