Direct cash looks 3-4x more cost-effective in a new GiveWell assessment

• Two dozen studies have documented that giving people in poverty cash can grow local economies, benefiting people in poverty who didn’t receive cash aid (“positive spillovers”). Because the money is spent in the community, it supports local small business owners like Mugabo in Rwanda →

• A robust study from Egger et al. (2022)¹ of Kenyans who received GiveDirectly cash in 2015 estimated a 2.5x multiplier – for every $1,000 transfer, $2,500 in local economic activity was created – without causing an offsetting inflation in prices.² Here’s more on the study →

• GiveWell assessed the 2.5x multiplier study: “Cash transfers have significant spillover benefits to other people in their geographic area, including non-recipients and other recipients, in the treated village and in nearby villages too… We estimate that recipient consumption returns explain ~30% of a 2.5x economic multiplier, implying ~70% of the multiplier comes from non-recipient consumption gains… Previously, we had a -5% adjustment for consumption spillovers to non-recipients – i.e. we assumed non-recipients would be made slightly worse-off by not receiving a transfer.”

• These results alone drove almost half the 3-4x increase to GiveWell’s cost-effectiveness estimate.

• GiveWell says both of their discounts are highly subjective: ”The ultimate adjustment we settle on is highly subjective, and we think reasonable people could place more or less weight on the Egger et al. estimate.”

• They discount the 2.5x study by 40% because of uncertainty in the result: ”We don’t think we should put all our weight on Egger et al. for several reasons: [1] The multipliers have large standard errors… [2] The paper doesn’t observe net imports… [3] The paper can’t observe ambient effects across the study area… [4] We don’t think it completely supersedes previous studies… [5] the multiplier seems surprisingly large. Our best guess is to apply a 60% internal validity adjustment (i.e. a -40% downweight) to the Egger et al. spillover results.“

• They discount this 2.5x study by another 20-30% because they believe it doesn’t replicate fully in other places where we work: ”Our best guess is that spillovers to non-recipients are likely to be smaller (but still positive) under more saturated program designs and in poorer and more remote contexts… We apply an 80% external validity adjustment in Kenya (i.e. a -20% downweight) to account for this, and slightly steeper 70-75% adjustments in Malawi/Mozambique/Rwanda/Uganda.“

• • Notably, new work from the authors of the 2.5x study predicts the opposite: “multipliers will be larger in more remote contexts… [and] real economic multipliers remain fairly constant as the size of the cash injection scales,“ which GiveWell reviewed and concluded “we find the arguments in favor of expecting smaller spillovers under more saturated programs in more remote contexts more convincing.“

• Consumption, a widely-accepted measure for determining poverty level, measures the total amount of goods and services consumed by individuals or households over a specific period.

• There are very few long-term (3+ years on) studies of any poverty interventions, but cash has more than any other. This limited literature finds the consumption impacts of cash diminish but do not dissipate in the long-term, with measurable positive impact even 12 years later.³

• GiveDirectly’s longest run published studies measure impacts <3 years later, including in the Egger et al. study which found ~2 years after receiving cash, families were spending 13% more than the control group. Unpublished data from the same study has found 5-7 years later these families are still spending 12% more.

• After consulting the evidence above, GiveWell determined, “recipients see large gains in consumption immediately after receiving the transfer, as they spend just over half (~60%) of the transfer in the first year. We expect consumption gains to persist for 10 years, but to fade out after this initial spike, as recipients spend down the rest of the transfer and the increase in local economic activity starts to fizzle out.”

• This is a more sustained increase than they’d previously assumed: “the direct consumption benefits of these transfers for recipients persisted more strongly than our original model predicted, further increasing our estimate of the impact of GiveDirectly’s cash transfers.”
• They reviewed evidence of impacts 5-7 years later that is higher than they’re predicting now, but did not weigh it heavily: “Our main uncertainty stems from preliminary results we’ve been sent from a long-run follow-up to Egger et al (2022)… [showing] 5-7 years later, recipient households had 12% higher consumption compared to households in distant control villages – almost the same difference as at the 2 year follow-up. We don’t put much weight on these results because they haven’t been subjected to external scrutiny, and seem generally inconsistent with other experimental findings. However, if these results make it through academic peer review, we would probably update towards them.” More how these results could increase their assessment here →

• GiveWell says their “previous model assumed that baseline consumption in Malawi, Mozambique, Rwanda and Uganda was similar to consumption levels we had calculated in Kenya. Now, we believe that consumption in those newer locations is lower than Kenya.”

