A caper is a group of people who raise a treasury together and govern it on-chain. The exit right is the part that decides what that membership is actually worth: at any point, a member who has taken part can take their share of the treasury and go, without asking anyone. This page answers what leaving involves, in the order people ask it.

Can I leave a caper?

Yes, at any time, and nobody can stop you – provided you have taken part at least once, by trading the caper's token or by voting in it. Leaving is a single method call on the caper's contract, named exit. It is not a proposal, so it is not voted on, not scheduled, and not subject to a council or a founder's approval. There is no motion to table and no quorum to reach.

The only thing the contract checks is that you hold the tokens you claim to hold. It asserts that your vote-token amount is greater than zero, that your governance-token amount is greater than zero, and that both are that specific caper's own resources rather than some other token. Your holdings are the credential. There is no membership list to be removed from.

Those two assertions are the whole gate, and the first is the one to plan around. Vote tokens are minted by voting: one for each ranked ballot you cast, and that is the only source a member has. A holder whose position arrived any other way – bought on the curve, swapped into, transferred in – holds none, and the call aborts on Exit requires vote tokens rather than paying out zero. Plan around it: if you intend to redeem a treasury share, cast a ballot while a proposal is open, because there is no way to acquire v after the fact except by voting. Trading minted v until 11 September 2026, when the redeploy removed the trade-side mint so that one transaction could not buy a position and its exit claim together. Where they come from is covered below.

What do I get when I leave?

Two separate payouts, and it is worth keeping them apart because they come from different places.

1. A weighted share of the treasury. Your share of the XRD the caper holds and of any of the caper's own tokens sitting in its treasury, plus a matching share of any other resource you name in the call. Those first two are a floor the contract applies whether you name them or not, because the vote burn is irreversible and an empty list must not cost you the two principal assets. Everything else, including the positions a caper holds in other capers, has to be named: a treasury's deposit surface is public, so walking an on-chain register of resources would let anyone pad the per-exit loop past the cost-unit ceiling and brick redemption for everybody. The exit the site builds for you names the full live set it reads off the Gateway.
2. The proceeds of selling your governance tokens back to the bonding curve. The contract prices the sale against the curve and takes the caper's standard trade fee, the same fee any sale on that curve pays. The sale emits an ordinary sell event on the caper's trade feed.

Both legs settle in the same transaction. You do not exit and then separately go and find a buyer.

How is my treasury share calculated?

By multiplying what you hand in by the vote tokens you have accrued. The expression is called a vote weight in the contract because it is exactly that: the same formula weighs a ranked ballot while you stay and prices your share of the treasury when you go. That identity is the point of the design – you cannot vote with more force than you could walk out the door with.

The contract computes:

share = (tokens_redeemed × votes_burned) / (vote_supply × tokens_in_circulation)

Read plainly: your share rises with how many governance tokens you redeem and with how many vote tokens you burn, measured against the caper's total vote supply and its circulating token supply. Both terms are multiplicative. Holding more tokens increases your share; so does having accumulated more vote tokens, which only voting mints (one per ballot cast). If either is zero the formula returns zero, and the contract rejects the exit before it gets that far.

The function is compute_vote_weight in contracts/common/src/lib.rs, and the logic tier calls it in three places, all in contracts/logic/src/lib.rs: exit() prices the redemption, fold_ballot_batch weighs each ballot as the crystallize crank folds it, and vote() computes the preview shown at the cast. exit() was the sole caller for three days only – the 26–28 August 2026 interregnum, when the deployed logic component carried no vote method at all – and the 29 August reinstatement brought the other two back with the ballot. Because the numerator reads the buckets you actually hand in, redeeming part of a position redeems that part's share – there is no all-or-nothing exit. See vote tokens for where the v term comes from.

Does it matter when I leave?

Yes, and this page would be misleading if it left it out. The shares do not add up to one hundred per cent. Both denominators – the caper's vote supply and its circulating supply – are read live at the instant exit runs, and every exit before yours has already burned vote tokens and retired governance tokens out of them. So two members holding identical positions do not take identical shares. Take a caper whose entire vote supply and entire circulation sit with two members in equal halves: the formula gives each of them a quarter. The first to leave redeems that quarter. The second is then measured against a supply that is only their own, so their weight is one, and they redeem all of the three quarters still there.

This is a property of the design rather than a fault in it. Nothing is stranded, the treasury cannot be over-drawn, and no member takes more than the weight they can prove. But a member deciding when to go should know that leaving first is the expensive option, and that the incentive runs the other way from the panic a redemption right is usually feared to cause. It is also the sharpest difference from the rage quit this right descends from, where a flat pro-rata slice of the treasury makes the order of departure worth nothing at all.

Where do vote tokens come from?

From one place: casting a ranked ballot. A caper's vote token is created with no initial supply and minting it is gated to the caper's own admin badge, and the only member-facing path to that mint is vote, which banks the 100 XRD vote fee, appends the ballot and mints the caster one whole token in a single call. The admin badge keeps a second path, mint_vote_to, restricted to the admin role rather than offered to members.

