Assessing Cost Effectiveness: malnutrition, famine, and cause prioritization

Tldr: Recent events have prompted me to reflect on whether tackling malnutrition and/or famine should be relatively more prioritised as a cause area. It also seems neglected as a discussion topic on the forum. I would like to prompt discussion on:

whether donating to tackle current famines is cost-effective, compared to a malaria medication

and to suggest that...

tackling malnutrition could be a major cause area

Write-up:

The recent events: A major consequence of the Ukraine war has been disrupting grain supplies, which in East Africa and other areas (e.g. Yemen) has exacerbated an already precarious problem in food production and famine (following four years of drought). (Bloomberg 17th Aug; the UN 22nd July; International Rescue Committee 11th May.)

More broadly, this is a problem of malnutrition, already a growing trend due to global warming, exacerbated by a temporary crisis.

Though those links show there has been media coverage, I was unaware of the famine, and only recently became aware after a colleague of mine recently visited Somaliland, where the famine is particularly hard-hitting, working with Save the Children.

Given that, I thought the issue may be neglected (by EA/more broadly). More broadly, I hear very little coverage of this. For EA, there are very few posts at all about ‘East Africa’/‘Somalia’ on the forum. On June 11th this year, one post introduced this risk , receiving 62 karma but just 1 comment. Eight days later on June 19th, a separate post soliciting donation suggestions for this cause received 29 karma and 1 comment. On ‘malnutrition’, there is similarly little discussion, mainly prior to 2020. This short post, with 8 karma & 5 comments, makes a similar suggestion to this, but I found little else.

Edit: on ‘famine’ there has been more discussion, e.g. this short post from May with 111 karma prompted 17 comments that partly compared the cost-effectiveness of tackling famines vs tackling malaria.

There are two causes here. One, the famine currently occurring in East Africa and similar regions. Two, malnutrition writ large. The uncertainty I have is around how cost-effective donating to these causes is (famine relief/malnutrition).

Famine, context

My understanding of the situation in East Africa is largely anecdotal, from my colleague who spent a week in Somaliland working with Save the Children to produce media material. Epistemic status, ⁴⁄₁₀?

He reports roughly as follows.

The people of this region travel around with cattle and the like, the produce of which they sell at markets.
Drought has meant it is hard to feed their animals, imposing a lower and biting cap on the number they can maintain, and thus the revenue they can receive at market (and what they can consume to live).
As a consequence, they have less to spend on grain at market, which has increased in price due to supply constraints.
The men traverse the land with the animals seeking water, hoping to return with resources to sell and leaving the women and children at home.
It is particularly the children who suffer, and charities (like Save the Children and the International Rescue Committee) help by providing malnutrition kits, such as special milk formula and high-nutrient peanut paste.
(There are other organisations targeting this cause. Please name any effective alternatives you are aware of.)

Assessing cost-effectiveness in famine relief and malnutrition

Mega charities

The most appropriate benchmark feels like like medicine to prevent malaria, which GiveWell estimates at $5,000 per life saved based on “exceptionally strong” evidence (nets come in at $5.5k). But cost-effectiveness decreases as the highest-priority funding opportunities are funded. (Malaria has received much (more) attention from EA, and I am interested in donations on the margin.)

Save the Children is not recommended by GiveWell, but this is largely because ‘mega charities’ (budgets of $250m+) are difficult to assess (though size may also obscure ineffectiveness). Scale has long been an obstacle to effective funding decisions and allocation (see GW’s 2011 explanation), and Save the Children is not exempt: GiveWell noted in Dec 2018 that StC was included in the ‘mega charity → hard to assess → uncertainty → less likely to recommend’ bucket. SoGive also rates Save the Children far down the list on cost-effectiveness, again due to absence of information (the AMF is rated ~5.5x as effective.)

As an aside, how GiveWell values uncertainty has an impact on cost-effective analysis, and the lack of ratings for mega charities has been considered problematic. For example, SoGive nonetheless rates Save the Children as ‘promising’, so we may be missing cost-effective opportunities. I would consider this post a success if it prompted discussion on how to mitigate uncertainty or improving how we value it.

Indeed, uncertainty regarding mega charities is not a sufficient reason to not donate to them at all. In Open Philanthropy’s overview of their main allocations for 2021, malaria and malnutrition, respectively, were the top two cause areas, and of the $27m allocated to malnutrition, $20m went to the International Rescue Committee, a mega charity (budget: $700m+). An equal amount ($27m) was allocated to malaria, for comparison of how Open Phil weighed up those cause areas (100% to the Malaria Consortium, a top-rated GiveWell charity).

Specific programmes: GiveDirectly

The specific programme matters as much, if not more, than the charity or the method. The attention paid by GiveDirectly to this cause in a Yemeni context prompted me to update upwards on my beliefs about how cost-effective this cause could be.

GiveDirectly currently has a banner on its homepage noting that food prices have doubled in the past year in Yemen. Not East Africa, but they are directing site visitors to their campaign for unconditional cash transfers. They have raised $4.7m out of a $6.5m funding goal.

(GiveDirectly was one of GiveWell’s top-rated charities from 2012 to Aug 2022, when it was demoted after a criteria change, “even though we still consider it an outstanding program.” See GiveWell’s summary here.)

A growing cause area?

In Open Phil’s 2021 overview (published November), GiveWell suggests that “malnutrition is a very promising area for charitable funding in the future.”

They clarify further: “To give a sense of what we expect, we would not be surprised if GiveWell directs as much funding to malnutrition in the future as we have to malaria programs in recent years.”

In their complementary post on this forum (14 karma, no comments), they note malnutrition as a “promising new area,” suggesting that it “may have large, cost-effective funding gaps” (emphasis added).

Opinions very welcome:

is donating to tackle current famines is cost-effective, compared to a malaria medication (which has received more attention from EA, though not necessarily the rest of the world)?
how should we assess the cost-effectiveness of donating on malnutrition and famine relief (particularly at mega charities like the International Rescue Committee)?

Links:

GiveDirectly (Yemen)
Save the Children (East Africa)
International Rescue Committee (generic)