[Written in a personal capacity, etc. This is the second of two comments, see the first here.]
In this comment, I consider how centralised EA should be. I’m less sure how to think about this. My main, tentative proposal is:
We should distinguish central functions from central control. The more central a function something has, the more decentralised control of it should be. Specifically, I suggest CEA should become a fee-paying members’ society that democratically elects its officers—much like the America Philosophical Association does.
I suspect it helps not just to ask “how centralised should EA be” but also “what should be centralised and what shouldn’t?”. Some bits are, as you say, natural monopolies in that it’s easiest if there’s one of them. This seems most true for places where people meet and communicate with each other: a conference is valuable because other relevant people are there. For EA, I guess the central bits are the conferences, the introductory materials, the forum, the name(?), maybe other things. In my post on EA as a marketplace, which you kindly reference but don’t seem sympathetic to, I point out you can think of EA on a hub-a-spoke model. Imagine a bicycle wheel, where the rim represents the community members and the spokes the connections. There are some bits we widely participate in, the hub, such as the conference. But, besides that, people have links to only a subset: longtermists hang out mostly with longtermists, etc.
Now, it does not follow that simply because something has a central function, it’s used by lots of people, that it should be centrally controlled, i.e. controlled by a few people. In fact, we often think the opposite is true. The more central something is, the more often we think it should be democratically controlled. The obvious example of this is the state. It has a big impact on our lives, it’s a natural monopoly, so we tend to think democracy is good to make it accountable. Rule of thumb, then: central role, decentralised control.
Another example of this is the one you gave, the American Philosophical Association. It’s pretty useful to have a place that convenes those who have a common interest—in that case, doing academic philosophy. The APA seems useful and unobjectionable (this is my impression of it, anyway). But, the reason for this is that it doesn’t and can’t do anything besides serve and convene its members. It doesn’t take sides in philosophical debates or try to steer the field. People would object if it tried. How is this unobjectionableness achieved? I imagine it’s to do with the fact it’s a fee-paying society where those in positions of power are elected by members, as well as the fact it doesn’t control lots of funding. Despite being philosophers, I doubt the members of the APA would want it to be run by unelected philosopher kings! Roughly, the moral of the story seems to be that, if something is so central you can’t avoid participating in, you probably want decentralised power.
You talk about various possible ways to centralise or decentralise EA. But why is there no suggestion of democratising the central element of EA, namely the Centre for Effective Altruism? Here’s a concrete suggestion: CEA becomes like the APA: a fee-paying membership society where the members elect the trustees or officials, who then administer various central functions, like a conference, journals, etc. Is this an absurd idea? If so, why? It’s hardly radical. Members’ organisations are a default coordination solution where people have a common interest. I’m not sure it’s a brilliant idea, but it’s weird it’s not been discussed, and I’d be happy for someone to tell me why would be terrible.
Worried no one would join? If you want to kick-start it, you could provide a year’s free membership to those who have signed the GWWC pledge or attended an EA conference. There may be other ideas. EAs already focus on much harder tasks like influencing the next million years and ending poverty. Surely setting up a membership society, a solved-problem, is not insurmountable.
There’s been a lot of discussion recently about people not feeling like they’re part of EA. Well, here’s a cheap solution: let people become members of CEA. Then, you’re in and you can have a say in how things are run. This has other advantages: it makes CEA accountable to its members. It also allows it to genuinely speak for them, which currently it can’t do, because it doesn’t represent them. By charging a membership fee, CEA can offset the costs of other things, so won’t be so reliant on other donors. Honestly, this membership scheme would probably work if it were just about the EA conferences (the “EA conference Association”?).
Who should be against this idea? I recognise some effective altruists are sceptical of applying democracy to philanthropy, but, to be clear, I am not advocating for communism, that all of EA or ‘EA resources’ should come under common control: that Open Philanthropy should give 1/Nth of its resources to the N people self-describing as EA, or that if you want to become part of EA you need to give the community all your (spare) money. That’s absurd. I am only in favour of democratising the central convening and coordinating parts; this is like saying the APA should be a democratic entity, not that all philosophy and philosophy funding should be run by a democracy. As far as I can see, all the central elements of EA fall under CEA (I’m also not against spinning off parts of EVF.)
I don’t think it makes sense to have a democracy for the non-central elements—the spokes. I believe in free enterprise, including free philanthropic enterprise, and private ownership.
Where does Open Philanthropy fit into this, given it has most of the ‘EA money’? I’m not sure it does fit in. It’s fortunate there’s one really big donor (because there’s more money), and unfortunate there’s only one (but that one will have outsized influence). I think society should control some of people’s income through taxation. But I also believe people should have private property, and where they spend that (assuming it’s inside the law) is best left up to them, rather than trying to also make that part subject to democratic control. Hence, insofar as worries about centralisation spring from their being a single, huge funder, I don’t have a neat solution to that. That doesn’t mean other things couldn’t be done, however, such as democratising CEA. Indeed, if CEA were a democratic society, I’d be fairly relaxed about Open Philanthropy providing funding to it, because control of CEA would still be decentralised, so that would mitigate concerns about undue influence. Of course, I have views about how, morally speaking, people should spend their post-tax wealth, but that seems a separate issue.
