Edit: Here is the talk
More news: Giles created a new #Moraltrade channel on Slack, which is being used during the Sunday US UK online EA workathons - in person for those in SF. To see the workathon Sunday event go here. And to enter the channel for Moral Trade enter here.
In these writings we propose the creation of a sub-field of knowledge, Moral Economics.
In this post I will make a detailed case for direct donation being more efficient than more common methods of funding EA work, and cite many other benefits on top of efficiency. I believe we can and should address its underuse.
Special Thanks to Eli Tyre, Dave Dekenberger, Matt Reyes, Giles Edkins and Ben Hoffman for edits and comments.
Moral economics series
-
Direct Funding Between EAs
-
Certificates of Impact, Doing It Right - Giles Edkins
-
Moral market failure: how COIs might help
-
Problems with COIs, and their solutions
-
Implementing COIs
-
-
Agential Identity in Moral Economics
Direct Funding Between EAs - Moral Economics
We will examine the current options that transform resources into labor within effective altruism, with a special emphasis in the least well-known idea of Direct Funding Between EAs, that is direct donations between EAs.
Previously we described:
Direct Funding Between EAs: A direct linkage between an individual who wishes to have their values promoted and another who attempts to promote these values in exchange for money.
This has been attempted within the Effective Altruist community by creating Gratipay and Patreon donation portals for EAs.
The Status Quo
Currently there are at least four main forms of monetary exchange between Effective Altruists
Institutional Donations: The vast majority of donations are going to institutions, this includes both recommended charities and institutions familiar to the EA movement such as CFAR, FHI, MIRI, CEA, Leverage etc…
Certificates of Impact: A very small fraction of resources is going to certificates of impact, which represent altruistic valuable actions taken in the past, and can be purchased, in which case the buyer becomes responsible for the good deeds purchased.
Direct Requests for Events: Some events like the HPMOR parties were arranged directly by asking donors to finance the event.
Direct Funding for EAs: A very small fraction of resources is being transferred via Gratipay and Patreon between EAs.
The initiative EA Ventures proposes a model of resource transfer to specific projects which combines some of the above, to facilitate connecting projects to potential funders.
Direct Funding - Details and Advantages
There are several important differences between direct donation, hiring someone in an Effective Altruist institution, and purchasing people’s certificates of impact. Direct donations are based on trusting what the person has done before, and what she claims she will do now. They are more stable than certificates - which are as stable as bounty hunting - and thus could be the entire source of income for individuals funded by donors who share their values, instead of institutional donations. Direct donations are also much cheaper per worker hour than institutional donations. As a back of the envelope calculation I estimated the difference as about threefold, and even after factoring 501c, charity status, Patreon’s fee, and all other discounts, it remains two times more expensive to donate to institutions than directly.
The current climate within the Effective Altruism community is one of skepticism of direct funding which receives a small fraction of resources, skepticism of certificates of impact, which also receive a small fraction of resources, and profound trust in institutions. This trust at a recent EA Think Tank meeting was called “sanctification” of institutions - which receive the vast majority of financial and volunteering resources from Effective Altruist donors. This is unjustified and could be taken as a temporary failure in the moral market. I expect this trend to change over time as the ideas of direct funding and currencies such as certificates of impact become more integrated into the strategic thinking of individual donors, and the snowball of increasing trust in long-term committed individuals generates a self-reinforcing incentive.
Direct donations would be a great addition to the moral market, since they offer many advantages, some of which are simply unavailable to institutions or certificates of impact. Let us look at some of these advantages:
Direct funding permits investigation of neglected causes: It requires merely two individuals for a cause to be investigated. If, for example, a single donor considers an investigation on the disruptive potential of Emulation Economics to the global economy, and one individual wants to undertake this research, this suffices for at least some attention to be allocated to that problem. Even causes neglected within EA can be evaluated.
