Nonprofit Boards are Weird

post by Holden Karnofsky (HoldenKarnofsky) · 2022-06-28T10:17:52.343Z · EA · GW · 4 comments

This is a link post for


  Why nonprofit boards are weird
    Great power, low engagement, unclear responsibility, no accountability
    What makes a good board member? Few people even have a hypothesis
  So what DOES make a good board member?
    The board's main duties
    Engaging on main duties, staying out of the way otherwise
    Who should be on the board?
    A few practices that seem good

Disclaimer: This post was cross-posted with the author's permission, under his account. He may not notice comments. It was originally posted on June 23. You can find the audio version here

Note: anything in this post that you think is me subtweeting your organization is actually about, like, at least 3 organizations. (I'm currently on 4 boards in addition to Open Philanthropy's; I've served on a bunch of other boards in the past; and more than half of my takes on boards are not based on any of this, but rather on my interactions with boards I'm not on via the many grants made by Open Philanthropy.)

Writing about ideal governance reminded me of how weird my experiences with nonprofit boards (as in "board of directors" - the set of people who formally control a nonprofit) have been.

I thought that was a pretty good intro. The rest of this piece will:

I am experienced with nonprofit boards but not with for-profit boards. I'm guessing that roughly half the things I say below will apply to for-profit boards, and that for-profit boards are roughly half as weird overall (so still quite weird), but I haven't put much effort into disentangling these things; I'm writing about what I've seen.

I can't really give real-life examples here (for reasons I think will be pretty clear) so this is just going to be me opining in the abstract.

Why nonprofit boards are weird

Here's how a nonprofit board works:

In my experience, it's common for the whole thing to feel extremely weird. (This doesn't necessarily mean there's a better way to do it - footnote has more on what I mean by "weird."2)

(Reminder that this is not subtweeting a particular organization! More than one person - from more than one organization - read a draft and thought I was subtweeting them, because what's above describes a large number of boards.)

OK, so what's driving the weirdness?

I think there are a couple of things:

I'll take these one at a time.

Great power, low engagement, unclear responsibility, no accountability

In my experience/impression, the best way to run any organization (or project, or anything) is on an "ownership" model: for any given thing X that you want done well, you have one person who "owns" X. The "owner" of X has:

When these things come apart, I think you get problems. In a nutshell - when no one is responsible, nothing gets done; when someone is responsible but doesn't have power, that doesn't help much; when the person who is responsible + empowered isn't engaged (isn't paying much attention), or isn't held accountable, there's not much in the way of their doing a dreadful job.

A traditional company structure mostly does well at this. The CEO has power (they make decisions for the company), engagement (they are devoted to the company and spend tons of time on it), and responsibility+accountability (if the company does badly, everyone looks at the CEO). They manage a team of people who have power+engagement+responsibility+accountability for some aspect of the company; each of those people manage people with power+engagement+responsibility+accountability for some smaller piece; etc.

What about the board?

So we have people who are spending very little time on the company, know very little about it, don't have much clarity on what they're responsible for either individually or collectively, and aren't accountable to anyone ... and those are the people with all of the power. Sound dysfunctional?4

In practice, I think it's often worse than it sounds, because board members aren't even chosen carefully - a lot of the time, a nonprofit just goes with an assortment of random famous people, big donors, etc.

What makes a good board member? Few people even have a hypothesis

I've searched a fair amount for books, papers, etc. that give convincing and/or widely-accepted answers to questions like:

In my experience, most board members just aren't walking around with any particular thought-through take on questions like this. And as far as I can tell, there's a shortage of good5 guidance on questions like this for both for-profit and nonprofit boards. For example:

To the extent I have seen a relatively common, coherent vision of "what board members are supposed to be doing," it's pretty well summarized in Reid Hoffman's interview in The High-Growth Handbook:

I use ... a red light, yellow light, green light framework between the board and the CEO. Roughly, green light is, “You’re the CEO. Make the call. We’re advisory.” Now, we may say that on very big things—selling the company—we should talk about it before you do it. And that may shift us from green light, if we don’t like the conversation. But a classic young, idiot board member will say, “Well, I’m giving you my expertise and advice. You should do X, Y, Z.” But the right framework for board members is: You’re the CEO. You make the call. We’re advisory.