Finally, you raise the point of who has the ‘legitimacy’ to centralise or decentralise EA, and say it should come from CEA or Open Philanthropy (the use of ‘legitimacy’ is somewhat interesting, because that has democratic connotations). At this point, I should probably mention I applied to be a board member of Effective Ventures, and in my application form, I explicitly stated I was interested in exploring bringing democratic elements into CEA, and so decentralising it. I didn’t make it to the first stage (I also asked for feedback and was told EV couldn’t provide any). Now, I am not claiming I am a ‘slam dunk’ choice for the board—EA has many highly talented people. However, I did find it discouraging, not least because I am interested in, and would have had legitimacy for, exploring institutional reforms. It reduced my confidence that, despite what you say, ‘central EA’ really is open to diverse voices or further decentralisation.
I think it’s a mistake to conflate making things more democratic or representative and making them more decentralised—historically the introduction of more representative institutions facilitated the centralisation of states by increasing their ability to tax cities (see e.g. here). In the same way I would expect making CEA/EVF more democratic would increase centralisation by increasing their perceived legitimacy and claim to leadership.
I take it you’re saying making things more democratic can make them more powerful because they then have greater legitimacy, right? More decentralised power → large actual power?
I suppose part of my motivation to democratise CEA is that it sort of has that leadership role de facto anyway, and I don’t see that changing anytime soon (because it’s so central). Yet, it lacks legitimacy (i.e. the de jure bit), so a solution is to give it legitimacy.
I guess someone could say, “I don’t want CEA to have more power, and it would have if it were a members society, so I don’t want that to happen”. But that’s not my concern. If anything, what your comments make me think is (1) something like CEA should exist, (2) actual CEA does a pretty good job, (3) nevertheless, there’s something icky about its lack of legitimacy (maybe I’m far more of an instinctive democratic that I thought), (4) adding some democracy stuff would address (3).
I’m confused about the mathematics of a a fee-paying membership society. I’m having a hard time seeing how that would generate more than a modest fraction of current revenues.
It’s not clear what the “central convening and coordinating parts” are. Neither Current-CEA nor Reformed-CEA would have a monopoly on tasks like funding community builders, funding/running conferences, and so on. They are just another vendor who the donors can choose to hire for those purposes. There is and would be no democratic mandate that donors who would like to fund X, Y, and Z are obliged to go through CEA.
I think your model is correct insofar as the membership society could assert independent control of certain epistemically critical functions that are relatively less reliant on funding (e.g., the Forum).
The extent to which “convening and coordinating” is effective may depend on whether there is money behind those efforts. Stated more directly, to what extent are CEA’s efforts in these areas boosted by the well-known (general yet strong) alignment between CEA and the major funder in the ecosystem? Would Reformed-CEA enjoy the same boost?
I used to work at EA Norway, which is a fee-paying membership society, and thought it might be useful to share more on how our funding worked. This is just meant as an example, and not as an argument for or against membership societies. (Here’s a longer comment explaining how we organise things at EA Norway.)
I can’t speak to EA Norway’s current situation, as I no longer have any position at EA Norway (other than being a paying member). However, I can say what it was like in 2018-2021 when I was Executive Director (ED). The total income from the membership fee roughly covered the cost of the general assembly. Most of our funding came from a community building grant from the Centre for Effective Altruism (CEA). However, the board made sure to fundraise enough from private donors for my salary. The two main reasons for this was to I) diversify our funding, and II) enable us to make longer term plans than CEAs grant periods.
When the board gave approval to accept the community building grant from CEA, we discussed that if at any point we did not want to follow CEAs guidelines and success metrics, we would pay back the remainder of the grant. This was definitely easier for us to say and truly mean when we had covered the ED’s salary from other sources, as it meant that if we were to return the funding, we would still have at least one employee. We never ended up disagreeing so much with CEA that we wanted to return the funds, though we were definitely very vocal about any disagreements we had with the groups team at CEA and did push for some changes.
I’m confused about the mathematics of a a fee-paying membership society. I’m having a hard time seeing how that would generate more than a modest fraction of current revenues.
I’m confused why you think this is required, I don’t think Michael implied it would.
The society wouldn’t be a good replacement for CEA unless it could attract significant major donor support. As the next paragraph implies, there’s no reason for major donors to support the society if they judge an alternative vendor to be more effective in delivering conferences, etc. As a result, the society would either have to adapt its programs to meet the scoring metrics of the big donors (in which case the democratic nature of the organization isn’t doing much work; the money is still calling the shots) or it would lack funding to perform those functions (in which case the organization isn’t effective on those functions).
As my third paragraph suggested, there are functions the membership society could potentially run on member revenue and small donations. But that is a significant tradeoff.
Yeah, I’ve not spent loads of time trying to think through the details. I’m reluctant to do so unless there’s interest from ‘central EA’ on this.
As ubuntu’s comments elsewhere made clear, it’s quite hard for someone to replicate various existing community structures, e.g. the conferences, even though no one has a literal monopoly on them, because they are still natural monopolies. If you’re thinking “I can’t imagine a funder supporting a new version of X if X already exists”, then that’s a good sign it is a central structure (and maybe should have democratic elements). There are lots of philosophy conferences, but that doesn’t take away from the value of having a central one.
Also, you make the point “well, but would reformed-EA be worth doing if the main funder wouldn’t support it?”. Let’s leave that as an open question. But I do want to highlight a tension between that thought and the claim that “EA is not that centralised”. If how EA operates depends (very) substantially on one what a single funder thinks, we should presumably conclude EA is very centralised. Of course, it’s then a further question of whether or not that’s good and what, if anything, should be done by various individuals about it.
Yes, I think the proposal effectively highlights that EA is significantly more centralized than some claim.
