Uniswap on-chain voting research: Insights on power, apathy, and evolution

Author: Chao

Source: X, @chaowxyz

Reprinted: White55, Mars Finance

It was supposed to be an ideal decentralized utopia, but the data reveals a digital oligarchy controlled by 1%. We reviewed all on-chain votes of Uniswap over the past four years, uncovering the shocking truth behind the Uniswap governance utopia.

In November 2021, Uniswap, the giant of decentralized finance, launched a highly anticipated governance mechanism: a digital democracy system where UNI token holders collectively decide the platform's future. It paints an enticing vision: a purely democratic utopia without a CEO, no board of directors, where power transparently belongs entirely to the token holders.

However, a four-year in-depth investigation into the Uniswap decentralized autonomous organization (DAO)—a comprehensive quantitative analysis based on 21,791 voters, 68 governance proposals, and 57,884 delegation events—revealed a surprising reality: digital democracy has evolved into a highly centralized digital oligarchy in practice, and the delegation mechanism intended to improve governance may instead be counterproductive, exacerbating inequalities and suppressing participation.

This research not only reveals the complex nature of digital governance but also challenges many of our fundamental assumptions about decentralized autonomy, providing profound insights for the future development of the cryptocurrency sector and even traditional democratic systems. It is not a romantic tale about pure democracy, but an epic about how humanity organizes itself under new tools, balancing efficiency and fairness.

1. The Harsh Judgment of Digital Oligarchy Politics

The judgment of data is ruthless. The average Gini coefficient of Uniswap governance reaches 0.938, which is more unequal than the wealth distribution in nearly any country on Earth. The facts are astonishing:

• The top 1% of voters control an average of 47.5% of the voting power, and in some extreme proposals, this can reach as high as 99.97%.

• The top 10% of voters stabilize control over 91.4% of the decision-making power, rendering the vast majority of token holders effectively irrelevant in the decision-making process.

The power structure of governance on the Uniswap chain

This concentration of power is not accidental; it is a natural manifestation of the token-weighted governance system in reality. Accompanying this is a worrying low participation rate: over four years, the median voter has only cast 1 vote, while the top 10 most active voters have an average of 54 votes each. The monthly participation rate plummeted by 61% from its peak in 2022-2023, indicating that the legitimacy of governance is facing a survival threat. We are approaching a critical point where possibly fewer than 200 people are routinely deciding the fate of a protocol worth hundreds of billions of dollars.

2. "Consensus Theater": Indifference is more dangerous than opposition

Despite the highly centralized power, Uniswap's proposal success rate is as high as 92.6%.

It must be acknowledged that proposals are mostly discussed in community forums and undergo off-chain voting for "consensus checks" via Snapshot before on-chain voting. This "negotiated consensus" mechanism is one of the reasons for its efficiency and high level of consensus. However, on-chain data still reveals a deeper issue:

94.2% of voters are loyal "supporters", with an average support rate of up to 96.8%.

The failure of the proposal is 100% due to the failure to meet the minimum voting weight threshold, rather than a majority opposition.

Controversial Analysis of Proposals

Meaningful dissent is exceptionally rare, with only 2 proposals facing more than 20% opposition. The failure of proposals is not due to opposition but rather indifference— all 5 failed proposals stemmed from a failure to meet the quorum, rather than a majority opposition. This reveals a profound truth: whether in digital or traditional democracy, the true enemy is not division, but the indifference of participants. Convincing people that you are right is less effective than persuading them to care enough to participate.

Three, The Hidden Structure Under Power and the Voter Ecology

The governance of Uniswap is not a single, flat structure, but rather a nested and intricate ecosystem.

Through network analysis, we revealed a "shadow governance structure" that operates through delegation. A total of 5,833 delegation events constructed a complex network, but it is highly fragmented, with 623 components of weak connections, forming **"island governance" — isolated islands of influence rather than a unified democratic system.

At the same time, the evolution of the network presents a "the rich get richer" model: 85% of new commissions flow to existing large agents, and the status of top agents remains stable over 3.8 years. The prominent feature of its "star structure" (87.5% are pure commissioners, 11.6% are pure trustees) clearly outlines the distribution of power around a few central nodes.

Voting Delegation Network Analysis

A deeper analysis also identifies different typical voter types, forming Uniswap's "Five-Tier Voter Ecosystem":

• Whale Voter (0.8%): Extremely high weight, low-frequency participation, yet possesses the ability to instantly determine the outcome.

• Active Governors (3.2%): High weight, high frequency participation, are the backbone of governance.

• Institutional Participants (1.5%): Medium to High Weight, Selective Participation.

• Technical Expert (4.1%): Medium weight, focusing on technical proposals.

• Follower (15.8%): Low weight, follow the mainstream.

• Silent (74.6%): Extremely low weight, minimal participation, representing untapped governance potential.

Voter Profile

These different tiers of voters operate with their own incentives, levels of information, and modes of participation. Interestingly, lifecycle analysis of voters shows that as experience grows, voters become more independent, but at the same time are more inclined to delegate—this explains why experienced participants tend to reduce direct voting. Additionally, different types of proposals exhibit different power structures: technical deployment proposals have the highest concentration of power (Gini coefficient of approximately 0.997), while governance reform proposals have the lowest concentration (Gini coefficient between 0.78 and 0.92). This indicates that Uniswap actually operates under "four different governance systems" based on the type of decision.

4. The Delegation Paradox: The Backlash in Good Faith Design

However, on top of all these findings, there is an even more shocking "plot twist": the entrusted system aimed at democratizing governance may be making the situation worse.

