This post has two parts: Part 1 outlines a type of donation that I think donors aiming for high expected impact should consider making: one-off donations to young organisations with highly uncertain futures. Part 2 is a case study that falls into this category, with my reasons for donating $3,000 to the Environmental Data & Governance Initiative (EDGI) this year.

Part 1 mainly provides context for part 2; you can skip/skim it without losing much. My motivation for writing this post is more to do with EDGI and with part 2: (I) EDGI is “off the beaten track” of where many effective altruists look for donation opportunities, and (II) I am aware of some people looking for donation opportunities related to the outcome of the US presidential election who may find this useful.

I hope that others investigate EDGI and related organisations further, and that this post acts as a pointer. The window for high impact donations in this area will close soon: Trump assumes office on January 20th, and the relevant federal departments will soon have new leaders with new priorities. If you want to investigate further, do it now!

Part 1: In favour of donating to young, promising organisations¹

I believe that the places I can make the highest value donations to are likely to be

(1) new organisations/projects, with an uncertain future, currently bottlenecked on funding, that

(2) if they worked out would be hugely impactful, where

(3) I have some information about the organisation that other donors who think in similar ways to me (e.g. the Open Philanthropy Project, effective altruist donors) don’t have, and that is difficult to communicate quickly or in sufficient depth (e.g. due to drawing on large amounts of specific background knowledge, or in-network trust that is non-transferrable).

Why (1)?

First, conventional donors - e.g. grant-making foundations - often avoid donating to new organisations/projects for reasons that don’t apply to me, so there may be relatively good opportunities left unfunded. For one, foundations often have a fixed cost of organisational overhead for each grant they make, so small grants may not be worth it for them. For two, my impression is that many foundations and individual donors are more risk-averse than I aim to be.

Second, the impact of donations to the long term trajectory of young organisations is often larger than to more established organisations, since the organisation’s very existence is on the line. There is also significant path dependency in how things turn out at organisations as a result of decisions made early on about initial hires and what to focus on, so providing nudges in directions you think are promising early on can have a larger impact than nudges provided later. Early donations give you some nudging power.

Third, lots of larger organisations I'd consider donating to will likely attract significant funding from the Open Philanthropy Project, since Open Phil’s values seem pretty in line with mine and they have a large amount of funding to disperse. Though this hasn’t stopped me from donating to larger organisations entirely, it has reduced the amount I donate, and I expect this will continue in the future.

(2) speaks for itself.

Why (3)?

I think it’s often hard to summarise judgments of how promising a young organisation or project is in a way that will provide enough information for other donors to make a decision. This is because young organisations have fewer objective indicators to go off (e.g. track record), so the decision mostly comes down to subjective and qualitative judgments about the project - to do with how effective you think the team is, whether the vision makes sense in the context of different pieces of background knowledge you have, and so on. These judgments also often draw on the judgments of other people you trust who are even closer to the action, in a way that is hard to transfer to other donors (since there are good reasons for this sort of trust not to extend much beyond one degree).

For this reason, I think donors who have thought carefully about an organisation or project and done a sufficient amount of research on it should be prepared to donate to things they have high conviction around but which don’t have the seal of approval of other donors, especially if timing is a significant factor in the donation.

I think this should be paired with as much information-sharing as is feasible: sharing donation ideas improves the idea set for other donors, even if you can’t convey all the information and judgment calls well; and over time, by sharing results/learning from how these donations turn out, donors will get a sense of how much they can rely on each other’s judgment.

Ways this approach may fail

The clearest way I can see this strategy leading to worse donations is if these donations are overly speculative, and the organisations rarely turn out to be excellent. It’s hard to predict the future success of small organisations, and it’s not clear individual donors can do it well enough to justify the gamble.²

Effective altruists to date have placed a great emphasis on evidence when making donation decisions, such that “effective” has almost become synonymous with “evidence-based”. Almost by definition this form of donation can’t meet that standard, since it’s betting on organisations and projects with little or no track record. The decision-making process also relies more on heuristics, gut feel, and system 1 judgments, all of which effective altruists have shown often to lead to bad judgments when used to pick between charities.