Red lights also very easy. Once you get to red light, the CEO—who, by the way, may still be in place—won’t be the CEO in the future. The board knows they need a new CEO. It may be with the CEO’s knowledge, or without it. Obviously, it’s better if it’s collaborative ...

Yellow means, “I have a question about the CEO. Should we be at green light or not?” And what happens, again under inexperienced or bad board members, is they check a CEO into yellow indefinitely. They go, “Well, I’m not sure…” The important thing with yellow light is that you 1) coherently agree on it as a board and 2) coherently agree on what the exit conditions are. What is the limited amount of time that we’re going to be in yellow while we consider whether we move back to green or move to red? And how do we do that, so that we do not operate for a long time on yellow? Because with yellow light, you’re essentially hamstringing the CEO and hamstringing the company. It’s your obligation as a board to figure that out.

I like this quite a bit (hence the long blockquote), but I don't think it covers everything. The board is mostly there to oversee the CEO, and they should mostly be advisory when they're happy with the CEO. But I think there are things they ought to be actively thinking about and engaging in even during "green light."

So what DOES make a good board member?

Here is my current take, based on a combination of (a) my thoughts after serving on and interacting with a large number of nonprofit boards; (b) my attempts to adapt conventional wisdom about for-profit boards (especially from the book I mentioned above); (c) divine revelation.

I'll go through:

(I don't claim any of these points are original, and almost everything can be found in some writing on boards somewhere, but I don't know of a reasonably comprehensive, concise place to get something similar to the below.)

The board's main duties

I agree with the basic spirit of Hoffman's philosophy above: the board should not be trying to "run the company" (they're too low-engagement and don't know enough about it), and should instead be focused on a small number of big-picture questions like "How is the CEO doing?"

And I do think the board's #1 and most fundamental job is evaluating the CEO's performance. The board is the only reliable source of accountability for the CEO - even more so at a nonprofit than a for-profit, since bad CEO performance won't necessarily show up via financial problems or unhappy shareholders.6 (As noted below, I think many nonprofit boards have no formal process for reviewing the CEO's performance, and the ones that do often have a lightweight/underwhelming one.)

But I think the board also needs to take a leading role - and not trust the judgment of the CEO and other staff - when it comes to:

Engaging on main duties, staying out of the way otherwise

I think the ideal board member's behavior is roughly along the lines of the following:

Actively, intensively engage in the main duties from the previous section. Board members should be knowledgeable about, and not defer to the CEO on, (a) how the CEO is performing; (b) how the board is performing, and who should be added and removed; (c) spotting (and scanning the horizon for) events that could reduce the board's powers, or lead to big enough problems and restrictions so as to irreversibly affect what future CEOs are able to do.

Ideally they should be focusing their questions in board meetings on these things, as well as having some way of gathering information about them that doesn't just rely on hearing directly from the CEO. (Some ideas for this are below.) When reviewing financial statements and budgets, they should be focused mostly on the risk of major irreversible problems (such as going bankrupt or failing to be compliant); when hearing about activities, they should be focused mostly on what they reflect about the CEO's performance; etc.

Be advisory ("stay out of the way") otherwise. Meetings might contain all sorts of updates and requests for reactions. I think a good template for a board member, when sharing an opinion or reaction, is either to (a) explain as they're talking why this topic is important for the board's main duties; or (b) say (or imply) something like "I'm curious / offering an opinion about ___, but if this isn't helpful, please ignore it, and please don't hesitate to move the meeting to the next topic as soon as this stops feeling productive."

The combination of intense engagement on core duties and "staying out of the way" otherwise can make this a very weird role. An organization will often go years without any serious questions about the CEO's performance or other matters involving core duties. So a board member ought to be ready to quietly nod along and stay out of the way for very long stretches of time, while being ready to get seriously involved and engaged when this makes sense.

Aim for division of labor. I think a major problem with nonprofit boards is that, by default, it's really unclear which board member is responsible for what. I think it's a good idea for board members to explicitly settle this via assigning:

This can further help everyone find a balance between engaging and staying out of the way.

