GiveDirectly

Based on Wikipedia: GiveDirectly

What If the Best Way to Help the Poor Is Just to Give Them Money?

Here's a question that haunted four graduate students at Harvard and MIT: What if everything we thought we knew about charity was wrong?

For decades, the development world operated on a fundamental assumption. Poor people, the thinking went, couldn't be trusted with cash. Give them money and they'd waste it on alcohol, cigarettes, or frivolous purchases. Better to give them things—mosquito nets, textbooks, seeds, training programs—chosen by experts who knew what they really needed.

Paul Niehaus, Michael Faye, Rohit Wanchoo, and Jeremy Shapiro weren't so sure. Their research kept pointing to an uncomfortable truth: the poor weren't stupid. They were just poor. And maybe the most respectful, most effective thing you could do was hand them money and get out of the way.

In 2012, they formalized this radical idea into an organization called GiveDirectly. The concept was almost embarrassingly simple. Find extremely poor people. Send money directly to their mobile phones. No strings attached. No requirements to spend it on approved items. No lengthy applications or humiliating interviews. Just cash.

The Mobile Phone Revolution Made It Possible

The timing wasn't accidental. A quiet revolution had been sweeping through the developing world: mobile money. In Kenya, a system called M-Pesa (the "M" stands for mobile, and "pesa" means money in Swahili) had transformed how people moved cash. Suddenly, anyone with a basic cell phone could receive, store, and send money without ever stepping inside a bank.

This mattered enormously. Traditional aid often lost huge chunks of funding to overhead—the salaries of administrators, the cost of shipping physical goods, the inevitable "leakage" as money passed through multiple intermediaries. Mobile transfers cut through all of that. GiveDirectly could send a dollar and have nearly a dollar arrive in someone's pocket.

The organization started in Kenya, targeting villages where people lived on less than a dollar a day. Their method was direct but dignified: identify the poorest households (often by looking at visible markers like whether a home had a thatch or iron roof), verify their identity, enroll them in the mobile money system if they weren't already, and transfer the funds.

Early Results Defied the Skeptics

What did people actually do with the money?

Not what the skeptics predicted. Study after study found that recipients didn't blow the cash on vices. Instead, they invested in exactly the things development experts might have chosen for them—but better tailored to their specific circumstances. They bought iron roofs that would last for years. They purchased livestock that could generate ongoing income. They paid school fees. They started small businesses.

One finding stood out as particularly striking. Researchers discovered that each dollar transferred generated about $2.60 in local economic activity. The money didn't just help the recipient—it rippled outward through the community. The family that received cash bought supplies from a local shop. The shop owner could then afford to buy more inventory. The supplier hired additional help. The multiplier effect was real and measurable.

Google noticed. In December 2012, the tech giant awarded GiveDirectly a $2.4 million Global Impact Award. Three years later, Good Ventures—the foundation started by Facebook co-founder Dustin Moskovitz and his wife Cari Tuna—contributed $25 million. The organization that started as a graduate student experiment was becoming a serious force in global development.

The Universal Basic Income Experiment

In April 2016, GiveDirectly announced something unprecedented: a twelve-year experiment testing universal basic income in rural Kenya. Universal basic income, often abbreviated as UBI, is a concept that has fascinated economists and social theorists for centuries. The idea is simple but radical: give everyone a regular, unconditional payment, regardless of their employment status or other circumstances.

This wasn't just another cash transfer program. This was a rigorous scientific test designed to answer fundamental questions about human behavior and economic development.

The experiment divided more than 26,000 people across 300 villages into carefully designed groups. Forty villages received long-term basic income—about 75 cents per adult per day, delivered monthly for the full twelve years. Eighty villages received the same monthly amount, but only for two years. Another eighty villages received a single lump sum payment equal to the total value of the two-year stream. And one hundred villages served as the control group, receiving nothing.

Why such elaborate design? Because researchers wanted to untangle complex questions. Does it matter whether money comes in a steady stream or all at once? Do people need ongoing security, or just a one-time boost? What happens when the payments eventually stop?

