Delegation is the mechanism that lets a token holder assign their voting power to another address — a delegate — who then votes on their behalf. What that delegate owes in return, and what happens when they stop voting without returning the weight, is covered in delegate accountability and dormant delegates. It is the single most common answer to the defining problem of token-weighted governance: most holders never vote, so without delegation a handful of active whales decide everything and quorums fail. By concentrating dispersed, apathetic weight onto engaged representatives, delegation is what lets large DAOs reach quorum at all.

Why delegation exists

Voter turnout in major DAOs is chronically low — most governance tokens sit in wallets that never cast a vote, whether from apathy, gas costs, or the sheer effort of evaluating every proposal. Delegation lets those holders stay passive without their weight going to waste: they hand it to someone who will read the forum, follow the technical debate, and show up for every vote. Some frameworks make this mandatory — under Compound's Governor, COMP confers no voting power until it is delegated, so even holders who want to vote themselves must first self-delegate (see voting power activation and delegation thresholds for what that step does to the size of the electorate). (a16z crypto: DAO governance FAQ)

How delegation works

A holder picks a delegate and signs a delegation; from then on their weight counts toward whatever that delegate votes, at the balance they hold when each proposal snapshots. Crucially, delegation is non-custodial and revocable — the delegate never touches the tokens, and the holder can re-delegate to someone else or take their weight back at any time, overriding the delegate on a specific vote if they disagree.

Around this primitive an entire layer has grown up: delegates publish delegate statements — platforms and voting histories — and holders browse and compare them on tools like Tally and Boardroom. ENS built its DAO delegation-first, encouraging every token holder to pick a delegate at claim time.

Cosmos SDK chains invert the default entirely. On the Cosmos Hub and every chain running its x/gov module, a delegator who does not vote automatically inherits their validator's vote; voting yourself simply overrides it. Delegation there is not something a holder opts into for governance, it is a side effect of staking for yield, and it means an inactive balance is counted as a vote rather than as an abstention. That single default is why Cosmos turnout figures are not comparable to the ones quoted for opt-in delegation DAOs. The pattern is not confined to Cosmos: Orbs runs the same inversion off-chain, where a Guardian votes its entire delegated stake unless the delegator opts out for that proposal – which is how one of its ballots passed on seven signatures carrying 151.1m ORBS.

SafeDAO runs a comparable delegate system and pushes further on the incentive side: its ongoing Outcomes-Based Resource Allocation (OBRA) track pays active delegates and working groups against measurable results — an explicit attempt to keep representatives engaged rather than absentee, since an unpaid delegate is as prone to apathy as the holders who delegated to them.

Liquid democracy and its cartels

When delegates can themselves re-delegate, delegation becomes transitive — a chain of trust that ends at whoever is actually voting. This is liquid democracy: a fluid middle ground between direct democracy (everyone votes) and representative democracy (fixed officials), where any holder can be a passive voter, an active voter, or a delegate, and can switch roles at will.

The flexibility has a cost. Delegation re-concentrates power: in practice a handful of large delegates end up controlling a majority of the active vote, so a low-turnout DAO can be captured by a delegate cartel coordinating across proposals — the pattern behind on-market governance raids like the Compound 'Golden Boys' incident. Delegation fixes apathy but does not, by itself, fix plutocracy; it can sharpen it. (Fritsch et al., 'Analyzing Voting Power in Decentralized Governance')

How Caper approaches this

A caper has no member-to-delegate system: every member votes their own ranked ballot directly, and voting weight is the product of a member's stake and their earned participation, not a balance that can be assigned away. That deliberately trades the expertise and higher turnout delegation buys for immunity to delegate cartels — there is no aggregated voting bloc to capture, because there are no delegates. The one delegation-like primitive a caper ever had was the VOTE proposal — a caper deciding as a body how to vote, then casting that pre-committed ballot inside another caper. It belonged to the pre-August 2026 escrow system and did not return with the ballots: kind 2 is retired, the contract rejects it by name, and vote() takes a Global<Account> voter it checks against the caller, so a treasury holding another caper's token has no path to cast it. Organization-to-organization delegation is therefore not something a caper can do today in either direction.

References

• a16z crypto, Lightspeed Democracy: what web3 organizations can learn from the history of governance.
• Compound, Governance documentation (delegation-to-vote requirement).
• ENS, ENS DAO documentation.
• Tally, delegate directory and voting platform.
• Fritsch, Müller & Wattenhofer, Analyzing Voting Power in Decentralized Governance (arXiv, 2025).