Who should be on the board?

One answer is that it should be whoever can do well at the duties outlined above - both in terms of substance (can they accurately evaluate the CEO's performance, identify big-picture irreversible risks, etc.?) and in terms of style (do they actively engage on their main duties and stay out of the way otherwise?)

But to make things a bit more boiled-down and concrete, I think perhaps the most important test for a board member is: they'll get the CEO replaced if this would be good for the nonprofit's mission, and they won't if it wouldn't be.

This is the most essential function of the board, and it implies a bunch of things about who makes a good board member:

In my experience, most nonprofits are not looking for these qualities in board members. They are, instead, often looking for things like:

I think a good profile for a board member is someone who cares greatly about the nonprofit's mission, and wants it to succeed, to the point where they're ready to have tough conversations if they see the CEO falling short. Examples of such people might be major funders, or major stakeholders (e.g., a community leader from a community of people the nonprofit is trying to help).

A few practices that seem good

I'll anticlimactically close with a few practices that seem helpful to me. These are mostly pretty generic practices, useful for both for-profit and nonprofit boards, that I have seen working in practice but also seen too many boards going without. They don't fully address the weirdnesses discussed above (especially the stuff specific to nonprofit as opposed to for-profit boards), but they seem to make things some amount better.

Keeping it simple for low-stakes organizations. If a nonprofit is a year old and has 3 employees, it probably shouldn't be investing a ton of its energy in having a great board (especially since this is hard).

A key question is: "If the board just stays checked out and doesn't hold the CEO accountable, what's the worst thing that can happen?" If the answer is something like "The nonprofit's relatively modest budget is badly spent," then it might not be worth a huge investment in building a great board (and in taking some of the measures listed below). Early-stage nonprofits often have a board consisting of 2-3 people the founder trusts a lot (ideally in a "you'd fire me if it were the right thing to do" sense rather than in a "you've always got my back" sense), which seems fine. The rest of these ideas are for when the stakes are higher.

Formal board-staff communication channels. A very common problem I see is that:

I've seen this dynamic improved some amount by things like a staff liaison: a board member who is designated with the duty, "Talk to employees a lot, offer them confidentiality as requested, try to build trust, and gather information about how things are going." Things like regular "office hours" and showing up to company events can help with this.

Viewing board seats as limited. It seems unlikely that a board should have more than 10 members (and even 10 seems like a lot), since it's hard to have a productive meeting past that point.11 When considering a new addition to the board, I think the board should be asking something much closer to "Is this one of the 10 best people in the world to sit on this board?" than to "Is this person fine?"

Regular CEO reviews. Many nonprofits don't seem to have any formal, regular process for reviewing the CEO's performance; I think it's important to do this.

The most common format I've seen is something like: one board member interviews the CEO's direct reports, and perhaps some other people throughout the company, and integrates this with information about the organization's overall progress and accomplishments (often presented by the organization itself, but they might ask questions about it) to provide a report on what the CEO is doing well and could do better. I think this approach has a lot of limitations - staff are often hesitant to be forthcoming with a board member (even when promised anonymity), and the board member often lacks a lot of key information - but even with those issues, it tends to be a useful exercise.

Closed sessions. I think it's important for the board to have "closed sessions" where board members can talk frankly without the CEO, other employees, etc. hearing. I think a common mistake is to ask "Does anyone want the closed session today or can we skip it?" - this puts the onus on board members to say "Yes, I would like a closed session," which then implies they have something negative to say. I think it's better for whoever's running the meetings to identify logical closed sessions (e.g., "The board minus employees"), allocate time for them and force them to happen.

Regular board reviews. It seems like it would be a good idea for board members to regularly assess each other's performance, and the performance of the board as a whole. But I've actually seen very little of this done in practice and I can't point to versions of it that seem to have some track record of working well. It does seem like a good idea though!


The board is the only body at a nonprofit that can hold the CEO accountable to accomplishing the mission. I broadly feel like most nonprofit boards just aren't very well-suited to this duty, or necessarily to much of anything. It's an inherently weird structure that seems difficult to make work.

