DAOrayaki: Evaluating the "Fairness" of DAO Governance Based on Data Perspective
Original Author: Sicong Zhao
The fairness of DAO governance has always attracted everyone's attention. DAOrayaki discusses how to quantify the fairness of DAO governance from the perspective of data science. Therefore, in this article, 3 indicators will be presented.
Equality of Token Distribution
In many cases, holding more governance tokens means having higher voting power in the DAO. In terms of token distribution, there may be different situations. For example, 1% of members own 90% of the tokens, or each member owns exactly the same amount of tokens. In the former case, the DAO is controlled by 1% of members and is very centralized. In the latter case, each member has equal voting rights in voting, which is in line with the nature of decentralization.
So how do we assess the dispersion of token holdings? This is the first indicator: the Gini coefficient (also known as the Gini index), developed by the Italian statistician Corrado Gini. This metric is widely used to assess income inequality, and I found it to be a good fit for our purpose.
Here is the formula for the Gini coefficient:
The numerator is the average absolute difference of all project pairs (token balances) across the population (DAO members). The denominator is the average token balance scaled by 2n². Factor(2n²) converts the Gini coefficient to a range of 0 to 1.
0 means perfect equality, each member has the same amount of tokens.
1 means full inequality, every member owns all tokens.
Confusion is normal. This video from Khan Academy might help you build an intuitive understanding.
Below are the DAOs with the lowest Gini coefficients. As you can see, all DAOs have a Gini coefficient greater than 0.76. In comparison, the U.S. Gini coefficient was 0.434 in 2017. So maybe all the DAOs in my database are more concentrated than US.
Token distribution sorted by Gini coefficient
voting percentage
Regardless of how evenly distributed the tokens are, if only a small percentage of members participate in governance, the DAO is still centralized. Vote percentage is a metric that changes from poll to poll, here is the definition.
The average percentage of votes in each DAO provides insight into the level of decentralization through participation in governance. You can even dig deeper by comparing the token distribution between voting members and the population. This discrepancy may indicate whether voting members represent wrt token holdings.
coefficient of agreement
In terms of voting, the consensus of the majority of members may lose to a smaller group with more governance tokens (i.e., voting power). Members or tokens, which one plays a bigger role in DAO governance? To quantify this comparison, I present here one final metric: the coefficient of agreement (AC).
There are several versions of the coefficient of agreement in statistics. For this upcoming indicator, here is a vision of mine, if you find it conflicts with a pre-existing concept, please let us know 🙏
Like vote percentages, ACs are vote-specific. First, let me introduce two basic concepts: token rate and turnout rate.
Token Rate and Voting Rate Formula
The above formula should be intuitive. To make it concrete, here's an example with 3 members and 2 options.
an example of voting
In this vote, option 2 wins. So the token rate is 6/11 and the voting rate is 2/3. To get the consensus coefficient, you divide the token rate by the voter rate. In this case AC = 9/11.
Coefficient of agreement formula
Here's how to explain AC. In a perfectly "fair" DAO, assuming tokens and members play the same role in governance, AC is expected to be 1. When AC < 1, it means that the winning option depends more on voters than tokens. When AC > 1, it indicates that the winning option depends more on tokens than voters.
Disclaimer: The metrics presented in this article are not comprehensive and may not be fair. I'll explain the thought process so you can understand the situation and make a judgment call.
