Bad Omens in current EA Governance

Hi Ludwig, thanks for raising some of these issues around governance. I work on the research team at Giving What We Can, and I’m responding here specifically to the claims relating to our work. There are a few factual errors in your post, and other areas I’d like to add additional context on. I’ll touch on:

1. Our recommendations (we do disclosure conflicts of interest). 
2. The Longtermism Fund specifically (payout reports are about to be published).
3. Our relationship with EVF (we set our own strategy, independently fundraise, and have little to do with most organisations under EVF). 

#1 Recommendations

With respect to our recommendations: They are determined by our inclusion criteria which we regularly link to (for example, on our recommended charities page and on every charity page). As outlined in our inclusion criteria, we rely on our trusted evaluators to determine our giving recommendations. Longview Philanthropy and EA Funds are two of the five trusted evaluators we relied on this giving season. We explicitly outline our conflict of interests with both organisations in our trusted evaluators page.

We want to provide the best possible giving recommendations to our donors. Unfortunately, given we are very connected to the effective giving ecosystem — and as you highlighted, part of EVF — this is regularly in tension with avoiding conflicts of interest. We did our best this giving season to highlight these conflicts, and justify why we chose the evaluators we did, but we want to do better next year (we touch on this in our most recent announcement of our new research direction). 
 

#2 The Longtermism Fund

The fund will disclose all of its spending in regular payout reports. Its first report will be released shortly (by the end of today! It’s been in production over the past weeks).
 

As shared in our announcement of the fund, the fund is a collaboration between Giving What We Can and Longview. We (GWWC) are responsible for the communications around the fund; Longview are responsible for the grantmaking and research.

We also publicly committed to sharing reports outlining the funds grants in our announcement of the fund.

#3 Relationship with EVF

Giving What We Can initially helped create EVF’s predecessor (CEA) back in 2011, alongside 80,000 Hours — read more about its history here. In short, EVF currently provides GWWC with:

• Operational support (e.g., finance, legal, HR) via EV Ops. 
• Board of Trustees (of which each organisation has historical had its own “Active Trustee” who has worked closely with the respective organisation’s leader on strategy and management). 
• Shared privacy policy (this facilitates a single sign-on for GWWC, EA Forum and EA Global). 
• Some limited shared communications and facilities (e.g., some shared Slack channels, Notion spaces, and access to Trajan House—though nobody at GWWC currently uses this).
   

Importantly, GWWC independently:

• Fundraises for its core expenses (i.e., we independently seek funding to pay for our staff and costs). 
• Sets its own strategy (we work as a team consulting GWWC members and other stakeholders to decide how we can have the most impact), does its own hiring, etc.  See our most recent strategy update  where we were seeking community feedback on our plans.
• Independently choose its approach to giving recommendations — we receive no benefit for recommending organisations within EVF; historically, we err on the side of avoiding this due to perceived/potential conflicts of interest). 

Happy to clarify any of the above.

I think some of this post's criticisms have bite: for example, I agree that EVF suborgs are at significant risk of falling prey to conflicts of interest, especially given the relatively low level of transparency at many of these suborgs, and that EVF should have explicit mechanisms for avoiding this.

However, I think this post largely fails to engage with the reasons so many suborgs have federated with EVF. Based on my experience^[1], members of many of these suborgs genuinely consider themselves separate orgs, and form part of EVF mainly because this allows them to be supported by EVF's very well-oiled ops machine. This makes it significantly easier for new EA projects to spin up quickly, while offering high-quality HR and other support to their employees. This is a pretty exciting proposal for any new EA project that doesn't place a high value on being legally independent.

"Breaking up" EVF could thus be very costly from an impact perspective, insofar as it makes the component orgs less effective (which seems likely to me) and necessitates lots of duplication of ops effort. You might argue that it's worth it for the transparency benefits, but I'd want to see serious engagement with the costs, and an attempt to explore promising alternatives, before I felt comfortable endorsing this kind of proposal.

1. ^{^}

   I've never worked for an EVF org, but have close personal ties to more than one. The orgs I'm affiliated with also collaborate closely with EVF ops on some projects. This is a potential source of bias, but also means I have a lot more information than some other commenters here.

Thanks Ludwig for raising the conversation around governance. It's something that is important to me and we're exploring how we can improve on this (and will share later any changes we plan to make). Michael has just left another comment which covers most of what I'd say right now so I don't have much more to add other than:

1. I'd to flag that I don't believe this comment is accurate, and seems very uncharitable:

I believe this structure was set up so the EVF board has central control over EA strategy.

CEA was setup before there was [added:much of] an EA movement (the term "effective altruism" was invented while setting up CEA to support GWWC/80,000 Hours). In recent years, several organisations have approached EVF so they can receive the same kind of operational and legal support. Some of these organisations have met EVF's bar for impact and thus been supported, and I'm aware some are in the process of spinning out of EVF after receiving initial support getting started.

2. In my experience leading GWWC for the past 2.5 years the EVF trustees have never "exerted influence" over our strategy.

During this time I have received helpful input from trustees (mostly working with Toby as our active trustee for 2 years, and with Toby and Will as founders, some advice from Nick on hiring researchers) from time to time. I am in regular contact with other organisation leaders inside of EVF as well as those outside of (in fact, I have much more contact and consultation with leaders of other effective giving organisations than I do with most other leaders of organisations within EVF or our board). My primary day-to-to-day contact is with my staff, donors, members and funders all who give a lot of input on the day-to-day of what we're doing at GWWC.

