Well, it's been a crazy calendar year in any number of ways...and here at the end of it, I have a few commitments to uphold. I remain committed to donating at least 10% of my income to the organizations that I think best make the universe a better place, and to talking about it on this blog. Here are my thoughts at the end of 2018.
While I've recently been conducting some independent research into investment strategies for effective altruists (results forthcoming), I haven't been particularly active in producing my own independent opinions on the effectiveness or value of organizations. So, as in 2017, my perspective here is primarily a synthesis of a raft of conversations I've had with a (uncredited) gaggle of friends and friends-of-friends.
disclaimers: I've made no particular attempt to be discriminating or fair in these conversations. Some of the friends who have helped me form my opinions here are involved in some capacity in the areas or organizations I'll mention. Some have their own positions on advisory or evaluator boards, or publish their own opinions separately.
Rather than get into the weeds of these conflicts, I'll just advise you to keep your brain engaged throughout. Not all of my reasoning that shapes these opinions was appropriate for this post, and not all of it will be covered here. I've erred towards providing more unexplained information, rather than restricting myself to what I can explain fully here.
I cover logistics (1a), donor lotteries (1b), my general approach to non-lottery donations (1c), the specific charities I’m supporting this year (2a, summary), and further reading in the form of evaluators’ reports and personal writeups (3).
I'll also drop links to further reading into individual sections, where topical. (In the interest of including more links, I'm linking them without explicit comment in many cases.)
Last year, I wrote about donor-advised funds; this year I'm donating a mix of long-term appreciated assets and cash to my DAF, for granting early in 2019. DAFs are a really useful tool for mitigating the end-of-year crunch, and for making it easier to donate things like stocks, mutual funds, and other non-dollar things. One or two organizations aren't supported by Vanguard Charitable (my DAF manager), and I'll donate cash directly.
While deductions are (I think) marginally more valuable to me in tax year 2018 than 2019 due to my part-year residency in New York, I'm not in as much of a rush to donate as I was in 2017, and am setting aside some funds that I ultimately intend to donate, to experiment with both tax-efficient investment-for-altruism and with donating opportunistically during the year. These aren't addressed below.
By far the largest single allocation I'm making from my 2018 donation pool is to entering a donor lottery run by the Centre for Effective Altruism. The basic motivation for donor lotteries is, I think, best expressed in this blog post by Ben Hoffman:
Let’s say that your charity budget for this year is $5,000, and your best guess is that it will take about five hours of research to make a satisfactory giving decision. You expect that you’ll be giving to charities for which $5,000 is a small amount, so that they have roughly constant returns to scale with respect to your donation... In particular, for the sake of simplicity, let’s say that you think that the best charity you’re likely to find can add a healthy year to someone’s life for $250, so your donation can buy 20 life-years.
Under these circumstances, suppose that someone you trust offers you a bet with a 90% probability of getting nothing, and a 10% probability of getting back ten times what you put in. In this case, if you make a $5,000 bet, your expected giving is 10% * 10 * $5,000 = $5,000, the same as before. And if you expect the same impact per dollar up to $50,000, then if you win, your donation saves $50,000 / $250 = 200 life-years for beneficiaries of this charity. Since you only have a 10% chance of winning, your expected impact is 20 life-years, same as before.
But you only need to spend time evaluating charities if you win, so your expected time expenditure is 10% * 5 = 0.5 hours. This is strictly better – you have the same expected impact, for a tenth the expected research time...
Of course, if you’re giving away $50,000, you might be motivated to spend more than five hours on this. Let’s say that you think that you can find a charity that’s 10% more effective if you spend ten hours on it. Then in the winning scenario, you’re spending an extra five hours to save an extra 20 lives, not a bad deal. Your expected lives saved is then 22, higher than in the original case, and your expected time allocation is 1 hour, still much less than before. (...)
After all of the thought that I've applied to this problem so far this year, I don't feel like I have much better marginal uses for donations than delegating my decision to evaluators (like GiveWell or ACE) or CEA's EA Funds. At that point, I have little problem squeezing my expected dollars into a few worlds where I can give the matter more thought -- or at worst, donate to selected EA Funds anyway.
I also see positive value in credibly signalling my support for donor lotteries as a mechanism, and in facilitating the existence of donor lotteries themselves, which I expect to improve the expected effectiveness of participants' donations. CEA has more to say about donor lotteries at their main info page.
more: Adam Gleave, winner of a $100k donor lottery drawn in early 2018, wrote up his allocation decision. I am encouraged by the depth of his report, and updated weakly in favor of donor lotteries after reading it (and after writing most of the above section).
