Background: I am a Research Fellow at Europe's largest foreign policy think tank. Some of the points may not translate to a non-UK context, though I hope that most do. I'm very keen to hear comments especially from others working in policy and also open to discussing bilaterally.

Instead of funding EA organisations to do EA policy research, we should consider funding existing, well-known think tanks to do this work. Often this might be preferable to doing the work 'in house', through an EA-organisation such as FHI or CSER. 

Examples of the kind of EA policy work we could fund think tanks to do include:

  • Assessing of the likelihood of WW3 (as per Stephen Clare's excellent post)
  • Comparing the merits of different voting systems
  • Researching the most effective policies to reduce air pollution in middle-income countries
  • Making the case for greater foreign aid contributions from developed countries
  • Promoting international, shared AI standards

This is not an original idea. A recent Founders Pledge report recommended funding Carnegie Endowment for International Peace to restart their Track II dialogue between the US and China. However, I think it is pretty rare for EAs to fund non-EA think tanks to do things, and I wanted to make the case from the perspective of a current think tanker. 

The power of funders

One thing that has surprised me since I began working in a think tank is how much influence funders have over the scope and activities of a research project. It is not unusual for funders to be involved in:

  • Setting the parameters of the research
  • Specifying exactly what the research is used for, or whom
  • Recruitment for the project
  • Which organisations/partners are involved
  • The tone of the research, and in some cases even the conclusions/recommendations

Since the EA movement is somewhat cash-rich, and most of the bottlenecks are around time, expertise and capacity, our prior should be that paying experts to do valuable things is likely to be a good use of money. Think tanks seem like a clear case of experts and time which can be bought through funding. 

Of course, being an active funder also takes time and some level of expertise, however it requires less of both compared to doing our own research.

Benefits of working through existing think tanks

Expertise

The most obvious benefit of working with existing think tanks is that they are full of experts who know a lot about issues that are of interest to EAs. This is particularly beneficial if we think that a distinction can be made between technical expertise and the ability to make good decisions. While EAs are good at the latter, they are often not the best at the former. For example, when it comes to US-China conflict (an issue I work on a bit), I would generally trust a well-informed EA to make better policy decisions, because they are likely to have good aims and do cost-benefit analysis well. However, I would not rank EA China experts above other China experts at Brookings or Chatham House in terms of their subject matter or technical expertise. 

This distinction implies that we want to get EAs to a position where they can make decisions and do cost-benefit analysis (e.g. it would be great to have an EA-aligned US Secretary of State, or EA-aligned foreign policy advisers in Beijing), but the technical research on China's hypersonic weapons, the chances of war, Beijing's intentions for Taiwan, US public opinion towards China, etc., does not need to be done by EAs. In fact it is likely to be performed better by experts at the best foreign policy think tanks.

Experts are also aware of other ideas in the space and what is already popular or politically feasible, and may possess useful soft skills, such as knowing how to speak the 'language' of particular policy areas and being connected to prominent individuals and organisations in the space. Again, I have particularly noticed this in foreign policy, where there is a lot of behind-the-scenes networking and discussions among experts (in the UK, a lot of this takes place at grand country houses) - and EAs risk being left out of these conversations if we confine ourselves to our own institutions.

Comms, publicity, advocacy

Think tanks tend to have media and/or advocacy teams, and there is evidence of a revolving door between think tanks and government in many countries, particularly the US.[1] When it comes to comms and influence, non-EA think tanks generally have worse objectives and incentives than EA orgs. For instance, I've noticed that think tanks value media hits, TV appearances and events with VIPs quite highly, even if these things don't translate towards policy impact. 

However, insofar as good comms and advocacy is important for real-world impact, non-EA think tanks tend to be better at comms than EA organisations: they are likely to have more connections in the political arena, a track record with the media and so on. I was invited to speak to the Washington Post and on CNBC about the Ukraine crisis, even though I have almost no expertise in this area, simply because my think tank is generally respected in the field. The same goes for political advocacy: MPs, congressional committees etc. are more likely to listen to well-known think tank brands.

Pooled resources

Think tanks tend to have core funding as well as project-specific funding. If EAs fund think tanks to do specific projects, they will also benefit from this core funding, which is often spent on comms, publicity and advocacy (as outlined above). 

Think tanks are also well-placed to identify and capitalise on linkages between different policy areas, and those which cover a range of issues may be able to promote research outside of usual networks. I have been surprised at how well-known think tanks in London have received funding for research in areas outside of their core focus. For instance, Chatham House has received a lot of climate-related funding, and the Institute for Government (which is normally focused on Whitehall operations and the technocratic side of government) received a lot of funding to look at trade policy after Brexit. In both instances, it is unlikely that the funders thought Chatham or IfG were the best placed experts on those issues, but instead they were looking for well-known institutions as gathering points for work on new issues.

Staffing

Funders can require that they have some say over recruitment, and EA funders could take advantage of this. This has multiple potential benefits. One benefit is that EAs can ensure that the people completing the research are competent and/or aligned, which presumably affects how useful/impactful the final product is. 

