Decentralization has been one of, if not the biggest promise of blockchain. Most of the technological progress made in the space is essentially geared to remove middle-men and democratize finance, payments, collaboration and participation. However, there is more to decentralizing a system than just good code. In the end, it’s still people who have to ensure it runs properly.

This is effectively where DAOs and their structures start to matter. One would think that, since it’s already in the name, a Decentralized Autonomous Organization would have a much flatter structure than, say, a corporation or a government. However, many DAOs today are effectively DINOs: Decentralized In Name Only. But what does it even mean to be decentralized? How can that even be determined? Luckily, a lot of research has been done into this topic and Decent is more than happy to contribute to this discussion. The purpose of this piece is to agree upon:

• How decentralization can be measured

• How DAO autonomy can be measured

• How DAOs can become more fair

Quantifying Decentralization

There has been a bunch of really good research done into measuring how decentralized a blockchain is. More notably, Balaji Srinivasan has co-authored a short article attempting to create a unit of measurement for decentralization, called the Nakamoto coefficient. For the purposes of understanding how we can approach DAO decentralization and decision-making parity, we will be adopting the same base logic that Balaji has.

The Nakamoto coefficient is intended to measure how many essential subsystems there are in any given decentralized network and how many entities one would need compromised in order to control the other subsystems. The higher the value of the coefficient, the more decentralized it is. While there are many variations of this coefficient used for measuring decentralization in mining, development, or nodes, for the purpose of understanding DAO decentralization, we will be looking at the two principles that guided the creation of the Nakamoto coefficient. Namely, the Gini Coefficient and the Lorenz Curve.

The Gini Coefficient

The Gini coefficient was created to measure income, wealth and consumption disparity in economics. A Gini coefficient of 0 represents absolute equality, while a Gini coefficient of 1 represents maximum inequality and centralization of resources in the hands of a single individual.

The formula has since been used to quantify the decentralization or entropy of other systems like networks, blockchain architectures or organizations. For DAOs, it would be a case of replacing income inequality with decision-making inequality, be it tied to voting weight, participation or share of voting power.

The Lorenz Curve

The Lorenz curve is essentially a graphic representation of the distribution described above. The graph shows the non-uniform distribution of resources, or in the case of DAOs control. With a Gini coefficient of 0, the graph is a straight line and represents complete uniformity of distribution. With a Gini coefficient of 1, the graph has a 90 degree angle, meaning all resources are contained by a single entity.

Decentralization is essentially a question of distribution. The more equal actors there are, the more decentralized a system is and vice versa. However, there are many other aspects to take into consideration when it comes to DAO governance and just overall democracy logistics. All users could be equal when it comes to their vote but not when it comes to participation or forward knowledge of an issue to be debated. How can we best gauge what makes a DAO not only decentralized but also as equally distributed as possible?

Measuring DAO Autonomy

There are several aspects to keep in mind while understanding how to measure DAO decentralization. The awesome paper listed below goes over a lot of aspects relating to the math behind measuring how dispersed the actual governance of a DAO really is. Combined with the other paper from above we can reach a reasonable measurement of DAO decentralization

Ownership Through Entropy

One common basis that researchers use to measure decentralization throughout DAOs is token ownership. In the end, most voting mechanisms and methodologies are relatively vulnerable to large scale actor collusion or being swayed by one large wallet holding a significant amount of the total supply. The distribution uniformity of the tokens therefore has a large impact on the actual decentralization.

As stated in the charts above, more wallets having more or less the same amount of tokens increases equality and decentralization. Conversely, the system is increasingly centralized. This is measured through one popular metric called “entropy.”

The above formula effectively assesses that for a set of addresses A = {a1, . . . , an}, with address a holding a t amount of tokens, high entropy corresponds to a more uniform distribution of governance power, whereas low entropy corresponds to high token concentration. As efficient as this method is, however, it does have its limitations.

Token ownership visibility ends where off-chain agency starts. Who actually owns those tokens is not something that can be derived from formulas. A high number of wallets governing a DAO might look like decentralization on the surface, only to, in fact, be >51% controlled by one single entity. There have been several attempts to take this factor into consideration when measuring uniformity.

