Ideal governance (for companies, countries and more)

Cross-posted from Cold Takes

Click lower right to download or find on Apple Podcasts, Spotify, Stitcher, etc.

I’m interested in the topic of ideal governance: what kind of governance system should you set up, if you’re starting from scratch and can do it however you want?

Here “you” could be a company, a nonprofit, an informal association, or a country. And “governance system” means a Constitution, charter, and/​or bylaws answering questions like: “Who has the authority to make decisions (Congress, board of directors, etc.), and how are they selected, and what rules do they have to follow, and what’s the process for changing those rules?”

I think this is a very different topic from something like “How does the US’s Presidential system compare to the Parliamentary systems common in Europe?” The idea is not to look at today’s most common systems and compare them, but rather to generate options for setting up systems radically different from what’s common today.

I don’t currently know of much literature on this topic (aside from the literature on social choice theory and especially voting methods, which covers only part of the topic). This post describes the general topic and why I care, partly in the hopes that people can point me to any literature I’ve missed. Whether or not I end up finding any, I’m likely to write more on this topic in the future.

Outline of the rest of the piece:

  • I’ll outline some common governance structures for countries and major organizations today, and highlight how much room there is to try different things that don’t seem to be in wide use today. More

  • I’ll discuss why I care about this question. I have a few very different reasons:

    • A short-term, tangible need: over the last several years, I’ve spoken with several (more than 3) organizations that feel no traditional corporate governance structure is satisfactory, because the stakes of their business are too great and society-wide for shareholder control to make sense, yet they are too early-stage and niche (and in need of nimbleness) to be structured like a traditional government. An example would be an artificial intelligence company that could end up with a normal commercial product, or could end up bringing about the most important century of all time for humanity. I wish I could point them to someone who was like: “I’ve read all of, and written much of, the literature on what your options are. I can walk you through the pros and cons and help you pick a governance system that balances them for your needs.”

    • A small probability of a big future win. The world today has lots of governments, but they seem to mostly follow a very small number of basic governance templates. At some point, there will be new states with new Constitutions—maybe via space settlements, maybe via collapse of existing states, etc. - but I expect these moments to be few and far between. A significant literature and set of experts on “ideal governance” could lead to a radically different kind of state government, potentially with radically different policies that the rest of the world could learn from.

    • A weird, out-of-left-field application. Some of my interest in this topic actually comes via my interest in moral uncertainty: the question of what it’s ethical to do when one is struggling between more than one theory of ethics, with radically different implications. This is hard to explain, but I try below.

  • I’ll describe a bit more what I think literature on this question could look like (and what already exists that I know of), partly to guide readers who might be able to help me find more.

Common governance structures today

All of these are simplified; I’m trying to illustrate the basic idea of what questions “ideal governance” is asking.

  • A standard (e.g., public) corporation works like this: it has shareholders, assigned one vote per share (not per person), who elect a board of directors that governs by majority. The board generally appoints a CEO that it entrusts with day-to-day decisions. There is a “constitution” of sorts (the Articles of Incorporation and bylaws) and a lot more wrinkles in terms of how directors are selected, but that’s the basic idea.

  • A standard nonprofit is like a corporation, but entirely lacking the shareholder layer—it’s governed directly by the board of directors. (I find something weird about a structure this simple—a simple board majority can do literally anything, even though the board of directors is often a somewhat random assortment of donors, advisors, etc.)

  • The US federal government is a lot more complex. It splits authority between the House of Representatives, the Senate, the Presidency and the Supreme Court, all of which have specific appointment procedures, term limits, etc. and are meta-governed by a Constitution that requires special measures to change. There are lots of specific choices that were made in designing things this way, and lots of things that could’ve been set up differently in the 18th century that would probably still matter today.

  • Other democracies tend to have governments that differ in a lot of ways (e.g.), while being based on broadly similar principles: voters elect representatives to more than one branch of government, which then divide up (and often can veto each other on) laws, expenditures, etc.

  • When I was 13, the lunch table I sat at established a Constitution with some really strange properties that I can’t remember. I think there was a near-dictatorial authority who rotated daily, with others able to veto their decisions by assembling supermajorities or maybe singing silly songs or something.

In addition to the design choices shown in the diagrams, there are a lot of others:

  • Who votes, how often, and what voting system is used?

  • How many representatives are there in each representative body? How are they divided up (one representative per geographic area, or party-list proportional representation, or something else)?

  • What term limits exist for the different entities?

  • Do particular kinds of decisions require supermajorities?

  • Which restrictions are enshrined in a hard-to-change Constitution (and how hard is it to change), vs. being left to the people in power at the moment?

