A Quick Overview of the Current Status, Aspirations, and Possible Evolution Directions of Star DeFi Protocol Governance

Author: Ellen Guo, Analyst at e^{i} Ventures

Compiled by: Perry Wang, Chain News

Humanity has been conducting grand experiments in governance systems to coordinate decisions that affect their collective future. Governance is a complex and challenging process that underpins nearly all human endeavors. Today, decentralized finance (DeFi) represents a convergence of thousands of years of human experience with the speed and scale of internet communities, incorporating powerful innovations of programmable money and trust-minimized transactions.

Over the past year, DeFi has fundamentally transformed the cryptocurrency world. DeFi applications not only pursue trustless and permissionless transactions but also achieve decentralized decision-making through governance. The goal is not for a centralized core team of trusted experts to make decisions for the user community, but for the community to directly determine changes through participation and voting.

To gain a deeper understanding of this latest chapter in humanity's grand experiment, I investigated the current state of decentralized governance to identify what works, what revolutions have yet to succeed, and where comrades still need to strive.

Background

Decentralized protocol governance refers to how participants in a protocol collectively make decisions about its future. Participants in DeFi protocols typically earn governance token rewards when they perform actions beneficial to the protocol. The number of governance tokens held by participants is usually associated with their voting power. So far, most decentralized protocols follow a one-token-one-vote approach, where each token grants its holder voting rights on community proposals. Proposals are typically aimed at modifying and upgrading the protocol.

The launch of COMP by Compound in June 2020 initiated a wave of yield farming and rapidly evolving DeFi governance. Since then, many other protocols have adopted various novel governance methods.

Forms of Governance

DeFi protocol governance can be divided into two main paradigms: on-chain and off-chain.

On-Chain Voting
The on-chain governance system pioneered by Compound involves the entire process of proposal, voting, and implementation occurring on the blockchain through smart contract interactions. Proposers use a list of executable commands to call the contract's propose function, and if the proposal is approved, these commands will execute. Voters interact directly with the smart contract to cast their votes. At the end of the voting period, the votes are tallied, and the proposal is either executed on-chain or declared failed.

Since on-chain voting requires proposals to be executable code, simply executing the code ensures that successful proposals are implemented. The result is secure, with the integrity of the final tally protected by the strong security of on-chain consensus.

However, submitting proposals in an on-chain governance system requires participants to have a high level of technical capability. Additionally, depending on the specific mechanisms of the underlying chain, proposers and voters may incur significant transaction fees to participate.

Off-Chain Voting
Off-chain governance systems allow users to submit proposals and vote without submitting blockchain transactions. The most common off-chain voting platform, Snapshot, is open-source software that stores proposals and votes as signed messages on IPFS. Snapshot allows protocol team members to flexibly create their own spaces on Snapshot, including setting rules to determine token holders' voting rights.

Off-chain voting avoids transaction fees, encouraging broader participation from community members, but it requires a trusted third party to tally the votes.
Moreover, implementing successful proposals is not as straightforward as executing code. Approved proposals may require the core team or other representatives to take off-chain actions, such as writing new code or documentation, hiring personnel, obtaining legal opinions, or conducting off-chain financial transactions.

Basic Information: Background of DeFi Protocol Governance

Each protocol's development history and token economics differ, which can help us determine what works and what has not succeeded.

Governance has been evolving, and the data in this article is updated as of August 19, 2021.

Compound
Compound's on-chain governance launched on April 16, 2020, with the GovernorAlpha smart contract. After Proposal 42 (March 2021), the community migrated to GovernorBravo.

Initially, any address that delegated more than 100,000 COMP to the protocol could create proposals (i.e., call propose). The 52nd proposal executed on July 14, 2021, lowered this threshold to 65,000 COMP. Proposals must pass by an absolute majority and require a quorum of > 400,000 COMP.

Compound also implemented Compound Autonomous Proposals (CAP) in September 2020 to lower the threshold for creating proposals, as the cost of acquiring the required COMP for proposals was substantial (over $41 million at the time of writing). Now, anyone holding 100 COMP can create a CAP, which is a smart contract that allows other community members to delegate their votes to it. Once the delegated votes meet the COMP threshold, a formal on-chain proposal is automatically initiated.

