DAOs Need a Vibe Check

Despite the Golden Boys’ motivations being unclear, it seems they were able to exploit CompoundDAO due to a persistent issue: voter apathy among its members. Compound, known for its early adoption of token-voting in decentralized systems and having a Nakamoto coefficient of 17, is more decentralized than many proof of stake networks. However, the Golden Boys managed to pass Proposal 289 with less than 7% of the total COMP supply, indicating that a significant portion of token holders did not participate in voting. This lack of participation is often overlooked when assessing decentralization. This, indeed, poses a serious challenge.

Today’s decentralization metrics

Today’s common measures of decentralization, like the Nakamoto Coefficient or the Gini Coefficient, primarily concentrate on the distribution of tokens across different addresses. In simpler terms, they are more concerned with how tokens are spread among various accounts rather than the overall structure of the network.

Let’s look at two fictional Decentralized Autonomous Organizations (DAOs) named WhaleDAO and MinnowDAO. In WhaleDAO, the total supply of tokens is divided equally among just five digital wallets. Conversely, in MinnowDAO, the token distribution is evenly split across 100 different wallets.

Initially, it might seem that MinnowDAO is more decentralized compared to WhaleDAO due to its structure. However, it’s essential to delve deeper into who actually holds these wallets. For instance, if 70 out of the total wallets in MinnowDAO are owned by a single whale, then the decision-making power in MinnowDAO would be concentrated among that one individual or group, making it more centralized than WhaleDAO. This single person would control around 70% of the voting power in MinnowDAO.

In the instance of the Golden Boys, it’s important to note that their collective “yes” vote, approximately 700k COMP, was dispersed among a multitude of wallets. Many of these individual wallets contained less than 25k COMP. Traditional measures of decentralization based on tokens, such as the Nakamoto coefficient, might incorrectly suggest that this decision-making process was decentralized, given that the vote was spread across various addresses.

Token-focused metrics overlook additional types of centralization as well. For instance, imagine a situation where 80% of the voters within a Decentralized Autonomous Organization (DAO) have been influenced to vote for a particular proposal. Bribery doesn’t impact token possession directly, thus a token-centric metric would fail to identify this manipulation. This same issue applies to groupthink, collusion, and other scenarios where control of the DAO’s governance may be concentrated within one or a few groups.

To accurately gauge the level of decentralization that truly reflects the effectiveness of a Decentralized Autonomous Organization’s (DAO) governance system, it is essential to develop a novel evaluation tool or metric.

A new decentralization metric

As an analyst, I’d like to introduce a novel set of metrics we’re calling Voting Bloc Entropy, or VBE for short. Unlike traditional methods that focus on individual token holdings, VBE delves into patterns of voting behavior among token holders grouped together as ‘voting blocs’. These blocs can range from loose alliances to self-identifying groups such as the Golden Boys. By assessing the presence of large, dominant voting blocs within a DAO, we aim to shed light on the extent of centralization in its governance structure.

In terms of the Golden Boys, a Value-Based Entropy (VBE) metric would group all the attackers’ votes together as one collective voting group, regardless of how they distribute their tokens across multiple wallets. Similarly, it would consider all non-participating voters within CompoundDAO as a single bloc since their behavior aligns in this context. This large bloc, consisting mainly of inactive voters, we refer to as an “inactivity whale.” If the Golden Boys were able to manipulate this massive bloc, it would suggest low VBE and consequently poor decentralization within Compound.

One of the most important benefits of a good decentralization metric is that it leads to useful, actionable guidance about how to reinforce decentralization. Our recent research enumerates a number of key lessons that VBE offers on DAO decentralization. For example, you might expect delegation to increase centralization by concentrating voting power. Surprisingly, our results show that when a DAO has a big inactivity whale, delegation can shrink the whale and lead to higher VBE, i.e., higher decentralization.

To clarify the level of decentralization across different Decentralized Autonomous Organizations (DAOs), we’ve developed a new tool for the community: The IC3 Value-Based Entity (VBE) dashboard. This resource allows users to compare VBE among various DAOs, but it’s important to note that comparisons between different DAOs should be made with care. Additionally, the dashboard enables individual DAOs to track changes in their VBE over time. As a stand-in for the overall “wellbeing” or health of a DAO, Value-Based Entity can assist operators in shaping governance strategies.

Additionally, we think our research could guide the assignment process within Decentralized Autonomous Organizations (DAOs). For instance, by grouping delegates into voting blocs akin to political parties, VBE provides token holders an effortless method for recognizing delegates (or even creating AI representatives) who can champion their causes. Consequently, this approach may help minimize voter disengagement within DAOs.

By utilizing tools like Voter Behavior Evaluation (VBE), Decentralized Autonomous Organizations (DAOs) can effectively prevent governance attacks that take advantage of voter inactivity, such as the recent incident with Compound. Going beyond this purpose, optimizing their use of VBE allows DAOs to enhance their governance processes, encouraging a broader range of opinions in voting circles and promoting constructive discussion. Although our work is ongoing, we’ve collaborated with significant DAOs and stakeholders to develop, test, and refine an open-source library that will serve as the foundation for VBE dashboards and various applications.

In many decentralized systems, the “D” in DAO might be the most challenging element to define accurately. However, a positive outlook could guide us through this process.

Keep in mind that the opinions shared within this article belong solely to the writer and may not align with the views of CoinDesk Inc., its proprietors, or their associates.