The results, as they've emerged, have been remarkable. Infant mortality rates dropped by nearly half in communities receiving basic income. Deaths among children under five fell by 45 percent. These weren't marginal improvements—they were transformative changes in the most heartbreaking statistics any community faces.

When Disaster Strikes

In 2017, Hurricane Harvey slammed into the Texas coast, dumping over 60 inches of rain in some areas and causing catastrophic flooding. For the first time, GiveDirectly deployed its model in the United States.

The organization distributed cash-loaded debit cards to residents of Rose City, a small community that had been devastated by the storm. It was a test of whether the same principles that worked in Kenyan villages could translate to American disaster relief. The traditional model—warehouses of donated goods, complex application processes, restrictions on how aid could be used—suddenly seemed cumbersome compared to simply giving survivors money and trusting them to know their own needs.

When COVID-19 struck in 2020, GiveDirectly scaled this approach dramatically. They raised $118 million for pandemic relief in the United States and another $76 million for African countries. The organization sent cash to 116,000 American families and 342,000 families across Kenya, Liberia, Malawi, Rwanda, and Togo.

Finding the Poorest of the Poor

One of the trickiest challenges in any aid program is targeting: how do you identify who actually needs help the most? In wealthy countries, tax records and government databases make this relatively straightforward. In places where many people work in informal economies and have no formal documentation, it's enormously difficult.

GiveDirectly developed an innovative solution in Togo, working with the government's existing cash transfer program called Novissi. The approach combined two cutting-edge technologies.

First, satellite imagery. Algorithms analyzed images from space, looking for markers of poverty that are visible from orbit: the material of roofs (thatch suggests poverty, metal suggests relative prosperity), the size of farm plots, whether roads are paved or unpaved. This helped identify the poorest villages.

Second, mobile phone data. Within those villages, the algorithm examined patterns of phone usage: how long calls typically lasted, how many incoming versus outgoing calls people made, how much mobile data they consumed. People in extreme poverty tend to have distinctive patterns—shorter calls, more incoming than outgoing (because receiving calls is free), minimal data usage. This helped identify the poorest individuals within already-poor communities.

The combination allowed GiveDirectly to extend the government program into rural areas where traditional methods of identifying beneficiaries had proven impossible.

Scaling Up in Rwanda

Since 2016, GiveDirectly has partnered with the Rwandan government on an ambitious effort to accelerate poverty reduction across the country. Working through the Ministry of Local Government, they've delivered unconditional cash transfers to over 200,000 households spread across eight districts.

A striking feature of this work: 84 percent of recipient households are led by women. This isn't accidental. Female-headed households are often among the poorest and most vulnerable in any society, and deliberately targeting them can have outsized impact. Research consistently shows that when women control household resources, more money flows to children's nutrition, health, and education.

Impact assessments tell the story of how families use the funds. About 43 percent goes to income-generating activities—starting or expanding small businesses, investing in agriculture or livestock. Roughly 31 percent funds home renovations—fixing roofs, adding rooms, improving sanitation. The rest spreads across education, health insurance, and increasingly, connecting homes to electricity.

In 2023, GiveDirectly launched a national scale-up, with eligible households receiving at least 1.1 million Rwandan francs (roughly $900) to support what the program calls "sustainable graduation out of poverty"—the idea being that a sufficiently large transfer can help a family permanently escape the poverty trap rather than just temporarily alleviating their situation.

When Things Go Wrong

No organization is immune to problems, and GiveDirectly faced its most serious crisis in 2023. An investigation revealed that in the Democratic Republic of Congo, where GiveDirectly had been operating since 2018, employees had diverted at least $900,000 from several hundred intended recipients in South Kivu province.

The fraud was particularly troubling because it involved the very teams tasked with preventing such abuses. Audit staff—the people supposed to catch theft—were themselves complicit in the scheme.

GiveDirectly paused operations in the country and launched a comprehensive review. The incident highlighted a fundamental tension in their model. The same simplicity that made cash transfers efficient also created vulnerabilities. Fewer intermediaries meant fewer checkpoints. Trusting local staff meant trusting local staff.

The organization's response—public disclosure, operational pause, expanded auditing—reflected a commitment to transparency that's rare in the nonprofit world. But the incident served as a reminder that even the most elegant solutions require vigilant oversight.

