Should Effective Altruists Focus More on Movement Building?

post by aaronb50 · 2020-12-30T03:16:11.757Z · EA · GW · 4 comments

Contents

  Motivation
  A Priori Considerations
  Importance
  Neglectedness
  Tractability
  Arguments Against MB
  Ideas 
  Conclusion
None
4 comments

The aim of this post is to begin a conversation about whether effective altruists should dedicate a larger proportion of marginal resources to growing and/or improving the EA movement itself. I’ll lay out my motivation for writing this post and explain some a priori reasons to expect this cause-area to be neglected, before offering a preliminary evaluation using the importance/neglectedness/tractability framework. 

Motivation

It seems to me a valuable principle of EA is that we should evaluate interventions by their net counterfactual impact irrespective of how “directly” they bring it about. An implication is that indirect and upstream interventions can be particularly high-leverage and thus particularly worthy of our time and money. I, and I suspect most EAs, can attribute most of their positive impact at least indirectly to the work of one or more intentional efforts to advance the movement. For example, my choice of major and tentative career plans are largely a direct result of the work of 80,000 Hours. Likewise, this document by Aaron Gertler of the Centre for Effective Altruism inspired me to write this post.  

For this reason, I intuitively suspect that EA movement building (henceforth MB) is one of the best ways to improve the world. However, my impression (based on the casual consumption of EA forum posts, 80k podcasts, speaking with a few other EAs, etc.) is that it receives a disproportionately small amount of attention relative to the standard set of more direct EA cause areas. While there may be good reason for this, I wanted to investigate a bit more whether MB is appropriately funded and attended to. 

A Priori Considerations

It is possible that some EAs may refrain from discussing or encouraging MB building in order to avoid appearing self-serving or “cultish.” I suspect this effect is particularly strong for the most knowledgeable, experienced, and/or influential EAs because these are the individuals most likely to be involved with the organizations that might benefit from MB-associated funding. For instance, might some of the folks at 80k or CEA be a bit biased, on the margin, against encouraging donations to places like 80k and CEA in order to avoid an appearance of bias in the other direction? 

If so, that leaves it to lowly undergrads with less knowledge and influence like me to write posts like this one. Don’t get me wrong, insofar as promoting MB really would send a negative signal - deservedly or otherwise - about EA, some degree of bias might be warranted. That said, it seems plausible that the amount of bias is much greater than would be appropriate. As a final note, I will sometimes refer to dollars donated as the metric of resource allocation in question, but would note that this should be taken as a stand-in for all relevant resources. 

As a separate point, it seems much easier to get non-EA individuals or organizations to support direct EA causes than to support EA movement building. For example, a company is much more likely to host a fundraiser for the Against Malaria Foundation than for 80,000 Hours. Therefore, those who identify as EAs should dedicate a larger proportion of their own resources to MB than they would if they could dictate how all causes and charities are funded in order to nudge these proportions towards their hypothetical ideal. 

Importance

To a very rough first approximation, MB seems unlikely to dramatically change how resources are apportioned across cause areas. In other words, if 50% of non-MB donations are currently being dedicated to global health and poverty reduction, 15% to animal welfare, 25% to existential risk reduction, and 10% to other causes, I would expect growing the movement to approximately maintain these proportions while growing the pie of resources, so to speak. If so, this implies that MB is important to the same degree that EA at large is important. 

It is almost necessarily the case that individual EAs apportion their resources in a manner they think superior to the movement as a whole; if not, a person could improve her expected impact by emulating her peers’ donations. Under these two assumptions, individual EAs should expect that growing EA by $1 via MB activities is less valuable than simply donating an additional dollar of her own. However, what MB loses in efficiency might plausibly be made up for in leverage.

For example, suppose EA was made of two direct-work organizations, Against Malaria Causing Extinction Foundation and Making Wild Insects P-Zombies Coalition, each of which get half of all EA resources at the moment. Maybe you think the former produces 10 utils/dollar, and the latter 4 utils/dollar, and so donate exclusively to AMCEF. Since EA at large generates 7 utils/dollar, movement building would have to grow the pie by 10/7=$1.43 per dollar spent on MB itself in order to justify donating here instead.

