Delegation lets a token holder hand their voting weight to another address while keeping the tokens. The delegate votes with that weight until the holder revokes it or points it somewhere else. Protocols adopt it because token-weighted governance asks every holder to form a view on collateral factors and contract upgrades, and almost none of them do. What delegation changes is who reads the proposal, not how much of the supply shows up.
Delegation does not fix turnout. It relocates it. In the four large DeFi DAOs where turnout has been measured directly on chain, roughly three addresses could carry most proposals on their own, and delegation is the machinery that puts weight in their hands.
Delegated weight, not delegated tokens
The token stays put. A holder calls a delegation function that assigns their voting weight to a chosen address, keeps custody and the ability to sell, and can revoke or reassign whenever they want.1 The delegate votes. The delegator does not, unless they take the weight back.
Revocability is the accountability model, and it is the whole model. A 2025 scoping review across thirteen studies frames delegation as a form of liquid democracy with reassignment available at any time, and reports that delegates do not reliably act in their delegators' interests.1 No slashing, no recall vote, no clawback. Your one lever is to leave.
The turnout numbers delegation is sold against
One study here states its method, sample and date, so use it instead of the figures circulating without any of the three. Falk, Pathan, Rigas and Tsoukalas pulled the full voting history of four DeFi DAOs from Ethereum and published in July 2024. Average share of total token supply cast per proposal: 3.2 percent at Aave across 130 proposals, 7.7 percent at Compound across 141, 5.6 percent at Lido across 149, 4.6 percent at Uniswap across 29.2
The same paper measures concentration through what it calls the minimal quorum, the smallest set of addresses that could decide a vote alone. In most votes the answer is around three.2 Put the two findings side by side and this stops looking like a participation problem delegation solves. Participation sits near 5 percent, three addresses are decisive, and delegation is how weight reaches those three.
What delegators actually reward
Bongaerts, Lambert, Liebau and Roosenboom studied who receives delegations inside the Uniswap DAO, using 84 wallets that appeared among the top five voters for at least one event between September 2020 and April 2022, across 48 proposals.3 Two results should change how you design a delegate program.
Delegation is partly a reputation market. Wallets that submitted proposals received roughly 59 times more delegated votes, and those that carried a proposal through the final on-chain vote received roughly 626 times more.3 Track record buys voting power, which is the mechanism working as intended.
Affiliation buys more. Delegates identified as affiliated with Andreessen Horowitz, 18.1 percent of the sample, received votes from others by factors between 2,689 and 3,547, and the paper also finds a negative association between a wallet's own holdings and the delegations it attracts.3 The authors read that as window dressing around decentralization by the DAO's largest venture investor. That reading is theirs. The coefficient is the finding.
We are not giving you a delegate concentration benchmark
Tables circulate showing what share of voting power the top ten delegates hold at Uniswap, Compound, Aave and Optimism. We checked the ones we could find. None state a pull date, a query or a method, and the dashboards underneath move weekly, so quoting them would be an unsourced number wearing a citation.
Your own number matters more anyway, and it is readable in an afternoon. Pull your delegate table, sort by voting weight, and count how many addresses it takes to reach a majority of votes actually cast across your last ten proposals. If the count is three, you are running a three-person board with a token-holder mailing list attached.
What to settle before delegation goes live
Three contract decisions. Whether delegation is single hop or can be passed onward, because chains move weight further from its owner. Whether voting weight is snapshotted at proposal creation or at vote time, which is a governance-attack question before it is a delegation question. And whether a party with a commercial position in your protocol can hold delegated weight without disclosing it.
Then the decision that is not a contract setting. Delegation treats a symptom. Holders do not vote because a collateral parameter change does not move their economics enough to be worth an hour of reading, and if your token's value comes from the business underneath rather than from governance rights, that is the honest explanation. No delegation design touches it. Put the decisions that need holder consent to a vote. Stop putting the rest.
Common questions
Can I take my delegated votes back?
Yes. Delegation is revocable at any time by the holder, who can pull the weight back or reassign it to a different address without asking the current delegate.1 The tokens were never transferred, so nothing has to be returned. Revocation is also the only hard accountability mechanism in these systems: there is no recall vote, no slashing and no penalty a delegator can impose on a delegate who votes badly.
Does delegating my tokens mean giving them away?
No. Delegation assigns voting weight, not ownership. You keep custody, you keep the ability to transfer or sell, and the delegate cannot move your balance. What you hand over is the right to cast the votes attached to that balance until you take it back. Selling the tokens ends the delegation along with the voting weight, because the weight follows the balance.
Why do DAOs use delegation instead of just letting holders vote?
Because direct token voting produces very low turnout. Across four large DeFi DAOs measured on chain, the average share of total token supply cast per proposal ranged from 3.2 percent to 7.7 percent.2 Delegation concentrates the reading and judgement into a smaller group who will actually do it. The tradeoff is concentration: in most votes in that same study, around three addresses could decide the outcome alone.2
See Tokenomics Design for how this applies in practice.
Sources
- Delegated voting in decentralized autonomous organizations: a scoping review
Weidener, Laredo, Kumar and Compton, Frontiers in Blockchain, vol. 8, 2025
Published 2 June 2025. Synthesises 13 studies. Source for the liquid-democracy framing, revocation at any time, and the finding that delegates do not reliably act in delegators' interests. - Blockchain Governance: An Empirical Analysis of User Engagement on DAOs
Falk, Pathan, Rigas and Tsoukalas, arXiv:2407.10945, 2024
Dated 16 July 2024. Voting history collected directly from Ethereum. Table 1 gives average percentage of total token supply cast: Aave 3.2 across 130 proposals, Compound 7.7 across 141, Lido 5.6 across 149, Uniswap 4.6 across 29. Section 5.2 defines the minimal quorum and reports it as around three in most votes. - Vote Delegation in DeFi Governance
Bongaerts, Lambert, Liebau and Roosenboom, Rotterdam School of Management, arXiv:2503.11940, 2025
Dated 13 March 2025. Sample of 84 Uniswap top-five voter wallets, September 2020 to April 2022, 48 proposals. Source for the 59x and 626x proposer multipliers, the 2,689 to 3,547 a16z affiliate range, the 18.1 percent affiliate share and the window-dressing interpretation, which is the authors' own.
Last reviewed 2026-08
Know the terms but not sure how they apply to your project? That is what an engagement is for. We design, document, and stress-test the whole token economy inside the Tokenomics Data Room.
100+ projects advised. Complete tokenomics in 4 to 6 weeks.