The funds that I'm not directing to my donor-lottery entry I'm allocating along lines broadly similar to 2017 -- in the face of deep uncertainty over how to weigh different kinds of moral and epistemic beliefs, I'm splitting my donation into differently-sized chunks that reflect something of my thinking about on a few at-least-plausible ways of doing good.
I consider this division primarily an exercise in practicing thinking for myself about charitable giving on a variety of margins, not primarily an exercise in aiming to do the absolute most expected good. (That's what the donor-lottery entry is for.) I readily acknowledge that my process here rests on provably-inconsistent heuristics and ad-hoc invented 'principles'. I haven't made extraordinary effort to eradicate either, and I don't expect my current thinking to be the final word so much as a jumping-off point for discussion and reflection going forward.
Like last year, I want to intentionally skew my efforts towards cashing out to "straightforward charity", rather than recursive investments, stochastic gambles, far-future moonshots, and other ungrounded bets with weak feedback loops -- even in cases where I’m not convinced that more direct, or more immediate, interventions are the most effective ways of doing good possible right now. (My intention here is to provide some kind of outside-view backstop against some kinds of decision-making failure modes.)
Practically speaking, I'm directing a substantial minority of my total 2018 donation pool to fighting global poverty and supporting non-human animal welfare, aiming to do good today, unconditionally, in straightforwardly knowable ways. (I'm making these donations outright, rather entering them into a donor lottery.)
I expect to direct any lottery sum towards efforts to make lives better in the far future or 'meta' opportunities to strengthen the community of effective altruists -- and in the face of my present uncertainty of the best way to do so, I'm donating comparatively less to those causes now, and will consider them in depth later if necessary.
But I want to continue a practice of directing some support to organizations who I believe are doing good work, and I'll lay those donations out in the remainder of this post. For the purposes of this post, I'll characterize the size of each donation as 'substantial', 'large', 'medium', and 'small'. These qualifiers are roughly comparable to my uses of the same words in 2017.
Substantial donation to GiveWell (unrestricted funds), fighting global poverty and disease to help humans living today.
My position here is little changed from 2017 (or indeed from 2014); I remain convinced that, among organizations working directly to make life better for humans living today, GiveWell’s top charities are some of the best-researched, highest-impact “sure bets”.
I note that this year, for the first time, GiveWell's list of top charities explicitly distinguishes between programs whose primary benefit is reducing deaths and programs that aim to increase recipients’ incomes and consumption. After reading GiveWell staff members' commentary on their personal donations for giving season 2018, I expect that their internal discussions on the moral tradeoffs between reducing deaths and increasing incomes to be deeper and more thoughtful that my current opinions, and aim to defer to the judgment of GiveWell staff on this moral point, at least for now.
As in the past two years, I feel best making my donation to GiveWell with no restriction on its use, as I trust GiveWell’s discretion to support its own operations where necessary and regrant excess funds to a well-considered balance between opportunities to save lives and opportunities to explicitly better them.
Substantial donation to The Humane League, advocating for non-human animal welfare.
I remain particularly uninformed about the landscape of effective animal charities -- though in the past year I have found no compelling evidence against THL or for any other effective animal charity, and I remain impressed by THL on the same grounds I was at the end of 2017.
I asked myself what I could do for <10 minutes that was most likely to change my mind, and I came up with this: I read Animal Charity Evaluators' reviews (1 2 3 4) of their 2018 top four charities, and tried to be generally on the lookout for anything that struck my systems-oriented intuition as particularly positive or negative. In the end, I was most impressed by ACE's view of THL's strengths:
In our view, THL’s most significant advantage is not any single program, but rather their general approach to advocacy. Among animal advocacy organizations, THL makes particularly strong efforts to assess their own programs and to look for and test ways to improve them. Their success in their corporate campaigns, and the publication of their research through Humane League Labs (HLL), has shifted the outlook and programming of several other advocacy organizations toward finding the best ways to advocate for animals.
THL’s organizational structure appears to be strong, with a cohesive and democratic culture promoting positive relationships between THL staff, board members, and volunteers. We think this is especially important for THL because part of the intention of their local offices is to build a grassroots movement, and setting a positive and results-oriented tone for those new to the movement is good for animal advocacy as a whole... (...)