My guess is that you would have to add a claim like “Funders should not fund ‘central convening and coordinating’ functions except as consistent with the community’s will” to get anywhere with your proposal as currently sketched. That’s a negative norm, less demanding than an affirmative claim to funding. But I haven’t exhaustively explored the possibilities either.
My own view is that a member-led organization is probably viable and a good idea, but has to be realistic about what functions it could assume.
Well, you’re not going to fund stuff if you don’t like what the organisation is planning to do. That’s generally true.
I don’t mind the idea of donors funding a members’ society. This happens all the time, right? It’s just the leaders have to justify it to the members. It’s also not obvious that, if CEA were a democratic society, it would counterfactually lose funding. You might gain some and lose others. I’m not sure I would personally fund ‘reformed-CEA’ but I would be more willing to do so.
I suggest CEA should become a fee-paying members’ society that democratically elects its officers—much like the America Philosophical Association does.
Okay, but the American Philosophical Association “was founded in 1900 to promote the exchange of ideas among philosophers, to encourage creative and scholarly activity in philosophy, to facilitate the professional work and teaching of philosophers, and to represent philosophy as a discipline” with a modern mission as follows ” promotes the discipline and profession of philosophy, both within the academy and in the public arena. The APA supports the professional development of philosophers at all levels and works to foster greater understanding and appreciation of the value of philosophical inquiry.” Seems like a membership structure works well.
Or if it were a charity that ultimately had a global mission, I’d hardly expect their mission to best be served by giving as much decision-making power to an intern as a co-founder, even if the charity had a lot of power over the lives of their staff (which presumably it would).
Besides, the APA is just one example of a centralized service in analytic philosophy—Will lists several others, none of which seem democratically run to me (but I admit I haven’t checked[1]).
The more central something is, the more often we think it should be democratically controlled. The obvious example of this is the state. It has a big impact on our lives, it’s a natural monopoly, so we tend to think democracy is good to make it accountable. Rule of thumb, then: central role, decentralised control.
Yes, it’s the obvious example. The state is extremely different to EA. It’s generally hard-to-impossible to escape the state you’re born into and it has an enormous effect of your life. I don’t think you should adopt a rule of thumb “central role, decentralised control” based on the example with the strongest case for democracy.
Edit: If you want to judge for yourselves, the other examples are journals; the Philosophical Gourmet Report—which surveys ‘leading’ philosophers—and a competitor; the Stanford Encyclopedia of Philosophy; a range of services created by two philosophers; surveys of journal rankings; two news aggregator blogs.
Fair—but you probably wouldn’t pick EA’s structure either.
We like our current main billionaires, but from an ex ante perspective relying on billionaires to discern who the right leaders and technocrats are seems dicey. And of course, from the ex post perspective, we’ve had one awfully bad billionaire.
I didn’t expect people to agree with this comment, but I would be interested to know why they disagree! (Some people have commented below, but I don’t imagine that covers all the actual reasons people had)
Having read this I’m still unclear what the benefit of your restructuring of CEA is. It’s not a decentralising move (if anything it seems like the opposite to me); it might be a legitimising move, but is lack of legitimacy an actual problem that we have?
The main other difference I can see is that it might make CEA more populist in the sense of following the will of the members of the movement more. Maybe I’m as much of an instinctive technocrat as you are a democrat, but it seems far from clear to me that that would be good. Nor that it solves a problem we actually have.
I think the standard arguments for democratic membership associations apply. Increases in: membership engagement, perspective diversity, legitimacy and trust (from POV of members), accountability, transparency, and perhaps also stability (less reliant on individual personalities).
EAs already focus on much harder tasks like influencing the next million years and ending poverty. Surely setting up a membership society, a solved-problem, is not insurmountable.
True, if there’s as much motivation for the latter as the former. Perhaps more relevantly, you already focus on the much harder task of ending depression. Surely you setting up a membership society, a solved-problem, is not insurmountable.
I also believe people should have private property, and where they spend that (assuming it’s inside the law) is best left up to them, rather than trying to also make that part subject to democratic control.
I agree. So better to send them a funding proposal for an EA membership society that you’re going to set up, rather than calling for one of their major grantees to be subject to democratic control?
More broadly, I want to push back on you thinking Open Phil doesn’t “fit in” here. What happens if CEA listens to you and completely restructures how their organization is run and Open Phil doesn’t want that?
There is a very big difference between “Surely setting up a membership society, a solved-problem, is not insurmountable” as directed at the collective leadership that controls about a half-billion in unrestricted spend a year, and the same comment directed at Michael personally. Many challenges are fairly solvable with access to significant monies, but have a low probability of success without that.
Moreover, Michael’s would-be task is harder than what he is saying the collective leadership should do. Legally forming a membership society is not difficult; equipping that society to actually do the stuff CEA/EVF does is the hard part. He would be creating an alternative meta structure that would have to compete with the existing one that has tens of millions per year in support. The potential donors are good EAs; they will look at a request to fund a global conference and consider the marginal value of an additional conference.
I found the second half of the comment to be helpful.
Yeah sorry I should have drawn a stronger link between the first and second half. As in, if Open Phil thinks it’s a good idea, they’ll let CEA do it or fund Michael to do it. If they don’t, CEA can’t do it and Michael can’t do it. That CEA currently has a lot more funding is not the issue.
But of course, Open Phil may well have greater trust in CEA’s general competence than Michael’s since they fund the former and not the latter, so maybe it wouldn’t be quite as easy as that (but maybe for good reason, hard to tell as an outsider). But the attitude of “This is so easy, why don’t you do it??” is so common on this forum and I think it’s holding EA back a lot, so I want to challenge it where I see it.
