Say 50, 100, 150 or 200 years ago
[Question] What would an entity with GiveWell’s decision-making process have recommended in the past?
Interesting question—I’m curious what makes you ask?
This looks super interesting to me. We can, in a sense, simulate a longer history of Effective Altruism and see what patterns there are.
Another angle (/ piece of the puzzle) to compare different decision-making processes
Going back a little more than 200 years, would it have recommended supporting the anti-slavery movement? Presumably it would agree that abolishing slavery was good, but the evidence that the movement would work would not be there beforehand, and it might have seemed very unlikely to succeed.
Another question: Would it have supported Christian missionary efforts because of education/healthcare they spread? Would it instead have competed with such efforts?
(I am assuming that we are talking of an organisation founded in the Western world. What a Chinese GiveWell in the 19th century would have done I have absolutely no idea about.)Mayyybe it would have bought slave’s freedom one by one instead? (I don’t know; just speculating)
If that was done before the slave trade was abolished it would have encouraged the enslavement of more people.
Good point, and GiveWell would probably have figured that one out
A couple relevant pieces: In this talk, Tyler Cowen talks about how impartial utilitarianism makes sense today since we can impact humans far from ourselves (in both time and space), but how deontology may have been more sensible in the distant past.
In this talk, Devin Kalish argues that utilitarianism is the correct moral theory on the basis of its historical track record. He argues that utilitarianism correctly “predicted” now widely recognized ethical positions (women’s rights, anti-slavery, etc).
So I think it’s interesting to ask, if GiveWell was around 200 years ago, what would they have recommended, and in hindsight, would that have been the correct cause to advocate for.
One common criticism of EA is that it focuses too much on incremental rather than Systemic Change. We might worry that in 1800, GiveWell would have advocated for better farming practices, but not for abolition, though in retrospect, the latter seems to have been more important.
This is more or less the point Patrick Collison makes here when he says: “It’s hard to me to see how writing a treatise on human nature would score really highly in an EA framework, and yet, ex-post, that looks like a really valuable thing for a human to do. And similarly, when we look at things that in hindsight seem like very good things to have happened, it’s unclear to me how an EA intuition would have caused someone to do so”
Overall, I don’t think this is a super damning criticism. The world has changed. It’s more legible, and more subject to utilitarian calculus.
But still, it’s an interesting quesiton.
Nitpicky point: Depending on how much better the farming practices were, and how wide they might have spread, the hypothetical comparison to abolition may not be as clear as it looks. If this list is even close to accurate, famines seem to have killed millions of people in the average decade of the 19th century. I’m not sure what better practices might have been possible to introduce “early” in that era, but I think EA circa 1800 might have had “famine” as a major cause area!
*****
I can’t easily find the link, but GiveWell’s early discussions of U.S. interventions focused on how difficult it can be to make a permanent change in someone’s life in the developed world. One example (this is mine, not theirs): some of the worst-off people in the U.S. are prisoners, and you can’t pay to get someone out of jail.
On the other hand, Open Philanthropy made multiple grants to the Brooklyn Community Bail Fund, with the goal of reducing the amount of time Americans spend in a state of imprisonment. The first of these came two years after the GiveWell Labs → Open Philanthropy transition, which means the organization was being seriously considered and researched even earlier.
If GiveWell of 1800 could pay to buy permanent freedom for slaves, it’s not crazy to think they’d have done quite a lot of that. (And perhaps advocated for abolition or funded the Underground Railroad; they seriously considered multiple U.S.-based “systemic” causes early on, all of which ran into the problem that it’s really hard to do as much good for people here as for people in the developing world with similar amounts of money. This picture looks different when there’s a huge slave population in your country, and clear measures can be taken to free them.)
*****
Now I’m fascinated by this question, so here’s an ugly BOTEC (based on the year 1850, since I wasn’t quickly able to find good data for 1800):
The average price of a slave in 1850 was $400, or $11,300 in 2009 dollars (roughly $14,000 in 2021).
GiveWell got ~$114 million in donations in 2020.
If GiveWell allocated 1⁄4 of that to buying slaves and setting them free (~$28 million), they could have done so for 2,000 people in a single year. With similar spending every year from 1850 to 1860, maybe 20,000 people total?
The price of slaves increased during that period, but GiveWell’s ability to move money would also have increased, if we assume a similar trajectory for the org as what we see today.
That would be ~0.5% of all slaves in the country. It’s a small number, but would spending the 1850 equivalent of $250 million over the course of a decade have advanced the cause of abolition by more than a tiny amount?
I don’t know how much money the most prominent abolitionist philanthropists gave, nor how long it would have taken them to actually achieve abolition without the Civil War.
One comparison might be the criminal justice reform movement; how much money has been spent by organizations in that space, for a relatively modest (albeit promising!) decline in total incarceration?
Of course, this is an ugly BOTEC, and there are many ways one could object. A few that come to mind:
Buying slaves and then transporting them back to the North + setting them up with the beginnings of a life would add a lot of cost.
People were much poorer in those days; $400 in 1850, or $14,000 in 2021, was several times the average income. GiveWell’s actual donations likely wouldn’t have been anywhere close to the 1850 equivalent of $100 million or more per year, unless they had enormous amounts of support from many of the country’s wealthiest people (not counting slaveowners).
The most notable philanthropist (I think?) of the age, George Peabody, donated nearly $200 million over his lifetime in 2021 dollars. That’s much, much less than Good Ventures has provided/will provide to GiveWell.
If international donations were at all feasible, it seems likely that preventing a death would still have cost much less than $14,000 in 2021 dollars; maybe GiveWell still would have focused on preventative healthcare. And given that the Civil War wound up creating abolition without the help of philanthropists, maybe they’d have been right to do so?
Heck, scrap the “international” part; even the United States had a major malaria problem at the time.
Or maybe there’s a reason that philanthropic support for abolition was important to the Civil War starting in the first place? I don’t know much about history and I’m sure this analysis is laughable on a few other dimensions. I just wanted to run the dumb numbers.
I think the systemic change point and the treatise on human nature points are meaningfully different. One presumes that “we” (leftist society) knows how to do good and EA is empirically mistaken, while the latter is saying that we’re lost on how to do good but having smart people explore their inclinations is plausibly a better path on getting there. Just addressing the latter point for now:
I find Hume a bad example from Collison, since empirically EA has a lot of philosophers and interest in psychology/philosophy, and “understanding human nature or we can better make decisions” feels right up the alley of EAs and people adjacent to us (eg rationality community).
If I wanted to make the point that EAs in history would have been insufficiently exploratory, I would’ve pointed to Newton instead of Hume. Newton ~ spent his life doing 4 things : astronomy/physics/calculus, Bible studies, alchemy, and managing the British banking system.
Arguably an EA I/N/T framework at the time would have said (given empirical beliefs at the time) that any of the latter 3 would be a better use of time than staring at stars and understanding how they move across the sky. And of course these days Newton is famous mainly as the inventor of calculus.
So I’d be more interested in whether EA would have stunted Isaac Newton’s intellectual development, more than Hume.