Share article

Decentralization is still doing a lot of marketing work for DeFi. The European Central Bank has now put numbers behind the old suspicion: many DAOs look less like leaderless internet collectives and more like systems where a small cluster of insiders keeps the steering wheel.
An ECB working paper released this week examined governance structures across four major DeFi protocols and concluded that governance token ownership is heavily concentrated. The core finding is simple and uncomfortable: a limited number of actors control most voting power, which weakens the claim that these protocols are governed by broad communities rather than by founders, funds, and other large token holders. [1]

Enjoy articles without ads?

Register for free and get unlimited access to all articles.

What the ECB found

The paper focuses on governance in decentralized autonomous organizations, or DAOs, the token-based structures many DeFi protocols use to approve upgrades, treasury decisions, and risk parameters. In theory, these systems spread control across token holders. In practice, the ECB says ownership concentration makes that decentralization claim hard to sustain. [2]
The researchers looked at four large protocols and found that governance power was not widely distributed. Instead, token balances clustered among a relatively small set of participants. That matters because DAO voting is usually proportional to token holdings. If a handful of wallets or affiliated actors hold enough tokens, they can dominate proposals, shape outcomes, or block changes without needing anything close to community-wide consensus.
The paper also pushes on a regulatory point that sits behind the headline. If governance is concentrated, supervisors may have an easier time identifying who actually exercises control over supposedly decentralized systems. That is important for Europe, where policymakers have been trying to work out when DeFi protocols fall outside traditional regulatory perimeters and when they do not. "No one is in charge" becomes a weaker defense if the token ledger suggests otherwise. [3]

Why this matters beyond the obvious

The ECB's argument lands in a market that has spent years treating decentralization as both a design goal and a legal shield. Protocols often present DAOs as evidence that control has been handed to the community. Sure. But if treasury allocations, code changes, and listing decisions can still be driven by a few large holders, the legal and operational picture looks different.

That has at least three consequences.

First, governance capture risk is higher than DAO branding suggests. Large holders can influence emissions, collateral rules, fee switches, and treasury deployments in ways smaller holders cannot realistically counter.
Second, regulatory attribution gets easier. A concentrated governance structure gives authorities a clearer path to ask who benefited, who voted, who proposed changes, and who effectively controlled the protocol during key decisions.

Third, decentralization theater becomes easier to spot. A protocol can distribute front-end branding, community calls, and forum chatter across thousands of users while still concentrating actual control in a far smaller group. Governance turnout data across DeFi has hinted at this for years. The ECB paper turns that long-running complaint into a more formal policy argument. [4]

The bigger policy angle

This is not just an academic swipe at crypto ideals. The paper appears aimed at a live policy problem: how regulators should identify accountable parties inside DeFi systems that claim to be decentralized by design.

That question matters more now because European regulators are moving from broad crypto frameworks to harder questions about infrastructure, governance, and liability. If token-based voting can be traced to core insiders, venture funds, foundations, or tightly linked entities, then the line between a decentralized protocol and a coordinated organization starts to blur.

For DeFi teams operating in Europe, the message is awkward but clear. Governance tokens alone do not prove decentralization. Neither do open-source codebases or on-chain voting portals. Regulators are increasingly likely to look at who holds the tokens, who submits the proposals, who controls the multisigs, and who can change the code path in practice. [5]

What the study does not prove

The ECB paper strengthens the case that many DAOs are more centralized than advertised, but it does not mean every protocol governance system is fake or purely cosmetic. Concentrated token ownership is common in young or capital-intensive networks, especially where founders, treasuries, and early backers still hold large allocations. Some protocols also deliberately use delegated voting, security councils, or emergency multisigs to avoid operational chaos.

That nuance matters. A system can be partially decentralized, or decentralized in some functions but not others. Treasury control, code deployment, front-end hosting, oracle dependencies, and voting power can all sit on different parts of the centralization spectrum. The ECB's core point is narrower: if a few actors dominate governance, then claims of broad-based control need a reality check.

What to watch next

The practical question is whether this line of research feeds directly into European enforcement or rulemaking. Watch for policymakers to lean more heavily on governance concentration, multisig composition, and proposal authorship when deciding whether a DeFi protocol has identifiable controllers.

Also worth watching: whether protocols respond by widening token distribution, limiting whale influence, increasing disclosure around affiliated wallets, or redesigning voting systems altogether. Quadratic voting, delegation reforms, and caps on voting power get discussed every cycle. Adoption is another matter, because of course the biggest token holders are often the ones who would need to approve those changes.
For investors and users, the takeaway is less philosophical than operational. When a DAO says it is decentralized, check the wallet distribution, voter turnout, delegate concentration, and upgrade keys. The branding is easy. The control map is the real story.