Direct funding is stable over time: Certificates of impact pay backward in time. An individual cannot anticipate to receive them beforehand because other projects may be judged more valuable. Direct donations done monthly could guarantee security for EAs that intend to zig-zag between institutions, researchers who want to research a specific cause, and other helpers who would undertake a function that isn’t defined by a job-role. Direct donations are a contract of trust between donor and donee, and the donee could establish ways for the donors to keep track of the work they are paying for, such as by sending periodic reports or having periodic meetings.
Direct funding permits unusual work: Performance is hard to measure. Individuals who travelled between EA institutions and influenced the EA movement via hard to assess means (e.g. Michael Vassar, Justin Shoevelain, Jasen Murray, Scott Aaronson...) are frequently volunteering their time and effort to EA causes. Some of these individuals have built a reputation of doing great work as EAs though they were always behind different projects. In theory, institutions could hire such people, but in practice, direct donations would be a much more efficient way of incentivizing more people to be EA generalists. Once a thread of trust is built, these EAs could have freedom to work in what seems more appropriate to them, as long as their backers still thought they were on the right track. More importantly, many more people could begin to work as EA generalists who currently don’t even consider that possibility, the labour market would become more fluid.
Direct funding prevents groupthink: Institutions need to have their focus well understood, and require some level of internal goal alignment. Bold new different ideas of what needs to be done by EAs, or what needs to be researched, or new methods of research won’t come easily out of institutions - this is less of a problem within the EA community than outside it, but the problem is not completely eliminated. Direct funding would attenuate it even further; If a person with a good track record of EA ideas in the past wants to pursue a very different project, their direct donors could guarantee their financial stability without jeopardizing either the person’s action flexibility, nor the reputation of the institution to which the person works for, since there would be none.
Direct funding is instantaneous: Hiring an individual can take a long time, especially a foreigner. Direct donations can be done as soon as there is an agreement on conditions and trust between donor and donee.
Direct funding facilitates international EA work: The reasons here are clear. It is bureaucratically less costly to transfer money internationally than to do it via institutions.
Direct funding makes moving to EA hubs easier: Many EAs move to locations with more EAs. It is much easier to do this if there are stable income sources that are region independent and provide a fallback during the period of transition. More importantly, donations are considered passive income, and as far as I know donations within a bounded range can be received by non-US citizens in US soil, even in non-favorable Visa conditions. Similar constraints are relaxed for donations to other countries as well. To move to Oxford, Berkeley, Switzerland or Melbourne would be much easier at the moment for EAs who are receiving the majority or entirety of income through direct donations.
Direct funding alters the mindset of the recipient: If you are hired by an EA institution to do a particular set of tasks, you are being paid to do those tasks, and help that organization do well in its function. If you are directly being paid to be an EA however, this is much more ingrained in your identity. There are people paying you to do the most good you can do. If you see an opportunity for doing good, you will be more likely to take it, since you are never done with the obligation of being a great EA.
Direct funding can be the entire source of income of an individual: This seems to be the untapped secret of direct funding. If someone wants to do EA work all the time, they can be fully funded by donors so as to not need any institutional affiliation, or have a token one. This threshold will be what determines the ultimate value of direct donations, since the point at which there are sufficient incentives for individuals to try doing this as their main income source will be the tipping point for most of the advantages above to manifest in full. When people can confidently undertake an EA career or a paid year on the basis of donation, then the labor market will have the necessary fluidity to accrue all of the advantages that can be had via direct funding.
Even though they are important all these advantages and details are completely dominated by the crucial factor when it comes to effective donation.
Direct Funding is Very Cheap
Institutions are made of people. When you pay an institution to do something, you are ultimately paying the people participating in it to do something. If you are an effective altruist donating to an EA charity, then you are paying someone because you want them to do something. This thing is what has value for you. Having institutions and charity status are very valuable for the EA movement, and it is one of the ways in which it causes a lot of people to do a lot of things.
The question then is, is there a niche or a space for causing people to do what needs to be done also outside institutional boundaries?
The response is a resounding yes!