I wish someone would do a great job studying and laying out how nonprofit boards should be assembled, how they should do their job and how they can be held accountable. You can think of this post as my quick, informal shot at that.

 Comment/discuss [LW · GW]



  1. I'm using the term "CEO" throughout, although the chief executive at a non profit sometimes has another title, such as "Executive Director."  [LW(p) · GW(p)]
  2. A lot of this piece is about how the fundamental setup of a nonprofit board leads to the kinds of problems and dynamics I'm describing. This doesn't mean we should necessarily think there's any way to fix it or any better alternative. It just means that this setup seems to bring a lot of friction points and challenges that most relationships between supervisor-and-supervised don't seem to have, which can make the experience of interacting with a board feel vaguely unlike what we're used to in other contexts, or "weird."  People who have interacted with tons of boards might get so used to these dynamics that they no longer feel weird. I haven't reached that point yet myself though. [LW(p) · GW(p)]
  3. The fact that the nonprofit's goals aren't clearly defined and have no clear metric (and often aren't susceptible to measurement at all) is a pretty general challenge of nonprofits, but I think it especially shows up for a structure (the board) that is already weird in the various other ways I'm describing.  [LW(p) · GW(p)]
  4. Superficially, you could make most of the same complaints about shareholders of a for-profit company. But:
    • Shareholders are the people who ultimately make or lose money if the company does well or poorly (you can think of this as a form of accountability). By contrast, nonprofit board members often have very little (or only an idiosyncratic) personal connection to and investment in the organization.
    • Shareholders compensate for their low engagement by picking representatives (a board) whom they can hold accountable for the company's performance. Nonprofit board members are the representatives, and aren't accountable to anyone.  [LW(p) · GW(p)]
  5. Especially "good and concise." Most of the points I make here can be found in some writings on boards somewhere, but it's hard to find sensible-seeming and comprehensive discussions of what the board should be doing and who should be on it.  [LW(p) · GW(p)]
  6. Part of the CEO's job is fundraising, and if they do a bad job of this, it's going to be obvious. But that's only part of the job. At a nonprofit, a CEO could easily be bringing in plenty of money and just doing a horrible job at the mission - and if the board isn't able to learn this and act on it, it seems like very bad news.  [LW(p) · GW(p)]
  7. The charter and bylaws are like the "constitution" of a nonprofit, laying out how its governance works.  [LW(p) · GW(p)]
  8. This is a judgment call, and one way to approach it would be to reserve something like 1 hour of full-board meeting time per year for talking about these sorts of things (and pouring in more time if at least, like, 1/3 of the board thinks something is a big deal). Some examples of things I think are and aren't usually a big enough deal to start paying serious attention to:
    1. Big enough deal: financial decisions that increase the odds of going "belly-up" (running out of money and having to fold) by at least 10 percentage points. Not a big enough deal: spending money in ways that are arguably bad uses of money, having a lowish-but-not-too-far-off-of-peer-organizations amount of runway.
    2. Big enough deal: deficiencies in financial controls that an auditor is highlighting, or a lack of audit altogether, until a plan is agreed to to address these things. Not a big enough deal: most other stuff in this category.
    3. Big enough deal: organizations with substantial "PR risk" exposure should have a good team for assessing this and a "crisis plan" in case something happens. Not a big enough deal: specific organizational decisions and practices that you are not personally offended by or find unethical, but could imagine a negative article about. (If you do find them substantively unethical, I think that's a big enough deal.)
    4. Big enough deal: transferring like 1/3 or more of valuable things the nonprofit has (intellectual property, money, etc.) to another entity not controlled by the board. Not a big enough deal: starting an affiliate organization primarily for taking donations in another country or something.
    5. Big enough deal: doubling or halving the workforce. Not a big enough deal: smaller hirings and firings.
  9. Sometimes the Board Chair is the CEO, and sometimes the Chair is an employee of the company who also sits on the board. In these cases, I think it's good for there to be a separate Lead Independent Director who is not employed by the company and is therefore exclusively representing the Board. They can help set agendas, lead meetings, and take responsibility by default when it's otherwise unclear who would do so.  [LW(p) · GW(p)]
  10. Nonprofits can get expertise on topic X by hiring experts on X to advise them. The question is: when is it important to have an expert on X evaluating the CEO [LW(p) · GW(p)]
  11. Though it could be fine and even interesting to have giant boards - 20 people, 50 or more - that have some sort of "executive committee" of 10 or fewer people doing basically all of the meetings and all of the work (with the rest functioning just as very passive, occasionally-voting equivalents of "shareholders"). Just assume I'm talking about the "executive committee" type thing here.  [LW(p) · GW(p)]