For these reasons it’s not clear to me that it would be better if lots of donors followed this approach. My guess is that more effective altruists should than currently do, though, and that this is a result partly of groupthink around donations and partly of fear of wasting money on things that don’t pan out. You should do it whenever you have particularly strong reasons to believe an organisation may be hugely impactful in the future, since the risk of you getting it wrong is less bad than missing out on a great donation opportunity.

Part 2: Donating to EDGI

Summary of EDGI’s activities

My largest donation for 2016 was $3,000 to the Environmental Data & Governance Initiative (EDGI). EDGI is an initiative founded in response to Trump’s election to preserve data and institutional knowledge at environmental and scientific institutions that are under threat of being defunded or de-legitimised by the new administration (e.g. the Environmental Protection Agency, Department of Energy, NASA: Earth Science). These organisations have spent large amounts of money over the course of decades collecting data and building systems that have helped us understand climate change; if the datasets or organisations are dismantled by the Trump administration, we risk losing a lot of the value from decades of work.

EDGI has tens of volunteers collaborating from a variety of different organisations that pre-date Trump’s election, including 314Action, Data Refuge, and the End of Term Web Archive. They received some early press coverage in December for their data archiving hackathons, and are continuing these “defensive” activities of speedily archiving important datasets. They are starting to work on “offense” tactics too: building web tools that track changes to government websites to see when funding, data, or staffing change, to mobilise the public around particularly worrying changes as they happen.

In the short term, EDGI is working on a problem that is time-sensitive and may prove useful (archiving datasets). In the medium-to-long term I believe EDGI could become a highly important organisation, devoted more broadly to protecting scientific institutional knowledge in the US government over the next 4-8 years under hostile conditions, and documenting changes carefully so that the public is more aware of what is worth caring about and fighting back on. EDGI is seeking donations; at the time of writing, there is a $23k funding gap left to their first target ($31k).

The political context behind EDGI as a donation opportunity

Broadly speaking, I believe that we should take risks posed by the new administration seriously: in order to reach a long term flourishing future for the world we need to “keep the show on the road”, and avoid any large derailments to human progress that would reduce our ability to deal with a range of threats and opportunities that face humanity in the future. I believe that, more likely than not, there won’t be a significant derailment in the next 4-8 years, but there is a high enough chance that we should guard against it.

Losing scientific institutional knowledge probably isn’t the largest issue to be wary of - nuclear war and creeping authoritarianism in the US rank higher, in my opinion. My sense is that these issues are already receiving a lot of attention, and may be harder to impact with donations; by contrast science is an area receiving less attention where there may be quick wins to be had cheaply, with large lasting benefits.

EDGI as a case study of the framework from part 1

EDGI ticks each of the three criteria above well. It has the bonus feature of being (0) excellently timed, which I think in this case is the most important factor in favour of the donation.

(0): Timing of the intervention: There are some actions the Trump administration may take quickly (e.g. in their first 100 days) related to institutions like the EPA, DoE, and Office of Science and Technology Policy, that would have large long term negative effects. As a result, there is a window of opportunity to impact some of these issues that will close relatively soon. EDGI is working on relevant counter-projects within this window: if they ramp up a few months too late due to lack funding, the window for impact may be closed. (In the other direction: I think this area would have been unimportant to donate to 3 months ago.)

I think it is important to take timing seriously when trying to figure out what projects and organisations end up impacting the world in the largest ways:³

• The importance of timing has been a theme in various of Open Phil’s investigations into the History of Philanthropy. The closest parallel to EDGI in their investigations is the founding of the Center on Budget and Policy Priorities in 1981 in response to Reagan’s election, to provide evidence-based policy recommendations related to poverty in the US.
• In the world of for-profit startups, timing may be the most important factor for success.⁴

(1): New, uncertain, bottlenecked on funds: EDGI has tens of energetic volunteers but little funding, as it was formed shortly after the US presidential election. Though it has received some press attention already, I believe it is working on a relatively neglected and technical issue that has received less funding so far than it should, as other worries about Trump’s administration are (often rightly) getting more press attention. EDGI is currently fundraising for $31,415, and I believe will be trying to raise significantly more money in the next few months when they have time to write grant applications.