Another benefit is that it can be a vehicle for giving EAs valuable work experience in well-known institutions. I have often wondered if EA funders should simply offer to pay for EAs to get work experience in big institutions that might otherwise be hard to get into (e.g. Google, UN, BBC, European Parliament...), and I know there are some fellowship schemes (e.g. a tech one in DC) which aim to offer something like this. If an EA funder has a choice between funding FHI to do work on climate change, and could alternatively fund E3G (a major climate think tank), one benefit of the latter is that it could lead to some EAs having E3G on their CVs, and any associated network benefits.

Downsides of working through non-EA think tanks

Misaligned objectives

I presume this is the main reason why EA funders tend to prefer paying EA organisations to do things - there is the assurance that they are aligned in terms of values and have better incentives to produce research which actually benefits the world. I think this is a good argument for many areas of research - e.g. if you are funding cause prioritisation work, having an EA worldview/framework is very important. However for other areas this alignment seems less important - such as some of the ones listed above, like the chance of US-China conflict, the number of deaths caused by air pollution, or the most effective voting systems.  

Also, as mentioned above, my experience is that funders get a lot of say over how think tanks go about their research, so an active funder could potentially make sure that a non-EA think tank's objectives for a particular project are aligned.

Bad publicity / politicisation

In some areas EAs may be cautious about working with public-facing or political institutions. There are a few risks here. One is that it brings EA into disrepute. For example, funding the Heritage Foundation or Adam Smith Institute (both right wing think tanks) could send the wrong message. Similarly, funding a more neutral think tank to do sensitive work could be risky: e.g. funding Carnegie to work on US-China track II diplomacy (as was recommended in the aforementioned Founders Pledge report) be interpreted by China hawks in DC as being too dovish. 

Another risk is that issues end up in the public space in an unhelpful way. For example, perhaps it's better for certain AI alignment or x-risk debates to remain outside the public sphere, to avoid panic, politicisation, bad policies or bad actors. This seems like a legitimate concern, and the obvious solution is to only fund policy research in areas that don't carry this risk. 

A related risk is that people don't like think tank funders, or may see it as a bit shady. For instance, people don't like it when oil firms or tech companies fund think tanks. If people wanted to discredited EA, funding think tanks could be seen as a case of trying to use money to influence politics. On the other hand, of all the ways to influence politics, funding think tanks research is surely one of the less shady ways. In most countries it is likely to be better perceived than funding candidates, parties or lobbyists directly. Working with reputable think tanks may even increase EA's credibility in the political sphere.

Think tanks (arguably) aren't actually that good at research

Most think tanks have weird incentives. Many are charities, and so they need to demonstrate that they are benefiting the world in some way. But they also need to satisfy and attract funders, who may have selfish or bad objectives. This latter point means that think tanks in general often produce bad research, or research that isn't so useful. For example, I have come across one think tank project where the funders were essentially paying to have meetings with senior politicians, and another where the funders were paying for the think tank to spend the money on services provided by the funders' company (!). 

However, this bad incentive system could work well for EA funders. Since think tanks will essentially prioritise what funders want, an altruistic and effective funder with lots of money can essentially dictate the rules of the game. To put it simply, I think that think tanks are often bad at research because of poor incentives rather than a lack of expertise or bad methodology, and the right funders can fix this.

Final thoughts / caveats...

  • It's very possible that what I am proposing already happens, but just in a more behind-the-scenes way. I suspect that it happens more in DC, where there seems to be a lot of money + ambition for EAs to work in the political ecosystem, than in London or Brussels (or anywhere else).
  • I haven't really compared funding non-EA think tanks to all the other things EAs can fund. I am just assuming that if funding EA policy research through EA institutions is worthwhile, then this also seems worthwhile.
  • I am clearly susceptible to motivated reasoning since I would be exactly the sort of person to benefit if EAs decided to fund think tanks like mine.  On the other hand, maybe it's not such a bad thing to provide a career boost to EAs working in policy/politics, especially if we want more EAs to pursue those routes.

 

  1. ^

    Abelson et al, Think Tanks, Foreign Policy and Geo-Politics, Routledge, 2017

109

19 comments, sorted by Click to highlight new comments since: Today at 1:12 PM
New Comment

Several think tanks in DC are or were funded through Open Phil (e.g. Johns Hopkins Center for Health Security, NTI, CSET, Center for New American Security, Carnegie, Wilson Center ). I think some program funding ended maybe because of it not being worth the money, or the specific person they were funding through these houses left, or because the program could get funding elsewhere and the counterfactual was no longer present, but this is conjecture . I'd take a look through Open Phil's database if you haven't.

Yes, we (Open Phil) have funded, and in some cases continue to fund, many non-EA think tanks, including the six you named and also Brookings, National Academies, Niskanen, Peterson Institute, CGD, CSIS, CISAC, CBPP, RAND, CAP, Perry World House, Urban Institute, Economic Policy Institute, Roosevelt Institute, Dezernat Zukunft, Sightline Institute, and probably a few others I'm forgetting.

I don't know why the original post claimed "it is pretty rare for EAs to fund non-EA think tanks to do things."

