Delegated Voting in DAOs: How Token Holders Transfer Governance Power
Decentralised autonomous organisations face an uncomfortable paradox. They promise governance by the many, yet participation rates routinely fall below ten per cent of eligible token holders. Delegated voting has emerged as the primary mechanism for bridging this gap — allowing passive holders to entrust their voting power to active participants who will exercise it on their behalf.
The concept borrows from representative democracy but operates with crucial differences. In a DAO, delegation is liquid, revocable, and transparent. Every delegation event is recorded on-chain, and the delegator can reclaim their voting power at any moment. There are no fixed electoral cycles, no geographic constituencies, and no formal campaign processes — though informal ones have certainly developed.
The Mechanics of Token Delegation
At its most basic level, delegated voting allows a token holder to assign their governance weight to another address without transferring custody of their tokens. The delegatee then votes with the combined weight of their own holdings plus all delegated tokens. The process is typically handled through a governance contract — most commonly an implementation of OpenZeppelin’s Governor framework or Compound’s GovernorBravo.
When a holder delegates, they sign an on-chain transaction that updates the delegation mapping in the governance token contract. From that point forward, the delegatee’s voting power reflects the sum of all delegations they have received. The original holder retains full ownership of their tokens and can transfer, sell, or stake them freely. However, the voting power follows the delegation, not the token balance, until the delegation is explicitly changed.
This creates an interesting dynamic. A delegate with modest personal holdings can accumulate significant governance influence purely through the trust of other holders. In Uniswap’s governance, for example, several delegates command millions of votes despite holding relatively few UNI tokens themselves. The power derives entirely from community confidence in their judgement.
Why Delegation Matters
The case for delegation rests on three interconnected problems that plague direct-democracy governance models in DAOs.
Participation fatigue is the most immediate concern. Active DAOs may produce dozens of proposals per month, each requiring technical analysis, economic modelling, and an understanding of the protocol’s strategic direction. Few token holders have the time, expertise, or inclination to evaluate every proposal. Without delegation, these holders simply abstain — and their governance power goes unexercised.
Expertise concentration compounds the problem. Many governance decisions involve complex smart contract upgrades, risk parameter adjustments, or treasury allocation strategies. A retail holder who purchased governance tokens for speculative reasons may lack the technical background to assess whether a proposed oracle migration introduces systemic risk. Delegation allows them to entrust that decision to someone with relevant expertise.
Quorum requirements create the third pressure. Most governance frameworks require a minimum percentage of total voting power to participate before a vote is considered valid. If participation is too low, even broadly supported proposals fail to pass — creating operational paralysis. Delegation concentrates voting power in active hands, making quorum achievement more reliable.
Models of Delegation
Not all delegation systems function identically. Several distinct models have emerged across the DAO ecosystem, each with different trade-offs.
Full Delegation
The simplest model allows a holder to delegate all of their voting power to a single address. This is the approach used by Compound, Uniswap, and most Governor-based DAOs. It is easy to understand, cheap to execute on-chain, and straightforward to implement. However, it forces an all-or-nothing choice — the delegator cannot split their power across multiple delegates or retain partial voting rights.
Partial Delegation
More sophisticated implementations allow holders to split their voting power across multiple delegates. A holder with 10,000 tokens might delegate 5,000 to a DeFi risk expert, 3,000 to a community advocate, and retain 2,000 for personal voting. This model is technically more complex, requiring fractional accounting in the governance contract, but it more closely mirrors how individuals actually think about representation.
Protocols like Agora have pioneered partial delegation interfaces, and the practice is gaining traction in larger DAOs where governance covers diverse policy domains.
Liquid Delegation (Transitive Delegation)
In liquid delegation, a delegate can further delegate the power they have received to another address. This creates chains of delegation — Alice delegates to Bob, who delegates to Carol, who ultimately casts the vote. The model mirrors liquid democracy proposals from political science and theoretically allows governance power to flow towards the most capable decision-makers.
In practice, transitive delegation introduces significant complexity. Circular delegation loops must be prevented, gas costs increase with chain length, and accountability becomes diffuse. Few production DAOs have implemented fully liquid delegation, though experimental frameworks like those explored by Gitcoin have tested the concept.
The Delegate Landscape
As delegation has matured, a distinct class of professional delegates has emerged. These individuals or organisations actively seek delegations, publish voting rationales, and maintain public track records. Some operate independently; others are affiliated with venture capital firms, protocol development teams, or governance-focused DAOs.
The professionalisation of delegation has brought both benefits and concerns. On the positive side, professional delegates tend to vote more consistently, publish detailed analyses, and engage constructively in forum discussions. They raise the quality of governance discourse and ensure that proposals receive meaningful scrutiny.
The concerns centre on power concentration. Data from major DAOs consistently shows that a small number of delegates control a disproportionate share of voting power. In some protocols, fewer than ten delegates command enough votes to unilaterally pass or block proposals. This concentration sits uncomfortably with the decentralisation ethos that motivates DAO governance in the first place.
Incentive Structures for Delegates
One of the most debated questions in DAO governance is whether delegates should be compensated for their work. The argument for compensation is straightforward: delegation demands significant time and expertise, and unpaid delegates either burn out or are captured by entities willing to subsidise their participation.
Several models have been trialled. Gitcoin implemented a delegate compensation programme that pays active delegates based on participation metrics. Optimism’s Citizen’s House experiments with retroactive rewards for governance contributions. ARB DAO has explored stipends for delegates who meet minimum activity thresholds.
The counterargument holds that compensation creates perverse incentives — delegates may vote frequently but superficially to meet activity requirements, or align their positions with the compensation committee rather than their delegators. The optimal design for delegate incentives remains an open question.
Risks and Failure Modes
Delegated voting introduces several risks that DAOs must actively manage.
Vote buying becomes easier when governance power is concentrated. A malicious actor need not acquire millions of tokens; they need only persuade or compensate a handful of large delegates. The transparency of on-chain delegation makes this somewhat detectable, but off-chain agreements are difficult to monitor.
Delegate apathy mirrors the voter apathy it was designed to solve. If delegates themselves become disengaged — failing to vote, skipping forum discussions, or voting without analysis — the governance system suffers doubly. The delegators assumed their power was being exercised; the delegate assumed no one was watching.
Capture by insiders is a persistent concern. Protocol teams, venture capital investors, and large token holders often appear on delegate leaderboards. When these entities accumulate delegated power in addition to their own holdings, the governance system may function as a plutocracy despite its democratic architecture.
Delegation inertia describes the tendency of delegators to set and forget their delegations. Even when a delegate’s behaviour changes — voting against community interests, becoming inactive, or facing conflicts of interest — delegators rarely reassign their power. This inertia reduces the accountability mechanism that delegation theoretically provides.
Best Practices for DAOs
DAOs implementing or refining delegation systems should consider several design principles.
Transparency mechanisms should require delegates to publish voting rationales within a defined period after each vote. Platforms like Tally and Boardroom have made this easier, but the norm should be codified in governance frameworks rather than left to voluntary compliance.
Delegation dashboards should provide delegators with clear information about their delegate’s activity, voting history, and alignment with community sentiment. If a delegate has voted against the majority position on several consecutive proposals, the delegator should be notified.
Term limits or periodic re-delegation requirements can combat delegation inertia. Requiring delegators to actively renew their delegations every quarter forces a moment of evaluation and prevents permanent power accumulation.
Delegate diversity initiatives — including delegation programmes targeted at underrepresented communities, geographic regions, or expertise areas — can counteract the natural tendency towards concentration.
The Road Ahead
Delegated voting is neither a perfect solution nor a temporary stopgap. It is an evolving mechanism that will continue to shape DAO governance as the ecosystem matures. The most promising developments lie in hybrid models that combine delegation with quadratic voting or conviction voting to mitigate concentration risks while preserving the participation benefits that delegation provides.
As DAOs govern increasingly significant treasuries and protocol decisions, the stakes of delegation design will only grow. The organisations that develop robust, accountable delegation frameworks will be better positioned to sustain decentralised governance at scale. Those that treat delegation as a set-and-forget feature risk recreating the very power structures they were designed to replace.
Donovan Vanderbilt is a contributing editor at ZUG DAO, the decentralised governance intelligence publication of The Vanderbilt Portfolio AG, Zurich. His work examines the intersection of governance design, institutional economics, and on-chain coordination.