Between 2024 and 2026 a Kenyan macadamia programme won warm reviews from the farmers it served and taught them nothing measurable. The organisation that ran it, Root Capital, said so in public, and its account sits in a March 2026 NextBillion article that this piece reads alongside two other items from the same weeks: the World Bank’s launch of a water in agriculture report that counts 245 million jobs, and its Africa jobs conference, which sets 1.2 billion new workers against 400 million expected jobs. The shared question is practical. When money flows to irrigation, advice and processing, how does an African cooperative or agribusiness know that the change it paid for actually happened? The evidence below is published before 14 November 2026.
Read this first. We read the Root Capital article in full, the World Bank press release and report summaries on water, the abstracts of the Kenyan and Ethiopian irrigation studies, a published summary of the Duke evaluation, and the four newsletters. We did not read the full Root Capital report, the full irrigation papers, the full World Bank report, the Fortell Impact article on AI data collection, or the OECD report on export restrictions, so those are taken from their descriptions. Company and agency figures are their own. Research stops at 14 November 2026.
Five findings
- Water is the largest job claim. The World Bank estimates converting rainfed farming to irrigation and modernising existing schemes could create about 245 million jobs, around 218 million of them in sub Saharan Africa.
- The money exists but is mostly spent elsewhere. Governments spend about US$490 billion a year on agricultural support, mostly subsidies, while irrigation needs an extra US$24 billion to US$70 billion a year.
- The jobs gap is wide. About 10 million to 12 million young Africans enter the labour force each year, against roughly 3 million new formal jobs, by older estimates.
- Satisfaction can mislead. Root Capital’s Kenyan advice service had more promoters than detractors, yet a randomised trial found no significant knowledge gain.
- Reach was narrow. Of more than 6,000 registered farmers, 6% recalled receiving materials and 9% answered the final survey.
- Measured irrigation works, modestly. A Kenyan randomised trial found cheap pumps raised net farm revenue by about 13% of the control mean and paid for themselves within three years.
- Funding is smaller than the need. The World Bank Group’s agribusiness commitments of US$9 billion a year plus US$5 billion mobilised are below the irrigation need alone, though they cover all agribusiness, so the comparison is one of scale only.
The case for water, and the jobs arithmetic
The World Bank’s Nourish and Flourish report, launched on 19 March 2026, argues that current agricultural water practices can sustainably feed less than half of the world’s people and that rebalancing water, crops and trade could create the jobs above and raise agricultural GDP by up to 46%. It says expanding irrigation where water exists, plus modernising current systems, needs an extra US$24 billion to US$70 billion a year to 2050, and that farmers will co invest when finance, equipment, markets and digital tools lower their risks. The starting point in Africa is low: the African Development Bank has put irrigation at only about 6% of cultivated land.
World Bank Group, Nourish and Flourish (19 March 2026), via report summary and press coverage. Estimates, not outcomes
US$ billion a year. Government farm support today (mostly subsidies) and the extra irrigation need, with the low and high estimate. World Bank
The jobs arithmetic is just as stark. A World Bank Africa jobs conference on 4 March 2026 noted that 1.2 billion people will reach working age in developing countries over 10 to 15 years while only 400 million jobs are expected. For sub Saharan Africa, older but widely cited estimates put annual entrants at 10 million to 12 million against about 3 million new formal jobs.
Millions a year, sub Saharan Africa. Entrants are a range; formal jobs are approximate. World Bank figures as reported by Princeton JPIA and African Leadership Magazine
What irrigation shows when someone measures
Irrigation is one of the few African farm technologies with a proper counterfactual behind it. In a randomised trial in three Kenyan counties, Julian Dyer and Jeremy Shapiro gave hand operated pumps costing about US$100 to women household heads by public lottery. After two years, net farm revenue was about 13% of the control mean higher, the pumps paid for themselves within three years, and women’s decision making power rose. That is a real gain, though far smaller than the headline jobs numbers imply.
Solar pumps show the same pattern with a longer lag. A Duke University evaluation of SunCulture’s pay as you go solar pumps, covering more than 750 households in six Kenyan counties from 2023 to 2025 and funded by Shell Foundation and the FCDO, found irrigation use rose from 72% to 95% and the share of plots irrigated roughly doubled. Net income and productivity were stable in the first year, which the funder’s summary treats as a transition period needing longer follow up. In Ethiopia a 2025 study of 161 households found solar pumps viable for only a few crops, with benefit cost ratios above 1 and net water values of US$0.21 to US$1.53 per cubic metre, and found that bundling other technologies lifted output and welfare.
Percent. Households with SunCulture solar pumps, Duke University evaluation 2023 to 2025 (more than 750 households, six counties), as reported by Shell Foundation and Sun Connect. Net income was flat in year one
US$ of net value per cubic metre, range across crops. Solar pump and water harvesting study, 161 households, Central Rift Valley and Upper Awash (2025). Viable for only a few crops
When the farmers said yes and learned nothing
Root Capital’s digital extension advisory service began in Rwanda in 2019, where rigorous assessments found it improved coffee farmers’ knowledge and adoption of agronomic practices. In 2024 it tried the same model for macadamia farmers in Kenya, this time with a randomised controlled trial and a perception survey. More farmers were promoters than detractors. But over 6,000 registered, only 6% recalled receiving materials and only 9% answered the final survey, and both the trial and a difference in differences analysis found no significant knowledge improvement. Root Capital stopped the expansion.