But the attitude of “This is so easy, why don’t you do it??” is so common on this forum and I think it’s holding EA back a lot, so I want to challenge it where I see it.
❤️
Also I wish people gave Giving What We Can more credit; it seems to me like they are basically doing this: membership org, relatively egalitarian donor base of 10k+ people, open access events, etc.
Same with EA Norway, Czech EA, and probably others.
As for the last sentence, I think it depends on the nature of the criticism/proposal. Here, I think it’s fair to critique Michael’s proposal on the grounds that it does not acknowledge that the plausible range of action for almost anyone but Open Phil is substantially constrained by Open Phil’s willingness to go along.
That being said, “this seems fairly easy, is there a reason you you don’t do it” can be a valid line of argument in appropriate circumstances.
I’d also like to call positive attention to Michael taking a concrete step that could involve a significant personal commitment of time (i.e., applying to be on the EVF board) in addition to writing on the Forum about the issue he sees.
EA Norway did this! (Set up a membership society which includes voting rights in the organization, make a newsletter, run conferences, etc.)
I don’t at all want to diminish how hard they worked, but I don’t think it was as challenging as you imply (e.g. I don’t know their budget but I’m sure it’s way less than $500 million/year).
Setting up a society can definitely be done on much less than $500M/yr! The point of my remark was to contrast what Michael had called on leadership to do with what sounded like a suggestion that Michael do the same thing personally. The reference to $500M was meant to underscore the extreme difference between Michael’s ask and what was asked of him in return, not to suggest $500M was necessary to form a society.
That being said, while I haven’t seen a recent budget for CEA, my assumption is that running an organization that could serve as a potential replacement for CEA (which is what this particular subthread was about in my view) would cost tens of millions USD per year. Michael’s view (as roughly/imprecisely summarized by me) is that many central coordinating functions (e.g., “natural monopolies”) currently handled by CEA should be instead run by a membership society. So the example of EA Norway doesn’t really update my belief in my asserion in the context of this discussion that “[l]egally forming a membership society is not difficult; equipping that society to actually do the stuff CEA/EVF does is the hard part.”
The main crux here might be the extent to which CEA has a monopoly on supporting people who want to do good effectively.
To the extent that it is a monopoly, it’s harder for people to start new projects in the space simply because they didn’t get there first.
To the extent that it isn’t a monopoly, anyone who thinks CEA could be much better can always try to start their own thing. Yes it would be very hard; it was very hard for the founders of CEA too.
But I think CEA is much less of a monopoly than it seems a lot of EAs think it is.
That’s part of the point of this post, right? There’s even an example of people starting a competitor to CEA in the ‘EA student group support’ space, getting funding from Open Phil, and having people like Will say they did a great job. And before Probably Good, there was only one org providing EAs with careers advice; but instead of calling for 80,000 Hours to make big changes to the way 80,000 Hours thinks they should run their free service, Omer and Sella started Probably Good, with financial support from Open Phil and encouragement from 80,000 Hours. In the ‘EA career support’ space, there’s also now Successif, Magnify Mentoring and High Impact Professionals, each focusing on areas they thought needed more attention.
“Ah, but the conferences are much more important as a centralized function and they are basically a monopoly.” In 2018, CEA gave a $10,000 grant to a competitor conference that had 100 attendees.
“But the EA Forum!” There are tons of Slack spaces and Facebook groups etc. not run by CEA—CEA is definitely not in control of all online discussion between EAs. But maybe a competitor forum is next on the list (not something Michael’s particularly concerned about though, so maybe someone else wants to have a go).
“Community Health!” Oh my god, if you found a successful competitor to the Community Health team, I will shower you with praise and gratitude. And I wouldn’t be surprised if they did too.
“Okay, maybe not CEA, but Open Phil!” Future Fund. Regardless of how FTX turned out, this was at least a proof of concept.
“Look, EA just needs to be radically different but there’s already an EA!” Start your own movement. Holden and Elie thought charity evaluators should be a lot better so they started GiveWell. The Oxford crew thought people should be doing good better so they started CEA. If you think EA just fundamentally needs to be more democratic but keep everything else the same, start a movement for Democratic Effective Altruism. I might even start one for Do-acratic EA.
There’s likely a second crux that influences how one views the extent to which CEA/EVF is a “monopoly” or has extreme advantages. That is whether it is advisible for the same organization (EVF) to be the primary provider of many different kinds of important coordinating functions, or whether that gives it too much power.
If that isn’t a concern, then pointing to the existence and viability of organizations that work in the same spaces at CEA/EVF orgs is a fairly good response.
“Start your own orgs” is still a possible response if one concludes that CEA/EVF’s dominant market position in numerous forms of coordination is a problem. However, I think the difficulty level is raised two orders of magnitude from most of the examples you gave:
The first raise is that the new org has to outcompete the EVF org to displace the latter from its role as the primary provider of the coordination system.
The second raise is that this would need to happen over several different coordinating functions to reduce CEA/EVF’s influence to an appropriate level.
(Although I would prefer a meta with less power concentration, “democratic” is not the primary word I’d use to justify that preference.)
It’s definitely hard to replace CEA! But this thread has an air of helplessness, like there are only 10 people in the world who can do anything in EA, and this seems immediately falsified by the large number of people who are doing things in EA, including specifically the stuff Michael suggests like having membership societies.
