À PROPOS DE CET ÉPISODE
Decades of agricultural development policy have chased yield. Bigger harvests, better seeds, more fertiliser. But how can we make farming more profitable?
Craig McIntosh of UC San Diego is academic lead on a J-PAL Policy Insight covering twenty-three randomised evaluations of credit and grants for farmers in low- and middle-income countries. He tell Tim Phillips that although yields and revenues often rise, profit rarely responds in the same way. When farmers are already running their farms close to the margin, costs rise at the same rate as income, and the household bank balance does not move much. What can we bundle with credit to change that situation?
The research behind this episode:
Abdul Latif Jameel Poverty Action Lab (J-PAL). 2026. "Can relaxing credit constraints boost farmers' profits?” J-PAL Policy Insights. Last modified February 2026. Academic leads: Craig McIntosh and Tavneet Suri; insight authors: Leonie Rauls and Rebecca Toole.
To cite this episode:
Phillips, Tim, and Craig McIntosh. 2026. “Boosting farmers' profits?" VoxDev Talks (podcast).
Assign this as extra listening. The citation above is formatted and ready for a reading list or VLE.
About the guest
Craig McIntosh is Professor of Economics at the School of Global Policy and Strategy, UC San Diego. His research spans development finance, agricultural credit, cash transfer design and the evaluation of large-scale anti-poverty interventions.
Research cited in this episode
Microcredit take-up among farmers. Across four randomised evaluations of traditional microcredit aimed at farmers, in Morocco, Ethiopia, Bangladesh and Malawi, take-up sat between 13 and 33 percent. Standard microcredit repayment begins a week or two after disbursement, which is incompatible with a crop cycle that pays out cash once or twice a year. Group liability also breaks down in agriculture, where shocks like drought or floods hit borrowers together rather than one at a time.
Tailoring credit to the agricultural cycle. Restructured loans push take-up much higher. Nakano and Magezi in Tanzania allowed rice farmers to defer 80 percent of repayment until harvest; 39 percent borrowed and over 92 percent repaid. William Jack and co-authors in Kenya offered dairy farmers asset-collateralised loans for a water tank; take-up reached 44 percent against 2.4 percent for a typical joint-liability product. Lambon-Quayefio, Manjeer and Udry in Ghana offered digital credit with a three-month grace period; 59 percent of farmers took it up.
Sell low, buy high. Burke and co-authors in Kenya showed that smallholders routinely sell at the post-harvest price trough and buy back grain at hungry-season prices 20 to 40 percent higher. Harvest-time loans that allowed farmers to delay sales had take-up of 64 percent and produced returns around 29 percent for borrowers. Treated villages also saw flatter price trajectories, generating spillover benefits for non-borrowers.
Lean-season credit. Fink, Jack and Masiye in Zambia found that lean-season loans let farmers stop hiring out their labour and instead work their own land. Output rose by 9 percent. Loan repayments were comparable to the gain, leaving farmers roughly even on profits.
Selection into credit markets. Beaman, Karlan, Thuysbaert and Udry in Mali first offered loans, then offered grants to those who had refused. Returns to capital among would-be borrowers were on the order of 130 percent. Returns among those who had refused the loan were close to zero. Credit appears to self-target toward farmers who can use it productively, which is regressive in welfare terms and also exactly what a capital-scarce economy needs credit markets to do.
Input subsidy programmes (ISPs). Jayne and co-authors reviewed eighty studies of fertiliser subsidies across sub-Saharan Africa. Yields rise while subsidies are in place; profitability is mixed; targeting is frequently politically distorted, often skewed toward better-connected or wealthier farmers. The standout randomised exception is Carter, Laajaj and Yang in Mozambique, where two-thirds of recipients had never used fertiliser before; the programme produced sustained gains and a high benefit-cost ratio. By contrast, Gignoux and co-authors in Haiti found a fertiliser-voucher subsidy crowded out farmers' own input spending and lowered yields once the subsidy ended.
Cash transfers and diversification. In six studies measuring both farm and non-farm outcomes, three found households doubled down on agriculture and three saw movement into non-farm enterprises. The Zambian Child Grant evaluation by Handa and co-authors saw women invest in seeds, fertiliser and livestock and start non-farm businesses, with household income roughly doubling.
Bundled input programmes. Four randomised evaluations bundled credit or a grant with information, training or market access. All four lifted revenues; three of the four lifted incomes or profits. Harou and co-authors in Tanzania showed that fertiliser vouchers alone and soil testing alone did nothing; only the combination raised yields and revenues. Ashraf, Gine and Karlan's Kenya study on French-bean and baby-corn export found credit increased programme participation from 27 to 41 percent, even where it did not further raise income among participants.