Share of the more than 6,000 registered Kenyan macadamia farmers. Root Capital DEAS evaluation, as reported by NextBillion (2 March 2026)
The authors draw four lessons. Satisfaction measures user experience, not effectiveness. Rigorous evaluation belongs at decision points such as entering a new market. SMS and voice surveys under represent poorer, rural and female farmers. And what works in one value chain may not work in another, because the relationship between aggregators and farmers differs. They add that SMS and voice contact built rapport but did not transmit knowledge in this setting.
Can the money close the gap?
The World Bank Group says it will double annual agribusiness financing to US$9 billion by 2030 and mobilise an extra US$5 billion a year under its AgriConnect initiative. Because a doubling to US$9 billion implies about US$4.5 billion today, the extra commitments sum to about US$14 billion a year across all agribusiness, below the US$24 billion to US$70 billion that irrigation alone needs. The two are not like for like, so the chart compares scale only. The Bank’s own answer is that farmers will co invest when finance, equipment, markets and digital tools cut their risks, and the Kenyan solar pump evidence points the same way: carbon market revenue to lower upfront pump costs and repayments matched to farm cash flows keep pay as you go credit workable.
US$ billion a year. Agribusiness figures from the World Bank Group press release; the 4.5 starting point is derived from “doubling to 9”. Not like for like, as the commitments cover all agribusiness
Measuring without a laboratory
A cooperative cannot run a randomised trial for every service, and does not need to. What it can do is keep a before and after record for members who received a service and for comparable members who did not, count everyone offered the service and not only those who answered, and test knowledge or yields rather than asking whether farmers liked the training. The Moonshots for Development challenge, which lists insurance, agro ecology, digital extension and fragility as tracks, shows funders are asking for solutions in exactly these areas, so evidence of what worked will matter. AgriKonnekt, which links farming cooperatives to mitigation and adaptation finance, depends on this sort of dated member record. The same logic applies to policy: the OECD’s review of export restrictions on cobalt, lithium and nickel found that domestic gains are often limited, a reminder to measure whether a raw material export limit actually raises local processing.
| Claim | Basis |
|---|---|
| 245 million jobs, 218 million in sub Saharan Africa | World Bank estimate under stated assumptions; not an outcome |
| US$490 billion a year of farm support | World Bank press release |
| 6% of cultivated land irrigated | African Development Bank, 2016 |
| 10 million to 12 million entrants, 3 million formal jobs | World Bank figures from older reports; not current |
| Hand pumps raised net farm revenue by about 13% of the control mean | Dyer and Shapiro randomised trial, Kenya; free pumps allocated by lottery |
| Irrigation use rose from 72% to 95% | Duke evaluation as summarised by Shell Foundation; adopters, first year income flat |
| Agribusiness financing to double to US$9 billion a year | World Bank Group press release; a commitment, not a result |
| No significant knowledge gain in Kenya | Root Capital randomised trial; one sector in one country |
Water, advice and processing can all raise African farm incomes, and the jobs arithmetic shows why they matter, but projections of millions of jobs become real only where someone checks the change on the ground. Root Capital’s willingness to publish a null result is the kind of evidence that stops money flowing into services that feel good but do not teach, while the World Bank’s figures show how large the prize is if irrigation and finance reach farmers who then verify their own gains. The cooperatives and agribusinesses that count outcomes rather than approval, and share what they find, will direct the next season’s investment to what works.
Sources & notes
SOURCES: NextBillion, The Impact Sector is Confusing Satisfaction with Impact (2 March 2026); World Bank, Smarter water use could feed 10 billion and create nearly 250 million jobs; World Bank, Nourish and Flourish summary; Engineering News, World Bank report highlights potential of smarter water use; People Daily, World Bank rebalanced water use; Princeton JPIA, Supporting Female Farming Cooperatives; African Leadership Magazine, Millions Entering the Workforce; NextBillion, authors page for Root Capital; Dyer and Shapiro, Pumps, prosperity and household power (Journal of Development Economics); Shell Foundation, How solar irrigation is transforming the lives of smallholder farmers in Kenya; Sun Connect, Solar irrigation in Kenya; Negera et al., Solar pump and water harvesting irrigation in Ethiopia (2025); AgriKonnekt.
DIFFERENCES BETWEEN SOURCES: the 245 million jobs figure is global and the 218 million Africa share comes from the report summary; the irrigated share of African cultivated land is 6% in AfDB based reporting and the comparison with Asia varies by source; the youth entrant range of 10 million to 12 million and the formal job figure of 3 million to 3.1 million come from older estimates and differ by author; the Kenyan pump trial covers free hand pumps allocated by lottery while the Duke work compares solar pump adopters with non adopters; Ethiopian solar pumps were viable for only some crops.
OUR OWN ARITHMETIC: 218 divided by 245 gives about 89%; 6% and 9% of more than 6,000 farmers are roughly 360 and 540 or more; the gap of about 7 million to 9 million a year is entrants minus formal jobs; the bar for the extra irrigation need is drawn from US$24 billion to US$70 billion; US$9 billion halved gives about US$4.5 billion today; 4.5 plus 4.5 plus 5 gives US$14 billion; the irrigation need is about 1.7 to 5 times that; irrigated plots rising from 23% to about 50% is about 2.2 times.