One way of thinking about the “ideal governance” question is: what kinds of designs could exist that aren’t common today? And how should a new organization/​country/​etc. think about what design is going to be best for its purposes, beyond “doing what’s usually done”?

For any new institution, it seems like the stakes are potentially high—in some important sense, picking a governance system is a “one-time thing” (any further changes have to be made using the rules of the existing system1).

Perhaps because of this, there doesn’t seem to be much use of innovative governance designs in high-stakes settings. For example, here are a number of ideas I’ve seen floating around that seem cool and interesting, and ought to be considered if someone could set up a governance system however they wanted:

  • Sortition, or choosing people randomly to have certain powers and responsibilities. An extreme version could be: “Instead of everyone voting for President, randomly select 1000 Americans; give them several months to consider their choice, perhaps paid so they can do so full-time; then have them vote.”

    • The idea is to pick a subset of people who are both (a) representative of the larger population (hence the randomness); (b) will have a stronger case for putting serious time and thought into their decisions (hence the small number).

    • It’s solving a similar problem that “representative democracy” (voters elect representatives) is trying to solve, but in a different way.

  • Proportional decision-making. Currently, if Congress is deciding how to spend $1 trillion, a coalition controlling 51% of the votes can control all $1 trillion, whereas a coalition controlling 49% of the votes controls $0. Proportional decision-making could be implemented as “Each representative controls an equal proportion of the spending,” so a coalition with 20% of the votes controls 20% of the budget. It’s less clear how to apply this idea to other sorts of bills (e.g., illegalizing an activity rather than spending money), but there are plenty of possibilities.2

  • Quadratic voting, in which people vote on multiple things at once, and can cast more votes for things they care about more (with a “quadratic pricing rule” intended to make the number of votes an “honest signal” of how much someone cares).

  • Reset/​Jubilee: maybe it would be good for some organizations to periodically redo their governance mostly from scratch, subject only to the most basic principles. Constitutions could contain a provision like “Every N years, there shall be a new Constitution selected. The 10 candidate Constitutions with the most signatures shall be presented on a ballot; the Constitution receiving the most votes is the new Constitution, except that it may not contradict or nullify this provision. This provision can be prevented from occurring by [supermajority provision], and removed entirely by [stronger supermajority].”

  • More examples in a footnote.3

If we were starting a country or company from scratch, which of the above ideas should we integrate with more traditional structures, and how, and what else should we have in our toolbox? That’s the question of ideal governance.

Why do I care?

I have one “short-term, tangible need” reason; one “small probability of a big future win” reason; and one “weird, out-of-left-field” reason.

A short-term, tangible need: companies developing AI, or otherwise aiming to be working with huge stakes. Say you’re starting a new company for developing AI systems, and you believe that you could end up building AI with the potential to change the world forever.

  • The standard governance setup for a corporation would hand power over all the decisions you’re going to make to your shareholders, and by default most of your shares are going to end up held by people and firms that invested money in your company. Hopefully it’s clear why this doesn’t seem like the ideal setup for a company whose decisions could be world-changing. A number of AI companies have acknowledged the basic point that “Our ultimate mission should NOT just be: make money for shareholders,” and that seems like a good thing.

  • One alternative would be to set up like a nonprofit instead, with all power vested in a board of directors (no shareholder control). Some issues are that (a) this cuts shareholders out of the loop completely, which could make it pretty hard to raise money; (b) according to me at least, this is just a weird system of governance, for reasons that are not super easy to articulate concisely but I’ll take a shot in a footnote4 (and possibly write more in the future).

  • Another alternative is a setup that is somewhat common among tech companies: 1-2 founders hold enough shares to keep control forever, so you end up with essentially a dictatorship. This also … leaves something to be desired.

  • Or maybe a company like this should just set up more like a government from the get-go, offering everyone in the world a vote via some complex system of representation, checks and balances. But this seems poorly suited to at least the relatively early days of a company, when it’s small and its work is not widely known or understood. But then, how does the company handle the transition from the latter to the former? And should the former be done exactly in the standard way, or is there room for innovation there?

Over the last several years, I’ve spoken with heads of several (more than 3) organizations that struggle between options like the above, and have at least strongly considered unusual governance setups. I wish I could point them to someone who was like: “I’ve read all of, and written much of, the literature on what your options are. I can walk you through the pros and cons and help you pick a governance system that balances them for your needs.”

But right now, I can’t, and I’ve seen a fair amount of this instead: “Let’s just throw together the best system we can, based mostly on what’s already common but with a few wrinkles, and hope that we figure this all out later.” I think this is the right solution given how things stand, but I think it really does get continually harder to redesign one’s governance as time goes on and more stakeholders enter the picture, so it makes me nervous.