However, there are some reasons to think that MB would not maintain the present distribution of resources in expectation. I have two conflicting intuitions about this, and there are likely other considerations I’m neglecting.

First, we might expect that increasing the number of hours, people, and dollars dedicated to figuring out how to do the most good would improve the distribution of resources. I think this is likely to be the case insofar as MB consists in or stimulates cause prioritization research. Indeed, the EA Infrastructure Fund has funded 80,000 hours and the Forethought Foundation to do just this sort of research. If this is the case, in the example above, $1 spent on MB would need to generate less than $1.43 in eventual donations to justify itself. As an aside, I think there is an important distinction to be made between prioritization research and movement growth, though as it currently stands the same organizations seem to be undertaking both and the two are frequently grouped together.  

Second, we might expect that growing the number of individuals involved in EA would worsen the distribution of resources. Intuitively, most social movements begin with a small group of particularly dedicated members and gradually attract more casual followers with time; perhaps more casual EAs would donate to less beneficial organizations on average.

 Likewise, perhaps MB activities encourage EAs to promote more broadly-appealing interventions over more cost-effective but “weirder”-seeming activities like preventing AI from turning the universe into a paperclip factory. If this is the case, in the example above, $1 spent on MB would need to generate more than $1.43 in eventual donations to justify itself.

Neglectedness

Using dollars as a proxy for all resources, let’s take a look at the total past and present amount controlled by each of the EA Funds

  1. The Global Health and Development Fund has given out $9,601,162 and controls about $1,880,000 for a total of $11,481,162
  2. The Animal Welfare Fund has given out $6,058,210 and controls about $395,000 for a total of $6,453,210
  3. The Long-Term Future Fund has given out $4,171,735 and controls about $325,000 for a total of $4,496,735
  4. The EA Infrastructure Fund has given out $3,917,264 and controls about $345,000 for a total of $4,262,264

So in total, EA Infrastructure (which I am taking as synonymous with movement building) constitutes 4,262,264/(4,262,264+4,496,735+6,453,210+11,481,162) or about 16% of the funds as a whole. 

This might be too low even if this was an unbiased estimate. However, I strongly suspect a positive correlation between (i) donating through CEA’s EA funds and (ii) donating to movement building activities, at least because both involve knowledge of and believing in the work of the Centre for Effective Altruism. If so, this would imply that MB receives less than 16% of all EA resources. 

Likewise, it seems more difficult for an individual donor to know which specific organizations to fund in the MB space in comparison to other cause areas because there is no meta-charity evaluator analogous to GiveWell which ranks EA infrastructure/MB organizations themselves (though perhaps there should be…). If so, this would further strengthen the correlation described above, in turn biasing the 16% estimate upward. 

Trying to improve this estimate is tough, but let’s try anyway. According to CEA's 2020 Review, CEA had a budget of $6.02 million which it used to, among other things: 

  1. $1.3 million in “community grants”
  2. $340,000 to run the EA forum and produce online content
  3. $560,000 to run events
  4. $300,000 on “community health”
  5. $600,000 to run itself, 80,000 Hours, the Forethought Foundation, and Giving What We Can

Given all that, I wouldn’t be surprised if CEA received at least half of all EA MB resources, which would imply a total MB expenditure of $12 million. The Open Philanthropy Project says they “recommended” over $200 million in donations in 2019, and granted over $100 million themselves (as an aside, it seems OpenPhil has the question of how to regard MB on their mind as well). 

For simplicity, let’s assume that CEA’s budget is directly comparable to donation quantity. Even if OpenPhil’s recommendations and donations constituted all of EA giving, EA MB would receive less than 4% of all EA resources. Intuitively, such an estimate indicates relative neglect of MB as a cause area. 

Tractability

My biggest doubt about MB as a cause area involves tractability - both the question of whether social movement growth is generally tractable and whether EA at present has the capacity to deploy more resources. 