Oh cool - thanks, I wasn't aware! I think it is potentially quite rare in the UK, which is my context. London has a lot of big think tanks and I don't know of any that have received EA funding. I agree that 'rare' is an unhelpful (and evidently inaccurate) word, though even in the US, as a proportion of EA policy research funding, am I right that the vast majority would still go to EA orgs?

What EA orgs do you have in mind? I guess this would be policy development at places like GovAI and maybe Rethink Priorities? My guess is that the policy-focused funding for EAish orgs like that is dwarfed by the Open Phil funding for CSET and CHS alone, which IIRC is >$130M so far.

Fwiw, I did some light research (hours not days) a few years ago on the differences between US and European think tanks and the (perhaps out of date) conventional wisdom seemed to be that they play a relatively outsized role in the U.S. (there are various hypotheses for why). So That may be one reason for the US/UK difference (though funders being in the US and many other issues could also be playing a role).

Yes, this is part of the reason I personally haven't prioritized funding European think tanks much, in addition to my grave lack of context on how policy and politics works in the most AI-relevant European countries.

I guess the US is generally more important also.

I'd guess the answer is "no", based in part on the $55m grant to CSET. (Though it's debatable whether CSET is EA vs EA-adjacent vs non-EA, and CSET was basically "incubated" by that grant rather than already having been a well-known think tank.) Also there's lots of funding to places like Nuclear Threat Initiative and the Centre for Global Development.

One could look at OP's grants database to get a more systematic sense of this. (Of course OP doesn't account for all of EA funding, but does account for a big chunk.)

(That apparent error aside, I still thought this was a good post and feel glad you wrote it :) )

Thanks both - this is really interesting and not what I expected. I think in the UK context there is policy-adjacent research being funded for CSER, CLTR and FHI, and a bit for CE, which I count as all in the EA org bucket, whereas I don't know of any EA funding going to non-EA policy think tanks. I had also put CSET more in the EA category, but it's great to hear that things are different in DC and there's a real interest in funding policy think tanks!

Yeah CSET isn't an EA think tank, though a few EAs have worked there over the years.

Interesting post, thanks for writing it!

I'm not very familiar with the inner workings of think tanks, but I think you may be understating one aspect of the bad research consideration: If the incentives are sufficiently screwed up such that these organizations mostly aren't trying to produce good, objective research, then they're probably not doing a good job of teaching their staff to do that, nor selecting for staff that want to do that or are good at that. So you might not be able to get good research out of these institutions by just locally fixing the incentives.

But this depends a lot on how bad the situation is, and how important researcher judgment is for the project. It seems likely that folks at these institutions genuinely know a lot of facts about their topics of expertise, and for some projects that would be more important than e.g. overall judgment about what is important or likely, which seems more heavily affected by bad incentives. But at least the first three of your examples seem like the kind of projects where overall judgment is really important.

Maybe having these people on a advising or on a team with EAs or good forecasters might also help offset this?

Would be curious to hear thoughts on this from more people who've worked at these places. How bad are these problems in practice?

"How bad are these problems in practice?"

At good think tanks, not very.

Yes, I agree, I don't think this is a big problem at the best think tanks, though there are plenty of (generally very ideologically motivated) half-rate think tanks in London and DC.

Any tips on the 'how' of funding EA work at such think tanks?

Reach out to individual researchers and suggest they apply for grants (from SFF, LTFF, etc.)? Reach out as a funder with a specific proposal? Something else?

For longtermist stuff directing think tank folk to the SFF would probably be best, and perhaps helping them think of a proposal that is aligned with longtermist ideas. I don’t think LTFF would fund this kind of thing (based on their payout reports etc).

For animal rights stuff probably the Animal Welfare Fund (or Farmed Animal Funders) might be interested in funding think tank folk outside of Europe / North America.

For other topics I have no good ideas.

 

If you have say £250k+ could just put out a proposal (e.g. on the EA Forum) that you would fund think tanks and share it with people in the EA policy space and see who gets in touch.

It's more helpful to have a good idea of which think tank and what groups there you should talk to for different things, since they have different output and different strengths. Happy to walk through details of how to do this if you have something specific, but yes, those general approaches would work. 

I think reaching out with a proposal (as Founders Pledge did with Carnegie) is probably the best bet, but it would also be worth ensuring think tanks are aware of the existence of the EA funds and that they can just apply for them. E.g. I don't think my think tank knows about them.

I'm no expert in this area but I'm told that European think-tanks are often strapped for cash so that may explain why the funders get so much influence (which is promising for the funder of course but it may not generalize to the US).

Thank you for sharing! This seems like a great idea. As Scronkfinkle states, there is indeed currently some EA funding going to existing think tanks through Open Phil--not sure about other large funders--but this definitely seems like an area where more funding may be highly beneficial.  

I particularly like your distinction between technical expertise and decision-making capability as well as your point that this is a promising avenue for spreading EA ideas. I share your concerns about EAs being overly siloed in EA orgs, and it seems like funding EA work at non-EA think tanks could be a good way to counteract some of those forces.