Rich countries give poor ones money through official development assistance, and private charities add more. Whether aid helps, and how much, is one of the liveliest debates in development. Since 1970 the United Nations has set a target for rich countries to give 0.7 percent of national income as aid; only a few consistently do.
The case for a big push: Sachs
In The End of Poverty (2005), Jeffrey Sachs argued, in paraphrase, that the poorest countries are caught in a poverty trap: people are too poor to save, invest or protect their health, so they stay poor. A large, coordinated increase in aid, spent on proven interventions such as bed nets, fertilizer, clean water, clinics and schools, could lift them over the threshold to self-sustaining growth. He helped launch the Millennium Villages project to demonstrate the idea.
The case against grand plans: Easterly
The economist William Easterly replied in The White Man's Burden (2006). In paraphrase, he argued that aid has too often been run by “planners”, who design big schemes from above without feedback or accountability, when it should rely on “searchers”, who try small things, find out what works locally, and answer to the people they serve. Large plans, he argued, ignore local knowledge and can strengthen bad governments. Others, such as the economist Dambisa Moyo, went further, arguing that heavy aid dependence can harm governance.
What the evidence suggests
Studies of aid's effect on whole countries' growth have produced mixed and contested results; it is hard to separate aid's effects from everything else going on. Evidence on specific programs is clearer. Campaigns against smallpox, measles and other diseases, oral rehydration therapy for diarrhea, and bed nets against malaria have saved millions of lives. Many economists now conclude that some aid works very well, some is wasted, and the question is which.
Randomized trials
To answer that question, a movement in development economics borrowed a method from medicine: the randomized controlled trial (RCT). Researchers randomly assign a program to some villages or people and not others, then compare outcomes. Because assignment is random, differences can be attributed to the program itself. In 2019 the Nobel memorial prize in economics went to Abhijit Banerjee, Esther Duflo and Michael Kremer for this experimental approach to fighting poverty.
Critics, including the economist Angus Deaton, have raised concerns. A result found in one place and time may not carry over to another. Trials suit small, testable questions better than big ones such as how to reform a tax system. And experiments on poor people raise ethical questions about consent and fairness. Supporters reply that trials should be one tool among many, not the only one.
Cash transfers
One of the most studied interventions is simply giving people money. Conditional cash transfers, pioneered by Mexico's Progresa program in 1997 and Brazil's Bolsa Família, pay poor families if, for example, their children attend school and get health checkups. Unconditional transfers give money with no strings attached. A common worry is that recipients will waste it. Reviews of the evidence have generally found that cash transfers do not increase spending on alcohol and tobacco, and often raise spending on food, schooling and small businesses. Results vary by setting, and cash cannot build roads or train nurses, so most experts see it as a complement to other services, not a replacement.
Cash also embodies a principle: treating poor people as the best judges of their own needs, rather than as passive recipients of what donors think they need.
Worked example: designing a program in a fictional district
A donor has funds to help a poor rural district, Kalendo, where many children are underweight and few finish primary school. How might the different approaches in this level shape the plan?
A big push planner might propose a comprehensive package: a new clinic, school buildings, fertilizer subsidies, a water system and a road, all at once, arguing that the problems are linked and must be tackled together.
A searcher would start by asking residents what they most need, try a few small programs, and keep only what works. Perhaps parents say the real barrier to school is the cost of uniforms, and the clinic is useless without reliable medicine supplies.
A randomista would propose a trial. Randomly choose 40 of Kalendo's 80 villages to receive, say, school meals, and compare attendance and test scores after two years with the other 40. If the results are good, scale up; if not, try something else. Once the trial is over, the comparison villages would receive the program too.
A cash advocate would suggest giving households money directly and letting them decide, perhaps comparing cash with the school meals program in the same trial.
A thoughtful donor might combine elements: consult residents, test two or three options, keep the best, and remember that the trial's results may not carry over to a neighboring district with different conditions. Most importantly, the donor should ask who is accountable if the program fails, and make sure the answer includes the people of Kalendo themselves.