Voting-Bloc Entropy (VBE) is one such solution proposed by a Cornell Tech paper. While only theoretical, it proposes an additional variable that takes into consideration how closely different addresses vote on various issues. This can be a metric that, at least in part, quantifies off-chain cartelization. Ultimately, the Oracle Paradox rears its head even in this context where there is no middle-man to ensure uncorrupted off-chain to on-chain data transfer. Individuals can ultimately create as many wallets as they wish in order to take over the minimum consensus requirement of a network.

Voting Power Distribution

Another paper from the Seoul Business School attempts a more holistic approach at quantifying decentralization. The authors attempt to take into consideration not just an overarching formula for entropy, but also take into account elements like quorum proposal pass rates or voting power index. All of these elements get measured as part of political, economic and administrative power. While focus is put on token ownership as well, this paper goes more into detail on specifically voting power distribution.

Voting power refers to the influence that an entity has on the voting outcome of a DAO. It disregards whether this control is achieved through owning multiple wallets specifically. By measuring the minimum level of voting power needed for each proposal to pass, this method helps clarify the number of participants needed for a successful outcome. Voting power can be directly tied to token-weighted voting systems, where different wallets have different voting powers, based on factors such as balance or special permissions.

A very relevant and powerful example is the recent Uniswap governance vote delay. In short, the Uniswap Foundation announced the delay of a vote intended to update the governance structure and fee mechanism. Of the 0.3% trading fees that liquidity providers receive, the proposal was to split this structure into a 0.25% for LPs and 0.05% to UNI token holders. The foundation cited concerns from a certain stakeholder, widely suspected to be a16z, regarding possible tax and securities law implications should the proposal go through.

On the other hand, there are several examples of how efficient decentralized governance can be. These examples shine literally everyday, as we see protocols like Curve, Compound and MakerDAO being able to successfully run and crowdsource decisions that have made these projects be at the forefront of DeFi development.

On the other hand, there are several examples of how efficient decentralized governance can be. These examples shine literally everyday, as we see protocols like Curve, Compound and MakerDAO being able to successfully run and crowdsource decisions that have made these projects be at the forefront of DeFi development.

Decentralization and Voting Models

Ultimately, decentralization can actually be quantified up to a certain point. Beyond on-chain data, we would need to assume good faith on the part of all participants in order to create any measurement. One solution that we at Decent particularly like is that of quadratic voting (QV). While there are many voting systems that other DAOs are experimenting with, QV has some advantages that could help mitigate uneven power distribution.

In short, QV is similar to rank-choice voting, where one individual can express the order in which they rate several issues or candidates, as well as give them the possibility to express their level of conviction for a certain issue, at the cost of having less voting power for others. Regardless of the person behind wallets, this method makes it possible for smaller wallets to effectively over power one potentially despotic wallet from dominating governance.

QV goes a long way when it comes to align incentives around several issues, balance power even if not completely and encourage voter dedication and participation. For these reasons and many more, Vitalik Buterin is also an outspoken supporter of quadratic resource distribution in governance and utility. Most of the time, this voting method is achieved by allocating a set number of “credits” to voters, let’s say 10. Considering a vote for 5 proposals, voters can allocate 2 credits per issue. For example, if they have a strong opinion on issue 2 and no opinion on issue 3, they could allocate 4 credits to issue 2 and no credits to issue 4.

QV has trade-offs, however the main issue keeping it back from being adopted by several DAOs is its risk when it comes to Sybil attacks. This is because the method considers not just how many credits were allocated to each option but also where those credits came from when calculating votes. One potential solution would be to publicly tie each voter to a single wallet via DID. DoraHacks proposes just such an alternative.

Ultimately, decentralization is different for every org. Onchain orgs navigating this for the first time should work with DAO tools and teams that have hands-on experience. Tooling partners are advisors is the model we believe wins long-term.

If you liked this piece:

Try the Decent app and refer to our docs. Questions? Message us at hello@decentdao.org and join our community on X, Farcaster, and Telegram.

As a reminder, we’re running a $5,000 giveaway for the best-governed DAO on Base as part of Onchain Summer. Learn more here: