Introduction.
DAOs can create more efficient, transparent, and objective forms of governance. They can also do the opposite.
The top ten voters in the last ten ApeCoin proposals accounted for 58% of total votes, despite representing only 6% of unique wallets participating.
The top three addresses on Compound control over 30% of the voting supply. All three are investment firms (Polychain, Bain Capital, and a16z).
How can DAOs create more equitable and sustainable structures? How can they optimize for an inherently volatile and speculative market?
This article will discuss the pain points of existing DAOs. And suggest multiple solutions to improve decentralized governance.
Voting Power.
Token-Weighted Voting
The most common voting structure in DAOs is token-weighted voting. Namely, fungible or non-fungible governance tokens as one-to-one voting sources. Examples include Uniswap, Compound, and ApeCoin.
Ideally, these governance tokens would have wide distribution and high participation rates — leading to diverse and informed voting. But token-weighted governance often creates plutocratic systems, where the wealthiest participants / earliest adopters have an outsized percentage of tokens. And where, therefore, smaller holders feel disenfranchised.
For example, 18 voters in Decentraland’s DAO account for 61% of the total voting participation. These 18 members are less than one percent of total voters.
There is no easy solution here — especially for existing DAOs. But several mechanics can be layered in to create more balanced governance.
Quadratic Voting
Quadratic voting — a mechanism that allocates voting power based on the square root of tokens — is one solution. It tempers minority interests and incentivizes participation amongst holders with fewer tokens. (Ethereum’s Vitalik Buterin discusses its merits at length.)
Unfortunately, quadratic voting hasn’t been tested at scale. Gitcoin Grants is the largest experiment, which used quadratic funding to seed over $20 million of digital public goods. However, this structure likely runs into a few challenges at larger scales.
- Vulnerability to Sybil attacks. That is, individuals can move tokens to multiple wallets at little to no cost to maintain their “token dominance.” Combining quadratic voting with verified personhood (i.e., requiring KYC) is a way to mitigate this risk. But in a space that values anonymity, few may be receptive to this structure. In the future, zero-knowledge KYC solutions — namely, being able to validate unique identities without revealing them — will likely be popularized.
- Buying elections. Quadratic voting doesn’t completely solve the problem of money-buying elections. Someone with 100 tokens will still have 10 votes. While it certainly levels the playing field, whales may continue to control voting in highly concentrated DAOs.
Until the drawbacks associated with quadratic voting are explored and resolved, existing DAOs should look to other solutions like merit-based voting to empower communities.
Merit-Based Voting
Merit-based voting — allocating more power or access to proven DAO contributors via reputation score or on-chain badges — may provide increased flexibility and improve operations. This can happen in several ways.
- DAOs can implement token multipliers for certain types of voters or proposal categories. For instance, the most committed members could have more say on general process votes, while working group members with specialized knowledge could have more say on vertical-specific funding proposals.
- DAOs can token-gate certain components of governance. They may limit proposal submissions to certain badge categories (i.e. technology-centric proposals to trusted third-party solutions providers). Or they may require voters to hold certain badges to request higher funding levels.
Designing merit-based structures need to be curated for each DAO. Every community will define top contributors differently. Some may base this on historical participation, while others may derive it from high-quality engagement via Discourse. But if structured correctly, DAOs could see significant improvements in efficiency.
Voting Structures.
Perfect Information Voting
Unlike traditional politics, DAO voting is often said to be a game of perfect information. Specifically, participants have access to information about every vote cast (which wallet, how many tokens, what time, etc.)
This can be beneficial for several reasons including:
- Increased transparency
- Public analytics and reporting
- Greater communication between voters and candidates
But this level of transparency may also create non-optimal outcomes in DAO elections and multi-choice proposals.
First, certain voters may not want to publicly vote for a particular side.
Second, seeing how other DAO participants cast votes may impact decision-making.
Third, with the ability to switch votes during multi-choice voting, large holders can strategically move tokens to influence how others vote.