(Note: I don’t know to what extent you endorse this view of helplessness, so feel a little like I’m picking on you here, but I feel fairly confident that the median reader would take away a sense of helplessness from this thread.)
[Written in a personal capacity, etc. This is the second of two comments, see the first here.]
In this comment, I consider how centralised EA should be. I’m less sure how to think about this. My main, tentative proposal is:
We should distinguish central functions from central control. The more central a function something has, the more decentralised control of it should be. Specifically, I suggest CEA should become a fee-paying members’ society that democratically elects its officers—much like the America Philosophical Association does.
I suspect it helps not just to ask “how centralised should EA be” but also “what should be centralised and what shouldn’t?”. Some bits are, as you say, natural monopolies in that it’s easiest if there’s one of them. This seems most true for places where people meet and communicate with each other: a conference is valuable because other relevant people are there. For EA, I guess the central bits are the conferences, the introductory materials, the forum, the name(?), maybe other things. In my post on EA as a marketplace, which you kindly reference but don’t seem sympathetic to, I point out you can think of EA on a hub-a-spoke model. Imagine a bicycle wheel, where the rim represents the community members and the spokes the connections. There are some bits we widely participate in, the hub, such as the conference. But, besides that, people have links to only a subset: longtermists hang out mostly with longtermists, etc.
Now, it does not follow that simply because something has a central function, it’s used by lots of people, that it should be centrally controlled, i.e. controlled by a few people. In fact, we often think the opposite is true. The more central something is, the more often we think it should be democratically controlled. The obvious example of this is the state. It has a big impact on our lives, it’s a natural monopoly, so we tend to think democracy is good to make it accountable. Rule of thumb, then: central role, decentralised control.
Another example of this is the one you gave, the American Philosophical Association. It’s pretty useful to have a place that convenes those who have a common interest—in that case, doing academic philosophy. The APA seems useful and unobjectionable (this is my impression of it, anyway). But, the reason for this is that it doesn’t and can’t do anything besides serve and convene its members. It doesn’t take sides in philosophical debates or try to steer the field. People would object if it tried. How is this unobjectionableness achieved? I imagine it’s to do with the fact it’s a fee-paying society where those in positions of power are elected by members, as well as the fact it doesn’t control lots of funding. Despite being philosophers, I doubt the members of the APA would want it to be run by unelected philosopher kings! Roughly, the moral of the story seems to be that, if something is so central you can’t avoid participating in, you probably want decentralised power.
You talk about various possible ways to centralise or decentralise EA. But why is there no suggestion of democratising the central element of EA, namely the Centre for Effective Altruism? Here’s a concrete suggestion: CEA becomes like the APA: a fee-paying membership society where the members elect the trustees or officials, who then administer various central functions, like a conference, journals, etc. Is this an absurd idea? If so, why? It’s hardly radical. Members’ organisations are a default coordination solution where people have a common interest. I’m not sure it’s a brilliant idea, but it’s weird it’s not been discussed, and I’d be happy for someone to tell me why would be terrible.
Worried no one would join? If you want to kick-start it, you could provide a year’s free membership to those who have signed the GWWC pledge or attended an EA conference. There may be other ideas. EAs already focus on much harder tasks like influencing the next million years and ending poverty. Surely setting up a membership society, a solved-problem, is not insurmountable.
There’s been a lot of discussion recently about people not feeling like they’re part of EA. Well, here’s a cheap solution: let people become members of CEA. Then, you’re in and you can have a say in how things are run. This has other advantages: it makes CEA accountable to its members. It also allows it to genuinely speak for them, which currently it can’t do, because it doesn’t represent them. By charging a membership fee, CEA can offset the costs of other things, so won’t be so reliant on other donors. Honestly, this membership scheme would probably work if it were just about the EA conferences (the “EA conference Association”?).
Who should be against this idea? I recognise some effective altruists are sceptical of applying democracy to philanthropy, but, to be clear, I am not advocating for communism, that all of EA or ‘EA resources’ should come under common control: that Open Philanthropy should give 1/Nth of its resources to the N people self-describing as EA, or that if you want to become part of EA you need to give the community all your (spare) money. That’s absurd. I am only in favour of democratising the central convening and coordinating parts; this is like saying the APA should be a democratic entity, not that all philosophy and philosophy funding should be run by a democracy. As far as I can see, all the central elements of EA fall under CEA (I’m also not against spinning off parts of EVF.)
I don’t think it makes sense to have a democracy for the non-central elements—the spokes. I believe in free enterprise, including free philanthropic enterprise, and private ownership.
Where does Open Philanthropy fit into this, given it has most of the ‘EA money’? I’m not sure it does fit in. It’s fortunate there’s one really big donor (because there’s more money), and unfortunate there’s only one (but that one will have outsized influence). I think society should control some of people’s income through taxation. But I also believe people should have private property, and where they spend that (assuming it’s inside the law) is best left up to them, rather than trying to also make that part subject to democratic control. Hence, insofar as worries about centralisation spring from their being a single, huge funder, I don’t have a neat solution to that. That doesn’t mean other things couldn’t be done, however, such as democratising CEA. Indeed, if CEA were a democratic society, I’d be fairly relaxed about Open Philanthropy providing funding to it, because control of CEA would still be decentralised, so that would mitigate concerns about undue influence. Of course, I have views about how, morally speaking, people should spend their post-tax wealth, but that seems a separate issue.