Why would we want to complement institutional roles with individual roles? Basically because individuals are cheap. It costs a lot in donations to hire a new worker at an institution, whereas there are many EAs who would (and do) work for much less money. Individuals who work in EA institutions frequently cost 75.000USD or more per year to their donors. Some EAs on the other hand would be happy and willing to work on their project for a third of that, 25.000USD. In fact for a couple years at least some EA organization paid less than that. The difference in cost to donors is striking. Even after accounting for the fees that Patreon and Gratipay charges (~3%), tax exemptions for charities and other advantages, the difference is still very substantial. Institutions have to bear many costs that individuals don’t have to, and if I recall correctly, direct donations are not subject to income tax on the donee’s part, which makes it even cheaper for the donors. Sometimes institutions also have to pay according to a table that determines how much people in different jobs should make, irrespective of whether both parties would prefer otherwise. No such constraint exists for donations.
For an altruistic donor, there are two numbers that matter: number of dollars that they have spent on the thing they value, and number of things that got done towards that value being accomplished. The cheapest way to do so seems to me hands down to be direct donation to individuals. It is at least twice as cheap.
Why Are Institutions Sacralized?
When discussing this strange phenomenon with other EAs, someone (who works for a high paying institution, so I’ll protect their name) suggested this interesting question. It seems to me that the reason why institutions are sacralized is that it feels like an institution is much closer to a cause or goal in mental space. The association between “saving the environment” and “Greenpeace” is much stronger than that of “saving the environment” and “Al Gore”. Al Gore may have strongly advocated for environmental causes, but a person seems less like a cause than an institution does. The forest, however, is made of trees, and Greenpeace is made, in great part, of people working in environmental causes.
EAs are notorious for their ability to reason through their felt sense, and do what they think is valuable even if it doesn’t necessarily feel immediately right. EAs are probably good then at combating the felt sense that donating to institutions is donating directly to a cause, instead of for people who work on/advocate/research on a cause. Donating to institutions is paying for the people who work there, plus taxes, rent, retirement benefits, legal fees, institutional costs, and many other collaterals. Maybe 70% or even 90% of EA paid work should indeed be done through institutions and wages, but I am very skeptical that not even 10 EAs should be being paid directly to work on EA projects, research, advocacy, etc… When we consider that this could also be done to pay individuals temporarily while they rearrange Visa situations, while their legal contract is processed, or that it can be used to finance EAs who live in countries with advantageous currencies, I find it impressive that this was not chosen by anyone yet as the problem they will solve.
Creating an EA donation network is one of the most valuable things we can do. Ozzie Gooen, Patrick Brinich-Langlois, and many others have begun to do this, but more needs to be done, and more donors should consider direct donations their primary place to donate. Also, we should concentrate our donations toward EAs with good portfolios of EA work completed, until donations amount to a significant fraction of their income. Having 10 individuals receiving 1/10th of the institutional cost for 1 worker won’t give any of them the security and trust needed to pursue an EA career and EA projects, but having 3 receive 1/3rd of that instead would likely cause all three of them to feel confident in this opportunity. At least two times more work would be done. If this were life-saving work, then twice the lives would be saved.
As the unnamed person above pointed out: It is as if EA institutions have become sacred, but there is no reason for this.
Improving Direct Donations Between EAs
Patreon versus Gratipay
Ozzie - who put up the Gratipay network - and I advocate a transition to Patreon as quickly as possible, since it is much more likely to be standing in the near future, Gratipay has gone down once already and has lower transaction volume.
Trust
One issue that comes up when discussing direct donations is “how to make sure the donors trust their donees to be doing good work?” This is one more reason to reduce the number of baskets in which donors should put their eggs. It is much easier to have one or two individuals you trust doing work you value, and keep track of what they are doing, than to invest at once in many people. Especially for neglected causes, it is valuable to have a setting where only one large or two smaller donors suffice to finance someone they trust.
Feedback
It is of course up to the dyad to establish how they will determine for how long the donations continue, how much freedom the donee has to change projects, and which types of report they will give. To make this easier, donees could have a standard way to report to their donors, and make contextual adjustments when requested.