Past Proposals

Total proposals: 57
Executed: 48
Failed: 4
Canceled: 4
Voting in progress: 1

Token Economics

Compound's governance revolves around the COMP token, with a maximum supply of 10,000,000. The initial distribution is as follows:

2,226,037 COMP (22.3%) allocated to founders and the team will be unlocked in phases over 4 years. COMP is released at a rate of 2,312 COMP/day, distributed proportionally to the interest accrued in each market (see here for current details). In each market, 50% goes to lenders and 50% to borrowers, with users earning COMP proportional to their balances.

Uniswap

Uniswap's on-chain governance is a fork of Compound's GovernorAlpha (although there are currently proposals to migrate to GovernorBravo), and its functionality is largely similar to Compound's version. When the protocol launched governance in September 2020, the proposal threshold was set at 10 million UNI, which was later reduced to 2.5 million UNI in Proposal 4. The required quorum has always been 40 million UNI.

In addition to on-chain proposals and voting, Uniswap encourages off-chain discussions before creating on-chain proposals. You can click here for more details. In summary, Uniswap designates a discussion forum and conducts Snapshot Temperature Checks (if the support votes exceed half and reach a quorum of 25,000 UNI, it passes), followed by forum and Snapshot Consensus Checks (if the support votes exceed half and reach a quorum of 50,000 UNI, it passes).

Past Proposals

Total proposals: 7
Executed: 3
Failed: 2 (both proposals met the support vote ratio but did not reach the UNI quorum)
Canceled: 1
Proposal in progress: 1
Total off-chain proposals: 29
Passed: 18 (13 Temperature Checks, 5 Consensus Checks)*
Failed: 11 (9 Temperature Checks, 2 Consensus Checks)

* Unmarked proposals are assumed to be Temperature Checks.

Token Economics

When the UNI token launched on September 16, 2020, a total of 1 billion UNI was minted. The initial 4-year distribution is shown in the following chart. UNI provides a phased exit plan for investors, team members, and advisors over 4 years. The inflation rate (2%/year) will be introduced after 4 years.

When the UNI token was first issued, the community's share (60%) included a quarter (i.e., 15%) airdropped to liquidity providers (LPs), users, and SOCKS holders' wallets, which was a well-known event that caused a stir in the crypto community. The governance treasury retains 43% of the share for allocations to contributor grants, community projects, and liquidity mining. This distribution is managed by community governance, with the last 2% of UNI held by the community released through liquidity mining from September to November 2020.

Sushiswap
Unlike Compound and Uniswap, Sushiswap's governance is entirely conducted off-chain using Snapshot. Currently, only proposals made by the core team—0xMaki (co-founder), Omakase (core developer), 0xJiro (core developer), or Ross Campbell (lawyer)—are considered binding if they pass the quorum (5 million SUSHIPOWAH—see token economics). Once a "binding" proposal is approved, the Sushi team must manually implement the changes, depending on whether the team is online.

As another security measure, any proposals using the devfund wallet require approval from at least 6 signers from the platform's multi-signature group (@SBF_Alameda, @rleshner, @0xMaki, @lawmaster, @cmsholdings, @mattysino, @mickhagen, @JiroOno, @zippoxer), and any operational changes, such as asset rebalancing and management of yield farming pools, require signatures from at least 3 of @0xMaki, @LevxApp, @OmakaseBar, @0xJiro, @0xKeno, @josephdelong, @MatthewLilley.

However, this off-chain Snapshot voting (initiated by core team members) and multi-signature structure is only temporary, as the team's 2021 development roadmap indicates that Sushi aims to transition to a mature DAO by the fourth quarter of 2021.

Past Proposals

Total proposals: 243
Total core team proposals: 61*

* All future analyses of Sushiswap proposals will only consider core team proposals. I also decided to exclude multi-signature election proposals, as the results pertain to multiple proposals rather than voting on a single proposal. For election statistics, see here.

Token Economics

Voting power is determined by SUSHIPOWAH balances. Each SUSHI in the SUSHI-ETH pool is worth 2 SUSHIPOWAH, while each SUSHI held through xSUSHI is worth 1 SUSHIPOWAH.

The supply of SUSHI is strictly capped at 250 million, with new SUSHI created at a rate of 100 SUSHI/block, produced from tokens staked in yield farming. After staking SUSHI, users receive xSUSHI and earn 0.05% of transaction fees from all pools on the exchange.

