A new working paper from the European Central Bank has challenged one of the core assumptions behind decentralized finance: that governance power is broadly distributed among communities of token holders. Instead, the study argues that governance across major DeFi protocols is dominated by a relatively small set of insiders, delegates, and trading platforms, leaving regulators and users with limited clarity on who is actually responsible for key decisions.
The paper, ECB Working Paper No. 3208, examined governance data from Aave, MakerDAO (now rebranded as Sky), Ampleforth, and Uniswap at two points in time, November 2022 and May 2023. According to the study, these protocols were selected because of their size and because they represented different segments of DeFi activity. At the time the data was collected, they collectively accounted for roughly 32% of Ethereum’s total value locked.
Token ownership is concentrated at the top
The report’s clearest finding is the scale of token concentration. Across all four protocols, the top 100 holders controlled more than 80% of the governance token supply. In Aave and Uniswap, the top five holders alone accounted for nearly half of all governance tokens, while Ampleforth showed even greater concentration, with its top five addresses controlling close to 60%.
That concentration matters because governance tokens are meant to represent voting rights over protocol changes, listings, risk parameters, and other strategic decisions. In theory, token-based governance is supposed to align control with a broad base of participants. In practice, the ECB paper suggests that power often remains clustered in a narrow group from the outset.
Researchers also attempted to trace who stood behind the largest addresses. Their findings showed that in many cases, roughly half or more of total holdings could be linked either to the protocols themselves, through treasuries, founder allocations, and developer holdings, or to centralized and decentralized exchanges. Among centralized platforms, Binance held the largest share across all four protocols in the sample, with positions ranging from 2% to 15% depending on the project.
Voting power is further concentrated through delegation
The distribution of voting activity was no less concentrated than token ownership. The paper found that the most influential voters were overwhelmingly delegates, meaning individuals or organizations that receive voting power from other token holders. Delegation is often presented as a practical tool for improving participation, especially in systems where many holders are passive. But the ECB study suggests it may also reinforce concentration by channeling influence toward a limited set of active actors.
Identifying those delegates proved difficult. Researchers relied on public web searches, Github profiles, social media accounts, governance forums, and blockchain analytics tools developed by Crystal Intelligence. Even with those methods, about one-third of the top voters in the sample could not be identified at all.
Among the voters researchers were able to classify, individuals made up the largest segment at around 21%, followed by Web3 companies at roughly 19%. Venture capital firms and university blockchain associations also appeared in the data. One of the most notable findings involved Uniswap: Andreessen Horowitz (A16z) was the protocol’s top voter in both periods studied. By May 2023, the firm had voting power delegated to it by 125 addresses.
Governance proposals focus on operations, not governance reform
The paper also reviewed the substance of 248 governance proposals across the four protocols. Proposals related to risk parameters represented the largest category, accounting for 28% of the total. These included decisions on loan-to-value ratios, debt ceilings, stability fees, and emergency shutdown mechanisms. Asset listing proposals made up another 23%.
By contrast, proposals directly addressing governance structure were rare, comprising only 1% of the sample. This is a notable point in the ECB’s analysis: while DeFi protocols advertise governance as a core feature, governance itself appears to be infrequently revisited as an object of reform. Most decision-making remains focused on operational maintenance rather than redesigning who holds influence or how accountability should work.
The study also observed that the underlying concentration of governance power remained broadly stable across the two snapshots. That continuity suggests existing power structures are durable rather than temporary. In other words, market turnover alone may not be enough to produce more distributed governance over time.
Regulatory accountability remains difficult to establish
For regulators, the paper presents a practical challenge. The authors concluded that under current conditions, governance token holders, developers, and centralized exchanges cannot reliably serve as regulatory entry points. The reason is straightforward: blockchain addresses are pseudonymous, delegation chains are opaque, and the actual decision-makers behind key governance outcomes are often difficult to identify with confidence.
This creates tension with existing European regulation. Under the EU’s Markets in Crypto-Assets Regulation (MiCA), services that are provided in a fully decentralized manner are exempt. But the ECB paper argues that this threshold is difficult to apply in practice. Based on the protocols studied, none came close to satisfying a robust standard of genuine decentralization. In most cases, meaningful control still rested with insiders or centralized intermediaries.
The implication is significant. If DeFi projects claim regulatory exemptions on the basis of decentralization, policymakers may increasingly demand evidence that governance, ownership, and implementation power are in fact distributed. The ECB paper suggests that many current protocols would struggle to meet that test.
Possible policy paths for DAO and DeFi oversight
Rather than simply criticizing existing models, the authors also outlined several possible ways forward. These include mandatory disclosure of token holder affiliations, legal structures tailored specifically for DAOs, and hybrid governance models that combine onchain voting with traditional legal accountability mechanisms. The paper also pointed to the Danish Financial Supervisory Authority framework as one practical reference for evaluating whether a crypto offering is genuinely decentralized.
The broader comparison made by the ECB is with traditional corporate governance. Both systems, the paper notes, suffer from low voter participation and outcomes shaped by a small group of active players. But traditional finance has mechanisms such as proxy voting rules, stewardship codes, and legal duties that help anchor accountability. DeFi governance, by contrast, often lacks those safeguards while also obscuring the identity of the most influential participants.
That gap may become more important as DeFi grows and as regulators focus less on slogans and more on verifiable control structures. The ECB paper does not argue that decentralized governance is impossible. But it does make clear that in the protocols examined, governance remains far more concentrated and far less transparent than the industry’s rhetoric often suggests.
For the DeFi sector, the study lands as both a warning and a policy signal. Projects that continue to present token-based governance as sufficient proof of decentralization may face closer scrutiny from regulators and market participants alike. And unless transparency around delegates, treasury control, and exchange-held voting power improves, the debate over who truly governs DeFi is unlikely to fade.