To be frank, the hardest thing about having so many organisations as part of one charity is that it's harder for a board to give lots of attention to the strategy of each organisation within it. It would be much easier (and IMHO more appropriate) to be critiquing the governance from the angle that not enough attention is given  (due to a small number of busy people who's expertise isn't governance being spread too thin) rather than too much (in the form of exerting influence).

3. What exactly do you mean/have in mind by "GWWC members should demand control over their institution"?

I'm not sure what you mean by this and would be interested to know more. 

Hi — I think this post overstates the level of program-level centralisation here. 

The EVF and CEA US boards provide overall oversight and governance of the projects and their executive directors, and will occasionally step in to change something important. But they have largely delegated program-level responsibility to each project’s executive director, who each set their own strategy for how to best have a positive impact on the world.

In practice, those strategies do differ: to give a couple of examples, CEA and 80,000 Hours have pretty different approaches to cause prioritisation; while Asterisk (a CEA US project) published a very critical review of What We Owe the Future (a book written by an EVF project lead and board member). The boards also very much want EA work to flourish outside of the EVF and CEA US governance structures — many of the grants made by the EA Infrastructure Fund support this work.

EVF's "EA Funds"-branded funds publicly disclose their spending, which is commendable!

This is, alas, false.

Even for the subset of EA Funds spending that is public^[1], 3/4 funds (1,2,3) haven't published a grant report for roughly the past year. ETA: This is incorrect/misleading. EA Funds isn't publishing grant reports anymore because they now have a public grant database. I support this change and regret the error. 

ETA: The potential for conflicts of interest with intra-EVF grantmaking isn't a problem I'd previously thought about here, but seems like an additional reason why EA Funds should not be making secret grants (or at least, why it should commit to not making secret grants within EVF).

1. ^{^}

   I believe the exact proportion that is public vs secret has not been disclosed, though would be happy to learn I was wrong about that.

With regards to effektiveraltruismus.de: The site has just been transferred to "Effektiver Altruismus Deutschland (EAD) e.V." (can be seen in the Imprint/Impressum already and will be mentioned in a newsletter that we will probably send later today). Donations will still be managed by Effektiv Spenden (officially knows as "UES – Gemeinnützige GmbH für effektives Spenden") since EAD can't do that at the moment (from a legal perspective and also from a technical/operations perspective). We already mention who is handling the donations on top of the donation form though. Happy to consider further clarification on the website (or elsewhere). Other than that I don't really see why the arrangement should be changed since it would lead to unnecessary overhead but happy to discuss.

Disclaimer: I'm the Co-Founder and Co-CEO of Effektiv Spenden.

I'd like to bring up a further question: What would happen to 80000hours' budget if CEA were to go bankrupt? What would happen to donated money being held by EVF through EA Funds?

Setting up your own legal entity is a lot of work. When we where setting up AI Safety Support I was surprised that we could not find anyone to offer EVF-type services to us. After we got our legal stuff set up, AI Safety Support is not offering this service to other orgs. 

Asking that every org is it's own legal entity is very costly and also would not solve any of the problems around power centralisation. I think this post is brining up a good point.

On https://www.effectivealtruism.org/, which serves as an introduction to EA, the EVF links to its own project 80000hours, but not to the competing Probably Good.

This is problematic, but I don't think it would be solved by forcing 80k to be a separate legal org. Both the fact that CEA signal boost 80k and not PG, and the fact that CEA shares ops team with 80k and not PG, is downstream from who is friends with whom. 

If I had anything to say in this, I would want EVF to offer their services to more orgs, not fewer.

I'm very thankful for EVF and associated orgs, and as referenced by others, it's understandable how/why the community is currently organized this way. Eventually, depending on growth and other factors, it'll probably make sense for the various subs to legally spin off, but I'm not sure if this is high priority - it depends on just how worried EAs are about governance in the wake of this past month. 

I will say, conflict of interest disclosures are important but seems like they may be doing a lot of work here.  As far as I can tell^[1], leadership within these organizations also function independently and they're particularly aware of bias as EAs so they've built processes to mitigate this. But being aware of bias and disclosing it doesn't necessarily stop [trustworthy] people from being biased (see: doctors prescribing drugs from companies that pay for talks.) Even if these organizations separated tomorrow, I'd half expect them to be in relative lock-step for years to come. Even if these orgs never shared funding/leadership again, they're in the same community, they'll have funders in common, they'll want to impress the same people, so they'll make decisions with this in mind. I've seen this first-hand in every [non-EA] org I've ever been a part of, in sectors of all sizes, so moving forward we'll have to build with this bug in mind and decide just how much mitigation is worth doing.

I'm aware that none of this is original or ground-breaking but perhaps worth reiterating.

1. ^{^}

   This is a little facetious, but does anyone else find themselves caveating more often these days, just in case...

Who sits on the board of the Effective Ventures Foundation? 

(Typo)

In your link here:

On https://www.effectivealtruism.org/, which serves as an introduction to EA, the EVF links to its own project 80000hours, but not to the competing Probably Good.

The comma is included, which makes the page go to an error site.

DFTBA :)