Crypto: ECB accuses DeFi of masking massive centralization

The ECB is clearly challenging one of the founding narratives of modern crypto. In practice, several large DeFi protocols remain concentrated in few hands, especially when we look at actual governance and not simple marketing talk. This is the meaning of the working paper published this week, which observes a high concentration of governance tokens and voting power on Aave, MakerDAO, Ampleforth and Uniswap.

A frontal attack on the crypto narrative of decentralization

According to the study, the 100 largest crypto holders control more than 80% of governance tokens across the four protocols studied. Even more striking, the first five portfolios concentrate between 36% and 59% of the offer depending on the case.

In other words, DeFi crypto can appear diffuse on the surface, with thousands of addresses visible on the blockchain, while remaining very tight at the top. The paper adds that the most active voters are often delegates, which can reinforce the weight of a small core of actors instead of actually broadening participation.

However, we must keep an important nuance. The ECB speaks here through a working paper, therefore a research document intended to fuel the debate. The institution itself specifies that these papers are work in progress and that the opinions expressed do not necessarily reflect its official position.

What the study really measures, and what it doesn't measure

The paper does not claim to demonstrate that all crypto DeFi is an illusion. It mainly measures governance. Clearly, it looks at who holds the tokens, who votes, who receives the delegations, and what decisions actually go through these mechanisms. This is not exactly the same as measuring the technical decentralization of a protocol.

The method is based on two observation periods, November 2022 and May 2023. The authors focus on Ethereum, which represented approximately 57% of the total value locked in DeFi at the time studied. The four protocols selected together accounted for around 32% of this ecosystem, with 248 governance proposals included in the analysis out of 1,051 identified.

But the study also has its blind spots. The data was collected manually from public and pseudonymous sources. The authors themselves acknowledge possible inaccuracies, missing information and the impossibility of integrating crypto protocols like Curve or dYdX due to lack of sufficient data. It's not a detail. This is even a major limitation when we want to draw a general conclusion about all of DeFi.

The real controversy concerns the chosen threshold

The sharpest criticism comes from Bill Hughes, lawyer at Consensys. According to him, the paper stacks up real numbers, then applies to them a subjective reading of the spectrum between centralization and decentralization. His criticism is therefore not that the data is entirely false, but that it results in a standard that is almost impossible to meet.

This is where the debate becomes political. The paper explains that crypto decentralization exists on a spectrum, while asserting that “full decentralization” is not achieved in the sample studied. He also emphasizes that there is no clear threshold to define what complete decentralization is.

But this vagueness matters a lot in Europe. MiCA anticipates that crypto services provided in a fully decentralized manner, without intermediaries, should not fall within its scope. At the same time, the AMF points out that the text also covers activities provided or controlled directly or indirectly by people or entities, including when part of the service is performed in a decentralized manner. The battle is no longer just about technique. It concerns proof of the absence of control.