• They assess that the poorer the recipient, the more good the cash does: “we think a $1,000 annual consumption gain for someone consuming $1,000 worth of goods and services a year is more valuable than a $1,000 gain for someone consuming $2,000 worth a year. This means it matters how poor we think recipients are before they receive cash transfers.”

• Because we delivered cash in these poorer contexts our Poverty Relief “program looks more cost-effective because the average recipient of GiveDirectly’s program is poorer than our previous estimate, which only assessed consumption in Kenya.”

• Researchers estimate through a process called meta-analysis that providing large cash transfers led to significant reductions in mortality among children under 5 years of age (14%) and women (30%). 

• This aligns with research on GiveDirectly’s work: a published study of our work in Rwanda found a 70% drop in mortality for children under 5 and an unpublished Egger et al. follow-up found a 46% drop for Kenyan children under 5.

• These are remarkable results, as our Poverty Relief program isn’t targeted to families with children under 5, but rather gives cash to every household in a poor village. And still, it is reducing child mortality per family at rates comparable to giving out antimalarial medicines or routine vaccines. We’re currently exploring a modified version of our program that would maximize reductions in child mortality per dollar spent by giving cash to recipients most at risk of losing a child to preventable causes.

• They assess that GiveDirectly work, “leads to a 23% reduction in the risk of all-cause under 5 mortality. This is based on published results of [studies of GiveDirectly’s programs], which imply large reductions in all-cause under 5 mortality.”

• They heavily discount this evidence because of uncertainty: “We apply steep discounts to these results because they seem much more than the mortality effects estimated from a meta-analysis of government cash transfer programs, and because we don’t have a clear theoretical understanding of how effects could be this large…” Egger et al.’s working paper, due out next year, will provide more insight on the mechanisms that created this large impact.

• For GiveWell “a large mortality effect size doesn’t translate to a large cost-effectiveness estimate because GiveDirectly’s program is relatively expensive – costing ~$1,200 (nominal) per household – and untargeted – as not every household that receives a transfer has young children.”

• As mentioned above, we’re exploring a more targeted cash program to families at risk of losing children prematurely, which we expect would be much more cost effective under GiveWell’s model as their staff weighs preventing the death of a child under 5 highly (~50x more important than doubling someone’s consumption for a year).

GiveDirectly believes in rigorously measuring the impact of poverty interventions and are happy to see direct cash used as a benchmark to set the bar for others. The depth of GiveWell’s cash evaluation is possible because of the 7 independent randomized control trials of our Poverty Relief programs – a rarity for charities. 

As you can see from the above, using that evidence to calculate cost-effectiveness is not a simple task – it requires nuance and subjectivity. We’re grateful GiveWell engages so thoroughly with the evidence on our work and shares [their analysis publicly] for you to consider.

Ultimately, you might draw different conclusions on the impact of direct cash based on how you weigh these research studies and what you believe are the most important outcomes (e.g., your own ‘moral weights’). 

At GiveDirectly, we prioritize the preferences of the recipients –– giving directly lets them choose what they think is best, responding to their diverse values and priorities. We believe primarily evaluating the effects of direct cash on improvements to consumption and health may undervalue the full, diverse scope of the impact on people’s lives like boosts to psychological well-being or multi-generational education benefits.

We’ve reached <4% of people living in extreme poverty in the countries we already work in

GiveWell’s chart below shows that even in countries where we’ve made meaningful progress, the remaining need is immense. In Rwanda alone, we’ve reached just ~130K people in poverty while 6M+ live on less than $2.15/day PPP.

We have the room for much more funding

This year, we expect to deliver ~$84M to Africans in poverty based on current funding, but we could have delivered up to $140M with the team we currently have. 

Over the next 3 years (2025-2027), we could scale to deliver ~$1.4B with additional investment in central functions (tech, HR, programs management, safeguarding, internal audit) and in-country delivery costs, and while maintaining or exceeding efficiency targets.