I'm also directing 10% of my total animal-welfare donation to Animal Charity Evaluators, to support their work that helped lead me to my decision. Unlike last year, I'm not able to find a public indication that they have little room for funding in the immediate term, and I want to continue a practice of supporting evaluators I trust (as I did for GiveWell before I began donating unrestricted funds to them instead).
Large donation to the Machine Intelligence Research Institute, conducting basic research on intelligence alignment to make better lives for humans in the future.
Small donation to the Future of Humanity Institute, focusing on political challenges arising from transformative AI to make better lives for humans in the future.
I’m far from certain, but my best guess for the most important problem facing us today is lowering the probability that powerful artificial intelligence agents with goals misaligned with human flourishing are developed (or allowed to develop). I want to continue my support of MIRI's contributions to a foundational understanding of intelligence, and of FHI's Governance of AI Program.
I considered an allocation to EA Funds' Far Future Fund, but I value the potential to instead consider the space of far-future donation opportunities first-hand, in the event I'm selected to allocate donor-lottery funds.
A number of friends I talked to mentioned the opportunity for large donors to donate to a major organization aiming to develop General AI systems, with the goal of influencing them favorably towards issues of AI safety and alignment. For donors on my scale, I find this a particularly attractive case for donor lotteries, as donor-lottery winners can build deeper communication channels and better exert influence than can a group of smaller donors.
nb: In mid-2018, I directed a small-for-2017 donation to the Road to AI Safety Excellence, when they were first seeking funding. I'm not supporting them further at this time.
more: MIRI's (2017) technical agenda: [pdf]
more: Ben Henry @ EA Forum | 2018 AI Alignment Literature Review and Charity Comparison -- long and detailed review of MIRI, FHI, CHAI, CSER, GPI, and 13 other organizations.
(I read this after I'd written most of this section, and at publication time I'm still weighing it against my prior opinions. I think it's weakly disfavored that my 2018 non-lottery donations will shift substantially from what I've written here.)
Large donation to 80,000 Hours, supporting efforts to grow the effective altruism community.
Medium donation to the Centre for Effective Altruism, providing continuing support to the effective altruism community.
My thinking about organizations in this section, more than any other, relies primarily on first- and second-hand subjective judgments of the track records and principals of the organizations involved.
In general terms, I'm impressed by instances where both of these organizations stepped up to execute an on-reflection-promising approach (that no one else was pursuing) to increasing the impact of aspiring effective altruists, or helping existing members of the community to think better about making the world better. In some cases I have positive subjective opinions about their principals or donors I may be funging donations with; in others I have relatively little. I'm happy to talk about these things more privately, but don't really wish to do so here in public.
nb: I was very excited specifically about CEA's leadership at the end of 2017; after some amount of personnel turnover in the past year, I now have little private information about their leadership in particular, and am primarily weighing their track record -- which I feel warrants the support I'm maintaining.
I expect that if I were to direct more to this 'meta' bucket from my current state of knowledge, I would donate to EA Funds' Effective Altruism Meta Fund, which has in fact donated to both of the above organizations in the past, as well as others that I considered positively but decided not to direct non-lottery donations to.
Summary (and direct links to donation pages):
- Small donation to the Future of Humanity Institute, focusing on political challenges arising from transformative AI.
- Medium donation to the Centre for Effective Altruism, providing continuing support to the effective altruism community.
- Large donation to the Machine Intelligence Research Institute, supporting basic research on intelligence alignment to make better lives for humans in the future.
- Large donation to 80,000 Hours, supporting efforts to grow the effective altruism community.
- Substantial donation to The Humane League, advocating for non-human animal welfare; a further 11% to Animal Charity Evaluators.
- Substantial donation to GiveWell (unrestricted funds; likely for operational expenses and/or regranting to Malaria Consortium for work on seasonal malaria chemoprevention), fighting global poverty and disease to help humans living today.
- Primary allocation to entering a donor lottery run by CEA.
Charity evaluators’ reports:
- GiveWell | Our top charities for giving season 2018
- Animal Charity Evaluators | Announcing our 2018 charity recommendations
- OpenPhil | Suggestions for Individual Donors from Open Philanthropy Project Staff - 2018
Personal writeups, roughly in the order I saw them:
- GiveWell | Staff members’ personal donations for giving season 2018
- Ben Henry @ EA Forum | 2018 AI Alignment Literature Review and Charity Comparison
- Adam Gleave @ EA Forum | 2017 Donor Lottery Report
- Ben.Kuhn | My 2018 donations
- Thing of Things | Donation Post: 2018
- (I'll try to update this space as more come in.)