(I agree with many points in this answer. But for communication succinctness and because this answer is highly upvoted, I will only point out my disagreements)
In general people think the community as a whole should donate 1-10% per year, so I’d suggest, as a starting point, you could pick a percentage in that range to donate each year.
There are lots of complications[...].But I think a good prior is to donate something close to the optimal percentage for the community as a whole
I don’t share this intuition at all fwiw, because of two considerations:
1) From purely an optimal investment vs philanthropy perspective, it seems like the prior should very strongly be to either donate everything if the community’s not donating enough, or donating 0 if the community donates too much. As an individual donor, your decision-making should involve thinking on the margin, and it will be surprising if 1 million just happens to be enough to swing the community from donating too little to donating too much, or donating too much to donating too little. [1]
2) A lot of these investment models are too conservative in their modeling of cash flows, because most people have incomes. In Ben’s case, I don’t know him or his life, but if he has a typical profile of an EA “unexpectedly given ~1M in stock”, I think it is very likely that he has >1M in net present value of future earnings.
Now in practice, I happen to agree that there’s a strong case against donating much this year because of “complications” like learning value and radical worldview shifts, especially for new EAs, but my intuitions come almost entirely from these complications and are not, to a first approximation, driven by macro-level movement considerations like optimal philanthropy. And I think it’s good for reasoning transparency for us to actually say our true reasons for believing things.
[1] Implicitly assuming Causal Decision Theory (CDT) here. I think your decision criteria makes more sense if you’re assuming FDT with tightly correlated decision makers, but “tightly correlated” is an additional assumption. (I don’t understand EDT too well but I think it ought to also support marginal analyses here as well)
2) I agree you should consider your future income, the percentage should be calculated as a percentage of current assets + NPV of future income.
1) I agree the approach of “work out if the community is above or below the optimum level of investing vs. saving, and then either donate everything, or save everything” makes a lot of sense for small donors. I’d feel pretty happy if someone wanted to try do that. (Another factor is that it could be a good division of labour for some to specialise in giving soon and some specialise in investing.)
But I feel a bit unsure about recommending it.
One issue is that it assumes you’re pretty coordinated with the rest of the community, which might not be true, especially of new people.
It’s very hard to know what the optimal level is for the community as a whole should be, and if the person gets it wrong, then they might donate everything, which is irreversible. The advice of “give about 4% per year” seems less likely to be drastically wrong for them.
From a community pov, ‘everyone donate what they think they ideal percentage should be’ vs. ‘donate all or nothing depending on your best guess at whether above or below the optimum’ would both end up converging to the ideal percentage donated.
Donating more gradually means you can take advantage of learning (and as you say, radical worldview shifts).
It also lets you self-insure if your personal circumstances end up worse than you expected.
PS A third approach would be more bottom up: do a search for the best thing you can donate to right now, then compare it to your best investment, and try to think about which is better.
Thanks so much for the fast and detailed response!
Sorry I should be clearer. Ben Wilder probably shouldn’t donate all of his money this year. EA is a marathon, not a sprint, and it’s good to take advantage of (eg) learning more and worldview shifts. I just don’t think optimal philanthropic timing is the main consideration here.
PS A third approach would be more bottom up: do a search for the best thing you can donate to right now, then compare it to your best investment, and try to think about which is better.
Agreed, I think there’s a lot to recommend for this approach, including but not limited to if this search works out well, relevant information can be passed along to the rest of this community.
(I agree with many points in this answer. But for communication succinctness and because this answer is highly upvoted, I will only point out my disagreements)
I don’t share this intuition at all fwiw, because of two considerations:
1) From purely an optimal investment vs philanthropy perspective, it seems like the prior should very strongly be to either donate everything if the community’s not donating enough, or donating 0 if the community donates too much. As an individual donor, your decision-making should involve thinking on the margin, and it will be surprising if 1 million just happens to be enough to swing the community from donating too little to donating too much, or donating too much to donating too little. [1]
2) A lot of these investment models are too conservative in their modeling of cash flows, because most people have incomes. In Ben’s case, I don’t know him or his life, but if he has a typical profile of an EA “unexpectedly given ~1M in stock”, I think it is very likely that he has >1M in net present value of future earnings.
(I have the same issue with using Kelly ).
Now in practice, I happen to agree that there’s a strong case against donating much this year because of “complications” like learning value and radical worldview shifts, especially for new EAs, but my intuitions come almost entirely from these complications and are not, to a first approximation, driven by macro-level movement considerations like optimal philanthropy. And I think it’s good for reasoning transparency for us to actually say our true reasons for believing things.
[1] Implicitly assuming Causal Decision Theory (CDT) here. I think your decision criteria makes more sense if you’re assuming FDT with tightly correlated decision makers, but “tightly correlated” is an additional assumption. (I don’t understand EDT too well but I think it ought to also support marginal analyses here as well)
2) I agree you should consider your future income, the percentage should be calculated as a percentage of current assets + NPV of future income.
1) I agree the approach of “work out if the community is above or below the optimum level of investing vs. saving, and then either donate everything, or save everything” makes a lot of sense for small donors. I’d feel pretty happy if someone wanted to try do that. (Another factor is that it could be a good division of labour for some to specialise in giving soon and some specialise in investing.)
But I feel a bit unsure about recommending it.
One issue is that it assumes you’re pretty coordinated with the rest of the community, which might not be true, especially of new people.
It’s very hard to know what the optimal level is for the community as a whole should be, and if the person gets it wrong, then they might donate everything, which is irreversible. The advice of “give about 4% per year” seems less likely to be drastically wrong for them.
From a community pov, ‘everyone donate what they think they ideal percentage should be’ vs. ‘donate all or nothing depending on your best guess at whether above or below the optimum’ would both end up converging to the ideal percentage donated.
Donating more gradually means you can take advantage of learning (and as you say, radical worldview shifts).
It also lets you self-insure if your personal circumstances end up worse than you expected.
PS A third approach would be more bottom up: do a search for the best thing you can donate to right now, then compare it to your best investment, and try to think about which is better.
Thanks so much for the fast and detailed response!
Sorry I should be clearer. Ben Wilder probably shouldn’t donate all of his money this year. EA is a marathon, not a sprint, and it’s good to take advantage of (eg) learning more and worldview shifts. I just don’t think optimal philanthropic timing is the main consideration here.
Agreed, I think there’s a lot to recommend for this approach, including but not limited to if this search works out well, relevant information can be passed along to the rest of this community.