Vote-escrowed tokenomics

How the lock prices voting power

A holder deposits CRV into the voting escrow contract and names an unlock time one week to four years out. The veCRV credited is the amount locked scaled by the fraction of the maximum term chosen: one CRV locked for four years yields one veCRV, the same CRV for one year yields 0.25.¹ The contract stores the maximum as four years in seconds and rounds future times to whole weeks.²

Two properties are easy to skim past and expensive to discover later. veCRV cannot be transferred, so voting power has no secondary market at the escrow layer. And the lock cannot be reversed before expiry, with no early exit and no penalty option.¹² A four-year lock is a four-year liquidity decision, not a governance preference.

Decay is the part founders underestimate

Voting power decreases linearly from the moment the lock is created and reaches zero on the unlock date. The contract records a slope and a bias per user so the decay is computed rather than checkpointed.² A balance you check today is smaller tomorrow whether or not you touch it.

Design consequence: a four-year lock made on day one and a two-year lock made two years later carry the same power at the same moment. Influence is a function of remaining term, not of when you committed. To hold a steady share of the vote a participant has to keep extending, which makes the real cost of governance power a perpetual commitment rather than a single one. That is the incentive the model is built to create, and it is why participation skews hard toward parties whose business depends on the outcome.

What the lock is actually paid for

veCRV carries three payoffs. A share of trading fees and of the interest from Curve's stablecoin markets. Boosted CRV rewards on your own liquidity, up to 2.5 times. And gauge weight votes, which decide how CRV emissions are split across pools.¹³

The third is the economic engine. A gauge vote directs a stream of newly issued tokens toward a specific pool, so the vote is worth money to whoever runs that pool. Once a governance vote controls a cash flow, someone prices it. If you are copying this model, that is the sentence to sit with, because everything downstream follows from it.

The market that grew on top

Convex pools CRV from depositors, locks it as veCRV under its own contracts, and issues CVX in return. CVX can be locked in 16-week terms as vlCVX, and vlCVX holders control how Convex's aggregated veCRV position votes on Curve gauge weights. Protocols that want emissions pointed at their pool pay vlCVX holders through marketplaces such as Votium and Hidden Hand.³

Read that as a mechanism rather than a scandal. A four-year lock meant to produce long-horizon voters produced a liquid 16-week claim on those votes, held by parties with no long lock at all, and a public market renting them per round. The escrow did what it said. The layer above converted the output into something tradable, which is what layers above things tend to do.

So the honest statement of the ve tradeoff: it removes the flash-loan vote and the drive-by voter. It does not remove vote buying. It relocates vote buying to a venue with better price discovery.

What to settle before copying a ve design

Ask what the vote controls first. A ve system earns its complexity when the vote allocates real economic weight, such as an emission stream or a fee split. If your vote decides parameters nobody would pay to influence, you have added a four-year lockup, a decay curve and a non-transferable balance for very little.

Then price the lock honestly for the person taking it. Four years of illiquidity is a large ask against a token whose value depends on a two-year-old business. Locks are worth taking when the business underneath generates something the lock gives you a claim on. Where it does not, the lock is a supply-suppression device with a governance story attached, and holders work that out faster than founders expect. The token is infrastructure. The business is the engine.