Unfortunately, both of these questions seem in particular need of more intensive research than I’m undertaking in this post, so any conclusions will be pretty tentative. Even still, it seems there are a few a priori considerations worth getting on the table. 

First, approximately 0% of people have even heard of EA. Perhaps in Cambridge (both of them) or San Francisco one gets a different impression, but my anecdata point is that in 20 years of living in a liberal, wealthy, well-educated suburb near D.C. and two years attending a selective (but non-Ivy) university, I cannot recall a single time that I’ve heard EA mentioned outside of an EA-adjacent event. 

If I’m not mistaken about this, it should be relatively easy to expand EA - if this is a thing we decide is good to do - simply by exposing more people to the idea. There are likely a lot of people predisposed to support EA who do not even have EA ideas in their mental space of possibilities. I’d bet a lot of people reading this post at one point fell into this bucket.

Compare this to the most salient and contentious social or political movements or ideas, which virtually everyone knows and has an opinion about. For example, most people know what veganism is, so mere exposure to the concept and a few basic arguments in its favor won’t significantly increase the number of vegans. No U.S. voter thinks “huh, just heard about this ‘Democratic Party’ a few days ago. Seems good, might vote for them instead.” 

Though I can’t recall precisely where, I’ve heard it argued that EA growth should be undertaken with caution lest the movement become irreversibly associated with a political tribe, just as once-bipartisan political ideas become controversial in proportion to their salience. I think this is a serious concern, but suspect that there is a lot of room for growth without this occurring. 

Back of the envelope calculation: There are about 6,000 GWWC members, so 50,000 might be a generous estimate of the number who identify as EAs in the U.S. I doubt more than 50 times this many, or 2.5 million people, have heard of the concept. So, only about 0.02% of the 250 million U.S. adults are EAs and 1% have heard of the concept. As a comparison, this is two about orders of magnitude lower than veganism. Is it unreasonable to expect we could expose 10% of the population to EA without undue tribal affiliation? 

Regarding the question of whether EA has the capacity to deploy MB resources at the moment, I don’t think I have much insight to add above soliciting the opinions of those who currently work at places like CEA. However, there are a few reasons why I’d expect there to be at least some good way to spend marginal dollars.

First, the GiveWell advertisements I’ve heard on several podcasts indicate that at least they think it’s worth spending resources to spread EA ideas. This speaks to a more general point; advertising (and perhaps other interventions?) seems to scale quite easily and doesn’t require existing organizations to expand their capacity. 

Unlike with salient movements and ideologies, advertising can be targeted at audiences particularly predisposed to be receptive to EA. For example, I’ve heard GiveWell ads on the Ezra Klein Show, which likely attracts listeners interested in using ideas to improve the world but is not directly associated with rationalism or EA. I’d have a hard time thinking either that these ads aren’t positive in expectation or that there aren’t similar opportunities that would be accessible with more money. 

To take another example, I am personally deeply impressed by the work of Kelsey Piper and Dylan Matthews et al at Future Perfect. But, why did its creation require a grant from the Rockefeller Foundation, which is not explicitly associated with EA? If there was a team at CEA flush with money and looking to spread EA ideas, would Future Perfect or a similar project have been started several years earlier? What other excellent projects don’t exist yet because MB has been relatively neglected?

Arguments Against MB

There are several plausible reasons that MB should not receive a larger proportion of EA resources. For one thing, I might be wrong about any or all of the specific points made in the subsections above. Relatedly, an “efficient market for ideas” hypothesis would suggest that if MB really was important, neglected, and tractable, then other more experienced and influential EAs would have already raised its salience.

At the object level, perhaps the argument that most stands out is the irreversibility and potential damage of movement building. Once an idea proliferates and acquires social and political connotations, there’s no going back. Thus, extreme caution might be warranted before, say, buying a Superbowl ad. Even if this is the case, it may likely be worthwhile to begin the research necessary to more confidently begin MB activities. 