Yearn
The governance mechanism of Yearn has undergone significant changes since its inception. The Yearn protocol adopted an on-chain governance structure, with YFI launched on July 17, 2020. Before the implementation of YIP-10 (July 24), two proposals—Yearn Improvement Proposal (YIP) 1 and 12—were passed to transfer governance to a new smart contract for security reasons. Before the community decided to migrate governance to off-chain Snapshot (click here for Snapshot archives, click here for current Snapshot), the new contract handled YIP 30-45, and Yearn governance migrated off-chain to avoid transaction fees and encourage higher community participation.

A notable feature of Yearn governance is the participation incentive mechanism. According to YIP-45, proposals that reach the "implementation" stage will receive a bounty of $500 worth of yCRV, paid from the protocol's treasury.

Past Proposals

Total proposals: 37
On the first governance contract (on-chain): 7
Executed: 3
Failed: 4
On the updated smart contract (on-chain): 16
Executed: 11
Passed: 12 (of which 1 proposal is pending implementation)
Failed: 4
Off-chain: 15 (all passed but not yet implemented)

Token Economics

The YFI governance token was launched on July 17, 2020, with a total of 30,000 YFI minted. Users earn YFI by providing liquidity to Yearn products, staking the tokens output in the YFI allocation contract, and participating in governance daily. Staking YFI not only grants governance voting rights but also earns a share of the profits generated by Yearn.

Gemini believes that the issuance of YFI was fair, as no YFI was reserved for the founding team, project investors, advisors, or other insiders.

Additionally, 6,666 YFI were minted as part of the YIP-57 proposal (⅓ to major contributors as phased exit rewards, ⅔ to the protocol treasury).

What is the Current State of Governance?

I set four parameters to evaluate how these decentralized protocols are functioning in governance practices, supplemented by qualitative research on community participation and discussion.

The four parameters I established are: distribution of voting power, influential voters, sources of proposals, and voting participation rates.

Severe Inequality in Voting Power Distribution

First, I examined how voting power is distributed among users in each protocol. Despite differences in token economics, each of these protocols exhibits a high and comparable Gini coefficient, indicating severe inequality in the relative voting power of protocol users.

In each protocol, less than 1% of voters hold 50% of the delegated votes. Since voting results are based on strict majorities, this means that less than 1% of voters can determine the outcome of each proposal in these protocols. The vast majority of votes (95%) are held by less than 3% of voters.

A Few Influential Voters Control Voting Outcomes

The unequal distribution of voting power has profound implications for the outcomes of proposals. For each proposal, I found the number of addresses that cast 50% or more of the total votes. Since the voting results of the protocols considered in this article are based on strict majorities, these voters have the greatest influence on the outcome of each vote.

We can see that the outcome of almost every proposal is determined by 2-7 voters. Yearn's governance (both on-chain and off-chain) shows a longer right tail in the distribution of influential voters. In other words, some proposals in Yearn require more top voters to determine their outcomes than other protocols. Despite this long tail, on average, 5-7 (up to 20) Yearn voters can influence the outcome of a proposal.

Proposal Initiators Tend to Be Core Teams or Whales

Naturally, I wondered who is submitting proposals? I wanted to see how many addresses in each protocol are proposing.

Keep in mind that there are technical and financial barriers to submitting proposals. For on-chain governance (Uniswap, Compound, Yearn's old system), writing and submitting proposals requires technical background (calling the propose function of the governance contract and writing executable functions in the proposal), and most protocols require proposers to hold tokens that meet the threshold requirements to prevent low-quality spam proposals.

As of mid-July, Compound's proposal threshold was 100,000 COMP, and Proposal 52 lowered this threshold to 65,000 COMP (approximately valued at ~$43,700,000 at the time of writing). However, with the introduction of CAP, anyone only needs to hold 100 COMP (~$43,700) to initiate the proposal process. Uniswap's proposal threshold was 10 million UNI (~$260 million) as of mid-June, and Proposal 4 reduced this threshold to 2.5 million UNI (~$65 million); meanwhile, the fish.vote platform allows anyone holding 400 UNI (~$10,400) to create a governance proposal. To create a proposal on Yearn's Snapshot, proposers must hold at least 1 YFI (~$36,900); however, if your proposal gains significant attention on the Discourse governance forum, Yearn team members will proactively help you submit the proposal. As for Sushiswap, anyone can propose suggestions on their Snapshot, but only suggestions created by core team members are considered binding (and will be implemented).

In this context, let's investigate the backgrounds of proposers in each protocol.