Trading minted it too until 11 September 2026, on every buy and every sell, at a flat rate the logic component held as vote_rate. The genesis redeploy of that date removed the trade-side mint and the field with it, so the deployed component's state now carries trade_fee, vote_fee, proposal_fee, proposal_fee_rate, collateralization_peak and the three windows, and no rate for v at all. The reason was that a buy minted its v in the same transaction that opened the position, so one transaction could acquire a stake, mint the record that prices the exit, and redeem on both — an exit right with a purchase price.

The token cannot be sent, airdropped, or bought off the account that minted it — its depositor role is locked to the caper's own component, which is what “soulbound” means here in practice. The resource's own on-ledger description says as much: soulbound stake token … drives exit redemption.

Two consequences follow, and neither is visible from the payout formula alone.

• A holder who has never voted cannot exit at all. Not “exits for nothing” — the call aborts on Exit requires vote tokens before any share is computed. A position bought on the curve, swapped into, transferred in, or brought through migration can still be sold back to the curve; only the treasury share is out of reach, and only a cast ballot opens it. There is no way to acquire v retrospectively, so a holder who intends to redeem should vote while a proposal is open rather than at the moment they want to leave.
• The mint is lenient, and can be forfeited. An account with a strict deposit rule that has not whitelisted v refuses the deposit; the contract burns the refused mint rather than aborting the call that carried it. The ballot still counts, the accrual is lost, and there is no retry call – whitelist the resource before casting.

The token has been paid for one thing or another at four different points, which is why third-party descriptions of it disagree. The pre-26-August design paid only for casting a ballot; for three days from 26 August it paid only for trading; from 29 August it paid for both at one rate; and since 11 September it pays only for the ballot again — this time because trade-side minting made the exit right purchasable, not because trading was thought unworthy. See vote tokens for the fuller account, and the FAQ for the short version.

What happens to my vote tokens?

They are burned. The contract pays your treasury share first, then destroys the vote tokens you handed over, then runs the sell leg. Vote tokens are soulbound, so they were never transferable in the first place; burning them on exit is what stops a departing member's accumulated governance weight from outliving their stake in the caper.

This is also why the treasury share is computed before the burn rather than after. The weight you leave with is the weight you had.

Is there a lock-up or a waiting period?

No. The contract's exit path contains no epoch check, no notice period and no vesting gate. There is nothing in it that defers your payout to a later block, and nothing another member can do to delay it.

Voting adds no qualification either, which was not always true. The retired escrow design required staking governance tokens into a ballot box, and a cast locked that stake until the proposal's ballots were released – so a voter's tokens were not in their account to hand to exit. The deployed vote is a plain wallet transaction with no escrow at all: your tokens stay liquid and stay yours the whole time. What keeps a voter honest instead is the crystallize crank's min-clamp, which credits each ballot min(held at cast, balance now) – selling down after casting shrinks the weight of the ballot you already cast, rather than being forbidden. Exiting mid-vote is therefore allowed and simply costs you the weight. See voting for how casting and crystallization work.

This matters more than it sounds. A great many DAO exit provisions are real on paper and unreachable in practice, because the exit itself has to be approved by the people it inconveniences, or because it runs through a multisig that can simply not sign. An exit that somebody has to authorise is a request. See how DAOs fail for the shapes that recur.

What is different about the $XRD caper?

One caper is a special case. $XRD has no bonding curve, so it has no sell leg: exiting it pays the treasury share and returns your governance tokens to you untouched, rather than selling them. Everything else about the exit is identical.

The reason is that its members hold native XRD, whose total supply cannot be read on-ledger at all: Radix creates XRD with supply tracking switched off, so the resource's total_supply() is empty on every network. The contract therefore substitutes a fixed 24 billion — the 12bn minted at genesis plus the roughly 12bn released as emissions over about forty years — for the circulating-supply term. The invariant that constant has to satisfy is that it never understates what members hold between them, which is what keeps the sum of every member's share inside the treasury; erring high only makes exits conservative. It cannot distort governance either, because the same term cancels out of the ratio a proposal is settled on.

How does this compare to a DAO rage quit?

The lineage is Moloch's rage quit, and the difference is what the right is attached to. A rage quit is usually scoped to a specific proposal and a specific window: you may leave because you object to something that just passed, during the grace period before it executes. Outside that window, there is nothing to exercise.

In a caper the exit is not attached to a proposal at all. It is a standing property of holding the tokens, available on any day, whether or not anything is being voted on and whether or not you object to it. There is no window to miss.

The essay The Exit Right works through why that difference changes how the rest of the governance behaves.

What if I want to reduce my stake without leaving?

Selling governance tokens on the curve is a separate operation from exiting, and it does not burn your vote tokens or touch your treasury claim. See trading. Exiting is the operation that settles everything at once and ends the membership.

Related

• What is a caper — the object this right belongs to
• Getting started — joining, the cashtag fee ladder, and what a caper costs to start
• Voting — the other half of the same weight formula
• FAQ — shorter answers, including divestment and upgrades