As a final point, the alternative to intentional MB isn’t stagnation, but rather the haphazard and undirected growth that is occurring right now. Perhaps this is superior to intentional MB, but I’m not sure what the arguments for this would be besides a general preference for slower change.

Ideas 

Merely as a starting point for further discussion, here are a few MB ideas that might warrant further consideration. 

  1. MB research to identify whether and how to build the movement.
    1. Literature review on social movement growth
    2. Analyzing case studies of successful and failed social movements in the past
    3. Identifying particular media and markets for targeted advertising
    4. Conducting surveys and focus group studies to identify the best messages and wordings (like the Good Food Institute’s intensive testing of “clean” and “cultivated” meat)
  2. Spreading EA ideas
    1. Advertising, especially targeted at receptive audiences.
    2. College outreach
      1. Note: it seems most of the top 10 or 15 U.S. colleges have EA clubs of some type, but many selective schools ranked 10-30 or so do not.
    3. High school outreach
      1. Note: this may be particularly important because major is easiest to change before starting college. The 80k career guide was one of the main reasons I chose my majors, but this was only possible because I read it in high school.
    4. More, bigger, and better versions of whatever CEA and other EA organizations are already doing, including support for local EA groups
    5. Documentary production
      1. See this Wikipedia section on An Inconvenient Truth.

Conclusion

I’m sure there are important arguments both for and against movement building that I’ve failed to consider. I hope this post serves as a starting point for a more informed discussion. To reiterate one interesting point from the intro, it’s worth noting that this very post was inspired by a movement building intervention - namely, an enthusiastic encouragement of EA Forum contributions from new writers. Perhaps even more of this type of activity would help EA do the most good that it can. 

4 comments

Comments sorted by top scores.

comment by Aaron Gertler (aarongertler) · 2020-12-30T11:50:59.406Z · EA(p) · GW(p)

What I liked about this post: 

  • While you began by speculating that movement building received a relatively small fraction of EA resources, you later backed that up with a lot of data. (It would have been nice to present the data earlier, though -- it lends more context to your early arguments.)
  • You bring up specific organizations that might invest more in MB strategy (e.g. GiveWell buying more ads) and list a lot of good, specific ideas.
  • You present reasons you could be wrong -- always a good call!

What I'd recommend changing:

  • You currently don't use any headings, which means you don't get the automatic table of contents that headings would generate. I recommend changing all your section titles to H2, or at least rendering them in bold (the Forum would then treat them as headings).

A few thoughts on the piece itself:

  • The share of funding directed at MB work may be a bit deceptive, as the main form of MB in EA is very cheap (running local groups, particularly student groups). I expect that a much higher percentage of EA hours goes toward MB than EA funding.
  • MB likely gets less attention from EA media (e.g. the 80,000 Hours podcast, the EA Newsletter) because it's not very interesting to broad audiences that aren't already invested in EA. The podcast has (I think) over 10,000 regular listeners, and the newsletter has over 10,000 regular readers. I expect that the majority of both audiences would not describe themselves as "active in the EA movement".
  • I think that many donors have much more extreme views on the relative value of charities/cause areas than in the example you present. For example, someone might think that the Long-Term Future Fund generates 100x the value/dollar as the Global Development Fund, such that trying to grow both funds by giving to the Infrastructure Fund isn't too likely to pan out. (I think you'd need something like a 4-5x multiplier on MB funding to get better value than with direct LTFF funding in my example, but I haven't worked out all the math.)
  • Funding non-MB causes may actually be a good way to build the movement. When I won a bunch of money in front of a large audience this year, I declared that I would donate it to GiveWell [EA(p) · GW(p)], partly because I thought GiveWell's work would be interesting to many people in the audience. Had I announced that I would give it to build awareness of something called "effective altruism", those same audience members likely would have been confused. I also have a much easier time pitching EA by discussing ways I've helped people in the developing world (going "meta" generally won't appeal to newcomers).

To take another example, I am personally deeply impressed by the work of Kelsey Piper and Dylan Matthews et al at Future Perfect. But, why did its creation require a grant from the Rockefeller Foundation, which is not explicitly associated with EA? If there was a team at CEA flush with money and looking to spread EA ideas, would Future Perfect or a similar project have been started several years earlier?