(2): High impact if it works out:

• I believe that US government investment in science has historically had a large positive impact on the world, and that US government investment in studying climate change has led to a better understanding of the issue as a global political priority.
• I believe that maintaining high quality datasets with no gaps in them is important for continuing to have impact at this scale in the future, and for maintaining public trust in science and evidence-based policy. I also believe that the people currently at these institutions and the institutional knowledge they hold collectively are important, probably more than the data: if a large number of the leaders of (e.g.) the EPA disband and move on to other jobs over the course of the next four years, the discontinuity in the organisation will lead to a large loss of value from the ideas, knowledge, and priorities in the heads of the people walking out the door.
• I believe there is a significant chance that existing government institutions doing important work related to basic science and the environment will face large funding cuts under the new administration, and/or become ineffective through having leaders installed that think the output of the organisations isn’t valuable (as has already happened at the Department of Energy and the Environmental Potection Agency).
• As a result, I believe that if EDGI executes well on their short terms goals of archiving data and tracking policy/funding changes on government websites, and their medium-to-long term goals of preserving institutional knowledge and mobilising public support, they will have a large positive impact on the world.

(3): Information I have that others don’t and is hard to communicate quickly:  After the US presidential election two months ago I spent a couple of weekends looking into areas to focus time or donations, and discussing this with friends. The friend I discussed things with most decided that he would devote significant time to investigating what to work on, and consider taking a leave of absence from his PhD program if he found something promising enough. By mid-December, from everything he'd seen he felt that EDGI was a strong bet. This is what put EDGI on my radar, and also what helped me get high conviction around the donation, since I trust his judgment in a way that allowed me to outsource a lot of the investigation of the cause area. This was a case of abnormally high levels of trust, that wouldn't occur often for me: I have spent a lot of time discussing ethics and politics with this friend in the past and have a good sense of (and respect for) his views, and he is a physicist so has better knowledge of how the US government interacts with science than I do.

Reservations about the donation

My reservations about donating to EDGI are mostly instances of my reservations about the donation structure from part 1: due to being a young organisation with no track record, it’s hard to predict how successful EDGI will be, and I may be bad at these sorts of predictions. Here are a few specific routes that could happen, that would mean the donation was not high impact:

1. EDGI’s leadership turns out to be ineffective or uncommitted over the long term, or the organisation ends up structured in a way that leads to poor decision-making or slow action.
2. The mission of the organisation drifts as it becomes more established, into something that (I think) is not as valuable.
3. The Trump administration moves quickly enough that EDGI misses the window of opportunity for action.
4. EDGI is unable to get press attention at the right times, or is unable to reach decision-makers.

Next steps for investigation

It would be useful for someone with a better understanding of the organisational landscape of US environmental/scientific government bodies to talk to current administrators and get a sense of where the largest threats lie. Then they could check whether EDGI’s plans are targeting the most important stuff - and if not, point them in a better direction.

 

Thanks to Ethan Barhydt and Andrew Bergman for comments on a draft of this post.

-----

1 For related ideas that have been nicely written up already, I’d recommend Open Phil’s post on high risk, hits-based giving in the context of foundation grants, and 80,000 Hours’ arguments for taking a growth approach to evaluating non-profit startups.

2 A good comparison point here would be angel investor returns in for-profit startups relative to returns for later stage investors, but I haven’t been able to find very robust data on this. If anyone has looked into this more I’d love to hear about it.

3 Maybe I should add “timing” as a fourth component of the framework in part 1, since I'm pretty sure it's generally important rather than just in this case. I haven’t thought about whether it’s as important as (1)-(3) in the general case, though, so I’ve left it out for now - but it may well be.

4 I find this claim hard to evaluate, as I think it’s likely you can’t predict how well something ranks on “timing” much beforehand, but only in hindsight - in which case it would convey a clearer message to say “luck is the most important factor in success”. However, I think there’s a strong argument that in the non-profit case, political timing does affect donation opportunities, in a way that’s predictive of impact ahead of time rather than just in hindsight.