Similar issues could apply to mega-corporations (e.g., FAANG) that are arguably more powerful than what the standard shareholder-centric company setup was designed for. Are there governance systems they could adopt that would make them more broadly accountable, without copying over all the pros and cons of full-blown representative democracy as implemented by countries like the US?

A small probability of a big future win: future new states. The world today has lots of governments, but they seem to mostly follow a very small number of basic governance templates (e.g., I believe you see almost none of the things I listed above), and probably relatedly, there seems to be remarkably little variety and experimentation with policy. Policies that many believe could be huge wins—such as dramatically expanded immigration, land value taxation, “consumer reports”-style medical approvals,5 drug decriminalization, and charter cities—don’t seem to have gotten much of a trial anywhere in the world.

At some point, there will be new states with new Constitutions—maybe via space settlements, maybe via collapse of existing states, etc. - but I expect these moments to be few and far between.

By default I expect future Constitutions to resemble present ones an awful lot. But maybe, at some future date, there will be a large “ideal governance” literature and some points of expert consensus on innovative governance designs that somebody really ought to try. That could lead to a radically different kind of state government, potentially with radically different policies that the rest of the world could learn from.

An out-of-left-field application for “ideal governance.” This is going to veer off the rails, so remember to skip to the next section if I lose you.

Some of my interest in this topic actually comes via my interest in moral uncertainty: the question of what it’s ethical to do when one is struggling between more than one theory of ethics, with radically different implications.

For example, there are arguments that our ethical decisions should be dominated by concern for ensuring that as many people as possible will someday get to exist. I really go back and forth on how much I buy these arguments, but I’m definitely somewhere between 10% convinced and 50% convinced. So … say I’m “20% convinced” of some view that says preventing human extinction6 is the overwhelmingly most important consideration for at least some dimensions of ethics (like where to donate), and “80% convinced” of some more common-sense view that says I should focus on some cause unrelated to human extinction.7 How do I put those two together and decide what this means for actual choices I’m making?

The closest thing I’ve seen to a reasonable-seeming answer is the idea of a moral parliament: I should act as though I’m run by a Parliament with 80 members who believe in “common-sense” ethics, and 20 members who believe in the “preventing extinction is overwhelmingly important” idea. But with default Parliament rules, this would just mean the 80 members can run the whole show, without any compromise with the 20.

And so, a paper on the “moral parliament” idea tries to make it work by … introducing a completely new governance mechanism that I can’t find any other sign of someone else ever talking about, “proportional chances voting” (spelled out in a footnote).8 I think this mechanism has its own issues,9 but it’s an attempt to ensure something like “A coalition controlling 20% of the votes has 20% of the effective power, and has to be compromised with, instead of being subject to the tyranny of the majority.”

My own view (which I expect to write more about in the future) is that governance is roughly the right metaphor for “moral uncertainty”: I am torn by multiple different sides of myself, with different takes on what it means to be a good person, and the problem of getting these different sides of myself to reach a decision together is like the problem of getting different citizens (or shareholders) to reach a decision together. The more we can say about what ideal governance looks like, the more we can say about how this ought to work—and the better I expect this “moral parliament”-type idea to end up looking, compared to alternatives.10

The literature I’m looking for

Ideal governance seems like the sort of topic for which there should be a “field” of “experts,” studying it. What would such study look like? Three major categories come to mind:

Brainstorming ideas such as those I listed above—innovative potential ways of solving classic challenges of governance, such as reconciling “We want to represent all the voters” with “We want decisions to be grounded in expertise and high engagement, and voters are often non-expert and not engaged.”

I’ve come across various assorted ideas in this category, including quadratic voting, futarchy, and proportional chances voting, without seeing much sign that these sit within a broader field that I can skim through to find all the ideas that are out there.

Economics-style theory in which one asks questions like: “If we make particular assumptions about who’s voting, what information they have and lack, how much they suffer from bounded rationality, and how we define ‘serving their interests’ (see below), what kind of governance structure gets the best outcome?”

Social choice theory, including on voting methods, tackles the “how we define ‘serving their interests’” part of this. But I’m not aware of people using similar approaches to ask questions like “Under what conditions would we want 1 chamber of Congress vs. 2, or 10? 100 Senators vs. 500, or 15? A constitution that can be modified by simple majority, vs. 23 majority vs. consensus? Term limits? Etc. etc. etc.”

Empirical research (probably qualitative): Are there systematic reviews of unusual governance structures tried out by companies, and what the results have been? Of smaller-scale experiments at co-ops, group houses and lunch tables?

To be clear, I think the most useful version of this sort of research would probably be very qualitative—collecting reports of what problems did and didn’t come up—rather than asking questions like “How does a particular board structure element affect company profits?”

