Vote-escrow tokenomics – often shortened to veTokenomics – is a governance design in which a member's influence is derived from locking a protocol token for a fixed period rather than simply holding it. The longer the lock, the more voting weight the token confers, and that weight decays as the lock runs down. Pioneered by Curve Finance with its vote-escrowed CRV (veCRV) in 2020, the model became one of the most widely copied and most contested patterns in decentralized governance, spawning the multi-year "Curve wars" and an entire secondary economy of vote-buying.
In a plain token-weighted DAO, one liquid token equals one vote, and a holder can sell the instant a vote closes. Vote-escrow breaks that by asking members to commit: they lock the token into a contract for a chosen duration and receive a non-transferable balance (veCRV, veBAL, and so on) that represents their voting power. Because the lock is time-bound, the design rewards long-horizon alignment over mercenary capital – a holder who locks for the maximum term signals they intend to live with the consequences of their votes.
The trade-off is illiquidity: locked tokens cannot be sold or moved for the duration, and the ve-balance itself is soulbound to the locking account. This is the central tension of the model – it manufactures commitment at the cost of composability, and much of what followed was the market's attempt to route around that illiquidity.
Using Curve's veCRV as the canonical implementation (Curve Resources):
The result is a flywheel: emissions are valuable, veCRV directs emissions, so acquiring veCRV becomes the highest-leverage move in the ecosystem.
Because directing emissions was so valuable – and because the size of the emission itself was on no ballot – protocols began racing to accumulate veCRV – the competition that became known as the Curve wars. The decisive escalation was Convex Finance, which let users deposit CRV and capture veCRV's benefits without personally committing to a four-year lock. Convex rapidly became the single largest veCRV holder, at times controlling roughly half of all veCRV.
Convex then handed the gauge-voting decision to holders of its own token who locked it as vote-locked CVX (vlCVX) for a minimum of 16 weeks. The war moved up a layer: rather than buying CRV directly, protocols competed to control vlCVX, since each unit of vlCVX steered several veCRV worth of gauge votes. A ve model designed to lock in commitment had, in practice, grown a liquid market on top of itself for renting that commitment.
Once gauge votes had a clear cash value, third parties built dedicated vote markets to price them openly. On platforms like Votium, a protocol that wants emissions routed to its pool posts an incentive ("bribe"), and vlCVX or veCRV holders who vote as directed collect it each round. At peak cycles these vote-incentive budgets reached eight-figure sums per week.
Bribe markets are the model's most debated outcome. Defenders argue they are simply an efficient, transparent price for governance influence – emissions go to whoever values them most, and lockers are compensated. Critics counter that they complete the model's drift from "aligned long-term stewardship" to "governance as a rentable commodity," a live example of the governance-capture failure modes that recur across DAOs.
The ve model was copied widely, each variant tuning the lock terms:
weeks remaining / 104, and what the veBIO balance buys is pro-rata allocation of tokens airdropped by new DeSci launches plus priority in oversubscribed sales, not a vote on where incentives flow. More striking, the ve token does not yet vote at all: Bio's docs state that BIO and vBIO remain the governance tokens until veBIO gains sufficient adoption and a formal governance proposal is passed. It is vote-escrow adopted for its distribution properties first and its governance properties later, which is close to an inversion of the original design's stated purpose.Each design above was written or forked by the protocol that runs it. Since 2025 Aragon has sold vote-escrow as a component instead: a library on its OSx framework split into separate contracts for the lock position (an NFT), the curve that turns a lock into voting power, epoch timing, the exit queue and the connection to voting, so a protocol changes parameters rather than rewriting the escrow (Aragon). Three deployments show how far those parameters now sit from Curve's:
Celo uses the same toolkit for public-goods gauges, which direct funding to ecosystem projects rather than to liquidity (Aragon, May 2025).
Caper shares vote-escrow's goal – reward participation, not passive size – but reaches it without a lock or a tradeable weight. A caper mints a soulbound vote token each time you cast a ballot – and nothing for a trade since 11 September 2026 – and your influence is your vote weight, which grows with the vote tokens you have earned rather than with your stake alone. Because vote tokens accrue only to the account that earned them and cannot be transferred between holders, there is nothing for a Convex-style wrapper to accumulate and no vote to sell on a bribe market – the thing that made ve weight rentable simply does not exist.
The same number that sets your voting power also sets your exit redemption from the treasury, so influence and skin-in-the-game are one quantity rather than two markets. A large token bag alone still can't capture a caper: without a vote record of its own, its weight is zero, and that record cannot be bought off anyone who already holds one – only minted, one token per ranked ballot its holder casts; buying and selling mint none. See voting and what is a caper.