Overall, we see that most proposals are initiated by 3-7 addresses; similar to the situation with influential voters, proposals have consistently been products of small circles.

For Compound's 57 proposals, there were 17 unique proposers. The top 3 proposers by number of proposals are whales in the DeFi space: blck proposed 19, Geoffrey Hayes proposed 8, and Gauntlet proposed 7.

For Uniswap's 7 on-chain proposals, Dharma proposed the first two, followed by two autonomous proposals, and then one each from Harvard Law School's Blockchain and FinTech Project, Dave Balter, and Getty Hill.

For Sushiswap, only 4 core team members (as defined in Snapshot) can create binding Sushiswap Snapshot votes.

For Yearn, each proposal also includes a list of proposers involved in drafting the proposal, not just the address submitting the proposal. Proposers are also concentrated among the founding/core team: banteg (who wrote 9 proposals), lehnberg (8), milkylkim (8), Andre Cronje (6), Substreight (4), lex_node (4)—all members of the founding/development/operations/legal team or current/former signers of the multi-signature.

Low Voting Participation Rates

Once proposals are activated, how engaged is the community?

Statistically, there is no significant difference between participation rates and the distribution of voting power, but the larger discrepancies indicate that some proposals achieve high participation while voters are less active on others.

For Compound and Uniswap, this is the participation rate of delegated tokens (you must delegate your tokens to vote with them; you can delegate to yourself). Based on the total circulating tokens, the participation rates of delegated tokens AKA are 43% and 19%, respectively. In other words, on average, 43% * 45.6% = 19.6% of circulating COMP tokens are used for voting, while this proportion for UNI is 19% * 52.7% = 10%.

However, I believe that data analysis cannot capture the spirit and cultural essence of decentralized governance, so I want to discuss two recent cases that have emerged in DeFi governance.

Case Study: Uniswap DeFi Education Fund

Uniswap's Proposal 5 to donate 1 million UNI to the DeFi Education Fund sparked controversy as a "governance theater" piece. Community members criticized the proposal for its lack of transparency—why would such a proposal, which hardly detailed what it would do, receive so much funding? Important questions raised by individuals like Chris Blec from DeFi Watch went unanswered, despite the proposal promising that 500,000 UNI would be liquidated over 4 to 5 years, with these tokens sold out in June. However, due to a few addresses controlling the voting power, the proposal passed: mainly supported by student groups (such as the proposer Harvard Law School's Blockchain and FinTech Project), but backed by a16z.

A user in Uniswap Discord wrote, "I think this shows that if an organization and its allies can arbitrarily draw blank checks from the treasury, then DAOs and treasuries face the risk of centralized control." In fact, dissatisfaction spread to the point of irony, initiating a meme-style proposal for a "haircut fund" on the governance forum. Yearn's core developer Banteg wrote, "Although the initial reaction was to dismiss it as a joke, I think it could be a good proposal to activate the voters of the protest group and change the governance token that has essentially died… I fully support you."

Case Study: Sushiswap "Phantom Troupe"

On the Sushiswap governance forum, 0xMaki proposed to sell 51 million SUSHI tokens to institutional investors at a discount. A complete report on this situation (referred to as the "Phantom Troupe") can be found in a DeFi Weekly article.

In summary: "Selling a large amount of treasury assets to investors with no clear value-add at a significant discount during the historical low of the $sushi token, without a clear reason, made the community very angry."

This emotional outburst led to more favorable negotiations for Sushi, but ultimately the proposal was completely withdrawn. As the Chief Investment Officer of Arca wrote in a forum post, "This is a case study in successful governance, and we are pleased to be part of both history and the future." Or more powerfully, as DeFi Weekly wrote, "The power of the community has now risen to a level where it can negotiate openly with many institutional investors."

The Future Path of Decentralized Governance

Most of the data I collected seems to indicate that current DeFi governance is still in a transitional state, yet to fulfill the promises of trustlessness, permissionlessness, and complete community control. I found that voting power, influence, and the right to initiate proposals are concentrated in the hands of a few, and the actual participation rates show a lack of widespread community engagement.

However, it is worth noting that decentralized governance is experimental and still in its infancy. Our findings support an intuitive conclusion that a certain degree of centralization is needed in the early formation stages of any project, as the community is still gradually approaching critical mass, and the project is still solidifying its position within the overall DeFi economy. Through the use of formal and informal techniques for coordination, DeFi governance is rapidly evolving while striving to balance decision-making efficiency with the decentralization of power.

