KYEGEGWA, Southwest Uganda — Economist Dean Karlan traveled to a remote district to review a large-scale graduation program he helped design. The initiative follows the Graduation Approach: give extremely poor households a small asset grant (typically around $200) and hands-on coaching so families can start microbusinesses — raising animals, farming, or offering services — and move toward steady livelihoods. Graduation programming has succeeded in some 20 countries and is often described as a promising ladder out of extreme poverty.
For the program in Uganda Karlan introduced an innovation: alongside the usual individual grants, the project put block grants of roughly $4,000 under the joint management of groups of about 20 households. Groups set borrowing limits, interest rates and repayment rules. Members could borrow larger sums than the standard $200, and interest paid on group loans would be shared among all members. The theory was that stronger entrepreneurs could scale faster, generating returns that would flow back to the whole group while protecting the most vulnerable.
The program, called SMILES (Sustainable Market Inclusive Livelihood Pathways to Self-Reliance), is run by the AVSI Foundation with a $28 million grant from the IKEA Foundation. It made some 14,000 households eligible, including refugees and host-community Ugandans. AVSI hires and trains coaches to work with groups; Innovation for Poverty Action monitors progress through the program’s planned end in 2027 and beyond.
At a two-year check-in, Karlan found something unexpected: roughly half of the block grant money remained untouched in bank accounts. Participants were borrowing far less than the design allowed. Conversations with participants and coaches revealed multiple reasons.
A young refugee from the Democratic Republic of Congo, 23-year-old Jacquerin Kabanyana, used about $74 from the program to buy two goats. Over two years he more than doubled his weekly income to about $13, repaired and expanded his home, and added sheep and chickens to his livestock business. When Karlan asked why he didn’t borrow enough to buy four goats at once, Jacquerin said he wanted to test the market and his own management capacity before scaling up.
That kind of caution was common. Several participants described a recent slowdown in local demand. Many refugees had previously received monthly cash assistance from the World Food Programme, funded in part by the U.S.; those payments were cut after a 2025 overhaul of foreign aid, reducing the money circulating in local markets and making larger investments riskier. Practical barriers also mattered: borrowing could require a long trip to a distant bank, sometimes a full-day errand, and some households distrusted formal financial institutions.
Emotional factors were equally powerful. At a group meeting a woman pointed to the small tin box the group kept for records and emergency cash and explained that this money was central to family survival — it was where their hearts were and it fed their families. For people living on the edge, the priority is preserving any safety net they have. The prospect of taking a loan and potentially losing even that small buffer creates acute reluctance. Karlan described this as a double whammy: the ultra-poor need capital to expand, but they are also the least able to absorb failure.
Karlan and AVSI coaches encouraged larger borrowing, arguing that more capital could raise incomes and produce interest payments to share across the group. But responsibility to family, fear of loss, market uncertainty and logistical friction often outweighed that calculus.
Why use a block grant in the first place? Karlan’s intention was to let productive members put in more capital more quickly while maintaining a group safety net. The block grant model is designed to let better-performing entrepreneurs ‘‘play more minutes’’ so the whole group benefits, with interest redistribution smoothing gains across members.
The program’s context heightened the stakes. Before January 2025, USAID had become a growing funder of graduation work, and funding for such programs rose while Karlan served as chief economist. But a major overhaul of U.S. foreign aid under the new administration cut or reprioritized many programs. NPR reported that a separate USAID-funded graduation program in Uganda was terminated just as it was about to launch, leaving thousands without expected support. With development assistance and poverty programs less prominent in the new U.S. strategy and USAID’s role reduced, there is pressure to make each dollar go further — an argument for experiments that could stretch scarce resources.
After hearing participants and coaches, Karlan and AVSI adjusted the design to lower barriers and build confidence. Instead of requiring trips to a branch bank, groups were enabled to access block-grant loans via mobile money wallets, widely used across many low-income countries. Coaches stepped up financial management training and trust-building, with the expectation that repeated interactions and stronger group cohesion would increase use of the shared funds over time.
Those conversations also showed ambition from participants. At one meeting a group leader jokingly addressed Karlan with a local honorific and suggested he should help them secure a tractor. Karlan pointed to the unused block-grant funds in the bank and said they could buy one themselves if they chose to borrow collectively, drawing laughter and nods.
The Uganda experience underscores a simple but important lesson: providing larger, collective capital does not guarantee that people living in extreme poverty will take it up. Decisions not to borrow are rooted in changing market conditions, distance and distrust of formal finance, and the psychological weight of risk when survival is precarious. Program design can adapt — simplifying access through mobile money, strengthening coaching and financial skills, and responding to macro shocks — but durable solutions must address both financial constraints and deeper behavioral and logistical barriers. In an era of shifting donor priorities, making interventions more responsive to these realities may be essential if scarce funds are to help more households climb out of extreme poverty.