Evaluation
EA donors have especially high regard for external evaluators such as GiveWell, GWWC and Animal Charity Evaluators. If individuals are competing with organizations for funds, it is likely that donors will place a similar burden of proof on each. As such, an evaluator for directly funded individuals might become necessary (although note that explicitly EA-aligned organizations which rely on donations, such as CEA, also fall outside the scope of current external evaluators).
Guarantees
Eventually, it would also be valuable to have financial guarantees for donees being funded. If a donor decides to stop funding a person responsible for a project (say because they had a financial emergency), the person still gets funded for 1 or 2 months by a guarantee fund, and gets some extra exposition to other prospective donors, who may choose to save the project from going under and the person from financial insecurity. Even if there were a guarantee bank, the whole process would still be much cheaper than hiring people institutionally.
In the next post on Moral Economics Giles Edkins will discuss several new ideas related to Certificates of Impact, and suggestions on how to create a moral market.
Don't forget to check the ignite talk on Moral Economics if you are going to EA Global at the Googleplex on Saturday.
It's worth spelling out why you think that individuals are cheaper than institutions. Reasons I can see:
Individuals are willing to work for less. But this doesn't really speak to the opportunity cost of their time, which (from the perspective of society, or "effective altruism" as a whole) is the main cost.
Tax treatment. Most of the time this is much worse for gifts to individuals (unless you arrange to make a grant to an individual, which is quite difficult for most donors and introduces significant additional reporting overhead). It can be better if (a) donors weren't able to deduct their donations anyway (e.g. because they are in an unfavorable jurisdiction or are donating > 50% of AGI in the US), and (b) people are willing to work for very low wages outside of institutions.
Fundraising. I don't see why this would be cheaper for individual donors, though I agree that in general we could as a community reduce fundraising costs by improving our collective decision-making.
Coordination. Groups spend time coordinating with each other, on training and recruiting, etc., but in general these seem like worthwhile projects that are pursued precisely because they are cost-effective.
Administrative overhead. This seems like the main game. Some of this is useful work that improves efficacy, while some is just overhead involved with hiring people, accounting, maintaining non-profit status. It would be useful to get a more precise estimate for the total "lost" overhead that can be avoided by working independently. I think this is the main thing you could do to make this case more compelling.
It's worth pointing out that the number of dollars spent on overhead as a fraction of total dollars isn't a directly relevant metric, given that the opportunity cost of employee time is often significantly higher (so e.g. a 20% financial overhead may be more like a 5-10% overhead if we include all inputs). Between this and some overhead being useful for productivity (e.g. having an office, saving employee time, recruiting...), I mostly don't feel concerned about institutional overhead.
Strongly second this comment. As someone currently doing a lot of things that more or less amount to 'pay other people to do thing x that I think should be done', I would be very interested in understanding the most efficient way of doing so. But from this post I really don't know why individuals should be cheaper overall; charitable status is a huge factor in favour of institutions (a much larger factor in the UK than the US admittedly, but still) and is currently the only significant factor I am aware of in either direction.
One difference in cost comes from institutions such as Oxford requiring their employees to get prestigious wages. That makes the $75k average misleading. More obscure charities can hire employees much more cheaply.
To give the example of a charity that's competitive with paying individuals (and indeed obscure, at least outside EA!), the employees at Charity Science try to live on fairly minimal expenses. And we're currently spending very little time and thus money on administrative overhead; the estimate of operations time in 2015 here is generous.
This isn't to say that I'm against funding individuals at all. Indeed the boundaries are fuzzy. People can fund me directly to work on the projects I run (albeit with some money then going to people I pay to work on them when this is more efficient), but they can do that tax-deductibly by donating to Charity Science (the legal umbrella entity, rather than the core fundraising projects).
As great as that all is, isn't this largely due to the culture at Charity Science, and the culture it tries to promote? I expect it would be difficult to normalize or replicate this in existing effective altruism charities. (It seems it would be difficult enough to spread among the existing non-profit world at large, where it isn't already being practiced, that I won't bother discussing it). Not that it wouldn't be worth trying, but we can dissect the problem to get a better sense of it.