The delegation mechanism is widely regarded as a remedy for the "lazy" problem of token holders. In theory, it should increase participation, improve decision-making quality, and reduce inequality by allowing token holders to delegate their voting rights to experts or community leaders. It sounds great, but the data tells a different story.

To understand the true impact of delegation, these four scenarios can be viewed as four "simulated replays" of the same vote, each time changing one key variable:

Scenario 1: Ideal Democracy ( theoretical benchmark ) assumption that all token holders vote personally. This represents the theoretically most democratic and equal upper limit.

Scenario Two: Current Situation ( Reality Benchmark ) refers to the actual occurrence: a portion of people vote directly, while another portion delegates their votes to "representatives" to vote.

Scenario 3: No Delegation in Reality ( Key Comparison ) This is a crucial thought experiment: Suppose the delegation function is disabled, and the original group of "representatives" can only vote with their own ballots. At the same time, we assume that 10% of the ordinary people who originally chose to delegate are activated and decide to vote in person. This represents the most realistic alternative.

Scenario 4: Voting only represents ( Minimized Benchmark ) Assumption that only the currently active "representatives" are voting, and they can only use their own tokens, with absolutely no delegated votes. This represents the lower limit of participation.

• Compared to the non-entrusted system, the current entrusted system has increased inequality by 6.6%, with the average Gini coefficient rising from 0.881 to 0.943.

• Compared to the non-delegated system, the delegated system reduced the number of participants by 88% (, with an average of 267 participants per proposal compared to 503 participants ).

• All 10 test proposals exhibited the same pattern, confirming 100% consistency of this finding.

The delegation paradox is born: the delegation system simultaneously reduces the equality and participation of governance.

Why does this happen? The fundamental reason for the paradox lies in the misunderstanding of human behavior. The traditional view holds that delegation can increase participation through representation, but the reality is:

1. Delegated centralization of power: It concentrates the voting rights of multiple token holders into the hands of a few delegates.

2. Reduce effective participants: Tens of thousands of delegators may ultimately be represented by only a few hundred active delegators.

3. Create artificial scarcity: there are only a limited number of "trusted" delegates.

4. Suppressing Direct Participation: The delegation mechanism creates a psychological effect where people think "others will handle it," thereby suppressing the willingness to participate directly.

In a real non-delegated system, delegators will still vote with their own tokens, while some token holders who would originally delegate will choose to vote directly. The end result will be more participants and more decentralized power. This system, designed to democratize governance, may actually move in the opposite direction.

V. The Dynamic Evolution of Democracy: The Self-Regulation of Oligarchy and the Glimmer of Hope

Despite the extreme inequality and delegation paradox, the study also found an encouraging trend: Uniswap is gradually moving towards democratization. Over the course of 3.8 years, the average Gini coefficient decreased from a peak of 0.990 in 2022 to 0.913 in 2025, achieving an 8.1% democratization, while the proposal success rate remained above 77%.

September 2024 is a coefficient change caused by a specific proposal, and does not represent the overall situation for the entire year of 2024.

This indicates that the token weight system has an inherent potential to evolve naturally towards greater equality without formal rule changes through learning and evolution. Perfect blockchain democracy may be an unattainable utopia, but digital oligarchy is not permanent; it may represent a transitional phase towards more democratic governance. (Important Note: The comparative data is based on actual votes from existing proposals, supplemented with simulated data formed under reasonable assumptions. It aims to provide a trend insight, but does not fully equate to the real situation; its model assumptions should be considered in interpretation.)

6. Profound Implications for Future Governance and the Path Forward

Considering all findings, Uniswap's governance model can be described as an efficient, stable, yet highly elitist "Plutocratic Republic". It excels in driving protocol technological iteration and fund management, but there is a significant gap between it and the democratic ideals of a decentralized community.

The governance of Uniswap, which combines the efficiency of oligarchic politics, broad legitimacy, economic consistency, and evolutionary capability, bears a striking resemblance to the historical Venetian Republic. The Venetian Republic sustained itself for a thousand years by balancing these forces, and perhaps Uniswap has inadvertently recreated a time-tested governance model—not purely democratic, but a form of functional democracy that is effective in practice.

However, they force the industry to rethink:

Is the default status of the delegation mechanism reasonable? It may not be a universal solution, but rather a "prescription" that needs to be used cautiously. Do not blindly assume that delegation can improve governance outcomes; its effectiveness should be verified through empirical analysis. Do not blindly assume that delegation can improve governance outcomes; its effectiveness should be verified through empirical analysis.

Should the optimization direction of DAO governance shift from "optimizing delegation" to "incentivizing direct participation"?

Should we design new governance modules, such as liquidity democracy and quadratic voting, to balance the systemic flaws of the existing delegation system?

The story of Uniswap is not a case of failure, but a valuable sample filled with real-world data and lessons learned. However, it is hopeful that these systems can evolve, improve, and gradually democratize. We are not stuck at the first version of digital governance, but can learn, adapt, and build better systems.

Uniswap's governance, despite its inequalities, has achieved extraordinary results: a 91% proposal success rate, ongoing democratic evolution, broad legitimacy, and consistency in economic interests. This may not be the perfect democracy we envision, but it could be a more valuable form of functional democracy.

The governance experiment of Uniswap provides an unparalleled real-world laboratory that allows us to study how human societies organize under new collective decision-making tools in a completely transparent manner. Digital oligarchy is not a design flaw, but rather a characteristic of how humans naturally organize when faced with new tools. Understanding and embracing this reality, rather than opposing it, may be the key to building a new generation of organizations and governance systems. Future governance, whether digital or traditional, will be built on the valuable lessons we learn today from these early experiments in decentralized democracy.