GiveDirectly’s cash transfer programming is quite a bit more diverse today than it was when we first started accepting donations, way back in 2011. In those days, we had a single program design: deliver $1,000 to households in rural Kenya living in poverty. Today we have programs that look similar to that, but we also have programs that are delivering…
- $1,000 to African refugees living in Nairobi, Kenya and Kampala, Uganda→
- $225 to vulnerable families in Mozambique ahead of predicted floods→
- Six months of $25 COVID-19 relief to residents of urban Kinshasa, D.R.C.→
…just to name a few.
That diversity is intentional. We’re trying new designs we think will better help specific populations in poverty. In doing so, we’re partnering with a wider range of funders today, including some that have specific objectives or design requirements. And we think they’re all great projects. But long-time donors may reasonably wonder what it means for them—should they have different expectations about what will happen to a dollar donated today than they did in the past?
Here is what we think you should know:
First, it doesn’t substantially alter how you should expect a typical dollar donated to “Poverty relief – Africa” to be used on dimensions we think are important for direct impact. Specifically, we allocate the great majority of these funds to…
- Transfers that are large. 85% of this spending in 2023 was on programs with total transfer amounts greater than $500 per person or household. This is important because there is evidence that larger amounts let households invest in larger assets that are otherwise difficult for them to purchase.
- Households that are very poor. Of the dollars we delivered last year, a conservative estimate is that over 80% went to households that are extremely poor (per the World Bank’s definition of living on <$2.15 per day) and the rest to households not substantially better off. For 3% of dollars delivered, reliable estimates are not available so we cannot of course be sure, but do not expect poverty rates to be very different.
- Programs with high marginal efficiency. In the last 3 years, 74% of dollars went to our large transfer programs, where 85% was delivered as cash, with the rest covering the cost of getting it there.
Second, the money that does go to alternative program designs (e.g. smaller transfers that are less efficient) does so because we think this will have indirect impacts that make the program even more appealing than our more traditional ones. Typically this involves building relationships and track record with new funders and in new countries, many of whom have pre-set outcomes they want to impact (e.g. nutrition, climate resilience, health).1 By demonstrating how cash can impact those predetermined outcomes, we can unlock funds for cash that might not otherwise have been spent or may have been allocated to less effective programs.
Third, and that having been said, if you wish to make a donation restricted to a traditional program design—large lump-sum transfers to very poor households with high efficiency— you can do so by emailing [email protected].
We recognize that there is the usual issue of fungibility: because GiveDirectly allocates some flexible funding across all of our programs, it is hard (even for us) to say with certainty how a single restricted donation changes the final allocation of funding. But one indication is that in recent board decisions about the allocation of flexible funding, the changes we have made between initial and final budgets have involved moving more money towards more traditional program designs. We think this is a reasonably credible indication that a marginal donation restricted to those designs is likely to “stick.”
Overall, we think the value proposition of a donation towards our anti-poverty work in Africa is the same or better as it was over a decade ago when we first started. But we also realize that this is not as easy to figure out from the outside, and that there is more we can do to make it clearer. Here is what we’re doing:
- Today, posting this note to outline our thinking
- In the near future, updating our FAQs and financials page to better answer the important basic questions touched on in this post
- Over the course of the next year, refining our internal systems for cost attribution and tools for understanding the marginal v.s. average efficiency of a program, so that we can provide more and more meaningful information about the use of marginal donation.
- For example, the program that allowed us to start working in Mozambique in 2022 was a 2 year, $40/month basic income for rural households, funded with $5M from USAID and a $3M match from private donors (i.e. people like you). This operated at a relatively lower efficiency (~60%), in part due to operational start-up costs of launching in a new country. Since then, we’ve started two more programs with USAID Mozambique with higher efficiency (~75%) – all told, $6M in match dollars from GiveDirectly has unlocked another $15M from USAID we wouldn’t have otherwise. In the first two years of this new country, we’ve reached 18k recipients, and will reach ~10k more by year end.