It's not necessarily that existing employees of various charities would be averse to engendering these norms, per se, but they might have a culture which already normalizes how perfectly acceptable it is to not live at a fairly minimal level, i.e., just covering basic living costs. I don't know what other organizations do or don't do this, so I won't speculate.
If Charity Science or its employees would want to change this, they can spread these values among other effective altruists or other supporters they meet, and who are enthusiastic about Charity Science. Then, when they go on to do some kind of direct work, they'll be more inclined to keep costs this low. One or more of you could also post to this forum, as its readership is amenable to (some) more personal stories. I think those values in how to run an organization or how to do the work won't stick unless some people who all shared those values founded an organization together.
Of course, Charity Science could always get more volunteers, expand to hire effective altruists doing independent work in line with its mission, or hire contractors(?).
Sure, that's totally fair. Though as I'm sure you know it doesn't affect the point about whether institutions or individuals are more efficient to fund: we're one example of people who could be funded cheaply whether as individuals or as an organisation. Admittedly an unrepresentative example, at one extreme.
That could be a good idea, though I don't know if we'll personally write this - we're happy for others to do so though! (It could be along the lines of that EA interview series you mentioned you might continue.)
I agree with Jay; most accessible essay in the series yet. Keep it up! Suggestions for building a moral market:
Encourage more potential donors to post what sort of research they'd seek to fund. If they then end up with multiple candidates coming forward, that can be resolved through a competitive hiring process.
Also encourage more potential donees to post their aspiring research agendas, to attract potential donors more easily.
Note: the prior suggestions might not be practical to enact until basic infrastructure of a direct funding moral market is set up.
Approach EA Ventures about vetting and normalizing direct funding. This could be one institution itself that doesn't receive funding from the direct funding market, which could generate trust in EA Ventures as an evaluator. It could also serve as a light version of Givewell or GWWC responsible for assessing the viability of individuals. EA Ventures could make clear in reports their assessments are their own opinions, and other donors and donees are by all means free to make their open funding arrangements.
Multiple donors could form coalitions to fund a single donee. For donors, this could decrease salient feelings of risk aversion otherwise preventing funding, because not one single donor takes on all risk. For donees, being funded by a coalition of multiple donors could increase security in two ways. Firstly, if one donor stops contributing funds, one or more funders are still around to provide funding. Hopefully, partial interim funding would be sufficient to meet living costs for the donee until full funding is restored. Anyway, secondly, funding coalitions of more members can seem more trustworthy and reliable to potential new donors. A new donor will be more likely to fill an existing funding gap if they can see one or more other donors already trust the donee enough to carry out the desired work.
Or to fund multiple donees.
At least on Patreon (which you seem to be suggesting that people use) you probably have to pay tax if you make over $600 per year. See their FAQ on taxes:
My preferred proposal would involve a conceptual division of funder responsibilities (though the same person might do both):
Provision of reliable income, to ensure that people can afford to work independently without worrying about financial issues, in exchange for some fraction of total output. I think that this fraction should probably be negotiated explicitly, as a way of helping us figure out how important money is and how reasonable it is to work independently. This step is profit-motivated, and can be split across several people.
Evaluation of outputs, and either resale of those outputs or internal recognition of altruistic gains.
[1] can be continued or eventually discontinued based on the results of [2]. The separation allows people well-positioned to evaluate work to be separate from those who have enough funding to support it. It also helps provide more explicit accountability, and to support better decisions about whether it's worth doing [1].
I and the impact purchase are potentially willing to participate in either half of this arrangement (or both), but I will continue to make evaluations which are noisy, transparent, and on the low side (and hence which are liable to be offensive).
Edit: after watching the EA Global livestream of Kerry Vaughan's talk about what he plans for the future of the effective altruism community as the director of EA Ventures, it's my impression he would be very open to being part of a partnership like the one you're proposing, Paul. I strongly encourage you to reach out to Kerry Vaughan.
I made a similar suggestion here.
Consider reaching out to EA Ventures to be the second half of the arrangement.
This is especiall useful and practical compared with other posts in your series so far - a lot of what you say about individuals and institutions is plausible.