If I recall correctly (I might not), Future Perfect launched pretty quickly after Dylan/Ezra Klein came up with the idea. The idea of starting an "EA media outlet" would have seemed very limited by available talent until the Vox folks got involved, and once they did, I think they obtained a Rockefeller grant before much time had passed, removing the need to raise further "EA" money. (I also wonder whether Dylan would have preferred not to fundraise from EA donors/organizations, so that Future Perfect could more easily remain neutral in its coverage of the movement.)

*****

In any case, I also think MB is pretty neglected and agree with the general thrust of your post. That's why I work in the area full-time!

comment by aaronb50 · 2020-12-30T23:55:13.628Z · EA(p) · GW(p)

Thanks so much for the feedback - just edited with the improved formatting. Regarding your thoughts:

  • Point well taken that MB likely receives a higher proportion of hours. However, it still seems plausible that its share of hours is too low; there a lot of people with full time positions dedicated to direct work (though insofar as these people are earning a salary for themselves that they'd have to earn in some position, not all of this time can be thought of as being spent on an EA cause unless we discount their salary from the 'donation' side of things). Also, seems that a reason that the main forms of MB are cheap is because MB isn't well-funded. If AI safety was underfunded, the main forms of AI safety work would be cheap too. 
  • Seems perfectly reasonable that the 80k podcast, etc. should consider entertainment, non-EA engagement, and similar considerations. That said, an unintended consequence might be that people like me get the wrong impression that more experienced folks have concluded that MB isn't super important/neglected/tractable. 
  • Yeah, the 10 vs 4 utils/dollar example might have been misleading and I agree with your point. One thought: perhaps this might be a sort of coordination problem, where it isn't rational for an individual to fund MB in isolation, but everyone would prefer that everyone give more funding to MB if they could coordinate. Haven't given this idea much thought though.
  • Good point that non-MB work has a sort of MB 'externality.' This has to be balanced against the obvious ways that MB helps direct causes.
  • Probably shouldn't have used the Future Perfect example, as it was fallacious to think that funding was its main constraint. Thanks for the correction.  
comment by Alex HT · 2020-12-30T13:13:35.019Z · EA(p) · GW(p)

Thanks for writing this and contributing to the conversation :)

Relatedly, an “efficient market for ideas” hypothesis would suggest that if MB really was important, neglected, and tractable, then other more experienced and influential EAs would have already raised its salience.

I do think the salience of movement building has been raised elsewhere eg:

Having said that, I share the feeling that movement building seems underrated. Given how impactful it seems, I would expect more EAs to want to use their careers to work on movement building.

One resolution to this apparent conflict is that the fraction of people who can be good at movement building long-term might be smaller than it first seems. For lots of the interventions that you suggest, strong social skills and a strong understanding of EA concepts seem important, as well as some general executional or project management ability. Though movement builders don’t necessarily have to be excellent in any of these domains, they have to be at least pretty good at all of them. They also have to be interested enough in all of them to do movement building. This narrows down the pool of people who can work in movement building. 

Another possible reason is that  within the EA community movement building careers are generally seen as less prestigious than more ‘direct’ kinds of work and social incentives play a large role in career choice. For example, some people would be more impressed by someone doing technical AI safety research than by someone building talent pipelines into AI safety, even if the second one has more impact.

Also, as Aaron says, a lot of direct work has helpful movement building effects. 

I also agree with Aaron that looking at funding is a bit complicated with movement building, partly because movement building is probably cheaper than other things, but also that it can be hard to tease apart what's movement building and what's not. 

comment by aaronb50 · 2020-12-31T00:03:08.003Z · EA(p) · GW(p)

Thanks for all those references. Don't know how I missed the 80,000 page on the topic, but that's a pretty big strike against it being ignored. Regarding your second point, I largely agree but there are surely some MB interventions that don't require full-time generalists. For example, message testing and advertising (I assume) can be mostly outsourced with enough money.