If I'm interviewing someone for a position my job is to assess their suitability as a potential employee, but if they're my cousin I might be tempted to give them an overly favorable review. Most organizations have Conflict of Interest (CoI) policies that describe how to handle this sort of situation: it's common that someone might have external relationships which lead to duties, interests, or desires that conflict with what's best for their organization.

It's reasonably common for non-profits to publish their CoI policies (Hewlett, Carnegie, Gates). Within effective altruism I do see some of this:

Historically people and organizations within the EA movement have prioritized transparency, and while there's been some shift away from the most enthusiastic versions of this as we've better understood the costs, there are still a lot of benefits. If you're already going to the effort of drafting a policy like this, making it public seems pretty useful:

  • EAs who are concerned about CoIs within the community and are thinking about what norms they might try to influence can see what's already formally in place.

  • Other organizations can reference it in trying to figure out what sort of policy they want.

  • People who are worried a situation can see what policy was (supposed to have been) followed.

On the other hand, many EA organizations don't seem to have public policies. This includes ones that work in community building or grant-making where they seem pretty important. Here are a few I checked:

I'll write asking for policies, and will update this post if I hear anything back.

Disclosure: my wife is a GiveWell board member, former president of GWWC, and works for CEA which is part of EVF. I haven't run this post by her and I don't know her views here. I work at an organization that has received funding from Open Philanthropy.

Comment via: facebook

New to LessWrong?

New Comment
7 comments, sorted by Click to highlight new comments since: Today at 2:31 PM

Note that the majority of public corporations have such policies, often published as part of SEC filings, as well as employee handbooks.   Is there something specific that's bothering you about some policy or another, or just "it's hard to find for some orgs"?  Being hard to find seems like a pretty weak objection, unless you suspect something is actually going wrong at some org.

Relatedly, is CoI really the biggest worry in terms of employee or officer conduct?  How about non-corruption/bribery?   Donor/customer privacy protection?  These are all part of many companies' codes of conduct - what is it that makes CoI come to the front of your concern?

[ Epistemic status: this came out more confrontationally than I intend.  I do find it curious to see a post on a fairly narrow part (policy verbiage in a published document, as opposed to actual behavior) of a narrow problem (CoI, as opposed to more general bad motives or actor/agent problems), and that makes me wonder if something is happening to cause the focus. ]

Is there something specific that's bothering you about some policy or another?

There's been a lot of discussion within EA lately about interpersonal relationships causing conflicts of interest. It's common for an org to employ people who live together or date, or for there to be people with a close relationship at organizations that have some sort of dependency. If handled well this is not necessarily a problem, but some people are concerned it hasn't been handled well.

I'm specifically interested in what approach different organizations have taken to disclosure, especially around romantic relationships.

Ah, I'm more aware of the tech corporate and non-tech non-profit worlds than EA specifically, but I don't think I've ever seen a "disclose to the public" policy around this.  There's often a "disclose to the org and work with HR and Legal" requirement.   I don't think that is enough to quell any distrust about improper dependencies,  I don't know which orgs outside of SBF and his cronies you're thinking about, but for that case, there's no amount of policy documentation that would matter.

Policies do no good for low-maturity orgs where the leaders don't care about such things.  Having a policy is only very weak evidence about "how an org handles it", and you're probably have to look at the specifics of who's making what decisions, and what relationships exist, not what policy says.  That doesn't scale, but to a great extent nothing around governance scales.

I'm more aware of the tech corporate and non-tech non-profit worlds than EA specifically, but I don't think I've ever seen a "disclose to the public" policy around this

Sorry! I'm talking about within-org disclosure, and concern over whether, say, orgs are getting notified about metamour conflicts.

Though some EA orgs do some public disclosure, for extra transparency: GiveWell, GWWC.

Can you tie your concerns to any specific worries from "a lot of discussion within EA lately".  I remain of the opinion that a policy (whether published or just part of the employee handbook) has almost no correlation to whether there is problematic behavior.  I'd love to have ANY specific concern or incident to analyze to see whether there's no policy, no publicly-visible policy, the policy was followed but didn't stop the problematic aspects, the behavior is not covered by policy, or the policy was violated in a way that matters.

I'd strongly bet that it's impossible to find a worry that's discussed within the community, which would be impacted by any policy I've seen so far.

One specific concern is about a grant where the person who would normally be the grant investigator and the head of the recipient charity were metamours. It's not clear whether this relationship was disclosed (internally to the granting org) or whether the investigator recused themself. And it's not clear whether metamour relationships should generally need to be disclosed and/or trigger recusal.

I'd start with the first question (was it actually disclosed internally, and did the investigator recuse themselves), and THEN the bridge question (did policy have any role in the disclosure/recusement, or was it "just" honorable actors), and only afterward the question about whether there is a policy, was it honored, and (if effective) do other places have similar policies.  

Honestly, this thread has gone on too long, and it makes it seem like I'm more opposed than I actually am.  I do look forward to your findings on the topic, and if you can make changes based on the investigation, you absolutely deserve credit.  I'm skeptical of policy-first approaches to things, as opposed to behavior-evaluation-first, but that shouldn't deter you very much.  I apologize if my argumentative nature has caused any problems.