The More a DAO Votes On, the Fewer Members Actually Decide

The people who built decentralized giving to remove the person in charge keep discovering that a new one has quietly formed. They did what the theory asked. They spread the tokens that carry voting power across many hands, and they published every proposal in the open where any stakeholder could inspect it and weigh in. And still, when researchers looked closely at how these organizations actually behave over time, the decisions kept flowing back to a shrinking few. A recent working paper studying decentralized autonomous organizations, which are the on-chain groups that run shared treasuries by member vote, found that this concentration is not a failure of the design. It is a product of it.

The mechanism is almost ordinary, which is what makes it so hard to wave away. As one of these organizations grows, the number of proposals its members are asked to judge grows faster than the members' capacity to read and act on them. Attention is finite because time is finite. When the flow of decisions outpaces the crowd's ability to keep watch, most members quietly stop trying, and authority settles onto the small group that still has the time and the incentive to pay attention. The paper names the tipping point a capacity breakpoint, which is a plain way of saying there is a threshold past which broad participation folds on its own. The shift happens without a takeover. The structure simply hands control to whoever is still watching.

This reaches well past crypto, because the real subject of the paper is how participation behaves under load, and participation under load is the unspoken problem inside nearly every promise the giving world has made this decade. We keep designing for the moment someone shows up to decide, and we keep ignoring the moment two hundred decisions show up at once and the same person has a job and a family and a finite evening.

For digital and crypto-native giving leaders, the finding takes away a comfortable story. The pitch for on-chain giving has often been that transparency plus token voting equals democratized control, so that removing the intermediary removes the concentration of power that intermediaries create. This research says the concentration returns through the back door, not because the tokens were distributed unfairly but because monitoring capacity was never distributed at all. A governance system that dumps every decision on every holder is not empowering them. It is quietly selecting for the few who can absorb the workload, which means the design has to budget for attention as carefully as it budgets for tokens.

For nonprofit and foundation professionals now experimenting with participatory grantmaking, which is the practice of letting the affected community rather than the program officer decide who gets funded, the warning is direct. A participatory process that keeps expanding the number of choices it puts in front of volunteers will hit the same breakpoint, and when it does the quiet result is that a couple of committed members carry the real decisions while everyone else rubber-stamps. The participation stays real on paper and thins out in practice, so the design question is not how many people you invited but how much you asked each of them to carry.

For donors and family-office advisors weighing where governance-heavy giving vehicles fit, the paper reframes what a governance right is actually worth. A vote you do not have the time to exercise is not ownership, it is the appearance of ownership, and advisors who steer clients toward collective or on-chain structures on the strength of their democratic design need to ask whether the client will ever realistically monitor the decisions those structures generate. The value of a seat at the table depends entirely on whether anyone can stay at the table once the meetings never end.

For NGOs and international organizations looking at DAO-style pooled funds as a way to move money collectively across borders, the lesson is that decentralization does not remove the need for delegation, it just changes who you are trusting and whether you admit it. A pooled structure that pretends every contributor governs equally will concentrate power anyway, so the honest move is to design the delegation openly, with clear mandates and the ability to revoke them, rather than to let concentration happen by exhaustion and then act surprised.

For corporate giving and CSR leaders, the breakpoint shows up in employee-directed giving, which is the program design that lets staff vote on where the company's charitable dollars go. Those programs are usually judged by participation rate, and the paper's finding is that participation rate is the wrong gauge, because a program that puts more choices in front of employees every quarter will see the same thinning, with a committed few carrying the decisions while the number on the dashboard stays flattering. The design fix is the one the DAO researchers imply, which is fewer decisions, bundled well, so that the votes employees do cast are the ones that change where the money lands.

The reframe underneath all of this is the one the dissertation has been circling. The old worry, sometimes called the iron law of oligarchy, holds that every organization eventually consolidates power in a small leadership no matter how democratic its founding. Decentralized giving was supposed to be the exception, the structure that finally escaped the law by writing openness into code. What the capacity-breakpoint finding suggests is subtler and more useful than either surrender or denial. Concentration is not inevitable because people are corrupt. It is likely because attention is scarce, which means it can be engineered against. You can cap how many decisions reach the full membership and bundle the routine ones so the crowd's limited attention goes to the choices that actually matter, and you can measure participation honestly instead of celebrating a turnout number that hides who is really deciding.

So the diagnostic question for anyone building participatory giving is not whether your members can vote. It is whether they can afford to. Design that treats attention as infinite will reproduce the very concentration it set out to abolish, and the reproduction will look, from the dashboard, like consent.

The point is not that decentralized giving fails. The point is that participation is a resource with a budget, and the organizations that will actually keep power distributed are the ones that spend that budget deliberately instead of pretending it has no limit.

"A vote that members have no time to cast is the picture of ownership rather than ownership itself, and the picture is exactly what concentration hides behind."

EILEEN HERNANDEZ, THE STRATEGIC STEWARD

Share this𝕏 / TwitterLinkedIn

This is why the DIA Protocol treats Participatory Ownership as something a system has to engineer and defend on purpose, not a property it inherits the moment it goes on-chain.