Finally, you raise the point of who has the ‘legitimacy’ to centralise or decentralise EA, and say it should come from CEA or Open Philanthropy (the use of ‘legitimacy’ is somewhat interesting, because that has democratic connotations). At this point, I should probably mention I applied to be a board member of Effective Ventures, and in my application form, I explicitly stated I was interested in exploring bringing democratic elements into CEA, and so decentralising it. I didn’t make it to the first stage (I also asked for feedback and was told EV couldn’t provide any). Now, I am not claiming I am a ‘slam dunk’ choice for the board—EA has many highly talented people. However, I did find it discouraging, not least because I am interested in, and would have had legitimacy for, exploring institutional reforms. It reduced my confidence that, despite what you say, ‘central EA’ really is open to diverse voices or further decentralisation.
I think it’s a mistake to conflate making things more democratic or representative and making them more decentralised—historically the introduction of more representative institutions facilitated the centralisation of states by increasing their ability to tax cities (see e.g. here). In the same way I would expect making CEA/EVF more democratic would increase centralisation by increasing their perceived legitimacy and claim to leadership.
Yes, I think there’s a lot of sliding between “decentralised” and “democratic” even though these have pretty much nothing to do with each other.
As a pretty clear example, the open source software community is extremely decentralised but has essentially zero democracy anywhere.
I take it you’re saying making things more democratic can make them more powerful because they then have greater legitimacy, right? More decentralised power → large actual power?
I suppose part of my motivation to democratise CEA is that it sort of has that leadership role de facto anyway, and I don’t see that changing anytime soon (because it’s so central). Yet, it lacks legitimacy (i.e. the de jure bit), so a solution is to give it legitimacy.
I guess someone could say, “I don’t want CEA to have more power, and it would have if it were a members society, so I don’t want that to happen”. But that’s not my concern. If anything, what your comments make me think is (1) something like CEA should exist, (2) actual CEA does a pretty good job, (3) nevertheless, there’s something icky about its lack of legitimacy (maybe I’m far more of an instinctive democratic that I thought), (4) adding some democracy stuff would address (3).
I’m confused about the mathematics of a a fee-paying membership society. I’m having a hard time seeing how that would generate more than a modest fraction of current revenues.
It’s not clear what the “central convening and coordinating parts” are. Neither Current-CEA nor Reformed-CEA would have a monopoly on tasks like funding community builders, funding/running conferences, and so on. They are just another vendor who the donors can choose to hire for those purposes. There is and would be no democratic mandate that donors who would like to fund X, Y, and Z are obliged to go through CEA.
I think your model is correct insofar as the membership society could assert independent control of certain epistemically critical functions that are relatively less reliant on funding (e.g., the Forum).
The extent to which “convening and coordinating” is effective may depend on whether there is money behind those efforts. Stated more directly, to what extent are CEA’s efforts in these areas boosted by the well-known (general yet strong) alignment between CEA and the major funder in the ecosystem? Would Reformed-CEA enjoy the same boost?
I used to work at EA Norway, which is a fee-paying membership society, and thought it might be useful to share more on how our funding worked. This is just meant as an example, and not as an argument for or against membership societies. (Here’s a longer comment explaining how we organise things at EA Norway.)
I can’t speak to EA Norway’s current situation, as I no longer have any position at EA Norway (other than being a paying member). However, I can say what it was like in 2018-2021 when I was Executive Director (ED). The total income from the membership fee roughly covered the cost of the general assembly. Most of our funding came from a community building grant from the Centre for Effective Altruism (CEA). However, the board made sure to fundraise enough from private donors for my salary. The two main reasons for this was to I) diversify our funding, and II) enable us to make longer term plans than CEAs grant periods.
When the board gave approval to accept the community building grant from CEA, we discussed that if at any point we did not want to follow CEAs guidelines and success metrics, we would pay back the remainder of the grant. This was definitely easier for us to say and truly mean when we had covered the ED’s salary from other sources, as it meant that if we were to return the funding, we would still have at least one employee. We never ended up disagreeing so much with CEA that we wanted to return the funds, though we were definitely very vocal about any disagreements we had with the groups team at CEA and did push for some changes.
I’m confused why you think this is required, I don’t think Michael implied it would.
The society wouldn’t be a good replacement for CEA unless it could attract significant major donor support. As the next paragraph implies, there’s no reason for major donors to support the society if they judge an alternative vendor to be more effective in delivering conferences, etc. As a result, the society would either have to adapt its programs to meet the scoring metrics of the big donors (in which case the democratic nature of the organization isn’t doing much work; the money is still calling the shots) or it would lack funding to perform those functions (in which case the organization isn’t effective on those functions).
As my third paragraph suggested, there are functions the membership society could potentially run on member revenue and small donations. But that is a significant tradeoff.
Yeah, I’ve not spent loads of time trying to think through the details. I’m reluctant to do so unless there’s interest from ‘central EA’ on this.
As ubuntu’s comments elsewhere made clear, it’s quite hard for someone to replicate various existing community structures, e.g. the conferences, even though no one has a literal monopoly on them, because they are still natural monopolies. If you’re thinking “I can’t imagine a funder supporting a new version of X if X already exists”, then that’s a good sign it is a central structure (and maybe should have democratic elements). There are lots of philosophy conferences, but that doesn’t take away from the value of having a central one.
Also, you make the point “well, but would reformed-EA be worth doing if the main funder wouldn’t support it?”. Let’s leave that as an open question. But I do want to highlight a tension between that thought and the claim that “EA is not that centralised”. If how EA operates depends (very) substantially on one what a single funder thinks, we should presumably conclude EA is very centralised. Of course, it’s then a further question of whether or not that’s good and what, if anything, should be done by various individuals about it.
