In light of Peter Singer’s TED talk and Dylan Matthews’s piece on “earning to give,” there’s been a fair amount of discussion recently of what one might call “Peter Singer’s challenge,” which I’d roughly summarize as follows:
- By giving $X to the right charity, you can save a human life.
- This fact has multiple surprising consequences, such as (a) you morally ought to give as much as possible (b) a reasonable path to doing as much good as possible is to pick a maximally high-paying job, to facilitate giving more to charity.
A common response to this reasoning – which one can see in Felix Salmon’s recent post – is to attack the first bullet point. This means disputing the robustness of the “$X saves a life” figure (a figure that is often quoted based on GiveWell’s analysis), and questioning the quantification exercise that generates this figure as being distortive and costly.
We believe that these objections to quantification have serious merit, and in fact we have produced a great deal of content that supports such objections. GiveWell is about giving as well as possible, not specifically about quantifying the expected value of donations. This distinction has become increasingly important to us since the start of our project, and we’ve continually moved in the direction of making our evaluations more holistic. (Some details on how we’ve done so below.)
But we also believe that these objections miss the real heart of Peter Singer’s challenge. In many ways we think that Peter and others do their own argument a disservice when they rely on the “$X saves a life” figure: such a figure is both open to reasonable attack and unnecessary to make the core point.
To us, the strongest form of the challenge is not “How much should I give when $X saves a life?” but “How much should I give, knowing that I have massive wealth compared to the global poor?” Perhaps the most vivid illustration comes not from Against Malaria Foundation (our #1-rated charity) but from GiveDirectly (our #2). If you give $1000 to GiveDirectly, ~$900 will end up in the hands of people whose resources are a tiny fraction of yours. GiveDirectly’s estimate – which we believe is less sensitive to guesswork than “cost per life saved” figures – is that recipients live on ~65 cents per day, implying that such a donation could roughly double the annual consumption for a family of four, not counting any long term benefits. We may not know exactly how many lives that saves, if any, but we find it a compelling figure nonetheless, and one that calls for far more generous giving than what’s “normal.”
Those figures aren’t precise, and we believe our #1 charity accomplishes even more good per dollar, but we believe the broad point to be quite robust: whether or not the money I spend on luxuries could have literally saved a life, it’s money that could do a lot more for someone else than it does for me. Jason Trigg’s attitude is, in my view, defensible based on this consideration alone.
This version of Peter Singer’s challenge relies not on the fragile estimates GiveWell produces, but on an extremely robust and nearly undisputed set of observations about extraordinary global inequalities. And it challenges us to give not just money, but time, thought, and whatever else we can spare.
We believe strongly in the value of healthy skepticism toward charities and toward cost-effectiveness estimates. What we don’t believe in is using such skepticism as an excuse to dodge questions about the appropriate level of generosity. We fear that Peter Singer and his advocates sometimes enable this dodge by relying so heavily on “cost per life saved” type figures.
The global distribution of wealth is mind-bogglingly uneven, and the readers of this blog are mostly on the privileged side of the divide. We have the informational and technological tools to help others enormously just by writing checks. These are facts that are hard to dispute, and they’re facts that raise some uncomfortable questions about how we should manage our lives and our budgets. We welcome (and instigate) debates over both our methodology and our particular recommendations, but such debates shouldn’t distract us from the moral case for giving.
Some notes on GiveWell’s relationship to “quantified giving”
We think it’s worth addressing some of the specific objections that Felix Salmon gave to the methodology of “quantified giving,” because in many cases we feel that we have not only acknowledged such objections but have put substantial work into fleshing them out, supporting them, and embracing their consequences. Specifically:
- We’ve gone well beyond listing caveats to our cost-effectiveness figures and publishing the full details of our analysis (though we have done this). We’ve put extensive work into formalizing some of the arguments against excessive formality, and illustrating in unusually detailed ways just how badly one can be led astray by model uncertainty (see posts here and here). We’ve argued that the project of giving primarily based on cost-effectiveness is potentially unsalvageable, even while defending the deeper principle behind it (maximizing good accomplished per dollar).
- Accordingly, we’ve recommended charities based on an increasingly holistic approach. Our most recent top charities announcement ranked our #2 and #3 charities in the reverse order that pure face-value cost-effectiveness analysis would suggest, because of issues around upside, accountability and confidence in the organizations.
- We have also acknowledged and embraced the idea that our current research is excessively focused on the measurable and quantifiable. For the past several years, we have been working on GiveWell Labs, which seeks to broaden our scope to encompass all philanthropic options, not just “direct delivery,” while understanding that many such options will require fundamentally different evaluation approaches (some discussion of the latter point can be seen at a transcript of our recent event in New York). This year, we have declared our intention to spend more time on GiveWell Labs than on our traditional research.
We believe that “systematically examining all options with the aim of doing as much good as possible, and being highly transparent about our reasoning” is often conflated with “making decisions based on explicit quantifications of good accomplished.” As long as the two are held equivalent, the project of “effective altruism” will be on shaky ground. But we believe the two are not equivalent – that it is possible to be simultaneously holistic, systematic and transparent. We will be writing more about the distinction.