Perhaps more detrimental is the role that timing can play in perfect information elections. As one example, if DAO electoral candidates do not receive enough votes at the beginning of an election, voters may assume they will lose and allocate votes to their second preference. In this case, even if a candidate theoretically has enough support to advance before voting begins, the perceived lack of support may sway voters elsewhere.
Anonymous Voting
Anonymous voting can avoid many of the problems listed above. Specifically, it can:
- Eliminate herd mentality
- Protect against intimidation or undue influence
- Limit external biases
Secret ballots have been tried and tested for centuries. Combining this age-old election model with technological advancements (zero-knowledge proofs, encryption, etc.) could lead to a more optimal voting structure.
Single Choice
In single-choice voting, each voter allocates tokens to one candidate, and the candidate with the greatest number of votes wins. Simple enough.
But if the fundamental requirement of an electoral system is to accurately represent the views of all voters, then single-choice voting may not suffice. The majority of constituents’ preferences are often discounted in favor of one (usually polarizing) choice. This is the case in many traditional political elections, such as the United States presidential primaries.
Ranked Choice
Ranked choice is a voting structure that can theoretically create a more representative vote. Voters can rank multiple candidates in order of preference, and each candidate is assigned a number of points based on that rank.
New Yorkers elected to use ranked-choice voting in 2019 (with almost 75% in support). The benefits were quickly realized. Only 15% of voters had inactive ballots in the final elimination round of the 2021 mayoral primaries. By contrast, 33% of voters cast ballots for candidates who didn’t make the top two spots in the 2013 mayoral primaries. Here, ranked-choice voting reduced “wasted” votes.
No structure is without its drawbacks, however. Ranked choice elections often fail to consider voters’ levels of conviction in elections. Results may appear much closer than voter preferences would suggest. Said another way, the difference in preference between one’s first and second favorite choice is probably much wider than the ranked-choice point system indicates.
Voting Participation
Delegation (Liquid Democracy)
Most DAOs are lucky to get over 10% participation. This means a minority of token holders are overrepresented in the DAO. To achieve democratic governance, DAOs need to encourage all holders to participate — or else set up mechanisms to avoid decision-making by minorities.
Delegation is one solution. While it doesn’t necessarily increase the unique number of voters, it does increase community representation.
ENS, which requires delegation upon the token claim, receives 10–15% participation on most proposals. Decentraland implemented delegation in July 2021 and has seen total monthly votes increase ~4x. Now, delegation accounts for ~76% of total votes cast in Decentraland DAO.
DAO Rewards
DAO rewards or incentive models are another way to encourage participation. These may take the form of financial rewards (DAO token distributions, bounties, gift token models, etc.) or cosmetic rewards (on-chain badges and reputation scores).
Financial rewards may be easily gamed by DAO participants. As such, most notable DAOs will likely lean on cosmetic rewards first. Reddit is a strong example of a platform tokenizing its reputation system (karma).
Quorum
Other than incentivizing participation, DAOs can require proposals to meet quorum to avoid minority decision-making.
Defining quorum thresholds can be difficult. Participation is volatile, and circulating token supply is difficult to gauge given the amount moved between exchanges and wallets. Moreover, for DAOs that have difficulty achieving a quorum, it may be in a voter’s best interest to abstain rather than vote against. We’ve seen this time and time again in Doodles’ DAO. To prevent this, Nouns DAO requires a floating quorum that increases based on the number of nays a proposal receives.
Motivation
Of course, the greatest way to incentivize participation and avoid minority decision-making is to shift holders’ motivations from extrinsic to intrinsic. That is, change the culture and ethos of the DAO to be defined by a common goal or mission. We want people to vote because they believe in something greater than themselves, rather than monetary incentives. Easy to understand, yet difficult to achieve.
Conclusion.
There is no single voting structure or ideology that works for every DAO. No copy-and-paste solutions. Platforms have different goals, requiring different governance structures. And communities have different compositions, requiring custom voting solutions.
But understanding the nuances of each structure can help DAOs better tailor their governance to the needs of their constituents. And ultimately, allow nascent communities to experiment and iterate on their governance models.
