KYEGEGWA, Southwest Uganda — Economist Dean Karlan visited a remote part of Uganda to check on a poverty program he helped design and found something surprising: half the money in a group bank account was untouched. Participants were not borrowing as much as the program allowed.
The program uses the “Graduation Approach,” whose basic logic is simple: you need money to make money. Typical graduation programs give ultra-poor households a small asset grant (around $200), training and coaching to start a microbusiness such as farming, animal husbandry or hairstyling. Those programs have shown success in about 20 countries; a U.N. report called them a “promising ladder from poverty.”
Karlan, founder of Innovation for Poverty Action and formerly chief economist at USAID, introduced an innovation in Uganda: instead of individual grants, groups of roughly 20 households received a block grant of about $4,000 to manage jointly. Groups set their own borrowing rules, interest rates and repayment deadlines. Members could borrow larger sums than individual grants allow, and interest payments would be pooled and shared among the group, so successful entrepreneurs could be rewarded while the most vulnerable weren’t left behind.
The program in Uganda — SMILES (Sustainable Market Inclusive Livelihood Pathways to Self-Reliance) — is run by AVSI Foundation with a $28 million donation from the IKEA Foundation. About 14,000 households, including refugees and local Ugandans, are eligible. Innovation for Poverty Action is monitoring outcomes through the program’s end in 2027 and beyond.
Karlan’s visit focused on why many groups weren’t tapping the block grant. He met participants in twice-monthly group sessions where coaches review businesses. One participant, 23-year-old refugee Jacquerin Kabanyana, had fled war in the Democratic Republic of Congo in 2018 and lived with extended family in a single-room hut. Before the program, occasional day labor left his family surviving on about $5 a week and often skipping meals.
Kabanyana used program funds to start raising and selling goats. Early in the project he reported a $74 grant helped kickstart the business; later he said, “I took out 500,000 shillings [about $140] to buy two goats.” In two years he more than doubled his weekly income to about $13 and began building a larger home and diversifying into sheep and chickens. Yet when Karlan asked why Kabanyana didn’t borrow enough to buy four goats at once, Kabanyana said he wanted to “first see how the market is” and only expand cautiously.
Other group members voiced similar caution. Many refugees had previously received monthly cash for food through the World Food Programme, funded in part by the U.S.; those transfers were cut under the Trump administration’s 2025 overhaul of foreign aid. Participants reported that after the aid cuts, local market activity dropped and customers had less money. That reduced demand made entrepreneurs wary of borrowing more to expand.
Practical obstacles also mattered. Some people said the bank that held the block grant was far away, requiring an all-day trip. Others distrusted the bank. In one meeting, a tin box with the group’s records and some emergency cash sat in the middle of the room; members pointed to it and said the money was their safety net. “It’s where our hearts are. It feeds our families,” one woman said, explaining why they were “more responsible” about withdrawals.
Karlan and AVSI staff recognized a deeper explanation beyond logistics: risk aversion. Households on the edge preserve what little they have, and taking financial risks can be terrifying when failure could mean missing meals. “What we’re seeing clearly is that it’s deeper than you need money to make money,” Karlan said. “You also need to be able to take on risk to make money. And that’s a double whammy for someone who’s poor.”
The block grant was intended to create a faster on-ramp to larger, more stable incomes. It’s designed to reward high performers—people who generate income from the grant share their profits via interest payments—while still supporting the most vulnerable. Karlan likened the approach to a basketball coach maximizing play time for the best players to win as a team. But the team dynamic also means trust, confidence and local economic conditions shape whether people will take loans.
The broader funding environment added urgency. Before January 2025, USAID had been ramping up support for graduation models; Karlan, as chief economist, saw financial support for such programs scale up by nearly half a billion dollars. After the administration’s foreign aid changes, some USAID-funded graduation projects were halted — one in Palabek was terminated just as it was about to launch. With U.S. development assistance deprioritized in the new strategy, program designers felt pressure to “squeeze more juice” from limited resources and to find ways to scale impact.
After listening to participants and coaches, Karlan and AVSI decided on practical tweaks. To remove the barrier of long trips to the bank, groups will be able to borrow from the block grant via mobile money, the electronic wallet system widely used in lower-income countries. Coaches will also continue building trust, encouraging members to experiment and helping them manage risk. AVSI’s director of programs, Rita Larok, gently urged groups to borrow more, reminding them that “this money is supposed to help you make more money.”
There were moments of humor and resolve. In one meeting a group leader called Karlan “jubu jubu,” a Runyankore term for a high authority leader, and asked if he could help them buy a tractor. Karlan replied that they didn’t need his authority; they already had block grant money sitting in the bank — if they could overcome their reluctance to borrow and coordinate, they could invest in larger assets themselves.
Karlan and the project team believe that as group members work together, trust will grow and the block grants will be used more. Over time, successful members can borrow more, generate income and share benefits, helping others graduate out of poverty. But the experience in Uganda shows how fragile those gains can be: global policy shifts, local market demand, distance to financial services, distrust and the intense downside of risk for the very poor all influence whether people will borrow to build businesses.
The SMILES program will continue monitoring outcomes, and Karlan hopes the block-grant model — with tweaks like mobile money and persistent coaching — can help programs be more efficient and reach more communities even as some traditional funders pull back. For participants like Kabanyana, the cautious steps they choose reflect both prudence and the harsh realities of life at subsistence income: building a business is not just about access to capital, it’s about being able to bear risk when the consequences of failure are so immediate.