A Different Theory of Change

What makes GiveDirectly philosophically interesting isn't just its effectiveness—it's the underlying theory of change. Traditional development work tends to assume that experts know best. Villages need wells, so we'll build wells. Children need education, so we'll construct schools. Farmers need better techniques, so we'll train them.

GiveDirectly flips this assumption. Poor people, the organization argues, are the experts on their own lives. They understand their constraints, their opportunities, their priorities better than any outside analyst ever could. A family might need a roof more urgently than a school if their children are getting sick from exposure. An entrepreneur might benefit more from inventory than from business training she doesn't need.

This connects to a broader intellectual movement called effective altruism, which applies evidence and reason to figure out how to benefit others as much as possible. GiveDirectly has been rated among the most effective charities in the world by GiveWell, a nonprofit that conducts rigorous evaluations of charitable organizations. From 2012 through 2020, they maintained top-rated status—a distinction reserved for the handful of organizations that can demonstrate exceptional impact per dollar donated.

The Academic Debates

Of course, not everyone is convinced. When GiveDirectly released a self-evaluation of its impact in October 2013, World Bank economist David McKenzie offered a nuanced response. He praised the study's design and the organization's transparency about potential conflicts of interest (after all, the people who founded GiveDirectly were studying whether GiveDirectly worked—an obvious concern).

But McKenzie raised two substantive concerns. First, the study relied heavily on self-reporting—asking recipients how they spent the money and how their lives had changed. Self-reports are notoriously unreliable. People might exaggerate improvements to please researchers, or misremember their spending, or report what they think they should have done rather than what they actually did.

Second, the statistical design divided participants into many different groups, which reduced the study's power to detect real differences. Imagine flipping a coin ten times to test if it's fair versus flipping it a thousand times. More observations mean more confidence in your conclusions.

Chris Blattman, a development economist who specializes in randomized controlled trials (studies where participants are randomly assigned to receive or not receive an intervention, the gold standard for causal inference), voiced additional reservations. He worried about the observer-expectancy effect—the phenomenon where people being studied might unconsciously change their behavior or answers based on what they think researchers want to hear.

Blattman also noted the absence of clear positive effects on long-term outcomes in some analyses, as well as the surprising lack of increased spending on health and education. If cash transfers are supposed to let people invest in their futures, why weren't recipients prioritizing these categories more heavily?

These aren't fatal criticisms—they're the normal back-and-forth of scientific inquiry. But they underscore that even promising interventions deserve continued scrutiny.

What Comes Next

In 2022, GiveDirectly appointed British politician and diplomat Rory Stewart as president—a choice that signaled the organization's growing ambition and public profile. Stewart, who had walked across Afghanistan in 2002 and later served in the British Cabinet, brought both development experience and media savvy. The following year, Nick Allardice took over as president and CEO.

The organization now operates across ten countries—Bangladesh, the Democratic Republic of Congo, Liberia, Kenya, Malawi, Mozambique, Rwanda, Uganda, and the United States—with past programs in the Bahamas, Morocco, Nigeria, Turkey, Togo, and Yemen. In January 2025, Canva, the design software company, announced a $100 million commitment to support GiveDirectly's work in Malawi.

That's a long way from four graduate students wondering if there was a better way to help the poor.

The Question That Remains

GiveDirectly hasn't proven that cash transfers are always superior to other forms of aid. Some problems—disease outbreaks, infrastructure deficits, conflict—require coordinated responses that individuals can't organize alone. No amount of mobile money can build a highway or fund medical research.

But the organization has fundamentally changed the conversation. For decades, the development world debated which goods and services to provide to poor communities. GiveDirectly introduced a new question: compared to just giving cash?

Any program that claims to help the poor should now be able to demonstrate that it produces better outcomes than simply handing people money. That's a high bar. And increasingly, organizations that can't clear it are having to explain why.

The simplest ideas are sometimes the most radical. Give money directly to people who need it. Trust them to spend it wisely. Measure what happens. It sounds obvious—which makes it all the more remarkable that it took so long for someone to try.