Comments sorted by top scores.

comment by Lorenzo Buonanno · 2022-06-28T16:22:04.691Z · EA(p) · GW(p)

Lots of comments on Hacker News and LessWrong [LW · GW].

I also found the section of Charity Entrepreneurship's Book on Boards (Chapter 26: "Wisdom") really interesting, from the perspective of the "CEO"s.

They suggest having two separate "boards":

  • An informal "Advisory Board" that helps with advice, mentoring, domain expertise, signaling, fundraising and personal support; but has no legal power. Example terms:
  • A formal Legal Board that handles finances, legal compliance, evaluates the executive team, and also provides advice
comment by Manuel_Allgaier · 2022-06-29T13:18:48.971Z · EA(p) · GW(p)

What about democratically elected non-profit boards?

Most national EA organisations with paid staff (like EA France, EA Norway or EA Germany just to mention a few) are registered associations that have their board (re-)elected by its members every 1-2 years. That way board members can be fired by the association members they represent.

I don't think this is perfect, the average member often does not have enough info to judge the performance of a board member and elections have their own downsides (like sometimes favoring popular and charismatic candidates over the best candidates for the job), but at least for national EA orgs it does seem like the best option to me (medium confidence).

This seems a lot more common in mainland Europe than in the UK or the US. Is this something we should explore more for other nonprofits as well? What other non-profits have clearly defined members (e.g. beneficiaries, stakeholders, ...) that could elect a board?

comment by MarcSerna · 2022-06-28T19:35:27.030Z · EA(p) · GW(p)

I recently became a board member and I was very confused because of many of the issues raised in this post. I have gained great ideas and a bit of clarity.

Thank you.

comment by Ulrik Horn · 2022-06-28T13:00:27.790Z · EA(p) · GW(p)

A couple of thoughts. And I only have some experience working with boards, so take my anecdotes here with a grain of salt.

Even for-profit boards can be weird as you note (perhaps even more weird - the profit incentive might bring out more feelings). Some public examples (lots of details in Jobs' bio and Bad Blood) are how Steve Jobs as the CEO reporting to the board basically fired many people from his board when he came back to Apple. Then there was (is?) Theranos - that seemed like an absolute nightmare. I guess where there is people there is drama. But Jobs and Theranos are likely edge cases but illustrate the extent to which boards can become weird.

Are board members compensated in non-profits? In for profits they are, so it feels natural to "put them to work." I think it works well when the CEO is dealing with something difficult, to work alongside a selected board member, in a transparent way, to solve a particular issue.

Also, perhaps the board should demand from the CEO a time-efficient way for them to report continuous process. Like a high level take on progress on important work, the financial status, the biggest risks and opportunities , evaluation metrics and similar. This should be sent to the board a week or so ahead of each board meeting so board members can prepare to ask questions etc. As in they should demand something that makes them understand the main activities and context of the organization they are responsible for. And perhaps also for the CEO to send some material ahead of the board meeting to dive a bit deeper into a specific topic, e.g. the plan to raise more funds in the coming months. At least this seems to work well in the for-profit sector. i think if the board does not feel they know what is going on, that is a big issue and one should work to resolve this by focusing on the communication that works best for the board and the CEO.

On the issue of laid-back vs. micro managing, perhaps one should not aim for every board member to have the right balance, but instead have a mix of approaches in the board? And that the composition of the board and the CEO works well together - it seems to be a lot about personal chemistry (that is why I understood Jobs wanted to replace certain members of the board before coming back as a CEO - I think this is reasonable to some extent).

Again, just some observations from my very limited experience with for-profit boards only.