Bananas for thought
In a personal experiment today, I distributed 23$ worth of food (water, bananas, peanuts) to approximately 15 people experiencing extreme poverty in Medellín.
Context: About 6% of Colombia live in extreme poverty (below $2.15 per day in consumption), i.e. about 3 million people. As of 2022, it is estimated that 735 million are undernourished worldwide (meaning caloric intake that is insufficient to meet the minimum energy requirements necessary for a given individual).
Neglectedness: In my limited personal experience, living in the area for over a year in aggregate, I’ve never seen systematic assistance to the local extreme poor. I’ve contacted GiveDirectly about their relation to Colombia – they do not have the capacity for now to serve the extreme poor here.
Tractability: Straightforward.
A quick comparative analysis:
GiveWell’s Malaria Consortium is estimated to save 1.92 lives per 3000$. They value preventing one under-5 death from malaria at 116.9 points, and preventing one 5-and-over death from malaria at 83.1 points. Let's assume an even split between under-5 and over-5 deaths prevented we get (1.92/2 * 116.9) + (1.92/2 * 83.1) = 112.2 + 79.776 = 191.976 points… and so 191.976 points / $3000 ≈ 0.064 points per dollar.
Our intervention was 23 USD spent on food (water, bananas, peanuts), distributed to about 15 people in extreme poverty, which is roughly equivalent to doubling consumption for one day for these 15 people. GiveWell’s benchmark value is one doubling consumption for one person for one year = 1 point, so for one person for one day = 1 / 365 ≈ 0.00274 points. Total for 15 people for one day = 0.00274 * 15 ≈ 0.0411 points. Cost-effectiveness = 0.0411 points / $23 ≈ 0.00179 points per dollar.
This analysis suggests that the Malaria Consortium is approximately 35.75 times more cost-effective.
It may be that this comparison doesn't capture several factors like addressing overlooked local needs, immediacy of impact, and understanding the complexities of extreme poverty everywhere it appears. Global optimization is crucial for large-scale impact, but there's also value in addressing poverty wherever it exists.
As for my personal relationship to the intervention: I believe my comparative advantage is to do technical work or realizing events towards reducing AI risk. That said, giving food is a good way to take a break, but I think others might find more joy in helping the poor directly.
In conclusion, we should continue supporting highly effective global charities while also exploring ways to address local poverty gaps, potentially by encouraging the expansion of effective charities into underserved areas.
Some questions:
It would be good to know more of the potentially effective charities in Latin America.
I’d like to have more ideas on effective altruistic direct actions I can take during my breaks.
References:
4 of them were sleeping on the pavement, so left the goods beside them. One offered me to smoke from his crack pipe in gratitude. Perhaps 7 of them were two indigenous families, moms and children. A selfish motorcycle driver was annoyed and honked at me because I was in their way to their parking while I was giving the supplies to someone that was sleeping on the ground.