Yes, I think the proposal effectively highlights that EA is significantly more centralized than some claim.
My guess is that you would have to add a claim like “Funders should not fund ‘central convening and coordinating’ functions except as consistent with the community’s will” to get anywhere with your proposal as currently sketched. That’s a negative norm, less demanding than an affirmative claim to funding. But I haven’t exhaustively explored the possibilities either.
My own view is that a member-led organization is probably viable and a good idea, but has to be realistic about what functions it could assume.
Well, you’re not going to fund stuff if you don’t like what the organisation is planning to do. That’s generally true.
I don’t mind the idea of donors funding a members’ society. This happens all the time, right? It’s just the leaders have to justify it to the members. It’s also not obvious that, if CEA were a democratic society, it would counterfactually lose funding. You might gain some and lose others. I’m not sure I would personally fund ‘reformed-CEA’ but I would be more willing to do so.
Okay, but the American Philosophical Association “was founded in 1900 to promote the exchange of ideas among philosophers, to encourage creative and scholarly activity in philosophy, to facilitate the professional work and teaching of philosophers, and to represent philosophy as a discipline” with a modern mission as follows ” promotes the discipline and profession of philosophy, both within the academy and in the public arena. The APA supports the professional development of philosophers at all levels and works to foster greater understanding and appreciation of the value of philosophical inquiry.” Seems like a membership structure works well.
If, on the other hand, the APA’s mission was to “help solve the greatest philosophical problems of our time by supporting philosophers” or some such, I personally think that a more meritocratic approach seems like a better fit. It’s certainly not obvious to me that a democratic membership structure would be superior.
Or if it were a charity that ultimately had a global mission, I’d hardly expect their mission to best be served by giving as much decision-making power to an intern as a co-founder, even if the charity had a lot of power over the lives of their staff (which presumably it would).
Besides, the APA is just one example of a centralized service in analytic philosophy—Will lists several others, none of which seem democratically run to me (but I admit I haven’t checked[1]).
Yes, it’s the obvious example. The state is extremely different to EA. It’s generally hard-to-impossible to escape the state you’re born into and it has an enormous effect of your life. I don’t think you should adopt a rule of thumb “central role, decentralised control” based on the example with the strongest case for democracy.
Edit: If you want to judge for yourselves, the other examples are journals; the Philosophical Gourmet Report—which surveys ‘leading’ philosophers—and a competitor; the Stanford Encyclopedia of Philosophy; a range of services created by two philosophers; surveys of journal rankings; two news aggregator blogs.
Fair—but you probably wouldn’t pick EA’s structure either.
We like our current main billionaires, but from an ex ante perspective relying on billionaires to discern who the right leaders and technocrats are seems dicey. And of course, from the ex post perspective, we’ve had one awfully bad billionaire.
I didn’t expect people to agree with this comment, but I would be interested to know why they disagree! (Some people have commented below, but I don’t imagine that covers all the actual reasons people had)
Having read this I’m still unclear what the benefit of your restructuring of CEA is. It’s not a decentralising move (if anything it seems like the opposite to me); it might be a legitimising move, but is lack of legitimacy an actual problem that we have?
The main other difference I can see is that it might make CEA more populist in the sense of following the will of the members of the movement more. Maybe I’m as much of an instinctive technocrat as you are a democrat, but it seems far from clear to me that that would be good. Nor that it solves a problem we actually have.
I think the standard arguments for democratic membership associations apply. Increases in: membership engagement, perspective diversity, legitimacy and trust (from POV of members), accountability, transparency, and perhaps also stability (less reliant on individual personalities).
True, if there’s as much motivation for the latter as the former. Perhaps more relevantly, you already focus on the much harder task of ending depression. Surely you setting up a membership society, a solved-problem, is not insurmountable.
I agree. So better to send them a funding proposal for an EA membership society that you’re going to set up, rather than calling for one of their major grantees to be subject to democratic control?
More broadly, I want to push back on you thinking Open Phil doesn’t “fit in” here. What happens if CEA listens to you and completely restructures how their organization is run and Open Phil doesn’t want that?
There is a very big difference between “Surely setting up a membership society, a solved-problem, is not insurmountable” as directed at the collective leadership that controls about a half-billion in unrestricted spend a year, and the same comment directed at Michael personally. Many challenges are fairly solvable with access to significant monies, but have a low probability of success without that.
Moreover, Michael’s would-be task is harder than what he is saying the collective leadership should do. Legally forming a membership society is not difficult; equipping that society to actually do the stuff CEA/EVF does is the hard part. He would be creating an alternative meta structure that would have to compete with the existing one that has tens of millions per year in support. The potential donors are good EAs; they will look at a request to fund a global conference and consider the marginal value of an additional conference.
I found the second half of the comment to be helpful.
Yeah sorry I should have drawn a stronger link between the first and second half. As in, if Open Phil thinks it’s a good idea, they’ll let CEA do it or fund Michael to do it. If they don’t, CEA can’t do it and Michael can’t do it. That CEA currently has a lot more funding is not the issue.
But of course, Open Phil may well have greater trust in CEA’s general competence than Michael’s since they fund the former and not the latter, so maybe it wouldn’t be quite as easy as that (but maybe for good reason, hard to tell as an outsider). But the attitude of “This is so easy, why don’t you do it??” is so common on this forum and I think it’s holding EA back a lot, so I want to challenge it where I see it.