One of the things I expect to be tricky about this sort of research is that I think a lot of governance comes down to things like “What sorts of people are in charge?” and “What are the culture, expectations, norms and habits?” A setup that is “officially” supposed to work one way could evolve into something quite different via informal practices and “soft power.” However, I think the formal setup (including things like “what the constitution says about the principles each governance body is supposed to be upholding”) can have big effects on how the “soft power” works.

If you know where to find research or experts along the lines of the above, please share them in the comments or using this form if you don’t want them to be public.

I’ll likely write about what I come across, and if I don’t find anything new, I’ll likely ramble some more about ideal governance. So either way, there will be more on this topic!


Footnotes

  1. Barring violent revolution in the case of countries.

  2. An example would be the “proportional chances voting” idea described here.

    • Proxying/​liquid democracy, or allowing voters to transfer their votes to other voters. (This is common for corporations, but not for governments.) This could be an alternative or complement to electing representatives, solving a similar problem (we want lightly-engaged voters to be represented, but we also want decisions ultimately made using heavy engagement and expertise). At first glance it may seem to pose a risk that people will be able to “buy votes,” but I don’t actually think this is necessarily an issue (proxying could be done anonymously and on set schedules, like other votes).

    • Soft term limits: the more terms someone has served, the greater a supermajority they need to be re-elected. This could be used to strike a balance between the advantages of term limits (avoiding “effectively unaccountable” incumbents) and no-term-limits (allowing great representatives to keep serving).

    • Formal technocracy/​meritocracy: Using hard structures (rather than soft norms) to assign authority to people with particular expertise and qualifications. An extreme example would be futarchy, in which prediction markets directly control decisions. A simpler example would be structurally rewarding representatives (via more votes or other powers) based on assessments of their track records (of predictions or decisions), or factual understanding of a subject. This seems like a tough road to go down by default, as any mechanism for evaluating “track records” and “understanding” can itself be politicized, but there’s a wide space of possible designs.

  3. Most systems of government have a sort of funnel from “least engaged in day to day decisions, but most ultimately legitimate representatives of whom the institution is supposed to serve” (shareholders, voters) to “most engaged in day to day decisions, but ultimately accountable to someone else” (chief executive). A nonprofit structure is a very short funnel, and the board of directors tends to be a somewhat random assortment of funders, advisors, people who the founders just thought were cool, etc. I think they often end up not very accountable (to anyone) or engaged in what’s going on, such that they have a hard time acting when they ought to, and the actions they do take are often kind of random.

    I’m not saying there is a clearly better structure available for this purpose—I think the weirdness comes from the fact that it’s so unclear who should go in the box normally reserved for “Shareholders” or “Voters.” It’s probably the best common structure for its purpose, but I think there’s a lot of room for improvement, and the stakes seem high for certain organizations.

  4. Context in this Marginal Revolution post, which links to this 2005 piece on a “consumer reports” model for the FDA.

  5. Or “existential catastrophe”—something that drastically curtails humanity’s future, even if it doesn’t drive us extinct.

  6. This isn’t actually where I’m at, because I think the leading existential risks are a big enough deal that I would want to focus on them even if I completely ignored the philosophical argument that the future is overwhelmingly important.

  7. Let’s say that 70% of the Parliament members vote for bill X, and 30% vote against. “Proportional chance voting” literally uses a weighted lottery to pass bill X with 70% probability, and reject it with 30% probability (you can think of this like rolling a 10-sided die, and passing the bill if it’s 7 or under).

    A key part of this is that the members are supposed to negotiate before voting and holding the lottery. For example, maybe 10 of the 30 members who are against bill X offer to switch to supporting it if some change is made. The nice property here is that rather than having a “tyranny of the majority” where the minority has no bargaining power, we have a situation where the 70-member coalition would still love to make a deal with folks in the minority, to further increase the probability that they get their way.

    Quote from the paper that I am interpreting: “Under proportional chances voting, each delegate receives a single vote on each motion. Before they vote, there is a period during which delegates may negotiate: this could include trading votes on one motion for votes on another, introducing novel options for consideration within a given motion, or forming deals with others to vote for a compromise option that both consider to be acceptable. The delegates then cast their ballots for one particular option in each motion, just as they might in a plurality voting system. But rather than determining the winning option to be the one with the most votes, each option is given a chance of winning proportional to its share of the votes.”

  8. What stops someone who lost the randomized draw from just asking to hold the same vote again? Or asking to hold a highly similar/​related vote that would get back a lot of what they lost? How does that affect the negotiated equilibrium?

  9. Such as “maximize expected choice-worthiness,” which I am not a fan of for reasons I’ll get to in the future.