WS MoreOrLess: Money for nothing?

Episode Summary

Executive Summary: The episode examines GiveDirectly’s experiment in Kenya: giving unconditional cash transfers to very poor households and testing their impact with a randomized control trial. Recipients used the money for food, livestock, housing, education, and marriage expenses, and after a year they were materially better off. The program is presented as both a practical aid model and a case study in the benefits and limits of RCTs in development.

Main Topics: Unconditional cash transfers as aid (Priority: 5/5): MIT students and Paul Niehaus founded GiveDirectly to test whether giving money directly to poor households, with no strings attached, could be a better form of aid than traditional charity projects. How the money was delivered and used (Priority: 4/5): The charity used mobile-phone banking to send funds to rural Kenyan recipients, who spent them on urgent needs like food, goats, housing, and school fees. Randomized control trial as evaluation method (Priority: 5/5): The project used an RCT, giving money to one group and nothing to a control group, to measure impact more rigorously than typical aid assessments. Measured outcomes after one year (Priority: 5/5): Researchers revisited the villages and found recipients had more assets and higher income from livestock investments, suggesting the transfer improved welfare. Ethical and social tensions (Priority: 4/5): The unequal distribution of aid caused resentment in villages, raising concerns about fairness and the ethics of control groups receiving nothing. Limits of randomized trials in development (Priority: 4/5): Experts caution that RCTs are useful but not a magic bullet; they can distort project selection and do not automatically prove the best intervention overall.

Key Arguments: Direct cash transfers can be an effective anti-poverty tool because poor recipients know their own needs better than outside aid workers do. Mobile banking makes large-scale direct payments feasible and relatively easy to administer. The Kenyan trial suggests recipients became better off, with stronger asset holdings and some income gains from livestock. RCTs help isolate the effect of aid from outside factors like rainfall or economic changes, improving causal inference. Ethically, a lottery for aid can be justified when not everyone can be reached, because everyone has an equal chance and the method generates useful evidence. However, positive trial results do not prove cash is always the best intervention, and evaluation methods may influence which aid projects are chosen.

Data Points: Cash transfer amount: Up to $1,000 per family - Amount GiveDirectly arranged to give poor rural Kenyan households Cash transfer amount in local example: £130 / 20,000 Kenyan shillings - Money received by Moses Lagunya in the transcript Trial duration before evaluation: 1 year - Researchers returned after a year to assess impact Asset increase: Almost 60% more assets - Reported difference for households that received cash versus control group Delivery method: Mobile phone banking service - Funds were sent directly to recipients via text-message-linked cash-out agents Household survey time: About 6 hours per villager - Researchers spent extensive time collecting follow-up data in the villages

Pivotal Quotes: "why not just give cash to the extreme poor and allow them to make their own decisions about how to spend it?" — Narrator: Introduces the core philosophy behind GiveDirectly "I would never look at one project in isolation, one experiment." — Paul Niehaus: Explains that one RCT is evidence, but not the whole case for cash transfers "the tail of evaluation starts wagging the dog of doing good development." — James Copestake: Critique that the need to test interventions may distort what aid programs do

Implications: The episode suggests direct cash transfers deserve serious consideration as a benchmark in aid, while reminding listeners that evidence from one RCT is limited and ethical design matters. It also signals that rigorous evaluation is reshaping development practice.