❤️
Also I wish people gave Giving What We Can more credit; it seems to me like they are basically doing this: membership org, relatively egalitarian donor base of 10k+ people, open access events, etc.
Same with EA Norway, Czech EA, and probably others.
Upvoted; thanks.
As for the last sentence, I think it depends on the nature of the criticism/proposal. Here, I think it’s fair to critique Michael’s proposal on the grounds that it does not acknowledge that the plausible range of action for almost anyone but Open Phil is substantially constrained by Open Phil’s willingness to go along.
That being said, “this seems fairly easy, is there a reason you you don’t do it” can be a valid line of argument in appropriate circumstances.
I’d also like to call positive attention to Michael taking a concrete step that could involve a significant personal commitment of time (i.e., applying to be on the EVF board) in addition to writing on the Forum about the issue he sees.
EA Norway did this! (Set up a membership society which includes voting rights in the organization, make a newsletter, run conferences, etc.)
I don’t at all want to diminish how hard they worked, but I don’t think it was as challenging as you imply (e.g. I don’t know their budget but I’m sure it’s way less than $500 million/year).
Setting up a society can definitely be done on much less than $500M/yr! The point of my remark was to contrast what Michael had called on leadership to do with what sounded like a suggestion that Michael do the same thing personally. The reference to $500M was meant to underscore the extreme difference between Michael’s ask and what was asked of him in return, not to suggest $500M was necessary to form a society.
That being said, while I haven’t seen a recent budget for CEA, my assumption is that running an organization that could serve as a potential replacement for CEA (which is what this particular subthread was about in my view) would cost tens of millions USD per year. Michael’s view (as roughly/imprecisely summarized by me) is that many central coordinating functions (e.g., “natural monopolies”) currently handled by CEA should be instead run by a membership society. So the example of EA Norway doesn’t really update my belief in my asserion in the context of this discussion that “[l]egally forming a membership society is not difficult; equipping that society to actually do the stuff CEA/EVF does is the hard part.”
The main crux here might be the extent to which CEA has a monopoly on supporting people who want to do good effectively.
To the extent that it is a monopoly, it’s harder for people to start new projects in the space simply because they didn’t get there first.
To the extent that it isn’t a monopoly, anyone who thinks CEA could be much better can always try to start their own thing. Yes it would be very hard; it was very hard for the founders of CEA too.
But I think CEA is much less of a monopoly than it seems a lot of EAs think it is.
That’s part of the point of this post, right? There’s even an example of people starting a competitor to CEA in the ‘EA student group support’ space, getting funding from Open Phil, and having people like Will say they did a great job. And before Probably Good, there was only one org providing EAs with careers advice; but instead of calling for 80,000 Hours to make big changes to the way 80,000 Hours thinks they should run their free service, Omer and Sella started Probably Good, with financial support from Open Phil and encouragement from 80,000 Hours. In the ‘EA career support’ space, there’s also now Successif, Magnify Mentoring and High Impact Professionals, each focusing on areas they thought needed more attention.
“Ah, but the conferences are much more important as a centralized function and they are basically a monopoly.” In 2018, CEA gave a $10,000 grant to a competitor conference that had 100 attendees.
“But the EA Forum!” There are tons of Slack spaces and Facebook groups etc. not run by CEA—CEA is definitely not in control of all online discussion between EAs. But maybe a competitor forum is next on the list (not something Michael’s particularly concerned about though, so maybe someone else wants to have a go).
“Community Health!” Oh my god, if you found a successful competitor to the Community Health team, I will shower you with praise and gratitude. And I wouldn’t be surprised if they did too.
“Okay, maybe not CEA, but Open Phil!” Future Fund. Regardless of how FTX turned out, this was at least a proof of concept.
“Look, EA just needs to be radically different but there’s already an EA!” Start your own movement. Holden and Elie thought charity evaluators should be a lot better so they started GiveWell. The Oxford crew thought people should be doing good better so they started CEA. If you think EA just fundamentally needs to be more democratic but keep everything else the same, start a movement for Democratic Effective Altruism. I might even start one for Do-acratic EA.
I think that’s one major crux.
There’s likely a second crux that influences how one views the extent to which CEA/EVF is a “monopoly” or has extreme advantages. That is whether it is advisible for the same organization (EVF) to be the primary provider of many different kinds of important coordinating functions, or whether that gives it too much power.
If that isn’t a concern, then pointing to the existence and viability of organizations that work in the same spaces at CEA/EVF orgs is a fairly good response.
“Start your own orgs” is still a possible response if one concludes that CEA/EVF’s dominant market position in numerous forms of coordination is a problem. However, I think the difficulty level is raised two orders of magnitude from most of the examples you gave:
The first raise is that the new org has to outcompete the EVF org to displace the latter from its role as the primary provider of the coordination system.
The second raise is that this would need to happen over several different coordinating functions to reduce CEA/EVF’s influence to an appropriate level.
(Although I would prefer a meta with less power concentration, “democratic” is not the primary word I’d use to justify that preference.)
It’s definitely hard to replace CEA! But this thread has an air of helplessness, like there are only 10 people in the world who can do anything in EA, and this seems immediately falsified by the large number of people who are doing things in EA, including specifically the stuff Michael suggests like having membership societies.
(Note: I don’t know to what extent you endorse this view of helplessness, so feel a little like I’m picking on you here, but I feel fairly confident that the median reader would take away a sense of helplessness from this thread.)