Who would some natural candidates for receiving funding be? I'll spare you from having to name yourself and do so for you! But are there actually any others? Maybe not?
Off the top of my mind and without consulting people:
Ruairi Donnelly, Max Carpendale, Kieran Grieg, Andrew McKnight, Oscar Horta
Had he not been hired by ACE: Jacy Anthis
The Foundational Research Institute (FRI) is a project aiming to be similar to the Future of Humanity Institute, except with a more explicitly utilitarian bent, based on the concerns of, e.g,. David Pearce, Brian Tomasik, and the EAs of Basel, Switzerland. They're in an awkward position where Brian would be a valuable researcher for them, but he considers his peers to be just as valuable and adept researchers, so he's taken to funding them and their organizations himself through earning to give. I believe he may be the only person doing so for (FRI), and also perhaps Animal Ethics. This seems to me an awkward state of affairs. I don't know if FRI is even legally incorporated in any country yet, so funding them and/or affiliated researchers might blur the lines between funding individual researchers and institutions. Either way, I believe Mr. Tomasik and his peers would appreciate a chance to discuss direct funding if or when such funding became available.
Thanks, Evan! FRI is an official charity in Switzerland. Donations would help us grow, though we'd probably need a lot of funding ($100K+) before I revised my personal thoughts on earning vs. direct research.
I sent 5 euros to you! :)
What I don't see is how FRI could actually change the future. No matter how much research you publish, there's no reason to predict it will actually influence real decisions which affect real future suffering.
So it is only a theoretical activity (still fun and legitimate, of course).
Off the top of my mind and without consulting people:
Justin Shovelain, Oliver Habryka, Malcolm Ocean, Roxanne Heston, Miranda Dixon-Luinenburg, Steve Rayhawk, Gustavo Rosa, Stephen Frey, Gustavo Bicalho, Steven Kaas, Bastien Stern, Anne Wissemann, and many others.
If they had not received FLI funding: Kaj Sotala, Katja Grace.
If they needed to transition between institutions/countries: Most of the core EA community.
I have mentioned it as an option for a while, but personally waited for less conflict of interest to actually post about it (at the moment I'm receiving much less funding than I did before, and am receiving some funding through institutions, I wanted to make clear that my goal is not to have people disliking institutions, but to fix the labor market).
I think the greater point here is that this has not been considered so far because people did not even envision it being an option. But now they will.
If this market can start taking off, encourage each of those persons you mentioned to post an outline of the research they're pursuing on the Effective Altruism Forum or LessWrong or something. Then, potential funders have somewhere to browse and ask questions about the research proposal(s).
Do you want to do this soon? You can help us get it done.
Yeah, let's do it. If others are rearing to go, I'll help edit or provide feedback or share links or do whatever I can. Between so many major foundations seeming poised to give more money than the rest of the community combined to all our current favored charities across causes, but them being unlikely to directly fund individuals' research, now is an ideal time to plant in the minds of 4-to-6 figure-per-year donors new acceptable targets for their donations.
It might help people willing to become funders, or who are seeking funding, if they update the details of such aspirations in a profile on the Effective Altruism Hub.
The goal of avoiding groupthink has the potential to be a very important reason for preferring direct funding. If the direct funding ends up substituting for donations to large, entrenched institutions, then I expect it to be valuable. But I expect that any groupthink associated with young charities that have a handful of employees comes from a broader community, not the specific institution.
While I agree that funding individuals is promising, I don't find this particular argument very convincing. If one is not being paid to do particular tasks, then "being paid to be an EA" is a lot like being paid to be cool, which probably only works if one has already demonstrated the ability to perform well under similar conditions. Saying that people are paying one to "do the most good one can do" sounds a lot like a rationalization where the whims of the donee are gratuitously tolerated.
Also I'd point out that constantly thinking of EA as an obligation and "never being done" may be suboptimal.
I think you mean AFAICT - as far as I can tell. It's normally better to spell these acronyms out for writing of this sort.
While I'm commenting, does this hold in other places EAs might want to move (I suppose this means the UK, Canada, Australia, Switzerland and perhaps Thailand?)