Token Economics

One of the loudest criticisms of DeFi governance is the unequal distribution of voting power, leading to a concentration of power in the hands of a few. While this is a characteristic of early DeFi governance, I believe there are many more stories to tell.

Although the token distribution structures in these protocols differ, they lead to similar concentrations of voting power, and how voters perceive the ballots they hold seems to be important. For example, Yearn's relatively fair distribution may not significantly reduce the Gini coefficient, but it may give ordinary voters confidence that large VCs or core team members will not dominate the voting; this assumption could explain why ordinary voter participation in Yearn (both on-chain and off-chain governance votes) is higher.

Another aspect of token economics is distribution. I observed that Uniswap has less community activity compared to Sushiswap, which may be due to this reason. The only way users received UNI during the initial distribution was through retrospective airdrops or a 3-month liquidity mining program (which still only accounted for 17% of the total issuance); all remaining UNI was either locked in the treasury or locked for the team/investors. This exclusive distribution may signal that participation is very difficult and only applies to early adopters, effectively excluding a large portion of users from governance. In contrast, SUSHI (and SUSHIPOWAH) is continuously distributed to users of the protocol.

Quadratic voting is one of the commonly suggested ideas to address the issue of unequal voting power distribution. Although it is not within the scope of this article, I believe it is something that those investing in governance should consider; Vitalik's "Quadratic Payments: A Primer" is a great starting point.

Identity of Proposers

Although proposers are clearly concentrated within small circles, voters are more inclined to trust proposals written by well-known DeFi figures rather than those created by autonomous community members, which is not without reason, especially when on-chain proposals require technical knowledge, or more importantly, when implemented proposals could significantly impact the protocol, for better or worse. Secondly, most protocol leaders respect governance forums, and official proposals (on-chain or off-chain) usually reflect community opinions. In the Sushiswap case study, the negative community reaction meant that core team members did not propose binding Snapshot votes.

I believe that the introduction of autonomous proposals in Compound and Uniswap is a positive mechanism to encourage more decentralized proposers— even if community members do not have enough tokens to directly propose something, they can still submit on-chain proposals as long as they gain sufficient community support. Autonomous proposals are a way for holders with fewer tokens to feel heard, which is crucial in decentralized governance: the feeling of being heard directly affects loyalty and participation. Just like in traditional politics, when voters feel their voices matter, they are more likely to show up to vote.

If we look at Uniswap, there have already been 2 autonomous proposals since the introduction of fish.vote. The introduction of autonomous proposals feels promising for democratizing proposers, but perhaps this is because the previous proposal threshold (10 million UNI) was excessively high.

Moreover, after the proposal threshold was lowered to 2.5 million UNI, I found that the UNI held by Dave Balter and Getty Hill was far below the initial proposal threshold—10 million UNI, indicating that lowering the threshold indeed allows participants with fewer tokens to propose suggestions. In other words, lowering the proposal threshold will allow for a greater diversity of proposers; of course, this may come at the cost of more spam proposals, but using Yearn as an example, its proposal threshold is the lowest compared to previous protocols (1 YFI is approximately $36,900), yet so far only productive proposals have been seen.

In addition to autonomous proposals, a strong governance forum seems to also promote the healthy development of decentralized governance. Anyone, regardless of how many tokens they hold, can comment and share ideas on the proposal process, making the forum a very powerful resource. However, it is also crucial that official proposals reflect the opinions of the forum; in the Uniswap case study, Harvard Law Blockchain ignored many concerns and complaints voiced in the forum, leading to a loss of trust in Uni governance and diminishing their enthusiasm to contribute to protocol governance.

Transparency

In my personal view, the most direct and impactful improvement that DeFi projects can make to further their decentralization goals is to reduce barriers to community participation by improving governance documentation and enhancing transparency. During my research, I found that this information can be difficult to find or outdated. Questions about how governance processes work are a common topic in governance chat channels, especially with more inquiries coming from new community members. In my opinion, protocols should focus on providing comprehensive and accessible governance process documentation and help new members get started with proposing, voting, or delegating votes.

Conclusion

In summary, I am excited to observe how DeFi governance continues to evolve from its current foundation. What I love most about DeFi is that everyone can participate in improving projects and the entire ecosystem. If you are reading this, I bet you are already interested in or involved in decentralized governance. What are your thoughts on DeFi governance? How do you think we can improve it?